10.1007/s10406-006-0175-4.pdf - Springer LINK

133

BFriday....................... 135Saturday .................. 189Sunday..................... 235Monday .................... 261Tuesday ................... 291

Scientific Sessions(SS)

Scientific Sessions

134 C D E FBA G

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 135C D E FA G

FFFFFridaridaridaridaridayyyyy, Mar, Mar, Mar, Mar, Marccccch 3h 3h 3h 3h 3

B

Scientific Sessions

136 A C D E F G

SY 1Satellite

SymposiumDigital

transformationin radiology:

Experiences andnew frontiers

(p. 546)

SS 301bGI Tract

Imaging in GItumors(p. 163)

SS 201bAbdominal

Viscera(Solid Organs)Contrast agents

for MR liverimaging(p. 130)

EP

OS

™ - scientific exhibition !

23:00

registration

technical exhibition

ECR OpeningConcert

CC 417EmergencyRadiology

Traumatic injuries(2)

(p. 20)

SF 4aSpecial Focus

SessionCancer screening

(p. 20)

RC 402Breast

Diagnostichighlights

(p. 21)

RC 401Abdominal andGastrointestinal

Liver imaging:Characterisation

and pitfalls(p. 22)

CC 416Infection in the

Adult TodayChest(p. 23)

RC 411Neuro

Brain tumors(p. 23)

RC 408Head and Neck

Radiologicalapproach to

stage head andneck squamouscell carcinoma

(p. 24)

RC 415Vascular

Imaging andintervention in

the carotid artery(p. 25)

OpeningCeremony/

InaugurationLecture/

Presentation ofHonoraryMembers

(p. 19)

SS 308Head and Neck

Larynx, oralcavity and neck

imaging(p. 165)

SS 303Cardiac

Detection ofcoronary plaqueand thrombus

(p. 167)

SS 306Contrast Media

Vascular andcellular targets

(p. 153)

SS 310Musculo-skeletal

Soft tissues(p. 155)

SS 301aAbdominal

Viscera(Solid Organs)

Contrast-enhanced US

studies(p. 157)

SS 309aInterventional

RadiologyLimb and dialysisaccess salvage

(p. 159)

SS 304Chest

HRCT anddiffuse lung

disease(p. 161)

SS 202BreastDigital

mammography(p. 132)

SS 204ChestImage

processing anddeveloping lung

imagingtechniques

(p. 134)


RadiologyLiver tumortherapy (1)

(p. 136)

SS 208Head and NeckOrbit, salivary

and thyroidglands(p. 138)

SS 203Cardiac

Detection ofcoronary artery

stenosis(p. 140)

SA 2State-of-the-Art

SymposiumAdvances in

peripheral arterialstenting(p. 17)


Arthritis/Infection(p. 126)

SS 201aAbdominal


Abdominalimaging:Technical

refinements andclinical

applications(p. 128)


Traumatic injuries(1)

(p. 8)

RC 110Musculo-skeletal

Bone marrowdisorders as a

manifestation ofdisease

(p. 8)

SF 1Special Focus

SessionPelvic floordysfunction

(p. 9)

RC 103Cardiac

Principles andtechniques of

cardiac imaging(p. 10)


Impact ofmultislice CT onimaging of the

upper abdomen(p. 11)

RC 104Chest

Digital imagingof the chest

(p. 12)

RC 111Neuro

Imaging ofnormal anatomy

and function(p. 12)

RC 108Head and NeckSalivary glands

(p. 13)

RC 105Computer

ApplicationsTele-imaging inEurope today

(p. 14)

room A2nd level

room B2nd level

room C2nd level

room E1entr. level

room E2entr. level

room F1entr. level

room F2entr. level

room Hlower level

room Glower level

08:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:00

14:00

14:30

15:00

15:30

16:00

16:30

17:00

17:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:30

14:00

14:30

15:00

15:30

16:00

16:30

17:00

12:3012:30

13:00

18:00

18:30

19:00

18:30

19:00

18:00

17:30

08:30

08:00

07:30

07:00

07:30

08:00

07:00

13:30

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 137A C D E F G

E3 220bHow to prepare acase that will be

published inEURORAD (p. 19)

WS 124Virtual

InterventionsCarotid

"Hands-on"Workshop

WS 224Virtual

InterventionsRenal

"Hands-on"Workshop

WS 222Vertebroplasty

"Hands-on"Workshop

E3 220aIHE (Integratingthe Healthcare

Enterprise) (p. 19)

SS 209bInterventional

RadiologyRadiofrequency

ablation oftumors of nonhepatic origin

(p. 143)

SS 205Computer

ApplicationsPACS in daily

practice(p. 145)

SS 211Neuro

Cerebralischemia and

trauma(p. 147)

SS 207Genitourinary

Prostate cancer/Male pelvis

(p. 149)

SS 215Vascular

Venous imagingand trauma

(p. 151)

WS 109Interventional

RadiologyHow I doperipheral

arterialinterventions

(p. 14)

CC 119Pediatric

RadiologyUrinary tractinfection in

children(p. 15)


Adult TodaySpinocranial

infection(p. 16)

RC 107Genitourinary

Imaging ofstones(p. 16)

E3 120Foundation

Course:AbdominalRadiology

Stomach andsmall bowel:

Plain films andbarium studies

(p. 17)

WS 324Virtual

InterventionsPeripheral"Hands-on"Workshop


RadiologyLiver tumortherapy (2)

(p. 169)

SS 305Computer

ApplicationsPACS andinformation

systems(p. 171)

SS 311Neuro

Diffusion MRimaging/Brain

pathology(p. 173)

SS 307GenitourinaryProstate cancer

MR imaging(p. 175)

SS 315Vascular

Carotid imaging(p. 177)

E3 320Foundation


Large bowel:Plain film and

barium studies(p. 19)

RC 409Interventional

RadiologyVertebroplasty:

Imaging andintervention

(p. 26)

CC 419Pediatric

RadiologyMusculoskeletal

infection(p. 26)

RC 414Radiographers

PACS(p. 27)

RC 407GenitourinaryProstate cancer

(p. 28)

SF 4bSpecial Focus

SessionRadiationprotection

(p. 28)

E3 420Interactive

image teachingThoracic

Calcifications(p. 29)

WS 424Virtual

InterventionsRenal

"Hands-on"Workshop

WS 624Virtual

InterventionsRenal

"Hands-on"Workshop

room Klower level

room L/M1st level

room N/O1st level

room Plower level

room Xentr. level

room Y1st level

The Globe2nd level

08:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:00

14:00

14:30

15:00

15:30

16:00

16:30

17:00

17:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:30

14:00

14:30

15:00

15:30

16:00

16:30

17:00

12:3012:30

13:00

18:00

18:30

19:00

18:30

19:00

18:00

17:30

13:30

08:30

08:00

07:30

07:00

07:30

08:00

07:00

room Ilower level

room Wbasement

room Z1st level

Scientific Sessions

138 C D E FBA G

10:30 - 12:00 Room B

Musculoskeletal

SS 210Whole body MRI and tumor imagingModerators:W. Jan; London/UKA.R. van Erkel; Leiden/NL

B-001 10:30

Bone marrow improvement in patients with type 1 Gaucher disease onenzyme replacement therapyW.L. Simpson, G. Hermann, S.M. Farkas, E. Kim-Melia; New York, NY/US([email protected])

Purpose: To demonstrate improvement in bone marrow infiltration in patientswith Type 1 Gaucher disease (GD) on enzyme replacement therapy (ERT).Methods and Materials: The Comprehensive Gaucher Treatment Program in-cludes over 200 patients with GD, of which 80 are receiving ERT. All patients areevaluated for the extent of bone marrow involvement of the lower extremitiesusing MR imaging, including T1, T2, chemical shift GRE and STIR sequences.Decreased signal intensity (SI) on T1 weighted images (WI) and increased SI onSTIR WI indicate marrow infiltration and disease activity respectively. Correlationis made with the genetic mutation expressed in this group of patients as well asthe length of time on ERT.Results: Ten patients (age range 30 to 63 years, mean age 41, 4 males, 6 fe-males) have demonstrable improvement in the bone marrow on serial studiesperformed at yearly intervals for 1 to 6 years (mean 2.7 years). These patientsdemonstrated a decrease in the amount of marrow with low SI on T1 WI involvingthe lower extremities. There is no correlation between the genetic mutation andmarrow improvement. There is a qualitative correlation between the time on ERTand the marrow improvement with greater improvement in those patients on ERTfor a longer period.Conclusion: ERT leads to improvement in bone marrow infiltration in adult pa-tients with GD. The amount of improvement is related to the length of treatment.There does not appear to be correlation between improvement or degree of im-provement and the genetic mutation.

B-002 10:39

Whole-body MR imaging using a single slab 3D T2 weighted turbo-spin-echo sequence with high sampling efficiency (SPACE) for high spatialresolution imagingM.P. Lichy1, J.P. Mugler III2, W. Horger3, M. Menzel3, B. Kiefer3, F. Schick1,C.D. Claussen1, H.-P. Schlemmer1; 1Tuebingen/DE, 2Charlottesville, VA/US,3Erlangen/DE ([email protected])

Purpose: Clinical evaluation of single-slab 3D T2-w TSE with high sampling effi-ciency (SPACE) for high isotropic body imaging with large FoV.Methods and Materials: A total of 50 patients were examined in clinical routinewith SPACE (regions of interest: pelvis n = 30, spine n = 18, extremities n = 4). A1.5 T MR scanner (Siemens Magnetom Avanto) was used. For achieving highsampling efficiency, parallel imaging with an acceleration factor of 3, high turbo-factor of 71 and magnetization restore pulses were used. Sequence parameterswere: FoV = 380x380 mm2, spatial resolution = 1x1x1 mm3, measurement timewas approximately 10 minutes (144 slices per slab). In contrast to a conventionalTSE imaging technique, the refocusing pulse train of the SPACE sequence con-sists of variable flip angle radio frequency pulses along the echo train. Imageswere compared with data acquired with conventional 2D T2-w TSE by evaluatingSNR and CNR impression, presence of artifacts and diagnostic quality. Addition-ally, clinical value of interactive multiplanar image reformations for image readingand demonstration was evaluated.Results: SPACE images were of sufficient diagnostic value in all cases. No pro-nounced artifacts were observed. The possibility of image reconstruction in mul-tiple planes was of clinical relevance in all cases and simplified data analysis.Conclusion: The achievement of single slab 3D T2-w TSE with isotropic voxels,high spatial resolution and the usage of interactive 3D visualization software forimage reading improve the diagnostic potential of MR imaging of the spine andthe pelvis.

B-003 10:48

Whole body imaging approach of patients with multiple myeloma:Comparing MR imaging with MDCTS. Buhmann, S. Schönberg, C.R. Becker, N. Lang, M.F. Reiser,A. Baur-Melnyk; Munich/DE ([email protected])

Purpose: The detection rate of bony involvement in whole-body multi-detectorCT (MDCT) was compared to whole-body MR imaging in patients suffering frommultiple myeloma.Methods and Material: Thirty patients with histologically confirmed multiplemyeloma were prospectively examined with whole-body, low dose MDCT andwhole-body MR imaging on a 1.5 Tesla system (Symphony, Siemens MedicalSolutions, Erlangen, Germany). The protocol consisted of T1-weighted SE andSTIR sequences. For data analysis, the entire skeleton was divided into 104 re-gions/patient. Images were evaluated in a consensus reading by two radiolo-gists, blinded to patients' history, with separate evaluation of each modality. Twoexperienced radiologists given histological, clinical and follow-up patient infor-mation evaluated the combination of MDCT and MR-images defining the goldstandard.Results: 722/3120 regions were affected. 239/722 regions were detected in MDCTand 715/722 lesions in MR imaging. 6 false positive results were found in MDCT,4 in MR imaging. Thus, the diagnostic accuracy for MR imaging was 99.6 % (sen-sitivity: 99.4 %, specificity: 99.7 %) whereas for MDCT 84.3 % (sensitivity: 33.6 %,specificity: 99.7 %).Conclusion: Although MDCT allows high resolution imaging of bony structures,MR imaging resulted in higher detection rates probably due to a direct visualisa-tion of the bone marrow before visually detectable osteoclastic lesions occur.

B-004 10:57

Whole-body MR imaging on a 32-channel system versus dual-modality PET-CT for the diagnosis of metastatic bone diseaseG.P. Schmidt, S.O. Schoenberg, P. Herzog, R. Schmid, R. Stahl, R. Tiling,C.R. Becker, M.F. Reiser, A. Baur-Melnyk; Munich/DE([email protected])

Purpose: To evaluate the diagnostic accuracy of bone metastases screeningusing whole-body MR imaging (WB-MRI) compared to dual-modality PET-CT.Methods and Materials: In a prospective blinded study, 32 patients (19 female/13 male, average 56 years) with different oncological diseases (breast cancern = 13, gastrointestinal cancer n = 7, cancer of unknown primary n = 6, othern = 6) and suspected skeletal metastases underwent FDG-PET-CT and WB-MRIwith the use of parallel imaging (PAT). T1-weighted and STIR imaging of the en-tire body and spine was performed. PET-CT was performed with a diagnosticcontrast-enhanced multislice-CT-scan covering the thorax, abdomen and pelvis.Two experienced radiologists read the MR imaging-scans, another radiologist incombination with a nuclear scientist read the PET-CT scans, each in consensus.The gold standard was constituted by histology and clinical-radiological follow-upwithin 6 months.Results: In 29 patients 102 malignant and 25 benign bone lesions were con-firmed. WB-MRI showed a sensitivity of 94% (96/102) and specificity of 76% (19/25), PET-CT resulted in a sensitivity of 77% (78/102) and specificity of 80% (20/25). Diagnostic accuracy was 91% (115/127) for WB-MRI and 77% for PET-CT(98/127). Cut-off size for the detection of malignant bone lesions was 2 mm forWB-MRI and 5 mm for PET-CT. WB-MRI revealed 10 additional bone metastasesdue to the larger field of view. Total examination time was 39 minutes for WB-MRIand 103 minutes for PET-CT.Conclusion: WB-MRI and PET-CT are reliable imaging modalities for systemicscreening for metastatic bone disease. PAT allows whole-body MR imaging bonemarrow screening with a diagnostic accuracy superior to PET-CT.

B-005 11:06

Whole body MR imaging for detecting metastatic bone tumor: Evaluatingthe usefulness of diffusion-weighted body imaging (DWIBS)K. Nakanishi1, M. Kobayashi1, M. Kyakuno1, K. Nakaguchi1, T. Murakami2,H. Nakamura2; 1Osaka/JP, 2Suita/JP ([email protected])

Purpose: To assess the usefulness of whole body MR imaging (WB-MRI), and inparticular diffusion weighted body scanning (DWIBS), for detecting bone metas-tasis.Materials and Methods: WB-MRI was performed on 13 patients for detectingbone metastasis (8 prostate, 3 breast, 2 lung cancers). In all these cases, bonescintigram (BS) was performed. At first, total spine was imaged on a sagittalplane using two image sets (fast T1WI and STIR). Then, the whole body was

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 139C D E FA G

imaged on the coronal plane using two image sets (fast T1WI and STIR). Finally,DWIBS (Single shot STIR, TR 6243, TE 59, TI 180 msec, b value = 600, 5 mmslice thickness, 112 x 112 matrix) was performed in the axial plain from lowerneck to proximal femur. When abnormal findings were detected, they were scoredas follows (Grade 1=benign, 2=probably benign, 3=equivocal, 4=probably malig-nant, 5=malignant).Results: Five cases of 6 lesions were diagnosed as benign (grade 1 and 2) onWB-MRI, and 2 cases were diagnosed as not having bony lesions. On BS, nocases were diagnosed as benign, 4 cases of 4 lesions were equivocal (grade 3)and 3 cases were diagnosed as not having bony lesions. On the other hand, 6cases of 28 lesions were diagnosed as malignant (grade 4 and 5) on WB-MRIand these lesions were in concordance with BS. In four of these 6 cases, multiplerib lesions were well depicted in DWIBS, but neither on T1WI nor STIR.Conclusion: WB-MRI including DWIBS was an excellent method both for detect-ing bone metastasis and benign lesion.

B-006 11:15

Magnetic resonance imaging of pelvic and femoral bones for detection ofbone marrow infiltration in patients with non-Hodgkin's lymphomaF.A.K.A. Al-Mulhim; Al-Khobar/SA ([email protected])

Purpose: This study was conducted to determine the advantage of femoral mar-row MR imaging as a non-invasive technique over bilateral iliac crest BM biop-sies in detecting BM infiltration before treatment and residual disease aftercompletion of treatment.Methods and Materials: Over a period of 27 months from January 2002 to March2004, a total of 30 patients with newly diagnosed NHL were included in the study.Of these, 25 were of B-cell type and 5 were of T-cell type. Coronal MR imaging ofthe pelvis and femoral marrow were obtained by T1, T2 SE sequences and STIRtechnique.Results: MR imaging showed BM infiltration in 17 cases (56.7%) before treat-ment and positive biopsy results were found in 9 cases (30%); all had abnormalMR imaging. There was a significant difference between both methods in detec-tion of infiltration (p = 0.037). MR imaging showed that 58.8% of cases have scat-tered pattern, 23.5% were uniform and 17.6% of cases have nodular patterns.MR imaging after completion of treatment showed residual BM infiltration in 6 outof 17 cases who previously had positive MR imaging and only one of these had apositive BM biopsy, with a significant difference between both methods, and allpatients relapsed within 6 months (p = 0.034).Conclusion: MR imaging of the pelvis and femoral marrow were superior to BMbiopsy on detection of BM infiltration before treatment and residual infiltrationafter treatment.

B-007 11:24

MR imaging of bone marrow edema: Analysis of bone marrow edemapattern with histological correlationS.L.J. James, R.J. Hughes, K.E. Ali, A. Saifuddin; Stanmore/UK([email protected])

Purpose: To identify the relationship between bone marrow edema on MR imag-ing and histological diagnosis in a range of benign and malignant neoplastic andnon-neoplastic conditions.Methods and Materials: Ethics review board approval and informed patient con-sent was not required for this study involving retrospective image review. Of 1998patients included on our Orthopaedic Oncology database between September1998 and March 2005, 392 demonstrated bone marrow edema and were includ-ed in the study. There were 227 males and 165 females, with patient age rangingfrom 1-87 years (Mean 29 years). MR images were retrospectively reviewed andassessed for the presence and extent of bone marrow edema. The amount ofedema was graded: Grade 1 - edema present but smaller than the lesion size;Grade 2 - edema equivalent to the lesion size; Grade 3 - edema greater than thelesion size. Histological correlation of the nature of the lesion was obtained in allcases.Results: Where bone marrow edema was identified, there were 192 Grade 1lesions of which 56% were malignant, 32% were benign and 12% were non-neoplastic; 74 Grade 2 lesions of which 19% were malignant, 50% were benignand 31% were non-neoplastic and 126 Grade 3 lesions of which 10% were ma-lignant, 45% were benign and 45% were non-neoplastic. There was a significantrelationship between edema grade and histological diagnosis (p < 0.0005).Conclusion: As the amount of bone marrow edema increases relative to the sizeof the underlying lesion, it is increasingly likely the cause is benign.

B-008 11:33

MR imaging of sacro-coccygeal tumoursT. Muthukumar, R. Guirguis, A. Saifuddin; London/UK([email protected])

Purpose: Evaluation of sacro-coccygeal tumours on MR imaging to aid in forma-tion of presumptive diagnosis and to identify features of malignancy.Methods and Materials: Analysis was made of MR imaging of 52 consecutivepatients with histologically proven sacro-coccygeal tumours from the RNOH on-cology database over the last 10 years. The following parameters were assessed:size of lesion, margin, central or eccentric position, sacro-iliac joint involvement,extra-osseous extension of tumour, signal characteristics on T2-W sequence,neuroforaminal involvement or any other relevant information.Results: Forty-five lesions originated from the sacrum and 7 lesions from thecoccyx. Forty-one lesions were malignant [22 chordomas, 7 metastasis, 5 chon-drosarcoma, 3 Ewing's sarcoma, 1 each of osteosarcoma, osteoblastoma, spin-dle cell sarcoma and MPNST] and 11 were benign [2 each of GCT, schwannoma,haemangioma, neurofibroma and 1 each of ABC, myxopapillary ependymoma,and arachnoid cyst]. The tumours are usually large at time of presentation withpredominant central location of origin. Tumours are well marginated, with the ex-ception of osteosarcoma. Twenty-one malignant and 6 benign lesions were lobu-lated. Twelve lesions crossed the SI joint, including a benign neurofibroma. Thesignal on T2-W sequences was generally heterogeneous and hyperintense, withthe exception of 1 case each of osteosarcoma and metastasis which returnedlow signal. Four benign lesions invaded the neural foramina [2 GCTs, 1 Schwan-noma and 1 neurofibroma]. Gluteal muscle invasion seemed unique to chordo-mas.Conclusion: Certain characteristic features assist in providing a more confidentradiological diagnosis. Features of aggression are well demonstrated, with impli-cations for management.

B-009 11:42

Osteoid osteoma in young childrenA. Zagarella1, F. Hornicek2, K. Raskin2, D. Rosenthal2; 1Milan/IT, 2Boston,MA/US ([email protected])

Purpose: To determine whether the appearance of osteoid osteoma or the rateof recurrence in young children differed from the general patient population.Methods and Materials: Informed consent was not required to include patientdata in this study. Patients less than 6 years of age were identified from compu-terized records of individuals referred for radiofrequency treatment. Lesion size,shape and location was evaluated from CT scans, and compared to a consecu-tive series of 131 patients over six years of age. Recurrence rate was determinedfrom office records after a minimum 2-year follow-up, using a previously pub-lished classification.Results: Out of 503 patients in the database, 23 (4.6%) were less than 6 yearsold at the time of tumor occurrence. Sex distribution, and affected bones were thesame as in older patients, but there was a striking absence of diaphyseal lesions,especially in the tibia. Tumors were often elongated, with the mean long axismeasuring 12 mm, while the mean transverse axis was only 5 mm. Long axes inexcess of 1.5 cm was demonstrated in 5/23 cases. The recurrence rate (20%)was higher than generally expected for osteoid osteoma, but with the numbersavailable the difference did not reach statistical significance.Conclusion: Osteoid osteomas may be markedly elongated in young children.Elongation of the lesion may contribute to a somewhat higher risk of tumor recur-rence, although confirmation would require larger numbers of patients.

B-010 11:51

Repeat evaluation of requests for MR imaging performed during two weeksin 2004 in Stockholm, Sweden. Comparison with the previous evaluation in2001B. Isberg, H. Jorulf, O. Flodmark, L. Zachrisson, Y. Palmquist, U. Svahn;Stockholm/SE ([email protected])

Purpose: To compare evaluations of requests for MR imaging in 2001 and 2004,adequacy of request information and appropriateness of MR imaging examina-tions.Methods and Materials: All requests for examinations performed during twoweeks in 2004 (n = 3482) in the County of Stockholm were reviewed. The infor-mation on each request form was evaluated by two to four experts; in all 10 ex-perts: two neurologists, two orthopaedic surgeons, two spine surgeons, twoneuroradiologists and two general radiologists. These were asked to evaluate ifthe information in the requests were adequate, if the indications for study were

Scientific Sessions

140 C D E FBA G

appropriate and if the choice of modality was correct. Four experts who enrolledin the previous study in 2001 also participated in 2004. The results were com-pared to those of requests for 3205 MR imaging examinations in 2001.Results: In 2004 compared to 2001, only a moderately higher proportion of re-quests were considered appropriate or probably appropriate (2004: 71%, 2001:64%). This proportion increased most for out-of-hospital practising physicians.The proportion was higher for specialists in hospital practice. In total, there wasonly subtle improvement of adequacy of information given in the requests of 2004compared to 2001. The experts disagreed whether MR imaging was the correctchoice of modality. The radiologists were more in favour of MR imaging than theclinicians.Conclusion: The results of this study support the need for further education andprecise guidelines in the proper utilisation of MR imaging as a diagnostic tool.The degree to which the use of diagnostic tests complies with guidelines mayrepeatedly be followed using the present study design.

10:30 - 12:00 Room C

Abdominal Viscera (Solid Organs)

SS 201aPancreatic disordersModerators:L. Cambj Sapunar; Split/HRL. Lincender; Sarajevo/BA

B-011 10:30

Echo-enhanced sonography: A new valid initial imaging approach for acutesevere pancreatitisS. Rickes, C. Uhle, S. Kahl, S. Kolfenbach, K. Mönkemüller, O. Effenberger,P. Malfertheiner; Magdeburg/DE ([email protected])

Purpose: This prospective study aimed to compare the accuracy of echo-en-hanced ultrasound (US) with computed tomography (CT) in assessing acute pan-creatitis and to explore the correlation between US findings and clinical outcome.Methods: Thirty-one patients with acute pancreatitis were investigated by CTand echo-enhanced US within 72 hours after admission. Echo-enhanced US (withinjection of 2.4 mL SonoVue®, pulse inversion, MI 0.1, Siemens Elegra) wasperformed following conventional US. Balthazar's grading system was used tomeasure CT and US severity indices (CTSI and USSI). Correlations betweenCTSI and USSI and between USSI and clinical parameters were tested by Spear-man's rank correlation coefficient.Results: A strong correlation was demonstrated between CTSI and USSI(r = 0.807, P < 0.01). US correlated with the following: the Ranson score (r = 0.401,P < 0.05), C-reactive protein levels 48 hours after admission (r = 0.536, P < 0.01),duration of hospitalization (r = 0.422, P < 0.05), and clinical outcome regardingmorbidity, including local and systemic complications (r = 0.363, P < 0.05). Basedon CT findings as the gold standard, sensitivity, specificity, positive predictivevalue, and negative predictive value of US for detecting severe acute pancreatitisbased on imaging criteria (Balthazar score D or E and/or presence of hypoper-fusion compatible with necrosis and/or SI ≥ 3) were, respectively, 82% (95% CI:61-93), 89% (95% CI: 57-98), 95% (95% CI: 75-99), and 67% (95% CI: 39-86).Conclusion: Echo-enhanced US produces excellent results in the staging of acutepancreatitis severity. The procedure is cheaper and has fewer contraindicationsthan CT.

B-012 10:39 !Early CT-scan is valuable in predicting outcome acute pancreatitisL.J. Delrue, J.J. De Waele, P. Duyck; Ghent/BE ([email protected])

Purpose: Early assessment of severity and prognosis is very difficult in acutepancreatitis (AP). The value of conventional clinical, biochemical and radiologicalparameters is controversial within the first 24h after onset of disease. We wantedto investigate whether signs of systemic inflammation on CT within the first 24h,rather than pancreatic abnormalities can predict outcome of AP.Methods and Materials: Forty patients with AP received an early CT (ECT) within24 hours after admission. An experienced radiologist, blinded for the outcomeparameters, reviewed all data sets. The outcome parameters were defined assevere AP (local complication or presence of organ failure for more than 48 hours)or in-hospital mortality. For each patient, we calculated an ECT-score based onpresence of pleural effusion, ascites and retroperitoneal fluid collections. Thisscore was evaluated by calculating ROC curves and the area under the curve(AUC) for predicting severe AP and mortality.

Results: Mean age of study group was 50 (±17.7) years. Fourteen patients (35%)developed severe AP. In-hospital mortality was 15% (6/40). The mean ECT-scorewas 3.6 (± 2.0). The AUC for predicting severe AP and mortality was 0.91 (95%CI 0.83 to 0.99) and 0.85 (95% CI 0.71 to 0.99) respectively. An ECT score of 4 ormore had a 100% sensitivity and 70.8% specificity for severe AP.Conclusion: The ECT score, which reflects extrapancreatic inflammation, canbe obtained within 24 hours of admission and allows early and accurate estima-tion of severity and mortality in patients with AP.

B-013 10:48

Multidetector CT in evaluation of the retroportal pancreatic lamina:Correlation with histopathological findingsC. Cappelli, S. Mazzeo, A. Pratali, G. Caproni, A. Belcari, A. Giannini,D. Campani, M. Del Chiaro, C. Bartolozzi; Pisa/IT ([email protected])

Purpose: To evaluate the role of multidetector CT in the assessment of the retro-portal pancreatic lamina by correlating CT findings with histopathological results.Methods and Materials: We evaluated 32 patients affected by ductal adenocar-cinomas of the pancreatic head or uncinate process, submitted to surgical resec-tion. The suspicion of infiltration (macroinfiltration or microinfiltration) of theretroportal lamina arose at CT when the fatty layer localized between the medialsurface of the pancreatic head and uncinate process and the mesenteric arteryappeared obliterated, irregular or with abnormal density.Results: CT suggested the infiltration of the retroportal lamina in 13/32 cases (4microinfiltration, 9 macroinfiltration) and the absence of infiltration in 19/32 cas-es. At histopathology the presence of infiltration was confirmed in 11/13 (85%)cases and the absence of infiltration was demonstrated in 17/19 (89%) cases.Two false positive cases were caused by inflammatory findings of the peripan-creatic fatty tissue. In all CT cases of microinfiltration the retroperitoneal resec-tion margin was not infiltrated, while all cases (4) with infiltration of theretroperitoneal margin were macroinfiltrated at CT. The sensitivity of CT was 85%,specificity of 89% with an overall diagnostic accuracy of 88%.Conclusion: The neoplastic infiltration of the retroportal fatty tissue by adeno-carcinomas of the pancreatic head or uncinate process represents a critical fac-tor in the tumor staging and in the surgical planning because it frequentlyrepresents a site of persistent and recurrence of disease. Multidetector CT isaccurate in the assessment of the neoplastic infiltration of the retroportal fattytissue.

B-014 10:57

MDCT and neuroendocrine pancreatico-duodenal tumors: Correlation withsurgical findingsE.D. Rappeport, C. Palnaes Hansen, A. Kjaer, U. Knigge; Copenhagen/DK([email protected])

Purpose: To investigate the accuracy of dedicated pancreatic MDCT in the diag-nosis of neuroendocrine pancreaticoduodenal tumors (NPT).Methods and Materials: MDCT and other imaging studies in patients with sus-pected NPT were identified. Imaging reports were reviewed and findings com-pared with surgical findings. Only patients who had surgical verification wereincluded. Fourteen patients (sixteen operations) with dedicated MDCT and sub-sequent surgery were included in the study. Eleven operations were for extra-pancreatic gastrinoma and five for pancreatic NPT.Results: Lesion-by-lesion-analysis: 19 NPT were identified at surgery. MDCTprospectively identified 16 and somatostatin receptor scintigraphy (SRS) 11 outof 19 tumors- sensitivity 84.2% and 57.9% respectively. Extra-pancreatic gastri-nomas: MDCT prospectively identified 13 and SRS nine out of 16 gastrinomas-sensitivity 81.3% and 56.3% respectively. Endoscopic ultrasound (EUS) was donebefore nine of eleven operations in this patient-group- EUS detected ten out of 13gastrinomas- sensitivity 76.9%. Patient-by-patient-analysis: Sensitivity for dem-onstrating presence of NPT in a patient was 100% for MDCT and 69.2% for SRS.Accuracy was 81.3% for MDCT and 66.7% for SRS.Conclusion: Dedicated MDCT of the pancreas provides excellent anatomicaldetails with high detection rates for NPT and also for duodenal tumors. Detectionrates for NPT by CT are better than stated in the older CT-literature.

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 141C D E FA G

B-015 11:06

Multi-detector row CT of the pancreas: Effects of contrast material flowrates and individualized scan delay on enhancement of the pancreas andtumor contrastG. Schueller1, W. Schima1, C. Schueller-Weidekamm1, M. Weber1,R. Prokesch2; 1Vienna/AT, 2Zell am See/AT ([email protected])

Purpose: Acquisition times in pancreatic CT imaging are considerably decreasedwith MDCT, affecting the time window for optimal pancreatic phase scanning.This study aims to optimize the pancreatic enhancement and the tumor-to-pan-creas contrast in 16-row MDCT by varying the contrast material flow rates, and toinvestigate the impact of individualized versus fixed delay imaging.Methods and Materials: Forty patients (21 females, 19 males) referred for MDCTbecause of suspected pancreatic tumor were randomized to receive 150 ml ofnonionic contrast (300 mg/ml) at 4 ml/s (n = 21) or 8 ml/s flow rate (n = 19). Pa-tients were dynamically scanned at a single level every 2 s for 66 s post IV ad-ministration. Contrast enhancement of pancreas and tumors was measured bycircular regions of interest.Results: At 8 ml/s flow rate, peak contrast enhancement in the pancreas wasobserved significantly earlier (28.9±3.4 s vs. 48.6±3.5 s, P < 0.05) and was sig-nificantly higher (119.9±29.2 HU vs. 94.5±39.1 HU, P < 0.05). Tumor-to-pancre-as contrast > 40 HU was achieved significantly longer in the 8 ml/s flow rate group(26.4±11.9 s vs. 8.6±8.3 s., P < 0.05). At 8 ml/s flow rate, an individualized scandelay of 19 s post aortic transit time revealed significantly higher tumor-to-pan-creas contrast than fixed scan delay imaging, and tumor conspicuity was better.Conclusion: In 16-row MDCT scanning, an increased contrast material flow rateof 8 ml/s and an individualized scan delay improve pancreatic enhancement andtumor-to-pancreas contrast compared to flow rates of 4 ml/s and fixed scan delayimaging.

B-016 11:15

Quantitative secretin MRCP: The definitive, non-invasive, pancreaticexocrine function test. Results in 216 patients with known or suspectedpancreatic pathologyA.R. Gillams, W. Lees; London/UK ([email protected])

Purpose: We have previously described quantification of pancreatic exocrinefunction using sequential secretin MRCP data sets. We present our results in 216patients with known or suspected pancreatic pathology.Methods and Materials: 170 patients were studied on a Siemens Vision 1.5 Tand 46 patients on a Symphony 1.5 T. A system dependent correction factor wasderived from phantom experiments with known volumes of water. A multi-slicessFSE MRCP sequence was performed before and after 0.1 ml/kg IV Secretin(Secrelux, Sanochemia). Change in small intestinal water volume was plottedagainst time and the flow rate derived from the gradient. Patients were classifiedby clinical history, MR parenchymal appearances and where available CT (150)and ERCP (56) findings. Statistical analysis was performed with the Student's t-test with significance at p < 0.05.Results: 76 patients had chronic pancreatitis, 26 were post surgical, 27 werenormal, 34 post acute pancreatitis, 6 atrophic, 8 obstructive, 15 divisum, 10 sphinterof Oddi dysfunction and 14 miscellaneous. There were significant differencesbetween normal patients (mean ± SD: 7.3 mls/min ± 2.9), those with mild(7.5 ± 2.5) or moderate (7.0 ± 3.0) chronic pancreatitis and those with severechronic pancreatitis (5.3 ± 2.4). There was a trend down with worsening degreesof pancreatitis. There was a significant difference between normals and thosewith pancreatic atrophy (3.8 ± 3.1) or obstructive pancreatopathy (5.5 ± 2.9).Conclusion: Secretin MRCP is the definitive pancreatic test providing ductal,parenchymal and quantitative functional information, i.e. complete assessmentof the pancreas, in a single investigation.

B-017 11:24

Intraductal papillary mucinous tumors of the pancreas: Evaluation withsecretin-stimulated MR pancreatographyF. Donati, P. Boraschi, R. Gigoni, S. Salemi, C. Bertucci, M.C. Cossu,F. Falaschi, C. Bartolozzi; Pisa/IT ([email protected])

Purpose: To assess the usefulness of Secretin-stimulated MR Pancreatography(SSMRP) for evaluating intraductal papillary mucinous tumors (IPMT).Methods and Materials: We retrospectively analysed thirty-two patients with di-agnosis of IPMT based on surgery or on a combination of diagnostic imagingfindings including ERCP, endoscopic ultrasonography, CT and MRCP. All thesepatients underwent MR imaging at 1.5 T (Signa Infinity, GE Healthcare). A super-paramagnetic suspension (Lumirem®, Guerbet) was orally administered a few

minutes before MR imaging. After acquisition of axial T1w and T2w sequenceswith and without fat suppression, MR pancreatography was performed using acoronal breath-hold, thick-slab (40-60 mm), SSFSE T2w sequence, repeated every30 seconds up to 15 minutes after intravenous administration of Secretin (Secre-lux®, Sanochemia). The type of lesion (main duct type, branch duct type, com-bined type), the changes of pancreatic ductal system, and the filling of duodenumafter Secretin-stimulation (graded on Matos' classification) were evaluated bytwo observers in conference.Results: MR imaging showed twenty-four cases of branch duct type (11 monofo-cal, 13 multifocal), four of main duct type, and four of combined type. After Secre-tin administration, in twenty-one patients a normal duodenal filling (grade 3) wasobserved, while in eleven patients (4 cases of main duct type, 4 of combinedtype, and 3 of markedly diffuse branch duct type) the pancreatic fluid outflowappeared reduced, limited into the duodenal bulb (grade 1) or partially filling theduodenum up to the genu inferior (grade 2).Conclusion: Secretin-stimulation increases the diagnostic value of MR pancrea-tography in both morphological and functional evaluation of IPMT.

B-018 11:33

Secretin enhanced MRCP in the evaluation of radiologically benign cysticpancreatic tumorsG.M. Carbognin, V. Girardi, L. Pinali, R. Salvia, R. Pozzi Mucelli; Verona/IT([email protected])

Purpose: To evaluate the utility of secretin administration in MRCP of radiologi-cally benign cystic pancreatic tumors.Methods and Materials: 91 patients with cystic pancreatic tumors (first diagno-sis) were studied with MRCP. 57 of them showed radiological findings suggestingnon-aggressive biological behavior: there were 32 SCA, 44 branch duct IPMTs,10 MCT, and 5 solid pseudopapillary tumors. All the diagnosis were verified byintervention or follow-up. When multiple lesions were present only the biggestwere considered. 34 patients were males and 57 females (mean age 56). In 24patients Secretin was administered in order to evaluate the relationships betweenthe lesion and the main pancreatic duct (MPD). Secretin was administered at adose of 1 IU per kilogram of body weight following which a total of 15 imageswere acquired.Results: The lesions ranged 8-200 mm (mean 33.3 mm). There were 41 lesionsin the head, 14 in the neck, 24 in the body and 12 in the tail. Secretin enhancedMRCP (S-MRCP) allowed precise definition of the relationships of the lesion withthe MPD in all patients. In 4 patients S-MRCP demonstrated changes of chronicpancreatitis, advocating surgical intervention.Conclusion: S-MRCP improves the definition of the relationships of the cysticlesion with the MPD. It is also useful in the demonstration of early changes ofchronic pancreatitis, whenever the lesion obstructs the MPD. S-MRCP thereforeis the imaging modality of choice in the first diagnosis and follow-up of radiolog-ically benign cystic pancreatic lesions.

B-019 11:42

Navigator-triggered isotropic three-dimensional (3D) magnetic resonancecholangiopancreatography (MRCP) in the diagnosis of malignantpancreatobiliary obstruction: Comparison with endoscopic retrogradecholangiography (ERCP) and/or PTCJ.-Y. Choi, J. Lee, J. Lee, S. Kim, K. Kim, S. Kim, S. An, J. Han, B. Choi;Seoul/KR ([email protected])

Purpose: To evaluate the accuracy of navigator-triggered isotropic three-dimen-sional MRCP using turbo spin echo for the malignant pancreatobiliary obstruc-tion and to compare ERCP or PTC.Methods and Materials: MRCP findings were evaluated in 23 patients with ma-lignant pancreatobiliary obstruction. Two observers independently evaluated 3D-MRCP and ERCP/PTC in a blinded fashion and recorded the level of obstructionand extent of tumor along with their confidence. Quality of images and morpho-logic appearance of stricture were also evaluated. The results of MRCP and ERCP/PTC were compared with surgical and histopathological data.Results: 3D-MRCP was of diagnostic quality without artifact in all patients, where-as ERCP/PTC examinations were non-diagnostic in 4 patients. Image qualitiesof 3D-MRCP were superior to ERCP/PTC (p = 0.009). For the evaluation of ob-struction level, average accuracy were 80.4 % for 3D-MRCP and 68.4 % for ERCP/PTC. For the evaluation of tumoral extent in hepatic hilum and distal common bileduct, 3D-MRCP was comparable with ERCP/PTC. 3D-MRCP was superior toERCP/PTC in assessment of extent of hilar bile duct involvement (Az = 0.925,0.870 for 3D-MRCP and 0.703, 0.627 for ERCP/PTC). Average sensitivity, spe-cificity, and accuracy for the evaluation of tumor extent were 85.0%, 82.4%, 82.4%

Scientific Sessions

142 C D E FBA G

for MRCP, and were 68.5%, 79.2%, 70.0% for ERCP/PTC. There was good agree-ment between morphologic appearance at MRCP and those at ERCP/PTC (kap-pa 0.684).Conclusion: 3D-MRCP can accurately assess the level of obstruction and ex-tent of tumor in patients with malignant pancreatobiliary obstruction. This MRCPtechnique has high resolution and good image quality.

B-020 11:51

Diffusion-weighted MR imaging in the differential diagnosis of focalpancreatic lesions: Work-in-progressF. Donati, P. Boraschi, R. Gigoni, C. Bertucci, S. Salemi, S. Spallanzani,F. Falaschi, C. Bartolozzi; Pisa/IT ([email protected])

Purpose: To evaluate the usefulness of diffusion-weighted MR imaging in provid-ing an objective value for the differentiation of focal pancreatic lesions.Methods and Materials: A series of twenty patients with focal pancreatic lesionsat US and/or CT and ten normal subjects underwent MR imaging at 1.5 T system(Signa Infinity, GE Healthcare), using a phased array coil. After the acquisition ofaxial T1w and T2w sequences with and without fat suppression (breath-holdingor respiratory triggering) and coronal MRCP, diffusion-weighted MR imaging wasperformed using a breath-hold single-shot echo-planar sequence with a b gradi-ent factor value of 500 sec/mm² in the three axes. Apparent diffusion coefficient(ADC) was calculated for cerebrospinal fluid, normal pancreatic parenchyma,and for each focal pancreatic lesion. Imaging results were correlated with ERCP,surgery, and/or imaging follow-up.Results: Final diagnoses included intraductal papillary mucinous tumor (IPMT)(n = 10), pseudocyst (n = 4), serous cystoadenoma (n = 3), and cancer (n = 3).Mean ADC value was 4.1x10-3 mm2/sec for cerebrospinal fluid, 1.73x10-3 mm2/sec for normal pancreatic parenchyma, 4.07x10-3 mm2/sec for IPMT, 3.62x10-3 mm2/sec for serous cystoadenoma, 2.61x10-3 mm2/sec for pseudocyst, and 1.57x10-

3 mm2/sec for ductal adenocarcinoma. Mean ADC values of different groups ofpancreatic lesions were statistically different (p < 0.05); particularly, ADC wassignificantly higher in IPMT than in the other pancreatic cystic lesions (p < 0.0001).Conclusion: Our preliminary results suggest that diffusion-weighted MR imag-ing may be helpful in the differential diagnosis of focal pancreatic lesions, spe-cially in the characterization of the cystic ones. Further investigations in largerseries of patients are necessary to establish its role.

10:30 - 12:00 Room E2

GI Tract

SS 201bMR and US in Crohn's diseaseModerators:P. Boraschi; Pisa/ITH. Czembirek; Vienna/AT

B-021 10:30

Clinical application study of true-FISP cine-MR sequences in evaluation ofCrohn's disease small bowel affected tracts, using polyethylene glycol(PEG) solution as oral contrast mediumF. Toso, E. Bruschi, R. Girometti, A. De Candia, G. Como, M. Bazzocchi;Udine/IT

Purpose: To investigate kinetic characteristics of small bowel tracts with sus-pected or diagnosed Crohn's disease using cine-MR sequences, and to comparethese data with radiological fluoroscopical findings.Methods and Materials: A prospective evaluation of 29 patients (22 affected byCD, 7 with suspected CD not histologically confirmed) were prospectively evalu-ated with MR-imaging and radiologic study, and endoscopic evaluation.Patients were prepared with 1.5-2 litres of PEG solution and scanned using 1.5 TMR system with trueFISP, Haste T2W, 3D-flash T1W sequences on coronal planes.Cine-MR sequence (trueFISP: TR 3.46 ms; TE 1.47 ms; thickness 4 mm; 30 con-secutive measurements on the same plan, 0.6 seconds each) was performed onoblique plans according to terminal ileum position. Images and MR-movies wereinterpreted by two expert observers without knowledge of the histologic results.Comparison of cine-MR and X-ray fluoroscopy was performed, and positive find-ings were correlated with distal ileum biopsies and surgery.Results: Terminal ileum was visualized in 27/29 patients (in two cases it wasdifficult to position the proper scan plan). Cine-MR sequences demonstrated in5/22 CD patients abnormal peristalsis, in 8/22 absence of peristalsis, while in 7/

22 peristalsis was found to be normal (histological diagnosis of early CD). Con-cordance of two radiologists between MR and fluoroscopic findings was excel-lent (k=1.0).Conclusion: TrueFISP cine-MR is a potentially excellent diagnostic tool to evi-dence absence or abnormal peristalsis of small bowel affected by CD, despitesometimes having difficulties with slice positioning on affected tracts.

B-022 10:39

MR imaging features of mesenteric lymphadenopathy in patients with smallintestinal Crohn diseaseS. Gourtsoyianni, N. Papanikolaou, A. Karantanas, J. Grammatikakis,J. Koutroubakis, N. Gourtsoyiannis; Iraklion/GR ([email protected])

Purpose: To describe MR enteroclysis (MRE) features of mesenteric lymphade-nopathy in patients with Crohn's disease, using ultrafast pulse sequences.Methods and Materials: 12 patients with Crohn disease underwent MR ente-roclysis examination. A high resolution true FISP sequence with fat saturationwas applied on coronal plane. In addition, a FLASH sequence with fat saturationwas acquired 75 seconds after the intravenous administration of gadobenatedimeglumine. Mesenteric lymphadenopathy was assessed with regard to preva-lence, anatomical distribution, size, shape and signal intensity characteristics onthe fat saturated true FISP images and enhancement ratio measurements of themesenteric lymph nodes after gadolinium administration on the FLASH images.Results: One hundred and sixteen mesenteric lymph nodes were depicted in 9out of 12 patients (80%). The maximum and minimum diameter of lymph nodeswere 0.93 cm (sd: 0.4 cm) and 0.6 cm (sd:0.2 cm), respectively, while the meansignal intensity on fat saturated true FISP images was 269.13 (sd:125). 71 oval-shaped and 50 round shaped lymph nodes were disclosed. Seven out of 9 pa-tients (77.7%) presented with homogeneous contrast enhancement while the meanenhancement ratio between lymph nodes and adjacent vessels was 0.77 (sd:0.2). The most common site of involvement was intermediate colic (n = 45, 35.3%),followed by ileocolic (n = 41, 38.8%), paracolic (n = 15, 12.9%), lateral aortic(n = 10, 8.7%) and inferior mesenteric (n = 5, 4.3%).Conclusion: Multiple mesenteric lymph nodes less than 10 mm in size can beregularly demonstrated on MR enteroclysis images. Homogeneous gadoliniumenhancement of mesenteric lymph nodes is a common finding in patients withCrohn disease.

B-023 10:48

Computer-assisted evaluation of perianal fistula activity with analultrasound in patients with Crohn's diseaseF. Caprioli, A. Losco, C. Viganò, P. Biondetti, D. Conte, L.V. Forzenigo,G. Basilisco; Milan/IT ([email protected])

Purpose: To assess, in a consecutive series of patients with perianal fistulizingCrohn's disease (CD); 1) the diagnostic performance of a visual assessment offistula activity at anal ultrasound (AUS) in comparison to magnetic resonanceimaging (MRI) T2-hyperintensity, 2) whether the application of a quantitative com-puterized analysis of grey-scale tones (GST) might improve the visual assess-ment.Methods and Material: 28 consecutive patients (mean age 37 years) with a CD-related perianal fistula were studied. Patients underwent MR imaging (Magne-tom SP63 1.5 Tesla, Siemens Avanto©, Germany) with T2-weighted imagesacquisition as reference standard for fistula activity. AUS was performed withFalcon 2101EXL equipped with a 1850 10Mhz rotating transducer (Brüel andKjær©, Herlev, Denmark). Visual assessment of fistula track activity at AUS wasperformed by two independent operators. Regions of interest (ROI) including thearea of fistula tracks were selected on AUS images and mean grey-scale tones(GST) of the selected ROIs were calculated with an image analysis software.Results: Sensitivity values for the visual assessment in the identification of ac-tive perianal fistulas were 0.85 - 0.90 and specificity 0.62-0.75; likelihood ratiosfor positive and negative test results were 2.4-3.4 and 0.16-0.20 for the two ob-servers, respectively. At the computer-assisted analysis of GST, hyperintensefistulas were significantly more hypoechoic than hypointense ones (P < 0.001).The computer-assisted analysis sensitivity was 0.85-0.95 and specificity 0.88 forboth operators; likelihood ratios for positive and negative test results were 7.1-7.9 and 0.06-0.17.Conclusion: Visual assessment of AUS images is a sensitive method to evalu-ate fistula activity in patients with CD. The application of a computer-assistedevaluation of GST confirms the results of visual assessment and improves itsspecificity.

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 143C D E FA G

B-024 10:57

Crohn's disease and color Doppler sonography: Response to treatment andits relationship with long-term prognosisJ. Ballestin, P. Calvillo Batlles, J. Cogollos Van der Linden, T. Ripolles,M. Martínez Pérez, M. Barrachina; Valencia/ES ([email protected])

Purpose: To evaluate if sonography detects changes in patients with Crohn'sdisease (CD) undergoing treatment, and if these findings are related to the pa-tient' long-term outcome.Methods and Materials: Twenty eight patients with confirmed CD (age; 15-51years, mean of 30 years) were prospectively examined with sonography beforeand during treatment. A total of three sonographic exams were made; the sameday of the beginning of treatment, 3 to 8 days thereafter and the last one about 6weeks later (or before, if clinical remission occurred in a shorter period of time).Sonographic examination included evaluation of bowel wall thickness and vascu-larity pattern. The sonographic data were compared with clinical and laboratorydata and its possible relation with the patients' long-term outcome.Results: In the initial sonographic examination, eighteen of twenty two patients(81%) with clinically active disease had moderate or marked parietal vascularity.Statistical significant reduction in color of the affected bowel was observed at thethird sonographic control (p < 0.05). Seventeen patients who were on clinical re-mission, had relapses and were treated with immunosuppressive therapy (n = 7)or surgery (n = 10) during the follow-up (mean: 33 months). Eighty-six percent ofpatients with persistent hyperemia (moderate or marked vascularity) after treat-ment, had a torpid clinical course, whereas only 30% of patients with absence ofor barely visible color Doppler flow had a torpid clinical course (p < 0.001).Conclusion: Sonography can identify bowel inflammation and its changes dur-ing treatment. Persistent hyperemia on color Doppler sonography during clinicalremission after treatment is predictive of relapse during the follow-up.

B-025 11:06

Assessment of Crohn's disease activity: Comparison between contrastenhanced MR imaging and contrast enhanced US harmonic imagingF. Maccioni, S. Parlanti, V. Cantisani, F. Siliquini, G. Luppi, M. Marini; Rome/IT([email protected])

Purpose: To compare contrast-enhanced ultrasound and contrast-enhanced MRimaging in the evaluation of CD activity.Methods and Materials: Fifteen consecutive patients with known Crohn's dis-ease underwent US and MR imaging. Patients were scanned using a 7.5 MHzprobe in harmonic mode with low mechanical index after 2.5 ml bolus injection ofSonoVue. Video-clips were examinated by two radiologists to determine and gradebowel wall thickness and wall enhancement, giving a 0-3 point score. Within 10days MR imaging was performed with a 1.5 Tesla system, after administration ofa negative oral contrast agent, using T2-weighted and T1-weighted sequencesbefore and after intravenous gadolinium injection on axial and coronal planes.MRI images were evaluated by two radiologists blind of the results of US.Results: Good correlation was found between MR imaging and US in the evalu-ation of bowel wall thickness. Wall enhancement resulted very high either in MRimaging and US in patients with hyperacute active disease (5 pts). Conversely,findings differed in active long-standing disease (7 pts), as we frequently observedat the level of the same affected bowel loops low enhancement at US and highenhancement at MR imaging, likely related to a different contrast agent behav-iour in presence of wall fibrosis. In non-active disease (3 pts) both imaging mo-dalities showed a low enhancement.Conclusion: MR imaging and US harmonic imaging give similar information onbowel wall inflammation in active disease. In the evaluation of non-active diseasepreliminary results show a different behaviour of contrast-enhanced US and MRimaging, likely due to different mechanisms of wall contrast-enhancement.

B-026 11:15

Dark lumen MR-enteroclysis in combination with MR imaging colonographyfor whole bowel assessment in patients with Crohn's disease: First clinicalexperienceA.G. Schreyer, N. Zorger, T. Finkenzeller, S. Feuerbach, H. Herfarth, J. Seitz;Regensburg/DE ([email protected])

Purpose: Magnetic resonance enteroclysis (MRE) is a recently introduced tech-nique to assess the small bowel with similar sensitivity and specificity as theconventional enteroclysis. MR imaging colonography (MRC) appears to be a prom-ising technique for polyp assessment in the colon. In this study we try to evaluatethe combination of small bowel MR imaging with unprepared MR imaging basedcolonography as an integrative diagnostic approach of the whole bowel in pa-tients with Crohn's disease.

Methods and Materials: 30 patients with known Crohn's disease were prospec-tively examined. No particular colonic preparation was applied. Applying the darklumen technique in all patients MRE as well as MRC were performed within onesession using an integrative examination protocol. T2 weighted as well as con-trast enhanced T1 weighted sequences were acquired. Inflammation assessmentof the colon was compared with conventional colonoscopy in 29 patient and withsurgery in 1 patient.Results: In 11 of 210 evaluated colonic segments feces hindered an adequateintraluminal bowel assessment. 23 of 30 patients had complete colonoscopy asgold standard. In 7 patients complete colonoscopy could not be performed be-cause of inflamed stenosis. Correct grading of colonic inflammation was per-formed with 55.1% sensitivity and 98.2% specificity in all segments. Consideringjust more extensive inflammation (grade 2) the sensitivity of MRC increased to70.2% with a specificity of 99.2%.Conclusion: The combination of MRE and MRC could improve the diagnosticvalue of abdominal MR imaging evaluation in patients with Crohn's disease. StillMRC can not replace conventional colonoscopy in subtle inflammation assess-ment.

B-027 11:24

MR of small bowel in patients with Crohn's disease: Evaluation ofinflammatory activity using standard and cine-MR sequencesF. Toso, E. Bruschi, R. Girometti, A. De Candia, D. Sorrentino, M. Bazzocchi;Udine/IT

Purpose: To assess the capability of MR to quantitatively evaluate the activity ofCrohn's Disease, using data from standard and kinetic sequences.Methods and Materials: 29 patients (20 with active CD of distal ileum, 2 withinactive CD, 7 with suspected CD not histologically confirmed) were prospective-ly evaluated with MR-imaging and radiological investigation using endoscopicbiopsy as standard reference.All patients were prepared with 1.5-2 litres of polyethylene glycol (PEG) solutionbefore evaluation and scanned using 1.5 T MR system with true-FISP and cine-True-FISP, Haste T2W and 3D-Flash T1W sequences on coronal and obliqueplans (pre- and post-Gd-DTPA). The data were interpreted by two expert observ-ers, without knowledge of the histological results. To assess activity grade of CDwe determinate a score ranging from 0 (absence of activity) to 14 (severe acuteCD) including information about contrast enhancement (quantitative and qualita-tive evaluation), wall thickness, bowel motility and stenosis, extravisceral find-ings.Results: At MR patients were classified as; 8/29 no activity, 9/29 low to moderateactivity and 12/29 high activity. We had 1 false positive case (moderate activity)due to contrast enhancement related to recent surgical procedure. Sensitivity,specificity, PPV and NPV of MR and X-ray were 100%, 87.5%, 95.2%, 100% and70%, 87.5%, 94%, 63% respectively. Concordance of two radiologists on MRscore calculation was excellent (k=1.0).Conclusion: Our score system for MR results to be feasible and reproducible toassess the activity grade of CD, with sensitivity and specificity values equivalentor superior to conventional radiological investigations.

B-028 11:33

Quantitative assessment of Crohn's disease activity by means of CE-MRIV. Pupillo, G. Frieri, M. Tanga, S. Delli Colli, E. Di Cesare, C. Masciocchi;L'Aquila/IT

Purpose: To report our experience using dynamic MR imaging after gadoliniuminjection to assess the inflammatory activity of Crohn's disease in a population ofadults and children.Methods and Material: We studied 30 patients with histologically proven Crohn'sdisease, who had MR imaging (1.5 T GE unit) using a phased array coil. Polyeth-ylene glycole solution oral contrast agent (10 ml/Kg) was administered. Our studyincluded the acquisition of T2-weighted SSFSE half Fourier sequences and FastSPGR T1-weighted fat saturated sequences before and after intravenous gado-linium administration (0.01 mmol/Kg); the acquisition was repeated 6 times. Wemeasured signal intensity of the normal and pathologic intestinal wall by meansof ROI evaluation. We performed time to signal intensity correlation to assessdynamic CE percentage. We compared this score with Crohn Disease ActivityIndex (CDAI).Results: We observed in terminal ileum intestinal wall pathological thickening in27 patients. MR imaging showed early and high CE between 140% and 310%(mean 210) in 22 patients in clinical acute phase of disease (15 pt) and in reacti-vation phase (7 pt). We observed 8 patients in post-therapy remission phase.After maximum contrast uptake, we observed a plateau phase in 22/30 cases for

Scientific Sessions

144 C D E FBA G

the total duration of measurement. A significant correlation existed between thepeak of contrast uptake and CDAI (r = 0.017982 p = 0.01).Conclusion: This data showed dynamic CE-MRI enables to quantify the inflam-matory activity of Crohn's disease. Therefore we propose CE-MRI in the follow-upof patients with Crohn's disease.

B-029 11:42

Dynamic contrast enhanced magnetic resonance imaging (MRI) in thedifferentiation of activity in Crohn disease patientsR.M. Del Vescovo, P. Pavone, R.F. Grasso, M. Cicala, R. Caviglia, I. Sansoni,P. d'Alessio, B. Beomonte Zobel; Rome/IT ([email protected])

Purpose: To prospectively investigate the correlation between Crohn DiseaseActivity Index (CDAI) and layered parietal enhancement in dynamic contrast en-hanced Magnetic Resonance Imaging (MRI) of the bowel wall.Methods and Materials: A study was performed in 17 patients who underwentserial bowel dynamic contrasted-MR imaging using an high-resolution protocolwith 1.5 T magnet. Systematic qualitative analysis of morphologic features andregion-of-interest-based analysis of enhancement kinetics were performed. En-hancement curves were obtained at the mucosa-submucosa layer and at thelevel of thickened bowel wall.Results: A statistically significant difference in enhancement curves of activeversus inactive lesions was found using the dynamic protocol. Patients with ac-tive disease always presented with a rapid contrast uptake of the mucosa-sub-mucosa layer; contrast of the mucosa-submucosa layer upsloped rapidly, with apeak enhancement at 80-130 sec following injection, then a decrease was ob-served. Thickened layer of the bowel wall presented a slow uptake of contrast,with a plateau in late phase. Enhancement curves crossed each other after 170-250 sec following dynamic injection. In patients with chronic disease a more ho-mogeneous type of contrast uptake was observed, with slow uptake in bothanatomical areas and no difference in the enhancement curves.Conclusion: Increased spatial resolution significantly improves diagnostic confi-dence and accuracy at dynamic MR imaging. Using high spatial resolution tech-niques at 1.5 T it was possible to differentiate between the enhancement of thedifferent layers of the bowel wall. A correlation between layered bowel wall en-hancement and Crohn disease activity index (CDAI) is suggested by the earlyresults.

B-030 11:51 !Assessment of Crohn's disease activity using quantitative perfusion MRimagingS. Aufort, B. Gallix, M.-A. Pierredon-Foulogne, M. Veyrac, P. Blanc, J.-M. Bruel;Montpellier/FR ([email protected])

Purpose: To determine if bowel wall perfusion, as measured with dynamic con-trast-enhanced MR imaging (DCE-MRI), reflects clinical activity.Methods and Material: Bowel wall perfusion was estimated with DCE-MRI atthe level of the more affected bowel loops, as determined by morphological MRimaging. All patients underwent clinical, biological assessment (Crohn's diseaseactivity index CDAI measurement) and MR examinations in the same week. Theevaluation was performed before treatment and repeated 2 weeks and 6 weeksafter Infliximab injection. DCE-MRI was performed using a 1.5 T MR-scannerwith multislice 3D-TFE perfusion sequences (TR/TE/FA 3.3/1.66/10; acquisitiontime 180 ms; flow rate 6 cc/s). Perfusion was analysed after deconvolution witharterial input function.Results: 23 patients underwent 57 MR examinations. Affected bowel loops hadhigher perfusion index with active Crohn's disease than during post-treatmentremission (local flow index: 0.158 vs 0.101, time to peak delay: 19.4 vs 26.2;p <.001).Conclusion: Bowel wall perfusion, as measured with DCE-MRI at the level of theGI tract affected loops, reflects disease activity in patients with Crohn's disease.

10:30 - 12:00 Room F1

Chest

SS 204Lung cancer: Screening and stagingModerators:P. Schnyder; Lausanne/CHD. Wormanns; Münster/DE

B-031 10:30

Baseline findings of a randomized Dutch-Belgian trial on lung cancerscreening by low-dose CT (NELSON study)D. Xu1, R. Vliegenthart1, Y. Wang1, T. Willems1, H. van der Zaag1,R. van Klaveren2, W.P.T.M. Mali3, H. de Koning2, M. Oudkerk1; 1Groningen/NL,2Rotterdam/NL, 3Utrecht/NL ([email protected])

Purpose: The NELSON study is a randomized screening trial in current and formersmokers, aimed at early detection of lung cancer by low-dose spiral CT, anddesigned to be large enough to show 20% reduction in lung cancer mortality.Methods and Materials: Men aged 50-75y are approached by mail through pop-ulation registries. After the first and second mailing, 11,000 current or ex-smok-ers (≥ 20 pack-years) are randomized to either undergo 16-slice MDCT screening(Siemens) versus a control group. Low-dose CT scans are performed in foursites, and LungCARE software is used for analyzing the scans. Nodules are di-vided into 3 categories: solid, non-solid and partly-solid. Solid nodules of 50-500 mm3, non-solid nodules with diameter > 8 mm, partly-solid nodules with theabove combined criteria are denominated as significant nodules, and a follow-upscan is performed after 3 months. Larger nodules (> 500 mm3) or growing nod-ules are defined as potentially malignant nodules. Participants are immediatelyreferred to a pulmonologist.Results: From April 2004 to March 2005, of the first 3,986 participants screened,80% (3167) had no significant pulmonary nodules. 21% (819) showed significantor potentially malignant nodules. 97% (743) had stable significant nodules after 3months. 2% (76) of the participants were referred to a pulmonologist. 0.5% (21)participants were diagnosed as having lung cancer.Conclusion: In an elderly population of male current and ex-smoker, the positivetest result of baseline screening was 21%. However, the cancer detection ratewas only 0.5%.

B-032 10:39

Morphological features of 453 indeterminate pulmonary nodules at threemonths follow-up scan in lung cancer screeningD. Xu1, Y. Wang1, H. van der Zaag1, T. Willems1, R. van Klaveren2, W. Mali3,H. de Koning2, M. Oudkerk1; 1Groningen/NL, 2Rotterdam/NL, 3Utrecht/NL([email protected])

Purpose: To analyze which type of nodules, found after baseline scanning, showgrowth within 3 months.Methods and Materials: 11,000 male current or ex-smokers (50-75y) with atleast 20 cigarette pack-years were randomized to low-dose CT screening (120kv, 18 mAs, 1.0 mm slice thickness, 0.7 mm reconstruction increment) versuscontrol. Nodules were divided into 3 categories: solid, non-solid and partly-solid.Solid nodules with volume 50-500 mm3, non-solid nodules with diameter > 8 mmand partly-solid nodules with the above combined criteria were considered asindeterminate nodules, and a follow-up scan was performed after 3 months.Results: From April.2004 to Sept.2005, 3,881 participants, who underwent base-line scanning, were retrospectively analysed. In 331 participants, 453 indetermi-nate solid nodules were observed. 20% (89/453) of the nodules were minor ormajor fissure-attached nodules, 15% (68/453) were pleural based, 18% (83/453)were vessel-attached nodules. 18% (82/453) of the nodules did not show typicalmorphological features (focal opacity). 29% (131/453) were solitary pulmonarynodules. Only 3% (14/453) of the indeterminate nodules found at baseline hadgrown more than 25% after 3 months. Six were solitary pulmonary nodules, fivewere interpreted as fibrosis, scar or even missed at the baseline scan. Two werevessel attached and one was fissure attached.Conclusion: Focal opacity and solitary pulmonary nodules have the highest pos-sibility to grow, whereas pleural based indeterminate nodules (well defined) donot grow within 3 months. With 95% confidence, the chance that fissure andvessel attached nodules will grow within 3 months is less than 8%.

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 145C D E FA G

B-033 10:48

Small nodules along the fissures in neoplastic patients: Incidence andclinical significanceM.L. Storto, R. Patea, M. Mereu, C. Torrione, A. Travaglini, B. Feragalli;Chieti/IT ([email protected])

Purpose: To assess the incidence and clinical significance of non calcified nod-ules along the fissures seen on CT scans of the chest in patients with knownmalignancy.Methods and Materials: Our study population included 163 neoplastic patients(124 with extrapulmonary neoplasms) undergoing follow-up CT, for whom 2 chestCT scans performed at least 12 months apart were available. CT scans wereacquired on a multidetector row CT scanner and were reviewed for the presence,size, shape, and location (either within the parenchyma or along a fissure) of noncalcified nodules. Change in size of nodules was also assessed on follow-upexams.Results: 817 nodules were identified in 141 patients; 690 nodules (84.5%; 136patients) were located within the parenchyma and 127 (15.5%; 42 patients) wereseen along the fissures. Most fissural nodules were seen along the major fissure(57 on the right; 40 on the left) and showed a triangular (55; 43.3%) or oval (48;37.8%) shape, whereas parenchymal nodules were more frequently rounded(92.1%). The mean diameter of fissural nodules was significantly lower than thediameter of parenchymal ones (3.3 mm vs 1.2 mm; p <.05). Fissural nodules re-mained stable on follow-up CT scans, whereas an increase or a decrease in sizewas observed in 457 (66.2%) parenchymal nodules.Conclusion: Small non calcified nodules along the fissures can be observed onCT scans in as many as 15.5% of patients with malignancy. They did not increasein size on 12-month follow-up CT scans and should not be considered metastaticlesions.

B-034 10:57

Growth of lung adenocarcinomas showing localized ground-glass opacityT. Aoki, M. Hanamiya, S. Kawanami, Y. Korogi, H. Hashimoto, K. Yasumoto;Kitakyushu/JP

Purpose: To assess the natural progression of peripheral lung adenocarcinomasthat presented with localized ground-glass opacity (GGO) on thin-section CT.Methods and Materials: A retrospective review was carried out on 22 patientswith lung adenocarcinomas smaller than 3 cm, in which tumor growth could beevaluated using thin-section CT prior to surgery. The CT observation period rangedfrom 112-2333 days (mean 740 days). Two thoracic radiologists evaluated tumorgrowth by analyzing GGO type (pure or mixed), pleural indentation, vascular con-vergence, increases in size, and tumor doubling time (Tdt).Results: On initial CT, mixed GGO was present in 7 cases and pure GGO waspresent in 15 cases. An increase in size occurred in 16 cases (73%) during anobservation period of 153-2333days; all 16 tumors showed a Tdt greater than241 days. In the remaining 6 cases (27%), no change could be seen during anobservation period of 112-1128days. Average Tdt was 523.5 days for cases withmixed GGO, and 760.7 days for cases with pure GGO. An increase of the solidcomponent (non GGO area) occurred in 6 cases (27%). The development of apleural indentation and/or vascular convergence occurred in 8 cases (36%) with5 cases consisting of a mixed GGO (71%), and 3 cases consisting of a pureGGO (20%) (P < 0.05).Conclusion: Lung adenocarcinomas showing a large GGO lesion often growslowly. Pleural indentations and/or vascular convergence occur more frequentlyin mixed GGO lesions than pure GGO lesions.

B-035 11:06

Bronchoscopic fine needle aspiration under guidance by CT-bronchoscopicnavigation in comparison to conventional orientation: Initial resultsG.M. Weiner1, B. Geiger2, K. Schulze1, P. Hortig1, S. Coupland1, K.-J. Wolf1,T. Albrecht1; 1Berlin/DE, 2Princeton, NJ/US ([email protected])

Purpose: Customarily bronchoscopists use axial CT-sections for orientation dur-ing transbronchial fine needle aspiration (TBNA). If the target lesion is not visiblebronchoscopically the insertion point for the needle can only be estimated. Asoftware-tool based on virtual bronchoscopy (CT-Bronchoscopic Navigation) wasdeveloped to facilitate the orientation during TBNA. The purpose of this studywas to evaluate if the yield of TBNA increases when using CT-BronchoscopicNavigation in comparison to the conventional method.Method and Materials: The software uses DICOM-data of multislice-CT andgenerates a real time virtual bronchoscopy. The target lesions are displayed be-hind the semitransparent bronchial wall on virtual bronchoscopy. Additionally, dis-

tance measurements can be performed to define the insertion point virtually. Thesoftware runs on a laptop to simulate the bronchoscopic puncture next to thebronchoscopy unit. 20 patients with suspect mediastinal masses underwent TBNA.In a first pass the orientation for needle positioning was done conventionally us-ing axial CT-sections, followed by a second pass with guidance by CT-Broncho-scopic Navigation. In total 39 lesions were punctured. The cytologic results ofboth methods were compared.Results: 14 of 39 lesions were classified as malignant by TBNA or resection. 5 of14 malignant lesions were discovered using the axial CT-sections, 12 of 14 le-sions were detected by punctures which were guided by CT-Bronchoscopic Nav-igation.Conclusion: CT-Bronchoscopic Navigation increases the yield of TBNA substan-tially in comparison to the conventional type of orientation with axial CT sections.

B-036 11:15 !The usefulness of super paramagnetic iron oxide (SPIO)-enhanced MRimaging for diagnosing nodal status of lung cancerM. Kitase, M. Hara, S. Mori, M. Iwasa, M. Shimohira, H. Suzuki, H. Oshima,N. Shiraki, Y. Shibamoto; Nagoya/JP ([email protected])

Purpose: To evaluate usefulness of SPIO-enhanced MRI for excluding false-pos-itive nodes on FDG-PET.Methods and Materials: Fifteen patients with lung cancer diagnosed to have N2or N3 nodes by FDG-PET underwent T2W spin-echo MRI with and without SPIO.The change of relative signal intensity (rSI) after SPIO administration was evalu-ated. In ten patients, the diagnosis was confirmed pathologically by operation,and in the other five patients, it was made by careful follow-up. On MR imaging,lymph nodes were considered positive when the rSI decreased by more than17% after SPIO administration, or the SI of lymph node was less than twice thatof adjacent muscle on unenhanced T2W images. Lymph nodes with higher accu-mulation of FDG than the surrounding mediastinum level were judged as posi-tive. The SPIO-enhanced MR imaging findings were compared with PET, CT andhistopathologic findings.Results: Sixty-seven lymph nodes in 15 patients were examined; 27 lymph nodes(40%) were positive and 21 (78%) were diagnosed correctly using SPIO-MRI. Of40 (60%) benign nodes, 34 (85%) were judged as negative using SPIO-MRI.Sensitivity, specificity and overall accuracy of SPIO-MRI were 78%, 85% and82%, respectively. Nine of 15 patients (60%) diagnosed to have N2 or N3 diseaseusing PET had overdiagnosis. On CT, sensitivity, specificity and overall accuracywere 65%, 61% and 63%, respectively.Conclusion: These results indicated that SPIO-MRI might improve accuracy ofdiagnosing nodal status in lung cancer patients, particularly with incorrect PETdiagnosis of having N2 or N3 disease.

B-037 11:24

Additional value of integrated PET-CT in the staging of lung cancer:Comparison with CT alone, PET alone and visual correlation of PET and CTimages side by sideW.F.M. De Wever, S. Ceyssens, S. Stroobants, J. Bogaert, J. Verschakelen;Leuven/BE ([email protected])

Integrated PET-CT is a new anatomo-metabolic imaging modality offering ana-tomic and metabolic information.Purpose: To evaluate the accuracy of integrated-PET-CT in the staging of pa-tients with lung cancer and to compare the accuracy of this technique with theaccuracy of CT alone, PET alone and visually correlated PET-CT images side byside.Methods and Materials: 50 patients undergoing an integrated-PET-CT for stag-ing of a suspicious lung lesion were studied. All these patients had surgical stag-ing and histopathological examination of the lesion. Tumor (T), lymph node (N)and metastatic (M) status were determined with CT, PET, visually-correlated PET-CT and integrated-PET-CT. These TNM stages were compared with the surgicalTNM status.Results: Integrated-PET-CT was most accurate in the assessment of the TNMstatus. Integrated-PET-CT predicted correctly the T-status in 86%, the N-status in80%, the M-status in 98% and the TNM status in 70% of the cases versus68%,66%, 88%,46% with CT, 46%,70%,96%,30% with PET and72%,68%,96%,54% with visually-correlated PET-CT. Integrated-PET-CT over-staged the T-status in 8% and N-status in 16% of the cases. These findings werefor CT 20% and 28%, for PET 16% and 20% while visual-correlated PET-CToverstaged in 12% and 20%. There was an understaging with integrated-PET-CTfor T-status in 6% and for N-status in 4% versus 12% and 6% with CT, 38% and10% with PET and 12% and 12% with visual-correlated PET-CT.

Scientific Sessions

146 C D E FBA G

Conclusion: Integrated-PET-CT improves the staging of lung cancer through abetter anatomic localization and characterization of lesions and is superior to CTalone, PET alone and visually-correlated PET-CT.

B-038 11:33

Preoperative lymph node staging in non-small cell lung cancer:Comparison between FDG PET and CTY.-H. Kim, Y.-M. Kim, H.-J. Seon, K.-J. Na, Y.-C. Kim, H.-S. Beom, J.-K. Kim,J.-G. Park, H.-K. Kang; Gwangju/KR ([email protected])

Purpose: To evaluate the accuracy of 18F-FDG PET and CT for preoperativelymph node (LN) staging of non-small cell lung cancer (NSCLC), correlated withhistopathologic findings.Methods and Materials: 46 patients (30 men, 16 women; mean age, 53 years)with NSCLC underwent surgical resection (tumor resection and LN dissection)after CT scans followed by FDG PET scans. Histopathologic results served asthe reference standard. The shortest (SD) and the longest diameters (LD) of LNswere measured on chest CT, and calculated roundness index (RI). FDG uptakeby the lesion was measured as the standardized uptake value (SUV). Statistical-ly significant differences for sensitivity, specificity, accuracy were determined usingthe McNemar test or with a generalized estimating equation.Results: 172 LN were detected on chest CT. Average of the SD was 8.24 mm,the LD was 12.5 mm, and the RI was 1.59. 29 LN revealed positive SUV (aver-age, 4.0) on FDG-PET. For the depiction of malignant nodes, the sensitivity, spe-cificity, and accuracy of CT were 33.3%, 85.5%, and 77.3%, respectively, whereasthose of FDG-PET were 44.4%, 76.8%, and 73.7%. If we considered cut-off val-ue for the SD of positive LN on CT as 8.8 mm, the sensitivity, specificity, andaccuracy were 72.3%, 71%, and 72.1%.Conclusion: In our study, the SD of the LN on chest CT might be better than theSUV of LN on FDG-PET for predicting LN metastases from NSCLC. The cut-offvalue of the SD of 8.8 mm on CT is useful for distinguishing LN metastases.

B-039 11:42

Detection of asymptomatic coronary artery disease prior to thoracicsurgery: Preliminary experience with 64-slice CT angiography in 99consecutive patients without specific medicationD. Delhaye, M. Rémy-Jardin, C. Rozel, C. Dusson, C. Khalil, A. Wurtz,J. Rémy; Lille/FR ([email protected])

Purpose: To screen for asymptomatic coronary artery disease prior to thoracicsurgery during presurgical staging of non-small cell lung (NSCL) cancer.Methods and Materials: 99 consecutive patients in sinus rhythm underwent (a)an ECG-gated acquisition over the cardiac cavities with administration of 120 mLof a 35% contrast agent; (b) followed by a nongated examination of the upperthird of the thorax. Data were reconstructed: (a) to assess the presence of coro-nary artery disease; and (b) to obtain presurgical staging of NSCL cancer on 3-mm thick transverse scans.Results: The mean level of enhancement within coronary arteries was278.49 ± 55.47 HU. Among the 396 proximal coronary arteries (CA) evaluated,376 arteries (95%) were rated as interpretable (main left CA: n = 97; 98%; leftanterior descending artery: n = 99; 100%; circumflex artery: n = 90; 91%; rightcoronary artery: n = 90; 91%) with 76% of all coronary segments rated as inter-pretable. The mean percentage of interpretable segments was significantly high-er in patients with a heart rate ≤ 80 bpm (n = 62) than in patients with a heart rate> 80 bpm (n = 37) (82 ± 11% vs 72 ± 13%; p = 0.008). CTA detected calcifiedplaques (n = 117), extensive calcifications (n = 22), mixed (n = 10) and soft(n = 10) plaques and stenoses (n = 8) in a total of 41 patients (41% of the studiedpopulation). Diagnostic image quality was achieved in all patients for staging ofNSCL cancer.Conclusion: Coronary arteries can be evaluated during presurgical staging ofNSCL cancer with 64-slice CT angiography without beta-blockers.

B-040 11:51

Comparison of survival for patients treated with radiofrequency ablationwith or without chemotherapy for inoperable non-small cell lung cancerG.-Y. Jin, Y. Han, Y. Lee, Y. Lee; Jeon-ju/KR ([email protected])

Purpose: We analyzed long term survival in patients treated with radiofrequencyablation (RFA) with or without chemotherapy for inoperable non-small cell lungcancer.Methods and Materials: From June 2000 to May 2004, fifty patients (mean age66.8 years: range 27-83 years) who had inoperable non-small cell lung malignan-cy underwent RF ablation under CT guidance. Sixteen patients (32%) had Stage

I lung cancer; twenty-five (50%) had unresectable stage II, III or IV cancer; andnine (18%) had single or two metastatic nodules. These patients were dividedinto three groups: Group I (n = 16) included stage I patients who underwent onlyRFA. Group II (n = 21) included the patients except for stage I, who underwentRFA and chemotherapy after RFA. Group III (n = 13) included the patients exceptfor stage I, who underwent only RFA. We assessed mean and median survivalsof the three groups using the ANOVA test and Kaplan-Meier test.Results: Of fifty patients, twelve patients are still alive and thirty-eight patientsdied. Mean and median survival of the group I, II, and III are 26.5 ± 17.5 and 19months, 15.8 ± 8.2 and 15 months, 13.6 ± 9.2 and 10 months, respectively. Meansurvival of group I is significantly longer than groups II and III (p < 0.05). Howev-er, mean survival difference between groups II and III was not significant statisti-cally.Conclusion: Radiofrequency ablation showed very promising survival in stage Ipatients with inoperable non-small cell lung malignancy.

10:30 - 12:00 Room F2

Cardiac

SS 203The atheromatous plaque and the calcium burdenModerators:I. Mastorakou; Athens/GRT.P. Willems; Groningen/NL

B-041 10:30

Enhanced MR imaging of atherosclerotic plaque with novel contrast mediagadofluorine in experimentally induced atherosclerosis of rabbit aortaW. Lee1, J. Chung1, W. Kim1, J. Park1, H. Weinmann2; 1Seoul/KR, 2Berlin/DE([email protected])

Purpose: To clarify the atheroma specific enhancement and the characteristicsof the enhancement pattern of gadofluorine.Methods: Thirteen male New Zealand white rabbits fed with a hypercholesteroldiet after inducing diabetes mellitus and two normal rabbits were used. The MRimaging was performed at 18 to 24 weeks after inducing diabetes mellitus. Theanimals were divided into four groups; Group I - atheroma model and 2 daysdelayed gadofluorine enhancement (n = 5), II - atheroma model and 9 days de-layed gadofluorine enhancement (n = 5), III - atheroma model and non contrastenhancement (n = 3), IV - normal aorta with 2 days delayed gadofluorine en-hancement (n = 2). Ex vivo MR imaging of the segments were obtained with a 14T micro MR system. Fat saturated T1-weighted spin echo images were obtained.Histologic examinations were performed after Hematoxylin-eosin staining. Thepresence of enhancement and the pattern of enhancement were analyzed ac-cording to the thickness and component of the atheroma.Results: The atheroma positive sectors showed enhancement in 100% of thesectors in group I and 95.4% of the sectors in group II. However, there was noenhancement in groups III or IV. The average signal intensity ratio between athero-ma and aorta wall were 1.418 in group I, 1.238 in group II and 0.863 in group III.The thick atheroma showed stronger enhancement, but the atheroma compo-nents did not affect the degree of enhancement of the atheroma.Conclusion: Gadofluorine shows specific enhancement in the early stages ofatheroma and there is no specific tissue component affinity of gadofluorine.

B-042 10:39

MR imaging signs of carotid plaque inflammation in patients with unstableanginaL. Natale, A. Bernardini, A. Porcelli, A. Meduri, A. Lombardo, L. Bonomo;Rome/IT ([email protected])

Purpose: Inflammation may contribute to destabilize vulnerable plaques in acutecoronary syndromes by promoting rupture and erosion. Systemic inflammatoryfactors could be related with widespread plaque activity in many vascular dis-tricts. We evaluated with contrast enhanced-MR imaging (CE-MRI) plaque in-flammation signs in carotid arteries of patients with unstable angina (UA) andrelated them to serum levels of C-reactive protein (CRP), a marker of systemicinflammation.Methods and Materials: 32 patients with carotid plaques, 19 with UA and 13with stable angina underwent carotid arteries MR imaging (1.5 T GE scanner),with SE, FSE and bb-FSE sequences, before and 10 to 15 minutes after Gado-linium-DTPA (Gd) i.v. administration (0.2 mmol/Kg). We evaluated wall thickening

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 147C D E FA G

(arterial wall edema or infiltration), increased T2 or FSE-STIR signal intensity(SI) (arterial wall or plaque edema) and arterial wall or plaque enhancement(increased capillary permeability). In total, 44 plaques were evaluated (20 monol-ateral and 12 bilateral). CRP levels were determined with the ELISA essay.Results: 24 plaques showed wall thickening and/or increased T2 and/or FSE-STIR SI and Gd enhancement, 5 plaques showed only CE, whereas 15 plaqueshad no signs of inflammation. CRP levels of patients with enhanced plaques weresignificantly higher than those of patients without enhancement (median values:12.4 vs 3.4, p < 0.05).Conclusion: Pts with UA showed signs of plaque inflammation more frequentlythan did controls. Patients with UA and inflammation signs showed higher CRPlevels than pts with UA and no signs of inflammation. Our results suggest a wide-spread plaque activity, possibly mediated by systemic inflammation, occurs inacute coronary syndromes.

B-043 10:48

Differentiation of coronary atherosclerotic plaque components by multi-row-detector computed tomographyF. Knollmann, M. Galonska, T. Kertesz, F. Ducke, H. Guski, R. Meyer, R. Felix;Göttingen/DE ([email protected])

Background: Recent studies suggest that multidetector computed tomography(MDCT) has a potential to identify vulnerable plaques. We investigated if 16-slice-MDCT allows reliable identification of calcified and noncalcified coronary plaquesand further classification of noncalcified plaques into either lipid-rich with a necroticcore or fibrous.Methods: Coronary arteries of 30 hearts were filled post mortem with a contrastmedium and scanned with a 16-slice-MDCT imager. The images were reformat-ted perpendicular to the axis of the main coronary arteries and analysed by es-tablishing the attenuation profiles of the coronary cross sections (NIH ImageJ1.33n software). Results were compared to the correlating histopathological sec-tions of the arteries.Results: Analysis of 195 CT-cross sections showed a sensitivity and specificityfor the detection of calcified plaques of 100% and 97.3% and for noncalcifiedplaques of 80.8% and 95.1% respectively. The attenuation of epicardial fat rangedfrom -119 HU to 23 HU (median: -71 HU), and from 93 HU to 625 HU (308 HU)for the contrast medium. Calcified plaques showed an attenuation between 333 HUand 1944 HU (1089 HU), noncalcified plaques between 26 HU and 124 HU(52 HU). Further sub-classification of noncalcified plaques showed attenuationvalues between 26 HU and 67 HU (median: 44 HU) for lipid-rich plaques with anecrotic core and from 37 HU to 124 HU (median: 67 HU) for fibrous plaques.Conclusion: Coronary atherosclerotic plaques can be reliably identified and clas-sified as either calcified or noncalcified by 16-slice-MDCT in post-mortem stud-ies. Further differentiation of noncalcified plaques in either lipid-rich or fibrous isnot reliably feasible due to substantial overlap of the attenuation.

B-044 10:57 !Effect of statins on non-calcifying coronary plaque: Non-invasivemonitoring with 16-multidetector computed tomographyG.M. Feuchtner, G.J. Friedrich, W. Dichtl, T. Schachner, S. Mueller,O. Pachinger, D. zur Nedden; Innsbruck/AT ([email protected])

Purpose: Statins may influence coronary plaque due to lipid-level-independent,pleiotropic effects. Aim of this study was to evaluate the effect of a de-novo statintherapy on non-calcifying coronary plaque with 16-multidetector computed tom-ography (16-MDCT) at a 6 months and a 1 year follow-up.Methods: 11 patients (from [1] Austrian Atrovastatin Study: randomized, doubleblind, prospectively and [2] coronary bypass surgery follow-up study) were pro-spectively assessed with 16-MDCT (Sensation 16™, Siemens)(16 x 0.75 mm;0.42 s; TF=6.7 mm/s; 120 kV,500 mAs; 100 ml Iodixanol (Visipaque™, Amer-sham), 3-4.5 ml/s flow; eff. slice 1 mm, inc. 0.6, retrospective ECG-gating at 60-80%) as follows: baseline and 6 months or 1 year after de-novo statin therapy(20 mg atrovastatin or 20 mg simvastatin/day). Plaque area [PA], lumen area [LA],mean CT-densities (Hounsfield Units [HU]) of non-calcifying plaque and intralu-minal CT attenuation (pre-lesion) was measured.Results: A total of 43 non-calcifying lesions (prox.coronary segments [1.5,6.11])were assessed. After 1 year of statin treatment [n = 20], a highly significant re-gression of plaque area (p < 0.001, t-test)(0.10 cm² ± 0.03vs.0.05 cm² ± 0.03)was observed. After 6 months [n = 24], a less significant (p = 0.02, t-test) de-crease of PA was noted (0.17 cm² ± 0.03 vs.0.14 cm² ± 0.04). CT-density of non-calcifying plaque showed a significant (p = 0.02) increase (69 HU baseline vs.91 HU) after 6 months/1 year. The difference in LA was not significant. The differ-ence in intraluminal CT attenuation values between baseline and follow-up wasmean 4.6 HU (range 0.6-14.4 HU)(absolute 288.9 vs.290.6 HU).

Conclusion: Statins may induce a regression of non-calcifying coronary plaqueand lead to an increase in CT-densities suggesting higher fibrotic plaque compo-nent which could lead to plaque stabilization. 16-MDCT has the potential to iden-tify patients who may benefit from lipid-level independent statin treatment.

B-045 11:06

Assessment of atherosclerotic plaque morphology with 16-slice MDCT:Comparison with histologyF. Knollmann, F. Ducke, T. Kertesz, H. Guski, R. Meyer, R. Felix; Göttingen/DE([email protected])

Purpose: To histopathologically evaluate the accuracy of 16-slice MDCT for as-sessment of plaque morphology.Methods and Material: 32 human hearts were obtained at autopsies and im-aged using 16-slice MDCT (LightSpeed 16pro, GE Medical Systems) with spatialresolution of 0.4x0.4x0.6 mm, gantry rotation 500 ms, 250 mA, 120 kV and512x512 pixel matrix. The coronary arteries were subdivided into segments us-ing anatomic landmarks. In every segment the lesion with the greatest plaqueburden was chosen. In cross-sectional images we manually measured plaquearea, noncalcified plaque area, calcified area and hypodense lipid-rich area.Plaque morphology was assessed upon visual inspection without using prede-fined density thresholds. Following the CT investigation, all coronary arteries weresectioned transversally and stained with hematoxylin-eosin and Elastica-van-Gie-son. Histopathology was correlated with the corresponding cross-sectional CTimages and the same morphologic parameters were documented.Results: 226 arterial segments were analysed. Calcium and Lipid were detectedin MDCT with a sensitivity of 86% and 46%, respectively, and a specifity of 100%compared to histology. Measurements of plaque area, noncalcified plaque area,calcified area and lipid-rich area as assessed by MDCT and histology correlatedwith r = 0.67, r = 0.56, r = 0.83 and r = 0.62 (P < 0.0001), respectively. MDCToverestimated all of these parameters.Conclusion: MDCT is able to detect and differentiate coronary atheroscleroticplaque components. Detection of lipid inclusions still suffers from the limited spa-tial resolution and is reliable only for large lesions.

B-046 11:15

MDCT of coronary arteries: Reproducibility and comparison with pathologyF. Knollmann, L. Krist, F. Ducke, T. Kertesz, H. Guski, R. Meyer, R. Felix;Göttingen/DE ([email protected])

Purpose: Our aim was to determine the reproducibility of the coronary artery CT-angiography with regard to detecting coronary artery stenoses and assessingatherosclerotic plaque burden.Methods: In 30 autopsy specimens, the coronary arteries were depicted on aclinical 16-row MDCT scanner, with the coronary arteries filled with a contrastagent. A clinical coronary helical CTA protocol was employed with a slice thick-ness of 0.63 mm and simulated EKG-gating. For each coronary artery segment,vessel area, lumen area, atherosclerotic plaque area, calcified plaque area, andlipid core area were determined by two independent observers. For comparison,the respective areas were determined from the pathologic specimen by the sametwo independent readers. The two CT readings were compared by linear regres-sion analysis and Bland-Altman plots.Results: On MDCT, the correlation between the two observers were highly sig-nificant for all comparisons (p < 0.0001), although the regression coefficient wasonly 0.43 for total vessel area, 0.35 for lumen area, 0.6 for plaque area, 0.7 forcalcified plaque area, and 0.65 for lipid core area. Upon histology, the reproduc-ibility of total vessel area (regression coefficient, 0.77), lumen area (0.74), andplaque area (0.68) were better than upon CT, while the reproducibility of calcifiedplaque area (0.63) and lipid core area (0.63) were worse than upon CT.Conclusion: We conclude that the reproducibility of coronary artery lumen sizeand the size of uncalcified plaque components are limited in 16-row MDCT, andworse than in pathological specimens. Interobserver variability is an importantlimitation of the technique.

B-047 11:24

Evaluation of coronary artery plaque in the patients having negativecoronary artery calcium scoring with 16- and 64-slice multidetector row CT;the preliminary results in 211 patientsM. Kim, J. Kim, S. Byun, C. Park, S. Choi, D. Yang, W.-H. Kim, H. Kim;Incheon/KR ([email protected])

Background: Absence of coronary artery calcification on unenhanced CT hashigh negative predictive value to exclude coronary artery obstructive disease

Scientific Sessions

148 C D E FBA G

(CAOD). Nevertheless, in about 5-10% of the patients, CAOD is present despitenegative calcium scoring due to so-called 'soft plaque'.Purpose: We analyzed the characteristics of coronary artery plaque in coronaryCT angiography in those patients.Methods: Two hundred eleven patients without history of known CAOD betweenMarch 2004 and July 2005 performed both precontrast CT (for calcium scoring)and coronary CT Angiography. In the patients showing negative calcium scoring,coronary CT angiography was retrospectively reviewed on the presence of coro-nary artery plaque and its characteristics (location, attenuation (HU), volume(mm3), degree of stenosis).Results: One hundred nine patients had no coronary artery calcium deposits onprecontrast CT. Coronary artery plaques were seen in 8 (7.3%) patients on coro-nary CT angiography. The location was right coronary artery (n = 3), left anteriordescending artery (n = 3), left main coronary artery (n = 2). Mean attenuationwas 76 HU (range, 52.2-107.1 HU). Average volume was 0.55 mm3 (range, 0.15-1.79 mm3). Stenosis degree of plaques was a mean of 40.7% (range, 16.0-68.0 %).Conclusion: Though in a small number of patients, detectable coronary arteryplaques were present despite negative calcium scoring. The plaque was locatedin main coronary arteries and had intermediate characteristics. Minimal volumeof the plaque was 0.15 mm3. Because the plaque led to significant stenosis, neg-ative calcium scoring does not seem to guarantee the absence of CAOD.

B-048 11:33

Intersystem variability of calcium scoreM.J.W. Greuter1, D.J.A. Kuijpers2, E.T. Scholten3, M. Oudkerk1; 1Groningen/NL,2The Hague/NL, 3Haarlem/NL ([email protected])

Purpose: The purpose of this study was to assess the intersystem variability ofcalcium scoring using three identical MDCT systems and an EBT system with acardio phantom.Methods: An anthropomorphic cardio CT phantom (QRM, Möhrendorf, Germa-ny) was scanned on three identical 16-detector MDCT units (Sensation 16, Sie-mens, Forchheim, Germany) in three different hospitals using a standardsequential and spiral acquisition protocol. In addition, the phantom was scannedon a 100 ms EBT-unit (GE-Imatron, San Francisco, California, USA). The Agat-ston and mass calcium scores were determined from the reconstructed imagesusing Syngo CaScoring (Siemens) and AW Smartscore (GE) from five differentcalcium inserts ranging in mass from 8.5 to 78.5 mgCaHA.Results: From EBT-data the Agatston score of the inserts resulted in values be-tween 35 and 305. All results of the MDCT-data were within the error margins,except for the highest Agatston score, which results in a value of 274 ± 18. Thevariation of the Agatston score per insert was 5.8-12.3%. The average mass scoresof the MDCT-data all are within error margins between +7% and -22% of thelisted values with a variation per insert between 2.5-6.9%. The use of a spiral orsequential acquisition mode yielded similar results and no substantial differencebetween the results from the three hospitals was observed.Conclusion: The Agatston score shows a substantially higher variation than themass score, and the use of a spiral or sequential acquisition mode yields similarresults on both scores, independent of the system site.

B-049 11:42

Different reconstruction intervals provide different calcium scores inretrospectively ECG-gated MDCT of the coronaries: Is the highest score thereal score?P. Hunold, J.-S. Sawatzki, T. Schlosser, A. Schmermund, H. Kuehl,J. Barkhausen; Essen/DE ([email protected])

Purpose: To compare the "real" calcium score provided by the highest qualityreconstruction with the calculated scores from other reconstructions.Methods: In 15 patients, coronary calcium was measured using retrospectivelyECG-gated MDCT (16-row scanner; temporal resolution, 210 ms). From eachacquisition 10 reconstructions were calculated at different times throughout thecardiac cycle. For each data set volume (VS) and Agatston scores (AS) wereassessed for each individual coronary artery and the coronary tree. By opticalanalysis the reconstruction with the least motion artifacts was stated the optimalreconstruction by definition; the corresponding scores were considered the "real"scores and compared to scores from other reconstructions.Results: Total VS and AS from the optimal reconstructions ranged from 0 to 811and 0 to 924, respectively. The "real" total VS and AS were distributed over theentire range of scores derived from the different reconstructions: VS: at 51 ± 37%of the score range; median, 53%; range, 0-100%. AS: at 53 ± 30%; median, 58%;range, 0-100%. Comparing "real" VS and AS with the ranges in the individualvessels, a similar distribution over the entire range was seen. The optimal recon-

struction intervals of the individual vessels were: LM and LAD, -30% or -40% in11/15 pts; LCX, -40% and -50% (11/15 pts); RCA, -40%, -50%, and -60% in 14/15 pts.Conclusion: The "real" scores are widely distributed over the range of scoresfrom all different reconstructions since they are highly dependent on the recon-struction interval. The highest score is not necessarily the "real".

B-050 11:51

Reproducibility of coronary calcium measurements using Multi SliceComputed TomographyS. Sabour, F. Atsma, A. Rutten, D.E. Grobbee, W. Mali, M. Prokop, M.L. Bots;Utrecht/NL ([email protected])

Purpose: To assess reproducibility of coronary calcium measurements from Multi-Slice CT (MSCT) images and to evaluate whether different measurement proto-cols, slice thickness and cardiovascular risk factors affect inter and intra- observerreproducibility. A cross-sectional design was used.Methods: The study population comprised 199 healthy postmenopausal women.Coronary calcium was assessed in these women using MSCT (Philips Mx 8000IDT 16). Images were made using 1.5 mm slice thickness and 3.0 mm slice thick-ness. To assess inter and intra-observer reproducibility, the images were read bytwo observers. One observer twice read 52 images. The Agatston score, a vol-ume measurement and a mass measurement were used to score coronary calci-um. Reproducibility was determined by estimation of mean differences, absolutemean differences and intra-class correlation coefficients.Results: The reproducibility of coronary calcium measurements between observ-ers and within observers was excellent with Intra-class correlation coefficients of> 0.95, and small mean and absolute mean differences. Reproducibility of themeasurements was similar for 1.5 mm slices as for 3.0 mm slices, and equal forAgatston, volume and mass measurements. Of the established cardiovascularrisk factors, none was significantly related to measurement error. However, in-crease in absolute calcium level appeared to be related with decrease in meas-urement error.Conclusion: Reproducibility of measurement of coronary calcium using imagesfrom MSCT is excellent, irrespective of slice thickness and type of calcium pa-rameter. The observations that coronary calcium levels are related to measure-ment error may have consequences for relations under study.

10:30 - 12:00 Room G

Contrast Media

SS 206Safety and effectivenessModerators:O. Clément; Paris/FRT. Gleeson; Dublin/IE

B-051 10:30

An evidence-based review of contrast-induced nephropathy (CIN)C.R. Becker1, F. Stacul2, J. Tumlin3, A. Adam4, C. Davidson5, N. Lameire6,P.A. McCullough7; 1Munich/DE, 2Trieste/IT, 3Atlanta, GA/US, 4London/UK,5Chicago, IL/US, 6Ghent/BE, 7Royal Oak, MI/US([email protected])

Purpose: In view of the clinical need for clear guidance on the issue of contrast-induced nephropathy (CIN), a systematic review of the literature on CIN wasundertaken as a basis for consensus statements on prevention.Methods and Materials: An international multidisciplinary expert panel met inNovember 2004 to agree the strategy and scope of the review. The EMBASE andMEDLINE databases were searched for English language papers on CIN pub-lished between 1966 and February 2005, using comprehensive search terms foriodinated contrast media and renal disease/failure/dysfunction. All abstracts werereviewed by 2 independent reviewers, following the Cochrane procedure. Datafrom studies that fulfilled the selection criteria were entered into data tables andused to prepare qualitative summary documents. The evidence was reviewed byall members of the group independently and statements agreed by a consensusprocess.Results: A total of 4370 references were retrieved of which 865 were consideredpotentially relevant. Consensus statements were agreed covering epidemiologyand pathogenesis of CIN, baseline renal function measurement, risk assessment,identification of high-risk patients, contrast medium use and preventive strate-gies.

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 149C D E FA G

Conclusion: Evidence-based statements and practical strategies that can beadopted in routine practice were agreed by the panel and will be presented.

B-052 10:39

Higher incidence of renal failure in 77,848 patients undergoing coronaryinterventions of Swedish hospitals using iso-osmolar contrast mediacompared to low-osmolar contrast mediaP. Liss1, P.B. Persson2, P. Hansell1, B. Lagerqvist1; 1Uppsala/SE, 2Berlin/DE([email protected])

Background: Contrast induced nephropathy (CIN) occurs as a complication af-ter radiographic examination using iodinated radiocontrast medium (CM). In or-der to evaluate the incidence of CIN in clinical practice we analysed retrospectivedata from the incidence of CIN and dialysis.Methods: Nineteen Swedish hospitals were examined that principally use a sin-gle CM for Percutaneous Coronary Interventions (PCI) and coronary angiogra-phy during 1990-2003. Iodixanol was used in 47,543 patients, ioxaglate in 24,577and iohexol in 5,728 subjects. Patients were evaluated from the Swedish Coro-nary Angiography and Angioplasty Registry (SCAAR) and rehospitalizations witha diagnosis of renal failure and/or treated for dialysis in the Swedish all-national"Hospital Discharge Register".Results: Incidence of the diagnosis renal failure within 12 months was greatestfor patients receiving iodixanol (2.3%) followed by ioxaglate or iohexol treatedpatients (1.2% and 1.2% respectively, p < 0.001). When adjusted for gender, age,diabetes, previous PCI and previous renal insufficiency the hazard ratio for iodix-anol treated patients remained significantly higher than ioxaglate but not thaniohexol. The incidence of treatment with dialysis within one year after the inter-vention was 0.2% after injection of ioxaglate and iodixanol (ns).Conclusion: In this register study a renal-failure diagnosis but not the use ofdialysis, within one year, was found more often after use of iodixanol than ioxaglate.A prospective study is required to rule out unknown cofactors and to confirm theassumed causal relationship of renal failure and type of CM.

B-053 10:48

Safety of iobitridol in the general population and at-risk patientsT.J. Vogl1, E. Honold2, M. Wolf3, R. Hammerstingl1; 1Frankfurt a. Main/DE,2Sulzbach/Taunus/DE, 3Püttlingen/DE ([email protected])

Learning Objective: The purpose of this post-marketing surveillance study wasto review the rate of adverse events after contrast medium administration in thegeneral population and at-risk patients under daily practice conditions.Methods and Material: 210 radiologists took part in this study; various X-rayexaminations were carried out in 52,057 patients. To document the safety of iobi-tridol in routine use, all patients who underwent X-ray examinations were includ-ed in the study. Exclusion criteria were contraindications listed in the locallyapproved summary of product characteristics.Results: The adverse event rate observed was 0.96% (at-risk patients 1.39%)and the rate of serious adverse events was 0.044% in all patients (at-risk patients0.057%). Adverse events occurred more often in women than in men (p < 0.001).In patients who had previously reacted to a contrast medium, adverse eventswere reported in 3.43% of all cases with generally mild to moderate symptoms. In47.76% of these patients, a premedication was administered. However, therewas no difference in the frequency of adverse events and serious adverse eventswhether premedicated or not (p = 0.311 and p = 0.295, respectively).Conclusion: Iobitridol was well-tolerated in 99.04% of cases. In the group of at-risk patients, the rate was 98.61%.

B-054 10:57

Prediction of GFR with Cockcroft-Gault's formula based on the NORIPplasma-creatinine reference intervals and using various body weightexpressionsU. Nyman1, J. Björk2, G. Sterner3, A. Grubb2; 1Trelleborg/SE, 2Lund/SE,3Malmö/SE ([email protected])

Purpose: Prediction equations, e.g. Cockcroft-Gault (CG) based on plasma-cre-atinine, weight, age and gender, should replace plasma-creatinine alone to eval-uate renal function (GFR) prior to contrast examinations (http://www.kidney.org/professionals/kdoqi/guidelines_ckd/toc.htm). This study analyses the accuracyof CG with various body weight expressions and the Nordic Reference IntervalProject (NORIP) recommendations of plasma-creatinine reference values forharmonization between laboratories (women: 50-90; men: 60-100 µmol/L).Methods: Data from 850 patients (18-93 years), consecutively referred for deter-mination of GFR by an invasive "gold standard" (iohexol-clearance), were used

for the analysis. CG based on total body weight (TBW), lean body mass accord-ing to Hume (LBM), ideal body weight according to Devine (IBW) or adjustedbody weight (ABW=lowest weight of TBW and IBW). Calculations of bias (medi-an percent error) and accuracy (percent estimates within ±30 and ±50% of meas-ured GFR) were used in the comparisons.Results: CG-TBW resulted in a 13% bias with 70% of the estimates within ±30%and 86% within ±50% of measured GFR. CG-LBM resulted in a -20% bias andonly 65% of the estimates within ±30% of measured GFR. CG-IBW or -ABWcaused no significant bias with the best accuracy for CG-ABW; 79% of the esti-mates within ±30% and 95% within ±50% of measured GFR (significantly betterthan CG-TBW, p≤ 0.001). CG-ABW yielded no overestimation of GFR in patientswith high BMI and only a weak underestimation in elderly.Conclusion: CG-ABW applied on the NORIP plasma-creatinine reference inter-vals resulted in an accuracy regarded as sufficient for good clinical decision-making according to NKF K/DOQI GUIDELINES.

B-055 11:06

Histopathological changes after renal arteriography using iodine andgadolinium contrast media in an ischemic porcine kidneyB.A. Elmståhl1, P. Leander1, D. Grant2, R. Doughty2, U. Nyman3, C.-M. Chai1,K. Golman1, T. Almén1; 1Malmö/SE, 2Oslo/NO, 3Trelleborg/SE([email protected])

Purpose: To investigate histopathological changes in the kidney caused by gado-linium contrast media (Gd-CM) in comparison with iodine contrast media (I-CM)after renal X-ray arteriography in an ischemic porcine model.Methods: Test solutions: 0.5M gadopentetate (1.96 Osm/kg), 0.5M gadodiamide(0.78 Osm/kg), 0.5M iohexol (190 mg I/mL, 0.42 Osm/kg) and mannitol solutionsiso-osmotic to each of these CM solutions, 1.0M gadobutrol (1.6 Osm/kg), iopro-mide (150 mg I/mL, 0.34 Osm/kg) iodixanol (150 and 320 mg I/mL, 0.29 Osm/kg)0.18M iohexol (70 mg I/mL, 0.29 Osm/kg) and saline (0.29 Osm/kg). Each testsolution (3 mL/kg) was injected into a balloon-occluded (10-minutes) right renalartery after left-sided nephrectomy in eight pigs. The kidneys were evaluatedhistomorphologically.Results: The main histological changes were proximal tubule necrosis and vac-uolation. The proximal tubule necrosis was statistically significant higher in thekidneys injected with 0.5M gadopentetate, mannitol solution iso-osmotic to 0.5Mgadopentetate and 1.0M gadobutrol compared to all other groups (p < 0.001).There was no proximal tubule necrosis after injection of iodixanol 150 mg I/mL oriodixanol 320 mg I/mL. The tubule vacuolation was statistically significant higherin the 0.5M gadodiamide group compared to all other groups (p ≤ 0.001).Conclusion: Gd-CM with high osmolality (gadopentetate and gadobutrol) causedtubular necrosis representing acute renal failure, clinically corresponding to CM-induced nephropathy (CIN). High-osmolar Gd-CM are more nephrotoxic than I-CM iso-osmolar to plasma. Gadodiamide with relatively high osmolality causedvacuolation of proximal tubule similar to I-CM. However, the clinical significanceof tubule vacuolation is unclear.

B-056 11:15

Effect of contrast media on regional renal hemodynamics and oxygenationin the rat. An evaluation of the medullary blood flow using two methodssimultaneouslyP. Liss, F. Palm, P.-O. Carlsson, A. Fasching, P. Hansell; Uppsala/SE([email protected])

Background: The renal medulla has been implicated as a key target for contrastmedia-induced nephropathy (CIN). The issue whether contrast media (CM) de-crease or increase renal medullary blood flow still remains controversial. Ourlaboratory consistently finds reduction in perfusion after CM administration, whileanother group, using iothalamate, has shown elevations in medullary perfusionusing the same technique. To reconcile this issue, two different methods wereapplied simultaneously to estimate the outer medullary blood flow (OMBF) re-sponse to injection of the CM iothalamate.Methods: The OMBF response was simultaneously estimated using laser-Dop-pler flowmetry and hydrogen gas washout (microelectrodes). Outer medullaryoxygen tension (pO2) was measured with Clark type microelectrodes. Iothala-mate was injected intravenously at a dose of 1600 mg iodine/kg body weightduring 2 minutes.Results: CM induced a transient 28% decrease in OMBF as measured with thelaser-Doppler. The hydrogen gas washout rate was reduced by 50% indicating areduced perfusion. CM induced a transient 60% reduction in pO2 while renal fluidand electrolyte excretion increased several fold.Conclusion: The CM iothalamate reduces outer medullary perfusion as estimat-

Scientific Sessions

150 C D E FBA G

ed by two different techniques applied simultaneously. The pO2 in the same re-gion was also reduced. The results strengthen the theory that CIN may have itsorigin in a reduced perfusion/oxygenation of an area which already during nor-mal physiological conditions works at the verge of hypoxia. Previous controver-sies regarding the effect of iothalamate on outer medullary blood flow can mostlikely be explained by extreme dosage and injection rates greatly exceeding clin-ical relevance.

B-057 11:24

Is the standard use of N-acetylcysteine (NAC) in preventing contrast-relatednephropathy evidence based?M. Gawenda, A. Moeller, G. Wassmer, J. Brunkwall; Cologne/DE([email protected])

Purpose: The need for contrast-based procedures is rising, especially in endo-vascular surgery, as is the incidence of postcontrast-renal failure. Clinical trialsevaluating N-acetylcysteine (NAC) for the prevention of contrast-induced neph-ropathy (CIN) have reported mixed results. Although previous meta-analyses haveconcluded that NAC is beneficial, the recent availability of additional trials callsfor reassessment of evidence.Methods: Computerized bibliographic search until June 2005, to identify rand-omized clinical trials that evaluate NAC and hydration with hydration alone forprevention of CIN.Results: 29 prospective studies (3629 patients) were found and further analysed.Contrast media was given mostly due to coronary angiographies, exceptionallyperipheral angiographies and/or computed tomographies were performed. Thekind and volume of contrast media varied (mean volume; min-max [in ml]: 145.4;75-230). Moreover, the trials showed variability due to application of NAC. Mostly,NAC was given orally, in three trials NAC was given intravenous.The pooled data demonstrated a relative risk reduction (RR = 0.67, 95% CI =0.47-0.95) for the application of NAC. However, by looking at the results of thesingle trials, 14 studies lacked to demonstrate any preventive effect of NAC.Conclusion: Although NAC is inexpensive, easy to use, and has a favourableadverse event profile, the analysis suggested that data are too inconsistent togive a definitive recommendation for the use of NAC in preventing contrast-in-duced nephropathy. The time has come for a large multi-center trial powered tolook for clinical end points such as the need for renal replacement therapy, death,and delayed hospital discharge.

B-058 11:33

Spurious hypocalcemia caused by interference of MR contrast agents withserum calcium colorimetric measurement techniqueM. Nguyen, C. Klassen, K. Kareti, O. Batmunh, A. Siuciak, N. Wilke;Jacksonville, FL/US ([email protected])

Purpose: Decomplexation of gadolinium-chelator complexes has been shown tocause spuriously low serum calcium measurement by ortho-cresolphthalein (OCP)colorimetric assay. Spurious hypocalcemic lab values can trigger unnecessarytreatments in oncology patients undergoing treatment for hypercalcemia or inpatients at risk for arrthymia or seizure disorders. The objective of this study wasto retrospectively investigate the incidence of spurious hypocalcemia caused byGd-DTPA and Gd-DTPA-BMA in vivo in cohorts of patients.Methods: Consecutive patients receiving Gd-DTPA (n = 647) and Gd-DTPA-BMA(n = 632) for routine MR imaging studies were included in the study if serumcalcium measurements were performed both before and within 24 hours of re-ceiving the MR contrast agents. Serum calcium level was measured using OCPassay with an automated analyzer (Modular Analytic Systems). Two consecutiveprecontrast serum calcium levels within 15 days closest to contrast administra-tion were used for statistic determination of whether the postcontrast calciumlevel is significantly different from typical variations.Results: The mean change in serum Ca post Gd-DTPA-BMA was -0.5 mg/dlcompared with 0.1 mg/dl precontrast (p < 0.005). The mean change in serum Capost Gd-DTPA was 0.1 mg/dl compared with 0.09 mg/dl precontrast administra-tion (p = 0.12). There was 9% incidence of decreasing Ca level greater than 2 mg/dl with Gd-DTPA-BMA compared with 2% for Gd-DTPA (X2 = 5.64, p < 0.025).Conclusion: There was significant increase in incidence of spurious hypocal-cemia after Gd-DTPA-BMA administration compared to Gd-DTPA. Physicians mustbe aware of this possible effect when reviewing hypocalcemic labs and to exam-ine patients for true clinical signs of hypocalcemia.

B-059 11:42

Brain MR imaging at 3 Tesla: Efficacy of gadobenate dimeglumine (Gd-BOPTA) in comparison to 1.5 Tesla and a standard gadolinium chelateB.J. Wintersperger1, V.M. Runge2, J. Biswas2, C.U. Herborn3, S.O. Schönberg1,M.F. Reiser1; 1Munich/DE, 2Temple, TX/US, 3Essen/DE([email protected])

Purpose: The purpose of the current study was to evaluate enhancement prop-erties of gadobenate dimeglumine at 3 T and to compare enhancement proper-ties of gadobenate dimeglumine at 3 T with a standard gadolinium chelate in arat brain tumor model.Methods and Material: Twelve CDF Fischer 344 rats with implanted gliomas(C6/LacZ) underwent two MR scans in a randomized fashion either comparinggadobenate dimeglumine (MultiHance) at 1.5 T and 3 T (n = 7) or comparinggadobenate dimeglumine and gadopentetate dimeglumine (Gd-DTPA) at 3 T(n = 5). Matched T1-weighted spin-echo techniques at 1.5 T and 3 T were used(TR 500 ms, TE 15 ms; BW 125 Hz/Px) at a spatial resolution of 0.2 x 0.2 mm2

and 2 mm slice thickness. Signal-to-noise ratio (SNR), contrast-to-noise ratio(CNR) and lesion enhancement (LE) were evaluated for all subgroups and com-pared. Harvested tumors were subsequently evaluated by histopathology.Results: Comparison of Gd-BOPTA at 3 T and 1.5 T revealed significant increasesregarding SNR, CNR and LE at 3 T. Average increases in brain and tumor SNRwere 93% (P < 0.0001) and 92% (P < 0.0001), respectively. CNR increased by121% (P < 0.0001). Comparison of Gd-BOPTA and Gd-DTPA at 3 T demonstrat-ed significantly higher CNR and LE with Gd-BOPTA, with the increase over Gd-DTPA being 35% (P = 0.002) and 44% (P = 0.03), respectively.Conclusion: The use of Gd-BOPTA leads to significantly higher lesion enhance-ment and tumor CNR compared with a standard Gd-chelate at 3 T. Gd-BOPTAmaintains superior enhancement properties at 3 T, as at 1.5 T, in neurologicalMR imaging.

B-060 11:51 !Bioeffects of pulse inversion ultrasound and contrast agent on "in vitro"culture of human lymphocytes TCD4+G. Salvaggio, F.E. Sorrentino, G. Fatta, F. Verderame, A. Campisi, V. Ricotta,C. La Mendola, G. Caruso, R. Lagalla; Palermo/IT ([email protected])

Purpose: To evaluate bioeffects induced by pulse inversion harmonic imaging(PIHI) with different low mechanical index (MI) and different insonation frequen-cies (3.5 MHz and 7.5 MHz) on human lymphocytes (T CD4+) exposed "in vitro"to different concentration of microbubbles (SonoVue), evaluating Interferon-γ (IFN-γ) release.Methods and Materials: Human lymphocytes (T CD4+) "in vitro" were classifiedinto groups as followed: one group as control group; two groups exposed to ultra-sound (3.5 MHz and 7.5 MHz); nine groups exposed to PIHI with different MI(0.01, 0.06 e 0.1) and different microbubbles concentration (0.01, 0.05 e 0.1 ml).The ELISA test was used to evaluate the IFN-γ production.Results: IFN-γ release was significantly lower in the groups exposed to PIHI withcontrast agent with respect to the control group and groups exposed to ultra-sound only. Moreover, IFN-γ release was lower for higher MI and contrast agentconcentration.Conclusion: Results suggest that PIHI with contrast agent can lead to cell dam-age. The cell damage is proportional to MI and microbubbles concentration.

10:30 - 12:00 Room H

Interventional Radiology

SS 209Tumor ablationModerators:J. Cast; Hull/UKY. Ni; Leuven/BE

B-061 10:30

Determination of treatment response of malignant liver tumors toradiofrequency ablation: Comparison of PET/CT and MR imagingH. Kuehl, H. Stergar, P. Veit, S. Rosenbaum, G. Antoch, J. Stattaus,A. Bockisch, J. Barkhausen; Essen/DE ([email protected])

Purpose: Accuracy of morphologic imaging (CT, MR imaging) after ablation ofliver tumors is impaired by rim-like enhancement of the ablation margin. Our study

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 151C D E FA G

aimed to compare PET/CT and MR imaging to determine accuracy at detectingpersisting or recurrent tumor after RF-ablation (RFA).Methods and Materials: 20 patients (mean age 63 years) with 28 hepatic malig-nancies (HCC 8, metastases 12) were examined by PET/CT prior to ablation.Post-interventional PET/CT was performed 24 h after ablation in patients withPET positive lesions and was repeated after 1, 3, 6 and 12 months. In addition allpatients were examined by MR imaging at the same time points. Images wereread by different reader groups blinded to the results of the other imaging modal-ities. The accuracy of the different modalities was assessed.Results: 9/20 patients had 16 liver PET positive lesions and were included in thisstudy. The mean follow-up time was 11 months. 10 lesions developed local recur-rence. 9/10 recurrences were identified by PET/CT, whereas MR imaging detect-ed only 6/10 lesions. The sensitivities and specificities for PET/CT were 82% and100%, for MR imaging 69.3% and 100%, respectively. The accuracy was meas-ured 91% for PET/CT and 89.7 % for MR imaging.Conclusion: PET/CT proved to be more accurate compared with MR imaging atdetecting local tumor recurrence after RFA. However, very small areas of residu-al tumor are difficult to detect with both modalities.

B-062 10:39

Percutaneous radiofrequency ablation treatment for palliation of recurrentrectal cancerT. Donnianni, E. Tedeschi, E.M. Covelli, G. Belfiore; Caserta/IT([email protected])

Purpose: To assess the role of CT-guided radiofrequency ablation (RFA) in thepalliative treatment of loco-regional recurrencies from rectal adenocarcinoma.Methods and Materials: The population consisted of 7 adult patients, previouslytreated with Miles surgery and radiotherapy, each with loco-regional recurrentrectal adenocarcinoma, who complained of persistent local pain. 3 lesions (maxdiameter: < 3 cm) underwent percutaneous RFA with single positioning of a 17-Gcooled-tip single electrode and a 200 W RF generator (Radionics, USA). In the 4remaining lesions (max diameter: > 3 cm) RFA was performed with single posi-tioning of a multitined expandable 15-G electrode and a 1500X StarBurst RFgenerator (RITA Medical Systems, USA). In all cases, mild sedoanalgesia wasobtained; follow-up was performed with clinical evaluation and contrast-enhancedCT at 3, 6, 12, 18 and 24 months.Results: All patients tolerated the procedure well. At 3 month follow-up, in 5/7cases, pain disappeared completely, while in 2 cases it remained unchanged. AtCT, the 4 lesions > 3 cm showed marked size reduction, while the 3 lesions < 3 cmwere reduced in two cases and disappeared in the remaining case. These find-ings were confirmed at 12 months. At 2 years, 4 patient are still alive and symp-tom-free, 1 is alive with mild loco-regional pain, 1 is dead and 1 is lost to follow-up.Conclusion: In our experience, RFA is a feasible therapeutic option for solidrecurrences of rectal adenocarcinoma with good efficacy on loco-regional pain.Furthermore, a significant cytoreduction can be obtained, which may be exploit-ed advantageously by further treatment modalities.

B-063 10:48

MR-guided laserinduced thermotherapy (LITT) of liver metastases of gastriccancerM.G. Mack, K. Eichler, T. Lehnert, D. Proschek, J. Flöter, J.O. Balzer, T.J. Vogl;Frankfurt a. Main/DE ([email protected])

Purpose: To analyze of the potential of LITT for the treatment of liver metastasesof gastric cancer.Methods and Materials: 58 metastases in 27 patients (22 male, 5 female, aver-age age 61 years, range 36-86 years) were treated. The mean number of treatedlesions per patient was 2.1 (range 1-5). Survival rates (SR) were calculated us-ing the Kaplan-Meier method. 7.4% of the patients had recurrent metastasesafter surgery, 48.1% metastases in both liver lobes, 25.9% refused surgical re-section, 7.4% had general contraindications for surgery, and 11.1% had metas-tases in difficult locations for surgery. 38.9% of the lesions were < 2 cm, 35.2%were 2-3 cm, 14.8% were 3-4 cm, and 11.1% were > 4 cm.Results: All ablations were performed under local anesthesia. The mean SR was2.3 years (95%-CI 1.5-3.1 years, median SR 1.2 years, 1-year SR 60%, 2-yearSR 36%, 5-year SR 22%) after the first LITT treatment and 2.7 years after thediagnosis of the metastases, which were treated with LITT (95%-CI; 1.9-3.5 years,median survival 1.4 years, 1-year SR 81%, 2-year SR 41%, 5-year SR 22%). Themean size of the lesion was 2.4 cm (range 0.7-7 cm, mean volume 13.9 ml), themean size of necrosis was 4.9 cm (range 2.0-8.0 cm, mean volume 50 ml). Themean applied energy was 85.1 KJ (median 56.9 KJ, range 3.1-258.8 KJ)Conclusion: Local ablation of liver metastases of gastric cancer using MR-guid-ed LITT is a safe minimally invasive treatment option in well selected cases.

B-064 10:57

CT-evaluation of in vivo liver regeneration post MR-guided laser inducedthermotherapy of liver metastases in comparison to surgical resection andin correlation with regeneration factors TNF-a and HGFT. Lehnert, A. Gazis, J.O. Balzer, A. Thalhammer, M.G. Mack, T.J. Vogl;Frankfurt/Main/DE ([email protected])

Purpose: To evaluate and compare the complexity of liver regeneration post liverresection and MR-guided laser induced thermotherapy (LITT) in correlation withthe regeneration factors TNF-α and HGF.Methods and Material: 40 patients were evaluated, 21 patients post surgicalliver intervention and 19 patients post LITT. Over a period of 12 months, the ef-fects of the therapy on normal liver parenchyma and possible liver regenerationafter surgical or interventional radiologic therapy were evaluated. Volumetric ex-amination (CT-volumetry) of non-involved liver parenchyma, tumor/metastasis preand post intervention and regional growth of liver segments were examined andthe influence of regeneration factors TNF-α and HGF evaluated.Results: After 5 months, an increase in liver volume up to the original liver vol-ume could be observed in 25.6% of all patients post surgical resection and in8.9% of patients after LITT treatment. In correlation of original liver volume, anincrease of liver volume by 2.8% after 2 weeks and 6.8% after 6 weeks could beevaluated. By comparison, surgical resection led to an increase of liver volumeby 3.7% after 2 weeks and 10.7% after 6 weeks. A significant increase of HGFcould be confirmed with loss of liver parenchyma resulting from surgery (on aver-age 245% 2 weeks after intervention) and LITT induction (on average 189% 2weeks after intervention).Conclusion: MR-guided laser induced thermotherapy (LITT) basically stimulatesliver regeneration and causes a significant increase of cytokine HGF, on average2.3 weeks post intervention. No affect on TNF-α can be observed.

B-065 11:06

Over 12 years experience of MR-guided laser induced thermotherapy (LITT)of colorectal carcinoma metastases in the liverM.G. Mack, K. Eichler, D. Proschek, T. Lehnert, J. Flöter, T.J. Vogl;Frankfurt a. Main/DE ([email protected])

Purpose: To evaluate long term results for MR-guided LITT for the treatment ofcolorectal liver metastases.Methods and Materials: MR-guided LITT was performed in 870 patients (meanage 61.6 years) with 2591 liver metastases of colorectal cancer origin, between1993 and 2005. Survival rates were calculated using the Kaplan-Meier method.33.3% of the lesions were 2 cm or less in diameter (mean applied energy 54.5KJ, range 8.2-249.7 KJ), 33.3% were between 2 and 3 cm (mean applied energy94.9 KJ, range 11.6-361.4 KJ), 18.1% were between 3 and 4 cm (mean appliedenergy 133.45 KJ, range 20.2-452.6 KJ), and 15.6% were larger than 4 cm inmaximum diameter (mean applied energy 189.1 KJ, range 10.4-515.0 KJ). 670patients were treated with curative intention, 270 patients were treated with pal-liative intention.Results: The mean survival rate for all treated patients, starting the calculation atthe date of diagnosis of the metastases which were treated with LITT, was 3.5years (95% confidence interval: 3.2-3.7 years, 1 year survival 93%, 2 year surviv-al 70%, 3 year survival 43%, 5 year survival 18%). In the curative patient groupthe mean survival was 3.7 years (95% confidence interval: 3.5-4.0 years, 1 yearsurvival 95%, 2 year survival 74%, 3 year survival 48%, 5 year survival 21%). Inthe palliative group mean survival was 2.5 years (95% CI: 2.2-2.8 years).Conclusion: MR-guided LITT is an excellent treatment option for liver metastas-es of colorectal carcinoma.

B-066 11:15

Complications after percutaneous radiofrequency ablation (RFA) of primaryand secondary hepatic tumors: A single centre experienceR.-T. Hoffmann1, T. Jakobs1, C. Trumm1, T. Helmberger2, M.F. Reiser1;1Munich/DE, 2Luebeck/DE ([email protected])

Purpose: The aim of this study was the assessment of complication rate afterlocal thermal ablative treatment of focal hepatic tumors.Methods and Materials: Within 52 months 422 patients (252 male, 170 female,age 26-83 yrs) were treated. 115 patients were suffering from HCC, 134 hadmetastases of colorectal cancer, 125 of breast cancer and 48 of other tumors.RFA was performed under conscious sedation and local anaesthesia using CT-fluoroscopy. All patients were observed for 24 hours on a ward and a control CTscan was mandatory before leaving the hospital to rule out an early adverseevent. Keeping in touch with the patients and their referring physicians enabled

Scientific Sessions

152 C D E FBA G

us to find out whether delayed complications occurred. Adverse events were di-vided into major and minor complications (further therapy/no further therapymandatory).Results: 905 procedures were performed in 442 patients. The size of the treatedtumors ranged from 5 to 60 mm (mean 28 mm). Major complications were diag-nosed after 10 of 905 treatments (3 pneumothorax, 2 bleeds, 4 abscesses, 1 ste-nosis of the common bile duct) and minor complications were diagnosed after 37of 905 treatments (24 small capsule hematoma, 10 prolonged pain, 3 small pneu-mothorax). No treatment related death happened.Conclusion: RFA with an overall complication rate in our series of about 5 % anda major complication rate of about 1% is a safe therapeutic option for patientssuffering from primary or secondary liver tumors.

B-067 11:24

Monopolar saline-enhanced radiofrequency ablation of malignant livertumors: Two year experienceZ. Brnic, B. Brkljacic; Zagreb/HR ([email protected])

Purpose: To evaluate technical success, safety and short-term patient outcomeof radiofrequency ablation (RFA) of malignant hepatic tumors under local an-esthesia and sedation vs. general anesthesia.Methods and Materials: During 2 years a total of 61 patients underwent saline-enhanced RFA of liver tumors (59 secondary, 2 primary; mean size 41 mm, range15-80, 1-4 nodules treated per session); 34 patients under local anesthesia withsedation (group 1) and 27 under general anesthesia (group 2).Results: Mean effective ablation time (min), ablation volume index (ablated vol-ume/total nodule volume), % of patients with complete tumor ablation and % ofpatients free of recurence after 6-12 months in groups 1 vs. 2 were: 12 vs. 22; 61vs. 94, 48 vs. 70, and 29 vs. 48, respectively. Incomplete tumor ablation and re-currence were more frequent in patients with larger tumors and in group 1. Apartfrom minor complications in 13%, only 5% of patients had major complication (1liver abscess, 1 track seeding, 1 gallbladder injury). No procedure-related mor-tality was observed.Conclusion: RFA is a safe and efficient technique for local control of growth ofliver tumors < 5 cm in diameter. Pain during procedure is a limiting factor andRFA under general anesthesia results in a better therapeutic response.

B-068 11:33

Additional concentric expansion of RF-coagulation volume: Efficacy ofhypertonic saline injection in area of vaporization around the electrode: Exvivo porcine liver studyA. Shimizu, H. Ishizaka, T. Tatezawa, S. Awata; Maebashi/JP([email protected])

Purpose: When coagulation volume is insufficient in initial radiofrequency (RF)ablation additional concentric expansion of coagulation would be an ideal solu-tion. However vaporization occuring around the electrode after RF applicationthat hinders current delivery limits further coagulation. Whereas hypertonic sa-line injection prior to RF ablation is known to increase current conductivity andthereby enlarge coagulation volume. We tested whether hypertonic saline injec-tion in the area of vaporization, followed by additional RF application could en-large the coagulation volume.Methods and Materials: RF coagulation of excised porcine liver was performedin five different areas. First an RF electrode (Cool-tip) was coaxially placed througha cannula and coagulated a maximal area at output power of 40 W. Temperatureof the liver near the electrode was continuously measured by a thermocoupleand the margin of maximal coagulation was determined as a 49 °C area after15 minutes. The RF electrode was then removed, 1.5 mL of 36% saline was in-fused through a coaxially placed PEI needle and RF application was resumed.Margin of maximal coagulation was then determined in the same way. Finally,coagulation areas were observed on cross-sections.Results: Enhancement of coagulation volume (by a distance of 4-15 mm; mean:9 mm from the electrode) was confirmed in all cases after injection of 36% saline.The coagulation area was left-right symmetrical with a margin immediately adja-cent to the thermocouple.Conclusion: Injection of a small amount of 36% saline in the area of vaporizationfollowed by additional RF application can concentrically expand the coagulationvolume, simplifying the treatment of larger lesions.

B-069 11:42 !Combination of percutaneous radiofrequency ablation and ethanol injectionfor renal tumours: Single center early experinceJ.P. Morales, T. Sabharwal, M. Georganas, R. Salter, R. Dourado, D. Hodgson,T. O'Brien, A. Adam; London/UK ([email protected])

Purpose: To evaluate clinical experience of percutaneous ablation of renal cellcarcinoma using combined radiofrequency coagulation and ethanol injection.Methods and Materials: Nine patients (5 males, mean 68 years) presented withrenal tumours. The tumour size was < 2 cm in 3, 2-3 cm in 4 and 3-4 cm in 3. Allpatients were either considered unfit for surgery or had refused it. Absolute ethanol(0.5-3 ml, mean 1.4 ml) was injected into the tumour immediately before treatmentwith radiofrequency. Radionics cool-tip electrodes were used in 8 patients and aRita electrode in 1. CT follow-up was carried out the day after the procedure andthree monthly thereafter for the first year. Follow-up 4-38 months, mean 20 months.Results: Complete coagulation necrosis was observed on CT in six patients af-ter the first procedure. Two patients were retreated because of enhancing areason CT, which may have represented residual disease. The mean hospital staywas 2 days. All the patients are well, except for one patient who has developed acolonic carcinoma. One patient had a subcapsular haematoma. Two patients hadtransient loin pain, treated conservatively.Conclusion: Combined therapy of percutaneous radiofrequency ablation andethanol injection of renal tumours is an effective method of treatment, which de-stroys the tumour. It is tolerated well and is associated with few complications.

B-070 11:51

Laser-induced thermal ablation of lung metastases: Follow-up imaging andfeasibilityC. Rosenberg, C. Weigel, S. Langner, C.-P. Fröhlich, N. Hosten; Greifswald/DE([email protected])

Purpose: Laser-induced thermal ablation of pulmonary metastases as a newoption in multimodal cancer therapy shows first technical and clinical results inan ongoing study.Methods and Materials: In 42 patients 64 lesions were treated (CT fluoroscopy,Monocath 5.5 F water-cooled applicator, MeoMedical, Germany; 14 W for up to15 min, 1 to 3 applicators per lesion). Different primaries, mainly colorectal carci-nomas (10/42 patients), had caused metastatic disease. Lesion size ranged from8 to 85 mm (median 27.6 mm). Follow-up CT was performed monthly for the firstthree months, then after 3 and 6 months. Median follow-up was 7.6 months.Results: Ablative therapy was technically succesful in 62/64 lesions. In 7 casestherapy was completed after 2 sessions and in 2 cases after 3 sessions. An initialincrease in size of the visualized correlate was seen in 25 cases and secondarycavities in 8 cases. Marked size reduction of the imaging correlate on furtherfollow-up was demonstrated in 32 lesions. These patients were diagnosed suc-cessfully treated. Residual CT findings were: ground-glass opacity or cavitarydefects. Complications were not significantly different to diagnostic lung biopsy.Incidence of recurrent tumor growth was significantly higher in lower segments(7-10), whereas there was no difference between central vs. peripheral location.Conclusion: Laser-induced thermal ablation is a promising option in multimodaltherapy of non-resectable lung metastases. Feasibility is correlated to target lo-cation. Interval reduction for therapy success evaluation is a future task.

10:30 - 12:00 Room I

Vascular

SS 215Carotid imagingModerators:P.E. Andersen; Odense/DKR. Szentgyörgyi; Szeged/HU

B-071 10:30

In vivo characterization and quantification of atherosclerotic carotid plaquewith 16-slice MDCT and histopathological correlationT.T. de Weert, M. Ouhlous, E. Meijering, P.E. Zondervan,M.R.H.M. van Sambeek, A. van der Lugt; Rotterdam/NL([email protected])

Purpose: To assess the potential of in vivo 16-slice MDCT to characterize andquantify atherosclerotic carotid plaque and its components (calcifications, fibroustissue and lipid) with histology as gold standard.

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 153C D E FA G

Method and Materials: Fifteen symptomatic patients with significant carotid ste-nosis (> 70%), underwent CTA with a standardized protocol prior to carotid en-darterectomy (CEA). After surgery CEA specimens were sent for histopathologicalanalysis. From each specimen 3 histologic sections and corresponding MDCTimages were selected. The Hounsfield values (HV) of the major plaque compo-nents (lipid, fibrous tissue and calcifications) were assessed. Based on this study,cut-off points of HV were defined to assess the different plaque components.Subsequently, plaque component quantities were assessed in the MDCT imagesand in the histologic sections.Results: The measured HV were: 657 ± 416 HU, 88 ± 18 HU and 25 ± 19 HU forcalcifications, fibrous tissue and lipid, respectively. The optimal cut-off value todifferentiate lipid from fibrous tissue and fibrous tissue from calcifications was setat 60 HU and 130 HU, respectively. Regression plots showed good correlationbetween MDCT and histology for total plaque, calcified and fibrous areas(R²> 0.73), but poor correlation for lipid areas (R²=0.24). However differentiatingbetween mildly calcified (0-10%) and more severely calcified plaques (> 10%)improved the lipid area correlation in the mildly calcified plaques (R²=0.77).Conclusion: This in vivo study showed that MDCT is capable of characterizingand quantifying total plaque burden, calcifications and fibrous tissue in athero-sclerotic carotid plaque and correlates well with histology. Lipid can only be ade-quately quantified in mildly calcified plaques.

B-072 10:39

Plaque imaging of the internal carotid artery: Correlation of B-flow imagingwith histopathologyM. Reiter, R. Horvat, S. Puchner, W. Rinner, P. Polterauer, J. Lammer, E. Minar,R.A. Bucek; Vienna/AT ([email protected])

Purpose: To evaluate the accuracy of B-flow imaging (BFI) in the characterisa-tion of internal carotid artery (ICA) plaques compared to B-mode and color Dop-pler Imaging (CDI), using histopathology of explanted plaques as the goldstandard.Methods and Materials: 28 consecutive patients with high-grade ICAS sched-uled for carotid endarterectomy were included in the present prospective pilotstudy. BFI, B-mode and CDI images were used to classify the plaques using thestandardised scores of Beletsky and the American Heart Association (AHA), tocalculate the gray-scale-median (GSM) and to detect potential ulcerations; theagreement with histopathology was calculated.Results: Based on the classification of Beletsky, BFI and histopathology agreedin 21 (75%, κ = 0.61, P < 0.001) patients. Corresponding results for the AHA clas-sification revealed inferior agreement for BFI (19 patients/68%, κ = 0.38, P=0.003).The median GSM for BFI and B-mode correlated significantly (r = 0.95, P < 0.001),the median BFI-GSM of symptomatic patients was 24 (IQR 52) and 40 (IQR 51)for asymptomatic patients, respectively. The sensitivity of BFI for the detection ofulcerated plaques was 100% and the specificity 95.8%, corresponding values forCDI were 100% and 92.7%, respectively.Conclusion: BFI provides a moderate correlation with histopathology and relia-ble measurements of the GSM and is highly sensitive for the detection of plaqueulceration.

B-073 10:48

Utility of USPIO as a magnetic resonance dual contrast agent able toidentify both fibrous cap and macrophage burden in-vivo in carotidatheroma: A tool for risk stratificationS.P. Howarth, J. U King Im, T. Tang, M. Graves, P. Kirkpatrick, J.H. Gillard;Cambridge/UK ([email protected])

Background: It is well described that "vulnerable" atheromatous plaque has athin, fibrous cap and extensive lipid core with associated inflammation. This in-flammation can be detected as areas of signal drop on MR imaging using a novelcontrast medium, Sinerem, an Ultrasmall Super-Paramagnetic Iron Oxide (US-PIO).Studies using USPIO have, thus far, not reported its additional T1 effect of en-hancing fibrous cap.Methods: 20 patients with carotid stenosis (10 symptomatic, 10 asymptomatic)were imaged with 1.5 T, pre and 36 hours post USPIO infusion.Images were anonymised, plaque seen was manually segmented into quadrants(CMR tools, London) and focal regions of signal drop post contrast. Signal changewas normalised to adjacent muscle.Results: There were 9 males and 11 females and mean ages were 71 (sympto-matic) and 73 (asymptomatic). Mean stenoses were 82% (symptomatic) and 72%(asymptomatic).Symptomatic patients had significantly more quadrants with signal drop post

USPIO than asymptomatic patients (75% vs 32%, p < 0.01)and more focal re-gions of signal drop (0.68 vs 0.26 regions/image, p < 0.01).Asymptomatic subjects had more quadrants with signal enhancement than thosewith symptoms (68% vs 25%, p < 0.05).Conclusion: Symptomatic plaques had more focal areas of signal drop thanasymptomatic plaques, suggesting that they harboured a greater inflammatoryburden. Asymptomatic plaques showed more enhancement than symptomaticplaques, suggesting thicker fibrous caps and greater stability, although some fo-cal areas of signal drop were noted suggesting vulnerability.If validated by larger studies, USPIO may be a useful dual contrast medium, ableto enhance risk stratification of patients with carotid stenosis, improving selec-tion for intervention.

B-074 10:57

Quantification and morphologic delineation of carotid artery stenosis usingmulti-detector row CT angiography (MD-CTA) and 3D MR angiography (3DCE-MRA) with parallel acquisition technique (PAT)W.H. Flatz, B.B. Ertl-Wagner, D.A. Clevert, G. Meimarakis, R. Brüning,M.F. Reiser; Munich/DE ([email protected])

Purpose: Our objective was to evaluate the diagnostic potential of MD-CTA and3D CE-MRA with parallel acquisition technique in evaluating carotid artery sten-osis.Methods: 30 previously untreated symptomatic patients (24m,6w) suffering fromsuspected carotid artery stenosis were included. Standardized MD-CT-angiogra-phy was performed using 16-row- and 64-row-scanners. 3D-CE-MRA was per-formed using a 1.5 Tesla scanner with 12-channel surface-coils. PAT using SENSEreconstruction with an acceleration factor of 2 was applied. MIP and MPR werecalculated for CTA and MRA. Two blinded neuroradiologists independently quan-titatively assessed the degree of stenosis separately for each modality followingmodified NASCET criteria. Furthermore visual assessment and evaluation ofplaque morphology were performed. Image reading results were compared tosurgical and duplex sonographic findings which served as a standard of refer-ence in all patients. For evaluation of plaque morphology, correlation to histopatho-logic data was performed. T-test and Kappa-test were applied for statisticalevaluation.Results: No significant differences were found between either examination mo-dality and the two standards of reference with regard to the quantitative evalua-tion of carotid artery stenosis (p> 0.05). Mean degree of stenosis was 87.9 ± 9.0%for MD-CTA, 84.8 ± 10.3% for MRA, and 90.2 ± 4.9% for Doppler sonography.Evaluation of the visual impression of the degree of stenosis showed no signifi-cant difference between MD-CTA and MRA, with a tendency toward underesti-mating the degree of stenosis with both modalities. Plaque ulcerations weredetected better with MD-CTA than with MRA (p < 0.05). A high interrater correla-tion was found (kappa=0.91).Conclusion: Vascular imaging using both CE-MDCT angiography and 3D-MR-angiography with parallel imaging technique have a good diagnostic potential inthe evaluation of carotid artery stenoses. Regarding plaque delineation, MD-CTAappears to be superior.

B-075 11:06

Common carotid artery intima-media thickness is independently associatedwith risk for transient ischemic attacks in addition to main risk factors andplaque presenceG. Terzis, A. Kalogeropoulos, A. Chrysanthopoulou, G. Gioldasis,D. Alexopoulos, D. Siablis; Patras/GR ([email protected])

Purpose: Sound evidence supports the independent value of common carotidartery intima-media thickness (CCA-IMT) as a risk factor for stroke. However, theassociation of CCA-IMT with risk for transient ischemic attacks (TIA) is less clear.Moreover, recent data suggest that incidence of stroke is related to CCA-IMTeven in the absence of carotid plaque. We investigated whether CCA-IMT andpresence of carotid plaque are associated with risk for non-embolic TIA afteradjusting for major risk factors in a case-control series.Methods and Materials: Our study population comprised of 80 consecutive pa-tients undergoing carotid ultrasonography during hospitalization for documentedfirst TIA (TIA Working Group Definition 2002) and 245 asymptomatic control sub-jects undergoing screening evaluation. Patients with known cardioembolic sourc-es were excluded. We selected 160 controls (2:1) by propensity score matchingin order to estimate with minimal bias the independent effect of CCA-IMT andplaque presence.Results: Patients and controls were matched for age (66 ± 10 vs. 65 ± 13 yrs,p = 0.492), gender (55.0% vs. 56.2% male, p = 0.854), smoking (28.8% vs. 30.6%,

Scientific Sessions

154 C D E FBA G

p = 0.686), hypertension (57.5% vs. 61.2%, p = 0.576), diabetes (18.8% vs. 20.0%,p = 0.817), and dyslipidemia (39.4% vs. 36.2%, p = 0.639). Compared to controlsubjects (0.68 ± 0.13 mm), CCA-IMT in TIA patients (0.73 ± 0.13 mm) was sig-nificantly higher (p = 0.01). The incidence of plaques did not differ in the twogroups (63.8% vs. 71.2%, p = 0.31). In multivariate analyses, only CCA-IMTemerged as independent predictor (p = 0.01).Conclusion: Increased CCA-IMT is associated with higher risk for non-embolicTIA in addition to established risk factors. This effect is independent of the pres-ence of carotid plaque.

B-076 11:15

64-MDCT angiography in the assessment of ICA stenosis: Comparison withcolor Doppler, power Doppler and B-flow ultrasoundD.-A. Clevert, T. Johnson, M. Stickel, B. Wintersperger, C.R. Becker, P. Flach,T. Strautz, M.F. Reiser; Munich/DE ([email protected])

Purpose: To compare the diagnosic accuracy of 64-slice multislice CT angiogra-phy (64-MDCTA) and color Doppler (CD), power Doppler (PD) and B-flow ultra-sound in the grading of extracranial high degree internal carotid artery (ICA)stenosis.Methods and Material: 35 patient with 43 known or suspected ICA-stenosesunderwent MDCTA (Sensation 64, Siemens). Collimation was 64x0.6 mm, slicethickness 0.75 mm, reconstruction interval 0.5 mm. 120 ml Solutrast 300 (Brac-co) was injected at a flow rate of 5 ml/s. Bolus tracking was used. CT angiogramswere analyzed for caliber irregularities and presence of calcified plaques. In CDsonography, PD and B-Flow (Logiq 9, GE), the degree of stenosis (NASCET-criteria) was evaluated. The results were correlated with CTA, CD, PD, B-Flowand the intraoperative findings.Results: Excellent visualization of vessels was achieved in all cases. 28/43 hadstenoses between 70-80 %, 10/43 had stenoses between 80-90 % and werecorrectly detected in CTA and CD/PD/B-flow. Filiform ICA stenoses were seenwith CD in 3, PD in 4 and with B-Flow in 5 cases. In 3 cases of filiform ICAstenoses, CTA overrated the degree of stenosis in the presence of extensivecalcified plaques. In these cases, no residual lumen could be detected at MDC-TA. Intraoperative findings in these cases showed filiform stenoses.Conclusion: 64-MDCTA is highly accurate in the detection and grading of ICAstenoses between 70-90%. There is an excellent correlation with CD, PD and B-flow ultrasound. In cases of filiform stenoses, ultrasound was more accurate indetection of a residual lumen.

B-077 11:24

The hemodynamic effects of carotid desobstruction in patients withsymptomatic carotid artery stenosis. A CT perfusion studyA. Waaijer1, M.S. van Leeuwen1, M.J.P. van Osch2, B.H. van der Worp1,F.L. Moll1, R.T.H. Lo1, M. Prokop1, W.P.T.M. Mali1; 1Utrecht/NL, 2Utrecht,Leiden/NL ([email protected])

Purpose: To evaluate whether pre-treatment CT perfusion analysis in patientswith symptomatic carotid artery stenosis (CAS) can be used to distinguish be-tween groups of patients that differ with respect to improvement in cerebral per-fusion after carotid intervention.Methods and Materials: Thirty-six patients with symptomatic CAS and no signif-icant contralateral disease were scanned before and 1 month after primary stentplacement (n = 21) or carotid endarterectomy (n = 15) using a multi-detector CTscanner. We calculated the difference in mean transit time (dMTT) and ratio ofcerebral blood flow (rCBF) between symptomatic and asymptomatic hemispheres.We classified patients into three groups according to the pre-treatment dMTT:group 1, dMTT < 1s, group 2, dMTT 1-2s and group 3, dMTT > 2s. We studiedchanges in brain perfusion for the whole group and separately for the three sub-groups by comparing pre- and post treatment dMTT and rCBF using a paired t-test.Results: For the total group mean dMTT decreased from 1.3s before to 0.1s aftertreatment (p < 0.001), mean rCBF increased from 0.85 to 0.93 (p = 0.02). In group1 (n = 16) dMTT and rCBF did not significantly improve, in group 2 (n = 12) onlydMTT improved significantly, and in group 3 (n = 8) both dMTT and rCBF im-proved significantly.Conclusion: Pre-treatment dMTT is a suitable CT perfusion measure to distin-guish between groups of patients with different improvement of cerebral perfusionafter carotid intervention. In 22% of patients significant improvement in dMTTand rCBF was seen. This might hold promise for selecting patients that benefitmost from carotid intervention.

B-078 11:33

Cervical artery dissection: Improved diagnosis by B-flow ultrasoundD.-A. Clevert1, E. Jung2, T. Johnson1, R. Kubale3, S. Schoenberg1, P. Flach1, T.Strautz1, M.F. Reiser1; 1Munich/DE, 2Passau/DE, 3Pirmasens/DE([email protected])

Purpose: The objective of this study was to evaluate the efficiency of B flowultrasound in diagnosing supraaortic vessel dissections compared with other ul-trasound techniques including B scan, color-coded Doppler (CCD) and powerDoppler (PD).Methods and Materials: 88 patients with suspected arterial dissection of theneck vessels were included in this prospective trial. All patients were examinedusing B scan, CCD sonography and PD. After documentation of the diagnoses,the patients were additionally examined by B flow ultrasound. Contrast-enhancedmagnetic resonance angiography (MRA) was used as reference standard in allcases.Results: Dissections of the carotid artery (n = 19) and of the vertebral artery(n = 35) were found in 44 patients. B-flow imaging identified 52 of 54 arterialdissections that were confirmed by MRA. The sensitivity of ultrasound examina-tion using B scan, CCD and PD in detecting all dissections was 95.9% and 99.1%with additional B flow examination. Sensitivity using B flow increased from 98.3%to 100% for carotid dissections and from 93.3% to 94.3% for vertebral arterydissections. The reduced effect of the ultrasound probe angle facilitated imagingof fissures, membranes and low flow phenomena and improved the identificationof low-reflection wall structures. Contrast-enhanced MRA in conjunction with ax-ial T1 and T2 weighted sequences enabled the best visualization of intramuralhematomas.Conclusion: B-flow imaging can significantly increase the sensitivity of ultra-sound examination for dissections of the neck vessels. It also improves the visu-alization of flow within the true and false lumen, of hypoechoic thrombi and ofintramural hematoma.

B-079 11:42

Comparison of contrast-enhanced ultrasound and magnetic resonance- andconventional angiography in the assessment of apparent carotid occlusionon ultrasoundC.J. Hammond1, J. Barber2, J. Patel1, D. Dellagrammaticas1, M. Gough1,S. McPherson1; 1Leeds/UK, 2Bradford/UK ([email protected])

Purpose: Patients with critical ICA stenosis benefit from endarterectomy but thosewith occlusion or long-segment tight stenosis (trickle-flow) do not. Distinguishingbetween these two on conventional ultrasound is problematic as it is insensitiveto very slow flow. The current reference standard is digital-subtraction angiogra-phy (DSA). The reliability of other non-invasive investigations is unproven.Methods and Materials: We prospectively studied the accuracy of contrast-en-hanced ultrasound (CE-USS), 2D time-of-flight (2D-TOF) magnetic resonanceangiography (MRA) and contrast-enhanced MRA (CE-MRA) against DSA in pa-tients with apparent carotid occlusion on conventional ultrasound. Dual reportingof CE-USS (unblinded) and MRA (blinded) was performed.Results: 60 vessels were studied.41 vessels underwent CE-MRA and DSA. On CE-MRA, 2 vessels were patentand 37 were occluded. It was difficult to differentiate critical stenoses from trickle-flow on MRA due to lack of dynamic flow information but no critical stenosis onDSA was misclassified as occluded by CE-MRA. One study was non-diagnosticand there was inter-observer disagreement in one study (trickle-flow on DSA).On 2D-TOF a critical stenoses on DSA was misclassified as occluded.54 vessels underwent CE-USS and DSA. Technical problems prevented confi-dent assessment of vessel patency with CE-USS in 7 patients. A critical stenosison DSA was misclassified as occluded by CE-USS.Conclusion: In the presence of apparent carotid occlusion on conventional ultra-sound, occlusion on CE-MRA can confirm the presence of a non-surgical lesionbut patency does not imply the presence of a surgical lesion. 2D-TOF and CE-USS can misclassify surgical lesions as being occluded.

B-080 11:51

How effective is a 3D-time-resolved echo-shared angiographic technique(TREAT) in the detection of hemodynamically significant carotid arterystenosis (CAS)?H.J. Michaely1, K.A. Herrmann1, H. Kramer1, K. Nael2, M.F. Reiser1,S.O. Schönberg1; 1Munich/DE, 2Los Angeles, CA/US([email protected])

Purpose: To evaluate the efficacy of TREAT in the detection of hemodynamicallysignificant CAS.

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 155C D E FA G

Methods and Material: 33 patients (15 female /18 male, age 69.7 ± 9.2) withsuspected CAS underwent high resolution (HR)-MRA of the supraaortic vessels.HR-MRA was performed on a 1.5 T MR-system (Siemens Sonata) using a fast3D-GRE sequence with parallel imaging (Voxel size 0.9x0.7x0.9 mm³) and 15 mlof gadobutrol. After injection of a 10 ml bolus of gadobutrol, time-resolved MRAwas performed with a 3D-TREAT sequence which combines view sharing andparallel imaging (Voxel size 2.0x1.4x2.0 mm³) and yielded a temporal resolutionof 2.3s.The presence of CAS and the vessel area within and 2 cm after the CAS wasdetermined on multiplanar reconstructions and the degree of stenosis (%) wasmeasured. CAS> 75% were considered hemodynamically significant. For assess-ment of delayed flow, signal-intensity-time curves were generated from the TREATdata. The contralateral side served as reference for the determination of delayedflow in the stenotic vessel.Results: All TREAT measurements were diagnostic. In the 8 patients withoutCAS no flow delay was seen. 11 patients had CAS< 75% and 13 patients hadCAS > 75%. Flow delay was identified in 2/11 patients with CAS< 75% and in 11/14 patients (p = 0.014) with CAS> 75%. Mean flow delay with CAS< 75% was0.3 frames whereas with CAS> 75% the mean flow delay was 1.6 frames.Conclusion: TREAT is an easily applicable and effective tool to monitor flowdelay in stenotic carotid arteries. The identification of delayed flow on TREATimaging is highly indicative of a hemodynamically significant CAS.

10:30 - 12:00 Room K

Genitourinary

SS 207MRI of the female pelvisModerators:L.C.O. Blomqvist; Stockholm/SEM. Kekelidze; Rotterdam/NL

B-081 10:30

MR staging of endometrial cancer: ESUR guidelinesK. Kinkel1, L. Oleaga2, F.M. Danza3, A. Bergman4, R. Forstner5, T.M. Cunha6,J.O. Barentsz7, C. Balleyguier8, B. Brkljacic9; 1Chêne-Bougeries/CH,2Bilbao/ES, 3Roma/IT, 4Uppsala/SE, 5Salzburg/AT, 6Lisbon/PT, 7Nijjmegen/NL,8Villejuif/FR, 9Zagreb/HR ([email protected])

Purpose: To define guidelines and to optimize MR protocols for staging of en-dometrial cancer.Methods and Materials: Critical review and expert consensus of MR imaging(MRI) protocols by the female imaging subcommittee of the European Society ofUrogenital Radiology, from 9 European institutions, and published literature be-tween 1999 and 2005.Results: High field MR imaging with pelvic phase-array coils should include threeT2 weighted sequences in sagittal, axial and coronal oblique orientation (shortand long axis of the uterine body) of the pelvic content. High resolution postcontrast acquisition in the optimal image orientation should always be applied forthe diagnosis of myometrial invasion and acquired between 90 and 150 secondsafter intravenous contrast injection; optimal timing of image acquisition dependson the chosen 2D or 3D technique. If cervical invasion is suspected additionalslice orientation according to the axis of the endocervical channel is recommended.Due to the limited sensitivity of MR imaging to detect lymph node metastasisretroperitoneal lymph node screening with pre-contrast sequences is optional.Possible lymph node invasion should be suggested indirectly based on tumorgrade at endometrial tissue sampling and myometrial and cervical invasion de-tected at MR imaging.Conclusion: Expert consensus and literature review lead to an optimized MRimaging protocol to stage endometrial cancer.

B-082 10:39

MR imaging of the female pelvis at 3 T compared with those obtained at1.5 T: Evaluation on high-resolution T2WI, HASTE, and diffusion weightedimageM. Kataoka, A. Kido, H. Isoda, S. Umeoka, K. Tamai, T. Koyama, Y. Maetani,Y. Nakamoto, K. Togashi; Kyoto/JP ([email protected])

Purpose: Despite excellent contrast /spatial resolution at 3 T Magnetic Reso-nance imaging (MRI) in neuro-imaging, its potential for body area remains con-troversial. This study aimed to evaluate MR imaging of the female pelvis at 3 T/

1.5 T on high-resolution T2 weighed image (T2WI), HASTE, and Diffusion weightedimage (DWI).Material and Methods: MR images of 19 healthy women were obtained by 3 T/1.5 T (Trio/Symphony:Siemens) MR unit using phased array coil with the follow-ing imaging parameters; T2WI (Axial/Sagittal): TR/TE=3200-3600/94-103 ms,FOV/voxel size = 260x195/0.8x0.6x2.0 mm, NEX=2. HASTE (sagittal): TR/TE=3000/100 ms, FOV/voxel size = 250x250/1x1x5 mm, NEX=1. DWI (axial): TR/TE=2200/80 mm, FOV/voxel size = 200x200/2.1x1.6x3.0 mm, NEX=3,b=0.500.1000 s/mm2. Parallel imaging (sense factor:2) was used.Qualitative evaluations of image quality (overall, uterus, uterine zonal structure,cervix, ovary) and artifacts were made independently by two radiologists using a5-point scale and mean scores were used for analysis. Quantitative evaluationswere made for contrast of uterine zonal structure and signal intensity of anterior/posterior fat (for signal intensity homogeneity). Intra-individual comparisons be-tween 3/1.5 T image evaluations were performed.Results: Image quality was superior at 3 T to 1.5 T for uterus (T2WI sagittal),uterine cervix (T2WI axial/sagittal and HASTE) and ovary (T2WI axial). Imageswere more inhomogeneous at 3 T (T2WI axial), while motion artifact was severerat 1.5 T (T2WI axial/sagittal)(p < 0.05). No significant difference was found be-tween 3/1.5 T MR imaging for overall image quality on all the sequences.Conclusion: Our results suggest that MR imaging of the female pelvis at 3 T hasa potential to provide excellent image for uterine cervix, and ovary on high-reso-lution T2WI. No significant improvement in overall image quality might be due toinhomogeneity at 3 T images.

B-083 10:48

3 T MR imaging in preoperative evaluation of depth of myometrialinfiltration in endometrial cancer: Preliminary resultsF. Fiocchi, S. Ferraresi, A. Barberini, N. Caproni, G. Ligabue, P. Torricelli;Modena/IT ([email protected])

Purpose: To evaluate the diagnostic accuracy of 3 T Magnetic Resonance Imag-ing (MRI) in assessing the depth of myometrial infiltration in endometrial cancer.Methods and Materials: Twenty-six patients (60 ± 13.4 years old) with histolog-ically proven endometrial carcinoma underwent preoperative MR imaging. AxialFSET1, axial, sagittal and para-coronal FSET2 and para-coronal FMPSGR Gado-linium-enhanced sequences were performed by a high field strength magnet (3 T,Philips Medical Systems, Best, The Netherlands) using an external phased-arraycoil. Within one month from MR imaging, all patients underwent hysterectomy.The exams were evaluated by two different radiologists and the results of MRimaging in assessing the depth of myometrial infiltration were compared with thehistological results. A statistical evaluation was performed.Results: Histology showed intramucosal neoplasm in 4 patients, myometrial in-filtration less than 50% in 6 patients, more than 50% in 12 patients and transmu-ral cancer in 4 patients. At MR imaging, intramucosal cancer was diagnosed in 3cases, myometrial infiltration less than 50% in 8 patients, more than 50% in 10patients and transmural in 4 patients. Concordance between MR imaging andhistological diagnosis was 88.4% (23 over 26 cases), with mean sensitivity, spe-cificity and accuracy respectively of 89.5%, 97.5% and 95.1%.Conclusion: Our preliminary results show that 3 T MR imaging proved to be areliable and accurate tool in assessing the depth of myometrial infiltration in en-dometrial cancer. Its accuracy falls in the highest rate of diagnostic accuracyachievable by 1.5 T MR imaging.

B-084 10:57 !MR evaluation of adnexal massesT.M. Cunha, A. Guerra, J. Caldeira, A. Félix; Lisbon/PT([email protected])

Purpose: To evaluate MR accuracy in the diagnosis of malignancy in histologi-cally confirmed adnexal lesions.Methods and Material: We retrospectively reviewed MR imaging of 131 patients(mean age, 52 years) with 154 surgical adnexal masses.MR readings were done by one senior and two radiologists in training. Masseswere divided in two subgroups (malignant/benign) according to their features inMR.Our inclusion MR imaging criteria for adnexal malignancy were: contrast-enhancedsolid lesions, contrast-enhanced solid component in mixed lesions (except thosewith low-signal intensity solid component on T2WI), contrast-enhanced papillaryprojections in cystic lesions (except those with low-signal intensity papillary pro-jections on T2WI) or septal thickness ≥ 3 mm. Ascites, peritoneal metastasis andpelvic enlarged lymph nodes were also inclusion criteria for malignancy.Results: By MR imaging evaluation 90 adnexal lesions were malignant and 64

Scientific Sessions

156 C D E FBA G

were benign. Histologically, 85 masses were confirmed to be malignant (ovarian:62 carcinomas, 15 epithelial borderline tumors (BT), 2 tumors with uncertainmalignant potential; 5 fallopian tube carcinomas and 1 mesosalpinx BT); 69 werebenign (63 ovarian tumors, 4 functional cysts), 1 peritoneal pseudocyst and 1epiploic appendix esteatonecrosis. Correlating MR imaging with histology resultswe confirmed 83 were malignant (true positive) and 62 were benign (true nega-tive) lesions. The MR sensitivity and specificity for malignancy were 90% and98%, respectively. MR reached an accuracy of 94% with a 92% positive predic-tive value and 97% negative predictive value for malignant adnexal lesions.Conclusion: MR has an excellent accuracy with a very high predictive value forthe diagnosis of malignant adnexal lesions.

B-085 11:06

MRI compared to intraoperative frozen-section examination for thediagnosis of adnexal tumors; correlation with final histologyM. Bazot, J. Nassar-Slaba, I. Thomassin-Naggara, A. Cortez, E. Darai,C. Marsault; Paris/FR ([email protected])

Purpose: The aim was to compare the accuracy of magnetic resonance imaging(MRI) and intraoperative consultation (IC) for the diagnosis of adnexal masses,with reference to final histology.Methods and Materials: MRI was performed in 136 women with sonographicallyindeterminate adnexal masses. IC consultation included macroscopic and fro-zen-section examination. Macroscopic examination and MRI determined size,nature, presence of vegetations or solid portions within masses. All masses werecharacterized as benign or malignant according to previously published MRI andhistopathological criteria. Sensitivity, specificity, and predictive values for the di-agnosis of malignancy by MRI and IC were assessed.Results: Histology revealed 168 adnexal masses (99 benign, 23 borderline and46 invasive). Frozen sections were examined in 151 cases. Among the 151 adn-exal masses studied by MRI and IC, the respective sensitivities, specificities, andpositive and negative predictive values of both methods for the diagnosis of ma-lignancy were 89.7% and 84.5%, 91.4% and 100%, 86.7% and 100%, and 93.4%and 91.3%. Sensitivities of MRI and frozen section for the diagnosis of serousversus mucinous borderline tumors were 33.3% and 93.3%, and 62.5% and 12.5%,respectively.Conclusion: MRI is less accurate than IC to characterize adnexal masses. How-ever, MRI may increase the relevance of intraoperative consultation for border-line mucinous tumors.

B-086 11:15

Characterisation of borderline ovarian tumours by MR imagingL.W. Turnbull, A.J. Sebastian, M. Sreenivas, K. Young, D. Poole; Hull/UK([email protected])

Purpose: Ovarian tumours with borderline malignant potential have a favourableoutcome. Characterisation of mixed solid and cystic ovarian masses pre-opera-tively can aid in surgical planning and may allow preservation of fertility in young-er patients. This study examines the imaging characteristics of borderline ovariantumours in comparison with other encapsulated ovarian masses.Methods and Materials: 49 patients with 59 encapsulated ovarian lesions wereretrospectively evaluated. All patients underwent MR evaluation of suspiciouspelvic adnexal masses between June 2001 and June 2004. MR imaging wasperformed at 1.5 or 3 T (GE LX Signa systems) using a flexible 4 or 8 channeltorso coil respectively. FRFSE images were acquired sagittally, axially and ob-liquely followed by T1W FSPGR ± fat suppression as determined clinically. Busco-pan and a dielectric pad at 3 T were used to minimise artefacts. Histopathologywas available in all subjects.Results: Ovarian cysts ranged in size from 0.9 to 26 cms. All cystadenomas(14), fibromas (2), a haemorrhagic cyst (1) and 2/3 cystadenocarcinomas werecorrectly diagnosed. MR correctly called 49% of cystadenofibromas, but 51%was misclassified as cystadenomas or fibroma (1/31). 3/6 borderline tumourswere correctly identified, 2 classified as cystadenoma or indeterminate cyst andone termed cystadenocarcinoma, although omental deposits were present in thiscase. Typical features seen in the correctly diagnosed borderline tumours includ-ed endocystic projections separately identified from an intact ovarian capsule.Conclusion: This study demonstrates the potential of MR imaging in pre-opera-tive assessment of ovarian masses and indicates the need for high resolutionimaging preferably at 3 T.

B-087 11:24

Multislice CT and MR imaging evaluation of ovarian massesA.C. Tsili, C. Tsampoulas, E. Arkoumani, E. Paraskevaidis, S.C. Efremidis;Ioannina/GR ([email protected])

Purpose: To evaluate the accuracy of multislice CT and magnetic resonance(MR) imaging in the detection and differentiation of benign and malignant ovarianmasses.Methods and Materials: Preoperative CT of the abdomen and MR imaging ofthe pelvis were performed in 40 women presenting with clinically or sonographi-cally detected ovarian masses. CT examinations were performed on a 16-sliceCT scanner, using a detector collimation of 16 X 0.75 mm and a pitch of 1.2. MRexaminations were performed on a 1.5 Tesla unit and T2 and T1-weighted se-quences, before and after intravenous administration of gadolinium chelate com-pounds were obtained. The accuracy of multislice CT and MR imaging in thedetection and characterization of ovarian masses was evaluated, using surgicaland histological results as reference.Results: Both multislice CT and MR imaging detected 50 (93%) out of 54 adnex-al masses, as proven histologically. The sensitivity, specificity, positive predictivevalue and negative predictive value of multislice CT and MR imaging in the differ-entiation between benign and malignant ovarian masses were 83%, 89%, 71%,94% and 92%, 95%, 85% and 97%, respectively.Conclusion: Multislice CT and MR imaging demonstrated a similar accuracy inthe detection of ovarian masses, but MR imaging performed better in the charac-terization of adnexal masses, mainly due to its superior tissue characterization.

B-088 11:33

Role of endocavitary MR imaging for posterior deep endometriosis withrectosigmoid involvement. Preoperative prospective study with surgicalcorrelationC. Roy, A. Biblocque, V. Thoma, A. Wattiez, J. Leroy, C. Saussine;Strasbourg/FR ([email protected])

Purpose: US and MR imaging are able to detect rectosigmoid infiltration of deependometriosis. The purpose was to assess added diagnostic value of TVUS withMR imaging obtained by phased-array (bMRI) and endovaginal (evMRI) to detectintestinal involvement.Methods and Material: From 2004 to 2005, 47 women underwent preoperativeTVUS, bMRI and evMRI before surgical resection of the nodule. MR imaging wasperformed using T2-weighted FSE and T1-weighted SE images before and afterGd. Examination results were correlated with findings during surgery and patho-logical analysis.Results: At pathology, 43 women demonstrated deep muscle layer invasion, 2an isolated serosal involvement and 2 no intestinal extension. All nodules hadwell-defined borders, low signal intensity with small hyperintense spots on T2-weighted MR (18 cases). All were homogeneous and hypoechoic on TVUS. Spic-ulated strandings were more often seen on MR (44 cases) than on TVUS (35cases). Deep wall extension was correctly diagnosed as a hypoechoic or hypoin-tense homogeneous thickening of muscle layer in 32 cases on TVUS (Se : 74%)and 41 cases on MR (Se : 95%). An additional pattern as a hyperintense layer ofmucosa surrounding the nodule was found on MR in all cases. EvMRI was clear-ly better at diagnosing intestinal involvement for small lesions (less than 2 cm)localized inside the pouch of Douglas or recto-vaginal septum.Conclusion: evMRI gives additional information over TVUS in the diagnosis ofrectal involvement preoperatively for small sized nodules. A hyperintense signallayer surrounding the nodule is specific for this purpose.

B-089 11:42

MR imaging of the pelvis in mullerian duct abnormalities: Is there a role inselecting patient for hysteroscopic metroplasty?P. Paolantonio1, M. Rengo2, M. Anzidei2, A. Carnevale2, F. Iafrate1,C. De Angelis2, A. Laghi1; 1Latina/IT, 2Rome/IT([email protected])

Purpose: Selection of patients for hysteroscopic metroplasty requires a com-plete assessment of uterine morphology, usually achieved by laparoscopy whichrepresents the gold standard in the evaluation of uterine morphology. The aim ofour study was to assess the role of MR imaging in selecting patients with mulle-rian duct fusion defects for hysteroscopic metroplasty.Methods and Materials: We performed a prospective study on 25 women withknown mullerian duct defects assessed at hysteroscopy. The examination wasperformed using an 1.5 T magnet equipped with a phased-array coil acquiring ahigh-resolution TSE T2W sequence (TR: 3000 msec/TEeff 190 msec/ETL: 29/

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 157C D E FA G

Matrix 256 x 512) parallel to the long axis of the uterus. Final diagnosis wasassessed in all patients at surgery using laparoscopy in the assessment of uter-ine morphology. In 22 patients hysteroscopic metroplasty was also performed.Image analysis was performed by two radiologist in a blind consensus fashion.They qualitatively evaluated morphology of both the uterus and endometrial cav-ity. MR findings were compared with laparoscopic and hysteroscopic findings.Results: MR imaging revealed uterine abnormalities in all patients. 15 patientswere diagnosed with an incomplete septate uterus, 5 with a complete septateuterus, four patients with an arcuate uterus and one patient with a bicornuateuterus. MR imaging findings were correct in all cases, as confirmed by hysteros-copy and laparoscopy.Conclusion: MR imaging is a valuable tool in the evaluation of uterine morphol-ogy. Its use will help spare patients from laparoscopic patient selection for hyster-oscopic metroplasty.

B-090 11:51

The value of magnetic resonance imaging in the diagnosis andmanagement of abdominal extra-uterine pregnancyP.D. Corr; Durban/ZA ([email protected])

Purpose: To determine the value of magnetic resonance imaging in the imagingdetection of placental location and adherence in patients with extra-uterine orabdominal pregnancy.Methods: A study of patients with a suspected diagnosis of extra-uterine preg-nancy was performed over a 12 month period. MR images on hard and soft copywere reviewed by two radiologists blinded to the clinical and operative findings.Results: Nine patients with ten fetuses were imaged. 9 fetuses were in an extra-uterine position, 3 were delivered dead, one from intrauterine growth retardationand one from a fatal congenital anomaly. The placenta was localized correctly inall 9 patients with placental adherence demonstrated in 4 patients. The placentawas safely delivered in 6 patients and left in situ in 3 in which there was MRevidence of placental adherence.Conclusion: MR imaging is valuable in accurately demonstrating the location ofthe placenta within the abdomen and the presence of placental adherence whichdirectly affects the decision whether to remove or leave the placenta in situ.

10:30 - 12:00 Room L/M

Neuro

SS 211Brain tumors: Impact of new methodsModerators:S.P. Morozov; Moscow/RUB. Ostertun; Osnabrück/DE

B-091 10:30

Comparison of flow-sensitive alternating inversion recovery (FAIR) and firstpass dynamic contrast enhanced MR imaging (CE MR imaging) techniquesin evaluation of intracranial tumor perfusion at 3 TeslaK. Zhang1, C. Li1, Y. Liu2, X. Meng1, D. Feng1, J. Zheng1, P. Cong1; 1Jinan,Shandong/CN, 2Hefei, Anhui/CN ([email protected])

Purpose: FAIR is a novel technique to evaluate blood perfusion. Compared withfirst pass dynamic contrast enhance MR imaging, it is non-invasive, more con-venient and cost-efficient. This study aims to determine whether quantification atFAIR imaging can be used to assess regional cerebral blood volume (rCBV) in amanner similar to CE MR imaging.Methods and Materials: Using a GE Signa Excite II 3.0 T MR scanner, 34 pa-tients, including 17 gliomas (10 low grade WHO I, II and 7 high grade WHO III,IV), 7 meningiomas, 3 metastases, 3 acoustic schwannomas, 2 medulloblasto-mas, 1 craniopharyngioma and 1 hemangioblastoma were examined by FAIRand CE MR imaging. Three ROIs were selected on the solid part of each neo-plasm on the post-processing images. Mean FAIR value and rCBV were calculat-ed. The results of benign and malignant gliomas were compared with independentsamples T test. The results of all cases were analysed with linear correlationanalysis.Results: There was statistically significant difference between the FAIR valueand rCBV of benign and malignant gliomas (P < 0.01). The FAIR value and rCBVof all cases showed positive linear correlation (P < 0.05).Conclusion: FAIR and CE MR imaging methods are both helpful in differentiat-ing benign from malignant glioma. The significant positive linear correlation in

detecting tumoral blood perfusion between FAIR and CE MR imaging methodssuggests that FAIR imaging, like CE MR imaging perfusion, is complementary toroutine MR imaging in the assessment of tumoral perfusion. Furthermore, be-cause there is no need for contrast enhancement, FAIR imaging, a non-invasivetechnique, will be more acceptable in clinical application.

B-092 10:39

Diffusion tensor imaging (DTI) in the normal appearing corpus callosumand contralateral white matter in gliomas showing potential microscopictumoral infiltrationE. Papadaki, M. Law, E. Clair, Y. Lui, M. Inglese, G. Johnson; New York, NY/US([email protected])

Purpose: Pathologically, there is microscopic infiltration of tumoral cells acrossthe corpus callosum (CC) in patients with high grade gliomas. The aim of thisstudy is to determine if there are DTI changes within the normal appearing CCand contralateral white matter (WM-CONT) and if there are differences betweenglioma grades.Methods and Materials: 30 patients with primary glioma - classified using theWHO classification as low-grade (n = 10), anaplastic (n = 12) and glioblastomamultiforme (n = 8) - had DTI performed using a pulsed gradient, double spin echo,echo planar imaging (TR/TE=5300/74; 128x128 matrix; 220x220 mm FOV; forty-five 3 mm contiguous slices; b= 1000 s/mm2) at 1.5 T. ROIs were placed in theCC in three locations, the ipsilateral callosum (CC-IPSI), central callosum (CC-CENT) and contralateral callosum (CC-CONT) to the side of the glioma, as wellas within the WM-CONT. FA and ADC were measured.Results: There is significantly increased ADC (p = 0.003) and decreasedFA (p < 0.001) in the CC-IPSI and in the CC-CENT (p = 0.007, and p < 0.001respectively) in grade III gliomas compared with grade II. There is also signifi-cantly decreased FA in the normal appearing WM-CONT (p = 0.017), as well asthe CC-IPSI (p = 0.02), CC-CENT (p = 0.001) and CC-CONT (p = 0.012) in gradeIV compared to grade II.Conclusion: DTI is able to demonstrate changes in FA and ADC in the normalappearing corpus callosum of grades III and IV and contralateral white matter ofgrade IV, indicating potential microscopic tumoral infiltration.Follow-up studies are underway to determine if there is tumoral infiltration onfollow-up imaging or correlation with clinical outcome.

B-093 10:48

Perfusion and diffusion MR imaging of glioblastoma: Prospectiveevaluation of a four year temozolomide clinical trialA. Leimgruber, S. Ostermann, E. Yeon, E. Buff, P.P. Maeder, P.A. Schnyder,R. Stupp, R.A. Meuli; Lausanne/CH ([email protected])

Purpose: New treatments for patients suffering from glioblastoma have shownpromising results. This study evaluates recent MR imaging follow-up techniqueson newly diagnosed glioblastoma patients and analyzes the potential for per-fusion and diffusion sequences in anticipating disease progression. Various MRimaging parameters are further studied as potential prognostic factors of diseaseprogression.Methods and Materials: This study included patients from a phase II clinical trialof temozolomide combined with surgery and radiotherapy. MR imaging follow-upwas carried out 1 month after radiochemotherapy and every 2 months thereafter.T1- and T2-weighted SE series as well as diffusion and perfusion EPI serieswere systematically performed on a 1.5 T magnet. The analysis used clinicalparameters combined with classical criteria on tumour size variation to focus ondisease progression.Results: Among the 41 patients included in the study (311 MR imaging examina-tions), 32 demonstrated disease progression at some time in their follow-up. Per-fusion and diffusion imaging did not anticipate disease progression. Furthermultivariate analysis determined T1 largest diameter and T2 size variation assignificant survival parameters at disease progression (p < 0.02, p < 0.05).Conclusion: Our study presents a simple time scheme combined with statisticalanalysis for the evaluation and potential scoring of brain tumours and treatments.T1 and T2 techniques remain the gold standard in setting the date of diseaseprogression. At disease progression, T1 contrast-enhancement larger than 3 cmin size combined with increased T2 hyper signal is a marker of inferior prognosis.Perfusion and diffusion techniques showed a complex behaviour and could notbe used to anticipate tumour progression.

Scientific Sessions

158 C D E FBA G

B-095 10:57 !Functional mean diffusion map: A noninvasive MR imaging biomarker forevaluating cell proliferation of brain gliomasZ.X. Li, J.P. Dai, T. Jiang, S.W. Li, Y.L. Sun, P.Y. Gao, J. Ma; Beijing/CN([email protected])

Purpose: To investigate the correlation between the mean diffusivity (MD) meas-ured using diffusion tensor imaging (DTI) and the Ki-67 labelling index in glio-mas.Methods and Materials: DTI was performed using a Siemens Trio MR scannerin fifty-six consecutive patients with histologically proven gliomas. The regions ofinterest (ROIs) were positioned on the solid portion of the lesion. The MD valueswere calculated. Ki-67 staining of paraffin embedded sections with monoclonalantibody MIB-1 was done. In the area with the highest number of labelled cells,the percentage of Ki-67 positive nuclei (labeling index: LI) was determined. Then,the results of Ki-67 LI were divided into three groups: High (LI > 40 %), median(15 % ≤ LI ≤ 40 %) and low (LI < 15 %). Then, the Ki-67 LI were compared withMD values.Results: The MD values of the high Ki-67 LI group (0.863 ± 0.197 × 10-3 mm2/s)were significantly lower than those of the median Ki-67 LI group (1.130 ± 0.275 ×10-3 mm2/s) (p = 0.015) or low Ki-67 LI group (1.278 ± 0.298 × 10-3 mm2/s)(p < 0.001). Negative correlation existed between MD values and Ki-67 LI (rs=-0.570, p < 0.001).Conclusion: In brain gliomas, Ki-67 LI bears a significant relationship to MDvalues. Its molecular pathological base is that Ki-67 can precipitate the tumor cellproliferation. The MD value of DTI is regarded as a noninvasive MR imaging bi-omarker for evaluating cell proliferation in gliomas.

B-096 11:06

Primary central nervous system lymphomas (PCNSL): MR imagingresponse criteria revisedW. Küker1, T. Nägele2, P. Pretorius1, E. Thiel3, M. Weller2, U. Herrlinger2;1Oxford/UK, 2Tübingen/DE, 3Berlin/DE ([email protected])

Purpose: To asses the validity of a treatment response evaluation system estab-lished for malignant gliomas to patients with primary CNS lymphoma (PCNSL).Methods and Materials: MR imaging examinations of 112 immunocompetentpatients with biopsy-proven PCNSL were retrospectively evaluated for responseto primary MTX-based chemotherapy according to the Macdonald criteria. Theresponse classification was compared to the further course and treatment of thedisease with the aim of identifying patients who did not achieve complete re-sponse (CR) but nevertheless remained progression-free for at least 6 monthswithout receiving further therapy.Results: Sixty-eight patients had persisting contrast-enhancing lesions after pri-mary therapy. Four patients of this cohort did not receive further tumour-directedtherapy. After follow-up of 9, 24, 32 and 54 months, all 4 patients did not presentwith progressive tumour growth. In all 4 patients, the residual lesions were 5 mmin diameter or smaller. In all patients, the residual lesion was found in the vicinityof the biopsy site.Conclusion: Residual contrast enhancement may be observed in PCNSL pa-tients after successful therapy and does not necessarily indicate residual, biolog-ically active lymphoma. On the basis of the morphology of the non-progressiveresidual lesions in the patients described above, we propose a modification ofthe Macdonald response criteria which, in the absence of perilesional oedema,classifies contrast-enhancing lesions < 3 mm in general, or contrast-enhancinglesions < 5 mm in the region of biopsy, haemorrhage, infection or a previouslylarger (> 5 cm) tumour lesion as CR.

B-097 11:15

Evaluation with MR imaging of patients with hematopoietic neoplasms andspinal epidural and leptomeningeal spread: Comparison with the results oflumbar punctureM. Di Girolamo, F. De Angelis, I. Ruggiero, L. Greco, L. Argnani, V. David;Rome/IT ([email protected])

Purpose: To evaluate with MR imaging the spinal epidural and leptomeningealinfiltration in patients with lymphoma and leukemia and correlate the results withlumbar puncture.Methods and Materials: 50 patients with ALL, AML and NHL and with clinicalsuspicion of spinal extra-medullary intra-dural spread underwent MR imagingusing a 1.0 T superconductive magnet. We performed sagittal-SE-T1-weightedscans (TR: 500 ms; TE: 20 ms; n-acq: 3; acq-time: 3'40") and FSE-T2-weightedscans (TR: 2840 ms; TE: 96 ms; acq-time: 3'30") at the level of the dorsal and

lumbar spine (sl. thickn: 4 mm). Then we performed contrast-enhanced sagittal-T1-weighted scans. In positive cases we also performed axial FSE-T2-weightedscans and contrast-enhanced axial T1-weighted scans. All the patients under-went lumbar puncture within two days after MR examination.Results: We detected 10 patients with positive features of spinal epidural andleptomeningeal infiltration on MR imaging. Among these 10 patients only 7 had apositive lumbar puncture ("blast" forms or lymphomatous cells in CSF). Lumbarpuncture was negative when the neoplastic infiltration involved the epidural spaceand the nerve roots. The most sensitive MR imaging-sequence in the detection ofspinal epidural and leptomeningeal involvement has been considered the sagit-tal T1-weighted scans performed after administration of Gadolinium-DTPA.Conclusion: In our limited experience MR imaging was more sensitive than lum-bar puncture in the detection of spinal epidural and leptomeningeal spread byhemopoietic tumors. The diagnosis of leptomeningeal spread is fundamental be-cause it modifies the therapeutic measures and its early symptoms are oftennonspecific. In patients with a hemopoietic neoplasm and strong clinical suspi-cion of spinal epidural or leptomeningeal infiltration, when lumbar puncture isnegative, MR imaging should be performed.

B-098 11:24

Detection of brain metastases with 3 T MR imaging: Comparison amongthree contrast-enhanced sequencesM. Hanamiya1, S. Kakeda1, Y. Korogi1, N. Ohnari1, J. Moriya1, T. Sato1, Y. Hiai2;1Kitakyushu/JP, 2Kumamoto/JP ([email protected])

Purpose: 3 T MR systems have some advantages of a greater relative T1-short-ening effect and an increased signal-to-noise ratio for contrast-enhanced MRimaging. However, it has not been clarified which T1-weighted sequence is ap-propriate for detection of brain metastasis at 3 T. The purpose of this study is tocompare the detection and conspicuity of enhancing brain metastases amongthree contrast-enhanced sequences, 3D fast SPGR with 1.4 mm isotropic voxel,conventional spin-echo sequence (SE), and inversion recovery fast SE (IR-FSE).Methods and Materials: In 20 consecutive patients with brain metastases re-ferred for enhanced brain imaging, axial images of SE, IR-FSE and 3D fast SPGRat 3 T were obtained after administration of gadoteridol (ProHance; Bracco) at0.2 mmol/kg. A 3D fast SPGR was acquired with 1.4-mm thick sections, and bothSE and IR-FSE were acquired at the same section locations with section thick-ness of 6 mm. Two neuroradiologists assessed the image quality of three se-quences with regard to the lesion detectability of brain metastases.Results: In nine (45%) of 20 patients, 3D fast SPGR revealed additional metas-tases. Significantly more metastases were detected with contrast-enhanced 3Dfast SPGR than with SE and IR-FSE (a total of 114 metastases with 3D fastSPGR vs.70 with SE and 69 with IR-FSE). In particular, 3D fast SPGR was supe-rior to the other two sequences in detection of the small lesions (< 5 mm).Conclusion: The post-contrast fast SPGR sequence with 1.4 mm isotropic voxelis clinically valuable for detecting brain metastases at 3 T.

B-099 11:33

High-resolution susceptibility weighted MR imaging in patients with braintumors during the application of exogeneous gaseous agentsJ. Sedlacik, A. Rauscher, C. Fitzek, B. Walter, A. Hochstetter, R. Kalff,W.A. Kaiser, J.R. Reichenbach; Jena/DE ([email protected])

Purpose: With susceptibility weighted imaging (SWI) it is possible to obtain func-tional and anatomical information with high spatial resolution. The aim was toinvestigate the response of brain tumors to a modulation of venous blood oxy-genation by breathing air, carbogen and oxygen.Methods and Materials: Data from six patients (four with glioblastoma multi-forme (grade IV), one with astrocytoma (grade II), and one with a cerebral metas-tasis; all histologically confirmed) were acquired using a 3D gradient-echo,first-order velocity compensated sequence on a 1.5 T scanner (TE/TR/alpha=40 ms/67 ms/25, FOV=256x192x64, matrix=512x192x64). One scan wasacquired for each breathing condition. SWI was computed by multiplying themagnitude images with phase masks calculated from the same scans. Addition-ally, venous vascularization was assessed on minimum intensity projections ofthe SWI scan acquired during air breathing. Signal changes were evaluated overROIs of typically 0.5 cm3 (15x15x10 voxels) and maps of relative signal changeswere computed.Results: Signal intensities in the tumor decreased during carbogen breathingcompared to oxygen inhalation in 3 cases. The glioblastomas displayed the strong-est response and high heterogeneity (signal changes between -7% and +28%).The astrocytoma and cerebral metastasis showed a negative response to carbo-gen in all ROIs ranging from -9.3% to -2.0% (astrocytoma) and -8% to -16%(cerebral metastasis).

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 159C D E FA G

Conclusion: The method proved its potential to probe the response of brain tu-mors to carbogen and oxygen and to visualize venous vascularity with very highspatial resolution. Systematic studies on different tumor types may allow for abetter classification of brain tumors or areas of abnormal blood supply.

B-100 11:42

Degree of mass effect by MR imaging in glioblastoma multiforme canpredict the gene expression level of a cell proliferation gene expressionprogramM. Kuo1, C. Nardini2, M. Diehn1, S. Cha3, L. Benini2; 1San Diego, CA/US,2Bologna/IT, 3San Francisco, CA/US ([email protected])

Purpose: To characterize the relationship between degree of mass effect on MRimaging and tumor proliferation rate as assessed by gene expression level of acell proliferation gene expression program in GBM.Methods and Materials: 22 GBMs with MR imaging and matching cDNA micro-array gene expression data were analyzed. Using unsupervised microarray anal-yses, a cell proliferation gene expression program consisting of 49 genes wasfirst identified from the ~2200 cDNA clones analyzed from the GBM microarrayexperiments. Degree of mass effect of each GBM on MR imaging was then grad-ed by 2 radiologists on a binary scale and in consensus. ROC curves were thenused to determine whether the gene expression level of the genes in the cellproliferation program could be used to classify the tumors by their degree of masseffect. Statistical significance was then evaluated.Results: 10/22 and 12/22 GBMs demonstrated high and low mass effect respec-tively. Gene expression levels of the cell proliferation gene expression programwere predictive of degree of mass effect with an AUC value of 0.72 and a p < 0.01.Overexpression of genes in the proliferation-associated gene expression pro-gram correlated with high mass effect while decreased expression levels corre-sponded with low tumor mass effect.Conclusion: Integration of genomic information and MR imaging may provideuseful insights into GBM tumor biology. Specifically, mass effect on MR imagingmay potentially be a broad, global indicator of GBM tumor proliferation rate basedon relative expression level of the cell proliferation gene expression program.

10:30 - 12:00 Room N/O

Breast

SS 202Breast cancer screeningModerators:G. Korzhenkova; Moscow/RUG.M. Villeirs; Gent/BE

B-102 10:30

Mammography screening for breast cancer in the Czech Republic: Resultsof analysis of 467,696 women and 2,147 breast cancer casesA. Svobodník1, J. Daneš2, H. Bartonková1, M. Skovajsová2, D. Klimeš1;1Brno/CZ, 2Prague/CZ ([email protected])

Purpose: As the efficiency of screening on reduction of breast cancer mortalitywas confirmed in several clinical trials, screening has been adopted in most Eu-ropean countries. In the Czech Republic, national-wide screening has started inSeptember 2002. First official results are presented.Methods and Materials: Data from a total of 57 independent screening centerswere analyzed. To assure high quality of data, thorough central system of datacollection and validation was used. The system is based on a) obligatory struc-ture of data collected by individual centers b) establishment of independent datacenter and c) development of user-friendly software for data management.Results: Czech National Mammography Screening Database contains detaileddata of 467,696 women. A total of 2,147 breast cancer cases were detected,hence the cancer detection rate was 46 cases per 10,000 women.The distribution of tumor sizes detected in screening was: Tis 7%, T1 71%, T220%, T3 1% and T4 1%. Before screening was implemented, the percentage ofbreast cancers ≤ 2 cm was 39 % (National Cancer Registry, 2002) in compari-son to 78 % of tumors ≤ 2 cm detected in screening.Conclusion: As mammography screening is enormously demanding project fromboth financial and organizational perspective, exact evaluation of its costs andbenefits should be conducted. The sophisticated system of screening data man-agement and validation is important motivational factor for all screening centersfrom the quality control perspective. Substantial quality control procedures are

crucial when chaotic system of cancer prevention is transformed into regular serv-ice screening program.

B-103 10:39

Over-diagnosis in screening: A 25-year follow-up of the MalmöMammographic Screening Trial, MMSTS. Zackrisson1, J. Garne2, J. Manjer1, L. Janzon1, I. Andersson1; 1Malmö/SE,2Aarhus/DK ([email protected])

Purpose: To assess the rate of over-diagnosis of breast cancer in screening.Methods and Materials: The Malmö Mammographic Screening Trial, MMST, wasa randomised controlled study of 42283 women, 45-69 years at entry, 21088were invited to screening and 21195 were controls. The study started in 1976and ended in 1986. The randomised design was maintained in the invited groupfor women under age 70. From 1990 onwards, all women between ages 50-70were invited to screening in Malmö. We assessed the detection rate of breastcancer during a 25-year follow-up, divided into period 1, where the randomiseddesign was maintained, and period 2, after termination of the randomised de-sign.Results: After period 1 there was an overall excess detection rate of 24%, whichwas more pronounced, 32% among the older age group (55-69 years at entry)compared with the younger group, 16% (45-54). During period 2 the excess de-tection rate in the invited group was balanced, when the control group was invit-ed in the age group 45-54 years at entry. In the age group where screening wasterminated (screening finished in the invited group and the control group was notinvited) an 8% drop was seen during period 2, with 10% excess remaining at theend of the 25-year follow-up (about 80 excess cases after the end of follow-up).Conclusion: There exists an over-diagnosis in screening which amounts to about10%. Screening in the older age group (55-69 at entry) generates most of theexcess cases.

B-104 10:48

Independent double reading of screening mammograms in the Netherlands:Impact of additional double reading by screening mammographyradiographersF.J. Nobrega1, L.E.M. Duijm1, J.H. Groenewoud2; 1Eindhoven/NL,2Rotterdam/NL ([email protected])

Purpose: To determine the value of independent radiographer double reading ofscreening mammograms in addition to standard radiologist double reading.Methods and Materials: From January 2003 to August 2004, all 45,037 screen-ing mammograms were independently read by two screening radiologists andtwo screening mammography radiographers, who for each case blindly recordedwhether additional diagnostic procedures were required. The radiographers werenot specially trained for the experiment. All positive radiographer readings werere-evaluated by the radiologists. Patients were referred for further assessment ifat least one radiologist considered recall necessary. One year follow-up of allpatients were obtained.Results: Radiologists referred 725 cases (1.6%), of which 246 (33.9%) werefound to be malignant. Review of 317 additional positive radiographer readingsled to another 56 referrals, which resulted in the detection of 12 additional can-cers. These extra referrals increased the initial referral rate (RR) from 1.6% to1.7% and the cancer detection rate (CDR) from 5.5 to 5.7 per 1,000 womenscreened. With radiographer double reading only, 507 women would have beenreferred (RR 1.1%), resulting in 210 malignant cases (positive predictive value44.0%; CDR 4.7), including 7 cases which now presented as interval cancers.Referral of all 317 additional positive radiographer readings would have increasedthe RR to 2.3% (1,042 cases) and led to 265 cancers found at screening (CDR5.9).Conclusion: Radiographer double reading would result in smaller CDRs thanradiologist double reading. Radiologist review of positive radiographer readings,however, may add to the sensitivity of breast cancer screening programs.

B-105 10:57

Late-detected cancers in the Dutch mammographic screening programmeH.J.T.M. Rijken, R. Holland, J.H.C.L. Hendriks†; Nijmegen/NL([email protected])

Purpose: The recall rate in the Dutch breast screening programme (1% in 2000for subsequent examinations) is the lowest worldwide, with possible consequencesincluding higher rates of late-detected (i.e. "missed") interval and screen-detect-ed cancers. To estimate the effect of changes in recall rate on earlier detection ofcancers, we carried out a blinded review of interval and screen-detected can-cers.

Scientific Sessions

160 C D E FBA G

Methods and Materials: A total of 495 sets of screen-negative mammograms(prediagnostic mammogram and the immediate previous mammogram) werecollected from women participating in the biennial Dutch screening programme(250 control- and 245 cancer cases). These mammograms were read by 15 ex-perienced radiologists who were blinded to outcome. Results were used to calcu-late the relationship among recall rate, detection rate and numbers offalse-positives.Results: Increasing the recall rate from 1% to 2% would increase the detectionrate from 4.2% to 4.5% due to the earlier detection of interval cancers. Moreover,0.5%of the screen-detected cancers would be detected 2 years earlier (late-de-tected cancers). For each 1% incremental increase in recall rate above 5%, thedetection rate would increase by approximately 0.03% only, with positive predic-tive values decreasing to below 10%.Conclusion: Breast cancer can be detected earlier by lowering the threshold forrecall, especially for recall rates of 1% - 4%. To benefit most from a higher recallrate, radiologists should be able to identify the cancers with subtle mammographicfeatures, which are often based on temporal changes.[JNCI 2005;97:748]

B-106 11:06

Interval breast cancer: Prognostic features and survival by subtype andtime since screeningG.J.R. Porter, A.J. Evans, H.C. Burrell, A.H.S. Lee, I.O. Ellis, J. Chakrabarti;Nottingham/UK ([email protected])

Purpose: We aimed to investigate the hypothesis that interval cancers arisingsoon after the previous screen and true interval cancers are biologically aggres-sive and have a relatively poor prognosis. We also wished to assess which prog-nostic features are relevant to interval cancers.Methods and Materials: We have analysed prognostic pathological features(grade, lymph node stage, size, vascular invasion, oestrogen receptor (ER) sta-tus, and histological type), radiological features (comedo/non-comedo calcifica-tion and spiculation) and survival for 538 invasive interval breast cancer cases bysubtype and time since previous screen.Results: Women presenting in the first year were more likely to be lymph nodepositive (47%vs40%, p = 0.02). Women presenting with true interval cancers weremore likely to be grade 3 than minimal signs and false negative intervals (52vs35%, p = 0.01). False negative cancers were more likely to have lobular fea-tures than other intervals (47%vs20%, p < 0.0001).There was no significant survival difference by classification sub-group (p = 0.64)or interval year (p = 0.83).At univariate analysis of all interval cancers; tumour size, grade, nodal stage, ERstatus, vascular invasion and comedo calcification correlated with survival.On multivariate analysis of prognostic features significant at univariate analysis,nodal stage (p-value = 0.009), tumour size (p = 0.004), ER status (p < 0.0001),and vascular invasion (p < 0.0001) were independently significant prognostic fea-tures.Conclusion: Our study shows that true intervals and interval cancers arisingquickly after screening do not have a worse prognosis and that interval cancershave a unique set of prognostic features.

B-107 11:15

The influence of mammographic background pattern on pathologicalfeatures, radiological features and survival in screened womenG.J.R. Porter, A.J. Evans, A.R.M. Wilson, H.C. Burrell, E.J. Cornford,J.J. James, L.J. Hamilton, A.H.S. Lee, J. Chakrabarti; Nottingham/UK([email protected])

Purpose: It is known there is an excess of breast cancer in women with radio-graphically dense breasts. Previous studies have shown poorer prognostic fea-tures in these tumours. No study has looked at survival.Methods and Materials: We have classified the background pattern (BI-RADSsystem) of 759 screened women who presented with a screen detected or inter-val invasive breast cancer. Pathological details (tumour size, grade, stage, vas-cular invasion and histological type) and mammographic appearances wererecorded. Breast cancer specific survival was ascertained with a mean follow-upof 9 years.Results: There was an excess of interval cancer in women with dense breasts(p-value < 0.0001). Screen detected (but not interval) tumours were significantlysmaller in fatty breasts (p = 0.014). Tumour grade, stage, vascular invasion andhistological type did not vary significantly with mammographic background pat-tern in screen detected or interval cancers. Screen detected cancer in fatty breastswere more likely to appear as ill defined (p = 0.003) or spiculate masses (p = 0.002)

in contrast to cancer in dense breasts which more commonly appear as distor-tions (p < 0.0001). However there was no significant breast cancer specific sur-vival difference by mammographic background pattern for screen detected cancers(p = 0.75), interval cancers (p = 0.82), or both groups combined (p = 0.12).Conclusion: The prognosis of screened women presenting with breast cancer isunrelated to dense mammographic background pattern despite an excess of in-terval cancers and larger screen detected tumours in this group. This data sup-ports the mammographic screening of women with a dense background pattern.

B-108 11:24

Delay in breast cancer diagnosis in the Netherlands: Performance ofhospitals involved in the evaluation of women with a positivemammographic screening examinationF.J. Nobrega1, L.E.M. Duijm1, J.H. Groenewoud2; 1Eindhoven/NL,2Rotterdam/NL ([email protected])

Purpose: To determine the inter-hospital variability in the diagnostic delay ofbreast cancer after breast cancer screening.Methods and Materials: We included all women aged 50-75 years who under-went biennial screening mammography in the southern breast cancer screeningregion of the Netherlands between January 1, 1996 and July 1, 2003. Clinicaldata, breast imaging reports, biopsy results and breast surgery reports were col-lected of all women with a positive screening result with a minimum of 2-yearfollow-up.Results: Of 196,317 mammographic screening examinations, 2,106 (1.1%) werepositive screens. Seven regional hospitals performed the workup of 2,097 wom-en with a positive screen examination and complete follow-up. Breast cancer wasdiagnosed in 992 women. A total of 58 (5.8%) women with a true positive screenexperienced a delay in breast cancer diagnosis exceeding three months. Thisdelay was due to erroneous mammographic interpretation of suspicious lesionsas benign or probably benign lesions (40 cases), benign biopsies from a malig-nant lesion (13), and omission to biopsy or remove a lesion, that was suspiciousat breast imaging (4) or core biopsy (1). The proportion of women who experi-enced a delay exceeding three months in breast cancer diagnosis after a truepositive screen ranged from 0-9.4% between the seven hospitals.Conclusion: Hospitals that are involved in the workup of women with a positivescreen examination show a large variation in the percentage of patients whoexperience a delay in breast cancer diagnosis. A majority of the delays are causedby radiological misinterpretation of suspicious lesions.

B-109 11:33

Role of routine breast ultrasound in women with negative radiologicallydense mammographyS. Ciatto1, V. Corsetti2, A. Ferrari2, M. Ghirardi2, R. Bergonzini2, S. Bellarosa2,O. Angelini2, C. Bani2; 1Firenze/IT, 2Calcinato (BS)/IT ([email protected])

Breast ultrasound (US) was performed routinely in 6,449 of 17,883 (36.5%) sub-jects with mammography reported as negative and radiologically dense (BI-RADSD3-4). While 138 cancers had been detected out of 257 suspicious mammogra-phies, 29 additional cancers were detected (overall detection rate: 0.44%, or 17.3%of total cancers detected) at US of D3-4 mammography negative breasts: 25 of29 US detected cancers were available for internal and external review, and neg-ative mammography and asymptomatic status was confirmed in 15 (detectionrate: 0.23%, or 8.9% of total cancers). Cancer detection rate was 0.11%, 0.22%,0.32% and 0.14 % in age groups < 40, 40-49, 50-59 and > 59, respectively. Thecost per additional carcinoma detected by additional US alone was € 25,847.85,whereas that per examined woman in the total series was € 21.68.The study confirms that US may detect breast carcinoma occult to mammogra-phy in dense breasts. Internal and external review of mammograms, which is notcommonly adopted in not controlled retrospective studies, is important to confirmthat mammograms were really negative, and to avoid overestimation of US ben-efits. Our data are not sufficient to support a general recommendation of addingUS to mammography in dense breasts, which should be based on the results ofongoing randomized trials (ACRIN, RIBES). Nevertheless we confirm that add-ing US in dense breast may be useful in the current practice, though associatedwith substantial costs.

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 161C D E FA G

B-110 11:42

Screening in women at high heriditary risk for breast cancer and history ofprevious of breast and/or ovarian cancerP. Viehweg, K. Kast, D. Haase, K. Friedrich; Dresden/DE([email protected])

Purpose: To evaluate the different breast imaging modalities for screening inwomen considered to be at high genetic risk for developing breast cancer.Methods and Materials: Thus far, a total of 34 women were enrolled in the ongo-ing screening programme for a mean period of 26.4 months. Entry criteria re-quired the patient to have: a previous history of breast cancer (n = 30), ovariancancer (n = 3) or both (n = 1) and to be either a proven BRCA 1/2 germ linemutation carrier (n = 11) or in women tested negative to have a calculated Cyrillicheterozygote risk > 20% (n = 23). The standardized protocol consists of multimo-dality examination (biannual clinical breast examination and ultrasound, annualmammography and MR imaging). Images were read prospectively and independ-ently by two radiologists.Results: A total of 8 lesions were detected in six patients. Four invasive cancersand one high risk lesion (ADH) were identified. Of the invasive cancers 3 (75%)were detected by clinical examination, 2 (50%) by mammography, 4 (100%) byultrasound and 4 (100%) by MR imaging. Tumor stage was pT1b (n = 2) and pT2(n = 2), high grade (n = 3). Both patients with contralateral breast cancer werefound to be lymph node-negative. Three of the four affected women are BRCA 1mutation carriers.Conclusion: In our first round of screening cancer detection rate was significant4/34 (11.8%). Ductal carcinoma in situ was not observed. Half of the cancerswere mammographically occult. Annual MR imaging did not enable additionalcancer diagnosis compared to conventional methods alone.

10:30 - 12:00 Room P

Computer Applications

SS 205CAD and advanced image processingModerators:M.E. Roddie; London/UKS. Weum; Tromso/NO

B-111 10:30

Evaluation of a new automatic shape tracking method for left ventricular(LV) function in multi-slice computed-tomography (MSCT)P.G.C. Begemann, F. Stahmer, F. Weiss, G. Adam, C. Lorenz, J. von Berg;Hamburg/DE ([email protected])

Purpose: Cardiac MSCT produces accurate LV function parameters such asend-diastolic volume (EDV), end-systolic volume (ESV), stroke-volume (SV), andejection fraction (EF) in comparison to MR imaging. Nevertheless, the commonsemi-automatic short axis method is time consuming. This study evaluates a newautomatic shape tracking method in comparison to the short axis method using aporcine model.Methods and Materials: 60 cardiac CT-datasets (16x0.75 mm, 120-140 kV, 100-500 mAs) were acquired for eight intubated healthy pigs on different days. Imag-es were reconstructed with slice thickness and increment of 2 mm every 10% ofthe cardiac cycle. LV function was evaluated using the common short axis meth-od and the automatic shape tracking method, where a three-dimensional triangu-lated deformable surface model was used to segment the endocardial border ofthe left ventricle and to track its motion through the cardiac phases. Results werecompared using the Bland-Altman-plot and the correlation coefficient.Results: All 60 data-sets could be post processed. Excellent correlation coeffi-cients were found for all LV parameters (EDV: 0.74, ESV: 0.79, SV: 0.77, EF:0.89). The mean difference comparing the short axis method to the automaticshape tracking method was -4 ± 10.9 ml for EDV, 0.9 ± 3.8 ml for ESV, -4.7 ± 8.8 mlfor SV, and -1.8 ± 3 ml for EF.Conclusion: This new automatic shape tracking method supplies us with reliablevalues for LV functional parameters in comparison to the common semi-automat-ic short axis method.

B-112 10:39

CT colonography: Multi-reader multi-case (MRMC) assessment ofperformance with and without computer-assisted-detection (CAD)S. Halligan1, D. Altman2, S. Mallett2, S. Taylor1, D. Burling1, M. Roddie1,J. Dehmeshki1; 1London/UK, 2Oxford/UK ([email protected])

Purpose: Computer-assisted detection (CAD) software for CT colonography hasrecently gained regulatory approval. Sensitivity for polyp detection is adequateusing internal cross-validation, but the most rigorous assessment is by multi-reader multi-case (MRMC) analysis of clinical decision making.Methods and Materials: 113 MDCT colonography examinations were accumu-lated: 63 (55.8%) cases had polyps, 50 (44.2%) were normal; all were endoscop-ically validated. Polyp coordinates were established by 3 experts using endoscopicand histologic findings. 10 readers, fully trained and accredited in abdominal CTbut unfamiliar with colonography underwent familiarisation and then read the 113cases using conventional software, identifying polyps using a 100 point confi-dence scale. After temporal separation, cases were re-read using CAD software(MedicColon 1.2, Medicsight PLC), which alerted the readers to potential polyps.Responses were categorised as true-positive (TP), true-negative, false-positive(FP), false-negative on a per-patient and per-polyp basis via comparison with theknown polyp coordinates.Results: Of 1130 individual responses, 311 (27.5%) were deemed positive with-out CAD, range 39/113 (35%) to 19/113 (17%). 261 (85%) of these were TP, sothat 41% of 630 TP cases were correctly identified, range 16/63 (25%) to 33/63(52%). With CAD, 427 (37.8%) cases were deemed positive, range 60/113 (53%)to 33/113 (29%). Overall, 357 (84%) of these were TP, so that 57% of 630 TPcases were correctly identified with CAD (relative increase 39%), range 27/63(43%) to 46/63 (73%). A MRMC ROC analysis will be presented.Conclusion: CAD increases sensitivity for non-expert CT colonography readerswithout increasing the FP rate.

B-113 10:48 !Polyp detection in virtual colonoscopy: 2D versus 2D + CADR. Ferrari1, M. Rengo2, M. Anzidei2, V. Vergari2, A. Laghi1, R. Passariello2;1Latina/IT, 2Rome/IT ([email protected])

Purpose: The aim of our study was to evaluate the advantages of reporting VCusing a colon CAD added to primary 2D approach compared to simple primary2D approach.Methods and Materials: Three radiologists with different levels of experience(expert, intermediate, non-expert) blindly evaluated 50 cases chosen from a stand-ard population of FOBT+ subjects. A primary 2D approach was performed on aVital Images workstation equipped with a Vitrea 3.7 software; 3D reconstructionwas performed for problem solving. A further 2D approach with the help of a CADsystem was performed on a Medicsight workstation equipped with Colon CAR1.3 software. We evaluated per patient accuracy, sensitivity and specificity foreach radiologist as well as mean reporting time.Results: No differences in accuracy, sensitivity and specificity for expert andintermediate radiologists using the two approaches were detected. Improvementin accuracy (94% only 2D approach, 100% for 2D + CAD) and sensitivity (80%only 2D approach, 100% for 2D + CAD) with the same results for specificity (100%)was revealed in the non-expert radiologist. Mean reporting time was 13 min forprimary 2D approach and 7 min 30 sec for 2D + CAD system.Conclusion: In our preliminary experience 2D + CAD approach does not im-prove accuracy, sensitivity and specificity for expert radiologists, but it reducesmean reporting time. Accuracy and sensitivity improve for the non-expert radiol-ogist.

B-114 10:57

CAD performance on cancer containing mammograms initially not detectedby the radiologistsA. Malich1, S. Schmidt2, D. Fischer3, M. Facius2, C. Marx4; 1Nordhausen/DE,2Jena/DE, 3Berne/CH, 4Greiz/DE ([email protected])

Purpose: Can CAD-systems detect mammographically non-detected malignantbreast lesions on priors and if so, how early?Methods and Materials: All available prior mammograms (PM) of patients suf-fering from breast cancer later detected at our department were retrospectivelyanalyzed by CAD and by radiologist with full knowledge (and availability) of themalignancy containing mammogram (MCM), that yielded cancer detection. 278mammograms (97 patients) were analyzed in consensus in order to verify wheth-er a lesion was already visible earlier. Images were analyzed using Second Look6.0 (iCAD) and correlated to MCM. The earliest mammogram with the correct

Scientific Sessions

162 C D E FBA G

detection of the (later diagnosed) malignancy was used for statistical analysis.Time intervals PM-MCM was scored: G1:< 0.5; G2:< 1y; G3:< 1.5y; G4:< 2 y;G5:< 3y; G6:> 3y.Results: 53/97 and 17/97 patients had a retrospectively visible mass and micro-calcification, respectively. 30/53 masses were detected by CAD in 1 and 4 in bothviews (64%). 8/17 MC were detected in 1 and 7/17 in both views (88%). Timeinterval vs. detection rate for masses was G1: 10/14 (71%), G2: 4/7 (57%), G3: 3/5 (60%), G4: 5/7 (71%), G5: 3/9 (33%), G6: 9/11 (81%); for microcalcificationsG1: 1/2, G2: 3/4, G3: 1/1, G4: 2/2, G5: 2/2, G6: 6/6.Conclusion: About half of the initially non-detected cancer containing mammo-grams had a mammographically visible lesion, being not suspicious according toBI-RADs. CAD detected the majority of such MC and masses. Earliest detectionsof later diagnosed malignancies were dated more than 3 years prior to the radio-logical detection. Thus, CAD might be a promising tool in this application.

B-115 11:06

Clinical role of a CAD-system for dynamic analyses of breast lesions in MRimagingA. Malich1, K. Huskobla2, J. Boettcher2, D.R. Fischer3, A. Hansch2, R. Gorna1;1Nordhausen/DE, 2Jena/DE, 3Berne/CH ([email protected])

Purpose: Breast MR-CAD systems automatically analyze and colour-code con-trast uptake of breast lesions. This was compared to ROI-based radiological dy-namic calculation.Methods and Materials: 110 patients undergoing breast MR imaging (6/2003-12/2003) with histopathological verification were included. Wash in was differen-tiated into < 50% (CAD: No colour coding), 50-100% and > 100% uptake; laterenhancement type into continued enhancement, plateau and wash out, as deter-mined by 3 radiologists and compared to CAD and histopathology. Worst curveevaluation analysis by CAD (Confirma, USA) was used.Results: 11/110 examinations were not transferable to CAD (inoptimal data stor-age, incomplete data). Out of 117 malignant and 58 benign lesions enhancementwas judged to be congruent in 71/117 and 23/58 lesions regarding wash in and74/117 and 20/58 regarding later enhancement type (radiological interpretationvs. CAD result). Radiologists scored enhancement of 5 malignancies/DCIS asunsuspicious, CAD only 1 (thus CAD detected 4 more malignancies/DCIS). Be-nign lesions were scored as rapidly enhancing in 20 cases (radiologist) and 19cases (CAD); wash out rate was 9 (radiologists) and 19 (CAD) (thus CAD detect-ed 10 more false positive cases). Using updated software transferability increased,volume calculation was possible in all except one lesion, and dynamic codingwas similar; no miscalculations of cardiac enhancement occurred.Conclusion: CAD offers for the very first time a reliable, automated analysis ofenhancement types within a breast lesion. Analysis of enhancement types is notalways congruent among radiologists and radiologists vs. CAD, suggesting a high-er rate of pathological enhancement types in benign lesions using CAD; CAD ishowever more sensitive for any enhancement, including that of premalignancies.

B-116 11:15

3D semi-automated segmentation and volumetry of lymph nodes in patientswith malignant melanomaM. Fabel-Schulte1, H. von Tengg-Kobligk1, F.L. Giesel1, L. Bornemann2,V. Dicken2, H. Naeher1, S. Delorme1, H.-U. Kauczor1; 1Heidelberg/DE,2Bremen/DE ([email protected])

Purpose: Malignant melanoma is a tumor with rising incidence, frequently asso-ciated with multiple lesions throughout the body. Therapy monitoring requiresreliable imaging and post processing methods for measuring tumor burden andtherapy response. The RECIST criteria are the current standard for therapy mon-itoring of solid tumors, with inherent disadvantages. This pilot study evaluates thefeasibility of volumetry as a more accurate measurement.Methods and Materials: Multislice-CT was performed in 25 patients with meta-static malignant melanoma, covering neck, chest, abdomen and pelvis with 1 mmcollimation using a standardized imaging protocol. Post-processing was performedusing semi-automated 3D-volumetric analysis software (OncoTreat, MeVis, Ger-many). Suspicious lymph nodes (LN), (n = 120) were evaluated by two independ-ent readers using volume [ml], t ime [s], quality of segmentation[unacceptable-excellent] and number of corrections [n]. Additionally, 20 LN weresegmented manually.Results: The software performance allowed a reliable and faster volumetry thanmanual segmentation. Quality of segmentation was rated acceptable - excellentin 81% by reader one and 79% by reader two, respectively. The correlation of thevolume was highly significant between both readers (r = 0.88). 15-20% of theLNs had to be manually corrected by drawing representative ROIs. The averagetime for automated segmentation per LN was 70-100s, including the manually

corrected ones. Average segmentation time for the primary manual segmenta-tion was 180-200s per LN.Conclusion: Semi-automated 3D-volumetric analysis allows a reliable, fast andconvenient segmentation of lymph nodes and should be further evaluated formonitoring tumor burden in a larger patient population.

B-118 11: 24

An integrated software assistant for therapy monitoring in oncologyL. Bornemann, V. Dicken, J.-M. Kuhnigk, S. Krass, H.-O. Peitgen; Bremen/DE([email protected])

Purpose: To develop a prototypical software assistant for the volumetric meas-urement of metastases and tumors in lung, liver, brain and lymph nodes to sup-port planning and monitoring of chemotherapy.Methods and Materials: Hybrid segmentation algorithms have been developedthat make use of contrasts in density as well as of morphology information tosegment the different types of lesions. For follow-up examinations the task oflocating the corresponding lesion in the new scan is simplified by automatic match-ing of corresponding lesion positions. Additionally, all information about the proc-essed lesions is summarized in a report, which includes a table containingvolumes, volume growth and doubling time for each lesion.Results: A reproducibility study was performed with 96 lung metastases (mini-mum diameter 4.6 mm, average 11.1, maximum 59.6) and showed a low inter-observer variability of the volumetry (median: 0.1%) and a low inter-scan (i.e.volume of the same lesions in different scans obtained within a few minutes andmeasured by different observers) variability (median: 4.7%). A second reproduc-ibility study was carried out with 86 liver metastases processed by 3 radiologistswith the developed application and showed a low median inter-observer variabil-ity of 7 %. The average processing time for each lesion is about 1-2 sec.Conclusion: The proposed software tool is able to measure oncological lesionsin lung, liver and brain as well as enlarged lymph nodes in the same application,and has the potential to improve treatment planning and monitoring with respectto accuracy, robustness, speed, and convenience.

B-119 11:33

Neural network based segmentation of multispectral MR imaging data forbrain volume quantification in patients with multiple sclerosisA. Wismueller1, A. Meyer-Baese2, J. Behrends1, M. Blazevic1, D. Auer3,M.F. Reiser1; 1Munich/DE, 2Tallahassee, FL/US, 3Nottingham/UK([email protected])

Purpose: In the literature, the percentage of brain volume (PBV) has been iden-tified as an important marker of disease progression in multiple sclerosis (MS).The aim was to develop, test, and evaluate a neural network-based quantificationsystem for PBV by computer-assisted segmentation of multispectral MR imagingdata.Methods and Materials: MR examination was performed in six women with re-lapsing-remitting MS (aged 23-46 years, 1.5-T system, T1w±C, T2w/PDw SE,FLAIR, MT±C). After preprocessing (image registration, automatic inhomogene-ity correction), PBV was computed by automatic CSF segmentation using a three-layer feed-forward radial-basis-functions (RBF) neural network: The voxel-specificgray-level intensity spectrum forms a 7-dimensional feature vector which is clas-sified by the neural network as belonging to CSF or not. As a reference, PBVresults were compared to the conventional angle-image method which only usesT2w and PDw data. For quality assessment w.r.t. observer-induced effects, train-ing data for the neural network were defined by two different observers inde-pendently. The same holds for threshold definition using the angle-image method.Results: Neural network-based PBV computation outperformed the convention-al angle-image method as expressed by inter-/intra-observer agreement indices(British Standards Institution) of 0.971/0.982 for RBF-neural network comparedto 0.937/0.952 for angle-image. The method-induced differences proved to bestatistically significant (univariate F-test, N=6, p < 0.01). The neural network spe-cifically performed better by avoiding misclassifications in white-matter-lesionsthat can hardly be distinguished from CSF by retrieving only T2w and PDw sig-nals.Conclusion: Computer-assisted machine learning based PBV quantification usingRBF neural network segmentation based on multispectral MR imaging data yieldscompetitive results w.r.t. inter- and intra-observer agreement levels.

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 163C D E FA G

B-120 11:42 !Automated labelling system for bronchial branches for intelligent virtualbronchoscopic navigationH. Natori1, M. Mori1, H. Takabatake1, K. Mori2, T. Kitasaka2, Y. Suenaga2,J. Toriwaki3; 1Sapporo/JP, 2Nagoya/JP, 3Toyota/JP

Purpose: This study shows an intelligent image diagnosis method for automatedlabelling of the bronchial tree, for navigation at virtual bronchoscopy by 3-D chestCT.Methods and Materials: The lumen of the bronchial tree was extracted fromvolumetric chest CT data by the region growing method, and the system givesanatomical names to segmental and sub-segmental bronchi by comparing themwith the bronchial tree models prepared in advance. We applied the proposedmethod to 25 cases of chest CT images in a virtual bronchoscopy system.Results: When an operator navigated through the bronchi on the virtual bron-choscopy mode, the system gave a proper anatomical bronchial name to them.The system extracted segmental and sub-segmental bronchi in the 25 casesexamined. The accuracy of labelling assignment was improved by employing themodel deformation method. The system extracted most of the bronchial struc-tures. The system extracted 84-100% of segmental bronchi and over 64% of sub-segmental bronchi.Conclusion: This automated labelling system for bronchial branches in virtualbronchoscopy is a useful guide for navigation on real bronchoscopic examina-tion, pre-bronchoscopic simulation, and also for bronchoscopy training systems.

14:00 - 15:30 Room B

Musculoskeletal

SS 310Spine and osteoporosisModerators:I. Boric; Zagreb/HRM. Gallucci; L'Aquila/IT

B-121 14:00

Osteoporotic compression fracture of the thoracolumbar spine: Analysis ofthe relationship with pelvic insufficiency fractureJ.-H. Kong, J. Park, K. Ryu; Seoul/KR ([email protected])

Purpose: To evaluate the incidence and characteristics of osteoporotic compres-sion fracture accompanied by pelvic insufficiency fracture in patients who under-went spine MR imaging.Methods and Materials: We reviewed 160 patients (27 men, 133 women, aver-age age 69.4 years) who underwent spine MR imaging and had thoracolumbarcompression fracture. Pathologic compression fracture was excluded. Out of thesepatients, we researched the frequency of accompanying pelvic insufficiency frac-ture. Between the group with and without insufficiency fracture, we analyzed thedifferences in sex, age, number of fractures, and the existence of bone edema ornecrosis.Results: Out of the 160 patients, 17 (10.6%) of them had pelvic insufficiencyfracture. Compression fracture occurred almost 5 times more in women (133:27),but there was no significant difference in those with insufficiency fracture (15:2).According to age (50's, 60's, 70's and 80's), the incidence of insufficiency frac-ture was respectively 0%, 10.6%, 12.5%, and 20.0%. In those with single andmultiple compression fracture, the incidence of insufficiency fracture was 8/65(12.3%) and 9/95 (9.5%), showing no significant difference. If there was com-pression fracture with bone marrow edema or necrosis, then insufficiency frac-ture occurred in 6/98 (6.12%) and if there wasn't, 11/62 (17.7%).Conclusion: About 10% of the patients who had osteoporotic compression frac-ture in the thoracolumbar spine also had pelvic insufficiency fracture, with anincrease in incidence as the patients got older. When reading thoracolumbar spineMR imaging of elder patients with osteoporotic compression fracture, the possi-bility of pelvic insufficiency fracture must be taken into consideration.

B-122 14:09

MR imaging of trabecular bone: 1.5 T versus 3.0 TS. Capuani1, F. Fasano1, C. Rossi2, G.H. Hagberg1, G. Mondello1, M. Di Mario1,B. Maraviglia1; 1Rome/IT, 2Tubingen/DE ([email protected])

Purpose: The aim of this study was to compare bone marrow relaxation times(T1,

T2

and T2

*) and other NMR parameters obtained from the human calcaneus at

1.5 T and 3.0 T and to correlate these measurements with imaging peculiar pa-rameters like SNR and contrast quality.Methods: Twelve calcaneus were imaged at 1.5 T and 3.0 T with a gradient echo(Flash) sequence and with a spin echo (TSE) sequence, using identical sequenceparameters. Relaxation times were measured. SNR in corresponding sections ofthe calcaneus were evaluated and correlated with relaxation times (T1,

T2, and T2

*).Results: Qualitative analysis showed a better contrast at 3 T than at 1.5 T. Quan-titative analysis demonstrated that magnetic susceptibility effects reduce the T2

*

and, because of diffusion effects, also reduce the apparent T2 relaxation time.Conclusion: Our results explain quantitatively the SNR gain obtainable in Flashand TSE images of trabecular bone at 3 T with respect to 1.5 T. Generally, in softtissues like brain tissue, the SNR gain at 3 T with respect to 1.5 T is about 2,while in trabecular bone tissue is less than 30% in Flash images. This is becauseof the internal gradient effects due to magnetic susceptibility difference betweentrabecular compact bone and bone marrow in spongy bone.

B-123 14:18

Comparison between digital radiogrammetry (DXR), dual energy X-rayabsorptiometry (DXA) and peripheral quantitative CT (pQCT) in childrenwith chronic inflammatory bowel diseaseH.-J. Mentzel, J. Blume, D. Tuchscherer, G. Lehmann, J. Böttcher, E. Kauf,W.A. Kaiser; Jena/DE ([email protected])

Purpose: Patients with chronic inflammatory bowel disease are at increased riskof developing osteopenia and osteoporosis. The aim of this study was to evaluatea radiogrammetrically based bone densitometric technology on children and ad-olescents.Methods and Materials: 25 patients (11 female, 14 male; 10y - 18y) sufferingchronic inflammatory bowel disease were included. Bone mineral density (BMD)was estimated from a plain radiograph of the non dominant hand using the Pron-osco X-Posure System for Digital radiogrammetry (DXR). DXR-BMD and DXR-MCI were evaluated at the three middle metacarpals. BMD was also evaluated byDXA (Hologic, QDR 4500 A) at lumbar spine and femur and by pQCT at the distalradius (Stratec, XCT 900). Spearman correlation between the results of the dif-ferent methods was performed for statistical analysis.Results: There was a significant correlation (p < 0.01) between DXR-BMD andDXA-BMD at the lumbar spine (r = 0.79) and DXA-BMD at the femur (r = 0.56).DXR-MCI correlated significantly (p < 0.05) with DXA-BMD at the lumbar spine(r = 0.75) and DXA-BMD at the femur (r = 0.50), and cortical BMD at the distalradius (r = 0.43).Conclusion: This study suggests that digital radiogrammetry shows a significantcorrelation to the standard method DXA. Practicable, non-invasive DXR-technol-ogy may be a possible adjunct to give information about the development of theindividual BMD and risk of fracture in patients with chronic inflammatory boweldisease.

B-124 14:27

Interdisciplinary approach of kyphoplasty in fresh and traumatic fracture:The Heidelberg conceptG. Nöldge, M. Libicher, K. Da Fonseca, I. Grafe, C. Kasperk, G. Kauffmann,P.-J. Meeder; Heidelberg/DE ([email protected])

Purpose: To evaluate long-term results (up to 3 years) for pain relief and stabili-zation of traumatic and osteoporotic fractures treated with CPC and PMMA bykyphoplasty.Methods and Material: Between June 2002 and September 2005, 355 patientsranging from 18 to 89 years (mean: 64.8 yrs) were treated.Results: CPC (Calcibon) was used in 103 and PMMA in 252 patients.3.7 - 9.9 ml of bone cement was used in 585 vertebral bodies. 92 fresh fractures(osteoporosis /trauma) and 8 unstable fractures (internal fixation prior to kypho-plasty) were treated. Good functional results (EVOS) of the treated segmentswith pain relief were achieved in nearly 93% 48 hours after the procedure andcould be maintained in about 70% for 12 months and in 48% after 36 monthsunder permanent accompanying medical antiosteoporotic therapy. Vertebral aug-mentation of the fracture could be achieved between 1 - 3 mm in about 65% andin fresh traumatic fractures up to 4 mm. Recurrence of pain occurred in about30% (12 months) and in 65% (36 months) was due to new fractures of neighbour-ing vertebral bodies. Cement leakage could be visualized in 18.5% without neu-rological symptoms.Conclusion: The Heidelberg interdisciplinary concept of more than 3 years ex-perience seems to be very convincing for the treatment of osteoporotic and trau-matic fractures with very good initial results in terms of stabilization, pain relief,and functional and clinical outcome.

Scientific Sessions

164 C D E FBA G

B-125 14:36

Assessment of the biomechanical strength of the proximal femur in-vitro:Evaluation of multi-detector CT-scans using local topological textureproperties based on the Minkowski functionals in 3DH.F. Boehm1, T.M. Link2, R.A. Monetti3, V. Kuhn4, F. Eckstein5, M.F. Reiser1,C.W. Raeth3; 1Munich/DE, 2San Francisco, CA/US, 3Garching/DE,4Innsbruck/AT, 5Salzburg/AT ([email protected])

Purpose: To introduce a novel algorithm for computation of the Minkowski func-tionals (MF) in 3D on a local scale in multidetector-(MD)-CT images of proximalfemur specimens and to compare their performance with histomorphometric tex-ture-measures and bone mineral density (BMD) in predicting biomechanicalstrength.Methods and Materials: Twenty-five femur specimens were imaged using MD-CT with spatial resolution of 0.25x0.25 mm2 in-plane and 0.5 mm in the z-direc-tion. BMD was measured with DXA in these specimens in standard regions ofinterest (femur head, neck, total femur, trochanter). Load-to-fracture was obtainedfor all specimens using a mechanical testing-device. From the MD-CT image data,the local MF in different volumes of interest (head, neck, trochanter) as well asthe histomorphometric measures in these regions were calculated. An algorithminvolving sliding windows for extraction of a characteristic quantity from the spec-tra of local MF was used.Results: Prediction of fracture-load by DXA ranged from R2=0.61 (shaft) to 0.69(trochanter). Correlations between strength and histomorphometric measuresranged from R2=0.12 to R2=0.54. Correlations for the local MF ranged from R2=0.50(head) to R2=0.81 (neck). Prediction of load-to-fracture significantly increased toR2=0.90 when combining the MF for more than one location in the proximal femurin a multiple-regression-model.Conclusion: Structural measures applied to MD-CT images of proximal femurspecimens by algorithms considering local topological properties expressed interms of the MF in 3D can successfully predict the strength in-vitro. Performanceis superior to that of 2D histomorphometric measures and, depending on theanatomical site within the femur, is in the same range as BMD.

B-126 14:45

MR imaging of early failing metal-on-metal total hip arthroplasty withsurgical and histological correlationA.P. Toms, T.J. Marshall, J. Cahir, C. Darrah, J. Nolan, S.T. Donell, T. Barker,K.J. Tucker; Norwich/UK ([email protected])

Purpose: To perform a retrospective review of all the conventional radiographicand MR imaging studies performed in patients with early post-operative pain fol-lowing cobalt-chrome metal-on-metal total hip arthroplasty, and to correlate withoperative and histological findings.Methods and Materials: The plain radiographs and MRIs of 18 symptomaticmetal-on-metal hip replacements in 17 patients were reviewed by three MSKradiologists and correlated with operative findings, and microbiological and his-tological investigations from 8 subsequent revision arthroplasties.Results: Plain radiographs were normal in all 18 hips. The median postoperativetime to MR imaging was 25.5 months. MR imaging abnormalities were demon-strated in all symptomatic hips. These comprised: periprosthetic fluid-signal col-lections encased in an irregular low signal wall (15), periprosthetic bone marrowedema (12), muscle edema (5), avulsion or atrophy of the short external rotators(10) and fracture of the medial calcar (1). Operative findings were similar in pa-tients who have undergone revision surgery (8): extensive periprosthetic soft tis-sue necrosis, fluid-filled cavities, proximal femoral diaphyseal necrosis and pittingand corrosion of the femoral stems which were, in all cases, firmly fixed to thecement mantle. Histology revealed non-specific necrosis (6), with a perivascularlymphocytic infiltrate (1) and features of an abscess cavity (1). Microbiologicalcultures failed to yield any organisms.Conclusion: A small proportion of patients with metal-on-metal hip arthroplast-ies present with early postoperative pain. This may be due to a unique hypersen-sitivity reaction. In these patients conventional radiographs are commonly normal.MR imaging demonstrates periprosthetic abnormalities in all these patients, manyof which correlate with surgical findings.

B-127 14:54

Multidetector-row CT in patients with suspected lumbar disc herniation:Comparison of standard-dose and simulated low-dose techniquesP.J.M. Bohy1, V. De Maertelaer1, A. Roquigny2, C. Keyzer3, D. Tack2,P. Gevenois1; 1Brussels/BE, 2Baudour/BE, 3Charleroi/BE ([email protected])

Purpose: To compare standard-dose and low-dose MDCT in patients with sus-pected lumbar disc herniation.Methods and Materials: 60 consecutive patients underwent MDCT with 4x1 mmcollimation at 140 kVp. Tube current was adapted to body mass index (BMI) asfollows: 200 mAs (BMI< 22Kg/m2); 300 mAs (BMI≥ 22Kg/m2 and < 30Kg/m2); and400 mAs (BMI≥ 30Kg/m2). We simulated doses at 65%, 50%, 35%, and 20% ofthe mAs used for acquisition. During two sessions, three independent radiolo-gists coded each of three caudal discs as "normal", "bulged", or "herniated", andgraded canal and foramen compromise. For discs coded as "herniated", readersdiscriminated protrusion from extrusion. For each reading the median numbersof discrepancies between the standard dose and reduced doses were compared.Agreements within and between readers were evaluated.Results: Dose reduction had no impact on a reader's ability to accurately andconsistently identify "bulged" discs (P=0.128) and left (P=0.413) and right(P=0.665) foramen compromises. However, for "normal" discs (P=0.002), for "her-niated" discs (P < 0.001), and for canal compromise (P=0.002), dose reductiondid have a significant impact. A reduction in dose impact was not detected at65% but at 50% (P=0.004 and P=0.008, respectively) for "normal" and canalcompromise, and at 35% for "herniation" (P=0.031). Dose reduction affected thediscrimination between protrusion versus extrusion for one reader (P=0.046).Agreements within and between readers ranged from poor to excellent and tend-ed to decrease with the dose.Conclusion: For patients with suspected lumbar disc herniation, mAs settingscould be reduced to 65% of the standard dose in MDCT of the lumbar spine.

B-128 15:03

Correlation among MR imaging, contrast-enhanced-MR imaging (CE-MRI),clinical course and serological tests in seronegative spondyloarthritis (SpA)P. Simoni, P. Cimini, F.M. Martina, I. Sansoni, A. Afeltra, R.F. Grasso, I. Baeli,B. Beomonte Zobel; Rome/IT ([email protected])

Learning Objective: To compare diagnostic reliability of unenhanced MR imag-ing sequences and CE-MRI in detecting inflammatory lesions of the lumbar spineand of SIj in SpA. To determine the correlation between MR imaging findings andclinical-serological results, using MR imaging as the gold standard.Methods and Materials: We enrolled 23 outpatients with proved SpA. We per-formed a standardized CE-MRI of the lumbar spine and the SIj (TSE T1-w andT2-w sequences, with and with out contrast medium and fat-saturation on multi-ple planes) using a 1.5 T MR-unit. Images were independently evaluated by tworadiologists with a final consensus. Two different reading sessions were performedfor unenhanced and enhanced studies to assess if CE-MRI could provide signif-icant advantages with respect to the basal MR study in assessing disease activ-ity. Local clinical symptoms, serological findings (e.g. erytrosedimentation-rate,C-reactive-protein) and disease duration, were also collected and statisticallyrelated to both basal-MRI and CE-MRI findings.Results: We found active inflammatory MR imaging changes in 74% of patients;5% of these lesions were detected only on CE-MRI. No significant differenceswere observed in determining disease activity of lesions detected at both unen-hanced and contrasted images. Among MR imaging positive patients, clinicalsymptoms allowed diagnosis of acute inflammation only in 16%. Among the se-rological tests, C-reactive-protein provided the highest sensitivity (88%) but alow specificity (33%), versus 64% and 100% of erythrocyte-sedimentation-rate.Conclusion: The limited advantages of contrast medium administration don't justifyits routine use in monitoring disease activity of patients with SpA. On the basis ofMR imaging findings, serological tests and clinical findings were showed to havea weak correlation with disease activity.

B-129 15:12

Kinetics of imaging lesions in spinal tuberculosisA. Feydy1, L. Le Page2, V. Dufour3, L. Rillardon3, N. Belmatoug3, B. Fantin3;1Paris/FR, 2Amiens/FR, 3Clichy/FR ([email protected])

Purpose: Our purpose was to study the kinetics of imaging lesions of spinaltuberculosis in relation to clinical and biological features.Methods and Materials: We designed a prospective longitudinal study including19 patients with proven spinal tuberculosis. Clinical and biological data were col-lected at inclusion and every 3 months during one year. Computed tomography

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 165C D E FA G

(CT) and magnetic resonance imaging (MRI) were performed at baseline. MRimaging was performed at least twice during the 12 months treatment and at theend. MR imaging and CT were reviewed by a senior musculoskeletal radiologist.Morphological changes of abscesses, signal intensity changes of vertebral bod-ies and discs were analysed and recorded.Results: Initial MR imaging showed oedema of 75 to 100% of the vertebral bodyin all cases. All patients improved clinically except one with cervical dislocation,who died immediately after surgery. Weight gain was almost totally achieved at 6months and pain relief at 9 months. Biological inflammatory syndrome vanishedat 3 months. On MR imaging, paravertebral abscesses disappeared respectivelyat 3, 6 and 12 months in 24%, 35% and 88% of the cases. Improvement of verte-bral body bone marrow oedema with conversion to a fatty signal occurred in 30%of cases at 6 months and 53% at 12 months. Epidural abscesses disappeared at9 months in 100% of cases.Conclusion: Significant imaging lesions persist at the end of treatment despite agood outcome. Therefore MR imaging should not be used to define the durationof therapy if clinical and biological evolution are favourable.

B-130 15:21

3D-CISS MR myelography for assessment of adult traumatic brachial plexusinjuriesM.A. Png, F.C. Yong; Singapore/SG ([email protected])

Purpose: Differentiation of pre-ganglionic from post-ganglionic brachial plexusinjuries (BPI) is necessary for adequate pre-operative planning. This study re-views our experience using MR myelography, based on a 3-D constructive inter-ference in steady state (CISS) technique, to assess for pre-ganglionic injuries inclinically suspected BPI.Methods and Materials: We retrospectively reviewed the 3D-CISS myelographyscans of 9 patients with suspected traumatic BPI who underwent MR imagingbefore surgical exploration. All scans were performed on a 1.5 T Siemens Visionand were reviewed by an experienced musculoskeletal radiologist. The status ofthe C5 to T1 roots on every MR myelogram was evaluated for: intact rootlets,nerve root avulsion or injury (partial or complete), and pseudomeningocele for-mation. The MR imaging findings were recorded and correlated with the findingsat surgical exploration for each root level.Results: Altogether 35 roots evaluated had surgical correlation. For detection ofroot avulsion injury, the overall sensitivity and specificity was 100% and 71%respectively, with positive and negative predictive values of 84% and 100%. Forcomplete avulsions, the sensitivity and specificity was 88% and 71% respective-ly. Twelve of 28 (43%) proven root avulsions had no associated pseudomenin-gocele on MR myelography. In this study, all cases with pseudomeningocelesdetected by 3D MR myelography, had root avulsions found at surgery.Conclusion: Our experience shows that 3D-CISS MR myelography is accuratefor detecting root avulsions in adult traumatic BPI. The ability of this technique todemonstrate root avulsions in the absence of pseudomeningoceles significantlyimproves the accuracy of MR imaging in detecting pre-ganglionic injuries.

14:00 - 15:30 Room C


SS 305PACS: New features and opportunitiesModerators:R. Chrzan; Krakow/PLA.G. Jurik; Århus/DK

B-131 14:00

The Health Optimum project: The standards application to telemedicineworkflowC. Dario1, S. Giovannetti1, G. Bettiol1, A. Ramelli2, C. Saccavini2, F. Di Paola1;1Treviso/IT, 2Padova/IT ([email protected])

Purpose: Health Optimum is a telemedicine project to validate systems interop-erability and the management of patient documents (creation, digital signature,transmission and publication in EHR), and to disseminate the results.Methods and Materials: Health Optimum, approved and co-funded by the Euro-pean Community, began in May 2004 and will end in January 2006. It involvesSpain-Aragona (oncology, haematology, radiology, dialysis and laboratory), Den-mark-County of Funen (cardiology and endocrinology) and Italy-Veneto regionas coordinator (12 healthcare structures for neurosurgical counselling, POCT-

laboratory and regional EHR), with other technological and organisational part-ners. Neurosurgical tele-counselling wants to satisfy the need to give fast andformalized answers for urgent cases of cranial and vertebro-medullary trauma,linking central neurosurgeons and neuroradiologists to peripheral ER, neurologyand intensive-care. The form management system is based on XML-CDAv2 andHL7 standard; digital signature follows law and rules; CT images are in DICOMformat. To share telemedicine documents we choose the IHE XDS profile (CrossEnterprise Document Sharing). To manage security we use IHE ATNA profile(Audit Trail Node Authentication) and XUA profile (Cross Enterprise User Au-thentication).Results: Physicians have reorganized workflow using a telemedicine approach,and they have defined a standard clinical data set. This demonstrates that theavoidance of unnecessary transfer of patients and documents, and the use ofEHR speed up diagnostic and therapeutic decisions.Conclusion: Telemedicine allows reorganization and optimization of processesand resources, with clinical and technological standardization. National and in-ternational rules and law, medico-legal liabilities, security and privacy must bedefined and respected.

B-132 14:09

Medical image viewers: Assessment of functionality and user-friendlinessS.D. Curran, L.H. Schwartz, B. Rajashanker, H. Hricak; New York, NY/US([email protected])

Purpose: Increasingly, radiology studies are distributed on compact discs (CDs)to referring physicians for review; CDs are cheaper and a potentially better optionthan producing X-ray film, however the software viewing interfaces used differconsiderably. Therefore, we analyzed the most common image viewing softwareprograms' functionality and ease of use.Methods and Materials: We prospectively analyzed 250 consecutive CDs sentto our Institution for review over a one month period in 2005. Functionality wasassessed by analyzing 5 basic viewing tasks 1) open a case, 2) navigate throughseries, 3) change window/level settings, 4) magnification, 5) open and comparestudies.Results: The 5 most commonly used image viewers accounted for 193 (77%) ofall CDs submitted. In descending order of frequency: 42%, 19%, 9%, 5%, 2%.Fifteen (6%) of CDs had no viewer or could not be opened. The opening displayformats ranged from 1 to 16 images. Left and right mouse button functions variedin 4 of 5 viewers. Image navigation steps were different in 5 of 5 viewers. Pre-setwindow settings were found on 2 of 5 viewers. Objective global functionality scoresranged from 26 to 38, mean 32 (out of 50) for the 5 most common viewers.Conclusion: There is great variability in commonly used functions among imageviewing programs. Additionally, many programs are not intuitive and are difficultto use. This variability, lack of standardization, and poor functionality hampers aphysician's ability to assess radiology studies, interrupting workflow and routineclinical care.

B-133 14:18

ASP for managing the PACS archiveE. Talini, A. Bertini, G. Cardia, R. Dell'Osso, D. Caramella, C. Bartolozzi;Pisa/IT ([email protected])

Purpose: Picture Archiving and Communication Systems (PACS) have provedable to assist radiologists by streamlining the digital workflow. However, the man-agement of PACS complex technology may cause additional strain to the hospi-tal personnel. Application Service Provisioning (ASP) may be a solution to thisissue, by means of the total or partial outsourcing of PACS functions. At our Insti-tution we evaluated the feasibility of using ASP for managing the PACS archive.Methods and Materials: Four high-volume DICOM modalities were configuredto send diagnostic exams to a cache server (Centricity Archive, GEMS) that for-warded them to a remote GEMS PACS by using a broadband wide area network.In order to evaluate the performance in different clinical and networking situa-tions we monitored the system when up to 14 reporting workstations were con-currently used and by setting the network speed at 2 Mbps, 4 Mbps, 6 Mbps,10 Mbps, and 100 Mbps.Results: Our assessment was done on the basis of objective measurement andof subjective evaluation collected by means of detailed questionnaires complet-ed by radiology residents. We found that with a network speed of 6 Mbps thenumber of concurrent reporting sessions ceased to have an impact on the delayin retrieving exams from the remote ASP archive.Conclusion: Our experience suggests that the PACS archive can be managedwith ASP. We found that no significant impact is perceived on the radiologicalworkflow, provided that a guaranteed bandwidth of 6 Mbps is available.

Scientific Sessions

166 C D E FBA G

B-134 14:27

Increased quality of patient care and balanced workload for technologists:Leveraging industry tools to analyze and optimize patient flow in a fulldigital regional hospitalH. Bartel1, E. Pisar1, A. Mueller2, P. Ahrens1, S. Liedke1, L. Mikolajewski1,K.-H. Zerbst2; 1Düsseldorf/DE, 2Stuttgart/DE([email protected])

Purpose: Analyze "current state" and identify waste in target process. Develop"future state" vision and implement optimized process within one week.Methods and Materials: Florence Nightingale Hospital in Kaiserswerth (Germa-ny) has a complete digital environment (PACS/RIS/HIS). However, patient wait-ing time to perform X-ray procedures was excessive and technologists' workloadswere not balanced. In a "Gemba" session workflow specialists observed the sit-uation, then analyzed RIS data and developed the "Current State Value StreamMap" and the "Future State Map "necessary to create the new process. "TargetSheets" showing major measurements/targets gave full visibility to the team. Af-ter training "on the job" a KAIZEN was developed, piloted and deployed by theradiology staff, implementing the new process.Results: Thanks to the "KAIZEN" tool, changes were implemented in one weekand improved performance is constantly monitored. The team moved from a "batch-ordering" process via "intercom" to a "single-piece-flow" of patients scheduledand picked-up by staff. Patient waiting queue was reduced from 3 to 1, waitingtime was significantly decreased; cycle-time was reduced by 60%, resulting in apotential 50% patient throughput increase as technologists workload was bal-anced. Results were obtained with no investment in additional resources/tech-nology; on the contrary, future department expansions may leverage the freedresources.Conclusion: Significant improvements can be implemented quickly provided staffunderstand and share the objectives. The KAIZEN concept was taught, under-stood and leveraged by staff to define, test and implement the new process di-rectly "on the job", which was successfully "up and running" within one week.

B-135 14:36

Introducing a redundant mini-PACS as a backup and migration tool in analready existing large PACS environmentU. Fronz, M. Nuefer, M. Forsting; Essen/DE ([email protected])

Purpose: Beginning in April 2002 a PACS was introduced into our radiology de-partment, supporting our already existing RIS. In March 2004 we installed a smallersystem as a redundant mini-PACS to support backup and upgrade needs in thelarge PACS system. The purpose of this study was to evaluate the advantagesand disadvantages of such a system in a large radiology department.Methods and Materials: Since 1999 we have used Medora from Innomed, Ger-many, as our RIS. In December 2001 we decided to introduce PACS Centricityfrom GE Medical Systems, USA, into our department. In March 2004 we intro-duced a second smaller system for backup and upgrade purposes in the originalPACS. This smaller PACS consists of an EMC CX200 computer system equippedwith 1.2 TB storage capacity in RAID technology. Based on the mini-PACS, re-ports can be generated and web-distribution of images and reports is possible.Results: We already used this system for two migration steps to upgrade thenormal PACS system. We also used it in one system failure with a downtime of 3hours. The mini-PACS proved to be fully functional and allowed the radiologydepartment to examine patients, generate reports and distribute the images andreports in the same way as with the normal PACS system.Conclusion: Running the redundant mini-PACS in a large scale environmentproved to be feasible and became useful in case of errors or upgrades within thenormal PACS system. We think that a filmless hospital needs to maintain a back-up and migration system.

B-136 14:45

Usability of speech recognition in a trauma hospital: One year follow-upstudyT. Kauppinen, J. Ahovuo; Helsinki/FI ([email protected])

Purpose: It has been demonstrated earlier by several study groups that speechrecognition (SR) reduces the time between imaging and final report. Because SRaffects the whole radiological process and turnaround, it would be important tototal the number of reports which are created by SR. The purpose of this workwas to evaluate usability and usage of SR in one radiological pilot unit, and try tofind reasons why the number of reports created by SR is not 100%. The aim wasalso to follow and collect experiences during this follow-up study.Methods and Materials: The radiological dictation process was followed in Töölö

trauma hospital in Helsinki, Finland, which has emergency unit working 24/7. Atthe beginning of this follow-up study three radiologists started to use SR and atthe end of follow-up time all senior radiologists are able to dictate using SR.Results: It was found by this study that the attitude of radiologists affects theusage of SR. At first the proportion of reports which were dictated by SR com-pared to the conventional way was 0%, after 2 months use by the first three usersthe proportion was over 50 %, and after 3 months over 60%.Conclusion: It was noticed that SR changes workflow and increases the qualityof the reports. The proportion of reports created by SR increased after the firstmonths. The total number of workstations, usability and dictation environmenthave an impact, increasing the usage.

B-137 14:54

Impact of PACS/RIS integrated speech recognition on report availability in athree-year follow-up studyC.G. Trumm1, C. Glaser1, C. Grosse1, B. Küttner1, S. Nissen-Meyer1, P. Popp2,M.F. Reiser1; 1Munich/DE, 2Erlangen/DE([email protected])

Purpose: To assess the long-term effect of a PACS/RIS integrated speech rec-ognition system on radiology report availability (RA) in a university hospital overa 3-year period.Methods and Materials: In a retrospective PACS/RIS database analysis, reportturnaround times (RTT) were calculated for 3 comparable 0.5 year time periods:1. 01-06/2003 (tape-based dictation with transcription), 2. 01-06/2004 (0.5 yearexperience with speech recognition system [SRS]), 3. 01-06/2005 (1.5 year ex-perience with SRS). Only remote (RSR) and online (OSR) speech recognitionwith transcriptionist and self-correction respectively were used in period 2 and 3.Cumulative RA (ratio of available reports vs. total number of examinations) within2/6/12h after examination completion was calculated for period 1/2/3 respective-ly including all modalities. For period 3, cumulative short-term RA was differenti-ated into reports available 10/30/60 minutes after completion of conventionalradiography (CR), computed tomography (CT) and ultrasound (US) examinations.Results: 33155/45136/60209 datasets were analyzed for time period 1/2/3. In-cluding all modalities, the number of reports available in less than 24h was 3796/13125/36031 respectively. Within 2/6/12h after examination completion cumula-tive RA was 2.3/7.9/11.2% for period 1, 14.9/24.5/28.3% for period 2 and 31.6/54.5/59.1% for period 3. Cumulative short-term RA after 10/30/60 minutes was4.7/19.3/23.8% for CR, 1.3/12.2/18.7% for CT and 6.2/31.3/47.3% for US exam-inations.Conclusion: Institution-wide use of a PACS/RIS integrated SRS facilitates a con-tinuous improvement of report availability over a 3-year period. Reasons are anincreasing acceptance of OSR by radiologists and a shift from RSR to OSR fordictation of routine reports.

B-138 15:03 !The status of structured reporting in imaging: Special focus on remotereportingH.K. Pohjonen1, P. Ross2, J.G. Blickman3, L. Christensen4; 1Espoo/FI,2Tallinn/EE, 3Nijmegen/NL, 4Svendborg/DK ([email protected])

Purpose: Structured reporting has been studied for at least 10 years to increasereport creation speed and the clarity, communicative value, and overall quality ofthe imaging information supplied for a modern hospital. Today, the digital work-place knows no boundaries: Remote reporting, out-of-hours coverage and the'virtual' radiologist are all ways to offer imaging services in a non-traditional way.The purpose of the paper is to study the status of structured reporting with aspecial focus on cross-border eRadiology.Methods and Materials: The presence of lists and hierarchical relationships,coded or numeric content in addition to plain text, relationships between con-cepts and embedded references to images and similar objects is characteristic ofstructured reporting. The study involves a survey on commercially available re-port templates in general, and especially those supporting DICOM structuredreporting. Special attention is paid to templates in multilingual reporting acrossborders.Results: To our knowledge there are no commercially available clinical templatesthat support DICOM structured reporting. A few companies exist - especially inNorth-America - that offer structured reporting as a service, but these are notimplemented in DICOM. A few research-oriented projects do exist however thataddress structured reporting in the DICOM world. The Baltic eHealth project usesstructured multilingual templates in remote reporting between Denmark, Estoniaand Lithuania.Conclusion: Structured reporting and multilingual templates are powerful tools

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 167C D E FA G

for remote reporting across borders. It enables the utilization of radiologist man-power from different countries, while using their own native language. Commer-cial applications to support these tools, however, are lacking.

B-139 15:12

Taking advantage of PACS/structured reports resources to generateteaching filesA. Delistamatis1, M. Mantatzis1, N. Tapus2, P. Prassopoulos1;1Alexandroupolis/GR, 2Bucharest/RO ([email protected])

Purpose: To develop a web-based system for teaching files (TF) creation usinginformation derived from DICOM structured reports (SR) documents and DICOMimages from a PACS.Methods and Materials: The TF system retrieves all the information from diag-nostic reports, in the form of SR documents, from a PACS to store it in a localdatabase via a synchronization procedure. Queries are applied to the systemlocal database using keywords related to specific pathology, imaging findings,diagnostic procedures or techniques for the selection of cases with educationalvalue. The system automatically retrieves from the PACS the images correspondingto the selected cases/group of cases. Representative images along with the di-agnostic reports, commentaries and MESH-keywords are edited to form the teach-ing file that is stored in the final teaching file database.Results: Review of teaching files is included in every-day training of medicalstudents and radiology residents accessing the TF system database through aweb-interface in the university intranet. In addition, the system incorporates aninterface for possible integration with the RSNA MIRC web network. The TF sys-tem expands by 2-3 new teaching files per day generated by attending radiolo-gists.Conclusion: The proposed TF system offers a platform for teaching files creationutilizing information already stored in a PACS.

B-140 15:21

A filmless radiology department in a full digital regional hospital:Quantitative evaluation of the increased quality of patient careA. Nitrosi1, G. Borasi1, F. Nicoli1, A. Botti1, G. Modigliani2, P. Notari1, E. Costa1,S. Bronzoni1; 1Reggio Emilia/IT, 2Cinisello Balsamo/IT([email protected])

Purpose: To quantitatively measure, assess and document the business andorganizational benefits that result from the transition to digital imaging.Methods and Materials: Reggio Emilia regional hospital installed PACS as thefinal step towards a completely digital clinical environment, completing the HIS/EMR and 1200 web/terminals for patient information access. Financial benefitsthroughout the hospital were assessed upfront and measured every 4 monthsafter the system's deployment. Key indicators (radiology exam turn-around-time(TAT), number of radiology procedures performed, inpatients length of stay be-fore and after the PACS implementation, etc). were analysed and values werestatistically tested to assess workflow and productivity improvements. All rawdata used for the analysis was extracted from the hospital HIS and finance sys-tems.Results: The hospital went "filmless," with the exception of analogue mammog-raphy in 28 weeks and within its original budget. Overall radiology departmentproductivity increased by 15%, TAT improved by more than 65%. Timelier patientcare resulted in decreased lengths of stay. Neurology alone experienced an 11%improvement in average patient stay. Annual financial upsides have exceeded€1.5 millions/year and PACS is being extended to 5 smaller county hospitals witha common clinical data repository that will be shared with the entire region nextyear.Conclusion: A well-planned PACS deployment simplifies imaging workflow andimproves patient care throughout the hospital while delivering substantial finan-cial benefits. Staff buy-in was key in this process and on-going training and proc-ess monitoring are a must.

14:00 - 15:30 Room E2

GI Tract

SS 301Disorders of the lower GI tractModerators:J. Olliff; Birmingham/UKS. Schmidt; Lausanne/CH

B-141 14:00

Imaging in acute abdominal pain: Can low-dose computed tomographyreplace plain abdominal radiographs?H. Geijer, B. Persson, P. Ingverud, K.-W. Beckman; Örebro/SE([email protected])

Purpose: In acute abdominal pain, plain abdominal radiographs are often ob-tained to identify bowel obstruction and free intra-abdominal gas. The purpose ofthis study was to compare low-dose computed tomography (CT) with plain ab-dominal radiographs in this setting, using the same radiation dose. The radiationrisk for the patient is thus the same with both methods.Methods and Materials: Fifty-eight patients were imaged with both methods.The effective dose for plain abdominal imaging was calculated at 1 mSv withMonte Carlo simulation. Low-dose CT was performed at 120 kV and 20 mAs/slice, giving the same effective dose. All cases were reviewed independently bythree radiologists. A gold standard was obtained by reviewing medical records,laboratory data and further radiological imaging.Results: There was good agreement between the two methods for diagnosis ofbowel obstruction. In three cases the cause of obstruction was found on low-dose CT (inguinal hernia, a low bowel obstruction and an abdominal wall hernia).Three cases were correctly diagnosed as having a diverticulitis and one as hav-ing a pancreatitis using low-dose CT while these diagnoses could not be madeon abdominal radiographs. One case of inflamed bowel was diagnosed on plainabdominal imaging but missed on low-dose CT. One case of diverticulitis andone appendicitis were missed with both methods. No case of free intra-abdomi-nal gas was found.Conclusion: Low-dose CT using the same radiation dose as plain abdominalradiographs is well suited as a replacement. The diagnostic potential is increasedand more information is gained.

B-143 14:09

The multislice CT-enteroclysis in bowel endometriosis detectionE. Biscaldi, S. Ferrero, V. Remorgida, G.A. Rollandi; Genova/IT([email protected])

Purpose: To demonstrate the effectiveness of multislice CT-enteroclysis (MSCTe)in detecting bowel endometriosis wall involvement.Methods and Materials: 40 patients (aged from 25 to 46) with ovarian endome-triosis already known (by MR imaging or laparoscopy) and symptoms suggestingcolorectal endometriosis (pelvic pain, dyspareunia, tenesmus) were evaluated.The colon was distended with methylcellulose and water (50%) before the con-trast enhanced MSCT (16 rows, collimation thickness 0.625 mm, reconstruction1.25 mm, pitch 1,375, 120 KV, 330 mAs; smart mAs; 1.5 cc per Kg of patientweight of iodinated c.m. (Iodixanol 320 mg I/100 ml) at a flow/rate of 3 cc/s, witha delay time of 40s after the arterial peak). Within 20 days after the radiologicalexamination, independently from the findings of MSCTe, all women underwentlaparoscopy.Results: Out of 40 women evaluated, 3 were found to have a normal colon onMSCT: These findings were confirmed at surgery. In 37 cases the MSCTe detect-ed the site of endometriosis and the colorectal wall involvement depth; the deep-er nodules infiltrated the serosa in 13 cases, the muscularis in 17, the submucosain 7. In one patient localisation in the last ileal loop was identified. One patientrefused surgery. In 8 cases the CT detected some lymphoadenopathy near thecolon.Discussion and Conclusion: The MSCTe is an effective method to detect theinvolvement of intestinal tracts by endometriosis and to evaluate the depth ofpenetration, for a preoperative assessment. Studies with a large number of pa-tients are needed in order to validate these preliminary findings.

Scientific Sessions

168 C D E FBA G

B-144 14:18

Assessment of disease activity of ulcerative colitis with magneticresonance imaging (MRI), in comparison with endoscopyF. Maccioni, S. Parlanti, F. Siliquini, G. Luppi, A. Bruni, F. Carrozzo; Rome/IT([email protected])

Purpose: To evaluate the degree of inflammatory activity in UC with MR imaging.Methods and Materials: Forty-five patients with UC underwent MR imaging.Results were compared with endoscopic data. MR imaging was performed afteroral administration of a superparamagnetic contrast-agent, with colon air insuf-flation, using T2-weighted and T1-weighted sequences, before and after intrave-nous gadolinium injection on axial and coronal planes. MR images were evaluatedby two radiologists blind of endoscopic results. The following parameters wereassessed and graded (0-3); extent of lesions, bowel wall thickness, T1 wall gado-linium-enhancement, T2 wall signal and T2 signal outside of the wall. MR imag-ing score of UC activity ranged between 0-15. In particular; 0-1 non active, 2-4moderately active, 5-10 active and 11-15 severe active pancolitis. Spearman non-parametric correlation was used for statistical analysis.Results. In 16/45 patients MR imaging was unable to find evident signs of UC(quiescent inactive UC). In 29 patients MR imaging detected signs of UC. TheMR imaging score was statistically correlated with the endoscopic score of activ-ity; bowel wall thickness (r:0.76), T1 wall gadolinium-enhancement (r:0.84), T2wall signal (r:0.84) and extent of colonic disease. In 12% of patients MR imagingshowed unsuspected complications (tight strictures, adhesions, 2 ovarian involve-ments). In 25% of patients MR imaging significantly affected the surgical plan-ning.Conclusion: MR imaging gives reliable information on the site and extension ofUC lesions in active or severe disease, whereas is unable to characterize thedisease in quiescent forms. MR imaging can play an important role in the man-agement of UC patients, particularly when endoscopy is contraindicated or diffi-cult to perform.

B-145 14:27

Ultrasonography in the diagnosis of typhoid feverM.A. Mateen, S. Saleem, P.C. Rao, D.N. Reddy; Hyderabad/IN([email protected])

Purpose: To determine the ultrasound diagnostic criteria in cases of TyphoidFever.To establish the efficacy of ultrasound of the abdomen as a diagnostic test inTyphoid Fever.Methods and Materials: The Widal test is the most commonly used method ofdetecting Typhoid fever, but does not provide results until a week after the onsetof fever due to the need for enough antibodies to develop to render a positiveresult. Abdominal Ultrasound was performed within three days of the onset offever in 80 cases suspected to be having Typhoid fever. Subsequently all 80 cas-es were found to be Widal positive. Subsequent follow-up scans were performedat five days, ten days and fifteen days.Results: The US findings were as follows: splenomegaly (n-80,100%); Bowelwall thickening (n-69, 86.25%); mesenteric lymphadenopathy (n-55.68.75%);hepatomegaly with normal parenchymal echotexture (n-22.27.5%); thickened gallbladder (n-53.66.25%); biliary sludge (n-8.10%); positive US Murphy's sign (n-14.17.5%); pericholecystic edema with increased vascularity (n-8 ,10%); mucos-al ulceration in the wall of the gall bladder (n-1.1.25%).Conclusion: In endemic areas like India, ultrasound findings of hepatomegaly,splenomegaly, ileal and caecal thickening, mesenteric lymphadenopathy and thick-walled gallbladder are diagnostic features of typhoid. Ultrasound can be a nonin-vasive, economical and reasonably sensitive tool for diagnosing Typhoid especiallywhen serology is equivocal and cultures are negative.

B-146 14:36

Magnetic resonance imaging for prediction of tumor-free circumferentialresection margins and long-term survival in patients with rectal cancerJ.C. Stollfuss, R. Rosenberg, H. Wieder, A. Sendler, K. Becker, K. Woertler,E.J. Rummeny; Munich/DE ([email protected])

Purpose: To investigate the prognostic significance of an involved circumferen-tial resection margin (CRM) on local recurrence and long-term survival predictedby Magnetic Resonance (MR) imaging before neoadjuvant treatment and totalmesorectal excision (TME) in rectal cancer.Methods and Materials: 68 patients (52male, 16 female; mean age58.9 ± 9.4years) with locally advanced tumors (cT3/4, N0-2, M0) at initial presen-tation were included. Contrast-enhanced and T2w MR-images were analyzed by

two radiologists according TNM categories. The observers also assessed theMR-scans for the shortest distance from the outermost part of the tumor to theadjacent mesorectal fascia (as the potential CRM in TME) at the level of themaximum depth of penetration through the rectal wall. Histopathology andfollow-up data was available in all cases. Mean follow-up time was 46 (8-79)months. The rates for 5-year-survial and local recurrence were calculated in threesubgroups depending on the minimum distance to the mesorectal fascia meas-ured by MR imaging (I: ≤ 1 mm; II: 2-5 mm; III: > 5 mm).Results: MR imaging accurately predicted a histologically involved CRM (Sensi-tivity: 1.00 (4/4; Specificity: 0.88 (56/64)). The rates for 5-year-survival (I: 0.40, II:0.74, III: 0.90, p < 0.001) and local recurrence (I: 0.41, II: 0.04, III: 0.03, p < 0.01)were significantly different between the subgroups depending on the distance tomesorectal fascia. There was no significant difference with respect to histomor-phological response to neoadjuvant treatment (I: 33%, II: 45%, 47%, n.s) or dis-tant metastases during follow-up (I: 33%, II: 27%, 15%, n.s).Conclusion: MR imaging allowed accurate identification of important prognosticrisk factors in patients with rectal cancer prior to neoadjuvant therapy and surgi-cal resection. This may provide better selection of patients undergoing preopera-tive treatment.

B-147 14:45

MR imaging directed multidisciplinary team pre-operative treatmentstrategy: The way to eliminate positive circumferential margins?G. Brown, S. Burton, I. Daniels, B. Mason, A. Norman, D. Cunningham;Sutton/UK ([email protected])

Introduction: Histopathological audit of positive circumferential margins (CRM)can be used as a surrogate for the success of rectal cancer treatment. We audit-ed CRM involvement in rectal cancer patients and the impact of the multidiscipli-nary team (MDT) on implementing an MR imaging-based pre-operative treatmentstrategy.Method: A database of consecutive rectal cancer patients treated over 4 years inour network were analysed for MR imaging and histopathology assessment ofCRM and MDT meeting treatment decisions. The CRM positive rate of thosediscussed at MDT versus those not discussed were compared. We re-auditedthe CRM positive rates one year after introducing a policy of mandatory preoper-ative MR imaging-based MDT discussion.Results: 298 patients were diagnosed with rectal cancer of which 39 were deemedpalliative, 178 underwent surgery alone and 81 underwent neoadjuvant therapy.Of these, 62/178 patients underwent surgery alone without MR imaging-basedMDT discussion, resulting in positive CRM in 16 cases (26%) as compared to 1/116 (1%) in those patients with MDT discussion of MR imaging. Overall CRMpositive rate, in all potentially curative patients with or without MDT discussionwas 12.5% (32/256) significantly lower than the < 20% rate (p < 0.001) quoted innational guidelines. Re-audit in 98 consecutive patients following a change ofpolicy produced a lower CRM positive rate of 3% (1/37) for all surgery alonepatients and an overall CRM positive rate of 7% (5/70).Conclusion: MDT discussion of MR imaging and implementation of a preopera-tive treatment strategy results in significantly reduced positive CRM in rectal can-cer patients.

B-148 14:54

Accuracy of FDG PET/CT in the detection of recurrent rectal cancerT. Kau, P. Reinprecht, P. Lind, M. Starlinger, K.A. Hausegger; Klagenfurt/AT([email protected])

Purpose: To assess the additional value of combined 18 F-FDG PET/CT in re-staging following resection of rectal cancer.Methods and Materials: Thirty-nine examinations in twenty-two patients wereevaluated retrospectively. All patients underwent surgical therapy before referralto PET/CT. CT and PET components were interpreted separately by one readereach, followed by a consensus reading. Sites of increased FDG uptake as wellas PET/CT findings were categorized as benign (1), equivocal (2) or malignant(3). Standard of reference was histology or clinical and imaging follow-up for atleast 6 months.Results: Sensitivity, specificity, positive and negative predictive value, and accu-racy for differentiating benign (14/30) from malignant (16/30) uptake sites in thesmall pelvis were 100%, 64%, 76%, 100%, and 83% for PET/CT, and 100%,29%, 62%, 100%, and 67% for PET, respectively. Regarding extrapelvic abnor-malities, PET/CT was able to distinguish benign (29/71) from malignant (42/71)with a sensitivity, specificity, positive and negative predictive value, and accuracyof 98%, 90%, 93%, 96%, and 94%, compared to 98%, 72%, 84%, 95%, and 87%for PET. Interpretation of PET images was false positive in a total of five sites,

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 169C D E FA G

however, equivocal in 14 sites. PET/CT was false-positive in both a single pelvicand extrapelvic site, and equivocal in seven. The rare case of an FDG uptakingadrenal adenoma will be presented.Conclusion: PET/CT is valuable in the assessment of recurrent malignancy fol-lowing surgical removal of rectal cancer. Particularly, this modality should ex-clude cancer recurrence suspected by PET interpretation alone in a considerablenumber of patients.

B-149 15:03

Assessment of colorectal tumour perfusion using MDCT: Comparison ofdistributed parameter analysis and Patlak analysisV.J. Goh1, S. Halligan2, C.I. Bartram3; 1Northwood/UK, 2London/UK,3Harrow/UK ([email protected])

Purpose: Quantitative perfusion techniques are advocated for assessing tumourperfusion, in particular to assess the effects of vascular modulating therapy. How-ever, analysis methods are not standardised causing problems for comparisonacross studies. We aimed to determine the level of agreement between tumorblood volume and permeability measurements using two commercially availableCT perfusion analysis methods.Methods and Materials: 20 patients (mean age 67.7 years; range 35 to 87 years,10 male and 10 female) with proven colorectal cancer were examined prospec-tively using a four detector row CT (LightspeedPlus, GE Healthcare technolo-gies). A 65-second perfusion study (4 X 5 mm slice thickness; 1 acquisition/second)was acquired through the tumour epicentre following intravenous contrast ad-ministration (100 mls 340 Niopam; 5 ml/s). Tumour blood volume and permeabil-ity was determined using two analysis methods: Distributed parameter analysisand Patlak analysis. Measurement agreement between perfusion parameters wasassessed using Bland-Altman statistics.Results: The mean and standard deviation (SD) for blood volume and permea-bility were 6.4 (1.3) ml/100 ml and 14.5 (3.4) ml/100 ml/min respectively usingdistributed parameter analysis and 11.3 (3.1) ml/100 ml and 17.8 (9.4) ml/100 ml/min respectively using Patlak analysis. The mean difference (95% limits of agree-ment) were -4.9 (-10.6 to +0.71) ml/100 ml and -3.3 (-22.7 to +16.2) ml/100 ml/min for blood volume and for permeability respectively. The within subject coeffi-cient of variation was 45.5% and 44.6% for blood volume and permeability re-spectively.Conclusion: Commercially available CT perfusion software may produce dis-crepant results. The level of agreement between methods must be taken intoaccount in cross comparison studies.

B-150 15:12

Magnetic resonance imaging in the assessment of neoadjuvantradiotherapy and chemotherapy in the treatment of rectal carcinomaD.J. Tuite, F. Smith, R. Stephens, M. Keogan; Dublin/IE([email protected])

Background: With the development of both Total Mesorectal Excision (TME) andneoadjuvant chemotherapy and radiotherapy in the treatment of rectal cancer,the role of MR imaging in local staging has never been more important. Accurateassessment of rectal carcinoma and its relationship to the mesorectal fascia isessential for appropriate patient selection for TME. Tumours that involve the me-sorectal fascia are associated with poor outcome. However, if such patients showa response to neoadjuvant chemotherapy and radiotherapy improvements in sur-vival are achieved.Methods and Materials: Twenty patients with rectal carcinoma were recruitedinto the study. All patients had either T3 disease or local lymphadenopathy withinthe pelvis. Patients received adjuvant chemotherapy and/or radiotherapy. Pre andpost therapy MR imaging studies were performed. The following parameters wereassessed; maximal tumour thickness, distance to circumferential resection mar-gin, position, size, and number of lymph nodes. In addition all lymph nodes werecategorized as group1; homogenous and smooth, or Group 2; irregular and/orheterogeneous. All results were correlated with the histopathological findings atthe time of surgery in all cases. Tumour regression scores were calculated byassessing the amount of residual tumour as compared with the amount of radia-tion induced fibrosis.Results and Conclusion: MR imaging proved highly accurate in assessing nod-al downstaging post neoadjuvant therapy. There was strong statistical correlationbetween nodal downstaging and the biological response of the tumour to therapy.MR imaging was not accurate in assessing T-stage downstaging post neoadju-vant therapy.

14:00 - 15:30 Room F1

Chest

SS 304Airways and lung parenchymaModerators:C. Schaefer-Prokop; Amsterdam/NLK. Steinke; Basle/CH

B-151 14:00

Visualisation of ex-vivo porcine bronchial wall layers with intraluminaloptical coherence tomography (OCT): Comparison with histologicalspecimenR. Eibel, M. Jaeger, G. Babaryka, H. Bitterling, O. Meissner, M.F. Reiser,U. Mueller-Lisse; Munich/DE ([email protected])

Purpose: To compare the delineation of the different bronchial wall layers in OCT,based on infrared light, with histology as the standard of reference.Method and Materials: Two investigators independently evaluated 124 sectionsof 31 ex-vivo porcine bronchial specimens. The OCT probe (0.014 inch diameter)was introduced into the bronchi via a 4 French catheter. After OCT experiments,the bronchi were cut into axial slices, fixed in formalin, H&E stained and evaluat-ed by a pathologist. The thicknesses of the different bronchial wall layers weremeasured and comparisons between the two methods were performed (BlandAltman Test).Results: The mean values of the bronchial lumen were 5.2 mm in OCT and 4.9 mmin histology. The tunica mucosa (TMUC) with ciliated epithelium was detectableby OCT as a medium-intense line with a mean thickness of 0.14 mm (0.11 mm inhistological specimen). The high signal tunica muscularis (TMUS) (mean thick-ness 0.24 mm in OCT, and 0.16 mm in histology) could be separated from thelow signal adventitia (0.23 mm in OCT, and 0.17 mm in histology). Visualizationof the cartilage in the adventitia, that produced no reflexions, was possible in all124 sections. Taking into account the magnification factor in OCT images due toshrinkage of the tissue by fixation, there was no significant difference in detec-tion of the different layers between the two methods.Conclusion: Intraluminal OCT allows a nearly microscope-like resolution of thedifferent bronchial wall layers in vitro. Therefore, utilization of this method in pa-tients appears very promising.

B-152 14:09

Free-breathing volumetric CT lung imaging using prospective non-contactgating in a large animal experimentJ. Zaporozhan1, S. Ley1, R. Unterhinninghofen1, Y. Saito2, M. Fabel-Schulte1,G. Szabo1, H.-U. Kauczor1; 1Heidelberg/DE, 2Tochigi/JP([email protected])

Purpose: Investigation of prospective respiratory gating for CT of the lung usingtwo prototype gating devices in a large animal experiment and evaluation of im-age quality.Methods and Materials: Eight anesthetized and ventilated healthy pigs under-went MDCT (1 mm, 120 kV, 300 mA) with prospective respiratory gating. Twonon-contact gating devices were used: (1) CCD camera in all animals and (2)laser sensor in five pigs. The output signal of both gating devices was connectedto the scanner instead of the ECG gating unit. Under free-breathing conditionsthe image acquisition was set to maximum inspiration and expiration. The qualityof images and MPRs was analyzed in axial, coronal and sagittal orientation forsharpness of bronchi, diaphragm, lung fissures, parenchyma and pericardial struc-tures using a 4-point-score (1 - excellent to 4 - severe artifacts).Results: Prospective respiratory gating worked perfectly well for inspiration andexpiration in all animals when using the CCD camera, and a little less convinc-ingly with the laser sensor: 100 % for inspiratory, but only 60% for expiratoryscans. All acquired images showed excellent sharpness (CCD camera vs. lasersensor) for bronchi (1.3 ± 0.5 vs. 1.6 ± 0.7), lung fissures (1.0 vs. 1.1 ± 0.3), pa-renchyma (1.0 ± 0.2 vs. 1.4 ± 0.6) and diaphragm (1.5 ± 0.8 vs. 1.9 ± 0.9). Thepericardial lung structures often showed minor to major cardiac motion artifacts(1.9 ± 0.6 vs. 2.3 ± 0.5).Conclusion: Free-breathing volumetric CT of the lung using prospective non-contact respiratory gating is feasible with high image quality. The CCD cameraprovided better results and slightly higher image quality than the laser sensordevice.

Scientific Sessions

170 C D E FBA G

B-153 14:18

Correlation between lung attenuation with high resolution computedtomography and COPD severity grading systemsM. Allona, I. Torres, B. Rodríguez-Vigil, F. García-Río, M. Pardo, V. Lores;Madrid/ES ([email protected])

Purpose: The severity of patients with COPD is usually graded with the GOLDindex, and recently with the BODE score. Our aim is to evaluate the accuracy ofquantitative HRCT in discriminating different severity grades of COPD obtainedwith these two indexes.Methods and Materials: Fifty-six emphysematous patients, all current smokers,underwent both HRCT and pulmonary function tests following a standardisedprotocol. Severity of COPD was classified according to the stages of the GOLDscore (depending on the forced expiratory volume in one second, FEV1) and thequartiles of the BODE index (based on FEV1, MRC scale, six-minute-walk-test,and BMI). For the statistical analysis, differences in attenuation values in HRCTbetween both severity grading systems were assessed by ANOVA with post-hoccomparisons by Bonferroni test.Results: No patients were included in the GOLD stage I. Seventeen patientswere included in stage II, 20 in III, and 19 in IV. For the BODE index, 17 were inthe quartile 1, 20 in the quartile 2, 13 in the quartile 3 and 6 in the quartile 4.Sig-nificant differences between GOLD stages II and III in the upper and the lowerlobes were found, and also between stages II and IV in the lower lobes. For theBODE index, only significant differences between quartiles 1 and 4 were found inthe lower lobes.Conclusion: HRCT allows differentiation of GOLD severity grades, but is onlycapable of detecting very severe disease with the BODE index (quartile 4).

B-154 14:27

High-resolution CT of the chest in cystic fibrosis: Is simplification ofscoring systems reliable?A. Oikonomou1, E. Hatziagorou2, M. Mantatzis1, J. Tsanakas2,P. Prassopoulos1; 1Alexandroupolis/GR, 2Thessaloniki/GR([email protected])

Purpose: The proposed scoring systems for the assessment of cystic fibrosis(CF) on HRCT, although accurate, are cumbersome and time-consuming, thusnot widely applied in every day clinical practice. The aim of this study is to simpli-fy these scoring systems without compromizing their reliability, by selectingrepresentative variables.Methods and Materials: Forty-two consecutive patients with CF (mean age: 12years) underwent a baseline and a follow-up chest HRCT (mean interval: 31months). Three radiologists deciding by consensus, evaluated 84 HRCTs, em-ploying 5 scoring systems (Bhalla, Hellbich, Santamaria, Brody, Oikonomou) -each one with 7 to 11 variables - resulting in 5 baseline and 5 follow-up "scores"for each patient. Subsequently four variables were selected (severity and extentof bronchiectasis, severity of thickening, air-trapping) from each scoring systemto form 5 new shorter "subscores". Each "subscore" was correlated with the cor-responding complete score, using Pearson correlation, and the baseline scores -subscores were compared with the follow-up ones.Results: There was excellent correlation between each score and its correspond-ing subscore (0.922<r< 0.0979, p < 0.0001). Strong correlation was found amongstthe 5 scoring systems (0.925<r< 0.0995, p < 0.0001) and the 5 subscores(0.828<r< 0.0993, p < 0.0001). Comparing the baseline scores or subscores withthe corresponding follow-up ones, significant worsening was observed(0.001<p < 0.0001), in agreement with patients' clinical deterioration.Conclusion: A simple and reliable HRCT scoring for routine use in clinical prac-tice may result by selection of representative variables from established scoringsystems.

B-155 14:36

Influence of low dose in MDCT of the chest on quantitative evaluation ofsevere emphysema: A simulation studyJ. Zaporozhan1, S. Ley1, O. Weinheimer2, R. Eberhardt1, I. Tsakiris1, Y. Noshi3,F. Herth1, H.-U. Kauczor1; 1Heidelberg/DE, 2Mainz/DE, 3Tochigi/JP([email protected])

Purpose: Using low dose protocols with higher noise levels, quantitative evalua-tion might be impaired or unreliable. Effects of low dose protocols on quantitativeemphysema evaluation were investigated using a raw data simulator.Methods and Material: 30 patients suffering from severe centrilobular emphyse-ma underwent MDCT (120 kV,150 mAs,1 mm). Original CT raw data were simu-lated using ten mAs settings (10-100 SIMmAs). Quantitative analysis provided

lung volume (LV), emphysema volume (EV), emphysema index (EI), MLD andfour emphysema clusters with different volumes (class 1-4). Simulated low doseresults were compared to original acquisition, given as percent variation.Results: Mean LV was 7.1 ± 1.4L and was significantly different (p < 0.05) onlyfor 20 and 10 SIMmAs. MLD was equal down to 20 SIMmAs. Mean EV was3.8 ± 1.4L and differed only up to 3.5 ± 3.2% from original data, though was sig-nificantly different for all dose settings. Mean EI was 52 ± 10% and showed vari-ation up to 3.4 ± 3.4%. EI values below 60 SIMmAs protocols were significantlydifferent from the original data set. Volume of large emphysema (class 4) clusterswas 3.6 ± 1.4L and differed up to 0.2 ± 1.2% with no significant difference until50 SIMmAs level. Although the absolute values of intermediate (class 3) andsmall clusters (class 2 and 1) was very small, the values showed mean variationup to 33-72% and were significantly different from 50 SIMmAs, 80 SIMmAs and90 SIMmAs. Down to 50 SIMmAs the intermediate and small clusters' volumeremained below 10% variation.Conclusion: Dose reduction down to 60 SIMmAs does not significantly alter 3Demphysema quantification results.

B-156 14:45

Is routine emergency room portable chest radiography in stable multipletrauma patients justified?G. Bartal1, B. Kessel2, B. Chaushu2, A. Breitgand2, U. Soimu2, R. Alfici2;1Kfar-Saba/IL, 2Hadera/IL ([email protected])

Purpose: According to ATLS, portable chest radiography is mandatory in anymultiple trauma patient, and is performed following initial clinical evaluation andresuscitation. Today, ultrafast multidetector CT scanners are readily available andwidely utilized in the evaluation of stable patients. Our aim was to assess a justi-fication of the portable chest radiography in stable multiple trauma patients with-out clinical evidence of chest injury.Methods and Materials: During 2004 109 consecutive stable multiple traumapatients, without clinical evidence of chest injury were admitted to the traumaresuscitation area of Hillel Yaffe Medical Center. Four certified radiologists (GB,BC, AB, and US) retrospectively evaluated and compared portable chest filmsand chest CT angiographies (CTA) of these patients. We recorded each casewhen the treatment policy was altered due to the results of either imaging modal-ity.Results: Mean injury severity score (ISS) of the patients was 19.6. Portable chestradiographs were normal in 92/109 (84.4%) and abnormal in 17/109 (15.5%) ofthe patients. Treatment policy was changed based on portable radiography find-ings in only 2 (1.87%) of the cases. Clinically significant injuries were diagnosedon CTA in 32/109 (29.9%), of which 7/32 (21.9%) were defined as potentially fataland required immediate intervention.Conclusion: Portable chest radiography in stable multiple trauma patients withblunt mechanism of injury did not change the therapeutic policy in our patients.Based on our results we believe that a current practice of routine portable chestradiography in these patients should be reevaluated.

B-157 14:54

T1 mapping of the entire lung in patients with emphysema and fibrosis:Influence of respiratory phase and correlation with lung function testsA. Stadler1, P.M. Jakob2, M. Barth1, E. Eisenhuber1, A.A. Bankier1; 1Vienna/AT,2Wuerzburg/DE ([email protected])

Purpose: To quantify the T1 of lung parenchyma of patients with fibrosis andemphysema measured in the entire lung, and to determine the effect of inspira-tion and expiration.Methods: Twelve patients with emphysema and 14 with fibrosis underwent MRimaging Investigation of the lung using a SNAPSHOT-FLASH sequence. Theacquired series of images were used for calculation of T1 relaxation times. Acqui-sition was performed both in inspiration and expiration. FEV1, total lung capacity,residual volume and vital capacity and blood gas tests were measured.Results: For the emphysema group, the average T1 value in inspiration was1034 ± 71 ms. The average of the mean T1 values in expiration was 991 ± 63 ms.The average T1 in expiration was thus 43 ms shorter than in inspiration. Thisdifference was statistically significant (P=0.021).For the patients with fibrosis, the average T1 value in inspiration was 996 ± 103 ms.Compared to that, the average T1 value in expiration was 1282 ± 170 ms. Theaverage T1 in expiration was 286 ± 152 ms longer than in inspiration. This differ-ence was statistically significant (P < 0.0001).Linear regression of T1 versus lung function parameters showed the highest re-gression coefficients for TLC and RV in expiration.Conclusion: Our study documents reverse respiratory phase dependence of T1

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 171C D E FA G

measurements of the entire lung parenchyma in patients with emphysema andfibrosis. Furthermore, expiratory measurements showed higher and reverse dif-ferences between patient groups compared to inspiratory measurements. Ourfindings underline the strong but nonuniform influence of the inspirational statusduring T1 measurements of the lung.

B-158 15:03

HRCT patterns in antisynthetase syndrome versus polymyositis/dermatomyositisR. Dore, V. Vespro, F. Bosoni, D.M. Savulescu, A. Valentini, L. Cavagna,M. Montecucco; Pavia/IT ([email protected])

Purpose: Antisynthetase syndrome (AS) is part of poly/dermatomyositis syn-dromes, characterized by the presence of antibodies against aminoacyl-tRNAsynthetases. The incidence of interstitial lung disease is higher (50%-80%) in ASthan in other PM/DM syndromes, but its prognosis is usually better in the former.Aim of this study is to verify if HRCT pattern is different between AS and PM/DM.Methods and Materials: We retrospectively evaluated 16 patients who under-went HRCT examination, 8 affected by AS and 8 by PM/DM (or overlapping dis-ease). HRCT was performed with 16 multidetector CT machine (Sensation,Siemens) with 1 mm slice thickness. We considered these findings: ground-glassopacity (GGO), honey-combing (H-C), septal lines, BOOP-like pattern, lymph-nodes, dilation of esophagus, enlargement of right heart or pulmonary artery.Results: GGO was present with the same incidence in the two groups (4/8 pa-tients with AS and 4/8 patients with PM/DM), but it was more extensive in ASpatients. GGO decreased substantially or completely resolved after treatment inall AS patients (100%), while only two patients in PM/DM group responded totreatment (50%). Moreover, one unresponsive patient in the PM/MD group devel-oped an important pulmonary fibrosis. Frequency of H-C was found in 6/8 pa-tients with AS (50% associated with GGO), 4/8 in other PM/DM (only 25%associated with GGO). Finally, we observed more lymphadenopathy in AS andmore dilation of the esophagus in overlapping than in other syndromes.Conclusion: In our experience GGO in AS is more extensive than in PM/DM, butit usually has a better response to therapy.

B-159 15:12

Prevalent radiological appearance of invasive pulmonary aspergillosis inthe immunocompromised patients after implementation of aspergillus PCRin the broncho alveolar lavage specimenO.R. Brook, E. Hardak, I. Oren, M. Yigla, H. Sprecher, A. Engel, L. Guralnik;Haifa/IL ([email protected])

Purpose: We examined how implementation of a new technique - aspergillusPCR in BAL specimens in immunocompromised patients influenced the domi-nant radiological appearance of invasive pulmonary aspergillosis.Methods and Materials: All consecutive immunocompromised patients that un-derwent BAL for pulmonary complications between 09/1998 and 12/2003 wereincluded in the study. One attending and one resident radiologist retrospectivelyreviewed chest CTs performed prior to the BAL. CT studies were acquired on 2,4 and 16-slice scanners. The review of the CT studies was performed withoutknowledge of clinical background. Final clinical diagnosis was determined byEORTC criteria and clinical results. Aspergillus PCR testing was introduced inour institution in November 2002. We compared the radiological picture of thedisease before and after implementation of the aspergillus PCR.Results: Twenty-one (26%) of 82 patients were diagnosed with invasive pulmo-nary aspergillosis before the advent of PCR testing (group I) and 22 (36%) of 62patients after it (group II). The patients of two groups did not differ significantly intheir clinical characteristics, but radiological findings differed significantly: in thegroup II CT revealed more lung nodules 11-30 mm in diameter with a halo (59%vs. 29%, p = 0.05), more focal ground glass (32% vs. 5%, p = 0.03), but lessdiffuse ground glass (0% vs. 19%, p = 0.04), less cavitation (0% vs. 33%,p = 0.005) and less consolidation (5% vs. 48%, p = 0.002).Conclusion: After the implementation of aspergillus PCR testing in the BAL spec-imen the prevalent radiological picture is that of focal disease, which might repre-sent an earlier stage of the disease.

B-160 15:21

Clinical relevance of high resolution multislice spiral-CT of the chest afterstem cell transplantationC. Weber, J. Pantzek, U. Wedegaertner, N. Kröger, H. Klose, A. Zander,G. Adam; Hamburg/DE ([email protected])

Purpose: To evaluate high resolution multislice computed tomography (HRM-SCT) in detection and monitoring of pulmonary complications after stem cell trans-plantation (SCX).Material und Methods: 80 patients (27 F/53M) with fever of unknown origin afterSCX, mean age 40y (18-62), were examined prospectively by 4 detector-rowHRMSCT. 18/80 patients received basic HRMSCT before SCX due to a high riskprofile. 77/80 patients received a follow-up HRMSCT after SCX (< 60days). Asgold standard for pneumonia a multifactorial system including bronchoscopy, bloodchemistry and microbiological data (bronchoalveolar lavage, sputum, titre, bloodculture) was established. Analysis of clinical and radiological data was performedby two investigators in a consensus procedure.Results: 40/80 patients developed pulmonary infiltrates in HRMSCT. Sensitivity,specificity and diagnostic accuracy of HRMSCT in detection of pneumonia was53%, 56% and 55% (8/40 aspergillus, 8/40 candida, 2/40 cytomegalovirus, 1/40pneumocystis carinii virus, 1/40 adenovirus, 3/40 staphylococcus, 1/40 pneumo-coccus, 16/40 unspecified). Diagnostic accuracy increased to 83% in the follow-upHRMSCT. Patients with a high risk profile showed a higher prevalence for pulmo-nary infiltrates and complications (p < 0.05). Highest diagnostic accuracy in themultifactorial system for revealing pneumonia was shown for the bronchoalveolarlavage with 48%. HRMSCT induced a change in therapy in 42/80 patients (36/80extended antibacterial and antifungal therapy etc).Conclusion: Increasing diagnostic accuracy in the follow-up HRMSCT and as-sociated therapeutic consequences allow clinical use of HRMSCT for detectionand monitoring of pulmonary complications as the imaging method of choice inthe complex multifactorial system of diagnosing pneumonia after stem cell trans-plantation.

14:00 - 15:30 Room F2

Cardiac

SS 303Coronary artery recanalization: Collaterals, stentsand bypass graftsModerators:D. Exarchos; Athens/GRK.-F. Kreitner; Mainz/DE

B-161 14:00

MR imaging for the assessment of collateral perfusion in patients withchronic occlusions of coronary arteriesA.M. Huber, M. Schweyer, O. Mühling, M. Näbauer, S. Schönberg, M.F. Reiser;Munich/DE

Purpose: The aim of the study was to investigate the extent of myocardial infarc-tions, perfusion defects and viable, hypoperfused myocardium dependent on thedegree of collateralisation.Methods: 26 patients with chronic coronary artery occlusions were examined byMR imaging. The MR imaging examination at 1.5 T included imaging functionwith a CINE SSFP technique, imaging perfusion at stress (140 µg/kg/min adeno-sine) and at rest with a saturation recovery gradient echo technique and viabilitywith a IR gradient echo technique. The volume and transmural extent of the my-ocardial infarct as well as the size of the entire perfusion defect and the viablehypoperfused myocardium were determined. The coronary angiography exami-nations were evaluated using the Rentrop Scale (0-1: group1, bad collateralisa-tion; 2-3: group 2: good collateralisation).Results: 12 patients were assessed as group 1 and 14 patients were assessedas group 2. The mean values for the volume of myocardial infarctions were 18.3 mland 8.4 ml, respectively, p < 0.001. The mean value for the transmural extent ofmyocardial infarction was 97 % (group 1) and 37 % (group 2), p < 0.00001. Themean percentage hypoperfused volume of LV myocardium with a MPRI < 1.4was 25 % (group 1) and 17 % (group 2), p < 0.05. Hypoperfused, viable myocar-dium was observed in 66 % of the patients in group 1 and 100 % of group 2.Conclusion: Patients with good collateralisation of chronic occlusions have small-er infarctions and perfusion defects compared to patients with bad collateralisa-tion, however a higher probability of hypoperfused, viable myocardium.

Scientific Sessions

172 C D E FBA G

B-162 14:09 !Non invasive assessment of collateral blood flow with multidetector helicalCT coronary angiographyM. Sammarra, P. Pavone, I. Sansoni, R.F. Grasso, P. Simoni, C. Di Capua,C. Catalano, R. Passariello; Rome/IT ([email protected])

Purpose: The existence and recruitment of collateral vessels within the coronarycirculation is one of the factors determining the clinical relevance and develop-ments of symptoms in coronary atherosclerosis. The purpose of our study was toprospectively evaluate diagnostic accuracy of multi detector row CT coronaryangiography for the detection of collateral coronary circulation.Methods: Out of a population of 283 patients submitted to multi detector row CTbetween November 2004 and August 2005, 21 patients with complete occlusionof at least one main vascular segment (anterior descending, circumflex and rightcoronary artery) were included in this study. The patients were 13 men and 8women, mean age 63 years (age range 44-78). Selective coronary angiographywas performed in all patients within 4 days from multi detector row CT evaluation.Images were evaluated for the presence of significant re-filling of vessels distal tothe occluded segment. The type of collateral circulation (ipsilateral vs controla-teral) and vessels involved in the collateral filling was also recorded.Results: In 15 patients filling of the segment distal to occlusion was observed,due to collateral circulation (Right Coronary artery n = 10, LAD n = 1 and circum-flex artery n = 3). Collateral circulation was directly evident in MDCT in 9 patientsand non evident in 6.Conclusion: Using MDCT, it is possible to get information on collateral circula-tion that are critical in the decision making regarding the performing of coronaryangioplasty.

B-163 14:18

64-slice multidetector coronary CT angiography: In-vitro evaluation of 68different stentsD. Maintz1, H. Seifarth1, R. Raupach2, T. Flohr2, M. Rink1, T. Sommer3,M. Özgün1, W. Heindel1, R. Fischbach1; 1Münster/DE, 2Forchheim/DE,3Bonn/DE ([email protected])

Purpose: To test a large sample of different coronary artery stents using 64-slicemultidetector row CT in vitro.Methods and Material: 68 different coronary artery stents (57 stainless steel, 4cobalt-chromium, 1 cobalt-alloy, 2 nitinol, 4 tantalum) were examined in a coro-nary artery phantom. Stents were imaged in axial orientation with standard pa-rameters: 32x0.6 collimation, pitch 0.24, 680 mAs, 120 kV, rotation time 0.37 s.Four different image reconstructions were obtained varying convolution kernelsand section thickness: 1. soft, 0.6 mm, 2. soft, 0.75, 3. medium soft, 0.6, 4. stentoptimized sharp, 0.6. To evaluate visualization characteristics of the stent, thelumen diameter, intraluminal density and noise were measured.Results: The high-resolution kernel offered significantly better average lumenvisualization (57% ± 10%) and more realistic lumen attenuation (222 HU ± 66 HU)at the expense of increased noise (15.3 HU ± 3.7 HU) compared to the soft andmedium-soft CTA protocol (p < 0.001 for all). Stents with a lumen visibility of morethan 66% were: Arthos pico, Driver, Flex, Nexus2, S7, Tenax complete, Vision (all67%), Symbiot, Teneo (70%), Radius (73%). Only ten stents showed a lumenvisibility of less than 50%.Discussion: Stent lumen visibility largely varies depending on the stent type.Even with the improved spatial resolution of 64-slice CT, a stent optimized kernelremains beneficial for stent visualization when compared to the standard medi-um soft CTA protocol. Using 64-slice CT and high resolution kernel the majorityof stent products show a lumen visibility of more than 50% of the stent diameter.

B-164 14:27

In vitro evaluation of metallic coronary artery stents with 64-rowmultidetector CT using an ECG-gated cardiac phantomG. Savino1, J.U. Schoepf1, C. Philip1, T.J. Vogl2, C. Herzog3; 1Charleston,SC/US, 2Frankfurt/DE, 3Kronberg/DE ([email protected])

Purpose: To assess the accuracy of 64-row multidetector CT in the visualizationof different coronary artery stents and the appraisal of in-stent stenosis.Methods: 5 different coronary stents types with diameters 2.5-4.0 mm were an-alyzed in an anthropomorphic dynamic cardiac phantom. All stents were mount-ed on polyurethran sticks (220 HU) of defined outer diameter and containing adefault concentric stenosis (70 HU) of 50% each. Imaging was performed at heartrates 60 bpm, 75 bpm and 90 bpm and with the sticks/stents being placed in asagittal-oblique orientation. Scanning parameters were: 64x0.3 mm collimation;0.75 mm slice thickness; 0.4 mm increment; 330 ms gantry rotation time; 165 mstemporal resolution; retrospective ECG-gating.

Results: In-stent stenosis was correctly identified in all cases, showing signifi-cant (p < 0.05) overestimation of the degree of stenosis (64.9% ± 24.6 vs.50% ± 0), underestimation of the stent lumen (1.6 mm ± 0.6 vs. 3.1 mm ± 0.6)and increase in in-stent attenuation (231 HU ± 27 vs. 200 HU ± 13) for all stentsand heart rates. Level of heart rate and stent type had no significant effect onresults. Diagnostic accuracy was significantly influenced by stent diameter, show-ing significantly (p < 0.05) less stenosis overestimation (15.3 mm ± 4) and stentdiameter underestimation (1.8 mm ± 0.2) in 4.0 mm stents than in 3.0 mm(23.1 mm ± 6/2.2 ± 0.2) or 2.5 mm (38.4 mm ± 6/3.2 mm ± 0.2) stents. In-stentattenuation was significantly (p < 0.05) lower in 4 mm (210 HU ± 3.4) and 3 mm(212 HU ± 3.4) stents than in 2.5 mm stents (225 HU ± 3.5).Conclusion: In an experimental setting 64-row MDCT allows for a reliable detec-tion of in-stent stenosis but significantly overestimates the degree of stenosis.Within the range of physiological heart rates diagnostic accuracy is rather re-stricted by spatial than by temporal resolution.

B-165 14:36

Assessment of coronary artery stent patency and restenosis using 64-slicecomputed tomographyC. Rist1, F. von Ziegler1, K. Nikolaou1, B.J. Wintersperger1, T. Johnson1,S.B. Reeder2, A. Knez1, M.F. Reiser1, C.R. Becker1; 1Munich/DE, 2Stanford,CA/US ([email protected])

Purpose: A new generation of high-resolution 64-slice Computed Tomography(64SCT) is now available. We compared the diagnostic accuracy incorporating64SCT for the feasibility of assessing the lumen of coronary artery stents. Rest-enosis remains a major limitation of coronary catheter-based stent-placement.Therefore, a reliable non-invasive diagnostic method for the evaluation of stentedcoronary arteries would be highly desirable.Methods: 25 consecutive patients after coronary artery stent placement under-went both 64SCT of the coronary arteries and quantitative X-ray coronary angi-ography (QCA). For coronary CT angiography, a 64SCT system (SOMATOMSensation 64, Siemens Medical Solutions, Forchheim, Germany) was employed,using the following parameters: spatial resolution, 0.4x0.4x0.4 mm; temporal res-olution, 83-165 ms; contrast agent, 80 cc; flow rate, 5 cc/s; retrospective ECG-gating. The MSCT-scans were analyzed for image quality and for signs of significantin-stent stenoses and pre- and post-stent stenoses. Sensitivity, specificity anddiagnostic accuracy of 64SCT in the detection or exclusion of significant sten-oses were evaluated.Results: 43 of 44 (98%) evaluated stents were of diagnostic image quality. Sen-sitivity, specificity, and diagnostic accuracy of 64SCT for the correct detection of≥ 50% diameter in-stent stenoses were 80% 97% and 93%. For the detection ofsignificant pre-stent stenoses 64SCT showed a sensitivity, specificity and diag-nostic accuracy of 75%, 95% and 93%, for post-stent stenoses 67%, 85% and84%, respectively.Conclusion: 64 SCT with improved spatial and temporal resolution shows prom-ise as a clinical tool for reliable, non-invasive diagnosis or exclusion of significantin-stent and peristent stenosis.

B-166 14:45

In-vitro evaluation of coronary stents and in-stent-stenoses using a 64-detector row CT scannerT. Schlosser1, O.K. Mohrs2, A. Magedanz2, B. Nowak2, T. Voigtländer2,A. Schmermund2, J. Barkhausen2; 1Essen/DE, 2Frankfurt/DE([email protected])

Purpose: Despite substantial advances in multi-detector CT (MDCT) coronaryimaging, the detection of coronary in-stent restenosis is currently not possible.We used an ex-vivo phantom model with a realistic CT density pattern, artificialstenoses, and dynamic movement mimicking the cardiac cycle to examine theability of 64-slice MDCT to detect in-stent stenoses in coronary stents.Methods: Four different stents (Liberté and Lunar ROX, Boston Scientific; Driver,Medtronic; Vision, Guidant) were examined using the dynamic cardiac phantommodel. The stents were placed on a polymer tube with a diameter of 3.0, 3.5, or4.0 mm. Different degrees of stenosis (0%; 30%; 50%; 75%) were created insidethe tube.Results: All stents lead to artificial reduction of attenuation, the least degree ofwhich was found in the Liber té (11.3 ± 10.2 HU) and the Vision stent(17.6 ± 17.9 HU; p = 0.25). A significantly higher reduction of attenuation wasfound in the Lunar ROX (53.5 ± 24.9 HU) and the Driver stent (36.2 ± 7.1 HU;p < 0.05). In a reconstruction interval thought to represent the diastolic phase ofthe cardiac cycle, it was possible to differentiate high-grade stenoses (> 70%) inall stents. In the Liberté stent high-graded stenoses were correctly diagnosed in100% of all cases (Driver: 83%; Vision: 66%; Lunar Rox: 33%).

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 173C D E FA G

Conclusion: Our ex-vivo data show that non-stenotic stents and low-grade in-stent stenoses can be reliable differentiated from intermediate and high-grade in-stent stenoses in vessels with a diameter of 3 to 4 mm. With regard to artifactsand grading of stenoses, the Liberté stent was best suited for CT coronary ang-iography.

B-167 14:54

Drug eluting coronary stent: Detection of restenosis with 64 MDCTAI. Carbone, E. Algeri, K. Lanciotti, M. Francone, A. Granatelli, A. Ascarelli,S. Cagioli, C. Catalano, R. Passariello; Rome/IT ([email protected])

Purpose: To evaluate the ability of 64 MDCT for the detection of significant res-tenosis in the follow-up of drug eluting coronary stent (DES).Methods and Materials: 51 patients (32m, age range 42-75 yrs) who previouslyunderwent coronary percutaneous transluminal angioplasty (PTA) and stent place-ment (112 stents) were evaluated. After 9 months all patients were studied withMDCTA one week prior to selective coronary angiography (SCA), in order to eval-uate stent patency. ECG-gated MDCT of coronary artery was performed beforeand after iv administration of contrast agent with the following acquisition param-eters: 64x0.6 mm slice collimation with a slice thickness 0.6 mm, 120 kV, 700 mAs.Images were analyzed by two radiologists on a dedicated workstation using areal time interaction approach. Stent patency was evaluated qualitatively andquantitatively by measuring the density value proximally, inside and distally to thestent. MDCT results were compared with SCA, considered the gold standard.Results: 3 patients (5 stents) were not evaluated because of heart rate (HR)artefacts due to abnormalities of HR during the administration of iodine contrast.107 stents were patent at SCA. MDCTA was able to detect in-stent re-stenosis in13/16 cases by the first observer and in 15/16 cases by the second observer(overall sensitivity 87%). At MDCT 86/91 stents were judged patent by the firstobserver and 89/91 by the second observer with an overall specificity of 96%.Conclusion: 64 MDCTA is becoming an excellent modality in the follow-up ofcoronary stent patency.

B-168 15:03

Noninvasive visualization of coronary artery bypass grafts using 64-slicecardiac CTG. Pache, T.A. Bley, U. Saueressig, A. Frydrychowicz, D. Foell, N. Ghanem,E. Kotter, M. Langer; Freiburg/DE ([email protected])

Purpose: The aim of this study was to evaluate the diagnostic accuracy in as-sessment of coronary artery bypass grafts using 64-slice computed tomographytechnology.Methods: CT coronary angiography was performed for 96 bypasses in 31 pa-tients with suspected coronary artery disease using a Siemens Sensation 64-slice CT-scanner and compared with invasive coronary angiography. ECG-gatedcardiac CT was performed at a mean heart rate of 63.3 ± 7.3 beats per minute(range 29 to 164). Patients with an irregular or fast heart rate despite ß-blockeradministration were not excluded from the study.Results: All bypass grafts and 94% of the distal bypass anastomoses could bevisualized by CT, non-evaluable distal arterial anastomoses was either due toclip material or calcification artifacts. 42 bypass graft occlusions and 3 significantstenoses were detected by CT and confirmed by invasive coronary angiography.Two venous grafts were missed and one arterial graft was not evaluable withinvasive coronary angiography, but both were clearly depicted by MSCT. Onefalse negative and two false positive CT-findings resulted in a sensitivity of 97.8%,a specificity of 89.3%, a positive predictive value of 90% and a negative predic-tive value of 97.7%.Conclusion: State-of-the-art MSCT coronary angiography allows detailed eval-uation of venous and arterial grafts with high diagnostic image quality even inpatients with higher heart rates. CT reveals complex graft anatomy readily andreliably. Calcification, poor opacification and clip material artefacts can still ham-per the evaluation of distal arterial anastomoses despite improved spatial andtemporal resolution.

B-169 15:12

Evaluation of coronary artery bypass graft patency in early postoperativeperiod: Non-invasive 64-slice multidetector row computed tomographyversus conventional angiographyY. Jeong, J. Kim, C. Park, S. Choi, C. Park, K. Park, D. Yang, H. Kim;Incheon/KR ([email protected])

Background: The purpose of this study was to evaluate the accuracy of 64-sliceMDCT, compared with conventional coronary angiography (CAG), in assessment

of the patency of coronary artery bypass grafts (CABG) in the early (< 1 month)postoperative period.Methods: From March 2005 to August 2005, CABG were performed in 41 pa-tients because of triple vessel disease. Sixteen of them underwent both 64-sliceMDCT and conventional coronary angiography within 15 days after bypass sur-gery to assess the graft patency (9 venous grafts and 27 arterial grafts). One ofthe arterial grafts and two of the venous grafts were excluded because of failurein catheter engagement during conventional CAG. The graft status was evaluat-ed according to the presence of significant stenosis (including obstruction). CTfindings were compared with CAG.Results: In 32 grafts, there was no significant stenosis in both CAG and MDCT.In two grafts, one significant stenosis and one obstruction was concurred in bothmodalities. Only two grafts were interpreted differently. With CAG findings as astandard of reference, the sensitivity, specificity, and positive and negative pre-dictive values were 100%, 97.0%, 75%, 100%, respectively, with an overall diag-nostic accuracy of 97.3%.Conclusion: Many problems such as respiratory motion and fast, irregular cardi-ac rhythm is present in early period after CABG. Although in a small study popu-lation, 64-slice MDCT showed high sensitivity and specificity compared withinvasive conventional angiography. CT coronary angiography is a promising al-ternative diagnostic modality in evaluating the patency of coronary bypass graft.

B-170 15:21 !Value of 16-multidetector CT angiography before total endoscopic coronaryartery bypass surgeryG.M. Feuchtner, T. Schachner, N. Bonaros, G.J. Friedrich, A. Mallouhi, G. Laufer,J. Bonatti, D. zur Nedden; Innsbruck/AT ([email protected])

Purpose: To assess the value of 16-multidetector computed tomography (16-MDCT) coronary and aortic angiography before total endoscopic coronary arterybypass surgery (TECAB) using the da Vinci telemanipulator™ on the arrestedheart with an intraaortic cardiopulmonary bypass (CPB) perfusion device (ES-TECH™ remote-access).Methods: 39 patients were examined with 16-MDCT (Sensation 16™, Sie-mens)(16 x 0.75 mm, rot. 0.42s; TF=6.7 mm/s; 120 kV, 500 mAs; 100 ml iodixa-nol (Visipaque™, Amersham), flow 3-4.5 ml/s; inc 0.6, eff.sl.1 mm; retrospectiveECG-gating). [1] Atherosclerosis of the ascending aorta; [2] myocardial bridgingof the left anterior descending (LAD)coronary artery was evaluated; [3] internalthoracic artery [ITA] graft length and fictive 3-d course to the LAD anastomoticsite was measured with a dedicated software (Vessel View™, Syngo, Siemens).Results: [1] Atherosclerosis at the ascending aorta (11/39) was associated withintraaortic CPB perfusion device related complications (intraaortic balloonmigration[BM] or rupture) (p = 0.007; Chi-square). CT-density of ascending aorticatherosclerotic plaque in patients with BM was 58 HU ± 51 suggesting non-calci-fying plaque. Plaque CT density in patients without BM was 526 HU ± 306 corre-sponding to calcifying plaque (p < 0.001; t-test). [2] Myocardial bridging of theLAD had 5/39 patients; intraoperative difficulties were noted in 4/5. [3] Length ofITA was mean 21.2 cm and fictive, curved 3-d graft course length 17.3 cm (n = 26).Among 13 patients who were not assessed, one required conversion to sternot-omy because ITA graft length was found too short intraoperatively.Conclusion: 16-MDCT angiography is useful prior to TECAB surgery. Ascend-ing aortic atherosclerosis is associated with intraaortic CBP system related com-plications. Measurement of ITA graft length and fictive 3-d course to LADanastomotic site is useful for preoperative planning.

14:00 - 15:30 Room G

Contrast Media

SS 306MRI and USModerators:L. Martí-Bonmatí; Valencia/ESA. Nilsson; Uppsala/SE

B-171 14:00

Contrast-assisted ultrasound for sentinel lymph node detection in normalnodes and in a melanoma modelE.R. Wisner; Davis, CA/US ([email protected])

Purpose: To compare US lymphography efficacy of microbubble formulations ofvarying diameter and composition using a normal dog model and to assess theeffect of tumor replacement on nodal contrast uptake in a swine melanoma model.

Scientific Sessions

174 C D E FBA G

Methods: Regional lymph nodes of 6 normal dogs were imaged using powerDoppler US following subcutaneous injection of one of three hard-shelled bub-bles (POINT Biomedical Corporation, mean particle size = 0.75, 0.91 and 1.33 µm)and one lipid-shell microbubble formulation (DefinityË, perflutren, mean particlesize = 1.8 µm) into the distal front or hindlimb. Regional nodes were imaged 15,30 and 45 minutes post-injection. Sixteen swine with cutaneous melanomas wereimaged in a similar manner following SQ injection of perflutren suspension aroundeach primary lesion. Lymphscintigraphy studies were performed on all dogs andpigs prior to US imaging.Results: Contrast enhancement was evident at a minimum of one time-point in95% (1.8 µm diam), 85% (1.33 µm diam), 75% (0.75 µm diam) and 70% (0.91 µmdiam) of normal canine nodes and occurred within 15 minutes in most instances.Enhancement involved 5-100% of cross-sectional area in all perflutren-enhancednodes. Regional nodes contrast enhanced in all pigs with excellent correlationwith lymphoscintigraphy even in nodes predominantly replaced by tumor.Conclusion: Microbubbles of varying mean diameter and shell composition areefficacious for sentinel node localization with contrast enhancement being moreconsistent when using larger particles. Furthermore, sentinel node detection us-ing this approach compares favorably to current localization techniques even whensignificant tumor replacement has occurred.

B-172 14:09

Comparison of CEUS with CECT in characterization of incidental focal liverlesions in non-cirrhotic patients: Results and economical aspectsM. Passamonti1, F. Berton2, G. Gola2, A. Vercelli2, F. Calliada2; 1Lodi/IT,2Pavia/IT ([email protected])

Purpose: After US detection of an incidental FLL in healthy patients the mostfrequent examination required is a CECT or a simple follow-up. Our aim is tocompare usefulness of CEUS to follow-up and to calculate cost-saving, usingCEUS instead of CECT to achieve a first characterization.Methods and Materials: 125 FLLs detected at US in 104 patients have beenstudied with CEUS using SonoVue (Bracco), and classified as benign not byusing hypoechoic pattern during portal phase. All patients underwent CECT. Re-sults were compared and hypothetical results of follow-up were estimated.Results: CECT were not considered as final test, but just as an accurate test toobtain a correct comparison to CEUS. Results included 15 malignant lesions and110 benign lesions. The diagnosis in CEUS and CECT agreed in all but 6 cases(4 HCCs and 2 adenomas); CEUS misdiagnosed as benign 4 lesions and asmalignant 2 lesions. Performing just CEUS instead of CECT to determinate iffurther examinations were required would have achieved a total saving of 6280 €(60, 4 €/patient). Performing just follow-up would have achieved a total saving of4264 € (41 €/patient), but with at least 15 malignant FLLs misdiagnosed.Conclusion: According to EFSUMB Guidelines, characterization of incidentalFLLs in non-cirrhotic liver by CEUS is possible at a high level of probability. Fur-thermore, the use of CEUS instead of follow-up is clinically safer and instead ofCECT achieves significant cost-saving; CECT remains irreplaceable in all of un-certain cases or when CEUS indicates malignancy.

B-173 14:18 !Image guided-controlled ultrasound delivery of cancer therapeutics usingtargeted nanoparticlesK.L. Watkin, S.L. Gosangari, T.M. Gallagher; Champaign, IL/US([email protected])

Purpose: A new ultrasound method of image guided drug delivery using target-ed nanoparticles will be described.Methods and Materials: The in-vitro efficacy of ultrasound mediated drug re-lease from specially formulated nanoparticles (NPs) containing chemotoxic agentswill be presented. Release of encapsulated drug from the nanoparticles usingultrasound was accomplished using a specially modified Aloka SSD5500 PHDRF ultrasound system using a 5.0 - 10.0 MHz linear ultrasound transducer (AlokaUST-5539). To determine the efficacy of this point and shoot ultrasound mediatedof drug delivery method using our NPs we cultured a set of cancer cells for 1). acontrol condition, 2). a raw drug condition and 3). an ultrasound mediated NPdrug release condition. Amount of cell death was determined using standard stain-ing methods.Results: The results of this experiment revealed that ultrasound mediated NPdrug release produced the same amount of cell death compared with the appli-cation of raw drug - doxorubicin; 80% raw drug; 79% ultrasound triggered NPdrug release. No cell death was observed in the control group.Conclusion: Ultrasound mediated release from our NPs produces cell death in asimilar manner as raw drug in-vitro. Consistent with published literature, there

appears to be an ultrasound advantage after six hours most likely due to theenhancement of the effects of ultrasound and drug uptake by the cells.

B-174 14:27

Early functional evaluation by contrast enhanced ultrasonography usingperfusion software of isolated limb perfusion with melphalan and TNF alphafor in-transit melanoma metastasesL. Chami, A. Cavalcanti, M. Lamuraglia, N. Elie, S. Bonvalot, J. Leclere,A. Roche, N. Lassau; Villejuif/FR ([email protected])

Purpose: The aim of this study was to evaluate if the tumor microvascular chang-es on contrast enhanced ultrasonography (CEUS) after a previous isolated limbperfusion (ILP) could predict the response.Methods and Materials: From march 2004 to august 2005 twenty patients qual-ified for first ILP were included in this prospective study. CEUS examination usingSonoVue (Bracco) was performed for each patient at D-1 before treatment, atD+1, D+7 and D+90. Using a 14 Mhz probe, the tumor size was measured andthe contrast enhancement was evaluated both visually and from raw data usingthe CHI-Q software (Toshiba). The results were compared with clinical responseat 3 months backward.Results: At this time, 72 CEUS were performed on 23 target lesions. One patientwas lost for follow-up, one had no visible lesion, one tumor did not exhibit anycontrast uptake and for 6 patients data at D+90 are not yet available. The medianfollow-up was 6 months (min 1; max 18). Regarding target lesions, the responserate was complete in 28.6% (n = 4), partial in 64.3% (n = 9) and progressionoccurred in 7% (n = 1). At D+1, 56% of target lesions showed an improvement ofcontrast uptake may due to inflammatory reaction. At D+7, a decrease of morethan 10% in tumor contrast uptake was observed in 69% (n = 9) of the respond-ers.Conclusion: CEUS shows early tumor microvascular changes which allow earlyprediction of tumor response after ILP and may help to select eligible patients forrepeated treatment.

B-175 14:36

Radiofrequency ablation of renal cell carcinoma: CEUS vs spiral CT toassess therapy outcomeF.M. Meloni1, T. Livraghi1, F. Bartolucci2, G. Abbiati1, C. Filice3; 1Vimercate/IT,2Lugano/CH, 3Pavia/IT ([email protected])

Aim: To compare the accuracies of contrast-enhanced ultrasound (CEUS) andmultislice CT in the evaluation of local recurrence following radiofrequency (RF)ablation of renal cell carcinoma (RCC).Methods and Materials: 20 patients RCCs ranging between 2.7-6.0 cm in diam-eter (mean 2.6 cm) underwent RF ablation with internally-cooled electrodes. 17had been judged inoperable because of other neoplastic disease or poor gener-alconditions, and the other three had refused surgery. Six of the 20 patients hadmildly elevated serum creatinine levels. Two of the RCCs were located in thecenter of the kidney, 16 were in the periphery, and two had mixed locations. Thetherapeutic response was assessed one month after the procedure and every 6months thereafter. At each visit, the patient underwent multislice CT or MR imag-ing and CEUS with a second-generation microbubble agent (SonoVue; Bracco,Milan, Italy) (2.4 ml IV bolus).Results: All studies were reviewed in a blinded fashion at a consensus confer-ence in which five of the authors participated. Residual viable tumor foci wereidentified in 8/20 lesions by CT and in 6/20 by CEUS.Conclusion: In the follow-up of patients with inoperable RCC treated with RFablation, CEUS is a safe and accurate imaging technique for assessment of ther-apeutic efficacy. The toxicity profile of the US contrast agent and the absence ofradiation exposure make CEUS particularly useful for repeated study of patientswith neoplastic disease and elevated creatinine levels.

B-176 14:45

Fast and efficient stem cell labeling with ResovistD. Golovko1, T. Henning1, J.S. Bauer1, T. Frenzel2, E.J. Rummeny3,H. Daldrup-Link1; 1San Francisco, CA/US, 2Berlin/DE, 3Munich/DE([email protected])

Learning Objective: To develop a clinically applicable, fast and efficient methodfor stem cell labeling with Resovist for subsequent depiction with 1.5 T and 3 TMR scanners.Methods and Materials: Efficient internalization of contrast agents into cells re-quires assisted labeling with transfection agents. Protamine sulfate is a FDA-approved heparin antagonist and a clinically applicable transfection agent. Labeling

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 175C D E FA G

media consisting of DMEM, FCS, Resovist and protamine at different concentra-tions and increasing incubation times of 1, 2, 6, 10 and 24 hours were used tolabel mesenchymal stem cells (MSC). After labeling, samples of 150,000 cells inficoll solution were imaged with T1, T2 and T2*-weighted sequences at 1.5 T and3 T. Respective relaxation times were calculated using self-written software. Ironuptake was confirmed by electron microscopy and quantified by spectrometry(ICP-AES).Results: Labeling medium consisting of DMEM, protamine and Resovist lead toformation of complexes that were too large for cellular uptake. Adding FCS intothe solution prevented any complex formation and allowed for efficient labeling ofMSCs. Simple incubation with Resovist only resulted in efficient labeling afterincubation for 24 hours while assisted labeling with protamine sulfate resulted ineven better results after only one hour. The labeling resulted in significantly short-ened T2 and T2*-relaxation times which were more pronounced at 3 T than at1.5 T. Electron microscopy confirmed intracellular iron oxide uptake.Conclusion: Optimized labeling of MSC with Resovist for MR imaging can beobtained within one hour when protamine sulfate is used as a transfection agent.

B-177 14:54

Safety and effectiveness of 1 molar gadobutrol as contrast agent forcomputed tomography: Results from a comparative porcine studyM. Kalinowski1, K. Goldmann1, M. Gotthardt2, K. Klose1, H. Wagner1;1Marburg/DE, 2Nijmegen/NL ([email protected])

Purpose: To evaluate the safety and efficacy of 1 molar gadobutrol as alternativecontrast agent for computed tomography.Methods and Materials: In nine domestic pigs (52 ± 4 kg body weight) nativeand contrast enhanced multidetector CTs (Volume Zoom, Siemens, Erlangen,Germany) of the chest and abdomen were randomly performed using either 1.2or 3 mL/kg BW gadobutrol (1 mol/L; n = 3 each). Changes in blood chemistry andhematologic parameters were monitored before, 24h, 48h, 72h, and 7 days postapplication. Urine samples were evaluated pre and 7 days post application. Be-fore and 7 days post contrast media application, a 99mTc-MAG3 renal scintigra-phy was performed. The animals were sacrificed and kidneys were removed forexamination by light microscopy one week after contrast media administration.Results: Differences in enhancement of vascular structures and parenchymalorgans was statistically significant between 1 mL/kg BW compared with2 and3 mL/kg BW. No serious adverse event occurred, and no clinically relevant changesin hematologic results, blood chemistry or urine analysis were detected duringthe observation period. Markers for glomerular filtration (creatinine, cystatin C,Beta2-microglobulin, creatinine clearence) and tubular function (Alpha1-microglob-ulin) were unaffected in all groups. No morphologic changes of the renal paren-chyma was found in histology.Conclusion: Contrast enhanced CT with use of gadobutrol resulted in an excel-lent vascular and parenchymal enhancement. Gadobutrol did not affect renal func-tion up to a dose of 3 ml/kg BW. Even in high doses, gadobutrol can beadministered as alternative contrast media in case of contraindications for iodi-nated contrast media, e.g. for CT.

B-178 15:03

Cell labeling with the positive MR contrast agent gadofluorine MO. Saborowski1, T. Henning1, D. Golovko1, J.S. Bauer1, H. Pietsch2,B. Sennino1, H.E. Daldrup-Link1; 1San Francisco, CA/US, 2Berlin/DE([email protected])

Purpose: To label human monocytes with gadofluorine M by simple incubationfor subsequent cell depiction at 1.5 and 3 Tesla.Methods and Materials: Gadofluorine M is a gadolinium-based T1-contrast agentwith a high R1 relaxivity, which forms micelles in aqueous solutions. The micellesare spontaneously phagocytosed by macrophages. Samples of human mono-cytes were incubated with gadofluorine M-Cy3 at varying concentrations, incu-bation times and either investigated directly or after culturing over increasingtime intervals. All cell samples and non-labeled controls were investigated at 1.5and 3 Tesla. R1-relaxation rates of the cells were determined and tested for sig-nificant differences using a t- test. Cellular contrast agent uptake was proven byfluorescence microscopy and quantified by ICP-AES.Results: Efficient cell labeling was achieved after cell incubation with gadofluo-rine M at a concentration of 25 µmol Gd/ml and an incubation time of 24 hrsresulting in a maximal Gd uptake of up to 3.0 fmol Gd/cell without impairment ofcell viability. Fluorescence microscopy confirmed an intracellular uptake of thefluorescent contrast agent inside the cells. The r1 relaxivity of the labeled cellswas 137 mM-1 s-1 at 1.5 T and 80.46 mM-1 s-1 at 3 T. Follow-up imaging studiesshowed a stable labeling for at least 7 days.

Conclusion: In vitro positive labeling of human monocytes for MR imaging iseffectively obtained with gadofluorine M. Potential subsequent in vivo cell track-ing applications comprise e.g. specific targeting of inflammatory processes.

B-179 15:12

Influence of contrast agent on 1H magnetic resonance spectroscopy at 3TeslaK. Zhang, C. Li, Y. Liu, D. Feng, X. Meng, J. Zheng, P. Cong; Jinan,Shandong/CN ([email protected])

Purpose: Using a high field (3 T) MR scanner, to investigate if the result of 1HMagnetic Resonance Spectroscopy (MRS) could be influenced by the paramag-netic contrast agent.Methods and Materials: 1H MRS examinations (PROBE-P, multivoxel, TR/TE:1000 ms/144 ms) were taken before and after administration of Gd-DTPA on 33cases (20 gliomas, 6 meningiomas, 2 metastases, 3 acoustic schwannomas, 1medulloblastoma, 1 NHL) on a GE Signa Excite II3.0 T scanner. Selecting thesame voxel, the peak areas of NAA, Cho, Cr and the ratio of Cho/Cr, Cho/NAAbefore and after administration of Gd-DTPA were compared. Paired-sample t testwas used for statistical analysis. p values less than 0.05 were taken to indicatesignificance.Results: The peak areas of NAA, Cho, Cr before Gd-DTPA administration weremore than those after Gd-DTPA administration in most cases. Only the decreaseof peak area of Cho had statistical significance (t = 2.401, P < 0.05). The ratios ofCho/Cr, Cho/NAA before and after administration of Gd-DTPA showed no statis-tical difference.Conclusion: According to the fact that it is ratios such as Cho/Cr, Cho/NAA butabsolute quantitation that are popular parameters used to analysis proton spec-tra so far, administration of paramagnetic contrast agent will not influence theresult of 1H MRS. To examine an ill-defined lesion, 1H-MRS could be performedafter contrast agent administration to put the ROI precisely without the concernover influence of contrast agent on spectrum analysis.

B-180 15:21

Morphological features of FNH (focal nodular hyperplasia) in Primovist®(Gd-EOB-DTPA)-enhanced liver MR imaging: Results of a multicenter trialC.J. Zech1, L. Grazioli2, S.O. Schoenberg1, M.F. Reiser1, J. Breuer3,E. EOB Study-Group3; 1Munich/DE, 2Brescia/IT, 3Berlin/DE([email protected])

Purpose: To describe morphological features and enhancement pattern of FNHin MR imaging of the liver with the hepatocellular-specific contrast agent Primo-vist®.Methods and Materials: In 182 patients, who were investigated in a phase IIImulticenter trial, 59 FNH were present (13=histopathology; 46=follow-up). MR-examination consisted of T1- and T2w-sequences prior and after bolus-injectionof 0.025 mmol Gd-EOB-DTPA (Primovist®, Schering AG)/kg BW. Morphologicalappearance and enhancement pattern of FNH was evaluated. CNR of FNH tonormal liver parenchyma was calculated. The number of correctly characterizedFNH was calculated and compared to spiral-CT.Results: FNH were described as homogenous, round and well-circumscribedlesions. A central scar was identified in 45% of lesions in T1w pre-contrast, in52% in T2w pre-contrast, and in 60% 20 minutes after injection of Gd-EOB-DTPA.Signal intensity in the hepatocyte-phase 20 min post was described as hyperin-tense in 38%, isointense in 32%, hypointense in 2% and as mixed in 28%. En-hancement in the hepatocyte-phase after 10 and 20 min was seen in 88% and90%, respectively. Median CNR was -5.9 in T1w pre-contrast, 14.0 in the arterialand 2.4 the in portovenous phase. Characterization of lesions was superior forcombined pre- and post MR imaging (88.1%) in comparison to biphasic-enhancedspiral-CT (84.7%, ns) and pre-contrast MR imaging (67.8%, p < 0.05).Conclusion: Gd-EOB-DTPA-enhanced MR imaging provides a better character-ization of FNH in comparison to pre-contrast MR imaging alone and spiral-CT. Inthe dynamic examination enhancement characteristics well known from otherextracellular agents can be seen, in the hepatocyte-phase a liver-specific en-hancement is visible.

Scientific Sessions

176 C D E FBA G

14:00 - 15:30 Room H


SS 309aTumor treatmentModerators:R.-T. Hoffmann; Munich/DEY. Nakajima; Kawasaki Kanaga/JP

B-181 14:00

Role of transarterial chemoembolization before liver transplantation: Aretrospective analysisI. Bargellini, C. Vignali, M. Perri, P. Petruzzi, R. Cioni, L. Urbani, F. Filipponi,C. Bartolozzi; Pisa/IT ([email protected])

Purpose: To retrospectively evaluate the long-term clinical results of transarteri-al chemoembolization (TACE) followed by orthotopic liver transplantation (OLT)in patients with hepatocellular carcinoma (HCC).Methods and Materials: Seventy five patients (M/F:66/9; mean age 55.6 years),with HCC in UICC tumour stage I (n = 10), II (n = 43), III (n = 17) and IV (n = 5),underwent OLT, 11 to 628 days after TACE. On follow-up, the overall and recur-rence free cumulative survival rates were analysed and stratified according tothe initial tumour stage.Results: Mean follow-up after OLT was 36.4 months. In the entire population, the5 year overall and recurrence free survival rates were 70.7% and 67.8%, respec-tively. A MELD ≥ 15 at the time of OLT was the only significant risk factor fordeath (RR = 2.67, P =.0066). At the pathological study on the explanted livers, nocases of tumour progression were observed in the early stage HCC patients(n = 53), in whom the 5 year recurrence free survival rate was 75.3%. Tumourdownstaging, assessed at pathology, was obtained in 11/22 (50%) patients withintermediate-advanced HCC. The 5 year recurrence free survival rates were 80.8%in the downstaged group (comparable with the early HCC patients, P=.81) andonly 18.2% in the non-downstaged group (P=.0004). In fact, among 8 cases(10.5%) of HCC recurrence observed on follow-up, 7 (87.5%) occurred in thenon-downstaged patients.Conclusion: TACE is able to delay tumour progression in HCC patients waitingfor OLT, with high survival rates. Comparable recurrence free survival rates canbe obtained in patients with intermediate-advanced HCC in whom TACE achievestumour downstaging.

B-182 14:09

p53 gene therapy in treatment of advanced hepatocellular carcinomaY. Guan; Chengdu/CN ([email protected])

Learning Objective: To evaluate the efficacy and safety of combination therapywith recombinant adenovirus-p53 injection (rAdp53) and transcatheter hepaticarterial chemoembolization for advanced hepatic carcinoma.Methods: A total of 82 patients with advanced HCC were treated only by tran-scatheter hepatic arterial chemoembolization (control group). Another 68 patientswith HCC were treated by TACE combined with recombinant adenovirus-p53 in-jection (p53 treatment group).Results: The total effective rate of the p53 treatment group was 58.3%, and thatof the control group was 26.5% with significant difference (P < 0.05). The occur-rence of gastrointestinal symptoms was lower than that in the control group(P < 0.05). 3 month and 6 month survival in the p53 treatment group was signifi-cantly higher than in the control group (P < 0.01).Conclusion: The approach for patients with HCC injected intratumorally withrAdp53 was safe and effective. Our results show rAdp53 is a potentially effectivegene therapy agent in HCC treatment.

B-183 14:18

Human growth factor (HGF) evaluation before liver intervention: Long termresults of a prognostic significance for liver regenerationT. Lehnert, A. Gazis, M.G. Mack, J.O. Balzer, W.O. Bechstein, T.J. Vogl;Frankfurt/Main/DE ([email protected])

Purpose: To evaluate and compare the complexity of liver regeneration in corre-lation with the regeneration factor HGF post liver resection and MR-guided laserinduced thermotherapy (LITT).Methods and Materials: Sixty-three patients were evaluated: 31 patients postsurgical liver intervention and 32 patients post ablation (LITT). During a period of18 months the effects of therapy on the cytokine HGF were evaluated and com-

pared to the initial, pre-interventional HGF value. In addition, the effect on livervolume after surgical and interventional radiological procedure was evaluatedusing unenhanced and contrast enhanced MR imaging and CT. Volumetric ex-amination (CT-volumetry) of liver parenchyma pre- and post-intervention wasaccomplished and the regional growth of liver segments was evaluated.Results: A significant influence on the HGF process could be confirmed on lossof liver parenchyma resulting form surgery or LITT induction. Pre-interventionalHGF values < 1500 pg/ml (n = 33) led to a significant increase in liver volume(115-584 ml). Patients with initial HGF-values > 2000 pg/ml (n = 21) showed adecrease in liver volume (5-409 ml) after intervention. Pre-interventional HGF-values between 1501 pg/ml and 1999 pg/ml (n = 9) showed no definite reactionof liver parenchyma. All patients younger than 55 years showed stronger dynam-ics than older patients. The technique used in surgery per se had no significantinfluence on the cytokine process and regeneration dynamics.Conclusion: Pre-interventional human growth factor (HGF) value has a signifi-cant impact on liver regeneration post MR-guided laser induced thermotherapy(LITT) and surgical resection. Pre-interventional low HGF-values correlate withincreased liver regeneration and stronger regeneration dynamics.

B-184 14:27

Experimental study of transarterial administration of GRGDSP combinedwith transarterial chemoembolization in rats with hepatic carcinomaJ. Qian, J. Yin, Y. Wang, G. Feng; Wuhan/CN ([email protected])

Purpose: To evaluate the effects of transarterial administration of GRGDSP (Gly-Arg-Gly-Asp-Ser-Pro) combined with transarterial chemoembolization (TACE)against hepatic carcinoma in rats.Methods and Materials: Subcapsular implantation of a solid Walker-256 tumorin the liver was carried out in 26 male Wistar rats. The following agents wereinjected in the hepatic artery: Group A - GRGDSP + TACE (mitomycin + lipiodol+ ligation of the hepatic proper artery) (n = 9); Group B - TACE alone (n = 8);Group C - 0.9% NaCl (0.2 ml) (n = 9). Tumor volume was assessed by MR imag-ing before (V1) and after (V2) the interventional treatment. The expression offactor VIII in tumor tissue and the tumor microvessel density (MVD) was ana-lyzed by immunohisto-chemical examination.Results: The mean ratio of V2/V1 in Groups A, B and C was 4.42 ± 0.48,6.98 ± 1.09 and 13.00 ± 1.68 respectively. There were significant differences inV2/V1 between different groups (p < 0.05). After interventional therapeutic pro-cedures, the optical density value related to the positive expression of MVD inGroups A, B and C was 0.18 ± 0.02, 0.22 ± 0.02 and 0.23 ± 0.02 respectively.There were significant differences between Group A, and Groups B and C(p < 0.05) but not between Group B and Group C (p> 0.05).Conclusion: Infusion of GRGDSP via the hepatic artery combined with TACEnoticeably inhibits the growth of hepatic carcinoma and its intrahepatic metasta-sis in rats.

B-185 14:36

Evaluation of an animal model for bone metastases of the human breastcancer cell line MDA-MB-231 in nude ratsD. Proschek, J. Balzer, T. Lehnert, K. Eichler, M. Mack, T.J. Vogl; Frankfurt/DE([email protected])

Purpose: To evaluate an animal model for local bone metastases of a humanbreast cancer cell line and to test different application systems for implantingtumor cells into the bone. Primary objective was to compare the different proce-dures.Methods and Materials: Human breast cancer cell line MDA-MB-231 was im-planted into the femur of 6 nude rats. Different concentrations of tumor cells (105

and 106) and different applicators were used. In order to prevent a leakage oftumor cells out of the bone, the drill hole was sealed after application of the tumorcells using bone wax, fibrin sealant and agarose gel (different concentrations:2%/3 %). CT-scans were performed to evaluate number, size and localization ofthe metastases. For statistical analysis, paired comparison procedures were used.Results: Irrespective of the different concentration of tumor cells (105 and 106,p = 0.08), 9 weeks (range 7 -10) after tumor cell implantation, local bone metas-tases were detected in all rats. Leakage of tumor cells and soft tissue metastaseswere seen in half of the rats (n = 3), when agarose gel and fibrin sealant wereused for sealing the drill hole. No leakage occurred when bone wax was used.Conclusion: The implantation of a human breast cancer cell line into the femurof nude rats is a simple and safe procedure. Sealing of the drill hole should becarried out using bone wax to prevent tumor cell leakage. Further interventionalprocedures can be evaluated using this animal model.

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 177C D E FA G

B-186 14:45

Treatment of extensive metastatic liver disease with intrahepatic yttrium 90microspheres: Toxicity assessmentT.F. Jakobs1, R.-T. Hoffmann1, A. Schmitz1, K. Tatsch1, T. Helmberger2,M.F. Reiser1; 1Munich/DE, 2Lübeck/DE ([email protected])

Purpose: Intraarterial injection of yttrium 90 microspheres is used in the treat-ment of otherwise non responding hepatic metastases from various tumors. Thisarticle presents an analysis of the incidence of liver toxicities associated withselective internal radiation therapy (SIRT).Methods and Materials: 34 patients with extensive hepatic metastases qualifiedfor SIRT. Liver toxicities were coded with use of standard oncology criteria (South-west Oncology Group Criteria). Laboratory (total bilirubin, aminotransferases,alkaline phosphatase) and clinical parameters (ascites, encephalopathy, liver fail-ure, hemobilia) were recorded and compared with pre-treatment levels. The SWOGcriteria grades liver toxicity from 0 (no toxicity) to 5 (death) with a clinical signifi-cance of a toxicity grade ≥ 3. Descriptive statistical methods were applied to de-termine the incidence of liver toxicities after SIRT.Results: 22 patients (65%) experienced 33 liver toxicities. At 3 month follow-up17 grade 0 and 16 grade 1 toxicities were observed. No toxicity grade ≥ 3 wasrecorded. At 6 month follow-up 6 grade 0, 15 grade 1, 4 grade 2 and 1 grade 3toxicities were noted. The most frequent toxicity was increased aminotransferaselevels, followed by alkaline phosphatase levels and bilirubin levels. Eight of the33 toxicities resolved on follow-up, 5 became worse. The grade 3 toxicity wasattributed to tumor progression.Conclusion: Whole liver treatment with intraarterial injection of yttrium90 micro-spheres in a single session is a safe palliative treatment option in patients withextensive hepatic metastases. No life threatening liver related toxicity was ob-served.

B-187 14:54

Transarterial chemoembolisation (TACE) for the treatment of hepatocellularcarcinoma (HCC) in a single UK center over a 17-year periodC. Holmes, S. Sookhoo, M. Bassendine, H. Reeves, D. Manas, J. Rose;Newcastle Upon Tyne/UK ([email protected])

Purpose: TACE was introduced as a treatment for HCC on our unit in 1987. Wehave compared and contrasted early versus recent experiences.Methods and Materials: Between 1987 and 1995, TACE was used to treat 80patients with HCC and all these patients (Group 1) have since died. Group 2comprises those cases treated between 2000 and 2004. Of 54 treated individu-als in group 2, case notes of 44 have been studied, 33 of who are dead. Age,mode of diagnosis, Child's Pugh stage, tumour features, treatment interventions,complications and survival were compared.Results: The median age at presentation was 62 years for Group 1 and 72 yearsfor Group 2. Major side effects were seen in 10% (versus 2% in Group 2); medianhospital stay was 26 days in the original group (versus 5 days for Group 2). Themedian number of treatments was 2 in both groups. Twice as many group 2 cases(37%) were identified by routine surveillance. Adjunctive percutaneous ethanolinjections (PEI) were given to 10 patients in Group 1, while 9 in Group 2 hadradiofrequency ablation (RFA). Median survival in patients with good liver func-tion improved from 8 months (Group 1) to 21.5 months (Group 2).Conclusion: Despite an increase in the incidence of HCC there has been animprovement in survival and complication rates, probably as a result of earlierdetection and better case selection. Other factors may include lead-time bias,evolving changes in the TACE procedure and the impact of RFA.

B-188 15:03

Hepatic intra-arterial chemotherapy with gemcitabine: A phase II study inpatients with liver metastases of breast cancerK. Eichler, M.G. Mack, S. Zangos, J.O. Balzer, A. Thalhammer, T.J. Vogl;Frankfurt/DE ([email protected])

Purpose: To evaluate the efficacy and tolerability of hepatic intra-arterial chem-oembolization with gemcitabine as a cytostatic agent in patients with inoperableliver metastases from breast cancer.Methods and Material: Ongoing, open-label, single-center study design; patientshad histologically confirmed breast cancer with inoperable liver metastases, ad-equate bone marrow reserve, sufficient liver/renal function, no active CNS me-tastases, KPS > 70%, and life expectancy > 12 weeks. A suspension ofgemcitabine 1.200 mg/m², 10 ml/m² of iodised oil (Lipiodol), and 5 ml of a degra-dable starch microsphere (Spherex) suspension, is administered intra-arteriallyup to 3 times every 4 wks. Dose-limiting toxicity (DLT) is defined as grade 4 throm-bocytopenia, neutropenia, or nonhematological toxicity > grade 3.

Results: 30 patients were enrolled (median 56 yrs, range 48-65). All patientstolerated the treatment well; with no DLTs. Tumor response was evaluated bymagnetic resonance (MRI) and computed tomography (CT) imaging. Imagingrevealed a decrease in the degree of vascularisation of the treated metastases,which was verified as a reduction in the Gd-DTPA enhanced T1-weighted trans-verse and sagittal MR imaging scans. Imaging follow-up according to the RE-CIST criteria revealed a partial response in nine patients, 15 stable disease andsix progressed, with an increase in size of the lesions or newly developing le-sions.Conclusion: Hepatic intra-arterial chemotherapy with gemcitabine is well toler-ated and achieves encouraging response rates in patients with liver metastasesfrom breast cancer.

B-189 15:12

Successful treatment of hepatocellular carcinoma (HCC) with transcatheterarterial embolization (TAE) with microspheres (embosphere)S. Crespi, A.F. Nicolini, A. Sangiovanni, P. Fasani, L.V. Forzenigo,P.R. Biondetti, M. Colombo; Milano/IT ([email protected])

Purpose: Transarterial embolization (TAE) with permanent embolic agent alone,carried out with superselective catheterisation, enhances tumor ischemia andspares potential toxicity of chemotherapeutic drugs. We assessed feasibility, tol-erability and efficacy of TAE with microspheres in patients with hepatocellularcarcinoma (HCC).Methods and Materials: Child/Pugh A/Barcelona Clinic Liver Cancer (BCLC)stage A or B patients showing a hypervascular HCC at CT with no more than 2tumor lesions, unsuitable for percutaneous therapies (N=25), or proposed forsurgical treatment (N=5), were enrolled. Superselective catheterization of thetumor was attempted in 30 patients: mean age 65 years, 77% HCV positive, 28(93%) BCLC stage A, 2 (7%) stage B. 23 (77%) patients had a single lesion witha mean diameter of 2.7 cm (range 1-6). 4 (13%) had serum alpha-fetoprotein(AFP)> 400 ng/ml. 40-100 or 100-300 micron embospheres were injected.Results: A total of 55 procedures were performed (range 1-3). 24 (80%) achievedcomplete necrosis. A complete necrosis was confirmed in 3 patients who com-pleted treatment and underwent liver transplantation or surgical resection. Nopatient had a deterioration in liver function after TAE. During a mean follow-up of10 months, 6 (20%) patients showed partial necrosis. Another 6 patients had alocal recurrence within 4 months, requiring TAE (N=5) or TACE (N=1), achievingcomplete tumor necrosis in all cases. New tumor lesions were reported in addi-tional 4 patients.Conclusion: TAE is a feasible and well tolerated treatment for HCC in patientwith compensated cirrhosis, causing complete necrosis in 80% of patients.

B-190 15:21

Imaging surviving tumor tissue after devascular treatment using hepaticartery ligationC.C. Riedl, Y.S. Chun, K.J. Hendershott, S. Carlin, L. Schwartz, A. Pugachev,H. Hricak, Y. Fong; New York, NY/US ([email protected])

Purpose: To investigate the potential of imaging hypoxic tumor areas after hepat-ic artery ligation with the hypoxia PET tracer 124I-IAZGP, by delineation of thesehypoxic areas with the hypoxia probe Pimonidazole.Methods and Materials: Hepatic artery ligation was performed on 8 out of 12Buffalo rats bearing solitary nodules of syngeneic Morris hepatoma. Pimonida-zole (an hypoxia probe with similar characteristics as the PET probe 124I-IAZGP)and Hoechst (an in-vivo marker for perfusion) were administered intravenouslyand all animals were sacrificed after 24 h. At the microscopic level, the relation-ship between the spatial distributions of tumor hypoxia (assessed by immunoflu-orescent detection of Pimonidazole), tumor perfusion (assessed by detection ofthe fluorescent marker Hoechst) and tumor tissue survival (assessed by H&Estains) was examined.Results: In animals without ligation of the hepatic artery (sham operation) tu-mors were well perfused and did not reveal areas of significant hypoxia. Hepaticartery ligation led to cessation of perfusion, with tumor necrosis in most areas ofthe tumors. Hypoxic areas with marginal perfusion could be delineated mainly inthe periphery of these tumors. These areas correlated with the surviving areas.Conclusion: Pimonidazole delineated marginally perfused surviving tumor are-as after hepatic artery ligation, suggesting the potential for detecting such areaswith 124I-IAZGP PET scanning.

Scientific Sessions

178 C D E FBA G

14:00 - 15:30 Room I


SS 309bPeripheral angioplastyModerators:B.-S. Tan; Singapore/SGI.P. Vulev; Bratislava/SK

B-191 14:00

Paclitaxel-loaded nanoparticles could reduce neointimal hyperplasia inrestenotic rabbit iliac arteriesM. Kalinowski, S. Schäller, J. Fuchs, B. Kleb, K. Klose, H. Wagner; Marburg/DE([email protected])

Purpose: To evaluate the efficacy of paclitaxel loaded nanoparticles adminis-tered by local administration over the channelled balloon catheter at preventingrestenosis after balloon angioplasty in restenotic iliac arteries of New Zealandwhite rabbits.Methods and Materials: Eight male New Zealand White rabbits were fed a 1%cholesterol diet and received denudation of both common iliac arteries with a3 mm balloon catheter to induce an arterial denudation. Four weeks after steno-sis induction, local application of paclitaxel loaded nanoparticles was performedrandomly on one common iliac artery. As a control saline was injected into thecontralateral side under the same conditions. The channelled balloon catheterallows simultaneous balloon dilation (8 atm) of the stenosis and local applicationof a drug solution (2 atm). Four weeks after local drug delivery the animals werekilled, the vessels were excised and computerized morphometric measurementswere performed.Results: In none of the animals acute complications e.g. dissection, vessel throm-bosis or perforation were seen. Application of paclitaxel loaded nanoparticlesresulted in a statistically significant reduction of neointimal area compared to thecontrol group of 50%. The neointimal area was 0.80 ± 0.19 mm2 using nanoparti-cles compared to 1.58 ± 0.6 mm2 in the control group.Conclusion: Local drug delivery of paclitaxel loaded nanoparticles using thechannelled balloon catheter could reduce neointimal hyperplasia in restenoticrabbit iliac arteries under the predominant conditions.

B-192 14:09

How do patients perceive the benefits and risks of peripheral angioplasty?Implications for informed consentS.B. Habib1, A.M. Groves2, T.C. See3, L. Sonoda3, P.J. Ell2; 1Nottingham/UK,2London/UK, 3Cambridge/UK ([email protected])

Purpose: In order for patients to give their informed consent they should fullyunderstand the benefits and risks of the procedure involved. A survey was per-formed to explore patients' perceptions of benefits and risks of peripheral angi-oplasty.Methods and Materials: Over 8 months, 206 patients undergoing peripheralangioplasty were surveyed. 30 patients formed a pilot study leaving 176 patients(123 male, 53 female) average age 69.6. The patients were randomly assignedinto 2 groups. One group was given a recognised "Risk-Assessment-Chart", theother group answered without this tool. Patients were questioned regarding therisks and benefits of the procedure immediately pre and post-angioplasty. Pa-tients were asked to express their likelihood of benefit and complication in termsof a percentage (definite benefit being 100%). Whether the patient had receivedinformation or had undergone angioplasty prior was noted. Statistical differenceswere calculated using students-t-test.Results: 158 (90%) patients thought that they would have at least a 75% (3 in 4chance) of benefiting from angioplasty. 39% thought they would definitely (100%)benefit. 85 (48%) patients thought the test would be painful, 59 (34%) patientsthought there would be no radiation exposure, 28 (16%) patients realised therewas a chance of uncontrolled bleeding and 143 (81%) patients thought the testwas easier than they had originally thought. The use of the "risk-assessment-tool" significantly altered patient's perceptions (t-test-p = 0.040).Conclusion: Most patients undergoing angioplasty were optimistic about thebenefits, but were unaware of the risks. The risk assessment tool would appearto be useful for the consent process.

B-193 14:18

An audit of pain control in interventional radiologyG.G.J. McKenzie, A. Downie; Glasgow/UK ([email protected])

Purpose: In the United Kingdom the majority of interventional cases are per-formed without an anaesthetist so making the interventional radiologist responsi-ble for pain control and sedation as well as performing the procedure. The aim ofthis audit was to assess the adequacy of pain control in interventional radiology.Method: Over a 6 month period, pain control was assessed using a Visual Ana-logue Scale. Both the pre-procedure and the highest intra-operative pain scoreswere recorded for a wide variety of procedures. The type of anxiolytic/analgesicadministered were also recorded. Statistical analysis was performed with theMinitab computer package.Results: Eighty patients took part in the audit. Analysis shows that 46 (57.5%) ofthese patients had insufficient intra-operative pain control. Particularly, there wasa significant increase in pain levels (p < 0.05) experienced intra-operatively, com-pared to the pre-procedure levels in angiography, angioplasty, Hickman line in-sertion, gastrostomies, biliary stenting and varicocele embolization. Furthermore,pain control was shown to be insufficient (p < 0.05) in Hickman line insertions,gastrostomies and biliary stenting. An anaesthetist was not involved in any of thecases and pre-medication was never given.Conclusion: Pain control is poorly managed in the majority of interventional cas-es, particularly Hickman line insertions, gastrostomies and biliary stenting. Thereis therefore considerable scope to improve patient analgesia perhaps by improv-ing the type and dose of drug administration including the use of pre-medicationand ideally if resources allow the presence of anaesthetic colleagues.

B-194 14:27

Effects of percutaneous transluminal angioplasty (PTA) and endovascularbrachytherapy (EVBT) on vascular remodeling of human femoropoplitealartery: Two year follow-up by noninvasive magnetic resonance (MR) plaqueimagingR. Wyttenbach1, R. Corti2, M. Di Valentino1, M. Alerci1, P. Santini1, L. Cozzi1,V. Fuster3, A. Gallino1; 1Bellinzona/CH, 2Zürich/CH, 3New York, NY/US([email protected])

Background: PTA of severely stenotic arteries is hampered by high restenosisrate. The effects of PTA on vascular wall as well as the anti-restenotic propertiesof endovascular brachytherapy (EVBT) remain unclear. We aimed to elucidatethe vascular effect of PTA and PTA+EVBT in vivo by serial MR imaging.Methods: Twenty symptomatic patients with severe stenosis of the femoropop-liteal artery were randomly assigned to PTA (n = 10) or PTA+EVBT (n = 10, 14Gy by γ-source) and imaged by high-resolution MR imaging before and 24 hours,3 months and 2 years after intervention. An independent observer blinded to theprocedure analyzed the MR imaging data.Results: At 24 hours, MR imaging revealed that lumen area (+86% and +67%)and total vessel area (+47% and +34%) similarly increased in the PTA andPTA+EVBT group. All patients showed severe splitting of the atheroscleroticplaque. At 3 months, MR imaging revealed a significant difference in lumen areachange between the PTA and PTA+EVBT group (+40% and +106%, respective-ly; p = 0.026) and in the total vessel area (+14% and +39%, respectively;p = 0.018). At 24 months, signs of restenosis and inward remodeling were alsopresent in patients treated with PTA+EVBT. PTA induced plaque disruption per-sisted at 3 months in 50% and at 2 years in 10% in PTA+EVBT patients only.Conclusion: PTA induces deep disruption of the atherosclerotic plaques andextensive remodeling process of the arterial wall. Luminal loss after PTA is par-tially due to inward vessel remodeling. Brachytherapy initially prevents resteno-sis, induces outward remodeling of the vessel and delays the process of restenosis.

B-195 14:36

SFA subintimal angioplasty: Experience in 77 patientsR. Gandini, L. Maresca, V. Pipitone, M. Stefanini, M. Chiocchi, A. Spinelli,G. Simonetti; Rome/IT ([email protected])

Purpose: To evaluate indications, feasibility and results of subintimal angioplastyof long femoral-popliteal occlusions in patients with severe claudication and/orcritical limb ischaemia.Methods and Materials: We treated 77 patients (52 males, 25 females, meanage 71.3 yrs) with long femoral-popliteal occlusion (> 10 cm). Forty-two patientswere symptomatic for claudication, 6 patients had rest pain and 29 patients pre-sented with gangrene. We studied all patients with Angio-MR and Doppler-USbefore treatment. Recanalization of the superficial femoral artery was performedin all cases. Contemporary recanalization of limb arteries was carried out in 17

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 179C D E FA G

patients. In 72 cases recanalization was obtained by antegrade femoral approach,in 2 cases by retrograde trans-popliteal approach and in 3 cases by combinedfemoral-popliteal approach. Follow-up was performed by physical examinationand Doppler-US 1, 3 and 6 months after the procedure and subsequently afterevery 6 months.Results: Technical success was 96.5% with resolution of rest pain or claudica-tion in 39/48 patients and complete resolution of gangrene in 25/29 patients. Threepatients underwent only a minor amputation. A mean follow-up of 10.3 months(range 3-18) demonstrated patency in 69/77 patients.Conclusion: Subintimal angioplasty of long femoral-popliteal occlusions is a safe,effective technique and a valid alternative to surgery.

B-196 14:45 !Infrapopliteal PTA or PTA with bail-out stenting in chronic critical limbischemia: One year clinical and angiographic resultsD. Siablis, P. Kraniotis, D. Karnabatidis, K. Katsanos, A. Diamantopoulos,N. Christeas, J. Tsolakis; Patras/GR ([email protected])

Purpose: To report the 1-year clinical and angiographic outcomes in patientswith chronic critical limb ischemia that underwent PTA alone or PTA with bail-outstenting during infrapopliteal interventions.Methods and Materials: Sixty-three consecutive patients with de novo athero-sclerotic infrapopliteal disease and chronic critical limb ischemia (Rutherford cat-egory 4 and above) were treated either with PTA alone (group A, 25 patients, 16males, mean age 66 ± 10 years) or with PTA and bail-out stenting (group B, 28patients, 18 males, mean age 68 ± 10 years). Baseline demographic variableswere comparable between the two groups.Results: A significantly higher number of initially occluded vessels necessitatedbail-out stenting after suboptimal angioplasty (23.8% in group A vs 44.3% in groupB, p = 0.0167). One year binary restenosis rate was 65.8% in group A and 60.6%in group B. Primary and secondary patency were 42.1% and 60.5% in group A.The respective patency rates were 48.5% and 63.6% in group B. The mortalityrate was 8.0% and 14.3% in groups A and B, correspondingly. The cumulativelimb salvage rate was 100% for both groups. A total of 88.0% of the patientsescaped rest pain in group A and 96.4% in group B. Comparing the aforemen-tioned clinical or angiographic results between groups A and B, none of themreached the level of statistical significance.Conclusion: Infrapopliteal PTA alone and PTA plus bail-out stenting portray equiv-alent 1-year clinical and angiographic results. Stenting should be reserved as abail-out procedure when PTA alone is not enough to restore vessel patency, es-pecially in occlusive lesions.

B-197 14:54 !Day-case diagnostic and interventional angiography: Experience with 331patients in a self-contained specialist radiological nurse-led andadministered unitD. Huang, H.L. Walters, C.J. Wilkins, D. Evans, R.D. Edmondson, H.I. Rashid,C.M. Ong, P.S. Sidhu; London/UK ([email protected])

Purpose: To prospectively evaluate feasibility and safety of day-case interven-tional angiography in a self-contained radiological nurse-administered unit.Methods and Materials: Data was recorded prospectively over an eight-yearperiod (1996 to 2004). Radiology Specialist Nurses (RGN), in a pre-angiographyevaluation clinic, assessed patient suitability for day-case angiography based ona protocol. Immediately prior to the procedure patients were reassessed and aninterventional radiologist performed all procedures. The RGNs were responsiblefor patient care before, during and after angiography and for follow-up post-pro-cedure. The data were summarised into either diagnostic or interventional (bal-loon dilatation/stent insertion/embolization) cases. Complications (minor or major)were divided into immediate or delayed (24 hour), either requiring hospital ad-mission or 'day-case' management.Results: Of 331 procedures performed,136 were diagnostic and 145 interven-tional. A total of 59/331 (17%) immediate post-procedure complications and 23/331 (7%) delayed complications were identified, of which local pain or haemato-ma and bleeding were the most common. In the diagnostic studies, 17 immediateand 8 delayed complications occurred, whereas in the interventional studies, 42immediate and 15 delayed complications were noted. Delayed complications wereall minor. Of the patients with immediate complications, 23/331 (7%) requiredhospital admission for a variety of reasons (5 in the diagnostic group and 18 inthe interventional group). No subsequent complications resulting in permanentdeficit were observed in either group.Conclusion: The incidence of complications for day case angiography in a nurse-led unit is comparable to physician-led units. A nurse-led day-case interventional

unit is a safe alternative and represents a considerable time saving for the inter-ventional radiologist.

B-198 15:03

"Cutting" balloon angioplasty (CBA) versus conventional balloonangioplasty (PTA) in short femoro-popliteal arterial stenosisR. Iezzi, A.R. Cotroneo, F. Quinto, R. Esposito, D. Pascali, M.L. Storto; Chieti/IT([email protected])

Purpose: To compare middle-term results of "cutting" balloon angioplasty (CBA)with PTA for the treatment of short femoro-popliteal arterial stenosis (TASC A).Methods and Materials: Between February 2004 and June 2005, 55 consecu-tive patients with 61 focal (< 3 cm) calcified and severe (> 70%) stenoses of fem-oro-popliteal arteries underwent endovascular treatment: 27 patients (29 lesions)were treated with PTA (Group A) whereas the last 28 patients (32 lesions) under-went "cutting" balloon angioplasty (Cutting-Balloon Ultra; IVT/Boston Scientific,USA; 3.5-5-mm diameter/10-15-mm length) (Group B). Baseline patient demo-graphic data, pre- and post-procedural patient clinical data and procedural re-sults of Group A and B were recorded. Follow-up consisted of clinical assessmentand color duplex ultrasonography (CDU) 1, 3 and every three months after theprocedure.Results: All treatments were successfully performed via an anterograde approachwith a technical success rate of 100%, without any major complications. In twotreated lesions of Group A (2/29, 6.9%) a stent was implanted because of a dis-section, whereas no patient of Group B required placement of a stent because ofrecoil, dissection, or arterial tears. In Group A (PTA) primary and secondary pa-tency rates were 85.2% and 93% at 6-months, 73.8% and 77.9% at 12-months,respectively. In Group B (CBA) primary and secondary patency rates were 90%and 96.8% at 6-months, 80.5% and 87.7% at 12-months, respectively.Conclusion: CBA seems to be a valuable tool in the endovascular treatment ofshort femoro-popliteal stenotic lesions allowing an increased patency and bettermiddle-term results as compared to conventional PTA.

B-199 15:12

The patient's perspective of percutaneous transluminal angioplasty: Howdoes it impact on quality of life?A.N. Keeling, P.A. Naughton, A.M. O'Connell, M.J. Lee; Dublin/IE([email protected])

Purpose: To prospectively assess the impact of percutaneous transluminal angi-oplasty (PTA) on quality of life (QoL) in patients with peripheral vascular disease(PVD).Methods and Materials: 71 patients (54% male, average age 71 yrs, 73% smok-ers, 28% diabetic) underwent lower limb PTA (66% intermittent claudicants, 34%limb salvage) over twelve months. QoL data were obtained in all patients beforeangioplasty, and at one and six months using the Short Form 36 (SF 36) ques-tionnaire.Results: SF 36 questionnaire was completed in 95% of cases pre-PTA, 89% atone month and 68% at six months to date. Baseline QoL was below that quotedfor the normal population. QoL average scores at baseline, one month and sixmonths were as follows: physical function 41, 50, 52; physical role 23, 38, 41;emotional role 67, 56, 61; social functioning 72, 73, 74; mental health 73, 76, 79;energy/vitality 49, 55, 55; pain 48, 65, 61; general health perception 61, 60, 58;change in health 38, 65, 66. There was an improvement in seven out of ninehealth domains at one and six months, being statistically significant in four: phys-ical function (p < 0.001); physical role (p < 0.005); energy/vitality (p < 0.01); pain(p < 0.0001).Conclusion: Percutaneous transluminal angioplasty results in improved qualityof life in patients as early as one month post-procedure, and is maintained at sixmonths. Quality of life assessment is a useful adjunct to clinical follow-up in pe-ripheral vascular disease.

B-200 15:21

Infrainguinal subintimal angioplasty in the management of diabetic footulcerI. Bargellini, P. Petruzzi, A. Auci, R. Cioni, A. Cicorelli, C. Vignali, A. Piaggesi,C. Bartolozzi; Pisa/IT ([email protected])

Purpose: To evaluate the safety and clinical outcome of infrainguinal subintimalangioplasty in a selected series of diabetic patients with lower limb chronic criti-cal ischemia and foot ulcer.Methods and Materials: From October 2003 to August 2005, 34 diabetic pa-tients (DM1/DM2=2/32, mean diabetes duration 19.5 ± 11.4 years, mean HbA1C

Scientific Sessions

180 C D E FBA G

8.6 ± 2%), (M/F=22/12, mean age 69.8 ± 8.4 years), from the Diabetic Foot Clin-ic, with Fontaine stage IV lower limb ischemia, underwent subintimal angioplastyto treat chronic long occlusions (length ≥ 5 cm) of the femoro-popliteal (n = 24)and/or infrapopliteal (n = 11) regions. All patients were unfit for surgical recanal-ization. Strict clinical and color-coded Duplex US follow-up was performed afterthe procedure.Results: Immediate technical success was achieved in 30/34 patients (88.2%).Three cases of minor immediate complications occurred at the site of the femoralpuncture (hematoma in 2 cases, pseudoaneurysm in 1) and were treated con-servatively. On follow-up (mean 6.5 ± 1.1 months, range 0-25 months), 8/30(26.7%) cases of restenosis were observed; among these, three patients under-went repeated angioplasty, with a secondary patency rate of 83.3% (25/30 cas-es). The cumulative primary patency rate was 74.9% and 61.6% at 3 and 6 monthsfollow-up, respectively. However, the clinical success rate (improvement/healingof the ulcer) was 93.3% (28/30 patients).Conclusion: Subintimal angioplasty represents a safe and effective treatmentoption in non-surgical diabetic patients with limb-threatening chronic ischemia.

14:00 - 15:30 Room K

Genitourinary

SS 307Upper urinary tractModerators:M.-F. Bellin; Villejuif/FRR. García Figueiras; Santiago de Compostela/ES

B-201 14:00

CT-urography (CTU) in the evaluation of upper urinary tract disease:Retrospective comparison of standard-dose and low-dose CTU protocolsC. Degenhart, T. Meindl, E. Coppenrath, U.L. Müller- Lisse, M.F. Reiser,U.G. Müller- Lisse; Munich/DE ([email protected])

Purpose: Comparison of routine and low-dose CTU protocols regarding diagno-sis of upper urinary tract (UUT) pathologies.Methods and Material: During a 1.5-year period, 92 patients with diseases sus-pected to affect the UUT underwent contrast-enhanced CT of the abdomen, in-cluding a portovenous/nephrographic and a delayed excretory phase. CTU wasobtained with 4x2.5 mm collimation after administration of i.v. saline. In group 1,CTU was performed with a standard dose protocol (n = 42, 120 kV, average175.6 mAs/slice) and in group 2 with a low dose examination protocol (n = 50,120 kV, 29 mAs/slice). Pathologies of the UUTs were corroborated by a combi-nation of portovenous/nephrographic scans and clinical information. CTU scanswere reviewed by two independent readers (R1, R2) blinded to corroboratingdata. UUT pathology (dilatation, compression, intraluminal lesions) was diagnosedusing a 5-point confidence scale: pathology definitely/probably absent (1.2), un-decided (3), probably/definitely present (4.5). Statistics were performed usingROC-analysis and Chi-Square testing.Results: Sensitivity and specifity for UUT dilatation was equal in routine and low-dose CTU (group 1, sensitivity 92.9%, specificity 99.7%; group 2, sensitivity 96.2%,specificity 99.4%; sensitivity, chi-square = 0.01, p = 0.92; specificity, chi-square = 0.062, p = 0.8). For UUT compression, no significant differences be-tween the groups were found (group 1, sensitivity 96.4%, specificity 98.8%; group2 sensitivity 90.9%, specificity 99.0%; sensitivity, chi-square = 0.013, p = 0.91;specificity, chi-square = 0.23, p = 0.6). Number of intraluminal lesions was low:In group 1 5/5, in group 2 1/1 (R1) and 0/1 (R2) lesions were detected.Conclusion: Dose-reduced CTU is adequate to determine UUT dilatation andcompression. However, it is unlikely to depict intraluminal lesions reliably.

B-202 14:09

Low-dose CT-urography: Image quality analysis during gradual tube currentreduction. Animal experienceJ. Kemper, M. Regier, M. Bansmann, P.G.C. Begemann, H.D. Nagel, A. Stork,G. Adam, C.C.A. Nolte-Ernsting; Hamburg/DE ([email protected])

Purpose: To intra-individually evaluate the possibility of reducing X-ray exposureduring multi slice CT urography (MSCTU) with consideration of image qualityusing a porcine model.Methods and Material: MSCTU was performed in 8 healthy pigs under generalanesthesia (weight ± 70 kg). Examinations were performed with a gradual reduc-tion of the tube current time product at 120 kV (200,125,100-20 eff.mAs). Three

blinded observers independently evaluated the image data for anatomical detail,subjective image quality and image noise using a five point scale. Overall imagequality was compared to radiation dose. Diagnostic reliability was tested usingthe intraclass-correlation-coefficient (ICC). Stratified comparisons of mean scoreswere calculated using Friedman and Wilcoxon-tests.Results: Adequate image quality for detailed visibility of the upper urinary tractwas obtained by dropping the tube current time product to 70 eff.mAs at 120 kV.Image noise did not impair image quality to a relevant degree using these param-eters. Mean score for image quality did not significantly advance with the use ofhigher currents (p > 0.05). A further reduction of the current time products below70 mAs leads to a significant impairment of image quality (p < 0.05). There was ahigh agreement between the 3 observers (ICC = 0.95).Conclusion: In the experimental setting, radiation dose in MSCTU could be sig-nificantly reduced. To maintain a sufficient image quality, a current time productreduction to 70 mAs at 120 kV proved to be feasible offering an advantageousdosage reduction. Systematic clinical studies are required for the assessment ofthe diagnostic impact of such protocols.

B-203 14:18

Contrast-to-noise ratio as an indicator of image quality in CT urographyE.M. Coppenrath, T. Meindl, U.L. Müller-Lisse, M.F. Reiser, U.G. Müller-Lisse;Munich/DE ([email protected])

Purpose: CT urography is based on contrast-enhanced delineation of the renalpelvis and ureter from adjacent tissues. As a quantitative parameter of imagequality, contrast-to-noise ratio (CNR) was determined in standard dose and lowdose CT urography protocols.Methods and Material: A standard dose (SD) patient group (n = 5, four-slicescanner, slice thickness 3.2 mm, 120 kV, 150-200 mAs, CTDI 10.2-14.1 mGy)was compared with a low dose (LD) group (n = 12, 120 kV, 15-29 mAs, CTDI 1.0-2.0 mGy). CNR was calculated from CT density in the renal pelvis and in proxi-mal, middle, and distal parts of the ureter divided by noise in adjacent fat tissueand muscle tissue (iliopsoas). Excretory phase was examined 15 minutes(+ 10 minutes) after intravenous contrast injection.Results: CNR (with noise in fat tissue) in SD vs. LD was 92.1 vs. 28.1 (renalpelvis), 65.7 vs. 25.0 (proximal ureter), 67.0 vs. 21.8 (middle ureter), 35.4 vs. 18.4(distal ureter). CNR (with noise in iliopsoas) in SD vs. LD was 67.8 vs. 18.8 (renalpelvis), 53.5 vs. 18.8 (proximal ureter), 48.4 vs. 18.9 (middle ureter), 41.7 vs. 17.4(distal ureter).Conclusion: CNR was higher in SD than in LD group by a factor of 2.3 - 3.2.However, even a CNR greater than 5 usually suffices for detection of the ureter,as previously demonstrated. Thus, CNR may be helpful in judging noise relatedimage quality in CT urography. Reasonable dose reduction in CT urography, to aCTDI of 1.0-2.0 mGy, seems possible with an acceptable loss of image quality.

B-204 14:27

Modification of multislice CT urography (MSCTU) technique: Analysis ofindividual acquisition delay and opacification using furosemide and low-dose test imagesJ. Kemper, M. Regier, A. Stork, M. Bansmann, P.G.C. Begemann, G. Adam,C.C.A. Nolte-Ernsting; Hamburg/DE ([email protected])

Purpose: To retrospectively analyze scan delay and opacification of the collect-ing system during MSCTU after furosemide and acquisition of low dose test im-ages.Methods and Materials: Excretory phase images obtained from MSCTU in 103patients (69 men, 34 women) were reviewed. MSCTUs were performed by usinglow dose furosemide. There was no fixed delay for urographic image acquisition.Urographic timing was individually adjusted by performing single slice test imag-es of the distal ureters to display their current opacification. Image analysis in-cluded grading of opacification and image quality. The average urographic delaywas calculated. Stratified comparisons of mean scores were assessed.Results: The median calculated scan delay for patients with normal creatine lev-els (n = 92) was 420 sec (mean 453 sec; SD, 121 sec). The median number ofacquired test images were 2 (range 1-6 images). The analysis of opacificationdemonstrated that 97% of the ICS, 89% of the proximal, 86% of the middle and81% of the distal ureter segments showed opacification greater than 90%. 7.8%of the distal ureter segments could not be visualized. There was no statisticallysignificant difference of opacification between proximal, middle, and distal uretersegments (p > 0.05). Overall image quality of MSCTU was shown to be high whenthe last test image demonstrated homogeneous contrast opacification in bothureters (r = 0.81). Interobserver κ were 0.78-0.8.Conclusion: Individual scan delay timing for MSCTU by means of test imageacquisition was demonstrated to be a reliable protocol modification to achieve

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 181C D E FA G

improved opacification of the upper urinary tract. It helps adapt MSCTU to indi-vidual renal excretory rates.

B-205 14:36

Initial experience and perspectives of a low-dose two-phase, "triple-bolussingle scan" multidetector CT urography (MDCTU) protocolM. Kekelidze, M.L. Dijkshoorn, S. Dwarkasing, S.M. Hussain, G.P. Krestin;Rotterdam/NL ([email protected])

Purpose: To evaluate the image quality of two-phase MDCTU protocol for evalu-ation of the urinary system.Methods and Materials: 110 patients with haematuria, flank pain and or sus-pected urinary tract abnormalities were examined on a 16-slice CT-scanner (So-matom Sensation 16, Siemens). MDCTU protocol consisted of two phases:Unenhanced and triple-bolus iv contrast injection. 1) patients were given 800 mlof water 20 min before the study and unenhanced (low-dose) scan of the fullabdomen was acquired; 2) 30 ml 2 ml/s of IV contrast (VisipaqueTM 320 mg I/ml,Amersham) was injected as a first bolus, 7 min. later a second bolus of 50 ml1.5 ml/s and after 20 seconds the third bolus of 65 ml 3 ml/s injected and the fullabdomen scanned. Coronal 5 mm MPR reconstructions were generated as stand-ard for both scans and in selected cases oblique MPR or VR.Results: Examination revealed urolithiasis (n = 24), hydronephrosis (n = 9), re-nal mass (n = 36), TCC (n = 2), accessory renal arteries (n = 7). High qualityparenchymal enhancement was achieved in 94 (85.4%) patients. Streak arte-facts were seen only in 4 (3.6%) cases caused by dense opacification of thecalyces. Opacification of the collecting system and ureters were scored on a 5-grade scale (1=not filled, 3=half filled, 5=all filled). Median score for upper urinarytract filling was 4 and 3 for lower segments of the ureters. In 2 cases MR followedMDCTU, but was not problem solving.Conclusion: Two-phase MDCTU with triple-bolus single scan IV contrast injec-tion is a simple protocol acquiring native, arterial, nephrographic and excretoryphases in two series. It provides satisfactory imaging scores, reduces radiationdose and radiological workload.

B-206 14:45

Multiphasic multidetector row CT (MDCT) in detection and staging oftransitional cell carcinomas of the upper urinary tractG.A. Fritz, H. Schoellnast, H. Deutschmann, F. Quehenberger, F. Thimary,M. Tillich; Graz/AT ([email protected])

Purpose: To evaluate the potential of multiphasic multidetector row CT (MDCT)in detection and staging of transitional cell carcinomas (TCC) of the upper uri-nary tract.Methods and Materials: We retrospectively reviewed the CT examinations of 39consecutive patients (18 women, 21 men) with histologically verified TCC of theupper urinary tract, performed between February 1998 and April 2005. Patientages ranged from 48-84 years (mean age: 68 ± 10 years). All patients were con-sidered candidates for open surgical management (partial ureterectomy: 14 pa-tients, nephrectomy: 14 patients, nephroureterectomy: 11 patients). The urinarytract was examined using MDCT performing unenhanced and contrast enhancedscans during the corticomedullary, nephrographic and pyelographic phase. Tu-mors were staged according to the TNM-classification. MDCT and histopatholog-ical findings were correlated.Results: In MDCT all 39 TCC including 2 multicentric TCC were detected. 10/39patients with TCC (25.5%, 4 renal, 6 ureteral) had stage 0 tumors, 9 patients hadstage I tumors (23.1%, 2 renal, 7 ureteral), 8 patients had stage II tumors (20.5%,4 renal, 4 ureteral); 11 patients had stage III tumors (28.2%, 6 renal, 5 ureteral)and 1 patient (2.7%) had a stage IV tumor of the renal pelvis. TCC confined to theorgan (stage 0-II) was correctly staged in 26/27 patients (96.3%). Stage III-IVtumors were correctly staged in 8/12 patients (66.6%). Overall, MDCT was accu-rate in predicting pathological TNM stage in 34/39 patients (87.1%).Conclusion: MDCT with its high spatial and temporal resolution is an accuratetool for detection and staging of TCC of the upper urinary tract.

B-207 14:54

Intraluminal optical coherence tomography: Demonstration of tissue layersin ureteral specimens ex vivo by new means of microstructural imagingU.L. Müller-Lisse, O.A. Meissner, G. Babaryka, M. Bauer, M.F. Reiser,U.G. Müller-Lisse; Munich/DE ([email protected])

Purpose: Based on near-infrared light, which is emitted from a catheter-mount-ed optical fiber, intraluminal optical coherence tomography (OCT) provides cross-sectional images of vessels and hollow organs, with a lateral resolution of

10-15 µm. We compared demonstration of different ureteral wall layers in OCTimages with histologic sections of porcine ureters ex vivo.Methods and Material: The OCT probe (diameter, 0.014 inch, LightLab Inc.,Westford, MA) was introduced into each of 6 porcine ureter specimens ex vivo.Ureters were dilated with normal saline solution. Single-slice cross-sectional OCTimages and H&E-stained histologic sections were obtained at marked positions.Two independent observers (O1, O2) each determined delineation of ureteralwall layers (urothelium, lamina propria, muscle layer) by means of OCT softwareand light microscopy in four different quadrants of each specimen, respectively.Results: OCT distinguished ureteral wall layers in all specimens. Urothelium andlamina propria were delineated in 23 (O1) and 24 (O2) quadrants, lamina propriaand muscle layer in 19 (O1) and 16 (O2) quadrants, respectively. While O1 delin-eated different muscle layers in 13 quadrants, O2 made this distinction in only 2.Conclusion: Intraluminal OCT demonstrates different ureteral wall layers similarto histologic sections. Results suggest that OCT fulfils essential requirements forin-vivo application.

B-208 15:03

Drug induced MR-pyelography in the evaluation of renal collecting systemmalformations, tumors, renal calculi and obstructed renal uretersM. Di Girolamo, M. Pesce, L. Argnani, A. Fregolino, S. Quartieri, V. David;Rome/IT ([email protected])

Purpose: MR-pyelography is a technique to evaluate the renal collecting sys-tems, performed after drug induced distention of the urinary tract.Method and Materials: 10 normal volunteers and 271 patients with renal diseas-es underwent drug induced MR-pyelography using 0.5 and 1.5 T superconduc-tive magnets. The examination was performed with a 3D non-breath holdingfat-suppressed TSE-sequence (TR:3000 ms; TE:700 ms; N.Ex.:4; ETL:128; ac-quisition time:3 min) in a coronal plane. These acquisitions were post-processedwith a MIP algorithm. To obtain maximum filling of the collecting system, diuresiswas pharmacologically induced by administering 250 ml of saline solution i.v.together with 20 mg of furosemide. One MR acquisition was performed 10 min-utes after diuresis induction. 5 normal volunteers and 230 patients had under-gone an IVU and ascending pyelography was performed in 15 cases.Results: We always obtained excellent anatomical evaluation of the renal col-lecting systems with drug induced MR-pyelography that was considered compa-rable to that obtained with an IVU. It was always possible to study the renalcollecting system malformations. Renal calculi larger than 2 mm were alwaysidentified by analysis of the 3D coronal scans. Drug-induced MR-pyelography isparticularly important for the detection of ureteric stones: In 23 cases with ob-structive uropathy the site of the obstruction was determined and using conven-tional MR images, abdominal plain radiographs and urinary cytology, we werealways able to determine the cause.Conclusion: Drug induced MR-pyelography allowed a morphological study ofobstructed and non obstructed renal collecting systems and is suggested forpatients with contraindications for i.v. administration of contrast agents; in non-functioning kidneys it is considered the best diagnostic imaging modality to per-form after US.

B-209 15:12

Vertical IVU is the «gold standard» to evaluate orthostatic reaction andanatomical condition of the upper urinary tractV.M. Builov, A. Borisanov; Yaroslavl/RU ([email protected])

Purpose: To evaluate the diagnostic value of orthostatic reaction of upper uri-nary tract (OR UUT) in vertical IVU, including combination of IVU with TV- roent-genoscopy (RTV) and in vertical IVU after CT examination.Patients and Methods: In 2001-2005 OR UUT was studied in 461 patients onvertical IVU, in 43 of them IVU was combined with RTV, in 11 IVU was performedafter CT with contrast enhancement.Results: Evaluation of OR UUT allowed assessment of the anatomy of the UUTtotal extents and the degree of impairment of urodynamic of UUT, caused bydifferent pathological entities such as renal and ureteric calculi, Fraley's syn-drome, pelvi-ureteric junction compression by renal vessels, extrinsic uretericcompression. The results of OR UUT to select the method of treatment of ob-struction of UUT: conservative therapy, distant lithotripsy, endoscopic or tradi-tional surgical operations.Conclusion: IVU in vertical position is a simple and efficient modality in the eval-uation of OR UUT. It gives important information, which can not be achieved withsophisticated modalities such as CTU and MRU.

Scientific Sessions

182 C D E FBA G

B-210 15:21 !Comparison of standard and low dose multidetector CT (MDCT) with 4Dtube current modulation in patients with urinary stone diseaseT.H.E. Mulkens1, R. De Wijngaert1, P. Bellinck1, S. Fieuws2, J.-L. Termote1;1Lier/BE, 2Leuven/BE ([email protected])

Purpose: To compare the use of tube current modulation in standard and lowdose MDCT in patients with renal colic.Methods and Material: 300 patients with a suspicion of urolithiasis were scannedwith 6- and 16-MDCT, 150 with standard dose (6-MDCT: 130 kV, 95 effectivemAs; 16-MDCT: 120 kV, 120 effective mAs) and 150 with low dose (6-MDCT:110 kV, 51 effective mAs; 16-MDCT: 120 kV, 70 effective mAs), all with 4D tubecurrent modulation (CareDose4D, Siemens).Results: Tube current modulation reduced effective mAs from 25% to 31% in allexams. Mean effective dose was 4.2 and 2.9 mSv for standard, 1.41 and 1.58mSv for low dose exams. Low dose exams reached an additional dose reductionof 45.5% to 62.5%. Excellent correlation exists between mean tube current andBMI: 0.85-0.86 in low dose, 0.87-0.88 in standard dose exams. There was a sen-sitivity of 96%-98%, specificity of 92%-94% and accuracy of 94%-95 % for lowdose exams, comparable with standard dose exams. Interobserver agreementwas excellent (mean kappa value of 0.925). 47% of the patients had an ureteralstone. An alternative diagnosis was identified in 16% and 15% by two experi-enced readers in low and standard dose, respectively, but only in 10%-12% instandard and 4%-5% in low dose exams by two inexperienced readers.Conclusion: Low dose MDCT with tube current modulation can be used as stand-ard procedure for evaluation of urolithiasis, even in obese patients.

14:00 - 15:30 Room L/M

Neuro

SS 311Stroke and cerebrovascular circulationModerators:R. Klingebiel; Berlin/DEP. Skjejø; Odense/DK

B-211 14:00

Relation between cerebral blood flow, cerebrovascular reactivity andseverity of leukoaraiosisF. Todua, D. Gachechiladze, M. Beraia; Tbilisi/GE

Purpose: Leukoaraiosis (LA) is a frequent finding among the elderly and in pa-tients with cerebrovascular disease. The aim of our study was to assess the rela-tionship between cerebral blood flow (CBF), cerebrovascular reactivity (CVR)and severity of LA.Methods and Materials: 114 patients (mean age 63.6 ± 12y) underwent brainMR imaging, MR-angiography and color Doppler of extra-intracranial vessels,and routine neurological examinations. Patients with prior stroke and high-gradecarotid stenosis were excluded from study. Total CBF was determined by addingvolume flow rates in both the internal carotid and vertebral arteries. CVR wasevaluated by means of TCD Diamox 1 g test. The severity of LA was rated bysemiquantitative rating scale by adding scores for periventricular and subcorticalareas.Results: Presence of multifocal hypointense cerebral lesions, revealed on MRimaging, was strongly correlated with LA. FLAIR sequences were most informa-tive for assessment of pathological changes. Examination revealed carotid arterymoderate stenosis in 43 (38%) of cases. CBF in LA patients was decreased (468- 1078 ml/min, mean 764 ml/min). A significant negative correlation was seenbetween periventricular white matter changes and CBF (r = -0.76). TCD showedincrease of pulsatile index (PI) -1.02 ± 0.22 and impaired CVR (+38.6% to Diamoxtest).Conclusion: Our data gives us the possibility to conclude that LA is not associ-ated with ICA atherosclerosis severity. There is a significant negative relationshipbetween total CBF and severity of LA. Due to increased vasomotor resistance inbrain circulation and decreased vasoreactivity the slowed blood flow may play animportant role in the development of LA.

B-212 14:09

Quantitative perfusion measurements in acute ischemic stroke usingfunctional CT and a distributed-parameter tracer kinetic model: Preliminaryresults and comparison with deconvolution-based perfusion measurementsS. Bisdas1, T.S. Koh2, L. Cheong2, M. Hartel3, J. Gurung1, T.J. Vogl1;1Frankfurt/DE, 2Singapore/SG, 3Katowice/PL ([email protected])

Purpose: To evaluate the potential of a distributed-parameter (DP) tracer kineticmodel for assessing cerebral perfusion in acute ischemic stroke using perfusion-CT.Methods and Materials: 15 patients with acute supratentorial stroke symptomsunderwent perfusion-CT measurements. Perfusion images were analyzed with aregion-of-interest-based method using a newly introduced DP model. Blood flow(F), fractional intravascular (V1) and extracellular extravascular blood volume (V2),transit time in vascular compartment (T), first pass extraction ratio (E) and per-meability-surface-area product (PS) maps were generated. A time-lag parameter(Tlag) was introduced so as to denote the time difference between the initial riseof the arterial and tissue density curves. Furthermore, a vascular pixel elimina-tion method was applied to avoid an overestimation of the F values. The kineticparameters as well as the infarct and penumbra size were compared to thosederived from a deconvolution-based tracer kinetic model.Results: The DP perfusion maps provided better resolution than the deconvolu-tion-based ones. Infarct and penumbra regions were delineated and demonstrat-ed distinct F, V1, and T values, which were significantly (P < 0.01) correlated inboth models. E and PS parameters were not specific to infarct and penumbradelineation. Significantly elevated E and PS values were found in two patientswith hemorrhagic transformation. Tlag maps allowed rapid identification of theareas with varying degrees of hypoperfusion and correlated better than T mapswith the F abnormalities (P < 0.01). V2 maps demonstrated abnormalities withinthe ischemic lesions.Conclusion: Our proposed DP model can reliably depict the hypoperfused brainlesions in acute ischemic stroke.

B-213 14:18 !CT-perfusion and angio-CT in acute ischemic stroke: Clinical and radiologymanagementL.F. Aguilera, D. Jimenez, C. Gomez Escalonilla, B. Sanchez, E. Fraile;Alcalá de Henares/ES ([email protected])

Purpose: To determine the accuracy of CT-perfusion (CTP) and angio-CT (CTA),and correlate results with the clinical evaluation during emergency admission topredict tissue outcome.Methods and Materials: Noncontrast-CT (NCCT), CTP and CTA were obtainedfrom 36 patients (mean age 64 years) with non-hemorrhagic hyperacute strokeand within 24 hours after beginning the episode (between 1 hour and 24 hours).The regional cerebral blood flow (CBF), cerebral blood volume (CBV) and meantransit time (MTT) maps were calculated to discriminate between infarcted, noninfarcted and penumbra areas.Head and neck CTA was performed from the origin of the aortic arch to the Circleof Willis. CTA was reviewed with MIP, MPR and vessel analysis program. Patientswere clinically evaluated according to the National Institutes of Health StrokeScore at the time of admission and discharge. Statistical analysis was performedby using the weighted kappa index (KI) with its 95% confidence interval (CI).Results: We analyzed NCCT, CTP and CTA images and assessed clinical out-come (NIHSS). The agreement was KI = 0.38 (CI: 0.14-0.43) in NCCT; KI = 0.59(CI: 0.29-0.75) in CTP; and KI = 0.48 (CI: 0.02-0.58) in CTA.Conclusion: CTP and CTA can accurately predict the short-term follow-up inf-arct size. Areas of ischemic tissue on CTP correlate with the condition duringclinical admission and allow the evaluation of clinical prognosis.This imaging protocol with CTP and CTA (24-hour availability) is a fast and easymethod to predict tissue outcome and select hyperacute stroke cases that couldbenefit from fibrinolytic treatment.

B-214 14:27

Apparent diffusion coefficient reflects crossed cerebellar diaschisis inpatients with ischaemic hemispheric strokeY. Liu1, J.O. Karonen2, M. Könönen1, J. Nuutinen1, E. Vanninen1, J. Kuikka1,R. Vanninen1; 1Kuopio/FI, 2Mikkeli/FI ([email protected])

Purpose: Crossed cerebellar diaschisis (CCD) refers to a matched hypoperfusionand hypometabolism in the a cerebellum due to remote effect from a conteralat-eral supratentorial lesion. The purpose of our study was to evaluate whether hem-ispheric ischaemic lesions could cause apparent diffusion coefficient (ADC)changes in the cerebellum.

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 183C D E FA G

Methods and Materials: 22 patients (70 ± 9 years) with first-ever acute ischae-mic stroke underwent diffusion-weighted MR imaging (DWI) and SPET within thefirst 48 hours and at day 8 after onset of stroke. CCD was considered to be presentwhen SPET interhemispheric asymmetry index (AI) was > 5% in cerebellum. ADC-values of bilateral cerebellar hemispheres were obtained from each patient and15 age-matched healthy volunteers (67 ± 6 years).Results: Within first 48 hours, the bilateral cerebellar ADC-values did not signif-icantly differ from volunteers or between the ipsilateral and contralateral cerebel-lar hemispheres (p = 0.867). However, on day 8, the ADC-value in the contralateralcerebellum was significantly higher than on the ipsilateral side (71 ± 5×10-5 mm2/s vs. 69 ± 4×10-5 mm2/s, p = 0.037), being also significantly higher than the aver-aged (right/left) ADC-value of volunteers (67 ± 4×10-5 mm2/s, p = 0.013). No suchdifference was found on the ipsilateral side. CCD in SPET was present in 11(50%) patients at acute stage, and in 14 (64%) patients on day 8. The differencein ADC between two sides significantly correlated with the AI in SPET within 48hours (r = 0.611, p = 0.003) and on day 8 (r = 0.617, p = 0.002).Conclusion: Elevated ADC-values indicate development of mild vasogenic oede-ma in the contralateral cerebellar hemisphere one week after stroke. ADC-valuescan potentially be used to investigate the mechanism of crossed cerebellar dias-chisis.

B-215 14:36 !Multidetector-CT densitometry of acute basilar occlusionD. Gadda, L. Vannucchi, F. Niccolai, A.T. Neri, L. Carmignani, P. Pacini;Pistoia/IT ([email protected])

Purpose: Increased density of the basilar artery on unenhanced CT may be asign of acute occlusion, whose early diagnosis, sometimes clinically challenging,is crucial to consider intra-arterial fibrinolysis. Slice thickness reduction, madeeasier by multidetector technology, decreases posterior fossa artefacts that canaffect the correct estimation of the attenuation values of the vertebro-basilar trunk.The aim of the study was to determine the attenuation values, measured on1.25 mm thick contiguous axial CT scans, that could indicate the presence ofacute basilar occlusion on unenhanced CT.Methods and Materials: We retrospectively studied 8 patients with basilar oc-clusion that underwent unenhanced-CT with a 4-detector scanner within 7 hoursof symptoms onset. The site of occlusion was defined by CT Angiography in 7cases and by MR Angiography in one. The attenuation values of the occludedarterial segments were measured and compared with those of patent ones of thevertebro-basilar trunk. Density differences were compared using Student's t-test.Results: The difference between the mean attenuation values of the occluded(65.1 HU, 95%CI 62.75-67.50, SD2.30) and the patent vessels (43.1 HU, 95%CI40.75-45.50, SD3.80) was significant (p < 0.001). In any patient a segmental in-crease in density of basilar artery of more than one third of the mean attenuationvalues of the other segments of vertebro-basilar trunk was found to be related toacute occlusion.Conclusion: Densitometric criteria to suspect on unenhanced CT an acute basi-lar occlusion can be defined, when they are consistent with neurological symp-toms and when thin-collimations are utilized.

B-216 14:45

Novel rehabilitation hand robots and fMRI in strokeA.A. Tzika, A. Khanicheh, A. Muto, C. Triantafyllou, L.G. Astrakas,C. Mavroidis; Boston, MA/US ([email protected])

Purpose: We combined functional Magnetic Resonance (fMRI), with a novel MRcompatible hand-induced, robotic device (MR_CHIROD) that employs a non con-ventional type of actuation, via electro-rheological fluids (ERF), which can changetheir viscosity in response to electric field.Methods and Materials: One optical encoder and one force sensor was attachedto the ERF resistive element's shaft to measure the patient induced motion andforce, respectively. The currents and voltages were recorded; localizer and T1EPI images were acquired.Results: The plot of current versus voltage for the actuated ERF inside the MRenvironment (zone 2) and outside the MR environment showed that the perform-ance of ERF was not affected by the MR environment. Images from zone 4 (thearea where the patient is located) showed that when ERF is actuated at 2.5 and4 kV no image shift exists. The measurements of SNR for six different voltageswere: 88.6 (SD=1.42) for slice 1, 90.2 (SD=1.39) for slice 2, and 88.1 (SD=1.46)for slice 3. Significant loss of SNR was not observed regardless of the ERF beingactuated in all zones where experiments were performed since 96% of the datapoints lie within ± 2 SD.Conclusion: A novel force-feedback MR compatible device for hand rehabilita-

tion was designed. Our results demonstrated that the MR environment does notaffect the ERF properties and that the ERF device (MR_CHIROD) did not de-grade the MR images. MR_CHIROD may aid in the evaluation of rehabilitationafter stroke.

B-217 14:54

Restenosis after carotid artery stentingM. Puglioli, M. Cosottini, M. Michelassi, S. Picchietti, I. Bargellini, G. Lazzarotti,G. Orlandi, C. Bartolozzi; Pisa/IT ([email protected])

Purpose: Carotid artery stenting (CAS) is an alternative method to endarterec-tomy in the revascularization of carotid artery stenosis.The long term efficacy of CAS is still debated. The aim of our study was to assessthe survival/neurological free event rates and the incidence of long term intra-stent restenosis after CAS and to identify some clinical or radiological predictorsof restenosis.Methods and Materials: We conducted a retrospective study selecting from ourdata base 191 patients treated between 1997 and 2004 that had a clinical andCD-US follow-up. The survival Kaplan-Meier curves were obtained for overallsurvival, for freedom from restenosis and freedom from stroke-TIA/all death. Weused a univariate analysis to correlate clinical, radiological and procedural datato stenosis recurrence and a multivariate analysis to determine independent pre-dictors of restenosis.Results: The overall peri-procedural complication rate was 3.1% (1.6% of minor,0.5% of major and 1 % of fatal stroke). In-stent restenosis was observed in 11.5%of procedures with a cumulative primary patency rate of 88.2, 82.6 and 73.4 at 1,3 and five years respectively. The cumulative overall survival rate at 1, 3 and fiveyears was 93.7, 82.4 and 70.9 with a stroke/TIA free survival rate of 87.8, 76.5and 66.4. The percentage of post-stenting residual stenosis was the only signifi-cant predictive factor for restenosis (P= 0.003).Conclusion: Encouraging results seem to derive from both neurological eventfree rate and restenosis incidence. Adequate recanalization of the vessel treatedwith CAS is important to limit the development of restenosis.

B-218 15:03

Cognitive changes after carotid artery stentingI.Q. Grunwald, T. Supprian, T. Struffert, M. Politi, K. Zercher, V. Vedder,M. Backens, C. Krick, W. Reith; Homburg/DE ([email protected])

Purpose: The aim was to test for changes in cognitive performance after carotidartery stenting (CAS).Methods and Materials: 30 patients were neuropsychologically tested at least24 hours before and 48 hours after carotid artery stenting. All patients receiveddiffusion and perfusion weighted MR imaging before and after CAS. The follow-ing neuropsychological tests were selected: MMSE, symbol digit test and sub-tests of the CERAD battery (verbal fluency test, constructional practice, word listmemory, delayed recall). Affective state was determined by the BECK Depres-sion Inventory.Results: None of the patients suffered from depression (BDI< 1) or dementia(MMSE 29.9 ± 1.5). 28/30 patients (p = 0.11) showed increased speed in ZVT(corresponding to trail making test). Most patients showed better or at least sim-ilar results concerning delayed recall (p = 0.31). No change was observed con-cerning digit symbol test, word list memory and verbal fluency and constructionalpractice. There was a correlation between perfusion (TTP) and ZVT.Conclusion: Better results concerning ZVT and delayed recall after carotid stent-ing might be due to improved brain perfusion. After carotid stenting cognitive andmemory performance seems to improve. Stenting of the ICA may offer more thanreduced stroke risk to patients with impaired perfusion.

B-219 15:12

High resolution MR imaging at 3.0 T in the short-term follow-up of patientswith proven cervical artery dissectionR. Bachmann1, I. Nassenstein1, R. Dittrich1, H. Kugel1, H. Kooijman2,G. KuhlenbÃ¤umer1, W. Heindel1, S. Kraemer1; 1Muenster/DE, 2Hamburg/DE([email protected])

Purpose: Assess short-term morphological changes in patients with proven cer-vical artery dissection (CAD).Methods and Materials: 66 patients with suspected CAD were examined on a3.0 T system (Gyroscan Intera, Philips). Imaging protocol consisted of 3D inflowMRA (TR/TE/FA = 25 ms/3.1 ms/16Â°, reconstructed voxel size 0.3x0.3x0.8 mm);black blood T1w 3D spoiled GE (TR/TE/FA = 31 ms/7.7 ms/15Â°, 0.3x0.3x1.0 mm)and fat suppressed T2w TSE (TR/TE/ETL = three heart beats/44 ms/7,

Scientific Sessions

184 C D E FBA G

0.3x0.3x2 mm). In 22 pts a dissection was proven by direct visualization of anintramural hematoma, 19 pts (5 male, 14 female, mean age 40.9 y) were availa-ble for follow-up. Images were assessed with regard to degree of stenosis, size ofintramural hematoma, presence of intraluminal thrombus, development of pseu-doaneurysms and incidence of new dissections.Results: Mean interval between initial study and follow-up was 14.7d. InitiallyCAD was proven in 21 vessels (14xICA, 7xVA). 18 vessels showed a stenosis, 3were occluded. On follow-up study the degree of stenosis increased in 3 vessels,remained unchanged in 11 vessels and decreased in 4. All 3 occluded vesselswere recanalized on follow-up. In 3 vessels a pseudoaneurysm was visible on theinitial study and remained unchanged on follow-up; in 1 vessel a new pseudoan-eurysm was observed. In 3 vessels new dissections were identified on follow-upstudies.Conclusion: High resolution MR imaging permitted excellent analysis of mor-phological features of CAD and clinically important information could be gained.Thus, short-term follow-up in pts with acute CAD is recommended. Further stud-ies are needed to assess the relationship between short-term results and defi-nite outcome.

B-220 15:21 !CT perfusion diagnosis of brain deathM. Torres, J. Arnáiz, T. Piedra, L. Martín, J. Jordá, E. Ruiz Pérez,E. Marco de Lucas, E. Sánchez, A. González Mandly; Santander/ES([email protected])

Purpose: To analyze the usefulness of perfusion computed tomography in theimaging diagnosis of brain death prior to organ donation.Methods and Materials: We prospectively examined 12 studies in 8 patients(17-67 years, 6 men) with suspected brain death where electrophysiological di-agnosis of brain death was impossible because of drug intoxication. A perfusionCT and angiographic CT study was performed and analyzed as confirmatorytests, and compared with subsequent clinical evolution of these patients.Results: Perfusion CT demonstrated a very good accuracy in assessing braindeath diagnosis, as a complimentary tool of the angiographic study performedwith CT. Angiography demonstrated non-filling of intracranial arteries at the entryof the skull in 6 studies. Perfusion CT confirmed an arrest of cerebral blood flowin both carotid and posterior circulation, and persistent flow in external carotidartery territories. In 4 patients reduced but persistent cerebral blood flow wasdemonstrated, and in three of them a new CT angio and perfusion study wasperformed several hours later that confirmed then a complete absence of bloodflow.Conclusion: Perfusion CT is a very good technique to study brain death be-cause it is rapid and easy to interpret. Diagnostic parameters may be discussedin future studies in order to obtain similar standard parameters between differenttechniques classically used in these patients.

14:00 - 15:30 Room N/O

Breast

SS 302MR imagingModerators:C. El Khoury; Paris/FRF. Sardanelli; Milan/IT

B-221 14:00

The diffusion-weighted magnetic resonance imaging: The role in thedifferential diagnosis of breast lesionsC. Altay, P. Balci, T. Canda, S. Saydam, B. Demirkan, I.B. Gorken, O. Dicle;Izmir/TR ([email protected])

Purpose: To evaluate the diagnostic value of diffusion imaging in distinguishingbenign and malignant breast lesions preoperatively.Methods and Materials: A total of 37 women between the age of 22-74 years(mean, 37 years) with 34 histopathologically verified breast lesions were investi-gated in this study. The patients were examined by a 1.5 T system using bilateralphased array breast coil. Spin echo EPI diffusion imaging was used to scan pa-tients. Images were obtained by b values 0 and 1000 seconds/mm2. The ADCvalues were calculated on normal fibroglandular tissue and breast lesions. Thecomparison between the histopathological diagnoses and the mean ADCs wereperformed by Mann Whitney U test.

Results: Out of 37 women, 4 had normally breast MR imaging findings. The diag-nosis of remaining 33 patients with 34 breast lesions were as follows; malignantlesions (n = 21), benign lesions (n = 10) and simple breast cyst (n = 3). The ADCvalues were as follows (in units of 10-3 mm2/sec): Normal fibroglandular tissue(range: 1.19-2.07, mean:1.61), benign breast lesions (range:1.08-2.46,mean:1.55), cysts (range: 2.29-2.38, mean: 2.33) and malignant breast lesions(range: 0.48-0.92, mean: 0.75). The mean ADC obtained from malignant breastlesions was statistically different from that observed in benign solid lesions(p = 0.00) and normal fibroglandular breast tissue (p = 0.00). Furthermore, themean ADC values of benign breast lesions were not statistically different fromcyst (p = 0.01) and normal fibroglandular breast tissue (p = 0.26).Conclusion: Diffusion imaging can be used in differentiation of malignant andbenign breast lesions.

B-222 14:09

Venous drainage of the breasts: Is there difference between right and left?H. Sherif, A.E. Mahfouz; Doha/QA ([email protected])

Purpose: To evaluate pattern of venous drainage of the breasts on gadolinium-enhanced MR angiography.Methods and Material: 150 women were examined by gadolinium-enhancedMR imaging for suspected breast masses (57 with breast carcinoma, 23 withbenign lesions, and 70 with normal breasts). At 1.5 T, coronal 3D T1-weightedGRE sequence was performed before and 60 s, 120 s, and 240 s after injectionof Gd-DTPA. Transverse T1-weighted SE images of the upper thorax and axillaewere obtained. Angiographic images were generated by subtraction and maxi-mum intensity projection. The shortest distance between sternum and arteries atlevel of left innominate vein was measured.Results: In 51 patients (34%), veins crossed from the left breast to drain in rightinternal mammary vein. In the remaining 99 women, veins from each breast drainedseparately in the internal mammary vein of the same side. Distance betweensternum and arteries at the crossing of left innominate vein was 5 ± 4 mm inpatients with crossing veins and 15 ± 6 mm in patients without crossing veins(p < 0.01). There was no difference in percentage of patients with crossing veinsamong patients with carcinoma, benign lesions, or normal breasts.Conclusion: Subcutaneous veins from left breast drain in the right internal mam-mary vein in 34% of women. No crossing occurs in the opposite direction. Inwomen with crossing veins, distance between aortic arch and sternum was sig-nificantly smaller than in other women. The crossing venous drainage of left breastcan be explained by left innominate vein compression between sternum and ar-teries.

B-223 14:18

Breast MR and microcalcification: Does MR help in managing clusters ofmicrocalcification detected by mammogram?C. Del Frate, A. Bestagno, V. Londero, M. Urbani, R. Pozzi Mucelli,M. Bazzocchi; Udine/IT ([email protected])

Purpose: To evaluate the role of breast-MR in the management of clusters ofmicrocalcification detected by mammography.Methods and Material: We retrospectively reviewed 68 patients, with clusters ofmicrocalcification, that underwent breast-MR (1 T system) one-week before breast-biopsy. Clusters of microcalcification were classified according to BIRADS-score.Enhancing areas, whenever detected by MR, were classified according to Fish-er-score. Comparison was finally made with pathologic findings.Results: Histology of 68 clusters revealed 33 benign (22 FCD, 3 TDH, 5 adeno-sis, 1 sclerotic tissue), 7 high risk proliferative benign lesions (2 LCIS, 3 ADH, 1papilloma, I sclero-elastotic nodule), 30 malignant lesions (12 DCIS, 10 IDC, 5IDC-DCIS, 3 ILC). MR correctly classified 23/30 malignant lesions and 32/38benign lesions, with 7 false-negatives (5 DCIS, 2 IDC grade 1) and 6 false-posi-tives (4 FCD, 2 adenosis), with sensitivity, specificity, PPV, NPV of 76.67%,84.21%, 79.31% and 82.05%, respectively. Mammography correctly classified29/30 malignant lesions and 20/38 benign lesions, with 18 false-positives and 1false-negative, with sensitivity, specificity, PPV, NPV of 96.67%, 52.63%, 61.70%and 95.24%, respectively.In 13/18 false-positives at mammography, MR correctly classified lesions as neg-ative, while the 1 false negative at mammography was classified as positive byMR. In 5/6 false positive at MR, mammography was false positive too.Conclusion: Breast-MR improves specificity and potentially reduces false-posi-tive rate of mammography, for clusters of microcalcification. Nevertheless, MRmissed 7 malignant lesions, even though low-grade lesions. Therefore, breastbiopsy cannot be avoided when breast MR results negative.

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 185C D E FA G

B-224 14:27

MR imaging of the breast post biopsy: Do post biopsy changes limit MRimaging interpretation subsequently?A. Mehdizade, S. Curran, T. Miller, L. Bartella, E. Morris; New York, NY/US([email protected])

Purpose: The purpose of this study was to assess the frequency and spectrumof post biopsy changes on MR imaging of breast following percutaneous needlebiopsy and their effect on subsequent MR interpretation. This has not been previ-ously described in the literature to our knowledge.Methods and Materials: From February 1998 to May 2005, the records of 415patients who had breast MR imaging was retrospectively reviewed. From these,51 had needle core or stereotactic biopsy prior to the breast MR imaging. Pa-tients who had lumpectomy, surgical excision or fine needle aspirations wereexcluded form this study. The average age was 55 ranging from 39 to 73 years.Lesion size ranged from 0.4 to 3.0 cm, median 1.4 cm. The mean interval timebetween the MR and biopsy was 213 days with a range of 1 to 545 days.Of these examinations, 42 had ultrasound guided breast core biopsies, and 11Stereotactic biopsy, prior to the MR imaging exam. From these; 14 were benignand 37 were malignant. Analyses of MR imaging was performed by two radiolo-gists blind folded to the results to detect hematomas, distortion, needle trackartefacts or abnormal enhancement, limiting subsequent MR diagnosis.Results: Post biopsy hematoma was observed in one case only. There were nosignificant enhancement following the MR biopsies, or needle track artefacts ob-scuring the lesions.Conclusion: Post biopsy changes do not limit MR imaging interpretation. MRimaging of the breast can be performed anytime after stereotactic or US guidedcore biopsy.

B-225 14:36

Relationship between MR imaging morphofunctional findings andhistopathological characteristics of breast cancerL. Martincich, G. De Rosa, F. Montemurro, V. Marra, I. Bertotto, D. Regge;Candiolo/IT ([email protected])

Purpose: To correlate MR imaging morphofunctional findings with histopatho-logical characteristics of breast cancer.Methods and Materials: 68 patients with breast cancer (clinical stage I in 7, IIAin 12, IIB in 23, IIIA/B in 26) underwent breast MR imaging followed by core-biopsy. MR imaging was performed at 1.5 T acquiring morphological (T2-weight-ed sequence) and dynamic (3DFSPGR sequence before and after intravenousadministration of Gadolinium chelates) studies.MR imaging morphofunctional findings were criteria described by Fischer in 1999(shape, margins, contrast material kinetic, early contrast uptake [ECU] and sig-nal/intensity curve morphology) evaluated as single criterion and as scoring sys-tem. Core-biopsy was performed for determination of histopathologicalcharacteristics: histological type, nuclear grade, estrogen and progesteron re-ceptors (ER, PgR), Ki-67 and cerbB2 expression. Cut off for positivity was 20%for Ki67, 10% for ER and PgR and 3+ for cerbB2.Results: MR imaging identified 68 single node breast cancers (57DCI, 11LCI).Fischer's scoring system resulted 2 in 1 case, 3 in 1, 4 in 10, 5 in 11, 6 in 16, 7 in12 and 8 in 17. ER, PgR, Ki67 and cerbB2 were positive in 38, 30, 40 and 15lesions, respectively. Significant association was found between higher Fischer'sscoring system and cerbB2 negative status (p = 0.002).No significant correlation between MR imaging single criteria and histopatholog-ical characteristics were found.Conclusion: Our study shows that combining MR imaging morphofunctional find-ings may identify breast cancer with different biological characteristics. In partic-ular a significant correlation was found between Fischer's scoring system andexpression of cerbB2 which is a marker of chemoresponsiveness and prognosis.

B-226 14:45

Is the grading of breast cancer influenced by the vascularization of themalignancy detectable in MR using MR-CAD?A. Malich1, K. Huskobla2, R. Gorna1, D. Fischer3; 1Nordhausen/DE, 2Jena/DE,3Berne/CH ([email protected])

Purpose: According to theoretical studies it could be assumed, that low-grademalignancies have a less aggressive contrast uptake vs. high-grade malignan-cies. CAD-systems automatically analyze contrast uptake and present a percent-age distribution of different enhancing types. Whether this helps to clarify gradingof breast cancer was evaluated.Methods and Material: All patients undergoing dynamic breast-MR imaging

(0.1 mmol Gd-DTPA/kgbw) using a standard protocol from 6/2003-12/2003 witha histopathological verification were included. CAD-based (Confirma, USA) col-our coded analysis of wash in was differentiated into intermediate (> 50%, C1)and strong > 100% uptake (C2), later enhancement into continued enhancement(C3), plateau (C4) and wash out (C5).Results: 11/110 MR imaging were not transferable to CAD. 53 enhancing lesionsremained without a lesion's automated volume calculation (exclusively technicalreasons). Enhancement distribution of G1, G2, G3 and Gx malignancies withinthe remaining 67 lesions were:C1: 88.6% / 83.0% / 76.7% / 87.8%; C2: 11.4% / 17.0% / 23.3% / 12.2%; C3:42.6% / 48.9% / 45.7% / 55.1%; C4: 32.1% / 19.4% / 26.1% / 19.0%; C5: 25.3%/ 31.7% / 27.7% / 25.9%Combination of fast initial wash in followed by a wash out was most common inG3: 12.9%; G2: 8.9%, G1: 2.1%; Gx: 5%. Mean curve peaks were: 172% (G1);121% (G2); 133% (G3) and 98%(Gx).Conclusion: Despite still existing technical problems CAD offers for the first timea reliable, automated analysis of percentage distribution of enhancing types withina suspicious breast lesion. A discrimination of the grading seems, however notyet possible using the dynamic enhancement distribution offered by the CAD-system.

B-227 14:54

Pure ductal carcinoma in situ of the breast (DCIS): Parametric MR imageevaluation and correlation with histological and biomolecular factorsS. Ganau1, M. Sentís1, L. Tortajada1, X. Andreu1, E. Saez1, M. Villajos1,J. Planas1, J. Camps2, S. Perez1; 1Sabadell/ES, 2Alzira/ES ([email protected])

Purpose: To assess the value of Parametric MR Images (PI) evaluation and itsrelationship with biomolecular factors in Ductal Carcinoma in Situ (DCIS).Methods and Materials: One hundred and forty-four patients with pure DCISwithout micro-infiltration or association to another infiltrating carcinoma (n = 146)were selected. Standard international dynamic protocol was used. Subtractedimages were used to increase lesion conspicuity, and ROIs analysis to evaluatecontrast uptake. Normalization of signal values was used in interpatient compar-ison.PI evaluation includes: wash-in, wash-out, time-to-peak, positive enhancementintegration, standard deviation over time. These parameters were calculated inall cases. Findings in MRM were compared to a range of histological and biomo-lecular factors including: ki-67, p-53, hormonal receptors (estrogen and proges-terone), HER2neu, VEGF and CD-31.Results: In 134 lesions enhancement was found (92 %). High grade DCIS showedsignificantly higher enhancement values than lower and intermediate grade anda more characteristic signal intensity curve. In parametric evaluation high gradeDCIS is related with high wash-in and short time to peak values.Delayed subtracted images are needed in order to depict DCIS with low enhance-ment profile. High signal intensity and high velocity contrast uptake are relatedwith high grade lesions This group of lesions shows overexpression of VEGF andincreased capillary density in the surrounding stroma.Overexpression of HER2-neu is associated with continuous and dendritic grow-ing pattern.Conclusion: Signal behaviour in DCIS appears to be related with biomolecularmarkers and quantitative evaluation is a promising tool in the diagnostic process.

B-228 15:03

Value of magnetic resonance imaging for the work-up of invasive lobularbreast carcinoma: Prospective and retrospective study of 39 cases.Comparison with physical examination, conventional imaging and histologyT. Caramella, C. Chapellier, F. Ettore, I. Raoust, E. Chamorey, C. Balu-Maestro;Nice/FR ([email protected])

Purpose: Determination of the value of MR imaging for invasive lobular carcino-ma (ILC), which remain a diagnostic challenge for the radiologist.Methods and Materials: 39 patients with histologically-proven ILC were studiedbetween 1998 and 2005. All patients underwent physical examination, mammog-raphy, ultrasound and an MR imaging. All anomalies detected were graded usingthe BI-RADS classification and careful research was made for the presence ofmultifocal/multicentric disease. The histologic tumor size was compared with theother techniques.Results: 77% of the tumors were palpable. The majority of the mammogramspresented had architectural distorsion (46%) or a mass (49%), most of them withspiculated contours. Sonograms demonstrated an irregularly-shaped mass. MRimaging Results: 67% were visualized as masses and 21% corresponded to asite of enhancement. Only 70% of the enhancements were typically malignant.

Scientific Sessions

186 C D E FBA G

Size correlations, based on comparison with histology findings were: 0.90 (MRimaging, p < 0.001), 0.56 (ultrasound), 0.45 (physical examination), and 0.31(mammography).The Kappa correlations for multifocal/multicentric disease were: 0.89 (MR imag-ing), 0.23 (ultrasound), and 0.14 (mammography). MR imaging had a sensitivity,specificity and positive predictive value of 100%, 89% and 100%, respectively.The therapeutic strategy was modified in 54% of cases.Conclusion: MR imaging is unquestionably valuable for the management of ILC.It provides the most accurate estimate of tumor size and is highly sensitive formultifocal/multicentric disease. Its main drawback is a lack of specificity that re-quires preoperative histologic verification. MR imaging often justifies modifica-tion of the initial therapeutic strategy, generally by ruling out a conservativeprocedure.

B-229 15:12

SPIO-MR lymphography for detection of axillary sentinel lymph node inpatients with breast cancerK. Ishiyama, Y. Katayose, N. Tomura, R. Sashi, K. Narita, J. Watarai,J.-I. Ogawa; Akita City/JP ([email protected])

Purpose: To identify the sentinel lymph nodes on interstitial MR lymphographyafter injection of superparamagnetic iron oxide (SPIO; ferumoxides) in patientswith breast cancer.Method and Materials: MR lymphography was performed in 11 patients withbreast cancer without palpable axillary lymph node before surgery. Coronal andaxial images were obtained before and after interstitial injection of SPIO. All thepatients had interstitial injection of 1.5 ml SPIO (16.8 mg iron per kg) adjacent tothe breast tumor. Post contrast MR lymphograms were obtained 20, 40, 60 min-utes after injection of SPIO. All images were obtained with a 1.5-Tesla MR scan-ner and anterior 5-inch and posterior 5-inch coils connected to the dual phasedarray system. 2D-fast SPGR sequence was used in all patients. Coronal andaxial images (TR: 180 msec, TE: 4.2 msec) using 2D-fast SPGR sequence wereobtained.Results: In all patients, MR imaging could show the axillary lymph nodes withdecreased signal intensity. One to six lymph nodes for each patient with decreasedsignal intensity, considered to be the sentinel lymph nodes, appeared within20 minutes. SNR of lymph nodes exhibiting influx of SPIO was significantly lowerthan that before injection.Conclusion: This study elucidated that MR lymphography could detect the sen-tinel lymph nodes with decreased signal intensity 20 minutes after interstitial in-jection of SPIO. This technique might offer potential for the evaluation of the sentinellymph nodes in patients with breast cancer.

B-230 15:21

Primary chemotherapy for locally advanced breast cancer using MRimaging: Parametric MR image evaluation and biomolecular factorscorrelation, monitoring responseM. Sentís1, L. Tortajada1, E. Saez1, X. Andreu1, S. Ganau1, M. Villajos1,J. Camps2, M. Segui1, A. Luttich1; 1Sabadell/ES, 2Alzira/ES ([email protected])

Purpose: To assess the value of parametric MR image evaluation and its rela-tionship with biomolecular factors monitoring cancer response in Primary Chem-otherapy (PCT).Methods and Materials: Prospective study of 85 patients diagnosed of LABCbetween 2002 and 2004 by core biopsy treated with four cycles of epirrubicina-cyclophosfamide, followed by 8-12 weekly doses of docetaxel. Surgery after PCT.Pathological evaluation using Miller & Payne histological grading MR studiesbefore, after the first cycle, and after PCT.Imaging Protocol: T2w and dynamic FLASH3D sequences with 0.16 mmols/bwKg.Measurements: Enhancement was quantified by ROI analysis. Normalization ofsignal values was used for interpatient comparison. Parametric MR image evalu-ation included : wash-in, wash-out, time-to-peak, positive enhancement integra-tion and standard deviation over time. MR findings were compared withbiomolecular factors: Hormonal receptors, proliferative index (KI67), expressionof TP53, Bcl2, EGFR and HER2neu and detection of gene amplification (HER-2and Topoisomerase IIα) by chromogenic in situ hybridization (CISH).Results: Response rate was 84%. Good correlation between MR and pathologyin 68%, underestimation 16%, overestimation 14%. Complete response (CR) isrelated with negative ER/PR tumors (p = 0.025). High proliferative index (Ki67> 25%) had significant association with CR (p < 0.05). Decrease of normalizedmaximum signal intensity values, wash-in and wash-out rates can be observed inthe group of tumours with CR in the intermediate evaluation. This profile seemsto be related with the basal phenotype tumours recently described.

Conclusion: Signal behaviour in PC appears to be related to some biomolecularmarkers and quantitative evaluation is a promising tool in the diagnostic process.

14:00 - 15:30 Room P

Vascular

SS 315CT angiographyModerators:W.P.T.M. Mali; Utrecht/NLA. Redheuil; Paris/FR

B-231 14:00

Multislice CT angiography of spinal cord feeding arteries before and afterendografting of the thoracic aortaH. von Tengg-Kobligk, T.M. Jose, D. Böckler, M. Ganten, S. Nagel, F.L. Giesel,S. Delorme, H.-U. Kauczor; Heidelberg/DE ([email protected])

Purpose: To study the diagnostic value of multi-slice CT angiography for imagingof the spinal cord feeding arteries in patients with complex thoracic aortic pathol-ogies undergoing endovascular aortic repair (EVAR). The impact of proximal pos-terior intercostal artery occlusion for the visualization of the spinal blood supplywas evaluated.Method and Materials: Eighteen patients (5 women, mean age 60 years) withthoracic aortic type B dissection or aneurysm underwent two CTA examinations(pre- and post-EVAR, 120 kV, 120-150 mAs, 1 mm collimation, 16-slice CT scan-ner). Contrast agent (400 mgI/ml) was administered at a flow rate of 5 ml/s. Pulserate and neurological status were documented. Qualitative analysis was performedby two experienced radiologists counting the posterior intercostal arteries, theirrami dorsales and the Adamkiewicz artery (AKA). Density measurements includ-ed the ascending and descending aorta.Results: Contrary to expectations visualization of the posterior intercostal arter-ies and rami dorsales could be visualized post-EVAR due to retrograde perfusion.On average 10/11 posterior intercostal arteries, 8/9 rami dorsales were retrogra-dly perfused post-EVAR within the stented aortic segment (mean length 237 mm).Additionally, 9/10 AKA were visualized. ROI analysis in the thoracic aorta con-firmed high mean HU values of 445 HU pre- and 385 HU post-EVAR. High pulserates and partially thrombosed lumen in type B dissection and aortic aneurysmlimited the visualization of the AKA in some patients. Neurological events werenot observed in any patient.Conclusion: Multi-slice CT angiography is capable of visualizing even small spi-nal cord feeding arteries. High contrast imaging showed patent post-EVAR spi-nal vessels due to retrograde perfusion.

B-232 14:09

Visualization of the artery of Adamkiewicz at ECG-gated multi-detector rowhelical CT in patients with Stanford type A aortic dissectionA. Kovacs, W. Schiller, T. Sommer, M. Hackenbroch, H. Schild, S. Flacke;Bonn/DE ([email protected])

Purpose: To assess the detectability of the arteria radicularis magna sive Adam-kiewicz at ECG-gated multi-detector row helical computed tomography (MDCT)in patients with Stanford type A aortic dissection.Methods and Material: 51 patients with Stanford type A dissection of the aortaunderwent contrast enhanced ECG-gated multi-detector row helical CT of theentire aorta and the iliac arteries (slice coll.16x1.5 mm; rotation time 0.42 s; 120 kV;300 mAs; 120 ml Ultravist 300Â®). The visualisation of the artery of Adamkie-wicz, as well as its branching level, side origin and provenience from the true/false lumen was investigated by two independent readers on source and multi-planar reformation images.Results: In 36 (70%) of the 51 patients a single artery of Adamkiewicz was vis-ualized from the stem of the intercostal or lumbar artery up to the anterior spinalartery. Two arteries of Adamkiewicz in the same patient were not observed. 30(83%) arteries of Adamkiewicz originated from the left side. 30 (83%) originatedbetween T8 and L1. 23 (64%) arteries originated from the true lumen and 13(36%) from the false lumen.Conclusion: Multi-detector row helical CT depicts the artery of Adamkiewicz in ahigh percentage of patients with Stanford type A aortic dissection. The providedinformation may help to develop strategies in protecting spinal cord ischemiarelated to thoracic and thoracoabdominal aortic surgery.

FF FFFrida

rida

rida

rida

rida

yy yyy

Scientific Sessions

B 187C D E FA G

B-233 14:18

Retrograde aortic dissection: An underestimated complication aftersupracoronary aortic replacement in patients with Stanford type A aorticdissectionA. Kovacs, W. Schiller, T. Sommer, M. Hackenbroch, H. Schild, S. Flacke;Bonn/DE ([email protected])

Purpose: To determine the presence and the extent of retrograde aortic dissec-tion in patients with Stanford type A dissection treated by supracoronary aorticreplacement.Methods and Material: 44 patients with treated Stanford type A dissection of theaorta underwent contrast enhanced ECG-gated multi-detector row helical CT (slicecoll.16x1.5 mm; rotation time 0.42 s; 120 kV; 300 mAs; 120 ml Ultravist 300Â®).CT -scans were obtained 10 to 70 months after supracoronary aortic replace-ment. The presence and the extent of a retrograde dissection into the sinus ofvalsalva was assessed by two experienced readers in consensus on axial slicesand multi-planar reconstruction images. The presence of a retrograde dissectionwas correlated to the time interval after operation.Results: 13 of 44 patients (30%) showed a retrograde dissection. This retro-grade dissection always extended into the non-coronary sinus. The diameter ofthe sinus of valsalva was significantly larger (47 ± 11 mm) in these patients com-pared to those without retrograde dissection (39 ± 8 mm; p < 0.05). The time in-terval between operation and follow-up CT scan was significantly longer in thegroup of patients with retrograde dissection (46 months vs 22 months).Conclusion: The presence of retrograde dissection is an underestimated com-plication of supracoronary aortic replacement. The high incidence in the studygroup may be related to the excellent image quality of ECG-gated MDCT. A retro-grade dissection may develop after failure of tissue glue used to fix dissectedmembranes. This would be in agreement with the longer time interval betweenoperation and follow-up in the group of patients with retrograde dissections.

B-234 14:27

MSCT of aortic aneurysm: Should we study the whole aorta? Unexpectedaortic lesions in 163 patientsD. Farina, I. Manildo, I. Siamouta, R. Maroldi; Brescia/IT ([email protected])

Purpose: To assess rate and impact on surgical planning of unexpected aorticdilatations in patients studied for suspected thoracic (AAT)/abdominal (AAA) aor-tic aneurysm.Patients and Methods: During the period June 2004-2005, 163 consecutivepatients (age range 51-92) were studied with 16-row MDCT for suspected aorticaneurysm (thoracic in 43/163, abdominal in 120/163). MDCT protocol entailedplain scan in all patients, extending from the aortic arch to the femoral arteries.After contrast application, arterial phase was focused on the suspected aneu-rysm. Delayed arterial phase was again extended from the aortic arch to thefemoral arteries, whenever the pre-contrast phase demonstrated an unexpectedaortic dilatation on the opposite side of diaphragm. According to the InternationalSociety for Cardiovascular Surgery guidelines, both suspected and unexpectedaortic dilatations were classified as ectasia, or aneurysm with diameter below orabove cut-off for surgical intervention.Results: Overall, 45 unexpected aortic dilatations were detected (21/45 ectasia,15/45 aneurysm below cut-off, 9/45 aneurysm above cut-off); unexpected dilata-tions of thoracic aorta were observed in 28.3% of patients with suspected AAA,unexpected dilatations of abdominal aorta in 25.6% of patients with suspect AAT.Nineteen unexpected AAT (8/19 above cut-off) and 5 unexpected AAA (1/5 abovecut-off) were observed: consequently in 9/163 (5.5%) cases the unexpected AAhad the potential to change surgical planning.Conclusion: The high rate of unexpected aortic dilatations (overall 27.3%) sug-gests that the whole aorta should be scrutinized in patients with suspect AAT/AAA. Detection of unexpected aneurysms may either require careful follow-up oralter surgical planning.

B-235 14:36

MDCT angiography for the assessment of lower extremity arterial occlusivedisease: 4- vs. 16-detector MDCTH. Ota, K. Takase, H. Rikimaru, A. Sato, T. Yamada, S. Takahashi; Sendai/JP([email protected])

Purpose: The purpose was to compare the diagnostic performance of 4- and 16-detector MDCT angiography for evaluating peripheral arterial occlusive disease.Methods and Materials: 50 cases with symptomatic peripheral arterial occlu-sive disease underwent both MDCT angiography and DSA. In 28 of the 50 cases,MDCT angiography was performed by 4-detector system with 2-mm collimation

with image reconstruction of 2-mm slice thickness at 1-mm intervals, and in theremaining 22 cases, it was performed by 16-detector system with 1-mm collima-tion and 1-mm slice thickness at 1-mm intervals. In evaluating MDCT data, cross-sectional images perpendicular to the vascular course were observed by scrollingmethod. The arterial system from the iliac to the calf was divided into 9 segments.Pedal arteries were not included. Each segment was classified as normal, mildlystenotic, moderately stenotic, severely stenotic, or occluded. The diagnostic per-formance of MDCT angiography was determined using DSA as the referencestandard.Results: Of the 911 segments studied, 892 were assessable on both MDCT andDSA. 312 steno-occlusive segments (73 mildly stenotic, 37 moderately stenotic,60 severely stenotic, and 142 occluded) were detected on DSA. In the subgroupperformed by 4-detector system, the sensitivity, specificity and accuracy of MDCTangiography for detecting segments with more than mild stenosis were 99.3%,98.6%, and 98.2%, respectively, whereas those in the 16-detector system sub-group were 98.9%, 98.0%, and 98.2%, respectively. These did not reveal a signif-icant difference.Conclusion: 4- as well as 16-detector MDCT angiography had high diagnosticaccuracy for evaluating peripheral arterial occlusive disease.

B-236 14:45

Peripheral angiography with 16-row MDCT: Optimization of acquisitionspeed using arterial enhancement quantification dependent on table feedB.C. Meyer, C. Ribbe, K.-J. Wolf, T. Albrecht; Berlin/DE([email protected])

Purpose: To assess the intraarterial enhancement of 16 slice MDCT peripheralangiography dependent on three different table feeds.Methods and Materials: 147 patients with PAOD underwent CTA of peripheralarteries using a Somatom Sensation 16 (Siemens, Germany) with 16 x 1.5 mmcollimation, 0.5s rotation time, 2 mm slice thickness, 1.2 mm reconstruction in-terval. Table feed was varied [30 mm/s, n = 29 (A); 40 mm/s, n = 87 (B) or 48 mm/s, n = 31 (C)]. Twelve patients showed an additional aneurysm of the abdominalaorta or the pelvic arteries. 100 ml Iomeprol 400 (Bracco, Italy) was injected at4 ml/s followed by 50 ml of saline solution. Scans were started 4 s after a thresh-old of 250 ∆HU in the aorta was reached. Arterial enhancement was measured inintervals of 10 cm.Results: The mean arterial enhancement shows a slight increase, reaches amaximum at 30 cm (A: 377 ± 93 ∆HU) 60 cm (B: 370 ± 147 ∆HU) and 70 cm (C:399 ± 99 ∆HU) and decreases slowly down to the distal calf and feet. The en-hancement in the calf arteries is significantly higher when table feed is increasedin the patients without aneurysms (n = 135; A: 162 ± 66 ∆HU, B: 210 ± 90 ∆HU,C: 256 ± 86 ∆HU, p < 0.05). In patients with arterial aneurysms the maximumcontrast enhancement decreases with increased table feed. In this group, thelowest enhancement was achieved with the highest table feed used.Conclusion: The best intraarterial contrast enhancement profile was achievedwhen the highest table feed was used. In this group the arterial enhancement inthe calf was significantly higher as compared to lower table feed settings. In pa-tients with aortic or pelvic aneurysms, table feed should be reduced.

B-237 14:54

Multidetector CT angiography in the assessment of peripheral arterialocclusive disease: Accuracy in detecting the severity, number, and length ofstenosesR.E. Schernthaner1, F. Lomoschitz1, M. Weber1, J. Lammer1, D. Fleischmann2,C. Loewe1; 1Vienna/AT, 2Stanford, CA/US([email protected])

Purpose: To determine the potential of multidetector CT angiography (MD-CTA)in the evaluation of peripheral arterial occlusive disease (pAOD). Digital subtrac-tion angiography (DSA) served as a standard of reference.Methods and Materials: 50 patients referred for DSA of the peripheral arteriesdue to known or suspected pAOD were prospectively included. All patients un-derwent MD-CTA prior to DSA using a 16-row multidetector CT scanner. All CTangiograms were reconstructed using a semi-automated reconstruction algorithmallowing visualization of entire vascular tree.For assessment, every leg was divided into 21 vascular segments, and scored in5 categories of disease severity. Additionally, length and multiplicity of lesionswithin every vascular segment was assessed and compared to DSA.Results: Adequate images could be obtained in all patients. Out of the 50 pa-tients, 1350 vessel segments were delineated by both imaging modalities. Sensi-tivity values in the detection of significant stenoses (over 50%) ranged between0.8 and 1 for different vascular segments. Specificity values ranged between 0.9

Scientific Sessions

188 C D E FBA G

and 1. Exact stenoses graduation using MD-CTA reached high kappa-values (0.79- 1) compared to DSA indicating high agreement between both methods. Even inassessment of lesion length and number, high kappa values (0.8 - 0.9) werereached.Conclusion: Non-invasive MD-CTA has been demonstrated to be a valuable toolfor imaging the pAOD. Extent of disease including number, length and severity ofstenoses was assessed with high accuracy compared to the DSA. We concludethat MD-CTA should play an important role in the diagnosis of pAOD in the future.

B-238 15:03

Grading of internal carotid artery stenosis: Can CTA overcome theconfusion?R.A. Bucek, S. Puchner, M. Haumer, E. Minar, J. Lammer; Vienna/AT([email protected])

Background and Purpose: Carotid endarterectomy and stenting in patients withhigh grade internal carotid artery stenosis (ICAS) significantly reduces the riskfor stroke. Exact grading of ICAS is therefore crucial, however, different gradingmethods as well as divergent results of different modalities often lead to confu-sion.Methods: 69 consecutive patients underwent multi-detector computed tomogra-phy angiography (CTA) and digital subtraction angiography (DSA) within 28 days.The images of both modalities were interpreted by two radiologists blinded to theresults of the other modality. The exact degree of ICAS was calculated for bothmodalities according to NASCET and ECST.Results: The agreement between both grading methods was comparable for CTA(R²=0.87) and DSA (R²=0.84), mean differences between ECST and NASCETwere 13.9% (CTA) and 12.9% (DSA; P> 0.05). Corresponding results for the in-ter-modality correlation were almost equal for NASCET (R²=0.59) and ECST(R²=0.55) with mean differences of 13.4% and 13.5%, respectively (P > 0.05).Sensitivity and specificity of CTA for the detection of occlusions was 100% forboth modalities and grading systems, for the detection of stenoses > 70% 90.9%and 54.9% for NASCET and 94.7% and 46.3% for ECST and of stenoses > 50%95.8% and 59.6% for NASCET and 96.4% and 42.5% for ECST.Conclusion: The introduction of multi-detector CTA cannot overcome the confu-sion in the exact grading of ICAS, because the application of both tested modal-ities as well as both grading methods results in clinical important differences.

B-239 15:12

Anatomic variations in the circle of Willis in patients with symptomaticcarotid artery stenosis assessed with multislice CT angiographyA. Waaijer, B.K. Velthuis, B.H. van der Worp, H. Verhagen, M. Prokop,W.P.T.M. Mali, M.S. van Leeuwen; Utrecht/NL ([email protected])

Purpose: To assess prevalence of anatomical variations in the circle of Willis thatmay hamper collateral supply in patients with symptomatic carotid artery steno-sis (SCAS) and compare this to literature findings in normal brains [%].Materials and Methods: Ninety-one patients with SCAS > 50% underwent CTAon a multislice scanner. Using consensus reading, two observers analyzed theanterior communicating artery (AcomA), A1 and P1 segments in relation to thecontralateral segment, and both posterior communicating arteries (PcomA) werescored separately in relation to the ipsilateral P1 segment.Results: The AcomA was not visualized in 5 patients (6%)[0-4]. The A1 was notvisualized in 14 patients (15%)[0-10], asymmetric (< 2/3 diameter of contralater-al segment) in 37 patients (40%)[4-21] of which 15 (16%) were hypoplastic (≤1 mm)[4-10]. Analyzing both sides (182 cases), the PcomA was not visible in 85 cases(47%), smaller than the P1 in 35 cases (19%), equal (transitional) in 9 cases(5%)[15] and dominant (fetal) in 17 cases (9%) [15-40] with the P1 segment notvisualized in 12 (7%). In 20% of all patients this fetal variant was located ipsilat-erally. In 55% the anterior part of the circle was asymmetric of which 31% washypoplastic or not visible, and in 10% there was also an ipsilateral fetal PcomA,thus isolating the carotid artery from collateral supply by the circle.Conclusion: Multi-slice CTA reveals more anatomical variation in patients withSCAS than described in normal brains and suggests a risk factor for symptomsin the presence of ICA stenosis.

B-240 15:21

Multidetector-row CT angiography in abdominal aortic aneurysm treatedwith endovascular repair: Evaluation of optimal timing of delayed phaseimaging for the detection of low-flow endoleaksR. Iezzi, A.R. Cotroneo, M. Santoro, B. Seccia, F. Di Fabio, A. Filippone,M.L. Storto; Chieti/IT ([email protected])

Purpose: To evaluate the optimal timing of delayed phase imaging for detectinglow-flow endoleaks.Methods and Materials: Thirty-four patients with unruptured abdominal aorticaneurysm treated with endovascular repair (EVAR) underwent follow-up MDCTperformed during unenhanced, arterial, and delayed phase (4x1-mm, a 1.25-mmslice width, and a pitch of 6). Delayed phase imaging, focused on stent-graft, wasperformed with a delay of 60 seconds (early delayed phase enhanced: set A) and300 seconds (late delayed phase enhanced: set B) after intravenous injection of120 ml of iodinated non-ionic contrast medium (Iomeprol 300 mgI/mL, Iomeron;Bracco, Milan, Italy), at a flow rate of 3 ml/s via an antecubital vein. Each set ofdelayed images was interpreted separately by two independent readers for thepresence of endoleaks. Sensitivity, specificity, and diagnostic accuracy of bothreading sessions were compared. The standard of reference was contrast-en-hanced ultrasound, as suggested by recent literature.Results: Sensitivity, specificity, and diagnostic accuracy of set B (late delayedphase enhanced) were significantly higher (P < 0.05) than those obtained withset A (early delayed phase enhanced), for both readers.Conclusion: In patients undergoing MDCT after EVAR of abdominal aortic an-eurysms, delayed scans should be performed 300 seconds after injection of con-trast medium, in order to improve detection of low-flow endoleaks.

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 189C D E FA G

SaSaSaSaSaturturturturturdadadadadayyyyy, Mar, Mar, Mar, Mar, Marccccch 4h 4h 4h 4h 4

B

Scientific Sessions

190 A C D E F G

SS 701bGI Tract

Function andmotility disorders

(p. 210)

SS 701aAbdominal


Focal liverlesions and

transplantationissuses(p. 206)

SS 710Musculo-skeletalShoulder(p. 204)

SS 709Interventional

RadiologySkeletal

interventions(p. 202)

SY 5Satellite

SymposiumInnovations inMR contrast

media imaging(p. 549)

SS 702BreastClinical

issues (1)(p. 208)

SS 711aNeuro

Methods/Techniques/

Software(p. 212)

SS 703Cardiac

Myocardialperfusion and

infarction(p. 214)

R&E Fund Awards

HL 1(p. 45)

SY 6Satellite

SymposiumDigital detector

systems for RAD,mammography,

and angiography

Non-invasive GEsolutions in

cardiac imaging(p. 550)

RC 815Vascular

Vascular imaging(p. 52)


Non traumaticthoracic

emergencies (1)(p. 46)


Osteoporosisand osteopenia

(p. 47)

SF 8aSpecial Focus

SessionStroke(p. 48)

RC 802Breast

Open questions(p. 49)


Anatomicalpathways of

spread ofabdominal

disease(p. 49)

RC 804Chest

Hotline Topic:SARS(p. 50)

SY 7Satellite

SymposiumMorphologicaland functional

myocardialimaging(p. 550)

SF 8bSpecial Focus

SessionTraining inradiology

(p. 51)

NH 6New Horizons

SessionDiffusion tensor

imaging(p. 43)

EM 1"ECR meets"

UK and IrelandHigh resolution

answers to chestproblems/Photons or

protons in the21st century

(p. 44)


Neoplasticdisorders(p. 184)

SY 2Satellite

SymposiumThe advantagesof high relaxivitycontrast mediafor the study

of organvascularization

(p. 547)

SS 601Abdominal

Viscera(Solid Organs)Bile ducts and

pancreaticimaging(p. 186)

SY 3Satellite

SymposiumClinical

advances indiagnosticultrasound

(p. 547)

SS 604Chest

Lung cancer(p. 188)

SS 608Head and NeckValue of 3 Tesla

and othertechnical

advances inhead and neck

imaging(p. 190)

SS 603CardiacCoronary

calcium scoring(p. 192)

RC 511Neuro

Cerebralaneurysms and

AVM’s(p. 38)

RC 514RadiographersNew challenging

techniques(p. 39)


RadiologyRenal


CC 517EmergencyRadiologyTraumaticinjuries (3)

(p. 34)


Inflammatoryjoint disease

(p. 34)

SF 5Special Focus

SessionColon: New

approaches toold problems

(p. 35)

RC 503CardiacClinical

applications:Myocardialperfusion,

viability, andstress testing

(p. 36)


Imaging of thepancreas: Key

questions(p. 37)

RC 504Chest

The chestradiograph:

What do we stillneed to know?

(p. 37)

EP

OS


22:00

registration


room A2nd level

room B2nd level

room C2nd level

room E1entr. level

room E2entr. level

room F1entr. level

room F2entr. level

room Hlower level

room Glower level

08:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:00

14:00

14:30

15:00

15:30

16:00

16:30

17:00

17:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:30

14:00

14:30

15:00

15:30

16:00

16:30

17:00

12:3012:30

13:00

18:00

18:30

19:00

18:30

19:00

18:00

17:30

13:30

08:30

08:00

07:30

07:00

07:30

08:00

07:00

SY 4Satellite

SymposiumUltrasound

elastography forclinical

diagnostic ofbreast cancer

(p. 548)

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 191A C D E F G

WS 824Virtual


"Hands-on"Workshop

WS 624Virtual

InterventionsRenal

"Hands-on"Workshop

WS 23A1Screening

mammographyinterpretation

test"Hands-on"Workshop

WS 23A3Screening



WS 23A4Screening



WS 23A2Screening



E3 720bE-mail for the

radiologist(p. 46)

EFOMPWorkshop

New technologyin diagnostic

radiologySession 2

WS 23B3Screening



WS 23B2Screening



WS 23B1Screening



WS 724Virtual

InterventionsPeripheral

"Hands-on"Workshop

JRFHighlighted

Lectures

SS 711bNeuro

Cerebralperfusion/Blood flow/

Hemodynamics(p. 217)

SS 712Pediatric

Abdominal andfetal imaging

(p. 219)

WS 818Workshops onInterventional

RadiologyBiliary, renal and

urologicalintervention

(p. 52)

SS 713Physics inRadiology

Digitalradiography/

Mammography(p. 221)

SS 707Genitourinary

MR and CTurography(p. 223)

SS 715VascularImaging:Pre- and

postendovasculartherapy(p. 225)

E3 720aRadiologyon the web

(p. 46)

CC 819Pediatric

RadiologyAbdominal pain

(p. 53)


Adult TodayMusculoskeletal

infection(p. 54)

RC 813Physics inRadiology

Special issuesof radiationexposure indiagnosticradiology

(p. 54)

E3 820Foundation


Upper abdominalstructures:

Ultrasound andCT/MR imaging

(p. 55)


"Hands-on"Workshop

WS 524Virtual


"Hands-on"Workshop

E3 620Foundation


Ultrasound of thelower abdomen

and CT/MRimaging of the

bowel(p. 45)

E3 520Interactive

image teachingCervical spine

trauma(p. 43)


RadiologyAbdominal

(p. 40)

CC 519Pediatric

RadiologyHead and neck

and genitalanomalies

(p. 41)


Adult TodayInfection in thegastrointestinal

tract(p. 42)

RC 507Genitourinary

Imagingproblematic

lesions (tumors)(p. 42)

EFOMPWorkshop

New technologyin diagnostic

radiologySession 1

SS 611Neuro

Brain tumors/Perfusion CT,

MR imaging, 1H-MRS

(p. 194)

SS 612Pediatric

Neuro, head andneck

(p. 196)


RadiologyTIPS and liver

chemo-embolization

(p. 198)

SS 607Genitourinary

Adrenal andrenal imaging

(p. 200)

room Klower level

room L/M1st level

room N/O1st level

room Plower level

room Xentr. level

room Y1st level

The Globe2nd level

08:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:00

14:00

14:30

15:00

15:30

16:00

16:30

17:00

17:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:30

14:00

14:30

15:00

15:30

16:00

16:30

17:00

12:3012:30

13:00

18:00

18:30

19:00

18:30

19:00

18:00

17:30

13:30

08:30

08:00

07:30

07:00

07:30

08:00

07:00

room Ilower level

room Wbasement

room Z1st level

Scientific Sessions

192 C D E FBA G

10:30 - 12:00 Room B

Musculoskeletal

SS 610Wrist and handModerators:G. Hadjidekov; Besançon/FRM. Shahabpour; Brussels/BE

B-241 10:30 !Dynamic carpal instability. Part I: The diagnostic use of kinematographyR. Schmitt, G. Coblenz, S. Froehner, G. Christopoulos, B. Pfaff, K.-H. Kalb;Bad Neustadt an der Saale/DE ([email protected])

Purpose: Dynamic stages of carpal instability are manifest only during wristmovement (dyskinematics). The aim of this study was to evaluate kinematogra-phy for detecting the radiographically occult stages of carpal instability.Methods and Materials: Twenty-four patients suffering from tenderness, clickand/or snap phenomena and grip weakness were examined with kinematogra-phy using pulsed fluoroscopy (15 pulses/s). The kinematographic movies werestored on the PACS. Fifteen patients underwent arthroscopy, and in 9 patientssurgical treatment was performed.Results: Standard radiograms were inconspicuous in all cases. Nineteen patho-logical movement patterns were found: 12 scapholunate dissociations (SLD), 2lunotriquetral dissociations (LTD) and 5 midcarpal instabilities (MCI), 2 of thesebilaterally. In the remaining 5 cases, no abnormal movement patterns were seen.There was a close correlation of kinematography to arthroscopic and surgicalfindings: 11/12 of the SLD and 2/2 of the LTD were confirmed, whereas in 1 casethe SL ligament was assessed in arthroscopy to be only incompletely torn. Onepartial tear of the LT ligament was detected in the 5 patients with normal kine-matograms. No information about the ligamentous state was available in the 5patients with MCI patterns.Conclusion: Aside from physical examination manoeuvres, kinematography isthe method of choice for detecting dynamic subcategories of carpal instabilitynon-invasively.

B-242 10:39 !Dynamic carpal instability. Part II: Is contrast-enhanced MR imaging helpfulfor detecting scapholunate or lunotriquetral dissociation?R. Schmitt, S. Froehner, G. Coblenz, K.-H. Kalb, S. Schmitt, G. Christopoulos;Bad Neustadt an der Saale/DE ([email protected])

Purpose: In dynamic stages of carpal instability, there is a normal arrangementof the wrist at rest, although the intrinsic SL or LT ligaments are completely rup-tured. The aim of this study was to evaluate contrast-enhanced MR imaging whendynamic carpal instability is suspected.Methods and Materials: Forty-five patients suspicious of having dynamic pat-terns of SL or LT dissociation were examined with contrast-enhanced MR imag-ing (0.2 mmol/kg) using axial T2*-w GRE, coronal fat-saturated PD-w FSE, plainand contrast-enhanced coronal T1-w SE, contrast-enhanced sagittal T1-w SE.All patients underwent arthroscopy.Results: Focal enhancement areas were found at the SL ligament in 25 patients,with the dorsal and volar segments involved in 16 patients and only one segmentinvolved in 9 patients. Arthroscopy revealed 24 complete tears of the SL liga-ment, but could not confirm 6 lesions (sensitivity 79%, specificity 71%, PPV 76%,NPV 75%, accuracy 76%). In 12 patients, enhancement patterns were seen atthe LT ligament, either dorsovolarly (7 cases) or volarly (5 cases). Arthroscopyrevealed 14 tears of the LT ligament, and did not confirm 4 lesions (sensitivity57%, specificity 87%, PPV 67%, NPV 82%, accuracy 78%). Statistics for MRimaging evaluation were significantly better in the acute/subacute rupture stagecompared to long-standing instability patterns.Conclusion: Contrast-enhanced MR imaging allows moderately good non-inva-sive visualization of dynamic SL and LT by enhancing fibrovascular reparationtissue at the rupture site. This mechanism is less significant in chronic instabilitystages when carpal malarticulation has induced diffuse synovitis.

B-243 10:48

4D dynamic rotational X-ray imaging of the wrist jointB. Carelsen1, G.J. Streekstra1, M. Maas1, S.D. Strackee1, N.H. Bakker1,S.N. Boon2, J. Sabczynski3, M. van Herk1, C.A. Grimbergen1; 1Amsterdam/NL,2Best/NL, 3Hamburg/DE ([email protected])

Purpose: To evaluate wrist functioning, the analysis of dynamic 3D movement ofthe carpal bones in the wrist joint is crucial. Video fluoroscopy and animated 3Dimaging of the wrist joint will not reveal abrupt changes as seen often with dy-namic joint pathology. We created a system for 4-Dimensional rotational X-ray(4D-RX); using a 3D-RX system to image an object that is forced into a cyclicmotion to obtain a series of time resolved 3D reconstructions. The image qualityof our setup is investigated by imaging a post mortem wrist.Methods and Materials: A prototype image intensifier based 3D-RX system (BVPulsera, Philips Medical Systems, Best, NL) is used together with a movementdevice. For 4D-RX, projection images from around a constant cyclic moving ob-ject are taken. The projection images are reconstructed and multiple 3D-RX re-constructions of the moving object at different motion phases are obtained.Results: The best quality was obtained using 41 projection images, obtaining 93D reconstructions. Although both the volume renderings and slice views showartefacts, the anatomical structures of the carpal bones are clearly visible. More-over, a clear representation of the 3D joint kinematics is obtained.Conclusion: A first prototype shows the feasibility of 4D imaging of the wristjoint. We anticipate that dynamic imaging 4D-RX of the wrist and other joints fordiagnostic purposes may be of great help for assessing functional disorders.

B-244 10:57

Peripheral triangular fibrocartilage tears: Depiction with MR arthrographyafter contrast injection into the distal radioulnar jointC. Rüegger1, M.R. Schmid1, C.W.A. Pfirrmann1, L. Nagy1, L.A. Gilula2,M. Zanetti1; 1Zurich/CH, 2St. Louis, MO/US ([email protected])

Purpose: To evaluate the accuracy of MR arthrography with contrast injectioninto the distal radioulnar joint (DRUJ) for depiction of peripheral triangular fibro-cartilage (TFC) tears at the ulnar attachment.Methods and Materials: Forty-one patients (18 female, 23 male, mean age 38years, age range 18-60 years) underwent MR arthrography and wrist arthrosco-py. For MR arthrography iopamidol and gadopentetate dimeglumine was injectedinto the DRUJ. A consensus readout using both MR arthrograms and convention-al arthrograms was performed by two experienced musculoskeletal radiologists.The presence or absence of communicating and non-communicating TFC tearswere recorded. Arthroscopy was used as the standard of reference to determinethe sensitivity, specificity, and accuracy for ulnar attachment tears.Results: A communicating tear at the ulnar attachment was diagnosed in threepatients, a non-communicating tear in 19 patients, and a normal attachment in 19patients. Arthroscopy revealed a tear in all three patients with a communicatingtear, in 14 of 19 patients with a non-communicating tear, and in three of 16 pa-tients with a normal attachment. Thus, the calculated sensitivity was 81% (17/20), specificity was 76% (16/20), and accuracy was 80% (33/41) respectively.Conclusion: MR arthrography with contrast injection into the DRUJ providesadequate accuracy for depiction of peripheral TFC tears at the ulnar attachment.

B-245 11:06

Acute dislocation of the extensor carpi ulnaris tendon in the tennis player:Ultrasonography and MR imaging evaluationJ.-L. Drapé, H. Guerini, A. Feydy, D. Le Viet, B. Montalvan, A. Chevrot;Paris/FR ([email protected])

Purpose: The rupture of the retinaculum of the extensor carpi ulnaris (ECU)tendon is an acute injury, more common because of lift and two hand technique,in professional tennis players. We study the value of US and MR imaging for thediagnosis and follow-up.Methods and Materials: Twenty-one professional tennis players were investi-gated by US and MR imaging after an acute injury of the wrist during a matchplaywith suspicion of a lesion of the ECU tendon or its retinaculum. All patients had adynamic assessment with pronosupination imaging. Imaging follow-up was per-formed over 2 to 3 months.Results: The most common injury combines a medial detachment of the ECUretinaculum and an avulsion pouch along the medial aspect of the ulnar head(n = 17), sometimes extending volarly to the pronotor teres muscle (n = 5). Theradial avulsion is less common (n = 4), with a risk of interposition of the retinac-ulum beneath the tendon and lack of healing. Tendinopathy is common with chroniclesions (n = 17) and tenosynovitis (n = 8) in these professional players. The supi-

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 193C D E FA G

nation maneuver is essential to highlight a dislocation of the ECU tendon (n = 15)which may be minor. Severe tendinopathies demonstrate a strong tendon de-formity during the stress. Follow-up imaging shows the decrease of enhance-ment and volume of the avulsion pouch.Conclusion: Diagnosis and healing of acute injuries of the ECU retinaculummay be assessed with US and MR imaging.

B-246 11:15

MR imaging of recurrent carpal tunnel syndrome: Comparison with EMGand a controlled groupR. Campagna, J.-L. Drapé, A. Feydy, P. Corlobé, D. Le Viet, D. Godefroy,A. Chevrot; Paris/FR ([email protected])

Purpose: The aim of the study is to evaluate the potential MR imaging findings ofpainful postoperative carpal tunnel syndromes (CTS) compared with electromy-ography (EMG).Methods and Materials: At 1.5 T, 34 wrists in 34 patients with previous CTSrelease were evaluated by two observers for median nerve (signal, morphologyand enhancement), for flexor retinacular regrowth, for fibrosis, for volar nervemigration and tunnel release measure, and for compressive abnormalities. Thepopulation is constituted of 11 controlled patients with normal electromyography(EMG), and 23 with electric abnormalities. There were 27 female and 7 malepatients, mean age 55.7 years. Mean time between surgery and MR imaging was29.9 months, and 3.8 months between EMG and MR imaging.Results: Retinacular section was seen in 12/23 with and 5/11 without recurrentCTS (p > 0.05). High T2 signal of the median nerve was seen in 20/21 with, and7/11 without recurrent CTS (p = 0.036), enhancement was seen in 10/23 with,and 1/11 without recurrent CTS (p = 0.06). Mean ratio was 2.40 with and 1.89without recurrent CTS (p = 0.018). Compressive abnormalities were present in 7/23 patients with, and in 5/11 without recurrent CTS (p > 0.05). Fibrosis was seenin 14/23 with, and 2/11 without recurrent CTS (p = 0.029).Conclusion: Fibrosis, proximal flattening and high signal of the median nerveare significant findings in recurrent CTS with electric abnormalities compared toa controlled population.

B-247 11:24

Sonographic diagnosis of carpal tunnel syndrome: Diagnostic value of thetriangular cross-section signK.T. Szopinski, T. Mazurczak-Pluta, B. Lakomiec, R.Z. Slapa; Warszawa/PL([email protected])

Purpose: In some patients with carpal tunnel syndrome, the median nervesqueezed between the flexor retinaculum and flexor tendons takes a triangularshape. The purpose of this study was to determine the value of the triangularcross-section sign in the diagnosis of carpal tunnel syndrome.Methods and Materials: In 50 patients with hand pain (41 females and 9 males),82 sonographic examinations of the median nerve were performed with an ATLHDI 5000 scanner, using a linear 7.5 MHz probe. All patients underwent an elec-tromyographic (EMG) examination. Patients with bifid median nerve were exclud-ed from this study.Results: Triangular shape of the median nerve in the carpal canal was observedin 7/63 (11%) cases of EMG-positive diagnosis of carpal canal syndrome. Nosuch deformation was seen in patients with normal results of the EMG examina-tion. Mean cross-sectional area of deformed median nerves in the carpal canalwas 17.7 mm2 (range 12-25 mm2), mean cross-sectional area of EMG-positivemedian nerves was 15 mm2 (range 6-30 mm2), mean cross-sectional area ofmedian nerves with normal EMG results was 10.1 mm2 (range 6-17 mm2). Sensi-tivity, specificity, positive predictive value (PPV) and negative predictive (NPV)value of the triangular cross-section sign in diagnosis of carpal tunnel syndromeas compared with the electromyography were 0.11, 1.0, 1.0 and 0.25 respective-ly.Conclusion: Triangular shape of the cross-section of the median nerve in thecarpal canal is a highly specific sign of carpal canal syndrome.

B-248 11:33

The role of US in affecting the management of patients referred to a handsurgeonG.M. Allen, M. Tan, M. Dhillon, V. Rajaratnam; Birmingham/UK([email protected])

Methods and Materials: This was a prospective study of all patients referred forUS of the hand, wrist and forearm conditions (including the elbow) in a hand unit.Referrals were made by three hand surgeons. The US was performed by one

dedicated musculoskeletal radiologist. The clinical diagnosis was recorded fromthe request card and clinic notes. The ultrasound findings were recorded and itwas noted whether the US made a difference to the decision in further manage-ment. The findings were compared with the final histological diagnosis for thosepatients who went to surgery.Results: A total of 31 patients are included in the study. The majority (24 pa-tients, 77%) of these were referred to US with a single proposed diagnosis. Ofthese cases US confirmed the diagnosis in 15 patients (63%) but in the remain-ing 9 patients (37%) other diagnoses were made. Of the remaining 7 patients, 6were referred with more than one clinical diagnosis and a single patient had noprovisional clinical diagnosis. US agreed with one of the possible differentials inone third, and excluded several of the possibilities in patients who had previouslyundergone complex hand surgery in another third. In the final third of patientswithin this subgroup the clinical and US diagnoses were at variance.Conclusion: Ultrasound is a very useful adjunct to clinical assessment, alteringpreoperative diagnoses in a significant number of patients potentially saving thepatient unnecessary surgery.

B-249 11:42

Value of pulse-inversion harmonic imaging of inflammation activity inrheumatoid arthritisC. Schueller-Weidekamm, G. Schueller, C. Krestan, F. Kainberger; Vienna/AT([email protected])

Purpose: To evaluate the value of contrast enhanced pulse-inversion harmonicimaging (PI-HI) on both the synovial vascularization and the therapeutic effectson the inflammation of finger joints in rheumatoid arthritis (RA).Methods and Materials: Two hundred and eighty finger joints (MCP I-V, IP, PIPII-V) of 14 patients (10 females, 4 males) with RA were investigated with PowerDoppler and PI-HI ultrasound after i.v. application of a second generation ultra-sound contrast media (SonoVue, Bracco, Milan, Italy). With power Doppler, onefinger joint per patient representing the strongest hypervascularization beforesteroid therapy was determined. Dynamic examination with PI-HI of the selectedjoint was carried out before and after 7 days of middle-high dosed steroid thera-py. Vascularization was quantified by calculating the area underlying time-inten-sity curves [dB]. The changes of the signal intensities before and after therapywere correlated with clinical examinations (DAS-score).Results: All patients showed a statistically significant reduction of PI-HI signalsafter steroid therapy (p < 0.05). The baseline and follow-up median values of thearea underlying time-intensity curves were 8.56 ± 1.28 and 7.65 ± 0.66 respec-tively. The median values of the DAS decreased significantly from 4.9 ± 0.86 to3.6 ± 1.0 (p < 0.01) 7 days after the steroid therapy.Conclusion: PI-HI enables the detection of synovial perfusion alterations aftersteroid therapy, and therefore may be an additional useful tool in the evaluation ofactive inflammation in RA and in the assessment of the therapeutic response.

B-250 11:51

Rheumatoid arthritis at 3 Tesla MR imaging: The proximal interphalangealjoints (PIPs), using a high-resolution microscopy coilI.-M. Noebauer-Huhmann, E. Jimenez-Boj, J. Zwerina, F. Kainberger,G. Schett, M. Bittsansky, H. Imhof, J. Smolen, S. Trattnig; Vienna/AT([email protected])

Purpose: To assess the PIP alterations in rheumatoid arthritis of the finger usinga high-resolution microscopy coil at 3 Tesla.Methods and Materials: High-resolution 3 Tesla MR imaging was performed in15 patients with rheumatoid arthritis, using a microimaging gradient insert withgradients up to 200 mT/m. Coronal STIR (TR/TE 1800/14.6, TI 170 ms, 4 averag-es, FOV 70x35 mm, matrix 256x96, slice thickness 2 mm), axial T1w (TR/TE538/15.3 ms, 10 averages, FOV 30x30 mm, matrix 256x192, slice thickness 1 mm),and contrast enhanced, fat-suppressed coronal and axial T1w sequences (TR/TE 808/15.4, 4 averages, FOV 70x35 mm, matrix 512x148, slice thickness 2 mm)were performed. The images were evaluated according to modified EULAR-OMERACT criteria.Results: Synovial swelling and enhancement, bone marrow edema, erosions,and the involvement of collateral ligaments, tendons and tendon sheaths wereclearly visualized. Interestingly, the most common findings were not only synovi-al swelling (15/15) and synovial contrast enhancement (14/15), but also thicken-ing and enhancement of the collateral ligaments (14/15). In all patients with bonemarrow edema (9 /15), the insertion zone of the collateral ligaments was involved.The erosions which were seen in 9/15 patients were exclusively located in theinsertion zone of the collateral ligaments in 3/9 patients, partly in 5/9 patients,and distant to this zone in only 1/9 patient.

Scientific Sessions

194 C D E FBA G

Conclusion: High-resolution 3 Tesla MR imaging of PIPs in rheumatoid arthritisallows detailed information about more anatomic structures involved than hasbeen described previously. The collateral ligaments and the bone of the insertionzone of the collateral ligaments are commonly involved.

10:30 - 12:00 Room C


SS 601aHepatocellular carcinomaModerators:E. Gyömbér; Zalaegerszeg/HUL.H. Ros Mendoza; Zaragoza/ES

B-251 10:30

Usefulness of SonoVue-enhanced pulse-inversion ultrasonography toassess hepatocellular carcinoma response after percutaneousradiofrequency thermal ablation therapyT.V. Bartolotta, M. Midiri, M. Galia, G. Runza, M. Bellia, A. Taibbi, R. Lagalla;Palermo/IT ([email protected])

Purpose: To evaluate therapeutic response of percutaneous radio-frequency ther-mal ablation (RFTA) of hepatocellular carcinoma (HCC) by SonoVue-enhancedpulse-inversion ultrasonography (CEUS) in comparison with multislice computedtomography (MSCT).Methods and Materials: 280 patients with 337 nodules (mean size 3.2 cm) diag-nosed as HCC by biopsy and treated with RFTA were evaluated by CEUS andcontrast-enhanced multiphase MSCT one month and three months after treat-ment respectively. On MSCT or CEUS we considered: a) complete ablation- whenthere was no enhancing portion within or at the margin of the ablation zone dur-ing the hepatic arterial phase; b) residual unablated tumor- an irregular peripher-al-enhancing focus in the ablation zone after one month; c) local tumor progression-the presence of growing enhancing tumors at the margin after three months butwith no evidence of residual tumor in the previous control.Results: CEUS and MSCT revealed residual tumor in 67 (19.8%) and 72 (21.3%)nodules respectively, local tumor progression in 18 (5.3 %) and 20 (5.9 %) re-spectively and complete response in 252 (74.8%) and 255 (75.7%) respectively.Local tumor progression was confirmed by US-guided biopsies of the areas ofgrowing enhancing tumors in all the 18 nodules assessed by CEUS.Conclusion: SonoVue-CEUS can be considered a reliable alternative techniqueto contrast-enhanced multiphase MSCT to assess therapeutic response afterRFTA of HCC.

B-252 10:39

Association between a wound-response gene expression program andhypodense tumor halos on CT in hepatocellular carcinomaM. Kuo1, J. Gollub2, C. Sirlin1, C. Ooi3; 1San Diego, CA/US, 2Stanford, CA/US,3Hong Kong/HK ([email protected])

Purpose: Elements of cancer have been compared to a wound healing responsegone awry. A canonical "wound-response" gene expression program has beenpreviously identified. We sought to determine whether hypodense tumor halosseen on CT were associated with the wound response gene expression programin HCC.Methods and Materials: 29 HCCs were analyzed across 29 CT scans for thepresence or absence of hypodense tumor halos by 2 radiologists in consensus.Tumor halo scores for each tumor were then evaluated for enrichment of thewound-response gene expression program against matched microarray geneexpression measurements across 3960 clones for the same 29 HCCs using su-pervised bioinformatic analyses. The strength of the enrichment for the woundresponse gene expression program was then determined and statistical signifi-cance evaluated.Results: Hypodense halos were present in 12/29 HCCs and absent in 17/29.Hypodense tumor halos were strongly associated with regulation of the wound-response gene expression program (p < 0.05). Further, HCC with absent haloswere associated with up-regulation of the wound-response gene expression pro-gram whereas presence of halos correlated with down-regulation. Additional anal-ysis also demonstrated that absence of hypodense tumor halos was also correlatedwith regulation of a p53-associated HCC gene expression program.Conclusion: Hypodense tumor halos may represent an imaging surrogate for a"wound-response" tumor transcriptional program in HCC and may also incorpo-rate elements of a p53 gene expression program.

B-253 10:48

MDCT in hypervascular liver tumors: Optimization of iodine concentrationR. Hammerstingl1, J.P. Valette2, D. Regent3, T. Vogl1; 1Frankfurt a. Main/DE,2Lyon/FR, 3Vandoeuvre/FR ([email protected])

Purpose: To optimise iodine concentration in Multidetector CT (MDCT) of hyper-vascular liver tumors.Method and Materials: In this multicenter, double blind, randomised, parallelgroup comparison clinical Phase IV study 60 patients with the diagnosis of he-patic tumors were included. All patients underwent MDCT using bi-phasic con-trast-enhanced technique. A late arterial and portal venous phase was performed.Iodinated contrast medium (Iomeprol) was administered intravenously in threedifferent concentrations (300, 350, 400 mg iodine/ml). The overall iodine dose(36 g) was equal within the groups. The absolute contrast between liver tumorsand surrounding tissue as well as contrast densities were analysed by summarystatistics. Moreover an exploratory Kruskal-Wallis test was performed.Results: Contrast density of normal liver parenchyma was 61.3, 64.8, 72.7 HUfor Iomeprol 300, 350, and 400 (p = 0.01) in the arterial phase. The highest con-centration led to higher densities for the enhancement of normal liver tissue. Forthe maximum of absolute contrast between overall lesions and lesion surround-ing tissue Iomeprol 400 group presented with highest densities (73.5 HU), fol-lowed by 350 (63.5 HU) and 300 (60 HU) subgroups, although this differencewas not statistically significant [exploratory Kruskal-Wallis test (p = 0.417)]. Con-cerning the maximum of absolute contrast a tendency to far higher contrasts ofIomeprol 400 versus 300 and 350 for hypervascularized tumors was document-ed: hepatocellular carcinoma (71, 48, and 45) and cholangiocarcinoma (84, 65,and 64 HU).Conclusion: High concentration contrast medium shows advantages in demar-cation and delineation compared to lower iodine-concentrated contrast agentsfor the diagnosis of hypervascularized liver tumors.

B-254 10:57

Three-phase helical CT findings of hepatocellular carcinoma in cirrhoticand non-cirrhotic liver of Asian patientsW. Lee, E. Lee, S. Kim, S. Kim, M. Kim, H.K. Lim, C. Park; Seoul/KR([email protected])

Purpose: To compare three-phase helical CT findings of hepatocellular carcino-ma (HCC) in cirrhotic and non-cirrhotic livers of Korean patients, and to find outthe difference of CT findings between noncirrhotic livers of Asian and non-Asianpatients.Methods and Materials: During a 5-year period, 191 patients (165 men and 26women, mean age; 52 years) who had three-phase helical CT were collectedamong 527 patients with surgically proven HCC. At pathology, 89 patients hadcirrhotic liver (88 HBV infection, 1 HCV infection) and 102 patients had noncir-rhotic liver (68 HBV infection, 5 HCV infection, 4 unknown chronic hepatitis, 25normal liver). All CT images were analyzed retrospectively for various morpho-logic and enhancement patterns of tumor, tumor thrombosis and intrahepaticmetastasis, and were correlated with pathologic findings.Results: More frequent morphologic patterns in noncirrhotic livers were ill-de-fined margins (p=.001), fibrosis (p <.001), necrosis (p <.001), and hemorrhage(p=.201). More frequently found patterns were capsule (p <.001), fatty metamor-phosis (p=.127), and mosaic pattern (p=.118) in cirrhotic livers. For enhance-ment pattern, arterial enhancement and delayed washout were more frequentlyseen in non-cirrhotic liver without statistical difference. Intrahepatic metastasisand tumor thrombosis were more frequently found in non-cirrhotic liver (p > .05).Conclusion: Some morphologic findings of HCC in cirrhotic liver were founddifferently in non-cirrhotic livers. Reviewing the literature, our results of patientswith non-cirrhotic livers were sufficiently similar to those of non-Asian patients.Therefore, we suggest that morphologic appearances of HCC are not dependenton the racial differences but on the presence of cirrhotic change.

B-255 11:06

Pharmaco CT offers improved sensitivity in detection of illusive lesions ofhepatocellular carcinomaE.K. Lang, M. Brammer, D. Slakey, L. Myers; New Orleans, LA/US([email protected])

Purpose: Assessment of improved detection of illusive lesions of hepatocellularcarcinoma by pharmaco CT (adrenaline modulated) on management and con-sideration for transplant surgery.Methods and Materials: 13 patients, 8 male, 5 female aged 37 to 58 with apresumptive diagnosis of hepatocellular carcinoma were entered. Routine 4 phase

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 195C D E FA G

spiral CTs obtained in all patients for initial staging were followed by a pharmaco-angio-CT. Pharmaco-angio-CT technique: 0.05 mg of adrenaline hydrochloridediluted in 10 ml of normal saline are injected through a 3 French catheter into thecommon hepatic artery, four seconds prior to injection of a contrast bolus (4 ml ofnon-ionic contrast medium in 20 ml of normal saline at a flow rate of 3-4 ml /sec).Multi-detector spiral CTs were obtained with a four second delay. A second CTseries was generated in the portovenous phase.Results: In 5 of 13 patients 2 to 7 additional lesions were demonstrated on thearterial phase adrenaline modulated CT. Lesions tended to fade in the portov-enous phase. Follow-up pharmaco-CTs after chemoembolization or RF ablationshowed 1 to11 additional lesions in 4 of 9 patients. Areas of necrosis in treatedlesions were better demonstrated on the porto-venous phase CT.Conclusion: The pharmaco-angio-CT showed additional lesions in 3 sectors,one in the left liver lobe, mandating modification of treatment options in three andeliminating one from the liver transplant list. The additional lesions demonstratedon follow-up pharmaco-angio-CTs called for further modification of treatmentoptions in 3 of these 9 patients.

B-256 11:15

Hepatocellular carcinoma in patients undergoing living-donor livertransplantation. Accuracy of multidetector computed tomography byviewing images on digital monitorsY. Maetani, H. Isoda, T. Koyama, K. Togashi; Kyoto/JP([email protected])

Purpose: To assess multi-detector row helical computed tomography (MDCT)findings with histopathology in patients with hepatocellular carcinoma (HCC) whohad undergone living-donor liver transplantation (LDLT).Methods and Materials: This prospective study comprised 41 patients (28 men,13 women, age 30-68 years, mean 55 years) with HCC admitted for LDLT. Re-sults of preoperative MDCT were correlated with histopathologic results after 5-mm-slicing of the explanted liver. After 2 mL/kg of 300 mg I/mL of contrast mediumwere injected, late arterial and equilibrium phase images were obtained. MDCTwas performed using following parameters: detector row configuration, 2 x 8 mm,reconstruction interval;1 mm, 200 mAs; and 120 kVp. Image analysis was per-formed in consensus by three abdominal radiologists using 2 mm thickness with1 mm overlap images with paging method as well as multi-planner reformat im-ages (i.e. axial, coronal and sagittal images).Results: A total of 134 HCC lesions (mean diameter: 21 mm) were confirmedpathologically. The sensitivity, specificity and positive predictive value for the de-tection of HCC was 87%, 96% and 84%, respectively for all lesions; 98%, 99%and 97%, respectively for lesions larger than 1 cm; and 46%, 76%, 41%, respec-tively for tumors smaller than 1 cm (n = 28). There were 12 false negative lesionswhereas only five false positive findings. Four tiny HCCs were misdiagnosed asarterial-portal shunts.Conclusion: Multi-detector row CT by viewing images on digital monitors is veryeffective in detecting HCCs larger than 1 cm with only a few false-positive le-sions.

B-257 11:24

Detection of hepatocellular carcinoma (HCC): Comparison of Gd-BOPTA-enhanced dynamic MR imaging and dynamic multislice CT (MSCT) inpatients with cirrhosisR. Iannaccone, C. Catalano, A. Laghi, D. Marin, M. Di Martino, A. Guerrisi,M. Baski, G. De Filippis, R. Passariello; Rome/IT ([email protected])

Purpose: To compare Gd-BOPTA-enhanced MR imaging with MSCT for the de-tection of HCC in patients with cirrhosis.Methods and Materials: 52 patients with 94 HCCs underwent MR imaging on a1.5 T scanner (Magnetom Vision Plus, Siemens) and CT on a multislice scanner(Somatom Plus 4 Volume Zoom, Siemens). For MR imaging, unenhanced T2-weighted HASTE and T1-weighted FLASH sequences were acquired. After intra-venous injection of 20 mL of Gd-BOPTA (0.1 mmol/kg; MultiHance, Bracco),arterial, portal-venous, and equilibrium phase T1-weighted images were obtainedat 25, 60, and 180 sec, respectively. A hepatospecific phase was obtained 1 hourafter the injection of contrast material. For MSCT, unenhanced, arterial, portal-venous, and equilibrium phase images were obtained before and 35, 70, and 180sec, respectively, after intravenous injection of 120-150 mL of nonionic contrastmaterial (300 mgI/mL). CT imaging parameters were: collimation, 4 x 2.5 mm,section thickness, 3.0 mm; effective mAs, 160; and kVp 120. Three observersindependently interpreted the images obtained with each technique. The alterna-tive-free response receiver operating characteristic (A (z)) method was used toanalyze the results. Sensitivity and positive predictive value were also calculated.

Results: Mean sensitivities of MR imaging and CT were 84% and 85.1%, re-spectively (p not significant). Mean positive predictive values of MR imaging andCT were 86.2% and 87%, respectively (p not significant). There was also nosignificant difference in A (z) values between MR imaging (0.81) and CT imaging(0.78) (p not significant).Conclusion: Gd-BOPTA-enhanced MR imaging and MSCT showed similar di-agnostic accuracy for the detection of HCCs in patients with cirrhosis.

B-258 11:33

Gd-EOB-DTPA enhanced MR imaging versus triphasic CT: Detection andclassification of focal liver lesionsC. Kölblinger, A. Ba-Ssalamah, F. Längle, T. Grünberger, F. Wrba, W. Schima;Vienna/AT ([email protected])

Purpose: To compare the performance of triphasic CT and Gd-EOB-DTPA (Pri-movist, Schering, Berlin, Germany) enhanced MR imaging in detecting and clas-sifying focal liver lesions.Methods and Materials: CT and Primovist enhanced MR imaging scans of 29patients, recruited from phase II and III studies, with focal liver lesions (a total of106 lesions) were evaluated. The gold standard was either defined by histopatho-logic correlation, intraoperative ultrasound or imaging follow-up. Images wereobtained by using triphasic CT and dynamic Gd-EOB-DTPA enhanced MR scans(1.5 Tesla). Additionally late phase MR scans 20 and 40 minutes after contrastmedia injection were acquired. Two radiologists, blinded to the gold standard,evaluated the images in consensus reading with respect to size, number, loca-tion and morphological characteristics of the detected lesions.Results: From a total of 106 lesions 68% could be revealed with CT and 89%with MR imaging, respectively. The correct diagnosis of the detected lesions wasaccomplished in 74% with CT and in 94% with MR imaging.Conclusion: MR imaging with Gd-EOB-DTPA markedly improves detection andclassification of focal liver lesions compared to triphasic CT.

B-259 11:42

Gd-EOB-DTPA enhanced MR imaging of hepatocellular carcinoma: Aredelayed enhanced images really useful for their detection?R. Cianci, A. Filippone, E. Grassedonio, F. Di Fabio, A. Gorgoglione,M.L. Storto; Chieti/IT ([email protected])

Purpose: To compare the diagnostic accuracy of Gd-EOB-DTPA enhanced dy-namic MR imaging and Gd-EOB-DTPA enhanced delayed MR imaging for thedetection of hepatocellular carcinoma (HCC).Methods and Materials: Thirty-three patients with known HCC underwent MRstudy on 1.5 T-scanner, before and after the administration of Gd-EOB-DTPA.Enhanced examination included the dynamic study (arterial, portal and equilibri-um phases) and the delayed study performed 20 minutes after contrast injection.Three radiologists independently interpreted the images for detection of HCCs,according to a four point confidence scale. The diagnostic accuracy was evaluat-ed using the receiver operating characteristic method and the sensitivity of eachimaging technique was compared.Results: The mean diagnostic accuracy of dynamic MR imaging was higher thanthat of delayed MR imaging (Az = 0.94 vs Az = 0.89), although the differencewas not statistically significant. The mean sensitivity of dynamic MR imaging wassignificantly higher (p < 0.05) than that of delayed MR imaging.Conclusion: Gd-EOB-DTPA enhanced dynamic study seems to be more accu-rate than Gd-EOB-DTPA enhanced delayed study in the detection of HCC.

B-260 11:51

Small focal hypervascular lesions seen in the arterial phase of Gd-enhanced MR imaging in cirrhotic patients: Differentiation of benign frommalignant based on pre-Gd SPIO-enhanced signal characteristicsK.S. Mangat, D. Aguirre, C. Sirlin; San Diego, CA/US([email protected])

Purpose: Unidentified bright objects (UBO) are < 2 cm, circumscribed foci ofarterial enhancement seen frequently on Gadolinium (Gd)-enhanced magneticresonance (MR) images of cirrhotic livers. We tested the hypothesis that we candifferentiate benign (BUBO) from malignant (MUBO) lesions based on super par-amagnetic iron oxides (SPIO) uptake.Methods and Materials: We retrospectively identified 50 UBO (34 MUBO, 16BUBO) with a reliable truth standard. All studies were acquired at 1.5 Tesla andincluded SPIO-enhanced spoiled gradient echoes (SGE) with time to echo (TE)of 2.65, 4.75 and 6.6 msec and T2w echo-train spin echo (ETSE) images fol-lowed by dynamic Gd-enhanced images. Three radiologists (R1, R2, R3), blind-

Scientific Sessions

196 C D E FBA G

ed to the truth standard reviewed the studies and UBO identified on dynamicimages were graded on a 6-point ordinal scale for SPIO uptake. Receiver operat-ing characteristics were assessed for differentiating BUBO from MUBO.Results: Depending on the sequence, area under curve (AUC) ranged from 0.67to 0.97 (R1); 0.62 to 0.85 (R2); and 0.64 to 0.88 (R3). Depending on the readerAUC for SGE ranged from 0.62 to 0.67 (TE=2.65); 0.83 to 0.88 (TE=4.76); and0.85 to 0.97 (TE=6.6). AUC for T2w ETSE ranged from 0.79 to 0.80. SGE withTE=6.6 msec provided the best performance. Using this sequence sensitivity forR1, R2 and R3 were 97%, 91%, 97% and specificity was 94%, 81% and 75%,respectively.Conclusion: Signal characteristics on SPIO-enhanced pre-Gd images help dif-ferentiate benign from malignant UBO in cirrhotic patients.

10:30 - 12:00 Room E2

GI Tract

SS 601bCT colonographyModerators:A.H. Freeman; Cambridge/UKC. Kulinna-Cosentini; Vienna/AT

B-261 10:30

The usefulness of CT colonography with fluid tagging in frail patients and inelderly patients with incomplete conventional colonoscopyF. Iafrate1, P. Paolantonio1, M. Rengo2, M. Anzidei2, M. Celestre2, A. Laghi1;1Latina/IT, 2Rome/IT ([email protected])

Purpose: To evaluate the value of CT Colonography in elderly patients and infrail patients in whom conventional colonoscopy (CC) could not be completed.Method and Materials: 68 patients (Age range:70-92) with incomplete Conven-tional Colonoscopy and 4 frail patients underwent CTC examination using a 64MDCT (LightSpeed VCT, GE) in order to study the remaining colon. Patient prep-aration was obtained drinking 200 ml of Diatrizoate Meglumine (Gastrografin,Schering) within two days prior to CTC. Colonic iodinated contrast was i.v. ad-ministered in 28 symptomatic patients. No antiperistaltic drug was administered.The amount of residual fluid was calculated on a three point grade scale as wellas patient acceptance using questionnaires.Results: CTC succeeded in 66/68 (97%) patients with incomplete colonoscopy.CC failed in assessment of 246/408 potentially visible colonic segments. StopPoint on CC examination was predominantly in the sigmoid. Additional lesionsidentified in proximal colon in 19 % of patients, include3 primary carcinoma, 12polyps in 8 patients, 2 lipomas. In one patient with incomplete colonoscopy atsigmoid a mesenterial liposarcoma was found. The main causes of incompleteCC was; redundant colon, tortuous colon loops and occlusive mass. Level oftolerance was rated good as well as patient compliance and patient preparation.Mean duration time of CTC was 9 minutes.Conclusion: CTC is feasible and effective technique in evaluating entire colorec-tal segments of the colon not visualized during colonoscopy. CTC has an undeni-able role in elderly and frail patients in whom conventional colonoscopy is difficultto be fully performed.

B-262 10:39

Multislice CT colonography (MSCTC) without cathartic preparation:Assessment of diagnostic accuracy in 180 asymptomatic average-riskpatientsR. Iannaccone, C. Catalano, D. Marin, A. Guerrisi, M. Di Martino, M. Baski,G. De Filippis, R. Passariello; Rome/IT ([email protected])

Purpose: To prospectively compare the diagnostic performance of MSCTC with-out cathartic preparation with that of colonoscopy for the detection of colorectalpolyps in asymptomatic average-risk patients.Methods and Materials: 180 asymptomatic patients referred to undergo screen-ing colonoscopy were recruited. For fecal tagging, an oral iodinated contrast agent(diatrizoate meglumine and diatrizoate sodium) was administered beginning 48hours prior to CT. No cathartic preparation was used prior to CT. All CT examina-tions were acquired using a MSCT scanner (Somatom Plus 4 Volume Zoom,Siemens) using collimation, 4 x 2.5 mm; section thickness, 3.0 mm; effective mAs,10; and kVp, 140. Colonoscopy with segmental unblinding was performed 3-7days after CT as reference standard. Three radiologists, unaware of colonoscop-ic findings, analyzed CT colonographic images using a dedicated software (Vit-

rea, Vital Images). Sensitivity was calculated on both a per-polyp and a per-pa-tient basis. In the latter, specificity was also calculated. Interobserver agreementwas calculated using the k statistic.Results: Colonoscopy detected two carcinomas and 76 polyps in 42 patients;the remaining 138 patients showed no abnormality. CT colonography yielded anaverage per-polyp sensitivity of 48.6% (92% for polyps ≥ 8 mm). Average per-patient sensitivity and specificity were equal to 92.8% and 94.2%, respectively.Interobserver agreement was high on a per-polyp basis, and high to excellent ona per-patient basis.Conclusion: In asymptomatic patients at average-risk for colorectal cancer,MSCTC without cathartic preparation has a diagnostic performance similar tothat of colonoscopy for the detection of colorectal polyps ≥ 8 mm.

B-263 10:48

Automatic unseen colon surface detection in 3D MDCT colonographyA. Graser1, C.R. Becker1, L. Guendel2, M.F. Reiser1; 1Munich/DE,2Forchheim/DE ([email protected])

Purpose: In primary 3D read of CT colonography (CTC), visualization of theentire colonic wall remains a problem. Even after fly-through from rectum to ce-cum and vice versa not all of the colon might have been visualized. We evaluatedthe percentage of the unseen surface after forward and backward fly-through andperformance of a computerized tool that automatically displays unseen areas tothe reader.Methods and Materials: 40 data sets with different patient positions (20 prone,20 supine) were randomly selected from a database and investigated on a CTCworkstation (syngo Colonography, Siemens). The percentage of unseen surfacewas noted after one single flight from rectum to cecum and after an additionalflight from rectum to cecum. Direct navigation to unseen areas was performed inall cases.Results: After one single fly-through from rectum to cecum, only 71.7-88.2%(mean, 81.3%) of the colonic surface had been visualized. A 2nd flight increasedthese values to 91.4-98.8% (mean, 96.2%). The "unseen areas" tool implement-ed in the CTC software sorted unseen areas according to their size and dis-played them to the reader.Conclusion: 3D reading with 2D for problem solving leads to an incomplete viewon the colon surface even if bidirectional fly-through is performed. A tool whichmeasures the percentage of colon surface that has been seen by the radiologistand automatically displays unseen areas prevents perception errors and may cutreading time.

B-264 10:57

CT colonography: Automated diameter measurement of colonic polypscompared with a manual technique: In vitro study using a human colectomyspecimenS.A. Taylor1, A. Slater2, L. Honeyfield1, M. Roddie1, S. Halligan1, J. Dehmeshki1;1London/UK, 2Oxford/UK ([email protected])

Purpose: To investigate the accuracy of automated and manual CT polyp meas-urement using a colectomy phantom.Methods and Materials: A colectomy specimen from a patient with familial ade-nomatous polyposis containing 27 polyps was insufflated, submerged, andscanned using a MDCT. A histopathologist measured the maximum dimension ofall polyps in the opened specimen. Digital photographs and line drawings wereproduced to document exact polyp location. Two observers (novice and expert)independently estimated polyp diameter using three methods: software calipers2D ('manual 2D'), software calipers 3D ('manual 3D') and automated polyp meas-urement software (automatic) [Medicolon 1.3, Medicsight plc]. The software re-quires the placement of 2 seed points in the polyp vicinity prior to automatedsegmentation and measurement. Data was analyzed using a paired t-test andBland-Altman limits of agreement.Results: 6 polyps (all flat and ≤ 6 mm) could not be extracted by the softwareand were excluded, leaving 21 polyps (4-15 mm) for analysis.Automated polyp measurement was not significantly different from true histolog-ical size for either reader. However both observers significantly underestimatedpolyp size using manual 2D (mean difference [md] -1.3 mm, sd 1.3, p = 0.001,and md -1.6 mm, sd 1.2, p < 0.001) and using manual 3D for the expert only (md-0.6 mm, sd 1.4, p = 0.04). Bland Altman limits of 95% limits of agreement weresimilar for all three methods (automated -3 mm to 3.5 mm, manual 2D -3.8 mm to1 mm, manual 3D (-3.2 mm to 2 mm).Conclusion: Automated polyp measurement is technically feasible, does not re-quire reader expertise, and is potentially more accurate than manual measure-ment.

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 197C D E FA G

B-265 11:06

CT colonography in the follow-up of patients with previous polypectomyE. Neri, P. Vagli, F. Vannozzi, A. Bardine, C. Bartolozzi; Pisa/IT([email protected])

Purpose: To evaluate the role of CT colonography (CTC) in the follow-up of asymp-tomatic patients that underwent polypectomy.Methods and Materials: Forty patients that underwent a previous polypectomy(mean follow-up at 1 year), were studied with CTC at low dose, with no contrastand a standard preparation (no feacal tagging); no muscle relaxants were used.Examination was performed in supine and prone decubitus. Patients were askedto undergo colonoscopy after CTC.Results: CTC found 34 polyps in 17 patients, that were all submitted to colonos-copy; in the remaining 23 patients (negative for CTC findings), only 9 underwentcolonoscopy that confirmed CTC results (no FP or FN results). The per polypssensitivity was 71% for those with diameter < 6 mm (6 FN) and 92% for largerpolyps (1 FN). The per polyps specificity was 65% for those with diameter < 6 mm(7 FP) and 91% for larger polyps (2 FP).Conclusion: CTC represents a valid alternative to colonoscopy in the follow-upof asymptomatic patients with previous polypectomy.

B-266 11:15

CT colonography: Automatic measurement of polyp diameter comparedwith manual assessment in-vivoD. Burling, S. Halligan, S. Taylor, L. Honeyfield, J. Dehmeshki, M. Roddie;London/UK ([email protected])

Purpose: The risk of malignant transformation of an adenomatous polyp is relat-ed directly to its diameter but conventional estimates of diameter made from CTmay vary due to measurement error. We investigated whether automated seg-mentation and diameter assessment was feasible for CT colonography.Methods: Two experienced observers independently measured the maximumdiameter of fifty polyps (colonoscopic reference size range 5-12 mm) from colonog-raphy data sets using conventionally placed software calipers and a variety of 2Dwindow settings (colon, abdominal, bone, lung CT), and also 3D perspective ren-dering. Polyps were re-measured subsequently using in-house automated polyp-segmentation software. Agreement between observers and with the colonoscopicreference were determined using Bland-Altman, Wilcoxen, and Mann-Whitney-Uanalyses.Results: Inter-observer agreement was similar for all window displays: Meandifference in mm (SD difference; 95% limits of agreement) ranged from 0 (1.7; -3.3, 3.3) for 2D colon to -1.1 mm (1.6; -4.3, 2.0) for 3D, compared to -0.5 (2.09; -4.6, 3.6) for automated measurement. When compared to colonoscopy; the largestmean size difference obtained by both observers occurred using the 3D display(1.3, 2.5 mm). There was also significant difference between observers and ref-erence size when using the 2D abdominal and 3D displays (P =.03, p <.001).Subset analysis of tagged cases revealed that one observer overestimated di-ameter for untagged cases measured using 3D (median difference 3 mm versus2 mm, p = 0.01).Conclusion: Automated and manual methods of polyp measurement results insimilar inter-observer agreement. The greatest measurement error is encoun-tered when using a 3D display for estimates of diameter.

B-267 11:24

Impact of computer-aided diagnosis (CAD) on radiologists' performance todetect colonic polyps at CT colonographyM. Cadi, O. Lucidarme, D. Touitou, P. Grenier; Paris/FR ([email protected])

Purpose: To determine potential improvement provided by computer-aided diag-nosis (CAD) system in the detection of polyps at CT colonography (CTC).Methods and Materials: Two different readers, one senior radiologist (R1), andone junior radiologist (R2), independently evaluated CTC exams of 40 patientsreferred for screening colonic polyp or cancer, obtained with a MDCT scanner(16 x 1.25 mm, 1 mm reconstruction interval, 120 kV, 80 mAs). Readings wereperformed first without CAD using a 2D first approach (Colonography software,ADW4.2, GEMS), then with CAD (Mediantechnologies Sophia Antipolis, France).The standard of reference was established from the results of classical colonos-copy that had been also performed in the same patients. Because the CAD toolwas not designed for the detection of polyps smaller than 5 mm, readers consid-ered only polyps above that threshold.Results: The classical colonoscopy identified 19 polyps greater than 5 mm in 14patients. CTC sensitivities to detect polyps without and with the CAD tool wererespectively 89.4% and 94.7% for R1, and 57.8% and 94.7% for R2.

CAD alone selected 195 candidates for polyps. 158 and 162 of these candidateswere correctly identified as false positives of CAD (bulbous fold 41%, stools 16%,ileocecal valve 9.7%) by both readers respectively.Conclusion: CAD increased sensitivity of CTC to detect colonic polyps. Effect ofCAD was greater with the beginner reader. CAD may help reduce false negativediagnoses at CTC, and improve the learning curve of junior radiologists.

B-268 11:33

Performance of a prototype computer-aided detection (CAD) system inMDCT colonographyA. Graser1, L. Bogoni2, M.F. Reiser1, C.R. Becker1; 1Munich/DE, 2Malvern,PA/US ([email protected])

Purpose: To evaluate the accuracy of a CAD system in CT Colonography ascompared to expert readers.Methods and Material: 145 individuals (82 men, 63 women, mean age 59 years)underwent MDCT colonography for colon cancer screening. Scans were performedusing 16-detector row scanners (Somatom Sensation 16, Siemens Medical So-lutions, Forchheim, Germany) at 0.75 mm collimation. Individuals were scannedsupine and prone without fecal tagging. 1 mm axial slices were reconstructed for3D visualization. Two expert readers read the scans using a primary endoluminalapproach. A prototype CAD system (polyp-enhanced view, PEV, Siemens Medi-cal Solutions, Malvern, PA) was used as "second reader". The system's sensitiv-ity for the detection of polyps and the number of false-positive findings per datasetwas evaluated. Average running time of CAD was documented. Polyps were clas-sified as small (≤ 5 mm), medium (6-9 mm) and large ≥ 10 mm).Results: In the study population, a total of 98 polyps (small-51, medium-32 andlarge-15 polyps) were found in 42 patients. CAD detected 69 of these polyps(overall sensitivity of 70 %) with an average of 3.8 false positives per dataset.Average CAD running time was 4.8 minutes. Sensitivity was 51 % (26/51) forsmall, 94 % (30/32) for medium and 87 % (13/15) for large polyps.Conclusion: Computer-aided detection shows high sensitivity in detecting clini-cally significant polyps > 5 mm with acceptable false-positive rates. These num-bers are within the range of expert readers. The short calculation time of the CADsystem makes it useable in routine MDCT colonography.

B-269 11:42

Effect of computer aided detection as a second reader in multidetector CTcolonography: A multiobserver studyT. Mang1, P. Peloschek1, C. Plank1, A. Maier1, C.J. Herold1, L. Bogoni2,W. Schima1; 1Vienna/AT, 2Malvern, PA/US ([email protected])

Purpose: To assess the effect of computer aided detection (CAD) as a secondreader for lesion detection in MDCT colonography. To compare the influence ofCAD on the performance of multiple readers with different levels of experience.Methods and Materials: 33 MDCT colonography data sets (21 patients: 30 en-doscopically confirmed polyps ≥ 5 mm, 5 carcinomas, 12 patients with no pol-yps) were reviewed by four radiologists (2 experienced, 2 inexperienced) on a 3Dworkstation. Virtual endoscopy was used for primary data evaluation with addi-tional 2D views for lesion confirmation. After primary data evaluation, secondreading in consensus with findings of a CAD system (polyp-enhanced view, Sie-mens) was performed. Sensitivities and false positive rates were calculated forthe different readers who were and were not in consensus with CAD findings.Results: Sensitivity of experienced readers was 97% each (29/30) for the detec-tion of polyps. Inexperienced readers reached a sensitivity of 90% and 77%. CADfound 27/30 lesions (90%) and increased sensitivity of inexperienced readers to97% each, with no effect on sensitivity of experienced readers. All carcinomaswere found by 4 readers, 3/5 was found by CAD. False-positive rate was one forboth experienced, and 5 and 6 for unexperienced readers, respectively. CADdecreased false positive rate of inexperienced readers from 5 and 6 to 2 each.CAD consensus reading increased reading time by a mean of 3 minutes.Conclusion: CAD used as a second reader improves sensitivities and reducesfalse positive rates for polyp detection of inexperienced readers. There was nosignificant effect on the performance of experienced readers.

B-270 11:51

CAD in virtual CT colonoscopy without bowel cleansing: Initial experiencesZ. Tarján, J. Koloszár, G.C. Kovács, J. Wacha; Budapest/HU([email protected])

Introduction: The thorough cleansing of the colon is usually a prerequisite ofvirtual colonoscopy similarly to conventional colonoscopy. Nevertheless, thecleansing may be one of the main causes of patient discomfort. We evaluated the

Scientific Sessions

198 C D E FBA G

possibility of a CAD system in detecting polyps in preparationless CT colonogra-phy.Methods and Materials: 21 patients with colonic polyps and 20 patients withoutcolon pathologies underwent preparationless CT colonography and optic colon-oscopy within a week. The patients had no food restrictions or colonic cleansingin the days before the CT examination, only 6 times 20 ml of ionated radiologiccontrast material (not containing meglumin) was administered orally 48 hoursprior to the examination. Native supine low-dose 2.5 mm collimated scans wereobtained on a 4 row scanner. The data were evaluated using an academic soft-ware (ColVis) equipped with electronic cleansing and CAD. The results generat-ed by the CAD and that of a radiologist were compared.Results: 4/41 patients had diarrhea due to the preparation. CAD detected 15/15polyps > 10 mm and 15/19 polyps between 5-10 mm with 17.3 ± 12.6 (mean,SD) false positives/case. The radiologist could detect 15/15 polyps > 10 mm and16/19 polyps between 5-10 mm with 0.5 (mean) false positives/case.Conclusion: Virtual colonoscopy without colonic cleansing minimizes patients'discomfort before the examination. CAD combined with electronic cleansing de-tects large and small polyps similarly to a radiologist but with more false positivesthan in a clean colon.

10:30 - 12:00 Room G

Cardiac

SS 603Myocardial infarction: Imaging assessment andclinical correlationModerators:M.J. Lipton; Boston, MA/USD. Tscholakoff; Vienna/AT

B-271 10:30

Phase-sensitive inversion recovery (PSIR) single shot trueFISP forassessment of myocardial infarction at 3 TeslaA.M. Huber, K. Bauner, B. Wintersperger, S. Schönberg, M.F. Reiser;Munich/DE

Purpose: The aim of the current study was to show if CNR can be improvedwithout loss of diagnostic accuracy for the assessment of myocardial infarction, ifa PSIR single shot trueFISP sequence is used at 3.0 T instead of 1.5 T.Methods and Material: 10 patients with myocardial infarction were examined ata 1.5 T MR system (Avanto, Siemens, Medical Systems) and at a 3.0 T MR sys-tem. Imaging of delayed contrast enhancement was started 10 minutes after ap-plication of contrast media. A phase-sensitive inversion recovery (PSIR) singleshot trueFISP sequence was used at 1.5 and 3.0 T and compared with a seg-mented IR turboFLASH sequence at 1.5 T, that served as reference method. In-farct volumes and CNR of infarction and normal myocardium were comparedwith the reference method.Results: The PSIR single shot trueFISP technique allows to image 9 slices dur-ing a single breathhold without adaptation of the inversion time. The mean valueof CNR between infarction and normal myocardium was 5.9 at 1.5 T and 9.5 at3.0 T. The CNR mean value of the reference method was 8.4. The CNR meanvalue at 3.0 T was significantly (p = 0.03) higher than the mean value of the ref-erence method. The correlation coefficients of the infarct volumes, determinedwith the PSIR single shot trueFISP technique at 1.5 Tesla and at 3.0 Tesla andcompared to the reference method, were r = 0.96 (p = 0.001) and r = 0.99(p = 0.0001).Conclusion: The use of PSIR single shot trueFISP at 3.0 Tesla allows for accu-rate detection and assessment of myocardial infarction with significantly improvedCNR.

B-272 10:39

Characterization of acute and chronic myocardial infarcts by multi-detectorCT. Comparison with contrast-enhanced MRB. Gerber, B. Belge, G. Legros, P. Lim, A. Pasquet, E. Coche,J.-L. Vanoverschelde; Brussels/BE ([email protected])

Background: Gd-DTPA contrast enhanced (CE) MR allows for the characteriza-tion of myocardial infarction (MI) by time dependent contrast enhancement pat-terns. We evaluated whether MI characterization is also possible usingMultidetector CT (MDCT) after injection of iodated contrast at the time of non-invasive coronary imaging.

Methods: Sixteen patients with acute MI, and 21 patients with chronic left ven-tricular ischemic dysfunction underwent MDCT and CE-MR. CE patterns on MDCTimages acquired early (at the time of coronary imaging), and 10 minutes laterwere compared with those obtained with MR.Results: Early MDCT images demonstrated areas of contrast hypo-enhance-ment in 11/16 patients with acute and in 7/21 patients with chronic MI. On seg-mental basis, there was moderately good concordance (92%, κ=0.54, p < 0.001)in the identification of early hypo-enhanced regions between MDCT and MR,Quantitative size of early hypo-enhanced regions was similar (6 ± 16 vs. 7 ± 16 g,p = 0.25), and highly correlated (r = 0.93, p < 0.001) to MR. Late MDCT imagesrevealed areas of delayed hyperenhancement 15/16 patients with acute MI andin 19/21 patients with chronic LV ischemic dysfunction. There was good concord-ance (82%, κ=0.61, p < 0.001), in the identification of these hyper-enhanced re-gions between both techniques on segmental basis. In absolute terms, size ofhyperenhanced area by MDCT was also similar (6 ± 34 vs. 31 ± 40 g, p = 0.14)and highly correlated (r = 0.89, p <.001) to MR.Conclusion: CE-MDCT demonstrates similar contrast enhancement patterns inacute and chronic myocardial infarcts to CE-MR. It may thus provide importantinformation on infarct size and viability at the time of non-invasive coronary imag-ing.

B-273 10:48

How to quantify infarct size on delayed enhanced MR images:A comparison between visual and quantitative approachF. Fiocchi, S. Ferraresi, G. Ligabue, L. Politi, A. Barbieri, M. Modena,R. Romagnoli, P. Torricelli; Modena/IT ([email protected])

Purpose: To evaluate the reliability of visual quantification of infarct extent indelayed enhanced images.Methods: Eighty patients (mean age 63.1 ± 8; 76.4% male) with previous myo-cardial infarction underwent cine and contrast enhanced Cardiac MR imaging.The IR-GE post Gadolinium images (20-28 8 mm thick overcontinuous slices)were evaluated using a segmental model with two different methods: a visual 5-points scale score (0=non hyperenhancement, 4=hyperenhancement > 76% ofmyocardial wall) and with quantitative analysis based on manual contours trac-ing followed by automatic threshold analysis (Cardiac Analysis, ViewForum 4.1,Philips, NL). Each segment was also assigned a wall motion score on B-TFEimages ranging from 0-normokinesia to 4-dyskinesia. Statistical evaluation wasperformed.Results: Mean LV ejection fraction was 49.3 ± 14.9 with EDV and ELV respec-tively of 149 ± 47.8 and 80.7 ± 44.3. Of 1280 segments 322 (25.1%) showed wallmotion abnormalities with enhancement in 327 (25.5%) evaluated with visualscore and in 414 (32.3%) quantitatively. Among segments with normal or mildhypokinesia 89.2% had a delayed enhancement score ≤ 1 whereas 80.2% ofakinetic a dyskinetic segments had a score ≥ 3, confirming the relationship be-tween transmurality of infarction and contractile impairment. Mean time request-ed by visual and quantitative approach was respectively 7 ± 3 and 18 ± 9 minutes.There was high agreement between visual and quantitative methods (K=0.92;p < 0.01); visual analysis underestimated 104 (8.1%) and overestimated 19 (1.5%)segments.Conclusion: Visual analysis of delayed enhancement is a time saving approach,sufficient to assess the transmurality and spatial extent of infarction; moreover ithas high correlation with wall motion abnormalities.

B-274 10:57

Potential of SHU555 A for delineation of ischemically injured myocardiumon MR imagingG.A. Krombach, Y. Temur, M. Dehles, R.W. Günther, A. Bücker; Aachen/DE([email protected])

Purpose: To determine whether ischemically injured myocardium can be deline-ated at MR imaging after intravenous administration of the SPIO (small particleof iron oxide) SHU555 A (Resovist, Schering, Berlin, Germany).Methods: In 7 pigs myocardial infarction was induced by occluding the left ante-rior coronary artery for 45 minutes. Two hours after reperfusion MR imaging wasstarted at a 1.5 T system. T1 values of ischemically injured and remote myocar-dium were measured with the Look-Locker technique prior and after injection ofSHU555 A at a dose of 0.16 mmol Fe/10 kg bodyweight serially over a period of3 hours. Postmortem, hearts were stained with TTC, to delineate the infarct.Results: After intravenous injection of SHU555 A the T1 value of ischemicallyinjured myocardium gradually decreases. ANOVA showed a significant differenceof regional T1 for remote myocardium and infarcted myocardium (F=6.2, df = 4,30, p < 0.001), starting 60 min after injection of SHU555 A. Three hours after

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 199C D E FA G

injection of SHU555 A the T1 values were 588 ± 54 for ischemically injured and759 ± 22 ms for remote myocardium. The size of ischemically injured myocar-dium remained constant over the last 2 hours of the observation period on theMR images. Compared with TTC staining size of infarction was overestimated by5.1 ± 1.6 % on MR imaging.Conclusion: SHU555 A has the potential for delineating the spatial extent ofischemically injured myocardium stably over a time period of at least two hours.This contrast medium may be useful as a marker for MR guided interventions.

B-275 11:06

Single shot phase-sensitive inversion recovery (PSIR) trueFISP fordetection of myocardial infarction during a single breathholdA.M. Huber, C. Hayes, B. Spannagl, A.-E. Werle-Rüdinger, J. Rieber,S. Schönberg, M.F. Reiser; Munich/DE

Purpose: To show that phase-sensitive detection and a single shot techniqueallows imaging of the heart for the detection of myocardial infarction during asingle breathold without adaptation of the inversion time.Methods: 35 patients two weeks to 3 month after Q-wave infarction were exam-ined at a 1.5 T MR System 10 min after administration of a double dose extravas-cular contrast agent. A IR-turboFLASH sequence with optimised TI was used asstandard of reference. A phase-sensitive inversion recovery single shot TrueFISPsequence, that allows to image 9 slices during one breathhold (TR/TE/FA/BW:2.2 msec/1.1 msec/60°,8°/1220 Hz/Px) was used with a nominal TI of 200 msec.160 K-Space lines were acquired during a diastolic acquisition window of352 msec. Spatial resolution was identical for both techniques: 1.3 mm x 1.8 mmx 8 mm. Infarct volumes, area of infarction on a selected slice and scan time werecompared.Results: The mean values for the time of imaging DCE were 10 minutes 32 sec(IR turboFLASH) and 17 sec SS PSIR TrueFISP), p < 0.001. The values for thecorrelation of infarct volume and infarct area were r = 0.95 (p < 0.004) and r = 0.97(p < 0.002). The regression equations were y=0.76+0.92*x and y=0.07+0.93*x,respectively. No significant difference was found for the infarct volume with meanvalues 18.7 and 17.3.Conclusion: Single Shot PSIR TrueFISP allows for accurate identification ofmyocardial infarction during a single breathhold with reduction of scan time by afactor of 37.

B-276 11:15

MR guided intramyocardial injection of substances to infarct borders formolecular therapyG.A. Krombach, Y. Temur, J.G. Pfeffer, S. Kinzel, R.W. Günther, A. Bücker;Aachen/DE ([email protected])

Purpose: To evaluate if interventional MR imaging allows targeting intramyocar-dial injections at the border of myocardial infarcts after intravenous administra-tion of the small particle of iron oxide SHU555 A.Methods and Materials: In six pigs with reperfused infarction three hours afterintravenous injection of SHU555 A a needle-catheter was advanced into the leftventricle under real time MR imaging guidance using a steady state free preces-sion sequence. 2 ml of a 0.1 mmol/ml gadodiamide solution, mixed with 3 % Evansblue dye, were injected at two regions at the border of the infarct. T1 values wereserially measured prior and 3, 15 and 30 min after each intramyocardial injection.Results: Injection of SHU555 A caused a decrease of the T1 value of the infarct-ed myocardium (778 ± 63 ms before, and 588 ± 54 ms 3 h after injection). Thecontrast between the infarct and remote myocardium was sufficient for delinea-tion of the infarct on real-time images (0.43 ± 0.02). In all animals, the cathetercould be directed into the border region of the infarct. In each case, two injectionswere performed at different locations. T1 values at the injection sites decreased(240 ± 25 ms) and slowly increased (468 ± 37 ms 30 min after intramyocardialinjection).Conclusion: Percutaneous, MR-guided intramyocardial injection of a gadodiamidecontaining solution into the border of infarcted myocardium is feasible after theischemically injured region is delineated following i.v. injection of SHU555 A. Dueto the strong decrease of T1 values at the injection sites repeated injections intothe same region can be avoided.

B-277 11:24

31P-MR-spectroscopy in patients with recent myocardial infarction:Correlation with wall thickness analysisW. Machann, H. Koestler, J.J. Sandstede, M. Spindler, D. Hahn, M.J. Beer;Wuerzburg/DE ([email protected])

Purpose: The purpose of this study was to compare energy metabolism in non-infarcted and infarcted myocardium using 31P-MR-Spectroscopy (MRS) and tocorrelate these changes with the outcome of the revascularization.Methods: 30 patients with subacute anterior myocardial infarction (MI) under-went MRS and 31P-spectra were obtained from non-infarcted/septal as well asinfarcted/anterior myocardium (voxel size 25 ml each) by means of 3D-CSI. Phos-pocreatin/adenosine triphosphate (PCr/ATP) ratios were determined using "AM-ARES". Left ventricular (LV) wall thickness was evaluated by cine magneticresonance imaging (MRI). To assess myocardial viability, MR imaging was re-peated 6 months after revascularization. 15 age matched volunteers served ascontrol.Results: 15 of 30 patients regained normal wall motion (VIAB), whereas theother 15 not (NVIAB). PCr/ATP-ratios of infarcted/anterior myocardium were sig-nificantly reduced for both groups (1.17 ± 0.25 (VIAB) and 0.71 ± 0.70 (NVIAB);1.51 ± 0.32 (control)). Regarding non-infarcted/septal myocardium, a reductionof PCr/ATP ratios was found only in NVIAB. Reductions of PCr/ATP ratios did notcorrelate with decreased wall thickness in any group.Conclusion: Irreversibly damaged myocardium is associated with severe de-pression of energy metabolism, which in turn affects even non-infarcted myocar-dium independent of wall thickness at the subacute stage of MI.

B-278 11:33

MR imaging before CABG surgery: Impact of scar extent in contrast-enhanced MR imaging on coronary graft flowP. Hunold, P. Massoudy, T. Schlosser, C. Böhm, J. Forkel, J. Barkhausen;Essen/DE ([email protected])

Purpose: To relate myocardial late enhancement (LE) in contrast-enhanced MRimaging (CE-MRI) to intraoperative graft flow in patients with impaired LV func-tion undergoing CABG surgery.Methods and Materials: Thirty-three CAD patients (mean EF 29%) underwentCE-MRI before CABG surgery. LE scans were acquired after 0.2 mmol/kg Gd-DTPA using an inversion-recovery turboFLASH sequence (TR, 8 ms; TE, 4 ms;FA, 25°; TI, 200-260 ms). Myocardial scar extent was evaluated using a LE score1-4 (score 1, no LE; 2, LE < 50% of wall thickness; 3, LE > 50%; 4, transmural LE)for each vessel territory (RCA, LAD, LCX). Intraoperative flow as determined byDoppler sonography in grafts supplying vessel territories with a score 1 or 2 (groupA) was compared to flow in grafts supplying vessel territories with a score ≥ 3(group B) which are not expected to improve function after revascularization.Results: Of the 99 vessel territories, 89 yielded score 1 or 2 (group A), 10 score≥ 3 (group B). In group A, 68 of 89, and in group B, 9 of 10 territories were graftedusing left intrathoracic artery (LITA) or venous graft. Mean LITA graft flow was54 ± 6 in group A and 28 ± 11 cc/min in group B (p=.04). Mean vein graft flowwas 74 ± 6 and 42 ± 8 cc/min in groups A and B, respectively (p=.05).Conclusion: Graft flow to vessel territories with no or little scar tissue is higherthan flow to territories with > 50% scar tissue. Since sufficient flow is a prerequi-site for graft patency, the prognosis of those grafts may be better.

B-279 11:42

Relationship between clinical parameters and infarct size as evaluated bymeans of cardiac MR imaging in patients submitted to primary coronaryangioplastyG. Ligabue, F. Fiocchi, S. Ferraresi, L. Rossi, A. Barbieri, M. Modena,R. Romagnoli, P. Torricelli; Modena/IT ([email protected])

Purpose: To evaluate if conventional infarct size clinical parameters at admis-sion correlates to infarct extension as depicted by cardiac MR imaging.Methods: Forty-six symptomatic patients (mean age 64.2 ± 13; 80.4% male)underwent cardiac MR imaging within 1 week after successful PCA performedfor acute myocardial infarction. Cardiac MR imaging was performed including thefollowing breath-hold ECG-gated sequences: Cine-TFE, IR-GE post gadolinium.Between clinical parameters measured at admission we choose the following: Q-wave presence, CK-MB peak, symptoms-to-balloon time (min). Cine and con-trast enhanced MR images were evaluated with a segmental semi-quantitativescore: wall motion score (WMS) and segmental scar score (SSS). Patients withmultivessel disease and/or previous infarction. Univariate and multivariate statis-tical evaluation was performed in the remaining 38 subjects.

Scientific Sessions

200 C D E FBA G

Results: The mean symptoms-to-balloon time was 192.8 minutes. Q-wave waspresent in 27 subjects (71%). Mean CK-MB peak was 235 ± 98.8 ng/ml in pa-tients who exhibit Q-waves and 171.4 ± 102.3 in the remaining. Most of segmentswith contractile impairment exhibit delayed enhancement (p = 0.03).Q-wave presence correlates with infarct extension as quantified by segmentalscar score (25.1% VS 12.8%; β=7.99; p = 0.01) and wall motion abnormalities(6.3% VS 18%, β=7.32; p = 0.02). Also high CK-MB peak positively correlateswith SSS (β=0.05; p = 0.00) and WMS (β=0.04; p = 0.00). MR imaging parame-ters did not significantly correlate with Symptoms-to-balloon time (p = 0.70 andp = 0.36 respectively). Multivariate analysis confirmed those correlations.Conclusion: Q-waves presence and CK-MB peak are better predictors of bothspatial and transmural extension of infarction and contractile impairment withrespect to other clinical parameters.

B-280 11:51

Multislice multiecho T2* MR imaging approach for the detection ofmyocardial iron distribution: Validation in healthy subjects and thalassemiamajor patientsA. Pepe1, E. Cracolici2, V. Positano1, L. Pitrolo2, L. Prossomariti3, D. De Marchi1,M. Midiri2, M. Lombardi1; 1Pisa/IT, 2Palermo/IT, 3Napoli/IT ([email protected])

Purpose: To validate the multislice multiecho T2*MR imaging approach, for thedetection of myocardial iron distribution, in healthy subjects and in thalassemiamajor (TM) patients.Methods and Materials: Three short axis views of the left ventricle were ob-tained using T2* multiecho sequences in 20 healthy subjects and in 53 TM pa-tients. The myocardium was segmented into 12 segments. T2* value on eachsegment and the global T2* value were calculated. To assess the endomyocar-dial and epimyocardial iron distribution, each segment was divided in the innerand outer half.Results: The coefficients of variation for intra-, inter-observer and inter-studyreproducibility were 3.9%, 5.5%, 4.7% respectively. In healthy subjects the per-centage of deviation (%D) of the slices from the global heart T2* mean was notstatistically significant. The %D of the segments from the global heart T2* meanranged from -20 to 18. Significant differences were observed only among thesegments with higher %D (5/66 cases). In TM patients, the epimyocardial ironconcentration was significantly higher in the slices and in 10/12 segments. Threegroups of patients were identified: 15% with homogeneous myocardial iron over-load (MIO); 47% with no MIO; 38% with a heterogeneous MIO (T2* segments< 20 ms and T2* segments > 20 ms). Serum ferritin levels, liver iron concentra-tions and urinary iron excretions were significant different among groups.Conclusion: Our data did not account for rejecting certain slices or segments.Multislice multiecho T2* MR imaging approach provides a reproducible way forassessing myocardial iron distribution, which may play a role in clinical arena.

10:30 - 12:00 Room H


SS 609aNew interventional treatmentsModerators:J.E. Jackson; London/UKF. Orsi; Milan/IT

B-281 10:30

US-guided percutaneous ethanol injection treatment of thyroid autonomousand toxic nodules: Efficacy and long-term follow-upB. Brkljacic, I. Cikara, V. Bozikov, M. Sucic; Zagreb/HR([email protected])

Purpose: To evaluate US-guided percutaneous ethanol injection (PEI) treatmentof thyroid toxic and autonomous nodules.Methods and Materials: Between September 1996 and June 2005 US-guidedPEI was performed in 65 patients (61 F, 4M) with thyroid autonomous (24) andtoxic adenomas (41). Nodular volumes were 2.5-38 ccm (mean 20.8). 96% etha-nol was used, with total volumes injected (4-57 ml) of 1.5 times nodule volume.Procedure was performed twice a week until total volume injected. 3-6 monthsafter completion of PEI outcome evaluation was performed by hormonal status(T3,T4,TSH), scintigraphy and US. The outcome was cathegorized as completerecovery, partial recovery or failure. In 21 patients regular follow-up was > 5 years(5-8.5 years) with hormonal status assessed regularly annually/biannually, US atleast biannually, and scintigraphy performed at least once.

Results: Technically successful PEI was performed in 61/65 patients (94%). Vol-ume reduction > 50% was observed in 57/61 patients (93%). Complete recoverywas noted in 37/61 patients (60.7%), partial recovery in 18/61 patients (29.5%)and failure in 6/61 (9.8%). Failure was noted in three patients with toxic nodulargoiters and three patients with toxic adenomas. Follow-up > 5 years in 21 pa-tients showed three cases of hyperthyroidism recurrence, after three, four andsix years, with scintigraphically hot nodules in treated (2) and non-treated (1)thyroid gland parts. All were successfully retreated with PEI. 18/21 patients wereeuthyroid (85.7%), with no new scintigraphic hot nodules.Conclusion: US-guided PEI treatment of thyroid autonomous nodules is an effi-cient method with very good initial and long-term results.

B-282 10:39

Evaluation of a novel navigation system with a patient fixation device forinterventional CT-guided proceduresM. Meier-Meitinger, M. Nagel, U. Baum, M. Uder, W.A. Kalender, W. Bautz;Erlangen/DE ([email protected])

Purpose: Evaluation of the clinical practicability and accuracy of a novel naviga-tion system for interventional CT-guided procedures with a patient fixation deviceand virtual navigation in a phantom and patient study.Methods and Material: The new navigation system (CAPPA IRad) was appliedto CT-guided biopsies in 3 patients and to 15 in a phantom. The system consistsof a standard PC, navigation software, a passive optical tracking system and apatient frame. For the phantom study we used a prepared spine phantom. First ofall we fixed the patient or respectively the phantom with a patient-fix-kit on the CTtable, positioned a patient frame with optical and CT markers above and the set-up was scanned with CT. All data were transferred to the navigation system in-side the scanner room. The target approach was defined and the trajectory wasadjusted to the surface of the corresponding CT marker. The needle holder wasorientated towards the trajectory. Finally a coaxial biopsy needle was pushedtowards the planned target. The accuracy of the procedures was evaluated usingnavigation software for measurement target-needle tip.Results: The deviation between planned and documented needle tip positionwas 0.9 mm in the phantom study (n = 15) and 0.8 mm in the patient study (n = 5).Time between the end of planning CT and needle positioning using the systemwas 29 minutes. There were no complications.Conclusion: CAPPA IRad allows interventions with high accuracy. Advantagesare the precision and less traumatic intervention even of the smallest lesions withan angulated 3D approach and reduction of radiation exposure.

B-283 10:48 !Clinical experience with a new retrievable vena cava filterF.G. Barral1, J.P. Beregi2; 1Saint-Etienne/FR, 2Lille/FR([email protected])

Purpose: Insertion of a permanent vena cava filter has been shown to reduceoccurrence of pulmonary embolism despite an increased risk in DVT over thelong-term. Retrievable or optional caval filters offer the ability to be removed ornot, according to PE exposure.Methods and Materials: A bi-centric retrospective study was performed to as-sess for filter efficacy and for ability to remove the ALN filter in patients present-ing with a severe acute venous thromboembolism associated with at least oneindication for a filter such as: contra-indication to anticoagulant (mainly due to anhaemorrhagic phenomena), acute major surgery or recurrence of venous throm-boembolism despite anticoagulant.Results: This study included 285 (225 + 60) patients presenting with DVT (26%),pulmonary embolism (13%) or both (61%). After a median follow-up of 11.4 months,there were two episodes of pulmonary embolism and three thromboses in thefilters. A decision to remove the filter was taken in 70 (55+15) patients (24.5%)and all filter retrievals were successful without any complications. The medianimplantation period was 48 days (range 6 days- 14 months).Conclusion: These preliminary results confirm the efficacy of the ALN filter. Theyalso demonstrate the feasibility and safety of retrieval up to 14 months after im-plantation.

B-284 10:57

A new method of treatment for deep vein thrombosis: Techniquedescription and clinical significanceC. Cronin, D. Lohan, N. Gough, C. Meehan, G. O'Sullivan; Galway/IE([email protected])

Purpose: Describe an institutional experience of acute ileo-femoral (IFV)/super-ficial femoral (SFV) deep venous thrombosis (DVT) treatment with the Bacchus

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 201C D E FA G

Trellis device which allows clot break-up/clearance. To outline its advantages:significant reduction in patient symptomatology, potential reduction of long termmorbidity, shortened hospital stay.Methods and Materials: Twelve patients presenting with acute IFV/SFV throm-bosis diagnosed on Doppler Ultrasound were referred for treatment with the Bac-chus Trellis Device. Symptoms were present for an average of 3.5 days prior todiagnosis. Prospective color Doppler ultrasonographic follow-up is on-going.Results: Subjective impression of reduction of thrombus load varied from 75-100% (6 patients 100%, 3 patients 75%, 3 patients 60%). Radiological resolutionof obstruction, reduction in pain within 24hrs and measurable reduction in legswelling within 72hrs, was seen in all. Based on retrospective hospital data, meanhospital stay reduced from 7 to 2 days. No symptomatic recurrence was noted at3-month.Conclusion: Standard DVT anti-coagulation regimes merely prevent further clotpropagation and diminish risk of pulmonary embolus. Catheter directed throm-bolysis is well established for IFV DVT particularly with phlegmasia cerulea do-lens (PCD), but is time consuming, usually requires ICU admission and involvesthe risk of systemic passage of thrombolytic agents. The Bacchus Trellis devicehas the major advantage of completion of treatment in a single sitting, without theneed for indwelling venous sheaths, ICU stay, repeat blood testing, systemic an-ticoagulation or thrombolytic administration. We show the feasibility of major clotclearance with no significant risks. Although the cost may not be inconsiderable,much quicker discharge from hospital is a major advantage.

B-285 11:06

Percutaneous intrahepatic transplantation of islets after kidney (IAK) andislets (ITA) in 34 type I diabetic patients: Technical aspects, complicationsand clinical outcomesM. Venturini, P. Maffi, E. Angeli, P. Fiorina, A. Secchi, A. Del Maschio; Milan/IT([email protected])

Purpose: Pancreatic-islets represent an alternative to whole-pancreas transplan-tation in diabetic kidney-transplanted patients (IAK). Recently, IA transplantation(Edmonton Protocol) has been adopted for non-uremic patients to prevent dia-betic complications, with a different immunosuppressive regimen. Our aim was tocompare IAK and ITA regarding technical feasibility, complications and clinicaloutcomes.Methods and Materials: Ultrasonography and fluoroscopy-guided intraportal is-lets transplantation was performed in 24 kidney-transplanted (IAK) and 10 non-uremic (ITA) diabetic patients. Immunosuppression was based on cyclosporine,azathioprine, mycophenolate-mofetil and low-dose-steroids for IAK and on a ster-oid-free regimen (sirolimus, tacrolimus) for ITA. For each procedure we evaluatedtechnical success, number of puncture attempts, complications and, at one year,the following clinical outcomes: well-function (C-peptide > 0.5 ng/ml), insulin-in-dependence, transaminases, creatinine, urinary albumin excretion (UAE).Results: The success of the percutaneous procedure was 100%, with a single-puncture attempt in 90% of the cases. Procedure associated complications were:1 peripheral portal thrombosis (ITA), 1 candida infection (IAK), 2 hemoperitone-um (1-IAK, 1-ITA). IAK and ITA showed similar 1-year rate of well-function andinsulin-independence (IAK: 58%, ITA: 62%). Transaminases were significantlyhigher in ITA than IAK at 1 month, normalizing during the follow-up. Creatinineand UAE were stable after islets transplantation (excepting 1 ITA with pre-exist-ing mild nephropathy, in which they increased).Conclusion: Percutaneous intraportal islets transplantation is a safe, low-com-plication procedure using a combined ultrasonography and fluoroscopy-guidedtechnique. A similar clinical outcome was found in IAK and ITA patients.

B-286 11:15

First experimental evaluation of an expandable percutaneous catheterpump for left ventricular supportT. Schmitz-Rode, J. Graf, J.G. Pfeffer, C. Brücker, R.W. Günther; Aachen/DE([email protected])

Purpose: To evaluate the performance of a newly designed percutaneous cathe-ter with expandable pump as a temporary support system for patients with acuteleft ventricular failure.Methods and Materials: The pump catheter (introduction diameter 9 F) is posi-tioned in the left ventricle. The rotor is driven by an external motor through aflexible drive shaft. A model circuit was used to assess pump performance, hemo-lysis tests, and particle image velocimetry. The feasibility of the catheter place-ment and pump operation were examined in 12 anesthetized sheep. In 7 of theanimals cardiogenic shock was induced. Cardiac output (CO) and mean aorticpressure (MAP) were recorded before and during shock, and during catheterpump action.

Results: The catheter pump delivered a flow of 4.1 l/min at a differential pressureof 60 mmHg. The average modified index of hemolysis (MIH) was 11.6 (optimum1.8). Fluoroscopically and echocardiographically guided in-vivo placement anddeployment of the device were quick and uncomplicated. Under simulation ofacute left ventricular failure (CO 43 ± 22% and MAP 55 ± 16% of the baselinevalue) the catheter pump significantly improved CO to 67 ± 12% and MAP to74 ± 18%. Maximum in-vivo duration of operation was 6 hours (average 3.1 ± 1.4hours). Animal studies revealed: 1. no significant hemolysis (average plasma freehemoglobin 26 ± 4 mg/l after 3 hours), 2. no thrombotic deposits at rotor or pumphousing, and 3. no damage of endocardium or aortic valve.Conclusion: A percutaneously insertable, expandable catheter pump is techni-cally and clinically feasible. First experimental results are encouraging.

B-287 11:24

Soft sclerosis for treatment of hepatic cystic diseasesF.M. Danza, A. D'Amico, M. Cirillo, A. Magistrelli, E. Bock, L. Bonomo; Rome/IT([email protected])

Purpose: To describe a new percutaneous treatment of symptomatic cystic dis-eases of the liver, with lower discomfort for patients and better handling for theoperator.Methods and Materials: 16 patients with a total of 19 cysts were treated with asoft method instead of classic alcoholization. Cysts ranged between 11 cm and19 cm in diameter. After US-guided percutaneous puncture 50% of the fluid wasextracted from the cyst, and a solution of minocycline hydrocloride (1 mg/ml ofnative volume of the cyst) was introduced and left in the cavity (single-shot). 2patients had polycystic liver disease (5 cysts). Follow-up was performed at 3, 6, 9and 12 months after treatment.Results: At 6 months treatment was successful in 17/19 cysts (89.4%). In 2/16 pts (10.6%) a re-treatment was needed, with a total success rate of 100% at12 months. Multiple shots technique through a small catheter was used in 1 ofthe recurred cases (brown, corpuscolated fluid at diagnosis). No complicationswere noted. No recurrences were noted at 2 years in 10 patients.Conclusion: This method is superior to alcoholization: The procedure is morerapid and less expensive (catheter set is not required). The discomfort for thepatient is minimal, the procedure is easier for the operator and results are better,without risks in case of extravasation into the peritoneal cavity or into the biliarytree.

B-288 11:33

Evaluation of an animal model for radiofrequency ablation of bonemetastases of human breast cancer cell lineMDA-MB-231 in nude ratsD. Proschek, J. Balzer, T. Lehnert, K. Eichler, M. Mack, T.J. Vogl; Frankfurt/DE([email protected])

Purpose: To evaluate an animal model for radiofrequency ablation of local bonemetastases of human breast cancer cell line MDA-MB-231. Primary objectiveswere number and size of bone metastases as well as the feasibility of radiofre-quency ablation in nude rats.Methods and Materials: Human breast cancer cell line MDA-MB-231 (106 tumorcells) was implanted into the femur of 10 nude rats. Weekly CT- and MR imaging-scans were performed to document number and size of bone metastases. Radi-ofrequency ablation (22 g bipolar and impedance-controlled RF-applicator, 2-4Watt, 3 min application time) was carried out. One week after RFA, the animalswere sacrificed and macroscopic and histological examination followed.Results: All animals developed local bone metastases. RFA was performed with-out complications. Imaging showed a complete ablation of the bone tumor in allrats. Histological findings confirmed a circular necrosis with extensive destruc-tion of tumor cells leaving a necrotic cavity.Conclusion: RFA in human breast cancer cell line in nude rats is a feasible anduseful possibility to evaluate and test different RF-procedures. Additional treat-ment options like local chemotherapy or chemoembolization can be performed.

B-289 11:42

Endovenous laser treatment (EVLT) for lower limb varicose veins due togreat saphenous vein (GSV) refluxK. Tay, S. Lin, B. Tan, S. Tan, M. Sebastian; Singapore/SG([email protected])

Learning Objective: To report our early experience with EVLT for treatment oflower limb varicose veins due to GSV reflux.Methods and Materials: Fifty-six GSVs in 55 patients (40 F:16M) with a meanage of 64 years (range 25-90 years) were treated between February 2004 and

Scientific Sessions

202 C D E FBA G

June 2005. All patients had symptomatic varicose veins with 80% (45/55) of pa-tients having CEAP Class 3 and above varicose veins, including 17 patients withnon healing venous ulcers. Treatment was performed with the Diomed 810-nmdiode laser via a 600-µm fibre introduced percutaneously into the GSV underlocal anaesthesia using ultrasound and fluoroscopic guidance. No adjuvant in-jection sclerotherapy or avulsion phlebectomy was performed. Patients were fol-lowed clinically and with duplex ultrasound. Mean follow-up period was 3.7 months(range 1 week to 15 months).Results: There was one technical failure as the GSV could not be cannulateddue to severe venospasm following guidewire insertion. The remaining 55 GSVswere successfully occluded with no evidence of recanalisation on follow-up du-plex ultrasound. All patients reported significant improvement or resolution ofsymptoms (cramps, swelling) by 1 week. The majority of venous ulcers (14/17)were healed by 3 months. Mild to moderate bruising and tenderness along thecourse of the treated vein, which resolved within 2 weeks, was noted in all pa-tients. There was one case of skin burn (blisters over the lower thigh) which healedwithin 1 week.Conclusion: Our early results suggest that EVLT is a safe and effective treat-ment for symptomatic varicose veins secondary to GSV reflux.

B-290 11:51

Implantable blood pressure monitoring system: First in vivo evaluationK.M. Ruhl, J. Tomko, R. Schlierf, J.G. Pfeffer, U. Schnakenberg, R.W. Günther,A.H. Mahnken, T. Schmitz-Rode; Aachen/DE ([email protected])

Purpose: To prove the feasibility of invasive arterial blood pressure (ABP) mon-itoring with a miniature pressure sensor designed for use in an implantable te-lemetry system for continuous ABP monitoring.Methods and Materials: After arterial puncture a miniature pressure sensor onthe tip of a 3 F catheter was inserted into the femoral arteries of 6 sheep. A microcable inside the catheter allowed the read-out of the measured values with anexternal reading unit. As a reference, a wire-mounted commercially available pres-sure sensor was placed into the contralateral femoral artery. To minimize the lossof blood during the insertion procedure, and to reduce bleeding complicationsafter removal of the catheter and the wire, a new peel-off arterial closure devicewas designed during this study.Results: In each sheep a continuous ABP curve could be obtained with the min-iature pressure sensor. The measured ABP values correlated well with those ofthe reference sensor. The arterial closure device is capable of preventing bleed-ing during insertion, led to a stable puncture site throughout several hours andprevented complications like haemorrhage or subcutaneous haematoma afterextraction of the catheter.Conclusion: Invasive ABP measurement with the miniature pressure sensor isfeasible. Further studies are needed for further miniaturization of the system,subcutaneous implantation of the telemetric unit, wireless data transfer and long-term results.

10:30 - 12:00 Room I


SS 609bBiopsy techniquesModerators:V. Chabbert; Toulouse/FRM. Dean; London/UK

B-291 10:30

10 G vacuum assisted biopsy under MR imaging: Feasibility andpreliminary resultsC. El Khoury, A. Tardivon, F. Thibault, L. Ollivier, B. Barreau, V. Servois,S. Neuenschwander; Paris/FR ([email protected])

Purpose: To assess incidental enhancing breast lesions without mammographicor clinical correlation, with normal second look ultrasound, by MR guided 10 Gvacuum assisted biopsy.Methods and Materials: 30 patients were referred to MR guided biopsy with MRfinding of a suspicious enhancing lesion (BI-RADS categories 4 or 5) with normalmammograms and normal second look ultrasound. 9 patients had a BRCA1 orBRCA2 mutation, 8 a personal history of breast cancer.Shape of enhancement was correlated with pathological results.Results: Biopsy was successful in 29 patients. One was rejected because the

lesion was too close to pectoralis muscle. 9 biopsies revealed malignant lesions(31%): 6 invasive carcinomas and 3 Ductal carcinoma in situ. One revealed fi-brous changes with lobular carcinoma in situ (LCIS). 19 biopsies were benign. Ofthe malignant lesions, 5 presented as an enhancing mass, one as ductal en-hancement and 3 as multinodular enhancement with segmental distribution. TheLCIS lesion presented as a mass and benign lesions presented as: 6 enhancingmasses, 4 ductal enhancements, 3 multinodular enhancements with segmentaldistribution, 2 linear enhancements and 1 multinodular enhancement. The posi-tive predictive value upon the shape of enhancement is: 45% for masses, 50% formultinodular with segmental distribution, 20% for ductal enhancement. Linear,focal or multinodular regional enhancements were benign in our study.Conclusion: MR guided vacuum assisted biopsy is a useful tool in managementof incidental enhancing lesions with negative second look ultrasound. It revealed31% malignant lesions among 29 suspicious enhancements (categories 4-5).

B-292 10:39

Comparison of ultrasound-guided fine needle aspiration of infracentimetricthyroid nodules according to sonographic criteriaR. Butros, F. Boyvat, C. Aytekin, B. Banu, N. Guvener Demirag, B. Demirhan;Ankara/TR ([email protected])

Purpose: Referrals to endocrinologists for incidentally discovered thyroid nod-ules have become increasingly common. The proper management of non-palpa-ble thyroid nodules has been debated over recent years. The purpose of ourstudy was to compare the biopsy results of two groups of infracentimetric nod-ules according to sonographic criteria.Methods and Materials: Ultrasonography of 491 patients showed 584 infracen-timetric thyroid nodules. Nodules were classified as positive or negative for sono-graphic features that suggest malignancy. These included: an irregular margin,microcalcifications, lymphanedopathy and marked hypoechogenicity. Fine nee-dle aspiration biopsy was performed on 317 positive and 272 negative nodules.Results: Of 317 nodules that were classified as positive, 9 were malignant. Of272 nodules that classified as negative, 4 were malignant. Sensitivity, specifity,and accuracy were calculated.Conclusion: Although we selected two groups according to the sonographic find-ings, we found a low prevalence of malignancy in both study groups. This raisesthe question; do we need to biopsy infracentimetric nodules? Studies are neededcorrelating pathological results with ultrasound appearance in order to identifynodules with a higher risk of malignancy.

B-293 10:48

Focal liver lesions: Evaluation of core biopsy with a multislice-CT andcomparison between 16-gauge and 18-gauge needlesJ. Stattaus, H. Kühl, E.A. Hauth, J. Kalkmann, H.A. Baba, M. Forsting;Essen/DE ([email protected])

Purpose: To evaluate MSCT-guided biopsy of focal liver lesions and to determinethe rate of diagnostic biopsies using 16G semi-automated core biopsy systemscompared to 18G systems.Method and Materials: From July 2001 to April 2005, 163 patients underwentMSCT-guided biopsy of the liver. Indication was based on the detection of a focalliver lesion on a preceding CT or MR imaging. In most cases a 16G automatedcore biopsy system was used (n = 121), partially in coaxial technique using a15G guidance needle (n = 68), partially without guidance (n = 53). An 18G corebiopsy system was used in 42 patients, under coaxial guidance in 27 cases.Results: Sensitivity, specificity and accuracy for MSCT-guided biopsy were 93.3%,100% and 94.5% respectively. Sensitivity and accuracy values were higher for16G biopsy compared to 18G biopsy: 97.2% vs. 78.6% and 97.5% vs. 85.7%respectively. A definitive histological diagnosis was obtained in 90.0% including94.0% in the 16G group and 75.8% in the 18G group. Using Chi-square-test allthese differences between 16G and 18G proved to be highly significant (p < 0.01).Median biopsy duration was 27 minutes. There was only one major complication:expansive subcapsular bleeding.Conclusion: MSCT-guided core biopsy provides fast and accurate diagnosis offocal liver lesions. The statistically significant higher rate of diagnostic resultsjustifies the use of biopsy systems with larger core diameter (i.e. 16G).

B-294 10:57

Ultrasound-guided biopsy of thoracic massesM.A.M.S. Abusedera; Sohag/E.G. ([email protected])

Purpose: To determine the utility of ultrasound as a guiding tool for biopsy ofthoracic masses.

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 203C D E FA G

Methods and Materials: The criteria for inclusion in the study were: thoracicmasses that were diagnosed by preliminary contrast enhanced CT scan and wereaccessible by ultrasound, and good patient coagulation profile.Results: Twenty-five patients had thoracic masses (two posterior mediastinal,one anterior mediastinal, two pleural, one chest wall, nineteen parenchymal).The lesion diameters ranged from 3-7 cm with average diameter of 5.3 cm andnumber of passes ranged from 2-4 with average of 2.9 passes per patient. His-topathological diagnosis of biopsy specimen was achieved in 92% (n = 23), in4% (one case) histology was inconclusive and the sample was insufficient in onecase (the biopsy was repeated). Flexible fibro-optic bronchoscopy was done in19 cases. It showed a central mass in 37% (n = 7) and extrinsic compression in10% (n = 2), could not be completed in 21% (n = 4) and was negative in 32%(n = 6). Cytological examination following bronchoalveolar lavage was positive in32% (n = 6) and negative in 47% (n = 9). No samples could be obtained in 21%(n = 4). Pleural effusion was found in 10 cases. Cytological examination of thepleural effusion was positive for malignant cells in 40% (n = 4) and negative in60 % (n = 6). There were no major complications or fatalities but minor complica-tions occurred in 8% (n = 2): mild hemoptysis in one patient and a small pneu-mothorax in another (neither required further intervention).Conclusion: Ultrasound can be used as image guidance for biopsy of thoraciclesions safely, with high diagnostic yield.

B-295 11:06

Navigation system for CT-guided percutaneous punctures: First phantomexperienceR. Bale1, M. Lang1, C. Lederer1, T. Lang1, W. Jaschke1, M. Nagel2;1Innsbruck/AT, 2Erlangen/DE ([email protected])

Purpose: Evaluation of a novel CT-guided navigation system for percutaneouspunctures.Method and Materials: The navigation system (CAS innovations AG, Erlangen,Germany) consists of an aiming device and a reference frame which allows forfully automatic image-to-patient registration. For evaluation of the accuracy aphantom consisting of a base plate and 20 rods with various lengths with glassspheres at the top of every rod was used. Helical CT scans of the phantom withthe reference frame with different slice thickness (1.25 mm, 2.5 mm and 5.0 mm)were obtained and the dataset was sent to the navigation system.Results: 300 punctures were performed. CT-scans were obtained and the dis-tance from the needle tip to the previously planned target point was measured.The distance was defined as normal deviation from the achieved position to theplanned target. The accuracy of all 300 punctures was 1.2 ± 0.67 mm. The meanaccuracy for 1.25 mm scans was 1.0 ± 0.6 mm (n = 100), for 2.5 mm scans1.1 ± 0.6 mm (n = 100) and for 5.0 mm slice thickness 1.5 ± 0.8 mm (n = 100).The deviation in the needle direction as evaluated in a separate study was0.31 ± 0.18 mm in 90 punctures.Conclusion: The presented device seems to be an ideal tool for various percuta-neous CT-guided interventional procedures including biopsies, local tumor thera-py, vertebroplasty etc.

B-296 11:15 !CT-navigation system for biopsy needle procedures: Initial evaluation in aphantom experimentM. Schiemann1, M. Nagel2, F. Khan1, T. Vogl1; 1Frankfurt/DE, 2Erlangen/DE([email protected])

Purpose: Experiment to assess accuracy of a novel CT-navigation system forbiopsy procedures in a phantom.Methods: The navigation system (CAPPA) consists of an optical tracking system(Polaris), device holder, a reference frame with CT-detectable markers as well asreflective markers visible for the optical tracking and user interface connected tothe CT scanner. The frame is placed over the target area prior to CT-scan. Theacquired CT volume data set is transferred to the user-interface for navigationplanning. To measure the technical accuracy a calibrated 10G needle was placedwith its tip exactly on 60 different targets of an individual acrylic phantom. Theoffsets calculated by the navigation system were noted. To evaluate the accuracyof the needle-positioning of the navigation system 6 radiopaque targets (d=2 mm)located in different positions in the phantom were punctured with a 7.5G needle(l=11 cm); control CT scans (Sensation 16) were made to verify needle tip posi-tions. This procedure was repeated 10 times with a total of 60 punctures. Thedeviation from the needle tip position to the target in the CT scans was analyzed.Results: The measurement of the technical accuracy of the system showed amean tip error of 0.66 mm ± 0.20 mm. In the needle positioning experiment thedeviation from the needle tip location to the target was mean 0.96 mm ± 0.26 mm.

Conclusion: The navigation system provided excellent targeting accuracy in thephantom experiments. It may offer advantages over conventional procedure whileeliminating the need for repetitive fluoroscopy.

B-297 11:24

Comparative study of FNAB and core needle biopsy of pulmonary andmediastinal lesionsA.I. Ikramov, N.M. Djuraeva, M.F. Maksudov; Tashkent/UZ([email protected])

Purpose: To compare the effectiveness of percutaneous fine needle aspirationand cutting biopsy of pulmonary and mediastinal lesions.Methods and Materials: We retrospectively studied 105 CT-guided percutane-ous interventional diagnostic lung (78) and mediastinal (27) biopsies that wereperformed between July 2002 and August 2005. For the FNAB, 20G needleswere used. For the cutting biopsy, automated 18G core biopsy needles with a 1-2 cm cutting edge were used. FNAB was performed in 54 patients and core biop-sy needles were used in the other 51 patients. In 14 cases of core biopsy and 30cases of FNAB, repeated manipulations were performed. The samples were an-alyzed both cytologically and morphologically.Results: Adequate samples for histological examination were obtained in 76 cases(72.3%), comprising 50 of 51 (98.03%) core biopsies and 26 of 54 (48.1%) FNABs.Morphologic analysis showed: metastatic lesions (39), primary tumors (37), tu-berculoma (3), hemangiopirocytomas (2), hemodectomas (1), inflammatory le-sions (17).Negative results of FNAB were 32.5%, verified after surgery. Specificity of corebiopsy was 87.5%. In 3 cases occurred complications (pneumothorax).Conclusion: Percutaneous transthoracic CT-guided core biopsy is a fast, safeand effective method for different localization of pulmonary and mediastinum le-sions and provides useful diagnostic information in the majority of cases. It repre-sents a valid alternative to more invasive techniques as thoracoscopy andmediastinoscopy.

B-298 11:33

Does non-sterile coupling gel used prior to disinfection process inultrasound guided percutaneous biopsy contribute to bacterial growth atthe biopsy site?K. Gurel1, O. Karabay1, S. Gurel1, C.F. Hildebolt2; 1Bolu/TR, 2St. Louis, MO/US([email protected])

Purpose: The aim of this study is to determine the extent to which non-sterilecoupling gel, used prior to disinfection process of US-guided percutaneous biop-sies, results in bacterial contamination of the biopsy site.Method and Materials: Patients referred for US-guided percutaneous biopsieswere included in this study. Transmission material used for sonographic evalua-tion before the disinfection process was either non-sterile coupling gel or sterilesaline. Patients were randomly assigned to two groups: gel (n = 30) or saline(n = 30). Specimens were obtained from the skin within a 10 cm2 area at the biop-sy site before using transmission material for sonographic evaluation and afterthe disinfection process. Specimens were also obtained from the coupling gel,saline and povidine-iodine. Inoculated specimen plates were incubated at 37º Cin aerobic conditions and the numbers of colony forming units were recorded.Nominal logistic regression analysis was used to calculate the odds of post-dis-infection bacterial growth based on group, gender, coincidental disease, biopsysite and specific factors (skin fold, hairy skin).Results: Controlling for the other factors, there was no significant difference inpost-disinfection growth between the gel and saline groups (p = 0.10). Biopsysite and specific factors did, however, significantly affect bacterial growth (p < 0.05).No bacterial growth resulted from samples obtained from the coupling gel, povi-dine-iodine, nor saline.Conclusion: Non-sterile coupling gel used prior to percutaneous biopsy doesnot significantly contribute to bacterial growth.

B-299 11:42

MR-guided biopsies in a low field system: An alternative option to CT-guided interventions?S. Zangos, K. Eichler, M.G. Mack, C. Herzog, R. Hammerstingl, P. Reichel,T.J. Vogl; Frankfurt/DE ([email protected])

Purpose: To evaluate the safety and precision of MR-guided biopsy in an openlow field system for different organs.Methods and Materials: In 357 patients (mean age 64.6 years, female = 104,male = 253) with unclear tumors MR-guided biopsies were performed in a low

Scientific Sessions

204 C D E FBA G

field system (0.2 T, Magnetom Concerto, Siemens). The biopsies were preformedin different organ systems (135 liver, 42 abdomen, 36 retroperitoneum, 86 pros-tate gland, 9 neck, 15 musculoskeletal and 34 other organs). For the monitoringof the procedure T1-weighted FLASH sequences (TR/TE=100/9; FA= 70°) in allpatients and additional different sequences were used. After positioning of theneedle tip in the tumors repeated biopsy specimens were acquired with cuttingneedles (Somatex) during the interventional procedures using coaxial technique.The visibility of the needles and the precisions of the biopsy procedure wereevaluated. Peri- and post-interventional complications were observed.Results: The size of the biopsied lesions ranged from1 to 15 cm. Adequate spec-imens for histological interpretation were obtained in 343 cases (96 %). In 14patients (4 %) the biopsy results were non-specific. Follow-up showed that thelesions were missed in 25 patients (7 %). The possibility of free slice orientationduring the biopsy facilitated the interventional procedures.The different organs, the tumors and vessels were determined safe and all biop-sies were performed without major vascular or organ injuries.Conclusion: Our results showed that MR-guided biopsy can be performed safe-ly and precisely and appears to be an alternative to CT-guided biopsy.

B-300 11:51 !Minimally invasive autopsy: Post-mortem navigated tissue sampling incomputed tomographyE. Aghayev, A. Christe, C. Jackowski, K. Yen, R. Dirnhofer, P. Vock, M. Thali;Berne/CH ([email protected])

Purpose: Post-mortem MSCT and MR imaging is being prospectively evaluatedin correlation with autopsy. Excellent visualisation of skeletal (MSCT) and softtissue (MRI) injuries and pathology as well as application of angiographic tech-niques has been recently demonstrated. Non-invasive full-body cross-sectionalimaging is so far insufficient for microscopic tissue analysis.Non-targeted needle autopsies have already been reported. We attempted toimplement navigated targeted post-mortem tissue sampling to enhance the po-tential of virtual autopsy.Methods and Materials: Clinically approved ACN-III biopsy core needles (14gauge x 16") with automatic pistol device (Bard Magnum, Medical Device Tech-nologies, Denmark) were navigated and tested in MSCT (Emotion 6, Siemens,Germany) using a navigation system (MEM Center/ISTB Medical ApplicationFramework, Berne, Switzerland) on ex-vivo porcine models and subsequentlyfive human bodies.Results: The navigation system enabled accurate placement of the needles withinthe brain, heart, lung, liver, spleen and muscle tissue. The ACN-III needles al-lowed for sampling ("biopsy") of tissue cores with a mean width of 1.7 mm (range1.2-2 mm) and a maximal length of 20 mm at all locations. The obtained tissuespecimens were of sufficient size and adequate quality for histological analysis.Conclusion: Post-mortem navigated "biopsy" can reliably provide tissue sam-ples from different locations of the body. The tissue cores sampled using ACN-IIIneedles seem to be adequate for microscopic tissue analysis.The minimally invasive navigated tissue sampling might significantly improve theinformation from post-mortem imaging alone in cases that, for whatever reason,are not amenable to autopsy.

10:30 - 12:00 Room K

Genitourinary

SS 607Lower urinary tract, pelvic floor and testesModerators:R. Ambrosini; Novara/ITG. Rottenberg; London/UK

B-301 10:30

Virtual cystoscopy with 64-slice MDCT in the detection of bladderneoplasms smaller than 10 mmV. Panebianco, M. Anzidei, F. Iori, M. Ciccariello, V. Vergari, M. Petroni,R. Rocca, R. Passariello; Rome/IT ([email protected])

Purpose: To assess the value of 64-MDCT cystography and virtual cystoscopyin the detection of bladder neoplasms smaller than 10 mm.Methods and Materials: Fifteen patients referred for hematuria, or follow-upsafter tumor resection, underwent examination of the pelvis on a 64-MDCT (Sen-sation 64, Siemens, Germany) with the following protocol: 50 mAs, 120 KV, thick-ness 1 mm, collimation 0.6, recon increment 0.6, kernel B30 smooth. Using a

Foley catheter the bladder was completely emptied of urine and distended withapproximately 350-400 cc of room air. Subjects were examined in supine andprone position. In patients with lesions larger than 1 cm, an additional scan wasperformed in the prone position after contrast injection. Datasets were download-ed to a dedicated workstation (Vitrea 3.7 Vital Images, Plymouth, USA) for imageanalysis. CT results were compared with those of conventional cystoscopy per-formed 1 week later.Results: Virtual cystoscopy showed 27 lesions, whilst conventional cystoscopycorrectly visualized 28. Twenty-three lesions were in the range of interest (1-10 mm). A 3 mm lesion situated on the bladder dome, in a patient with poor blad-der distension, was missed. Overall CT sensitivity for lesion detection was 91%with a specificity of 100%.Conclusion: Virtual cystoscopy is a non-invasive and accurate technique for thedetection of bladder lesions smaller than 1 cm. The attenuation gradient betweenair and bladder wall allows a reduction of tube current, lowering patient exposure.Conventional cystoscopy is still required for biopsy and lesion removal.

B-302 10:39

MR imaging use of flair sequences for detection and local staging ofbladder tumors with MR imagingM. Di Girolamo, B. Sessa, S. Annibali, G. Grauso, I. Pochesci, V. David;Rome/IT ([email protected])

Purpose: To increase the diagnostic accuracy of MR imaging in the detectionand local staging of bladder tumors by using Fluid-attenuated Inversion Recov-ery (FLAIR) sequences.Methods and Materials: 32 patients with bladder tumors detected by US under-went MR imaging using 0.5 and 1.5 T superconductive magnet. We performedSE T1-weighted (TR:500 ms, TE:30 ms), TSE T2-weighted (TR:2500 ms,TE:120 ms, ETL:40) and FLAIR-sequences (TR:6000 ms, TE:150 ms, TI:2000 ms,N.Ex.:4; Acq.time:7'30") on axial scans. The contrast to lesion ratio was evaluat-ed in all sequences. All the patients underwent cystoscopy with transurethralbiopsy and 14 had subsequent cystectomy.Results: In comparison with other sequences, FLAIR-sequence was more sen-sitive in the detection of bladder neoplasms. This sequence demonstrates thehyperintense signal of bladder neoplasms from the filled bladder lumen with nosignal. The sensitivity in the identification of bladder neoplasms was 100% withFLAIR sequences, 89.6% with TSE T2-weighted sequences and 86.2% with SET1-weighted sequences. That was due to the higher signal to lesion ratio of theFLAIR sequences in comparison with the others. In fact on FLAIR-sequences themean value of contrast to lesion ratio of bladder neoplasm was 33.1 while on SET1-weighted sequences and TSE T2-weighted sequences was respectively 15.2and 29.2. FLAIR-sequences allowed the detection of small papillomas (less than2 mm). TSE T2-weighted sequences were more sensitive than other sequencesin the study of bladder wall infiltration.Conclusion: FLAIR-sequences were more sensitive than others in the detectionof bladder neoplasms, thanks to their higher contrast to lesion ratio and can bevery helpful in the visualization of small papillomas, especially when multifocal.

B-303 10:48

Detection of vesico-ureteral reflux with MR cystography using dynamicbolus tracking technique: Preliminary resultH.S. Teh1, B.M. Yeh2, S.J.J. Gan1, F.C. Ng1; 1Singapore/SG, 2San Francisco,CA/US ([email protected])

Purpose: To evaluate the feasibility of using MR as an imaging technique fordetection of vesico-ureteric reflux (VUR).Methods and Materials: Study received institutional review board approval, andwritten informed consent was obtained from all patients. Six patients (age range,24 - 35 years, 5 female and 1 male) with VUR detected on fluoroscopic voidingcystourethrography (VCUG) were recruited into the study. One patient was ex-cluded when MR imaging revealed an intra-uterine pregnancy. These yielded 10kidney ureter units (five were normal, and 5 had VUR detected on VCUG). MRcystography was performed using a 1.5 T MR scanner. Patients were catheter-ized and gadolinium enhanced saline was infused into the bladder. A gadoliniumdetection pulse sequence technique was used to allow real time visualization ofthe ureters as a means to detect VUR. When the contrast was seen refluxing intothe ureter, a three-dimensional (3D) MR gradient echo sequence was triggered.The result was compared to VCUG.Results: Findings at MR cystography and VCUG were concordant in 8 (Fournormal; 4 with VUR) of the 10 kidney ureter units (80%). There was discordancebetween the two techniques in the remaining 2 kidney ureter units; one had VURdetected on MR cystography only and the other on VCUG alone. There was alsogood concordance for the extent of hydronephrosis.

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 205C D E FA G

Conclusion: Dynamic MR cystography is a novel and feasible technique for func-tional and anatomical evaluation of VUR without exposure of the patient to ioniz-ing radiation. This novel technique may potentially be of value as a screeningexamination.

B-304 10:57

Voiding MR-cystourethrography: A new diagnostic imaging technique forevaluation of the male lower urinary tractM. Di Girolamo, B. Sessa, M. Molinari, A. Rebonato, Y. Puce, V. David; Rome/IT([email protected])

Purpose: To evaluate the diagnostic accuracy of voiding MR cystourethrographyin visualization of the male urethra.Methods and Materials: 10 volunteers and 86 patients underwent voiding MRcystourethrography. The MR examination was performed with a 1.5 Tesla super-conductive magnet. The urinary bladder was filled with paramagnetic contrastagent by administering 20 mg furosemide followed by ¾ of the normal dose of aparamagnetic contrast agent. When the bladder was full, the patient was askedto urinate. An urisheath was placed on the penis in order to collect urine. Duringmicturition T1-weighted spoiled 3D-GE acquisitions in the sagittal plane wereperformed (TR: 12 ms; TE: 2.7 ms; flip-angle: 40°; slice thickness: 2 mm; acquisi-tion time: 12 s). Two consecutive acquisitions were performed and the 3D imageswere post processed with a MIP algorithm. In 10 patients we also performedurethral virtual endoscopy using VR-technique.Results: 10 patients were unable to perform the MR examination. In other caseswe obtained homogeneous opacification of the bladder lumen and perfect evalu-ation of the male urethra. The visualization of the urethra with MIP reconstructedimages was considered comparable to conventional cystourethrography. We de-tected 24 cases of bladder neck obstruction, 36 cases of urethral strictures, 2urethral papillomatosis and 4 cases of benign prostatic hypertrophy. The site,length and the number of urethral strictures were accurately determined. Theanalysis of 3D sagittal scans and urethral virtual endoscopy allowed a betterevaluation of the morphology of the urethral strictures compared with conven-tional cystourethrography.Conclusion: Voiding MR cystourethrography can replace standard retrogradeand micturating cystourethrograms. This novel technique avoids both radiationexposure to the gonads and urinary catheterisation.

B-305 11:06

Transurethral ultrasonography guided injection of adult autologous stemcells versus transurethral endoscopic injection of collagen in treatment ofurinary incontinence: A randomised trialF. Frauscher, L. Pallwein, R. Marksteiner, H. Strasser, A. Klauser,D. Zur Nedden; Innsbruck/AT ([email protected])

Purpose: We investigated transurethral ultrasonography guided injection of adultautologous stem cells versus transurethral endoscopic injection of collagen intreatment of urinary incontinence in a randomised trial.Methods and Materials: Between September 2002 and October 2004, 42 wom-en and 21 men suffering from urinary stress incontinence (age: 36-84 years)were treated with transurethral ultrasonography guided injections of autologousmyoblasts and fibroblasts obtained from skeletal muscle biopsies. Fibroblastswere injected into the urethral submucosa, while myoblasts were implanted intothe rhabdosphincter. In parallel, 7 men and 21 women (age: 31-84 years) alsodiagnosed with urinary stress incontinence were treated with standard transure-thral endoscopic injections of collagen. Patients were randomly assigned to bothgroups according to capacity in the cell culture laboratory. A defined incontinencescore, urodynamic parameters, quality of life and morphology and function ofurethra and rhabdosphincter were evaluated pre- and postoperatively.Results: After a follow-up of 6 to 30 months incontinence was cured in 39 womenand 11 men after injection of autologous myoblasts and fibroblasts. Mean qualityof life score (51.38 preoperatively, 104.06 postoperatively; p-value < 0.001), thick-ness of urethra and rhabdosphincter (2.103 mm preoperatively, 3.303 mm post-operatively; p-value < 0.001) as well as contractility of the rhabdosphincter(0.56 mm preoperatively, 1.462 mm postoperatively; p-value < 0.001) were im-proved postoperatively. Only 2 patients treated with injections of collagen werecured of incontinence.Conclusion: Our results demonstrate that, in contrast to injections of collagen,urinary incontinence can be treated effectively with ultrasonography guided in-jections of autologous myoblasts and fibroblasts.

B-306 11:15

Evaluation of pelvic floor disorders: Comparison between dynamic MRimaging of the pelvic floor and conventional defecographyF. Maccioni, S. Parlanti, F. Siliquini, G. Luppi, F. Carrozzo, M. Marini; Rome/IT([email protected])

Purpose: To evaluate the role of dynamic MR imaging in pelvic floor disorders, incomparison with conventional defecography. As it is impossible to reproduce areal defecation in a supine position in a MR unit, we propose a dynamic MRtechnique performed without defecation, by distending the rectum with air.Methods and Materials: Fifty-five patients with evidence of posterior pelvic floordisorders underwent conventional defecography and dynamic MR imaging in asupine position. After rectal air insufflation, the Foley catheter was the benchmark of the ano-rectal junction. T2-weighted HASTE sequences were acquiredat rest, contraction and straining in axial and sagittal planes. A dynamic studywas performed using a modified T2-weighted HASTE sequence (1image/sec-ond) in rest, squeeze and straining phases. Conventional defecography was con-sidered the gold-standard. We adopted the same standard reference lines andangles of conventional defecography, as well as modified values specific for MRimaging, obtained from a control group (16 normal volunteers). MRI images wereanalyzed by two radiologists blind of the results of conventional defecography.Results: MR vs defecography identified 26/28 rectoceles (86%), 23/28 (82%)ano-rectal junction descents, 5/9 invaginations (56%), 6/6 dyskinetic pubo-recta-lis syndrome; 4 enteroceles, 9 uterine descents, 10 cystoceles, undetected orunderestimated at conventional defecography.Conclusion: MR imaging of the pelvic floor is a rapid and well tolerated examina-tion. In comparison with defecography, it was valuable in the diagnosis of func-tional and pathological pelvic floor disorders, with a tendency to underestimateinvaginations. On the other hand it provided additional information on the sphinc-ters, levator ani, puborectalis muscles, uterine prolapses, cystoceles and ente-roceles, underestimated at conventional defecography.

B-307 11:24

MR imaging in perineal fistulasM.E. López Puchau, V. Belloch Ugarte, R. Broseta Torres, F. Mata Escolano,M.J. Romero Gómez; Valencia/ES ([email protected])

Purpose: 1. To illustrate the normal anatomy in MR imaging of the perineum. 2.To correlate the perianal fistula classification with MR imaging.Methods and Material: 24 patients underwent pelvic MR imaging (1.5 T) per-formed with phased-array coil. Images were obtained in the sagittal (T1, T1 +Gd), axial (TSE PD fat sat, high resolution TSE T2, in phase Gr T1 fat sat, T1 +Gd) and coronal (STIR and T1, T1 + Gd) planes.Results: The most frequent cause of perineal fistulae (19), was as a result ofinfective processes: secondary to surgery (urethral diverticulectomy, hysterecto-my, colectomy, vaginal surgery) and to inflammatory gastrointestinal conditions(Crohn's disease). Only in 5 patients were fistulae caused by tumors and radio-therapy (cervix cancer). MR imaging's good contrast and spatial resolution, inconjunction with direct image acquisition in any plane, permit us to identify high-signal intensity tracts in transverse T2-weighted images and coronal STIR, where-as the enhancing sinus tract walls are demonstrated on the T1-weighted fat-satpostgadolinium images. In addition MR imaging is able to evaluate the relation-ship of fistulas to the levator ani muscle and soft tissue.Conclusion: MR imaging is the diagnostic imaging technique of choice in thestudy of, surgical planning, follow-up and post treatment evaluation for patientswith perianal fistulae.

B-308 11:33

Testicular microlithiasis: A prospective study. Is ultrasound follow-upnecessary?L. Siddique, J. Richenberg; Brighton/UK ([email protected])

Purpose: Testicular microlithiasis (TML) has been reported to have a high asso-ciation with testicular cancer. Our aim in this study is to show that following upTML patients with ultrasound on a yearly basis to monitor development of Testic-ular Cancer is unnecessary.Methods and Materials: In Brighton and Sussex University Hospitals NHS Trust,2656 men were referred for scrotal ultrasound over the period January 2000 toDecember 2004. Of the 2656 men, 51 had TML. The 51 men were followed up bya single radiologist on a recommended yearly basis. The results were recordedas TML grade at presentation, unilateral or bilateral, any adjunct pathology foundon the scan and whether there was any change on follow-up, especially whetherany developed testicular cancer.

Scientific Sessions

206 C D E FBA G

Results: 51 out of 2656 men (1.92%) had TML on presentation. Their age rangeat presentation was 15 to 83 years (mean 41 years). The number of ultrasoundsperformed on each patient ranged from 1 to 8 (mean 3.33). They were followedup between 3 to 76 months (mean 33.29 months). During the time period offollow-up none developed testicular cancer.Conclusion: This prospective study has shown that regular sonographic surveil-lance is not required for patients with TML.

B-309 11:42 !The prevalence of testicular microlithiasis in an asymptomatic populationS. Serter, M. Ünlü, Ö. Tunçyürek, B. Gümüs, S. Tarhan, V. Ayyildiz, Y. Pabusçu;Manisa/TR ([email protected])

Purpose: Testicular microlithiasis (TM) is a rare, usually asymptomatic diseaseof the testes associated with various genetic anomalies and infertility. The aimsof this prospective study are to determine the true prevalence of TM in an asymp-tomatic population by ultrasound screening and to quantify whether there areany Doppler ultrasonographic (DU) findings specific to TM.Methods and Material: 2179 healthy male volunteers 17-42 years old, were in-cluded in the study. A screening scrotal US was performed. All men diagnosedwith TM underwent a complete clinical evaluation, physical examinations, deter-mination of tumor markers and scrotal DU. DU was performed for subjects withTM and intratesticular max velocities, min velocities and resistive indices weredetermined. Scrotal DU was also performed in 50 subjects without TM for a con-trol group.Results: 53 men with TM were identified in the 2179 ultrasound scans, giving anincidence of 2.4% for TM in this asymptomatic population. There were no signif-icant differences regarding the Doppler parameters and spectral examinationsbetween subjects with or without testicular microlithiasis.Conclusion: Our results demonstrate the prevalence of testicular microlithiasison ultrasound screening as 2.2% in 2179 healthy men with a mean age of 24years. Additionally our study shows there are not any effects on Doppler param-eters by microlithiasis in TM.

B-310 11:51

Testis magnetic resonance spectroscopy: Preliminary findingsA.K. Firat, M. Ugras, H.M. Karakas, B. Kahraman, C. Baydinc; Malatya/TR([email protected])

Purpose: To evaluate the magnetic resonance spectroscopy patterns of testes inprepubertal and postpubertal males.Methods and Material: Fourteen patients who were referred to the urology de-partment with various symptoms and signs were included in the study group. Theconventional MR imaging sequences of all revealed normal findings. 22 MRS oftestes were obtained from the remaining 12 patients. There were 6 adults (age:19-35 years, mean age: 26 years) and 6 pediatric patients (age: 9-14 years, meanage: 12 year). Lipid and choline peaks were measured and choline/lipid ratioswere calculated. Mann Whitney-U test was used for statistical analysis.Results: After all MR imaging and spectroscopy findings were obtained, theywere evaluated by two radiologists. Two peaks were prominent and diagnostic:choline (3.21 ppm) and free lipid peak (1-1.5 ppm). There were statistically sig-nificant differences between adult and pediatric age groups (p < 0.001). Choline/lipid ratios were ranging between 0.47-4.20 (Mean: 1.64) in adults and between0.06-0.94 (Mean: 0.32) in the pediatric age group. Blood testosterone levels wereevaluated in both study groups.Conclusion: Choline/lipid ratios were high in adults and low in the pediatric agegroup. This statistically significant difference represents the increasing amount ofspermatogenesis in adults. We recommend using choline/lipid ratio for the evalu-ation of unilateral testicular spermatogenesis. To our knowledge, there was norelevant study in the literature dealing with the normal MR spectroscopy patternof testes. This article will provide the preliminary findings and be the initiator ofMRS studies of testes with various pathologies.

10:30 - 12:00 Room L/M

Neuro

SS 611AVMs and aneurysms: Diagnosis and treatmentModerators:U. Ernemann; Tübingen/DEB.F. Schuknecht; Zurich/CH

B-311 10:30

Can individual brain AVM factors predict nidus obliteration afterstereotactic radiosurgery?S. Nagaraja, K.J. Lee, S. Coley, D. Capener, I.D. Wilkinson, P.D. Griffiths;Sheffield/UK ([email protected])

Purpose: There is limited data available in the monitoring of brain arteriovenousmalformations (AVM) after stereotactic radiosurgery (STRS) due to the need forrepeated catheter angiography. Using a combination of static and dynamic MRangiographic sequences we have studied the rate of AVM obliteration at 12 monthsand correlated this with baseline AVM characteristics.Methods and Materials: Forty patients (41 AVMs) consented to a dedicated 1.5 TMR protocol 12 months after receiving STRS for a brain AVM. CE-SLINKY (stat-ic) and MR-DSA (dynamic) were performed. Nidal volumes were calculated us-ing CE-SLINKY data in patients with a persisting arteriovenous shunt. The rate ofchange of nidus volume was computed by statistical methods using the planningangiographic data and the MR angiographic data. The AVM factors including 1)maximum linear dimension, 2) nidal volume, 3) AVM location, 4) nidal morpholo-gy, 5) venous drainage, 6) "high-flow angiographic change" and 7) prior emboli-zation were used in the statistical analysis to determine their role in nidusobliteration.Results: Complete nidal obliteration was found in 9 patients, 26 showed greaterthan 50% nidal reduction and 6 had less than 50%. Without Bonferroni correctionfor multiple comparisons, only AVM venous drainage (p = 0.018) and AVM loca-tion (p = 0.032) were significantly correlated with rate of obliteration (Fisher'sExact Test).Conclusion: Certain AVMs show a rapid and complete response to STRS; earlyidentification of limited response is important in order to consider further treat-ment options. Our data suggests AVM venous drainage and AVM location arepossibly the only significant factors.

B-312 10:39

Functional network reorganization in cerebral AVMs: A functional MRimaging/DTI studyJ.L. Ulmer, C.V. Salvan, L. Hacein-Bey, E.A. DeYoe, D.L. Daniels, R.W. Prost;Milwaukee, WI/US ([email protected])

Purpose: To determine the relationship between functional MR imaging corticalorganization and DTI white matter architecture in patients with AVMs; to predictfunctional network reorganization and to improve preoperative planning.Methods and Materials: BOLD functional MR imaging and DTI data were ob-tained in 7 patients (14-60y; 4M, 3 F) presenting with AVMs. Functional sensori-motor, language and visual systems were tested. A cross-correlation analysis ofthe fMRI data was used, with a p < 0.01. Supraselective WADA testing was per-formed in 5 patients for functional assessments. Fractional anisotropy (FA) anddirection sensitive colour-coded FA maps were used to study white matter tracts.Results: Four of the seven patients showed functional MR imaging activationpatterns within language, sensori-motor or visual cortex that were consistent withcortical reorganization. Altered architecture of white matter tracts was demon-strated in each of these four patients. Cortical reorganization was supported,though not proven, by WADA testing. In one case clear white matter tract abnor-malities were identified without altered cortical organization, while two patientsshowed no cortical reorganization or altered white matter architecture.Conclusion: There is high association between cortical reorganization detectedwith functional MR imaging and altered architecture of white matter tracts detect-ed with DTI in AVMs, with impact on preoperative planning.

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 207C D E FA G

B-313 10:48

Integration of 3D MR-DSA using keyhole sampling techniques into theradio-surgical treatment of arteriovenous malformations (AVM) by gammaknifeC.A. Taschner, V. Le Thuc, N. Reyns, S. Blond, X. Leclerc, J.-P. Pruvo; Lille/FR([email protected])

Purpose: The definition of target areas in radio-surgical treatment of cerebralAVMs by gamma-knife is currently based on MR, CT and conventional catheterangiography data. The purpose of our research work was to evaluate the feasibil-ity of integrating a 3D MR-DSA sequence with stochastic k-space acquisition(CENTRA Keyhole) into a treatment planning system for gamma-knife therapy inpatients bearing cerebral AVMs.Methods and Materials: Time resolved MR-DSA was performed under treat-ment conditions with patients carrying a stereotactic frame, on a 1.5 T MR sys-tem (Achieva, Philips) using a circular head coil. The imaging protocol consists ofa contrast enhanced 3D fast field-echo sequence with CENTRA Keyhole (18%Keyhole) acquisition. Post-processing of the source images with the data visual-ization software (IDL, Research Systems Inc). included definition of the outline ofthe head on non-subtracted MR-DSA images. The image subtraction was re-stricted to the defined areas of the image in order to maintain the MR markersnecessary for the gamma-knife planning system (GammaPlan, Elekta).Results: The proposed MR sequence yields an in-plane resolution of 1 x 1 x1.8 mm covering an imaging slab of 126 mm at a temporal resolution of 1.8s pervolume. The integration of imaging data into the GammaPlan allows for directcomparison between projection MR-DSA data and conventional digital subtrac-tion angiography data with regard to the radiosurgical target volume and its pre-cise anatomical localisation.Conclusion: The integration of 3D MR-DSA into a treatment planning system forgamma-knife therapy in patients bearing cerebral AVMs is feasible.

B-314 10:57

Cavernous sinus dural arteriovenous fistula with an intracranial lesioncaused by retrograde leptomeningeal venous drainageS. Takahashi, I. Sakuma, T. Otani, K. Yasuda, N. Tomura, J. Watarai;Akita City/JP ([email protected])

Purpose: To evaluate venous drainage patterns of cavernous sinus dural arteri-ovenous fistula (CSDAVF) with an intracranial lesion seen on MR imaging.Methods and Materials: DSA and MR imaging findings of 18 patients (4 malesand 14 females, age range 38-77 years, mean age 63.8 years) with angiograph-ically proven CSDAVF were retrospectively reviewed to evaluate the relationshipbetween retrograde leptomeningeal venous drainage (RLVD) and abnormal sig-nal intensity of the brain parenchyma on axial T1-weighted images before andafter intravenous administration of Gd-DTPA and T2-weighted fast spin-echo im-ages (T2WI).Results: RLVD was identified on DSA in 10 of 18 patients (55.6%), includingdrainage into the superficial middle cerebral vein (n = 5), uncal vein (n = 3), pet-rosal vein (n = 3), lateral mesencephalic vein (n = 1) and anterior pontomesen-cephalic vein (n = 2). In 4 of 18 patients (22.2%), T2WI revealed abnormalhyperintense lesions in the region with RLVD. Of these 4 patients, one presentedwith putaminal hemorrhage, while the other three presented with hyperintensityof the pons or the middle cerebellar peduncle on T2WI, reflecting venous con-gestion, and 2 presented with enhancing lesions of the pons on the contraststudy. All of these 4 patients showed concentration of venous outflow from cav-ernous sinus into RLVD and other venous drainage routes were obliterated. Thepatients without abnormal signal intensity of the brain parenchyma on MR imag-ing had other patent venous drainage routes in addition to RLVD.Conclusion: Brain parenchymal hyperintensity on T2WI or the contrast enhance-ment in CSDAVF is a crucial finding that indicates concentration of venous drain-age into RLVD, causing venous congestion or parenchymal hemorrhage.

B-315 11:06

Predicting successful coil embolization of intracranial aneurysms with 16-row multidetector CTAK. Papke, D. Sauner, M. Schlunz, C. Haupt, M. Fruth, A. Bani, F. Brassel;Duisburg/DE ([email protected])

Purpose: To assess the feasibility of deciding on the basis of 16-row multidetec-tor CTA with advanced postprocessing techniques whether endovascular coilembolization may be successfully performed in an intracranial aneurysm.Methods and Materials: 89 patients with suspected subarachnoid hemorrhageunderwent 16-row MDCTA and DSA. For each aneurysm diagnosed by this com-

bined standard of reference, two neuroradiologists assessed the MDCTA datasets using thin-slice multiplanar reformats, thin-slab MIP and volume renderingtechniques. They rated in consensus whether simple endovascular coil place-ment was considered possible (i.e. without assisting techniques such as remod-elling or stenting), based on the morphology of the aneurysm neck and on thepresence of vessels originating from the aneurysm. The ratings were comparedto the actual therapy results in each aneurysm.Results: Out of 84 aneurysms diagnosed per standard of reference, 72 wereconsidered as potentially ruptured and amenable to acute treatment. 49 aneu-rysms were successfully coiled without stenting, 5 required additional stenting, 9were primarily clipped, 5 were clipped after unsuccessful coiling attempts, 4 re-ceived no further therapy after unsuccessful coiling attempts. The sensitivity, spe-cificity, PPV and NPV of MDCTA for successful coil embolization is 94%, 92%,96% and 88%.Conclusion: 16-row MDCTA with advanced postprocessing techniques is highlysensitive in the identification of intracranial aneurysms that can be successfullytreated by endovascular coil embolization.

B-316 11:15

Post-embolization growth of cerebral aneurysms: Incidence and risk factorsM. Han, H.-S. Kang, B. Kwon; Seoul/KR ([email protected])

Purpose: Although aneurysmal recanalization after coil embolization is a well-known phenomenon, information on the enlargement of embolized aneurysms islimited. The purpose of this study was to evaluate the fate of embolized aneu-rysms with a focus on the incidence and risk factors of aneurysmal growth aftercoil embolization.Methods: A total of 529 patients harboring 591 intracranial aneurysms were treat-ed by endovascular coil embolization from 1992 to 2004. We reviewed 586 angi-ographic studies of 214 cases who underwent follow-up angiographic studies.Radiological and clinical data were reviewed to identify the risk factors for en-largement of aneurysms after embolization.Results: Thirty-four cases showed enlargement of embolized aneurysms, andall of them showed major recanalization as well. Aneurysms showing minor reca-nalization revealed no increase in their sizes on the follow-up angiographic stud-ies. The incidence of aneurysmal growth was 81% among cases with majoraneurysmal recanalization, 35% among cases with any aneurysmal recanaliza-tion, and 8.5% among cases in which follow-up imaging studies were performed.The location of aneurysm at the terminal bifurcation was an important risk factorfor aneurysm enlargement at the time of major recanalization. All bifurcation an-eurysms except one showed enlargement in the presence of major recanaliza-tion. Older age and presence of a secondary pouch were found to be related tono growth of recanalized aneurysms.Conclusion: Aneurysm enlargement after coil embolization is common in thepresence of major recanalization. Our data suggest that bifurcation aneurysmshave a higher risk of enlargement after coil embolization when they are recanal-ized.

B-317 11:24

The role of micro-Doppler ultrasound in aneurysm clipping intraoperativelyE. Kapsalaki1, K. Fountas2, I. Fezoulidis1, P. Toulas3, J.S. Robinson2;1Larissa/GR, 2Macon, GA/US, 3Athens/GR ([email protected])

Purpose: We present our experience with the utilization of micro-Doppler ultra-sonography during aneurysmal clipping.Methods and Materials: In a retrospective study, 54 patients surgically treatedfor 69 intracranial aneurysms were studied. Five patients harboring 7 aneurysmswere surgically treated on an elective basis while 49 patients acutely presentedwith subarachnoid hemorrhage. Nine patients (9 aneurysms) had a posterior cir-culation aneurysm while the remaining harbored anterior circulation aneurysms.Blood flow velocities of the parent, adjacent vessels and aneurysmal sac weremeasured by using a Consorma shapable micro-Doppler flow probe (Cook Vas-cular, Inc., Leechburg, PA, USA).Results: In 11 aneurysms (15.9%) there was decreased or absent flow in theparent vessel or on one of the adjacent vessels after clipping. The clip was repo-sitioned and in 9 of those the blood flow in the previously compromised vesselwas restored. In the remaining 2 cases repositioning of the clip could not improvethe MDU findings. Interestingly, in 9 aneurysms (13%), MDU demonstrated flowthrough the aneurysmal dome although the aneurysmal neck appeared to betotally obliterated. After repositioning of the applied clip in 7 cases or applicationof a booster clip in 2 cases, the obtained MDU showed no blood flow.Conclusion: MDU seems to be a non-invasive, reliable alternative methodologyto intra-operative angiography. This inexpensive method might lend itself to the

Scientific Sessions

208 C D E FBA G

routine usage of aneurysm clipping for confirming patency of the parent and ad-jacent vessels, as well as complete obliteration of the aneurysm.

B-318 11:33

Comparison of perfusion in patients with a perimesencephalic hemorrhageand patients with aneurysmal subarachnoid hemorrhageI. van der Schaaf, M. Wermer, B. Velthuis, G. Rinkel; Utrecht/NL([email protected])

Purpose: The origin of the bleeding in perimesencephalic hemorrhage (PMH) isstill unknown but clinical and radiological characteristics suggest a venous source.If PMH is of venous origin, less elevation of the intra-cranial pressure and lessperfusion deficits will occur than in aneurysmal subarachnoid hemorrhage (aSAH).We compared perfusion in the acute stage after subarachnoid hemorrhage (SAH)in patients with PMH and patients with aSAH.Methods and Materials: Patients scanned within 72 hours after SAH were in-cluded. Patients with aSAH were matched to patients with PMH on clinical condi-tion on admission. Cerebral blood flow (CBF) was assessed in eight predefinedregions of interest. Differences in CBF values and 95% confidence intervals werecalculated. An additional analysis was performed in which patients with aSAHwere also matched on amount of cisternal blood and age.Results: We included 11 patients with PMH and 24 patients with aSAH. CBFvalues were statistically significantly higher in patients with PMH (mean CBF:72.7) compared to aSAH patients (mean CBF: 65.1); difference of means -7.6 (-1.2 to -14.0). For the additional analysis after matching for amount of blood andage, 11 aSAH patients were included. Again CBF values were statistically signif-icantly higher in patients with PMH than in aSAH patients (mean CBF: 64.9);difference of means: -7.8 (-0.3 to -15.2).Conclusion: Patients with PMH have a higher CBF than patients with aSAH. Thefindings of this study offer further support for a venous origin of the bleeding inpatients with PMH.

B-319 11:42

Prognostic value of cerebral perfusion CT in the acute stage after SAH forthe development of delayed cerebral ischemiaI. van der Schaaf, M. Wermer, Y. van der Graaf, B. Velthuis, C. van de Kraats,G. Rinkel; Utrecht/NL ([email protected])

Purpose: Delayed cerebral ischemia (DCI) is an important cause of death anddisability after subarachnoid hemorrhage (SAH). We studied the additional prog-nostic value of brain perfusion to three established predictors (age, clinical con-dition on admission and amount of subarachnoid blood) for the development ofDCI.Methods and Materials: 69 patients scanned with perfusion CT within 72 hoursafter SAH were included. For each patient we determined cerebral blood flow(CBF) ratios of prespecified opposite regions of interest and the three estab-lished predictors. We calculated adjusted hazard ratios (HRs) for the CBF ratioand the three established predictors by means of multivariate analysis (Cox re-gression). The additional prognostic value of CBF ratios was assessed by com-paring the area under the ROC curve (AUC) of two models: One with and onewithout addition of the CBF ratio to the three established predictors.Results: The CBF ratio was an independent predictor for the development of DCI(HR: 0.63; 95% CI: 0.46 - 0.86) as was clinical condition (HR: 1.47; 1.01 - 2.13).By adding the CBF ratio to the model with the three established predictors, theAUC of the ROC curve increased from 0.76 (95% CI 0.65 - 0.89) to 0.81 (0.71 -0.91). This trend towards an increased AUC suggests an improved predictivevalue.Conclusion: The CBF ratio is an independent predictor for the development ofDCI and can contribute to a better identification of patients at high risk for DCI.

B-320 11:51

Temporomesial volume loss after subarachnoid haemorrhage:Quantification by MR imaging volumetry and correlation toneuropsychological test performanceP. Bendel, T. Koivisto, T. Hänninen, M. Könönen, A. Kolehmainen, R. Vanninen;Kuopio/FI ([email protected])

Purpose: Subarachnoid haemorrhage (SAH) leaves many patients with neuropsy-chological disturbances. Aims of this study were to assess whether: 1. SAH andthe treatment of the ruptured aneurysm is followed by volume loss in temporome-sial structures (hippocampus, HC, and amygdala, AM), detectable by MR volum-etry and 2. The neuropsychological outcome is associated with quantitativetemporomesial volume loss.

Methods and Materials: MR imaging was performed on 77 patients one yearafter SAH. The point counting method (EasyMeasure), was used to calculate theHC and AM volumes. A comprehensive neuropsychological test battery was usedto evaluate the cognitive performance of the subjects. Thirty healthy individualswere imaged as MR imaging controls.Results: In SAH patients, the normalised HC volumes (23.2/21.3, right/left,p = 0.072/0.002) and normalised AM volumes (18.4/18.7, p = 0.012 /0.045) weresignificantly smaller than in the matched control population. The AM ipsilateral tothe ruptured aneurysm was smaller in patients treated surgically (15.7) vs. en-dovascularly (20.3, p < 0.001). The hippocampal volumes correlated with a visu-al memory test and several tests of attention, flexibility of mental processing,intellectual ability and psychomotor speed. The volumes of amygdalas did notsignificantly correlate with neuropsychological test results.Conclusion: In spite of the seemingly good outcome in a majority of our patients,the present study clearly demonstrated neuropsychological impairments and re-duced temporomesial volumes one year after SAH. The mechanism of the vol-ume loss may be attributable to cerebral hypoperfusion caused by the initialbleeding, treatment procedure, or secondary vasospasm. A clear correlation wasdemonstrated between the neuropsychological defects and the reduced tempo-romesial volumes.

10:30 - 12:00 Room N/O

Breast

SS 602Angiogenesis imagingModerators:W.A. Kaiser; Jena/DER. Sinnatamby; Cambridge/UK

B-321 10:30 !Role of computed tomography laser mammography (CTLM) in theevaluation of breast lesionsA. Özdemir, I. Tunçbilek, S. Öztürk, S. Gültekin, V. Kahveci, F. Elbüken;Ankara/TR ([email protected])

Purpose: To assess the value of CTLM in the evaluation of breast lesions, todetermine the role of parenchyma type and vascular anatomy on image patternand to evaluate the impact of bilateral scanning.Methods and Materials: CTLM scans (CTLM Scanner 1020, IMDS Inc, Fl) of258 breasts in 191 patients were evaluated, including 26 patients with radiologi-cally normal breasts and 165 patients with lesions diagnosed on mammography,US-Color Doppler, and MR imaging (n:179). Two assessments were made by tworeviewers, first without, second with knowledge of lesion size, location and BI-RADS category.Results: Lesions (mean lesion size: 16 mm) were categorized as BI-RADS 2-3-4-5 (n:23, 89, 28, and 39); (41 malignant, 138 benign). Interobserver agreementswere 95.1% and 99.5% for two CTLM assessments. CTLM findings were com-pared to histopathology in BI-RADS 4-5, and to follow-up in BI-RADS 3 catego-ries. Sensitivity, specificity, ppv and npv of CTLM for the first and secondassessments were 47.5%, 69.8%, 27.5%, and 84.7% and 70%, 82.5%, 49%,and 91.9%, respectively. CTLM finding wasn't in the same quadrant with lesion in28% of true positive cases. When used as adjunct to the combined radiologicalmodalities, CTLM decreased the specificity (81% vs. 69%) and ppv (61% vs.50%). Parenchyma type and vascular anatomy were not correlated with the CTLMimage pattern. CTLM vascular patterns were asymmetric in 76% of normal breasts.Conclusion: CTLM has currently a limited value in differential diagnosis. Imagepattern does not reflect the mammographic density or vascular anatomy, andbilateral scanning isn't helpful.

B-322 10:39

The value of CT-laser-mammography (CTLM) as an adjunct tomammography and ultrasound in the diagnosis of breast cancerD. Flöry1, C.C. Riedl2, S. Jaromi1, M. Fuchsjaeger1, C. Reiner1, T.H. Helbich1;1Vienna/AT, 2New York, NY/US ([email protected])

Purpose: To assess the value of Computed Tomography Laser Mammography(CTLM) as an adjunct to mammography (MG) and ultrasound (US) in the detec-tion of malignant breast lesions.Methods and Materials: In a prospective study, CTLM scans (CTLM scannerModel 1020 by IMDS, Inc. FL), MG and US of 310 patients with 325 breast le-

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 209C D E FA G

sions were obtained. Histopathological diagnosis (biopsy and/or surgery) wasperformed in all cases. The CTLM images were read as an adjunct to MG andUS.Results: Of 325 lesions, 147 (45.2%) were malignant and 178 (54.8%) benignincluding 78 invasive ductal carcinoma (IDC), 20 invasive lobular carcinoma (ILC),40 ductal carcinoma in situ (DCIS) and nine cancers of other types. CTLM achieveda sensitivity of 65.3% (96/147 true positive) and a specifity of 53.4% (95/178 truenegative). Positive and negative predictive value were 53.6%, and 65.1% respec-tively. Sensitivity, specifity, PPV and NPV were 91.4%, 40.8%, 57.1%, and 87.3%for MG and 66.3%, 52.7%, 53.1%, and 65.0% for US.Conclusion: The present results indicate that CTLM may have potential adjunc-tive value to MG and US in the diagnosis of breast cancer.

B-323 10:48

Contrast-enhanced digital mammography: Initial results of a multi-centricstudyC. Marx1, R. Jong2, C. Dromain3, F. Diekmann4, E. Pisano5, M. Yaffe2; 1Jena/DE,2Toronto, ON/CA, 3Villejuif/FR, 4Berlin/DE, 5Chapel Hill, NC/US([email protected])

Purpose: To compare the diagnostic accuracy of contrast-enhanced digital mam-mography (CEDM) as an adjunct to X-ray-mammography (MX), with the diag-nostic accuracy of MX alone.Methods and Materials: In 4 centres (Europe and North America) 77 womenwith 90 MX-lesions underwent one-view CEDM additionally to MX, using special-ly modified digital mammography systems. Histology results were available for alllesions detected by MX (68 malignant, 22 benign). 5 radiologists interpreted theimages using high resolution softcopy workstations. Confidence of presence (5-point scale), probability of cancer (5-point scale) and BI-RADS score (MX imag-es) were evaluated for each lesion. Sensitivity, specificity and ROC curve areaswere estimated for each reader and overall.Results: 52/68 (76.5%) malignant lesions enhanced with contrast in CEDM while9/22 (41%) benign lesions did not. Overall, 12.3% vs. 8.8% of malignant and 18%vs. 11.8% of benign lesions were missed by MX vs. MX+CEDM respectively. Theaverage of sensitivity/specificity values across the readers were 0.83/0.40 forMX alone and 0.84/0.44 for MX[Unsuppor ted Character - Codename]+CEDM. Individual and overall areas under the ROC curves for MX andMX+CEDM did not differ significantly.Conclusion: CEDM resulted in fewer missed lesions than MX alone. Clinicalaccuracy did not differ significantly between MX and MX+CEDM. However, differ-ent inclusion criteria and examination techniques influenced the results. Furtherstudies including standardized examination techniques and analysis of contrastkinetics are necessary to prove the clinical utility of CEDM.

B-324 10:57

Contrast-enhanced digital mammography versus MR imaging: Initial clinicalexperienceR. Schulz-Wendtland, E. Wenkel, T. Wacker, W. Bautz; Erlangen/DE([email protected])

Purpose: To investigate the potential of using intravenous contrast material withfull-field digital mammography in comparison to MR imaging in the question ofmultifocality/-centricity at breast cancer.Methods and Materials: 17 women (age 32-58 yrs, median 42 yrs) having ab-normalities (BI-RADS 5) (diameter 9-31 mm, median 23 mm) at clinical breastexamination, sonography and digital mammography (Novation; Siemens. Expo-sure parameters: Automatic Exposure Control). For determination of possible multi-focality/-centricity, all patients received a contrast-enhanced digital mammography(CEDM). Six sequential images (Senoscan; Fischer Imaging) (same positioningprocedure but exposure parameter > 35 kV) of the affected breast were obtainedafter a contrast agent (Ultravist 370; Schering) was injected. Additionally MR im-aging (n = 17) was performed. All suspected lesions (BI-RADS 4/5) were histo-logically evaluated.Results: In 13/17 patients we found in complementary breast diagnosis as wellas in contrast-enhanced digital mammography and MR imaging only one lesion;in 3/17 multifocality only in CEDM and MR imaging; in 1/17 only in MR imaging(multicentric lesion).Conclusion: The results of this preliminary study suggest the possibility thatcontrast-enhanced digital mammography (CEDM) potentially may be useful inthe question of multifocality/-centricity at breast cancer. Further investigationsare necessary.

B-325 11:06

Contrast-enhanced three-dimensional power Doppler ultrasound in thecharacterization of tumor vascularity in solid breast massesW. Moon; Seoul/KR ([email protected])

Purpose: To evaluate the diagnostic value of 3D power Doppler (PD) US com-bined with a microbubble contrast agent in characterization of tumor vascularityin solid breast masses.Methods and Materials: Sixty consecutive patients with solid breast masses(mean size, 1.2 cm) visualized by gray-scale US, were evaluated with 2D and 3DPDUS before and after injection of the contrast agent SHU 508 A. At each USstudy, lesion vascularity (avacular, hypovascular, hypervascular) and morpholo-gy (regular vs irregular) of vessels were prospectively assessed by two radiolo-gists.Results: At unenhanced 2D PDUS, 19 of 27 (70%) cancers were vascular whereas12 of 33 (36%) benign lesions were vascular. After injection of contrast agent, allcancers (100%) were vascular whereas 20 (61%) benign lesions were vascular(P <.001). Irregular vessels were seen in 7 cancers (26%) and 1 benign lesion(3%) at unenhanced 2D PDUS and 16 cancers (59%) and 3 benign lesions (9%)at contrast-enhanced 2D PDUS (P <.0001). By using the presence of irregularvessels in the mass as the diagnostic criterion for malignancy, sensitivity, specif-icity, NPV, and PPV of 3D PDUS changed from 30% (8/27), 94% (31/33), 80% (8/10), and 62% (31/50) to 81% (22/27), 91% (30/33), 88% (22/25), and 86% (30/35), respectively after contrast agent injection.Conclusion: Compared to unenhanced 2D PDUS, contrast-enhanced 3D PDUSwas useful to demonstrate and characterize tumor vascularity in solid breastmasses.

B-326 11:15

Microbubble contrast harmonic ultrasound for differentiating breast tumors:First resultsM. Jung1, R. Kubale2, K.-P. Jungius3, J. Tacke1; 1Passau/DE, 2Pirmasens/DE,3Brig/CH ([email protected])

Purpose: To investigate whether fibroadenomas and carcinoma can be differen-tiated by injection of a new contrast agent (Optison).Methods and Materials: 50 patients with 53 impalpable breast tumors (20 fi-broadenomas, 33 carcinomas- diameter 4-15 mm) were examined by contrastenhanced ultrasound. Three of the cases had two carcinomas in the same breast.Histological confirmation was obtained in all cases. Ultrasound was performedusing a multifrequency linear probe (9-14 MHz). B-mode and tissue harmonicimaging (THI) and power Doppler imaging without and with intravenous contrastadministration were employed. Contrast behavior in the tumor was observed forover 20 min.Results: In plain ultrasound, marginal, penetrating or central vessels were foundin 24 of 53 lesions (18 carcinomas, 6 fibroadenomas). After contrast size andnumber of detectable vessels increased. Both in benign (17/20) and malignanttumors (30/33) marginal and/or intratumoral vessels were identified in the earlyphase. In the late phase a diffuse contrast accumulation was observed in 30 of33 malignant tumors. No accumulation occurred in benign tumors. For the carci-noma, the sensitivity was 93 % and the specificity 100% in the late phase com-pared with a sensitivity of 100 % and a specificity of 25% the early phase, whenall findings were combined.Conclusion: The use of Optison as contrast agent allows the detection of tumorenhancement for up to 20 min. Whereas a differentiation between fibroadenom-as and carcinomas is difficult in the early phase, contrast agent is only taken upby carcinomas in the late phase.

B-327 11:24

Benign and malignant breast lesions: Efficacy of real time contrast-enhanced US vs magnetic resonanceP. Ricci, E. Pagliara, L. Ballesio, V. Cantisani, L. Coletta, R. Passariello;Rome/IT ([email protected])

Purpose: To assess the efficacy of low mechanical index (MI) real time grayscale contrast-enhanced US (CEUS) in the differentiation of breast lesions incomparison to Magnetic Resonance (MRI).Methods and Material: 100 lesions, previously detected at mammography orconventional US, have been evaluated by means of CEUS and MR imaging. Con-trast-enhanced examinations have been performed with a dedicated equipment(Esatune, Esaote, Genoa, Italy), before and after injection of 4.8 mL of SonoVue(Bracco, Milan, Italy). MR imaging has been performed with a 1.5 T equipment(Siemens Vision Plus, Erlangen, Germany) with bilateral dedicated superficial

Scientific Sessions

210 C D E FBA G

coil, on T2w STIR and 3D Flash T1w before and 1.2,3.4,5 minutes after the ad-ministration of mdc (Gd-DTPA, 1.5 ml/Kg). Wash-in and wash-out curves havebeen assessed for both procedures. A specific US quantification software (Qon-trast, Bracco, Milan, Italy), based on pixel by pixel signal intensity over time, hasbeen used to obtain CEUS-perfusion maps for each lesion.Results: 48 invasive ductal carcinomas, 38 fibroadenomas, 8 fibrocystic dyspla-sias, 2 mucinous carcinoma, 2 invasive ducto-lobular carcinoma, and 2 phylloideswere diagnosed. Contrast-enhanced US patterns correlated well with those pro-vided by MR imaging. According to the different contrast enhancement patternsand the resulting perfusion maps, all the malignant lesions and 42 out 48 benignlesions were correctly diagnosed, with 89.5 % sensitivity and 100 % specificity.Conclusion: Contrast-enhanced US seems to be a reliable method to differenti-ate breast lesions, since it provides typical enhancement patterns. Contrast USperfusion curves well correlate with MR imaging wash in-wash out curves.

B-328 11:33

Accuracy of CT perfusion in assessing metastatic involvement of axillarylymph nodes in patients with breast cancerY. Liu1, M. Bellomi2, X. Ping1; 1Yinchuan Ningxia/CN, 2Milan/IT([email protected])

Methods and Material: 26 cases (one patient had bilateral breast cancer withbilateral axillary lymph nodes enlarged) with locally advanced breast cancer andwith clinical palpable axillary lymph nodes underwent dynamic MDCT (Lightspeed16 slice, GE: 1 scan per second for 100 at the same table position, after 40 ml i.v.contrast injection at 3.5 ml/sec. Semi-automatic calculation of perfusion parame-ters (Blood flow: BF, blood volume: BV, mean transit time: MTT and permeabilitysurface: PS) was performed. Results were compared with pathology and withHer-2/neu and Ki-67 levels in surgical specimen of the primary tumor.Results: Examined lymph nodes resulted negative in 8 cases and metastatic in18. Mean values of perfusion parameters in normal and metastatic nodes wererespectively: BF 76.18 + 16.09 and 161.6 ± 20.89 ml/sec (p <.001), BV 5.81 ± 1.28and 9.15 ± 1.54 ml (ns), MTT 6.80 ± 0.79 and 5.49 ± 0.94sec (ns), PS58.32 ± 33.03 and 25.96 ± 3.81 ml/g/sec (ns).Ki-67 levels in breast cancer, expressing primary tumor activity, was significantlycorrelated (p <.001) to PS in patients with metastatic nodes.Conclusion: Perfusion CT is an effective tool for predicting the status of axillarylymph nodes in patients with breast cancer. PS may reflect tumor activity.

B-329 11:42

Contrast enhanced MR imaging of the breast: One-side increasedvasculature as a predictor of invasive cancerA. Fausto, A. Iozzelli, L. Menicagli, F. Sardanelli; Milan/IT ([email protected])

Purpose: To test the diagnostic value of an increased ipsilateral vascular mapsas MR imaging (D-MRI) sign of breast cancer.Subjects and Methods: 204 consecutive D-MRI exams using a 1.5 T equipmententered the analysis. Seventy-two exams were excluded due to unilateral imag-ing or previous treatment. For the remaining 132, maximum intensity projections(MIPs) of the first subtracted images of a T1-weighted 3D GE sequence (coronal1-mm 128 par tit ions; TR/TE=11/4.8 ms; FA=25°; FOV=384 mm;matrix=384x192 mm; 1-mm3 voxel; 0.1 mmol/kg Gd-DOTA; 120-s time resolution;1 pre- and 4 post contrast phases) were reviewed. A one-side increased vascula-ture was reported when two or more vessels (of 30 mm or greater in length and2 mm or greater in maximal transverse diameter) was found per breast. Goldstandard was pathological examination or at least negative 1-year follow-up (fornegative exams).Results: Excluding two bilateral cancers (with symmetrical vascular maps), aone-side increased breast vascular maps was found in 46 exams: 39 malignan-cies of 22.9 ± 12.9 mm in diameter and 7 benignancies (22.6 ± 14.0 mm) at pa-thology. Vascular maps were symmetrical in 86 exams: 35 benignancies and 10malignancies (11.4 ± 3.2 mm), three of them being pure DICS (8-15 mm) at pa-thology, and 39 negative follow-up. Sensitivity was 80% (39/49), specificity 91%(74/81), PPV 85% (39/46), NPV 88% (74/84).Conclusion: One-side increased vascular maps is highly associated with ipsilat-eral invasive breast cancers, not with DCIS.

B-330 11:51

Contrast-enhanced magnetic resonance mammography (CE-MRM):Improvement in breast lesion detection and characterization withgadobenate dimeglumine (Gd-BOPTA) vs gadopentate dimeglumine (Gd-DTPA)F. Pediconi1, C. Catalano1, S. Padula1, F. Altomari1, A. Roselli1, E. Moriconi1,M.A. Kirchin2, R. Passariello1; 1Rome/IT, 2Milano/IT([email protected])

Purpose: To prospectively and intra-individually compare equivalent 0.1 mmol/kg doses of Gd-BOPTA and Gd-DTPA for accuracy of detection and characteri-zation of breast lesions on contrast-enhanced MR mammography (MRM).Methods and Materials: Twenty-six consecutive women with a suspected breasttumor on mammography and sonography underwent two identical MRM exami-nations at 1.5 T separated by 48-72h. A T1w 3D gradient-echo sequence wasused and images were acquired pre and several minutes after injection of eitherGd-DTPA or Gd-BOPTA at an identical flow rate of 2 ml/s. Separate and com-bined assessment of non-enhanced, enhanced and subtracted images was per-formed blindly by two readers in consensus. The accuracy for lesion detectionwas determined against final diagnosis, based on findings from conventionalmammography, sonography and surgery. The sensitivity, specificity, PPV, NPVand overall accuracy for malignant lesion identification was determined againsthistology.Results: Gd-BOPTA-enhanced MRM detected significantly more lesions (45/46)than Gd-DTPA-enhanced MRM (36/46) and detected lesions were significantlymore conspicuous with Gd-BOPTA. The confidence for characterization was sig-nificantly greater with Gd-BOPTA. Comparison of the agents for their ability toidentify malignant lesions revealed significant superiority for Gd-BOPTA: the sen-sitivity, specificity, PPV, NPV and overall accuracy for malignant lesion identifica-tion was 94.7%, 100%, 100%, 80.0% and 95.6% on Gd-BOPTA-enhanced MRMand 76.3%, 100%, 100%, 47.1% and 80.4% on Gd-DTPA-enhanced MRM. Quan-titative evaluation of SI-time curves revealed significantly greater lesion enhance-ment with Gd-BOPTA.Conclusion: The detection of breast lesions and the accurate identification ofmalignant lesions is significantly superior with Gd-BOPTA compared to Gd-DTPA.

10:30 - 12:00 Room P

Molecular Imaging

SS 606Basic researchModerators:M.A. Funovics; Vienna/ATT.F. Massoud; Cambridge/UK

B-331 10:30

Using magnetic resonance imaging (MRI) and gadolinium-loadedimmunomicelles to detect and assess macrophages found inatherosclerosisS. Amirbekian, V. Amirbekian, M.J. Lipinski, J.C. Frias, J.S. Aguinaldo,Z.A. Fayad; New York, NY/US ([email protected])

Purpose: The capability to quantify biological activity of atherosclerosis usingMR imaging holds great promise for non-invasive risk-stratification based onplaque composition. Macrophages are known to play a central role in atheroscle-rosis pathogenesis. Gadolinium-containing immunomicelles (micelles linked to aspecific antibody targeting macrophages) and micelles have been shown to im-prove in-vitro and ex-vivo assessment of macrophages using MR imaging. Thegoal of the current study was to evaluate whether immunomicelles improve in-vivo imaging of atherosclerotic plaque in Apolipoprotein-E-Knockout (ApoE-KO)mice using MR imaging.Methods and Materials: Gadolinium-containing micelles were synthesized. Im-munomicelles were made using a biotin-avidin-bridge to bind micelles to a spe-cif ic antibody targeted to macrophage-scavenger-receptor (MSR).Immunomicelles, micelles and standard (Gd-DTPA/Gd-DOTA) paramagnetic con-trast-agents were tested in ApoE-KO-mice. Mice were imaged at baseline with a9.4T MR system using a high-spatial-resolution sequence. The mice were thenimaged at intervals fol lowing a tai l injection of micelles (n = 6),immunomicelles (n = 7), or standard contrast agent (n = 4).Results: Using micelles the ratio of signal-intensity in the aortic wall post andpre-contrast, normalized to muscle, was an average of 1.4 ± 0.3 (40%enhance-

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 211C D E FA G

ment) at 1 hour post-contrast, 1.3 ± 0.2 (30% enhancement) at 24 hours post-contrast, 1.2 ± 0.2 (20%) at 48 hours, 1.15 ± 0.12 (15%) at 72 hours and 1.00(0%) at 1 week. Using immunomicelles the same ratio was an average of 1.6 ± 0.3(60% enhancement) at 1 hour post-contrast, 1.69 ± 0.2 (69% enhancement) at24 hours post-contrast, 1.25 ± 0.2 (25%) at 48 hours, 1.17 ± 0.12 (17%) at 72hours and 1.00 (0%) at 1 week. In the control ApoE-KO group there was no sig-nificant enhancement of aortic wall using the non-specific Gd-DOTA contrast agent.Conclusion: The current in vivo study shows promising results for detection ofatherosclerotic vascular disease using MR imaging. Immunomicelles may proveuseful in the detection of high macrophage density typical of high-risk plaques.

B-332 10:39

Early prognosis of anti-angiogenic treatment efficacy by iron oxideenhanced parametric MR imagingT. Persigehl, L. Matuszewski, T. Kessler, A. Wall, R. Bieker, W. Berdel,R. Mesters, W. Heindel, C. Bremer; Muenster/DE ([email protected])

Purpose: To evaluate iron oxide enhanced parametric MR imaging for outcomeprognosis of anti-angiogenic tumor treatment.Methods and Materials: Fibrosarcoma bearing nude mice were injected with alow dose of a recently developed thrombogenic vascular targeting peptide (VTP;n = 9). Control animals received saline only (n = 8). For the assessment of anti-angiogenic treatment effects USPIO enhanced MR imaging was performed 4-8 hafter single i.v. injection of the VTP or saline respectively. Changes of the relaxa-tion rate ∆R2* and the vascular volume fraction (VVF) as a surrogate marker fortumor perfusion were calculated for all animals. The extension of tumor thrombo-sis/necrosis was graded histologically on a six point scale (grade 0-2 = non-responder; 3-5 = responder).Results: Histological analysis revealed low grade histological necrosis/thrombo-sis (non-responder) in 5 out of 9 animals, while 4 animals showed a good treat-ment response. Responders as defined by histology showed a significant decreasein ∆R2* compared to the control group (∆R2* control group: 0.0175 ± 0.0021 msec-

1 versus responder: 0.0039 ± 0.0018 msec-1; p < 0.01). However for non-respond-ers the ∆R2* remained unchanged (∆R2* control group: 0.0175 ± 0.0021 msec-1

versus non-responder: 0.0167 ± 0.0029 msec-1; p > 0.5). Moreover the VVF ofthe non-responders was significantly higher compared to the responders (non-responder: 0.0167 ± 0.0029 msec-1 versus responder: 0.0039 ± 0.0018 msec-1;p < 0.02).Conclusion: Parametric MR imaging allows a non-invasive, early assessment ofanti-angiogenic treatment efficacy using a thrombogenic vascular targeting agent.This imaging paradigm requires further evaluation for other anti-angiogenic treat-ment modalities.

B-333 10:48

Optical imaging of experimental arthritis using allogenous leukocyteslabelled with a near-infrared fluorescent probeG.H. Simon1, B.J. Pichler2, J. Kau1, S. Metz3, G. Piontek3, H.E. Daldrup-Link1;1San Francisco, CA/US, 2Tübingen/DE, 3Munich/DE([email protected])

Purpose: To depict differences in leukocyte migration to arthritic joints with opti-cal imaging, before and after anti-inflammatory therapy.Methods and Materials: An immune-mediated monoarthritis was induced in theright knee of 12 Sprague Dawley rats. Six rats were investigated before therapy,six rats were investigated after seven days of systemic cortisone therapy. Legswere shaved prior to imaging. Optical images of both knees were acquired be-fore, at 5 minutes, 1 hour, 4 hours, and 24 hours after injection of DiD-labelledleukocytes. Images were evaluated by calculating signal intensity ratios betweenthe arthritic (A) and contralateral normal (N) knee. A/N-ratios were tested for sig-nificant differences between baseline values and values after cell injection usinga paired t-test and between the untreated and cortisone-treated group using anunpaired t-test. Imaging data was compared with fluorescence microscopy.Results: Optical images before leukocyte injection showed a minor, but signifi-cantly increased autofluorescence of arthritic knees compared to normal knees.At 4 h and 24 h after intravenous injection of DiD-labeled leukocytes, the fluores-cence increased significantly in the arthritic knees (p < 0.05), but not in the nor-mal knees. Arthritic-to-normal knee ratios were significantly higher before treatmentthan after cortisone treatment (p < 0.05). Fluorescence microscopy confirmed anaccumulation of DiD-labelled leukocytes in arthritic knees, but not normal knees.The arthritic knees contained a higher quantity of fluorescent leukocytes beforethan after cortisone treatment.Conclusion: An accumulation of DiD-labelled leukocytes in autoimmune-arthri-tis can be detected and quantified with optical imaging. This leukocyte accumula-tion is significantly reduced after anti-inflammatory therapy.

B-334 10:57 !Pilot study of iodine-124 labelled chimeric antibody G250 (124I-cG250) in thepresurgical diagnosis of clear cell-type renal carcinoma in patients withoperable renal massesE. Quaia1, P. Russo2, D. Pryma2, N. Pandit-Taskar2, A. Nagel2, C. Pierre2,V. Reuter2, L.J. Old2, S. Larson2, C. Divgi2; 1Trieste/IT, 2New York, NY/US([email protected])

Purpose: Chimeric G250 antibody (cG250) targets the carbonic anhydrase IXantigen in clear cell-type renal carcinoma. The aim of this pilot study was to eval-uate 124I-cG250-based PET/CT scanning in the presurgical diagnosis of clearcell-type renal carcinoma in patients with operable renal masses.Methods and Materials: 15 renal tumors (diameter 1-5 cm) identified by base-line US in 7 patients (3 male and 4 female; age range 43 - 78 years; mean age ± SD,56 ± 11) were evaluated by dual-phase contrast material-enhanced CT, and byPET/CT scanning performed 2-4 hours and 4-10 days after i.v. infusion of 5mCi/10 mg of 124I-cG250. Dual-phase CT and PET/CT images were evaluated inde-pendently and in a blinded fashion by different panels of readers, and all renaltumors which showed malignant features at cross-sectional imaging were surgi-cally resected.Results: Clear cell-type (n = 5) or papillary cell-type carcinomas (n = 2), angi-omyolipomas (n = 5), and renal cysts (n = 3) were finally diagnosed according toimaging and surgical findings. A selective uptake of 124I-cG250 (StandardizedUptake Value > 3) was found in 4/5 (80%) clear cell carcinomas, while all otherrenal masses did not show any significant uptake. The overall accuracy of 124I-cG250-based PET/CT scanning in the diagnosis of clear cell phenotype was 93%(14/15).Conclusion: Preliminary results demonstrate that 124I-cG250 has potential appli-cation in the noninvasive diagnosis of clear cell phenotype renal carcinomas.This could provide precise surgical planning in localized tumors by increasingthe limits of kidney-sparing surgery, and avoid unnecessary surgery for patientswith benign renal masses.

B-335 11:06

FA difference among astrocytomas of different grade at 3.0 T DTI andcorrelation between FA and tumorous microarchitectureY. Liu1, C. Li2, K. Zhang2, J. Hou2, Q. Wang2, X. Meng2, D. Feng2, J. Zheng2,P. Cong2; 1Hefei/CN, 2Jinan/CN ([email protected])

Purpose: To compare FA of different grade astrocytomas at 3 T DTI and analyseif there is correlation between FA and tumorous microarchitecture.Methods and Materials: DTI was performed on 14 patients with astrocytomaconfirmed histopathologically (5 grade I/II, 4 grade III, 5 grade IV) by using a3.0 T MR scanner. Postprocessing was done with DTI specific software to gainFA images, and FA values in tumors were recorded. Microvascular density (MVD)and vascular endothelial growth factor (VEGF) of tumor tissue were detected bystreptavidin peroxidase (S-P) immunohistochemistry. Cd34 was used to explorethe MVD as a marker of endothelial cells. The positive expression of VEGF, MVD,and cell density were analysed. Pearson correlation analysis was used.Results: The mean FA value (0.1021 in grade I/II, 0.1708 in grade III, 0.2001 ingrade IV) significantly differed among different grade astrocytomas. Positive cor-relations were observed between FA value and VEGF, MVD, cell density (r = 0.768,P < 0.01; r = 0.868, P < 0.01; r = 0.825, P < 0.01 respectively).Conclusion: The results of this study suggest that FA value in DTI can reveal themicroarchitecture of gliomas and may be used as a predictor of grading astrocy-tomas.

B-336 11:15 !Differentiation of human stem cells in high and low malignant xenografts ofkeratinocyte derived tumorsF.M.A. Kiessling, J. Dittrich, T. Moehler, J. Peter, R. Schulz, E. Woenne,M. Mueller, W. Semmler; Heidelberg/DE ([email protected])

Purpose: To investigate the accumulation and differentiation of human stem cellsin highly infiltrative (HaCaT-ras) and low malignancy (HaCaT-A5) human tumorxenografts.Methods and Materials: Tumors were grown subcutaneously in nude mice.AC133+ human stem-cells were obtained from peripheral blood. Stem cells werelabelled with 111In-oxine, SPIO and quantum dots. Stem cells of control groupswere irradiated prior injection to prevent active migration or fixated to destroysurface receptors. Labelled cells were injected intravenously and tracked over 4days by MR imaging. Organ distribution was assessed by gamma counting. Dif-ferentiation of progenitor cells in tumors was investigated by double-immunofluo-rescence-microscopy.

Scientific Sessions

212 C D E FBA G

Results: The highest stem cell fractions were always found in liver, spleen andkidney. There were no significant differences in the progenitor cell accumulationbetween low and high malignancy tumors, and no differences between vital andirradiated progenitor cells. However, in both tumor models fixation reduced theamount of progenitor cells in tumors by a factor of > 10. While 20 ± 9% were stillpositive for AC133 in HaCaT-ras-tumors, no positive cells were found in HaCaT-A5-tumors. 24 ± 19%/24 ± 3% (HaCaT-ras/HaCaT-A5) of the progenitor cells werepositive for smooth muscle actin. Surprisingly, only 12 ± 10%/8 ± 2% of stem cellswere positive for CD31 (HaCaT-ras-tumors/HaCaT-A5-tumors). No positive cellsfor human lysozyme were found in both tumor models.Conclusion: Accumulation of human stem cells in both tumor models does notseem to underlay active migration, however, adhesion may be receptor mediat-ed. The results further indicate that the progenitor cells predominantly differenti-ate to smooth muscle cells or myofibroblasts and only to a lesser degree toendothelial cells.

B-337 11:24

Diagnosis of peritonitis using near infrared optical imaging of in vivolabelled macrophagesM.-R. Lisy, E. Schueler, I. Hilger, W.A. Kaiser; Jena/DE([email protected])

Purpose: To non-invasively detect early stages of peritonitis with fluoro-opticalmethods by in vivo labelling of macrophages using a near infrared fluorescentdye.Methods and Materials: Optical properties of the near infrared fluorochromeDY676 (Dyomics, Jena, Germany) and uptake by mouse macrophages J774 A.1have been tested. Peritonitis was induced in 6 NMRI-mice by intraperitoneal in-jection of 3 mg zymosan A. After intravenous injection of 55nmol fluorochromeper kg body weight, imaging was performed using the whole body small animalnear infrared fluorescence (NIRF) imaging system (bonSAI, Siemens, Erlangen,Germany) at short acquisition times of 0.5s. In vivo data were confirmed by exvivo investigation of isolated peritoneal macrophages with NIRF imaging andconfocal laser scanning microscopy.Results: Absorption and emission maxima of DY676 were determined at 676 nmand 701 nm respectively. The molar extinction coefficient was approximately 90,000M-1 cm-1. Fluorochrome was found to be taken up into macrophages by endocyto-sis. Qualitative and quantitative analysis of NIRF images of whole mice with peri-tonitis revealed a significant increase in fluorescence intensities in the ventralcavity by a factor of 4.6 from 188 ± 41 arbitrary units in control mice to 869 ± 151arbitrary units in diseased mice that have received fluorochrome intravenously.Confocal laser scanning microscopy of isolated peritoneal macrophages showedevidence of involvement of those cells in imaging.Conclusion: Diagnosis of peritonitis by near infrared optical imaging of in vivolabelled macrophages is feasible, indicating that early stages of inflammationcould be detected by the proposed diagnostic method.

B-338 11:33

In-vivo detection of magnetically labelled mesenchymal stem cells inkidneys after selective intra-arterial injection: Results in aglomerulonephritis model in rats at 3 TH. Ittrich, F. Thiass, C. Lange, H. Dahnke, E. Kilic, G. Adam,C.C.A. Nolte-Ernsting; Hamburg/DE ([email protected])

Purpose: To evaluate MR imaging for in-vivo monitoring of selective intra-arteri-ally injected iron oxide-labelled mesenchymal stem cells (rMSCs) in a rat modelof glomerulonephritis.Methods and Materials: 106 bone marrow derived and magnetically labelledrMSC (Resovist) were injected in the left renal artery in rats after induction of ananti-Thy1 glomerulonephritis (2 control groups: Saline or free SPIO injection). In-vitro viability of labelled rMSC was checked by trypan blue exclusion and prolifer-ation assays. In-vivo MR Imaging (T2*wGRE) and T2*/R2* measurements wereperformed before and repetitively up to day 22 after administration using an ani-mal coil in a clinical 3 T scanner (Philips Intera). MR imaging data were matchedwith histology (HE, Prussian blue). Statistical analyses of signal intensity (SI)and T2*/R2* changes were performed (student's t-test).Results: Magnetic labelling of rMSC with Resovist was highly effective. In-vitrono changes in cell viability were found. Significant SI decrease (96.8-58%) andshortening of T2* (3.5-21.4 ms, day 0-22) were observed in the cortex and theouter medulla of left kidneys in rMSC group up to 3 weeks after cell administra-tion (control1: SI decrease 2.9 ± 3.3%, T2* 39.4 ± 3.1 ms, p < 0.05) with glomer-ular predominance of labelled rMSC in histology. Aberrant rMSC were detectedin spleen, liver and lungs. Free SPIO injection (control2) resulted in a significantSI and T2* decrease only in liver and spleen.

Conclusion: Effective cell labelling of mesenchymal stem cells is feasible withclinically approved SPIO. This initial study demonstrates the potential of MR im-aging for in-vivo cell-tracking of therapeutically administered, magnetically la-belled rMSC in the kidneys after selective intra-arterial perfusion.

B-339 11:42

Imaging for enhanced hypoxia driven gene expression after hepatic arteryocclusionC.C. Riedl, I. Serganova, Y.S. Chun, B. Wen, K. Dobrenkov, R. Blasberg,H. Hricak, Y. Fong, V. Ponomarev; New York, NY/US ([email protected])

Purpose: To verify enhanced hypoxia driven gene expression after hepatic ar-tery ligation by use of genetic imaging reporter systems.Methods and Materials: RH7777 Morris hepatoma cells were transduced with aconstitutively expressed "beacon" reporter gene (monomeric red fluorescent pro-tein/Renilla luciferase double fusion) and a "sensor" reporter gene (HSV1-tk/greenfluorescent protein/Firefly luciferase triple fusion) which was under the control ofa HIF-1-inducible hypoxia responsive element (HRE). Cells were sorted for redfluorescence and hypoxia induced green fluorescence by flow cytometry. In sort-ed cells luciferase activity was assessed in vitro with and without 12h hypoxiatreatment (pO2=2%). Isolated Morris hepatoma tumors were established in liversof Buffalo rats and the effect of hepatic artery clamping on vascularization andoxygen levels in liver tumors and normal liver tissue was assessed by in vivoHoechst staining as well as with continuous in situ real time oxygen measure-ments using the Oxylite probe.Results: Bioluminescence assays after hypoxia treatment showed upregulationof the HRE driven Firefly luciferase and no positive effect on the expression ofthe constitutively expressed Renilla luciferase. In vivo experiments revealed com-plete cessation of tumor perfusion and severe hypoxia in liver tumors during he-patic artery clamping.Conclusion: These preliminary results show that hypoxia driven gene expres-sion can be detected with imaging reporter genes and suggest that in transducedliver tumors upregulation of such genes after hepatic artery clamping could beimaged in vivo.

B-340 11:51

Twofold labelled macrophages and their characterization in MR imagingA. Hartung, R. Trost, M.R. Lisy, I. Hilger, W.A. Kaiser, J.R. Reichenbach;Jena/DE

Purpose: The aim of this study was to label macrophages with a bimodal con-trast agent. This approach facilitates detection of inflammation, both in MR imag-ing and near infrared fluorescence (NIRF) spectroscopy, and combines thesensitivity of NIRF with the anatomical resolution of MR imaging.Methods and Materials: Bacterial magnetosomes with a diameter of 42nm wereused as the contrast agent. To make them observable in NIRF spectroscopy,fluorochrome DY676 (Dynomics, Jena) was bound on their surface. T1- and T2-relaxation times were determined on a 1.5 T MR-scanner (Siemens, Erlangen)with IR and multi-echo spin-echo sequences, respectively. Different concentra-tions of pure magnetosomes as well as samples with different cell counts of mag-netosome-containing macrophages were investigated.Results: The magnetosomes had only a small effect on T1, while a distinct short-ening of T2 was observed. Furthermore, NIRF spectroscopy showed that fluores-cence labelling of the macrophages was successful. Labelled macrophages hadlower relaxivities than samples containing pure magnetosomes. For pure magne-tosomes, values of R1 = 2.3 ± 0.4 (mM*s)-1 and R2 = 407 ± 86.8 (mM*s)-1 wereobtained, for labelled macrophages R1 = 0.57 ± 0.07 (mM*s)-1 and R2 =170.4 ± 20 (mM*s)-1.Conclusion: Macrophages were successfully labelled by a bimodal contrast agent.The different relaxivities obtained for pure magnetosomes and labelled macro-phages may be caused by shielding effects of the surrounding cytoplasm or aconglomeration of the nanoparticles inside the cells. Further studies with miceshall verify the applicability of this contrast agent in vivo.

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 213C D E FA G

14:00 - 15:30 Room A

Breast

SS 702UltrasoundModerators:M. Locatelli; Gorizia/ITE. Rabanal; Sabadell/ES

B-341 14:00

To determine the diagnostic benefit of undertaking bilateral breastultrasound examinations on patients under the age of 35 years withunilateral symptomsJ.A. Bull; Leeds/UK ([email protected])

Purpose: Due to a yearly increase on the demands of the breast ultrasoundservice and subsequent increase in pressure on staff and resources, certain as-pects of service provision were analysed. The aim of this audit was to investigatethe diagnostic efficacy of undertaking bilateral breast ultrasound scans on a spe-cific patient group.Methods and Materials: A retrospective audit of the asymptomatic breast of 100bilateral breast ultrasound examinations on patients under 35 years was under-taken over a 6 month period.90% were GP referrals. 10% were referred from non-consultant breast clinicians.The information collated was; patient identification number, referral source, refer-ral reason, further recommendations/management.Results: Ultrasound abnormalities were detected in only 2 cases out of 100 sam-pled (2%). Case 1: Bilateral cysts. Case 2: Bilateral fibroadenomata.Both were reported as radiologically benign with no further patient managementrecommended.Conclusion: From the results obtained departmental protocols were amended.Patients presenting under the age of 35 with unilateral breast symptoms havingunilateral breast ultrasound only.As a result the efficiency of the department has been enhanced by reducingappointment times and increasing the throughput of patients.

B-342 14:09

Evaluation of solid breast masses by four sonographic technologies andtheir combinations: Conventional, tissue harmonic, compound and specklereduction imagingA. Ozdemir, S. Gultekin, V. Kahveci, F. Elbuken, I. Tuncbilek; Ankara/TR([email protected])

Purpose: To evaluate the efficiencies of three recent ultrasound technologies,compound (CI), tissue harmonic (tHI), speckle reduction imaging (SRI) and theircombinations (CI+tHI and CI+tHI+SRI) compared to the conventional B-modeimaging, in the evaluation of solid breast masses.Methods and Material: 126 solid breast masses (mean size 14 mm) with BI-RADS 3-to-6 categories were evaluated. Cine views obtained with single andcombined technologies for each lesion was recorded. SRI function was used atthe maximum level. Static images were obtained from the same planes of cineviews for each lesion. Images were reviewed and graded by three observers fordelineation of lesion boundaries, clarity of internal texture, and conspicuity ofposterior echo pattern, edge-shadowing and peripheral echogenic halo. Fried-man's test was used for multiple statistical comparisons between the techniques.Results: In the evaluation of lesion boundaries and peripheral halo, tHI+CI+SRIwas the most valuable method, followed by SRI, tHI+CI, tHI, CI, and B-mode(p < 0.001). In the evaluation of internal texture, CI was the most sensitive meth-od followed by tHI, while SRI, single or combined, was the least valuable one(p < 0.001). In the evaluation of posterior echo pattern and edge shadowing, SRIwas the most, and tHI+CI+SRI was the second most useful method (p < 0.001).Conclusion: SRI, CI, and tHI have significant advantages over B-mode imagingin the evaluation of solid breast masses. SRI is useful in delineating lesion bound-aries, peripheral halo, posterior echo and edge shadowing. CI and tHI are usefulin clarifying internal texture.

B-343 14:18

Comparison of three-dimensional and two-dimensional ultrasound fordifferentiating benign and malignant solid breast masses: A performancestudy by radiologists and computer-aided diagnosis systemsW. Moon1, R.-F. Chang2, D.-R. Chen3, N. Cho1; 1Seoul/KR, 2Taichung/TW,3Chiayi/TW ([email protected])

Purpose: To evaluate the performance of radiologists and CAD systems on 3DUS and conventional 2D US for the characterization of solid breast masses.Methods and Material: 150 solid breast masses including 68 nonpalpable le-sions were prospectively collected using both 2D and 3D US with a 5-10 MHzdedicated volume transducer. Sixty-one cancers and 89 benign lesions were con-firmed with biopsy. Six radiologists who specialize in breast imaging independ-ently and blindly reviewed static 2D US images and stored 3D volume data. For aCAD study, the extracted texture features of the 2D and 3D US images wereused to classify the tumor as benign or malignant using the artificial neural net-work.Results: For all radiologists, 3D US was superior to 2D US in sensitivity (97.8 ± 2.8vs 95.8 ± 3.9), specificity (69.5 ± 15.2 vs 64.5 ± 16.8), and PPV (70.0 ± 10.6 vs66.3 ± 11.6) and NPV (98.5 ± 2.1 vs 96.5 ± 3.0) values. The negative predictivevalue of over 98% was obtained in five of six reviewers at 3D US whereas it wasobtained in only two reviewers at 2D US. For a CAD study, accuracy, sensitivity,specificity, positive predictive value, and negative predictive value were statisti-cally significantly improved using 3D instead of 2D US images. At ROC analysis,3D and 2D CAD schemes yielded Az values of 0.97 and 0.85 in distinguishingbetween benign and malignant lesions, respectively.Conclusion: Our results suggest that stored 3D US information may allow radi-ologists and CAD systems to do better classification of solid breast masses thanis possible with current 2D US images.

B-344 14:27 !Ultrasound elastography: Results of a French multicentric prospectivestudy about 345 breast lesionsA.A. Tardivon1, A. Delignette2, S. Lemery3, B. Baratte4, L. Levy1, P. David5,A. Le Mouel6, C. Balu-Maestro7, B. Scheffer8; 1Paris/FR, 2Dijon/FR,3Clermont-Ferrand/FR, 4Amiens/FR, 5Albertville/FR, 6Besancon/FR, 7Nice/FR,8Nantes/FR ([email protected])

Purpose: To evaluate the performance of ultrasound elastography in breast le-sions.Methods and Materials: 345 lesions (228 benign, 117 malignant, 61% < 10 mmin size) were analyzed with the EUB 8500 Logos ultrasonic unit (Hitachi, Japan)and a linear array transducer of 7.5-13 MHz. Diagnosis was obtained by FNA,core or surgical biopsies, follow-up or comparison with previous US studies (be-nign lesions). The elastic score was classified according a five-point color scale(Ueno classification, 1-3 = benign and 4 -5 = malignant).Results: Sensitivity, specificity, positive predictive value, negative predictive val-ue and accuracy were respectively 79.5, 93, 85.3, 89.8 and 88.4%. There were11 stiff benign lesions (fibrous mastopathy and sclerosis adenosis) and 22 softmalignant lesions (DCIS, mucinous subtype, poor-differentiated IDC, and ILC). Ifthe11 soft malignant lesions categorized as Bi-RADS category 5 were excluded,the sensitivity of elastography was 88%. When the elastographic score 3 (centralpart of the mass stiffer than the peripheral normal tissue) was considered asmalignant, sensitivity, specificity, positive predictive value, negative predictive valueand accuracy were respectively 90, 81, 70.5, 93.9 and 83.8%. When only mass-es categorized as Bi-RADS category 3 and 4 were analyzed, the sensitivity ofUS elastography was 90% (malignancy = scores 3-4-5) with a specificity of 72%.Conclusion: US elastography is a valuable adjunct to increase the specificity ofthe conventional B-mode, especially in breast masses categorized as Bi-RADScategory 3 and 4 by adding new criteria for benignity and avoiding unnecessarydiagnostic procedures.

B-345 14:36

Elasticity imaging of 234 breast lesions shows that it could halve biopsyrates of benign lesionsW.E. Svensson1, D.G. Amiras1, S. Shousha1, A. Al-Kufaishi1, D. Chopra1,D. Sinnett1, T.J. Hall2, C. Lowery3; 1London/UK, 2Madison, WI/US, 3Issaquah,WA/US ([email protected])

Purpose: To evaluate the potential of real time ultrasound elasticity imaging toalter clinical management and reduce the false positive rate associated with breastultrasound examinations particularly in the screening situation. Earlier work hasshown that elasticity imaging profiles of benign lesions are smaller while malig-nant lesions are larger than the grayscale appearance.

Scientific Sessions

214 C D E FBA G

Methods: Patients attending for routine breast ultrasound for focal breast abnor-malities had ultrasound breast strain imaging (implemented on a Siemens Elegraultrasound scanner) at the end of their ultrasound examination if there was afocal abnormality. The strain images were assessed and correlated with subse-quent histology or follow-up.Results: 234 lesions were assessed and correlated with the subsequent diagno-sis; 67 cancers, 5 intermediate lesions had an Elasticity to B-mode ratio > 0.75.162 benign lesions (103 fibroadenomas, 21 cysts, 10 inflammatory masses, 6intramammary lymph nodes, 6 normal breast, 2 gynaecomastia and 14 otherlesions had an Elasticity to B-mode ratio < 0.75. For Cancers and intermediatelesions:- Sensitivity = 100 %, specificity = 51%, negative predictive value = 100%,positive predictive value equals 48%.Conclusion: All lesions with an elasticity to grayscale ratio of less than 0.75were benign. If a ratio of less than 0.75 was used to determine the necessity forbiopsy in the screening situation this would halve the number of biopsies of be-nign lesions (false positive biopsies). Increasing experience with, and improve-ments in, elasticity imaging may lead to a further decrease in false positive resultsthan achieved in this study.

B-346 14:45

Role of ultrasonography in detection of wide intraductal component ofinvasive breast cancer: A prospective studyS. Sundararajan; Tsukuba/JP ([email protected])

Purpose: Efficiency of ultrasound (US) in detection of extent and direction ofintraductal component in comparison with magnetic resonance imaging (MRI)and histopathological findings.Patients and Methods: In 59 patients with 60 invasive breast cancers, US fea-tures of the intraductal component were classified as (a) ductal dilatation be-tween main tumor and satellite lesion, (b) presence of satellite lesion in the samesegment (c) solid ductal dilatation radiating from the tumor. The direction of intra-ductal component was classified as towards nipple and towards periphery. Crite-ria for MR imaging are as follows: (a) a satellite lesion around the main tumor, (b)bridging enhancement between main tumor and satellite lesion. The direction ofintraductal component was similar to that of US.Results: In 17 of 59 (28.8%) patients, wide intraductal component (15 or > 15 mm)towards nipple and 3 out of 60 (5%) patients, wide intraductal component to-wards periphery, was proved histopathologically. US and MR imaging could ac-curately detect wide intraductal component towards nipple in 14 and 8 patientsrespectively out of 17 patients. Sensitivity, specificity and accuracy for detectionof wide intraductal component towards nipple by US were 82.3%, 90.5% and88.1%, respectively. Sensitivity for detection of wide intraductal component to-wards nipple by MR imaging was 47.0%, specificity was 72.9% and accuracywas 71.2%. When the results of both diagnostic methods namely US and MRimaging were combined, sensitivity rose to 93.7%, specificity was 72.1% andaccuracy was 80.0%.Conclusion: Ultrasound findings of intraductal extension and its direction corre-lated well with histological intraductal extension and the direction.

B-347 14:54

Ultrasound elasticity imaging is more accurate than grayscale ultrasoundfor assessing the extent of invasive breast cancers: Early results ofpathological correlationW.E. Svensson, A. Usupbaeva, S. Shousha, S. McLaggan, A. Al-Kufaishi,P.T.R. Thiruchelvam, J.S.K. Lewis, D. Sinnett; London/UK([email protected])

Purpose: To determine how accurately freehand elasticity images and measure-ments correlate to the pathological appearances and dimensions of breast can-cers. An earlier study suggested that it is more accurate than conventionalgreyscale ultrasound for demonstrating breast tumour extent. Reported seriesshow reoperation rates of as high as 50% when wide local excision is based onconventional breast imaging techniques.Methods: Freehand elasticity images (using real time elasticity software imple-mented on a Siemens Elegra ultrasound scanner) were obtained at regular inter-vals through the tumour the day before surgery. The skin of the breast was markedto allow orientation sutures to be placed on the tumour specimen as it was ex-cised. Histology sections were made in the same plane as the imaging planes atright angles to the greatest diameters of the greyscale tumour appearance. Theedges of the specimen were examined to ensure tumour margin extent was ac-curately measured. The skin marking at imaging, and specimen marking at sur-gery, allowed ultrasound and specimen image correlation.Results: One of nine cases, of elasticity imaging, under measured maximumtumour diameter. B-mode under measured in seven cases. Seven cases of elas-

ticity imaging had closer correlation to pathology diameters than greyscale imag-ing. On average greyscale under measured by 15% while elasticity imaging overmeasured by 9%. The strain image of Case 1 extended to the edge of the strainbox, which could explain the under measurement.Conclusion: Initial results suggest that elasticity imaging provides a more relia-ble method of assessing tumour extent.

B-348 15:03

Axillary echography in determining state of breast cancerJ.A. Lopez-Ruiz1, J.A. Mieza2, M. Ruiz-Delgado1, Z. Fernandez1, I. Zabalza1,I. Basarrate2; 1Galdakao/ES, 2Barakaldo/ES ([email protected])

Purpose: To evaluate the diagnostic value of Mode B axillary echography in de-termining the state of the lymph nodes in breast cancer.Methods and Materials: 102 cases of infiltrating breast cancer have been re-viewed, involving pre-surgical axillary echographical evaluation and subsequentpathological verification following lymph-adenectomy (gold standard). The echo-graphical study was performed with Mode B echography, high-frequency sound(7.5-10 Mhz), and was considered "positive" or "negative", according to the pres-ence/absence of pathological findings in the nodes. No percutaneous biopsy wasperformed on the nodes in the "positive cases". The following parameters havebeen evaluated: Sensitivity, Specificity, Positive Prediction Value, Negative Pre-diction Value and Exactitude.Results: The echographical study considered 41 cases to be "positive" and 61cases "negative". In all "positive" cases, disease was confirmed by surgery. How-ever, metastases was observed in 9 of the 61 cases considered "negative". Theparameters used were: Sensitivity 82 %, Specificity 100 %, Value Positive Pre-diction Value 100%, Negative Prediction Value 85 %, and Exactitude 91 %.Conclusion: The echographical evaluation of the armpit, should be routinely indetermining the state of breast cancer, especially in those Medical Centres wereSentinel Lymph Node Technology (S.L.N.B). is used. A "positive" axillary echog-raphy, even if not accompanied by percutaneous lymph node biopsy, could makeS.L.N.B. unnecessary. Conversely, a "negative" echography" would require S.L.N.Bto be performed.Axillary echography (with or without percutaneous biopsy), in determining thestate of breast cancer, allows for a more rational and efficient use of S.L.N.B.

B-349 15:12

Axillary lymph node sonography prior to sentinel node biopsyM. Bernathova, M. Duenser, G. Bodner, B. Zelger, W. Jaschke; Innsbruck/AT([email protected])

Purpose: The aim of our study was to assess the negative predictive value ofaxillary lymph node (LN) sonography in patients with histologically proven breastcancer that underwent sentinel lymph node biopsy (SNB).Methods and Material: From January 1998 to December 2004 101 patients withhistologically proven breast cancer, without any sonographic evidence for meta-static axillary LN involvement, were admitted for SNB. Sonographic criteria fornormal or reactive LN were the longitudinal transverse axis ratio less then 2,presence of a normal central echogenic hilum with colour Doppler sonographicmapped hilar vessels, without vessels on the periphery of the LN or asymmetricthickened lymph node cortex.Results: Retrospectively we reviewed the final histological findings and correlat-ed the definitive tumor nodal staging with preliminary sonographic findings.In 30 patients (30%) histology of Sentinel LN showed positive result, with size ofinvolvement varying from 0.2 mm to 16 mm. In 9 patients (from 30) the size wasless then 0.2 mm. In one patient a sentinel LN was negative, whereby in axillarylevel 2 LN were metastatic involved.Only 2 of 71 patients with negative sentinel LN, later developed LN metastasis inhigher axillar levels. The negative predictive value of sonography was 0.70.Conclusion: Sonography of the axillary LN has a high negative predictive valuefor axillary LN involvement in breast cancer. The image resolution is currently themost limiting factor to detect smaller size LN metastasis.

B-350 15:21 !Usefulness of contrast-enhanced US in the evaluation of axillary nodes ofbreast cancer patientsJ. Camps Herrero1, M. Sentís Crivellé2, V. Ricart Selma1, C. Martinez Rubio1,M. Lloret Martí1, B. Ballester Sapiña1, A. Casterá March1; 1Alzira -Valencia/ES,2Sabadell/ES ([email protected])

Purpose: To assess the usefulness of a microbubble contrast in the evaluation ofaxillary lymph nodes in patients with a diagnosis of breast cancer. The study was

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 215C D E FA G

carried out within the Perfusion Ultrasound Multicenter European Breast (PUMEB)protocol.Method and Materials: Contrast-enhanced US (CEUS) with a microbubble con-trast agent (SonoVue) was performed in 26 patients with suspicious findings inaxillary US and/or negative fine needle aspiration biopsy (FNAB) in the contextof breast cancer staging to confirm or rule out the presence of axillary metastas-es. In all positive lymph nodes a FNAB was performed. Findings of FNAB, senti-nel node procedures or lymphadenectomies were considered the gold standard.Results: CEUS showed hypoechoic areas consistent with metastatic deposits in11 patients and homogeneous enhancement consistent with normal reactive lymphnodes in 9 patients. The smallest metastatic deposit seen with CEUS was 4 mm.There were four false positive cases (one due to a granuloma and three attribut-ed to the low resolution of the images) and two false negative cases. Depiction ofmetastatic deposits with CEUS changed therapeutic approach from surgery toneoadjuvant chemotherapy in four patients (15%). Sensitivity was 73% and spe-cificity 82%.Conclusion: These preliminary results show that perfusion imaging with CEUScan enhance the already high positive predictive value of axillary US guided bi-opsy in initially negative cases and increase diagnostic confidence in negativeexams. Low resolution of the images may limit its interpretation and further stud-ies as well as longer temporal windows are needed to gain experience and in-crease its accuracy.

14:00 - 15:30 Room B

Musculoskeletal

SS 710MRI and US of the kneeModerators:E.G. McNally; Oxford/UKP.W.J. Vincken; Leiden/NL

B-351 14:00

Ultrasonography of meniscal tears in patients with equivocal physicalexaminationJ.W.J. de Rooy, S.W.T.P. Heijmink, A.L. van Kampen, C. van Kuijk,J.G. Blickman; Nijmegen/NL ([email protected])

Purpose: To evaluate the accuracy of ultrasound (US) in the detection of menis-cal tears in a selected group of patients with equivocal physical examination,correlated with arthroscopic findings.Methods and Materials: 124 consecutive patients (49 female; 75 male; meanage 44.0 y) with a positive clinical anamnesis for a meniscal tear (49 lateral, 75medial) but equivocal physical examination, underwent US of the knee. US wasperformed with a linear array transducer (12-5 MHz) by an experienced muscu-loskeletal radiologist. All patients subsequently underwent arthroscopy. The find-ings of US and arthroscopy were compared and correlated.Results: US had a sensitivity of 91% for the medial meniscus (95% CI: 0.79-0.97) and 94% (95% CI: 0.79-0.99) for the lateral meniscus. Specificity was re-spectively 75% (95% CI: 0.51-0.90) and 66% (95% CI: 0.39-0.88). US revealedan overall sensitivity of 92 % (95 % CI: 0.84-0.97) and a specificity of 71 % (95 %CI: 0.53-0.85) with a +PD value of 89 % and a -PD value of 78 % for both menis-ci. Five out of 7 patients with false negative results had a small tear of the deepmedial posterior horn. Seven out of 10 patients with false positive results hadother pathologic findings at arthroscopy, including chondropathy of the femoralcondyle (5), medial meniscocapsular dehiscence (1) and ACL-tear (1).Conclusion: US seems a promising, inexpensive tool in the pre-operative evalu-ation of meniscal tears in a selected group of patients with equivocal physicalexamination.

B-352 14:09 !Comparison of US and MR in meniscus degeneration and tearsN. Unlu, H. Ozcan, E. Ustuner, I. Erden; Ankara/TR

Purpose: The aim of this study was to investigate the role of US in the diagnosisof meniscus degeneration and tears compared to MR imaging and arthroscopy.Methods and Material: Thirty-five patients (ages 19-65) with knee pain and lockingwere included in the study. After MR imaging and US exams, 22 of the patientswere also evaluated with arthroscopy and the results were compared. Then pa-tients were divided into 2 groups according to their ages (≤ 35 years and > 35years) and studies were compared using k test.Results: US (n:15) was not as efficient as MR imaging (n:23) in detection of

degeneration and the correlation between the two studies was low (k < 0.05)except in the anterior horn of the lateral meniscus where moderate correlationwas present (k=0.50-0.75). For tears, the sensitivity and specificity of US were90.9% and 63.6% and for MR, 93.3% and 100% respectively. The posterior hornof the medial meniscus was the best visualized portion of the meniscus by USwhile visualization of the main bodies were limited. In the young group, the sensi-tivity and specificity of US were 80% and 100% compared to 66.7% and 75% inthe elder group. Moderate correlation was present between US-MR, and US-arthroscopy while MR showed better correlation with arthroscopy.Conclusion: US is not a suitable substitute for MR imaging in the routine diag-nostic evaluation of meniscus lesions. However, in selected cases such as youngage group, traumatic cases and cases with a contraindication for MR imaging,US may find a role to stratify patients for further evaluation.

B-353 14:18

Open 1.0 Tesla MR system: Kinematic studies of the kneeC. Bangard1, A. Gossmann1, K. Bloemers2, C. Bos2, E. Siakavara1, K. Lackner1;1Cologne/DE, 2Eindhoven/NL ([email protected])

Purpose: To perform kinematic studies of the knee with high spatial resolutionimages on an open high field MR system.Methods and Materials: The knees of ten volunteers were examined on an openhigh field MR scanner (Panorama 1.0 T, Philips Medical Systems, Eindhoven,The Netherlands). High spatial resolution T2-w and PD-w TSE sequences weregenerated in seven different positions between full extension and approximately90 degrees flexion of the knee. During the whole extent of flexion and extension,all knees could be placed in the isocenter of the magnet because of the wideopen patient aperture. All assessed images were analyzed by two experiencedradiologists for different clinically relevant anatomical structures (cruciate liga-ments, menisci, cartilage, collateral ligaments, position of patella and femoralcondyles).Results: The high spatial resolution images showed comparable SNRs with thatof cylindrical 1.5 T scanners in all different positions of the knees. The excellentimage quality allowed a precise delineation of the anatomical structures with vis-ualization of the movement of the cruciate ligaments, the slide of the femoralcondyles and the patella. Depending on the position of the knee, the contact ofthe cartilage of the femoral condyles related to the menisci and the tibial plateaucould be shown.Conclusion: Kinematic studies on an open high field MR scanner allow a pre-cise assessment of clinically relevant anatomical structures of the knee. Highspatial resolution images in combination with kinematic studies could be helpfulfor the diagnosis of various diseases of the knee, especially transient patellardislocation.

B-354 14:27

Applications of a new upright weight-bearing dedicated MR imaging unit inthe evaluation of ligamentous, meniscal and osteochondral disorders of theknee: Preliminary resultsA. Barile, N. Limbucci, S. Carducci, L. Satragno, C. Masciocchi; L'Aquila/IT([email protected])

Purpose: To assess the role of a new upright weight-bearing dedicated MR im-aging unit in the evaluation of knee pathology in comparison with conventionalsupine MR imaging.Methods and Materials: We selected 60 patients (38M, 22 F) with clinical andMR imaging evidence of knee disorders. At standard MR imaging 20 patients hadsigns of ACL instability (group A), 20 had meniscal lesions (group B) and 20presented femoro-tibial osteochondral pathology (group C). We only enrolledpatients with an isolated disorder. In the same session, after conventional supineMR imaging, patients underwent upright weight-bearing examination with the samededicated MR imaging unit (G-Scan, Esaote; 0.25T; FOV 25 cm; dedicated DPAknee coil). The conventional examination comprised sagittal SE T1-w, axial SET2-w and coronal GE T1-w sequences. The upright examination was performedwith sagittal and coronal SE T1-w sequences and (in case of osteochondral dis-orders) coronal GE T1-weighted sequence. We evaluated the differences of kneedynamics between supine and upright imaging in the same patients.Results: In group A and B weight-bearing MR imaging was useful for evidence ofinstability not visible with conventional MR imaging (stump deflexion, translationof the tibia, widening of meniscal tear). In group C weight-bearing MR imagingallowed the evaluation of signs of overload (thinning of the articular rim and thecartilage).Conclusion: Our preliminary observations allow affirmation that upright weight-bearing MR imaging has wide potential for the study of knee stability and dynam-ic modifications in case of ligamentous, meniscal and osteochondral lesions.

Scientific Sessions

216 C D E FBA G

B-355 14:36

High resolution MR imaging versus conventional MR imaging compared toarthroscopy in patients with a suspected meniscal tearC. Krestan, S. Marlovits, M. Weber, W. Matzek, H. Imhof, S. Trattnig; Vienna/AT([email protected])

Purpose: To evaluate the diagnostic performance of high resolution compared toconventional MR imaging of the knee with arthroscopic findings in patients with asuspected meniscal tear.Methods and Material: Twenty-nine patients (15 female, 14 male) from the de-partment of trauma surgery, with a suspected meniscal tear, were included intothe study. MR imaging exams were performed on a 1.0 T MR imaging scanner(T10-NT, Philips Medical Systems, Best Netherlands) with 2 protocols: standardMR imaging using a knee coil consisting of sagittal dual T2 TSE sequence (FOV200 mm, matrix 512x512, slice 3 mm, TR 2500 ms, TE 11 ms, voxel size 0.39 x0.39 x 3 mm; and high resolution MR imaging with a dedicated surface coil (FOV160 mm, matrix 512x512, TR 1200 ms, TE 10 ms, voxel size 0.31x0.31x3 mm).The menisci were independently evaluated by 2 blinded musculoskeletal radiolo-gists on the basis of an adapted score (0=normal, 1=intrameniscal hyperintensesignal, 2= discontinuity of the surface, 3= fragmentation). Athroscopy with evalu-ation of the menisci was performed by an experienced trauma surgeon and doc-umented by videotaping.Results: Fifteen of the 29 patients underwent arthroscopy and all revealed ameniscal tear. The correct diagnosis of a meniscal tear was 80% for the standardMR imaging and 86.7% for the high-resolution MR imaging-exam. The kappa-value between reader 1 and 2 was 0.58 for the standard MR imaging and 0.42 forthe high-resolution MR imaging protocol.Conclusion: Our preliminary data suggest that high resolution MR imaging ofthe menisci leads to an improved sensitivity for diagnosis of meniscal tears, withan acceptable interobserver agreement.

B-356 14:45

Incidence of meniscal root injuries in patients with non-displaced tibialplateau fracturesJ.H. Yun, J.P. Clement, M.A. Bredella, W.E. Palmer, A. Kassarjian; Boston,MA/US ([email protected])

Purpose: To analyze the incidence of meniscal root injuries in patients with non-displaced tibial plateau fractures.Methods and Materials: MR studies performed in 2004 showing non-displacedtibial plateau fractures were analyzed retrospectively. Patient information includ-ing age, sex, mechanism of injury, and surgical correlation was obtained. Thepresence of meniscal, ligamentous, myotendinous, and cartilaginous injuries wasrecorded. Attention was given to the presence of meniscal root injury defined asmeniscal tear involving the central 1 cm of the posterior meniscal attachment.Results: 21 patients had medial tibial plateau fractures (MTPF), and 11 patientshad lateral tibial plateau fractures (LTPF). In female patients with MTPF, 14 of 15(93%) had medial meniscal injury (MMI), 10 (67%) of which involved the menis-cal root, and 2 of 15 (13%) had lateral meniscal injury (LMI). In female patientsage > 60 with MTPF, 9 of 9 (100%) had MMI, 7 (78%) of which involved themeniscal root, and 4 of these 9 (44%) had popliteus myotendinous injury. In pa-tients with LTPF, 4 of 11 (36%) had MMI, 1 (9%) of which involved the meniscalroot, and 4 of 11 (36%) had LMI. Only 1 of 17 (6%) knee radiographs of patientswith MR evidence of non-displaced tibial plateau fractures showed evidence offracture, while 10 of 17 (59%) showed degenerative changes.Conclusion: Radiographically occult medial tibial plateau fractures, particularlythose in older women, have a high incidence of medial meniscal root injury andpopliteus myotendinous injury.

B-357 14:54

Sonographic imaging of the posterior cruciate ligament using convextransducerM.P. Dzianach1, A. Czwojdzinski2; 1Poznan/PL, 2Warsaw – Otwock/PL([email protected])

Purpose: To describe the US appearance of entire length of the posterior cruci-ate ligament (PCL) using a convex transducer.Methods and Materials: Fifty healthy individuals were examined using ultra-sound with a 2-5 or 5-7 MHz convex transducer (100 knees). Thickness of theproximal and distal part of the PCL and the curvature of its bow (PCL index, PCLbowing ratio) were assessed. Echostructure of each part was analyzed. Addition-ally five studies of fresh amputated knee specimens were performed. Eye visiblecontrast agent was injected into the regions of origin and insertion of the PCL

under US guidance. Then knee was dissected to expose PCL and deposition ofcontrast inside.Results: In sagittal popliteal view the PCL presents as a hypoechoic bow withsharp margins, mean 4.7 mm thick proximally, mean 5.3 mm distally, mean bow-ing ratio (PCL index) 4.5. In the distal part anisotropic effect is possible to achieve.Cadaveric studies showed the feasibility of the entire length of PCL in ultrasoundevaluation.Conclusion: High resolution sonography is valuable in PCL imaging. Convexprobes enable evaluation of the entire length of this ligament.

B-358 15:03

Grading of anterior cruciate ligament graft injury after ligamentreconstruction surgery: Diagnostic efficacy of oblique coronal MR imagingof the kneeS. Moon, S. Hong, J.-Y. Choi, W. Jun, J.-A. Choi, H. Kang; Seoul/KR([email protected])

Purpose: The purpose of this study was to evaluate the diagnostic efficacy ofadditional oblique coronal MR imaging of the knee for the grading of anteriorcruciate ligament (ACL) graft after ligament reconstruction surgery.Methods and Materials: We retrospectively reviewed 43 consecutive MR imag-es of the knee in 41 patients with follow-up arthroscopy. The MR examinationsincluded routine sequences and oblique coronal T2-weighted images, which ori-ented in parallel to the course of the femoral intercondylar roof. By consensus,four readers evaluated the status of the ACL graft by routine knee MR imagingand then by additional oblique coronal imaging. The severity of the ACL graftinjury was graded using a 3-point system from MR images namely normal, suspi-ciously abnormal, and graft failure, and results were compared with arthroscopicfindings. Weighted k statistics were used to analyze the diagnostic accuracy ofroutine knee MR imaging with and without additional oblique coronal imaging. Ineach evaluation, observers reported a confidence level for the grade of ACL graftinjury in two image groups.Results: The weighted k scores (kws) were 0.481 by routine knee MR imagingonly; with additional oblique coronal imaging, the kws increased to 0.618. Themean confidence level was significantly higher (P < 0.01, Wilcoxon signed ranktest) by an additional oblique coronal imaging than by routine knee MR imagingsequences.Conclusion: Additional use of oblique coronal MR imaging of the knee improvesboth diagnostic accuracy and confidence in the grading of ACL graft injury.

B-359 15:12

US and MR imaging of peroneal intraneural ganglia: Emphasis on thearticular branchM. Damasio1, N. Stagnaro1, N. Pignataro1, S. Bianchi2, G. Bodner3,C. Martinoli1; 1Genova/IT, 2Geneva/CH, 3Innsbruck/AT([email protected])

Purpose: To describe US and MR imaging findings in patients with peronealintraneural ganglia.Methods and Materials: 17 consecutive patients with dysfunction of the com-mon peroneal nerve were evaluated with 12-5 MHz and 17-5 MHz US and 1.5 TMR imaging. All patients had abnormal nerve conduction studies.Results: In all but 1 patient, US and MR imaging identified ganglia of variablesize and shape in relation to the superior tibiofibular joint. US and MR imagingdivided these cysts into extraneural ganglia (n = 9/16), which developed outsidethe nerve, and intraneural ganglia (n = 7/16) developing within the nerve. Intraneu-ral ganglia had a stereotypical appearance: US found the bulk of the ganglion onthe anterolateral aspect of the superior tibiofibular joint, located remotely fromthe position of the peroneal nerve. The ganglion was invariably associated with adilated articular branch. This branch appeared as a long tubular process cours-ing along the posterolateral aspect of the fibula to join the bulk of the ganglionwith the nerve. It was markedly enlarged and assumed a cystic appearance with-out detectable fascicles. More proximally, the fascicles of the deep and commonperoneal nerve were displaced eccentrically by the cyst growing within the epineu-rium. The overall size of the ganglion and the entity of nerve deficit varied withtime. Two intraneural ganglia regressed at 6-months follow-up.Conclusion: In the evaluation of peroneal neuropathy caused by superior tibi-ofibular joint ganglia, US and MR imaging provide unique information on the in-traneural location of the cyst and the status of the involved nerve.

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 217C D E FA G

B-360 15:21

Assessment of bone marrow edema and subchondral lesions of the knee:MR imaging controlled outcome following oral treatment with Iloprost orTramadolM.E. Mayerhoefer1, M.J. Breitenseher2, N. Aigner3, S. Hofmann4, C. Norden5,A. Vakil-Adli6, J. Kramer6; 1Vienna/Oberwart/AT, 2Vienna/Horn/AT, 3Vienna/AT,4Stolzalpe/AT, 5Berlin/DE, 6Linz/AT ([email protected])

Purpose: To determine the outcome of bone marrow edema (BME) and subchon-dral lesions of the knee after treatment with either the prostacyclin analogue Ilo-prost or Tramadol in a double-blind, randomized MR imaging study.Methods and Materials: Thirty-one patients with painful BME of the knee causedby early-stage osteonecrosis, osteoarthritis, bone bruise or stress were includ-ed, and Institutional Review Board approval was obtained. Coronal STIR imagesof the affected knees were acquired. After the initial MR examinations, patientswere randomized either to Iloprost (n = 16, group 1) or Tramadol (n = 15, group2). The treatment duration was 4 weeks. Follow-up MR images were obtainedafter 3 months and again after 1 year. For baseline and follow-up examinations,the mean volume and signal contrast of the BME were quantified using a compu-ter-assisted method. The presence of subchondral lesions immediately adjacentto the subchondral bone was also assessed.Results: In affected bones, mean BME volume regression at the 3-month follow-upindicated a better treatment effect for Iloprost (Iloprost: -47.1%; Tramadol: -38.6%).Mean BME signal contrast regression was also higher in group 1 (Iloprost: -42.8%;Tramadol: -38.5%). Subchondral lesions in the Iloprost group decreased from 11at baseline to 1 at final follow-up, while in the Tramadol group, subchondral le-sions only decreased from 5 at baseline to 3 at final follow-up.Conclusion: BME regression was more pronounced in patients treated with Ilo-prost. Under Iloprost treatment, there was also a clear trend for better healing ofsubchondral lesions, which have been reported to indicate lesion irreversibility.

14:00 - 15:30 Room C


SS 701Diffuse liver diseaseModerators:S. Bohata; Brno/CZS.D. Yarmenitis; Iraklion/GR

B-361 14:00

Long-term natural history of focal confluent fibrosis (FCF) in the cirrhoticliver: A CT serial analysisG. Brancatelli1, R. Baron2, M. Federle3, G. Sparacia1, K. Pealer3; 1Palermo/IT,2Chicago, IL/US, 3Pittsburgh, PA/US ([email protected])

Purpose: To assess the natural history of FCF in the cirrhotic liver.Methods and Materials: Two radiologists reviewed 118 CT (obtained before, at30s and 70s after contrast injection [5 mL/sec]) in 26 patients. Number, location,attenuation and presence of trapped vessels within FCF were evaluated. Sizevariation was assessed with an arbitrary retraction index obtained by using astandard linear regression analysis with the mixed-effects model. The retractionindex was defined as a variation of the ellipsoid volume of FCF over time. Aclassification and regression tree analysis performed by using the Fisher exacttest, was utilized to build the most efficient classification algorithm combiningretraction index and time interval.Results: Each radiologist identified 41 FCF. Twenty-eight lesions were locatedeither in segment IV, VII or VIII. Thirty-two lesions were hypoattenuating on NC,25 were hypoattenuating on HAP and 20 were isoattenuating on PVP. Sevenwere hyperattenuating on PVP. Trapped vessels were found in six lesions. Theretraction index showed overall significant increase over time (r=.423, P= <.0001).Classification and regression tree analysis demonstrated an increase of the re-traction index of 1% in 40% of lesions at ≤ 24 months, of 7% in 32% of lesions at24-40 months, of 52% in 17% of lesions at 40-60 months and of 17% in 11% oflesions at > 60 months (F =51.9, p =.0001).Conclusion: The liver changes from FCF occur substantially and rapidly in somepatients and more slowly in others. Follow-up imaging may be beneficial in char-acterizing these lesions.

B-362 14:09

Computed tomography for the assessment of macrovesicular hepaticsteatosis in living liver donors: Comparison of visual grading and liverattenuation indexS. Lee, S. Park, P. Kim, K. Kim; Seoul/KR ([email protected])

Purpose: This presentation is to determine and compare the performance ofvisual grading and liver attenuation index (LAI) of unenhanced computed tomog-raphy (CT) in the diagnosis of ≥ 30% macrovesicular steatosis (MaS) in potentialdonors for living donor liver transplantation (LDLT).Methods and Materials: Of the LDLT donor candidates who underwent same-day unenhanced CT and ultrasound-guided liver biopsy, two case (i.e. ≥ 30%MaS)-control (i.e. < 30% MaS) combinations between a case group (n = 24) andtwo control groups (n = 24, respectively), i.e. an overtly-negative control(mean ± SD of MaS=4 ± 6%) and a rather inexplicitly-negative control (mean ± SDof MaS=10 ± 8%), were chosen to represent diagnostically easy and difficult sit-uations, respectively. Two radiologists independently measured liver attenuationindex (LAI) on CT, derived and defined as the difference between mean hepaticand mean splenic attenuation, and graded hepatic steatosis using a five-pointscale. Interobserver agreement for visual grading and LAI was determined. Re-ceiver operating characteristic (ROC) analysis was done to compare the accura-cy of the two methods in the diagnosis of ≥ 30% MaS.Results: Weighted kappa values for visual grading were 0.972 and 0.957 in easyand difficult situations, respectively. Intraclass correlation coefficients for LAI were0.970 and 0.962 in easy and difficult situations, respectively. In easy situations,areas under the ROC curve (AUCs) for visual grading and LAI were 0.947 and0.960, respectively, showing no significant difference (P=.583) and, in difficultsituation, AUCs were 0.927 and 0.929, respectively, showing no significant differ-ence (P=.973).Conclusion: Both visual grading and LAI showed high accuracy and reliability inthe diagnosis of ≥ 30% MaS in potential LDLT donors.

B-363 14:18

CT perfusion studies of liver lesions with a 64-detector row scanner: Initialclinical resultsE. Squillaci, S. Fabiano, M. Carlani, M. Di Roma, N. Fusco, G. Simonetti;Rome/IT ([email protected])

Purpose: To explore the possibilities of first-pass perfusion imaging in patientswith focal liver lesions with a 64 channel multidetector row helical CT (MDCT).Methods and Materials: In a total of 20 patients with histologically known liverlesions (3 hemangiomas, 2 FNH, 10 metastases, and 5 HCC) perfusion imagingwith a MDCT (GE Lightspeed VCT) of liver was performed. Scans began simulta-neously with a contrast bolus injection of 60 ml (370 mgI/ml) at a rate of 5 ml/sec.Arterial and portal perfusion values for each ROI were calculated from the timeintensity curves. Blood flow (BF), blood Volume (BV), mean transit time (MTT)and capillary permeability surface area product (PS) were calculated in the nor-mal parenchyma and lesions.Results: Perfusion values of normal liver tissue corresponded to expected phys-iological ranges. A statistically significant difference was demonstrated for BVand MTT between benign and malignant lesions (hemangioma BF: avg. 819.39,dev. 296.62; BV: avg. 25.187, dev. 21.98; MTT: avg. 2.28, dev. 2.20 PS avg. 52.49,dev. 50.61. HCC BF: avg. 265.22, dev. 25.187; BV: avg. 20.04, dev. 20.18; MTTavg. 7.60, dev. 7.06; PS: avg. 38.85, dev. 39.73.cirrhotic liver BF: avg. 27.43, dev.81.41; BV: avg. 2.86, dev. 6.51; MTT: avg 4.37, dev. 8.65, PS: avg. 14.73, dev.31.77. Metastases BF: avg. 34.10, dev. 34.76; BV: avg. 10.181, dev. 10.93; MTT:avg. 20.502, dev. 18.224; PS: avg. 45.66, dev. 45.74).Conclusion: Initial results of our study imply that perfusion MDCT may be help-ful to assess differential diagnosis between liver lesions.

B-364 14:27

An evaluation of cirrhotic recipient's liver function before transplantation byperfusion MR imagingH. Zhang, A. Gu, J. Miao, G. Zhang; Shanghai/CN([email protected])

Purpose: To evaluate if perfusion MR imaging can help assess hepatic functionof cirrhosis recipients before liver transplantation.Methods and Materials: 60 normal adults and 52 pre-liver transplantation recip-ients with cirrhotic liver underwent perfusion MR imaging by using a multi-slice2D FSPGR sequence (6/1.5/60º, 1.0s/acquisition). Artery and portal venous phaseof liver was determined by analyzing the time-intensity curves of abdominal aortaand portal vein, and then positive enhancement integrals (PEI) of each patient's

Scientific Sessions

218 C D E FBA G

right and left hepatic parenchyma were calculated separately in both phases.Hepatic artery fractions [HAF= artery PEI/(artery PEI + portal vein PEI) ×100%]of the two lobes were calculated and lab results of each recipient were collected.Statistics were performed to analyze the HAF and their relationship with lab re-sults.Results: The HAF of right and left hepatic parenchyma are 19.2 ± 4.6%,19.5.0 ± 5.4% in normal and are 25.3 ± 4.9%, 25.3 ± 5.0% respectively in cirrho-sis. The HAF difference between normal and cirrhosis shows statistically signifi-cant (t-test, p < 0.01) in both right and left hepatic lobes. The HAF of cirrhosishave close correlation with ALT, AST, GGT, total bilirubin, albumin/globulin ratio,PT, RBC, PLT levels in right lobe (the correlation coefficients are 0.47, 0.49, 0.44,0.50, -0.46, 0.62, -0.47, -0.53 respectively, p < 0.05) and have close correlationwith ALT, total bilirubin, PT, PLT levels in left lobe (the correlation coefficients are0.49, 0.53, 0.58, -0.49 respectively, p < 0.05).Conclusion: Perfusion MR imaging is useful in evaluating the hepatic function ofcirrhotic recipients before liver transplantation.

B-365 14:36

MR elastography: Non-invasive assessment of liver fibrosisL. Huwart1, F. Peeters1, R. Sinkus2, L. Annet1, N. Salameh1, L. ter Beek3,B. Van Beers1; 1Brussels/BE, 2Paris/FR, 3Best/NL ([email protected])

Purpose: To assess the feasibility of determining the degree of liver fibrosis withMR elastography.Methods and Materials: Forty-two consecutive patients who had liver biopsy forchronic liver disease were included in the study. The degree of fibrosis on thebiopsies was assessed according to the METAVIR scoring system from F0- nofibrosis, to F4- cirrhosis. MR elastography was performed with a 1.5 Tesla scan-ner. Low-frequency (65 Hz) mechanical waves were transmitted into the liver bya transducer placed at the back of the patients. A motion-sensitized spin echosequence, phase-locked to the mechanical excitation was used. The phase mapswere processed to obtain elasticity and viscosity maps. The elasticity and viscos-ity values were compared to the results of the biopsies with a Kruskal Wallis testfollowed by two by two comparisons with Wilcoxon rank-sum tests.Results: The mean elasticity was 2.25 ± 0.22 kPa in the 16 patients without sub-stantial fibrosis (F0 - F1), 3.15 ± 1.39 kPa in the 8 patients with substantial fibro-sis (F2 - F3), and 4.38 ± 1.26 kPa in the 18 patients with cirrhosis (F4). Thedifferences between groups were statistically significant (P < 0.05). The meanviscosity was significantly higher in the patients with cirrhosis (4.57 ± 1.75 Pa·s)than in the patients without cirrhosis (2.12 ± 0.74 Pa·s in F0 - F1 and 2.69 ± 1.66Pa·s in F2 - F3 patients).Conclusion: Non-invasive MR elastography is a feasible method to assess thedegree of liver fibrosis.

B-366 14:45

In vivo phosphorus-31 magnetic resonance spectroscopy of the humanliver in patients with cirrhosisS.S. Bagnenko, G. Trufanov, V. Fokin; St.Petersburg/RU ([email protected])

Purpose: To investigate the value of in vivo phosphorus-31 magnetic resonancespectroscopy (31P-MRS) in the assessment in patients with cirrhosis using 1.5-T scanner.Methods and Materials: Nineteen patients with compensated cirrhosis wereexamined with 31P-MRS, which was performed in the right hepatic lobe. Refer-ence data were acquired from 20 healthy volunteers who had no history or evi-dence of liver disease. Spectra were acquired using a 1.5 T spectroscopy system.Peak area ratios of phosphomonoesters (PME), inorganic phosphate (Pi), andphosphodiesters (PDE) and Adenosine triphosphate (ATP) were calculated.Results: The PME/PDE (0.82 ± 0.5), PME/Pi (5.08 ± 2.7) ratio and the PME sig-nal height ratio were significantly higher in patients with cirrhosis compared withcontrol group (0.25 ± 0.15 and 1.7 ± 0.91 respectively). These changes suggestregenerative activity in cirrhotic liver. Furthermore in patients with small liver thePDE/βATP (0.66 ± 0.14) was significantly lower than the healthy volunteers(3.28 ± 1.19), while this ratio in patients with hepatomegaly was normal(3.28 ± 1.1). These changes reflecting the energy state in the cells.Conclusion: The results of this study indicate that biochemical changes in he-patic energy and phospholipid metabolism exist in patients with cirrhosis and31P-MRS can help us to detect them. Of course a longitudinal study with largernumber of patients is required to better confirm these findings and to understandthe biochemical changes underlying this phenomenon.

B-367 14:54

Diffusion-weighted MR in evaluation of liver fibrosis: Role of apparentdiffusion coefficients (ADC) measurement in diagnosis and determinationof a threshold for pathologyR. Girometti, A. Furlan, G. Como, M. Bazzocchi, F. Toso, C. Zuiani; Udine/IT([email protected])

Purpose: To investigate diagnostic performance of ADC measurement as mark-er of hepatic fibrosis in cirrhotic patients and to establish a threshold for patholo-gy.Methods and Materials: 12 cirrhotic patients with histologically proven liver fi-brosis (8 male and 4 female; mean age 59.6; 11 A and 1 B Child-Pugh score) and13 healthy volunteers (9 male and 4 female; mean age 43) underwent Diffusion-weighted MR of the upper abdomen using a 1.5 T magnet (maximum gradientstrength 40 mT/m), equipped with a six-elements phased-array body coil. Diffu-sion study with parallel imaging (GRAPPA) was performed with a single breath-hold, single-shot Spin Echo Echo-planar sequence (TE 74 msec), using increasingb-values: b= 0 sec/mm2, b= 150 sec/mm2, b= 250 sec/mm2, b= 400 sec/mm2 andapplying a unidirectional diffusion gradient along slice direction. ADCs of hepaticparenchyma were calculated on ADC maps.Results: ADC was significantly lower in cirrhotic livers than in controls (mean1.12 ± 0.17 vs 1.54 ± 0.11 x 10-3 mm2/sec) (p < 0.0001). ROC analysis showedan area under the curve (AUC) of 0.949 (C.I. 95%: [0.84; 1.05]), showing bestsensitivity and specificity (respectively 91% and 100%) for an ADC cut-off of 1.34x 10-3 mm2/sec.Conclusion: Liver Diffusion-weighted MR is a potentially highly accurate tool inevaluating and monitoring liver fibrosis in cirrhotic patients. Using a single-shotSE Echo-planar sequence with maximum b= 400 mm/sec2, the best ADC thresh-old in discriminating liver fibrosis resulted 1.34 x 10-3 mm2/sec.

B-368 15:03

Comparison of two single-shot spin-echo echo-planar diffusion-weightedsequences with increasing b-values in diagnosing liver fibrosisR. Girometti, A. Furlan, G. Como, M. Bazzocchi, F. Toso, C. Zuiani; Udine/IT([email protected])

Purpose: To determine the agreement of two Diffusion-weighted sequences withincreasing b-values in diagnosing liver fibrosis, in order to assess optimal b-val-ues set.Methods and Materials: Diffusion study on 12 cirrhotic patients (histologicallyproven fibrosis) and 13 healthy volunteers was performed on a 1.5 T magnetusing a breath-hold, single-shot SE echo-planar sequence with two sets of in-creasing b-values, respectively: 0 sec/mm2, 150 sec/mm2, 250 sec/mm2, 400 sec/mm2 (first sequence); 0 sec/mm2, 100 sec/mm2, 200 sec/mm2, 400 sec/mm2, 600sec/mm2, 800 sec/mm2 (second sequence). Hepatic parenchymal ADCs werecalculated on ADC maps. Signal-to-noise ratio was measured at b=400 sec/mm2

and b=800 sec/mm2 images.Results: A significant difference in ADCs was found comparing cirrhotic patientsto controls with the first sequence (ADCs lower in cirrhotic than in controls) (mean1.12 ± 0.17 vs 1.54 ± 0.11 x 10-3 mm2/sec) (p < 0.0001), but not with the second(mean 0.93 ± 0.15 vs 1.06 ± 0.21 x 10-3 mm2/sec) (p > 0.05). First and secondsequences significantly differed in ADC measurements (p < 0.0001) and in diag-nostic performance at ROC analysis, with AUC respectively of 0.94 (p < 0.001)(best cut-off: 1.34 x 10-3 mm2/sec) and 0.71 (p > 0.05). According to Bland-Alt-man analysis, the limits of agreement between sequences were 0.76 and 0.08(C.I.95%:[0.25; 0.42]); interclass correlation coefficient resulted moderate 0.71(C.I. 95%:[0.34; 0.87]). Signal-to-noise ratio significantly decreased in secondsequence compared to first sequence (mean 20.98 vs 25.07) (p < 0.0001).Conclusion: Diffusion-weighted sequence with maximum b-value = 400 sec/mm2

is notably superior to the sequence with maximum b-value = 800 sec/mm2 in dis-criminating liver fibrosis, probably due to signal loss as b-values increase.

B-369 15:12

All in one diagnostic work-up of the liver: Evaluation of liver parenchyma,vascular and biliary system using double contrast technique in MR imagingW. Schwarz, R.M. Hammerstingl, C. Herzog, T. Lehnert, K. Eichler, A. Wetter,D. Proschek, T.J. Vogl; Frankfurt am Main/DE ([email protected])

Purpose: To evaluate prospectively a new concept in the diagnosis of liver dis-ease using a combined MR protocol including double contrast-enhanced MRimaging (Gd-Chelate post SPIO) as well as MR angiography and MR cholangi-ography.Methods and Materials: 50 patients suffering from focal liver lesions (n = 35)

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 219C D E FA G

and cirrhosis (n = 15) were investigated using a 1.5 Tesla MR imaging scanner.Sequence protocol included unenhanced (T2-w TSE, T1-w FLASH 2D, MRCP)as well as dynamic (T1-weighted) and static SPIO-enhanced studies. MR angi-ography was performed 10 minutes post SPIO-injection using a double contrasttechnique (i.v. bolus of Gd-DTPA post SPIO). Delayed static double contrast T1-weighted imaging was performed to conclude the study. Three independent radi-ologists evaluated raw data and 3D-images with regard to detection andcharacterization of liver lesions, liver parenchyma as well as vascular and biliarysystem.Results: The combined MR-protocol was well tolerated in all patients. Good cor-relation with intraoperative findings was observed in 21 patients. Best imagingprotocol for detection of lesions was the SPIO-enhanced T2-weighted sequence(95%). Improved characterization was documented using dynamic SPIO-enhancedcombined with delayed T1-weighted double-contrast imaging. MR angiographydepicted anatomical variants in 6 patients (right hepatic artery out of superiormesenteric artery) and portal thrombosis in 3 patients. Dilated bile ducts weredelineated in 5 cases.Conclusion: The combined MR-protocol with the use of double contrast tech-nique is an excellent tool for the diagnosis of the liver to identify the exact mor-phology, vascular system, biliary anatomy as well as detection and characterizationof focal liver lesions in one approach.

B-370 15:21

Comparative evaluation of abdominal MR in cirrhosis and schistosomiasisA.S.A. Bezerra1, G. D´Ippolito2, M.V.A. Soares1, D.D. Leopoldino1,G.R. Batista1, R.P. Caldana2, J. Szejnfeld2; 1Brasília/BR, 2São Paulo/BR([email protected])

Purpose: To compare the abdominal MR findings between cirrhotic and schisto-somotic forms of chronic liver diseases.Methods and Materials: Abdominal MR was performed in 24 patients with chronichepatosplenic schistosomiasis and 27 patients with alcoholic or viral cirrhosis(Child-Pugh class A). Images were evaluated for the presence of periportal fibro-sis, irregularity of hepatic contours, fissure widening, heterogeneity of hepaticparenchyma, splenic siderotic nodules, ascites and collateral circulation. Meas-urements of the right, left and caudate lobes of the liver were performed alongwith measurements of the longitudinal, anteroposterior and transverse diame-ters of the spleen allowing for calculation of the splenic index.Results: Chi-square tests showed that schistosomotic patients more frequentlypresented periportal fibrosis and splenic siderotic nodules (P <.001) while as-cites was more frequent in cirrhotics (P <.01). Tukey-Duncan tests showed thatschistosomotics had smaller right lobes (P <.0006) and larger caudate-right loberatios (P <.0079) than cirrhotics and larger splenic diameters and index (P <.01).Conclusion: The presence of periportal fibrosis, splenic siderotic nodules, smallerright lobe and larger caudate-right lobe ratio and splenic diameters can help dif-ferentiate between schistosomotic and cirrhotic forms of chronic liver diseaseson MR imaging.

14:00 - 15:30 Room F1

Chest

SS 704Pulmonary nodules: Detection, CAD and volumetryModerators:S. Diederich Düsseldorf/DEI. Hartmann; Utrecht/NL

B-371 14:00

Lung nodule volume measurement from CT images: Effect of variation inimage plane along Z-axisJ. Goo1, K.T. Bae2; 1Seoul/KR, 2St. Louis, MO/US ([email protected])

Purpose: To assess the measurement variability in lung nodule volume that isassociated with the image plane.Methods and Materials: 121 non-calcified pulmonary nodules (≤ 30 mm) from44 consecutive patients were included. CT scans were acquired using a 16-de-tector row CT scanner with 16 × 0.75-mm collimation. Two sets of overlappingimages were reconstructed by two section thicknesses (1-, and 3-mm) at 0.2-mmreconstruction intervals. From these highly overlapping data sets, at each sec-tion thickness, we sub-sampled and prepared multiple contiguous, non-overlap-ping image sets by shifting the z-position of the images. For the non-overlapping

data set of 1-mm section thickness, the starting point in z-axis was either 0 or0.6 mm. Likewise, for the non-overlapping data set of 3-mm section thickness,the starting point in z-axis was either 0 or 1.6 mm. Thus, in each set, a nodule iscentered at a different image plane. A commercially available software package(Vitrea 2; Vital Images) was used for the volume measurements. The volumemeasured from two different image planes along the z-axis was compared fortwo section thicknesses respectively with a paired t-test.Results: Due to failure of measurements in some pairs, 82 pairs of 1-mm sectionthickness and 71 pairs of 3-mm section thickness were compared. There were nostatistically significant differences between two different image planes along thez-axis for both 1-mm and 3-mm section thickness (P > 0.05).Conclusion: Measurement variability by different image planes along the z-axiswas not a significant factor in volumetric measurement of lung nodules.

B-372 14:09

Inadequacy of manual measurements compared to automated CTvolumetry in assessment of treatment response of pulmonary metastasesusing RECIST criteriaK. Marten1, F. Auer1, S. Schmidt1, G. Kohl2, E.J. Rummeny1, C. Engelke1;1Munich/DE, 2Forchheim/DE ([email protected])

Purpose: Comparison of relative values of manual uni-dimensional measure-ments (MD) and automated volumetry (AV) for longitudinal treatment responseassessment in patients with pulmonary metastases.Methods and Materials: 50 consecutive patients with pulmonary metastasesand repeat chest multidetector-row CT (median interval=2 months) were inde-pendently assessed by two radiologists for treatment response using ResponseEvaluation Criteria in Solid Tumours. Statistics included relative measurementerrors (RME), intra-/interobserver correlations, limits of agreement (95%LoA),and Kappa.Results: 202 metastases (median volume = 182.22 mm3; range = 3.16-5195.13 mm3) were evaluated. RMEs were significantly higher for MD than forAV (intraobserver RME=2.34- 3.73% and 0.15-0.22% for MD and AV, respective-ly; p < 0.05. Interobserver RME=3.53- 3.76% and 0.22-1.19% for MD and AV,respectively; p < 0.05). Overall correlation was significantly better for AV than forMD (p < 0.05). Intraobserver 95%LoA were -1.85-1.75 mm for MD and -11.28-9.84 mm3 for AV. The interobserver 95%LoA were -1.46-1.92 mm for MD and -11.17-9.33 mm3 for AV. There was total intra-/interobserver agreement on responseusing AV (κ=1). MD intra- and interobserver agreements were 0.73-0.8 and 0.77-0.80, respectively. 28/200 MD response ratings (14/50 patients) were discordant.Agreement using MD dropped significantly from total remission to progressivedisease (p < 0.05).Conclusion: AV allows for better reproducibility of response evaluation in pulmo-nary metastases and should be preferred to MD in these patients.

B-373 14:18

Volumetric evaluation of therapy response in patients with lung metastases.Preliminary results with a computer system (CAD) and comparison withunidimensional measurements.F. Fraioli, L. Bertoletti, A. Napoli, F. Calabrese, C. Catalano, R. Passariello;Rome/IT ([email protected])

Purpose: To assess the efficacy of a CAD system in the identification of lungmetastases and to compare CAD volumetric with observer unidimensional meas-urements in the evaluation of treatment response in oncology patients.Methods and Materials: Two observers (A, B) evaluated 9 patients undergoinglung CT just before and immediately after having received chemotherapy. TheMDCT scans were performed before and after the injection of contrast media,with a high resolution protocol (collimation 4x 1 mm, 100 mAs, 120 Kv). RECISTcriteria have been used to consider the disease as stable, increased, or decreased.Subsequently, target lesions (most significant lesions identified before and afterchemotherapy) were evaluated with a CAD system to establish volumetric meas-urements. Observers' unidimensional and CAD volumetric measurements wereanalyzed for comparison.Results: 24 nodules (DT:5-18 mm in the first study and 4-20 mm in the follow-upstudy) were included in this study. Observers agreed in the assessment of thera-py response in 21 nodules:8 were considered increased in dimensions and 13were judged stable. Observers and CAD measurements disagreed in three nod-ules; two considered as stable by the radiologists and increased by CAD; oneconsidered increased by the radiologists and stable by CAD. In judging patientsresponse, radiologists disagreed in two cases. CAD and observers did not agreein one patient.Conclusion: Our preliminary data suggest that volumetric measurements canmodify diagnostic and therapeutic evaluation of oncology patients under chemo-

Scientific Sessions

220 C D E FBA G

therapy. CAD volumetric measurements allow an easy and objective evaluationreducing interobserver variability in the evaluation of chemotherapy response.

B-374 14:27

Effect of 3D wavelet image compression on computer aided detection(CAD) lung nodule volumetryP. Raffy1, Y. Gaudeau2, D.P. Miller3, J.-M. Moureaux2; 1Sunnyvale, CA/US,2Nancy/FR, 3San Francisco, CA/US ([email protected])

Purpose: To assess the effect of 3D lossy image compression on CAD nodulevolume measurement.Methods and Materials: Two MDCT chest examinations, one normal and onepathological (emphysema) were acquired at 16 x 1.25 mm with a 0.6 mm recon-struction interval. 89 computer-generated nodules with known volumes (31.7-896.7 mm3) and mirroring characteristics of solid nodules typically found in chestCT examinations were inserted at random solitary, juxtavascular or sub-pleurallocations in the datasets. CT scans were compressed using the 3D SPIHT com-pression method to 24:1, 48:1, 96:1 levels. A repeated measure ANOVA modelwas fit to evaluate the impact of compression level and nodule characteristics onCAD volume estimates.Results: Mean nodule volume measurement error was +3.4 % ± 7.6 % [SD] inuncompressed datasets, and +0.5 % ± 9.1 % [SD] at level 96:1. No significantdifference in volume measurement error for any of the compression levels wasidentified. Even though compression did not predict the size of the volumetricerrors, 96:1 level did yield statistically significantly smaller volumes on average(average volume loss 4.1 mm3) than the other amounts of compression. Nodulelocation was found to be a significant predictor of volume measurement errorwith larger percent errors on average for juxtavascular (+7.1 %) and sub-pleuralnodules (+1.6%). Nodule size and presence of emphysema were not identifiedas significant predictors of volume measurement error.Conclusion: CAD volume estimates of computer-generated nodules in lossycompressed data showed no significant difference up to 96:1 when comparedwith known volumes in reference database.

B-375 14:36

Computer aided detection (CAD) of solid lung nodules in lossy compressedMDCT chest examinationsP. Raffy1, Y. Gaudeau2, D.P. Miller3, J.-M. Moureaux2; 1Sunnyvale, CA/US,2Nancy/FR, 3San Francisco, CA/US ([email protected])

Purpose: To assess the effect of 3D lossy image compression on computer aid-ed detection.Methods and Material: 120 cases were selected from MDCT chest scans ac-quired with collimation of 1.25 mm, and collected from different sites. The data-base was composed of 66 low-dose cases and 109 cases with a high-frequencyreconstruction kernel. Two expert radiologists established that 37 cases had noactionable nodules and the remaining 83 cases contained 169 nodules. The meannodule size was 5.8 mm ± 3.0 [SD] (3.8 - 35 mm). All cases were compressedusing the 3D SPIHT algorithm to 24:1, 48:1, 96:1 levels. A study of the effect ofcompression on sensitivity was performed at different operating points (2.5, 5,10 FP/case) using McNemar's test. Logistic regression models were used to eval-uate the impact of compression level, nodule and scanner characteristics on CADsensitivity.Results: Compared to no compression, there was no significant degradation insensitivity found at any of the proposed compression levels and operating points.However, between compression levels there was marginal association with sen-sitivity. Specifically, 24:1 level was significantly better than 96:1 at all operatingpoints, and occasionally better than no compression at 10 FP/case. Based onmultivariable analysis, nodule location was found to be a significant predictorwith a lower sensitivity associated with pleural nodules. Nodule size, dose, re-construction filter, and contrast medium were not significant predictors.Conclusion: CAD detection performance of solid lung nodules was found to berobust across all compression levels up to 96:1.

B-376 14:45

Pulmonary nodules detection with MDCT: Comparison of thin axial images,maximum-intensity-projections (MIP) and computer-assisted detection(CAD)A. Jankowski, T. Martinelli, J. Timsit, M. Coulomb, G. Ferretti; Grenoble/FR([email protected])

Purpose: To compare 3 different methods for detection of lung nodules with MDCTin subjects included in a lung cancer screening study (DepiSCAN).Methods and Material: 28 subjects were randomly selected from the database.

CTs were acquired with a 16-row MDCT using 1 mm slice thickness. Three radi-ologists independently reviewed in a random order on a workstation 3 types ofimages: axial 1-mm images, axial 6-mm MIP, CAD system's detections (Lung-Care, Siemens). An independent experienced chest radiologist decided for eachdetection method which nodules were true positive nodules.Results: 287 nodules of 711 (40.4%) nodules ≥ 1 mm were identified as truepositive. The detection rates of the three independent observers with 1-mm axialimages were 64/287 (22.3%), 86 (30%), and 136 (47.3%). When MIP imageswere added to 1-mm images, these rates were 123/287 (42.9%), 140 (48.8%),and 205 (71.4%) (p < 0.0001). When CAD was added to 1-mm images, theserates were 145/287 (50.5%), 154/287 (53.7%), and 190/287 (66.2%) (p < 0.0001).CAD detected 298 opacities; 107 of them were true positive nodules (detectionrate: 107/287; 37.3%). CAD showed 6.82 false positive per CT scan. 31 noduleswere detected by CAD only; they were smaller than other nodules (2.24 ± 0.9 vs3.14 ± 1.7, p < 10-4).Conclusion: MIP and CAD reduced the number of overlooked nodules on 1-mmimages. The additional nodules detected by these 2 techniques were different.

B-377 14:54

Is the computer-aided detection scheme for lung nodules also effective indetecting lung cancer?J. Goo1, H. Kim2, J. Lee3, H. Lee1, C. Lee1, K. Lee4; 1Seoul/KR, 2Koyang-si/KR,3Daejeon/KR, 4Seongnam-si/KR ([email protected])

Purpose: To evaluate the performance of a computer-aided detection (CAD) sys-tem in the detection of both lung nodules and lung cancer.Methods and Materials: 150 chest CTs were evaluated. Database included 23lung cancers (long diameter < 20 mm), nodules that were stable for at least 2years, and normal cases. 5 chest radiologists and 5 radiology residents inde-pendently recorded the locus of each nodule candidate, and assigned each con-fidence scores for the likelihood of nodule and malignancy without CAD; thenthey repeated analysis with use of CAD. Jackknife alternative free-response re-ceiver operating characteristic (JAFROC) analysis was used to compare the ob-servers' performance of detecting nodules and cancer with and without CAD.Results: The JAFROC analysis demonstrated that the performance of detectinglung nodules was significantly increased with CAD for all radiologists and sub-groups of chest radiologists and radiology residents (P < 0.01). CAD itself de-tected 18 of 23 lung cancers. Four lung cancers missed by three radiology residentson initial reading were additionally detected with CAD. However, because thenumber of false positive detections for lung cancer increased with the use ofCAD, the overall performance of detecting lung cancers was not significantly dif-ferent with and without CAD for all radiologists and subgroups (P > 0.05).Conclusion: The detection rate of lung nodules was significantly improved withCAD. However, the overall performance of detecting lung cancer by radiologistswas not significantly affected with CAD, although the use of CAD contributed todetecting additional lung cancers by less experienced radiologists.

B-378 15:03

CAD detection of early lung cancer: Performance in conventional dose CTcompared with low dose CT of the thoraxM.E. Roddie1, H. Roberts2, G. Bastarrika3, J. Zulueta3, X. Ye1, J. Dehmeshki1;1London/UK, 2Toronto, ON/CA, 3Pamplona/ES ([email protected])

Purpose: Assessment of computer-aided detection (CAD) software performancein detection of small, biopsy proven lung cancers from a lung cancer-screeningprogram compared with those found on conventional, diagnostic MDCT of thethorax.Methods and Materials: Twenty-seven multi-detector row CTs of the thorax withsmall biopsy proven lung cancers were collected; 16 cases from a low-dose CTlung cancer screening program (size range 7-19 mm, median diameter 15.5 mm,10/16 solid) and 11 detected as indeterminate or suspicious nodules on conven-tional body CT (size range 7-23 mm, median diameter 18 mm, 11/11 solid). Do-nor institutions provided reference truth coordinates and histological reports forthe cancers and the CTs were analysed with commercially available CAD soft-ware optimised for detection of spherical, solid nodules 5 - 20 mm diameter (LungCAD, Medicsight plc, London) using 12 different CAD filter settings.Results: For conventional dose thoracic CT (mA 100 -200) CAD detected be-tween 8-9/11 (73 - 82%) of cancers while on low-dose CT (mA 40-60) only 3-5/16(19 - 31%) of cancers were detected (p < 0.001). Removal of the non-solid andpart-solid cancers in the screening group from the analysis reduced CAD lungcancer detection to 1/9 (11%) indicating that the CAD algorithm was more affect-ed by mA and noise than by cancer density.Conclusion: The performance of a CAD algorithm for detection of small lungcancers on conventional diagnostic CT varies when applied to low-dose lung

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 221C D E FA G

screening CT suggesting that optimized algorithms differ according to the clinicalscenario.

B-379 15:12

Sensitivity of computer-aided detection software for pulmonary nodules inlow dose vs. full dose MDCT chest examinationsM. Das1, M. Salganicoff2, G. Mühlenbruch1, A.H. Mahnken1, J. Stoeckel2,A. Bakai3, L. Gündel3, R.W. Günther1, J.E. Wildberger1; 1Aachen/DE, 2Malvern,PA/US, 3Forchheim/DE ([email protected])

Purpose: To compare the sensitivity of a Computer-aided detection (CAD) soft-ware for pulmonary nodules in low dose vs. full dose settings.Methods and Materials: 20 routine chest examinations from patients with knownintrapulmonary nodules were chosen. All examinations were acquired on a 16-slice MDCT scanner (SOMATOM Sensation 16, Siemens) with a standard proto-col (120 kV, 100 mAseff., 0.5 s rotation time, 16x1.5 mm collimation, table feed/rotation 24 mm, 2 mm thick sections with 1.5 mm reconstruction interval, recon-struction algorithm B50f). Artificial noise was added to the raw data, simulating adosage of 10 mAseff. A consensus panel of two radiologists, plus all findings fromthe CAD software were used to assess the gold standard.The CAD software prototype (LungCAD, Siemens) was applied to both radiationdose settings, all findings were verified as true positive or false positive and re-sults were compared to the gold standard.Results: In 20 datasets 116 pulmonary nodules were found (range 0-15; average6). At standard dose settings the CAD software found 90 nodules (Overall sensi-tivity: 77.6%). Sensitivity per case was 82%. Average false positive per case was1.8. At low dose settings the CAD software found 64 nodules (Overall sensitivity:55%). Sensitivity per case was 63%. Average false positive per case was 1.5.Conclusion: Computer-aided detection of pulmonary nodules is feasible also inlow dose settings, with overall good detection performance, although detectionperformance is slightly reduced, possibly due to increased image noise in lowdose settings.

B-380 15:21

MSCT performed at a chest radiograph dose: Influence of reader'sexperience on detection and characterization of pulmonary nodulesV.C. Romano, P. Hein, U. Zaspel, G. Diederichs, B. Hamm, P. Rogalla;Berlin/DE ([email protected])

Purpose: Assessment of the influence of reader's experience on detection andcharacterization of pulmonary nodules (PN) in ultra-low-dose MSCT (ULD-MSCT).Methods and Materials: 135 patients referred to chest CT for suspected PNsentered the study. The scoutview was replaced with a chest ULD-MSCT at follow-ing parameters: 3.5eff. mAs, 120 kV and 1 mm collimation (Toshiba-Aquilion),resulting in a radiation-dose of 0.2-0.3 mSv (same dose-range as CXR in 2 views).A standard dose MSCT followed with 52.2eff.mAs. Two blinded radiologists (R1-experienced; R2-resident) read axial images first in ULD- and consecutively inSD-technique: number, size, location, morphology and characterization of eachPN was analyzed. If more than 5, PNs were grouped in "many" (6-20), "multiple"(21-50), "diffuse" (> 50) and max./min. Ø was determined. In each technique,subjective diagnostic confidence for detection and characterization was scoredbetween 1-10. After analyzing images in SD-technique, the readers were askedto say whether ULD technique hindered either PN detection or characterization.Results: The findings of SD-technique served as reference standard: sensitivityand specificity of ULD-MSCT were 100%, 95.7% (R1) and 96.3%, 100%(R2)respectively. After evaluating images in SD the readers believed ULD-techniqueto be a hindrance to PN detection (over- and undercounted PNs < 2 mm) andcharacterization in 9.7%(R1), 5.2%(R2) and 12.7%(R1) and 1.5%(R2) respec-tively. Significant differences in reader's subjective confidence in PN detectionand characterization between the 2 methods were found for R1 and R2 (p < 0.05).Nevertheless objective characterization showed very good agreement betweenthe 2 methods for both R1 (k=0.94) and R2 (k=0.96).Conclusion: Even though the less experienced radiologist reacts to a dose re-duction of 93.3% with a subjective diagnostic confidence fall, reader's experi-ence does not influence or compromise PN detection and characterization inULD-MSCT.

14:00 - 15:30 Room G

Cardiac

SS 703Myocardial perfusion and viabilityModerators:S. Ghods; Basle/CHD. Hahn; Würzburg/DE

B-381 14:00

Dobutamine stress magnetic resonance imaging followed by a perfusionsequence optimizes the interpretation of cardiac MR imaging in thedetection of myocardial ischemiaD.D. Lubbers1, D. Kuijpers2, P.R.M. van Dijkman2, J. Overbosch1,C.H.C. Janssen1, T.P. Willems1, M. Oudkerk1; 1Groningen/NL, 2The Hague/NL([email protected])

Purpose: To assess the additional value of a perfusion sequence after DobutamineStress MR imaging (DS-MRI) for the interpretation of cardiac MR imaging in thedistinction between ischemic and non-ischemic segments.Methods: Fifty patients with an inconclusive diagnosis of myocardial ischemia,without a previous history of myocardial infarction or rest wall motion abnormali-ties on baseline DS-MRI were included. DS-MRI was performed on a 1.5 T sys-tem (Magnetom Sonata, Siemens Medical systems, Erlangen, Germany).Short-axis cine and grid levels were scanned at three levels of the left ventricle.Images were acquired during rest and stress, and followed by first pass myocar-dial perfusion. An experienced radiologist and cardiologist graded the images,according to the sixteen segment model by Cerqueira, on a four point scale.Results: Forty-four out of 50 patients (88%) showed no new wall motion abnor-malities (NWMA) and the subsequent perfusion sequence showed absence ofperfusion defects. The remaining 6 patients out of 50 had NWMA, 4 of whomcould be confirmed by a perfusion sequence as myocardial ischemia. Two pa-tients with NWMA did not have perfusion defects.Conclusion: Good agreement exists between the absence of NWMA and myo-cardial perfusion. Furthermore, a perfusion sequence can facilitate to decide if aNWMA is caused by ischemia or is "non-ischemic". Combining DSMR with "first-pass" myocardial perfusion optimizes the interpretation of cardiac MR imaging inthe detection of myocardial ischemia.

B-382 14:09

Guidance of interventional revascularization in patients with coronaryartery disease using myocardial MR stress perfusion imagingS. Miller, T. Hövelboern, B. Klumpp, M. Fenchel, A. May, U. Kramer, M. Gawaz,C.D. Claussen; Tübingen/DE ([email protected])

Purpose: Cardiac MR imaging (MRI) was used in order to detect relevant coro-nary artery stenoses (CAS) and guide re-intervention in patients after myocar-dial revascularization.Methods: Prior to invasive coronary angiography 75 consecutive patients (meanage 64 ± 12, range 24-85y) with recurrent symptoms of myocardial ischemia af-ter interventional or operative coronary revascularization were examined (1.5 TMagnetom Sonata, Siemens, Erlangen). Myocardial perfusion imaging was ob-tained at stress (140µg Adenosin per kg BW per min) and rest using a turbo-FLASH 2D sequence. A dose of 0.1 mmol Gd-DTPA (Magnevist, Schering, Berlin,Germany) was injected at both times. Viability was assessed by delayed enhance-ment imaging (segmented IR turbo FLASH 2D sequence). Qualitative consensusreading of two experienced observers who were blinded for clinical data wasapplied for image analysis. Catheter coronary angiography was used as stand-ard of reference for the diagnosis of hemodynamically relevant CAS (> 75% ofluminal narrowing).Results: Stress induced myocardial malperfusion was detected in 53/75 patientsand corresponding to relevant CAS requiring re-intervention in 43/45 patients(=sensitivity 96%, specificity 70%). In 22 patients absence of relevant CAS wascorrectly diagnosed. Sensitivity/specificity for MR imaging detection of the targetregion corresponding to the ischemia related artery was in a range of 56-94%/80-85%. Delayed contrast enhancement was observed in 49 patients and con-sistent with unsuspected myocardial infarction in 21 cases.Conclusion: Hemodynamically relevant stenoses can be accurately assessed inpatients with known CAD after revascularization. In this group of patients MRimaging prior to invasive coronary angiography is helpful to guide re-intervention.

Scientific Sessions

222 C D E FBA G

B-383 14:18 !Utilization of stress magnetic resonance first pass perfusion imaging inchest pain patientsC. Klasssen, M. Nguyen, K. Kareti, N. Wilke; Jacksonville, FL/US([email protected])

Purpose: The purpose of this study was to use MR imaging to detect perfusionand viability defects in patients with chest pain and report on its utilization in atertiary hospital.Methods: Patients evaluated for chest pain at the University of Florida/ShandsJacksonville with a cardiac MR imaging extending from 05/03 to 03/05 were in-cluded. A Siemens Sonata 1.5 Tesla magnet was used for all studies. Localizerimages were acquired followed by rest perfusion using TurboFLASH. Gadoliniumcontrast (Magnevist 9 ml, 0.03 mmol/kg) was injected with and serial images wereobtained to capture the first pass of contrast through the myocardium. Adenosinewas then infused (140 mcg/kg/min) for 4 minutes and another dose of contrastadministered. After 10 minutes, a delayed enhancement scan was acquired. Ex-ams were interpreted qualitatively and reviewed for perfusion and viability de-fects.Results: 357 patient exams were acquired and reviewed (age was 60 ± 13 yearswith 53% female). 46% of these exams were normal. Ischemia was detected in65 patients (18%) and acute or chronic infarction in 68 patients (19%) additional-ly of these, 14 patients had both ischemia and infarction. Twenty-two (6%) pa-tients had hypoperfusion in a circumferential subendocardial distribution indicatingmicrovascular disease.Conclusion: This study demonstrates the potential utility of MR in diverting somepatients from unnecessary coronary catheterization (with 46% being normal) aswell as detecting microvascular disease that would evade detection by angiogram(6%). Additionally, ischemia was detected in 18% that could be referred for fur-ther diagnostic studies including CT angiography or catheterization.

B-384 14:27

Myocardial tagging at 3.0 T: Improved tag persistence and contrast-to-noiseratio in comparison to 1.5 TU. Kramer1, V. Desphande2, M. Fenchel1, B. Klumpp1, C.D. Claussen1,S. Miller1; 1Tübingen/DE, 2Los Angeles, CA/US([email protected])

Purpose: The aim of this prospective study was the optimization of a gradient-echo (GRE) MR tagging sequence at 3.0 T.Methods: 17 consecutive healthy subjects (mean age 43.2 yrs) and 3 patientswere scanned at a 1.5 T and a 3 T whole body scanner (Siemens). A prospec-tively triggered GRE sequence was used. Sequence parameters were as follows:TR 5.2 ms, TE 3.7 ms, matrix 192, field of view 350x350 mm, temporal resolution46 ms. Sequence parameters like flip angle (8-16°), tag size (4-8 mm) and slicethickness (4-8 mm) were varied systematically at both field strengths. For eachparameter, the signal-to-noise (SNR), contrast-to-noise (CNR) and relative con-trast myocardium-to-tag (RCMT) were evaluated individually.Results: While the conspicuity of the tags is greatly reduced in the diastolic phaseat 1.5 T, tags are clearly visible in the myocardium in all phases at 3 T. Based oninitial results, a slice thickness of 6 mm and a tag thickness of 8 mm were foundto be suitable. At 1.5 T, the optimal flip angle was found to be 12 deg while that at3 T was found to be 8 deg based on overall image quality, SNR and tag conspicu-ity (RCMT). Blood SNR was significantly higher at 3 T despite the lower flip an-gle.Conclusion: Myocardial tagging at 3 T has shown a superior image quality incomparison to 1.5 T due to a higher baseline SNR and an improved CNR as wellas RCMT. As a consequence of suppressed fading of the tags, the accessibilityto the diastolic phase of the cardiac cycle becomes possible.

B-385 14:36

Intraindividual comparison of myocardial MR-perfusion imaging at 3.0 T and1.5 TB. Klumpp, T. Hövelborn, M. Fenchel, A. May, U. Kramer, C.D. Claussen,S. Miller; Tübingen/DE ([email protected])

Purpose: Myocardial MR-perfusion imaging (MPI) is increasingly gaining atten-tion for the assessment of hemodynamically relevant coronary stenoses. Short-comings of MPI at 1.5 T are limited SNR and image contrast. The aim of ourstudy was to intraindividually compare MPI at 3.0 T and 1.5 T.Methods: MPI at rest is performed in 20 patients with proven chronic myocardialinfarction at 3.0 T and 1.5 T in our ongoing study. For MPI a Turbo FLASH 2Dsequence (1.5 T/3.0 T: TR 2.2/1.9, TE 1.08/1.04, flip 12°/12°, matrix 94x192/

115x192) is used after injection of 0.05 mmol gadobenate dimeglumine/kg bw at4 ml/s (MultiHance, Bracco-Altana, Konstanz, Germany). Contrast dynamic wasanalyzed visually and quantitatively applying least square fit of the signal-timecurve. SNR measurements were performed in precontrast and contrast enhancedmyocardium. Image quality was rated by two independent observers (score 0-3).Results: The image score was significantly higher at 3.0 T (2.0 ± 0.3/2.4 ± 0.3p < 0.05). SNR of contrast enhanced myocardium was significantly higher at 3.0 T(14.9/20.9, p < 0.005). In three of six patients perfusion deficits at rest correlatingwith myocardial infarction were detected at 3.0 T whereas only two perfusiondeficits were seen at 1.5 T.Conclusion: MPI at 3.0 T yields significantly improved image quality with supe-rior SNR and CNR compared to1.5 T.

B-386 14:45

Assessment of myocardial viability by delayed enhancement MR-imaging:Intraindividual comparison of 3.0 T and 1.5 TB. Klumpp, T. Hövelborn, U. Kramer, A. May, M. Fenchel, C.D. Claussen,S. Miller; Tübingen/DE ([email protected])

Purpose: Determination of myocardial viability using delayed enhancement MR-imaging (DE MR imaging) becomes increasingly relevant. The aim of our studywas the intraindividual comparison of DE MR imaging at 3.0 T and 1.5 T.Method and Materials: In this ongoing study cardiac MR imaging is performedin 20 patients with proven chronic myocardial infarction at 3.0 T and 1.5 T. Forevaluation of myocardial function SSFP sequences are acquired in four chamberview (FC), two chamber view (TC) and short axes views (SAV). 15 min after con-trast injection (0.1 mmol gadobenate dimeglumine/kg bw, MultiHance, Bracco-Altana, Konstanz, Germany) segmented inversion recovery prepared Turbo FLASHsequences (1.5 T/3.0 T: TR 11/9.9, TE 4.4/4.91, flip 30°/30°, matrix 160x256/154x256) are acquired in FC, TC and SAV. Image quality is rated by two inde-pendent observers (score 0-3). Measurement of SNR and CNR is performed ininfarcted and normal myocardium.Results: In our first patients image quality at 3.0 T was significantly higher thanat 1.5 T (score 1.5 T/3.0 T=2.1 ± 0.2/2.6 ± 0.3 (p < 0.05) with high image qualityat both systems. SNR/CNR of infarcted myocardium was 28.2 ± 6.7/24.1 ± 6.5 at3.0 T and 18.4 ± 4.4/14.7 ± 4.2 at 1.5 T (p < 0.001/p < 0.001). SNR of normalmyocardium did not differ significantly (4.1 ± 0.9 at 3.0 T and 3.7 ± 0.7 at 1.5 T(p > 0.05)).Conclusion: DE MR imaging at 3.0 T yields significantly increased SNR, CNRand image quality compared to 1.5 T allowing potentially for acceleration of ac-quisition time higher spatial resolution or reduction of contrast agent.

B-387 14:54

Myocardial viability assessment using 2-D and 3-D wall motion analysis ingated SPECT with reference to cardiac MR imaging in patients with highlyimpaired left ventricular functionS. Mehl, H. Amthauer, M. Froehlich, M. Grothoff, B. Spors, J. Ruf, R. Felix,M. Gutberlet; Berlin/DE ([email protected])

Purpose: Wall motion provides important additional information for the viabilityassessment of left ventricular myocardium. A valid method for standardized wallmotion analysis (WMA) in combination with perfusion analysis using gated SPECTwas tested.Methods: Myocardial viability was assessed in 540 segments/45 patient studies(mean age 64.7 ± 7.6) performing 2-D and 3-D wall motion analysis using gatedSPECT - all patients with CAD and highly impaired left ventricular function (< 35%EF). "Low dose" dobutamine stress MR imaging WMA of each segment afterrevascularisation represented the gold standard. In the 3-D cine mode each seg-ment was defined as viable when showing proper or hypokinetic movement, where-as in the 2-D short axis view the partial volume effect during the endystolic phasewith visual uptake augmentation was defined as a sign of viability.Results: For every method sensitivity (SE), specificity (SP), positive (PPV) andnegative predictive values (NPV) were calculated. For 2D-WMA we found: SE=0.9;SP=0.55; PPV=0.9; NPV=0.57 compared to SE=0.85; SP=0.73; PPV=0.94;NPV=0.52 for 3D-WMA. For perfusion analysis (cutoff > 55% for viability) thefollowing results were evaluated: SE=0.87; SP= 0.65; PPV: 0.91; NPV= 0.54. Forcombined perfusion and 2D-WMA we found SE= 0.92; SP=0.35; PPV= 0.86;NPV=0.51 and for perfusion and 3D-WMA there was SE=0.98; SP= 0.27;PPV=0.85 and NPV=0.74.Conclusion: Both WMA methods using SPECT, especially if combined with per-fusion analysis represent an extremely sensitive tool for the assessment of myo-cardial viability even in patients with severe CAD and highly impaired left ventricularfunction.

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 223C D E FA G

B-388 15:03 !Myocardial viability assessment using multidetector CT: Comparison withsingle photon emission computed tomography and FDG-positron emissiontomographyI.H. Lee, Y. Choe, K.-H. Lee; Seoul/KR ([email protected])

Purpose: To evaluate capability of MDCT in detection of nonviable myocardiumin patients with myocardial infarction (MI).Methods: The study included 19 patients with acute (n = 13) (within 1 month)and chronic (n = 6) (mean, 8.3months; range, 4-27 months) MI. 16-slice-MDCTwas performed at early phase and 10 minutes after administration of 120 ml no-nionic contrast media. SPECT was performed in 19 patients with 201Tl (n = 5) or99mTc- sestamibi (n = 14). FDG-PET was performed in 10 patients during acutephase and in 2 patients during chronic phase. Short axial images were recon-structed from CT for comparison with SPECT and PET. CT, SPECT and PETimages were analyzed using a 17-segment model. Myocardial damage on CTwas defined as delayed hyperenhancement. Severity of myocardial damage wasgraded by four-point scale according to depth of involvement. Grade 3 was de-fined as nonviability.Results: MDCT and SPECT were concordant in localization of MI in 85.8% (277/323 segments) and discordant in 14.2% (46/323). MDCT and PET were concord-ant in localization of MI in 89.2% (182/204) and discordant in 10.8% (22/204). Inconcordant segments, MDCT overestimated severity in 16.2% compared withSPECT and 9.9% compared with PET while MDCT underestimated severity in5.1% (SPECT) and 8.2% (PET). Sensitivity, specificity and diagnostic accuracyof MDCT in determining nonviable segments were 68.5%, 86.5%, 81.4% as com-pared with PET and 67.3%, 84.3%, 83.6% as compared with SPECT, respective-ly.Conclusion: MDCT is useful in detecting myocardial infarction and in determina-tion of nonviable segments.

B-389 15:12

Exercise training in patients with chronic heart failure: Effects onmyocardial energy metabolism, RV and LV morphology and function asassessed by quantitative 31P-MR-spectroscopy and MR imagingM. Beer1, D. Wagner2, J. Sandstede1, H. Köstler1, P. Dubach2, D. Hahn1;1Würzburg/DE, 2Chur/CH ([email protected])

Purpose: The beneficial role of exercise training (ET) is increasingly acknowl-edged for patients with non-ischemic cardiomyopathy. Using MR-imaging (MRI)and a quantitative 31P-MR-spectroscopic technique (SLOOP-MRS), we investi-gated possible improvements of long-term ET in patients with dilated cardiomy-opathy (DCM).Methods and Materials: 22 patients were randomized to an exercise (Ex) or acontrol (Co) group. Examinations were performed three times, at baseline and at2-and 8-month follow-up periods. The Ex group had 5 supervised training ses-sions/week lasting 40 minutes. Concentrations of phosphocreatine (PCr) andadenosinetriphosphate (ATP) were determined by SLOOP-MRS, right and leftventricular parameters were measured with cine MR imaging.Results: ET increased the peak VO2 in Ex patients (13.2 ± 4.0 to 19.3 ± 5.3 ml/kg/min; p < 0.01). Eight months of ET decreased LV end-systolic volume (150 ± 66to 136 ± 65 ml; p < 0.05) and increased LV EF (33 ± 15 to 37 ± 14 %; p < 0.01).No changes were observed in RV parameters. 36% of patients in the Ex groupshowed a more than 50% increase of PCr correlating to an increase of LV EF.Conclusion: Long-term exercise training improves LV systolic function in pa-tients with DCM. In more than a third of the patients a concomitant improvementof energy metabolism was observed. Thus, myocardial work appears to be moreeconomically performed after ET.

B-390 15:21

Impact of parallel imaging (iPAT) on image quality and visual vs.semiquantitative evaluation of MR perfusion of the myocardiumA.M. Huber, M. Schweyer, S. Schönberg, O. Mühling, M. Näbauer, M.F. Reiser;Munich/DE

Purpose: To assess the ability of parallel imaging with iPAT (GRAPPA, AF: 2) forimprovement of image quality and diagnostic accuracy of MR perfusion of themyocardium.Methods: 64 patients were examined with coronary angiography and MR imag-ing. 34 patients were examined with a SR turboFLASH sequence under adenos-ine infusion (140µg/min/kg over at least 3 min) and at rest (128 phase encodinglines, 192 msec acquisition window, 3 slices). 32 patients were examined with aSR turboFLASH sequence combined with parallel imaging (iPAT: GRAPPA-algo-rithm, AF 2, 128 lines, 150 msec acquisition window, 4-5 slices). Two experi-

enced radiologist assessed the images for image quality (susceptibility artifacts,ECG-artifacts etc) and for detection of hemodynamically relevant coronary ar-tery stenosis (75%) using a five-point scale. In addition, semiquantitative evalua-tion of the image data sets was performed.Results: The use of iPAT reduces the percentage of susceptibility artifacts from15.1% of the segments of the myocardium to 4.8 %. ECG-artifacts from 15.6% to11.5%. The values for sensitivity, specificity, accuracy and area und curve (ROC-analyses) for the visual evaluation were 69%, 83%, 79% and 0.84 for SR turbo-FLASH and 93%, 89%, 92% and 0.97 for the SR turboFLASH technique withiPAT. The values for the semiquantitative evaluation were 100%, 83%, 95% and0.95 without and 88%, 100%, 92% and 0.97 with iPAT.Conclusion: IPAT improves the image quality and diagnostic accuracy of a visu-al evaluation by reduction of artefacts.

14:00 - 15:30 Room H


SS 709Embolization therapyModerators:M.-S. Grammenou-Pomoni; Athens/GRM. Köcher; Olomouc/CZ

B-391 14:00

Preoperative gastric embolization: A means to reduce the incidence ofesophagogastric leaks in cervical esophagoplastyE. Alba, E. Escalante, J. Dominguez, C. Sancho, S. Lamas, M. Sanz;L´Hospitalet de Llobregat/ES ([email protected])

Purpose: Anastomotic leakage after esophagoplasty using a gastric tube is oneof the most serious complications. One of the major causes is the sudden de-crease in gastric blood supply leaving the right gastroepiploic artery as the solesupply of the gastric tube. Our purpose was to evaluate the effect of embolizationin preoperative redistribution of gastric blood supply in the prevention of anasto-motic leakage.Methods and Materials: The study was prospective from July 2002 to November2004. Transarterial embolization (TAE) of the left gastric artery (LGA), right gas-tric artery (RGA), and splenic artery (SA) was performed in 17 consecutive pa-tients. The procedure was carried out 2-3 weeks prior to surgery. Coeliacarteriography was performed before and after embolization with 0.035'' platinumcoils and 0.018'' microcoils through a 4 or 5 F catheter and a 2.7 Fr coaxial micro-catheter. Embolization was considered successful when the LGA, RGA and SAwere occluded. The results were compared with those previously obtained in ourinstitution in 100 patients who underwent this type of surgery without TAE: 14cases of anastomotic leakage and one death.Results: In 16 out of 17 patients embolization was successfully performed. Itwas unsuccessful in 1 patient because of anatomical difficulties. There were noearly complications. Many patients experienced mild local pain, nausea or vom-iting. One developed acute self limited pancreatitis. No anastomotic leak wasobserved.Conclusion: Preoperative vascular redistribution by TAE is a safe procedure thatcontributes to decreasing the incidence of postoperative anastomotic leakageafter esophageal reconstruction.

B-392 14:09

Superselective preoperative embolization of bone neoplasia facilitatesextended tumour resection and stabilizationC. Waldherr, R. Kickuth, K. Ludwig, M. Beck, K. Siebenrock, J. Triller;Berne/CH ([email protected])

Purpose: The aim of this study was to evaluate the impact of preoperative devas-cularization of hypervascular bone tumours before resection and orthopaedicstabilization in relation to the degree of postembolization tumour vascularity andthe following intraoperative blood loss.Methods and Material: We retrospectively analyzed charts and images of 17consecutive patients (m: 11, f: 6; median age: 65y, range: 31-85y), who under-went preoperative transluminal embolization of long bone, hip and spinal tumoursbefore resection and stabilization. The following bone tumours were identified:Metastases of renal cancer (n = 12), of malignant melanoma (n = 1), of leiomy-osarcoma (n = 1), of hepatoma (n = 1), of prostate cancer (n = 1) and a giant celltumour (n = 1). Embolization was performed with polyvinyl alcohol particles(n = 13), with coils (n = 1), and with both coils and particles (n = 3). Each tumour

Scientific Sessions

224 C D E FBA G

was assigned a postembolization vascularization grade (I-III) according to tu-mour blush after contrast agent injection in the main feeding arteries.Results: Intraoperative blood loss was correlated with the degree of postembol-ization vascularization. In patients with complete devascularization (grade I; n = 8),it was median 500 ml (range: 200-4000 ml), whereas in patients with partial em-bolization (grade II; n = 8), it was 1750 ml (range: 350-3500) and 2500 ml (gradeIII, n = 1). The mean intraoperative blood loss was 1585 ml. Embolization wascomplicated by tissue necrosis in 1 patient.Conclusion: Preoperative embolization of bone neoplasia is an effective methodto reduce intraoperative blood loss, which facilitates and enables extended tu-mour resection and bone stabilization.

B-393 14:18

Segmental hepatic arterial embolization for the induction of contra-lateralliver hypertrophy: Experiences with a porcine modelA. Koops, C. Wilms, J. Herrmann, N. Heits, S. Brilloff, D.C. Broering, G. Adam,G. Krupski; Hamburg/DE ([email protected])

Purpose: Transarterial embolization for preoperative induction of contra-lateralliver hypertrophy had been proposed as an alternative to portal vein emboliza-tion. The aim was to evaluate clinical outcome and efficiency of this technique.Methods and Materials: In 11 mini-pigs all but the right-lateral segmental hepat-ic arteries were embolised with lipiodol/histoacryl-mixture using transfemoralaccess. In-vivo angiography and ex-vivo specimen angiography were repeatedat termination after 4 weeks. Arterial patency and flow-parameters were obtainedby intraoperative ultrasound before and after the embolization and after 4 weeks.Results: The radiological intervention was successful in 10/11 cases. In one an-imal subtotal occlusion of all arterial branches occurred. To avoid right-lateralarterial occlusion, proximal minor sub-segmental branches of the right-middlesegment remained patent in 30% of angiograms and 40% of ultrasound exami-nations performed peri-interventionally. After 4 weeks all 9 surviving animals haddeveloped extensive collaterals with arterial reperfusion of all segments in angi-ography as well as ultrasound. The animal with subtotal embolization died aftertwo days of liver failure, one other animal after 7 days. 78% of the surviving ani-mals presented with bile duct necrosis with hepatic abscesses at autopsy. Allhad signs of cholangitis with irregular dilatation of the bile ducts in the specimencholangiography.Conclusion: Compared to portal vein embolization, the distinctive arterial collat-eralisation observed limits chances of effective hypertrophy in the contra-lateralsegment. Because of the high rate of necrotic and inflammatory damage to thebiliary system, hepatic artery embolization for induction of hepatic regenerationis not considered suitable for human application.

B-394 14:27

Embolization procedures in acute massive hemorrhages from gastric andduodenal ulcers (24 patients)R. Loffroy, B. Guiu, D. Ben Salem, J. Mazaud, B. Hecquet, J.-P. Cercueil,D. Krause; Dijon/FR ([email protected])

Purpose: To assess the effectiveness and outcome of transcatheter gastroduo-denal artery embolization in the control of massive upper gastrointestinal tractbleeding, in a selected group of poor surgical patients (24). Mean follow-up last-ed 25 months.Methods and Materials: 24 consecutive patients (men 15, women 9, mean age73 years) between September 1999-January 2005 underwent arterial examina-tion and embolization, for acute upper gastrointestinal tract bleeding due to gas-tro-duodenal ulcers, following endoscopic treatment failure. In 16 cases, selectiveembolization of the thin arterial pancreatico-duodenal branch, that was supply-ing the ulcer was feasible with coils and/or histoacryl (glue). For the 8 otherspatients, the "sandwich" technique was systematically performed with coils in thegastroduodenal artery, especially in the absence of contrast medium extravasa-tion. The coaxial technique was mainly used (17/24) with a 3 F microcatheter.Results: In this group (24), no major complications related to the procedure wereobserved. Immediate, durable and permanent haemostasis was achieved in 22patients, without a second catheterisation during the following days; only 2 pa-tients were operated on with success after rebleeding. On the other hand, re-gardless of their clinical conditions at the endovascular intervention, severalvariables did impact on the primary success, including multiorgan system failure(6/24), with a pejorative outcome during the first three months.Conclusion: Angiographic selective embolization with a microcatheter may bean effective way to stop massive life threatening bleeding from gastro-duodenalulcers, especially in the case of debilated non operable patients. For the mid andlong term outcome (3 months- 6 years), the patient's survival is directly depend-ent on their initial clinical condition, at the embolization procedure.

B-395 14:36

Superselective arterial embolization for the treatment of lowergastrointestinal hemorrhageT. Martini, A. Dessl, P. Waldenberger, A. Chemelli, W.R. Jaschke; Innsbruck/AT([email protected])

Purpose: To evaluate the safety and effectiveness of superselective emboliza-tion for the treatment of lower gastrointestinal hemorrhage.Methods and Materials: A retrospective review was conducted on 40 patientswho underwent superselective arterial embolization for acute lower gastrointesti-nal bleeding. In 19 procedures embolization was achieved by using microcoils asthe sole embolic agent, polyvinyl alcohol particles were used in 8 proceduresand a combination of microcoils and particles in another 13. Postembolizationischemia was evaluated objectively in 20 patients by colonoscopy (n = 13), surgi-cal specimen (n = 3), and barium enema (n = 4). All patients were followed forclinical evidence of bowel ischemia.Results: Transcatheter embolization achieved immediate hemostasis in 37 of 40patients (92.5%). Recurrent bleeding occurred in 8 patients (20%), 3 patientsunderwent surgery, 4 patients were managed medically and 1 patient underwentsuccessful repeat embolization. One patient who underwent surgery was foundto have necrosis of the appendix after embolization of the ileocolic artery. Follow-upcolonoscopy demonstrated 2 small ischemic ulcers of the colon in another 2 pa-tients. Both patients remained asymptomatic and developed no sequelae as aresult of ischemic changes on long term follow-up.Conclusion: Superselective arterial embolization is a safe and effective treat-ment for lower gastrointestinal hemorrhage.

B-396 14:45

Uterine artery embolization for symptomatic fibroids: Short-term versusmid-term changes in disease specific symptoms, quality of life and MRimaging resultsC. Scheurig, A. Gauruder-Burmester, C. Kluener, A. Lembcke,E. Zimmermann, T. Islam, B. Hamm, T.J. Kroencke; Berlin/DE([email protected])

Purpose: To evaluate changes in fibroid specific symptom severity and health-related quality of life (HRQL), patient satisfaction and volumetric changes afteruterine artery embolization (UAE).Methods and Materials: 82 women undergoing UAE completed a validated fi-broid-specific symptom and quality of life questionnaire (UFS-QOL) and under-went magnetic resonance imaging (MRI) prior and 3-20 months after UAE. Primarystudy objectives were clinical outcome as measured by the changes in the UFS-QOL after short- (5.5[3-7] months) and mid-term (14[8-20] months) follow-up.Secondary outcome measurements included dominant fibroid (DFV) and uterinevolume (UV) changes and patient satisfaction.Results: 71/82 (87%) patients completed follow-up after UAE. Seven/82 patients(8.5%) had a re-intervention to control persistent symptoms. Symptom severityscores decreased by a median of 57.1% (95%CI: 35.7-71.4; p < 0.001) in theshort-term versus 64.3% (50.0-85.7; p < 0.001) in the mid-term follow-up group.HRQL-scores increased by a median of 30.3% (15.1-59.1; p < .001) and 33.5%(25.8-54.2; p < 0.001), respectively. UV decreased by 26.7% (95%CI: 20.3-33.8;p < 0.001) and DFV by 61.9% (47.8-71.6; p < 0.001) for the short-term collectivewhile a decrease by 36.1% (23.3-55.0; p < 0.001) and 66% (54.3-71.9; p < 0.001)was noted at mid-term follow-up. Decrease in UV was correlated with change insymptom severity score (p < 0.005). 67/71 (94%) answered on satisfaction andrecommendation of the procedure with 63/67 (94%) of women being satisfied.65/67 (97%) would recommend UAE to others.Conclusion: Women who undergo UAE report significant improvements in fi-broid-specific symptoms and health-related quality of life. Comparing short- tomid-term follow-up results, there is evidence of further improvement while overallclinical failure rate appears to be low. Patients´ satisfaction with the procedure ishigh.

B-397 14:54 !Clinicopathologic features and results of transcatheter arterial embolizationfor Graves' diseaseJ. Chu, W. Chen, J. Yang, W. Zhuang, H. Xiao; GuangZhou/CN([email protected])

Purpose: To evaluate the effect of transcatheter arterial embolization (TAE) forGraves' disease and to describe the clinicopathologic features produced by TAE.Methods and Materials: TAE was carried out in 41 patients with Graves' diseasewho were unable to tolerate, or chose not to accept, current therapies (oral med-

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 225C D E FA G

ication, radioactive iodine or surgery). Six of the patients received both arterialembolization and surgery at 2-3 weeks after embolization. Polyvinyl alcohol orbletilla microspheres and micro-coils were used as effective embolic materials.Histological examination of the thyroid tissue after surgery and long-term clinicalfollow-up were carefully evaluated.Results: During successful follow-up (24 - 57 months), 36 of 41 patients (88%)became euthyroid. Four patients required a maintenance dose of antithyroid drugtherapy and one patient had a relapse twelve months after TAE. No hypothyreo-sis or hypoparathyroidism was encountered. At the end of the follow-up period,the size of thyroid gland had decreased by one third to one half of its originalvolume, but 4 patients continued to require antithyroid drugs. Histologic analysisof thyroid tissue from 6 patients who had embolization followed by surgery showedocclusion of both the superior and/or inferior thyroid arteries plus most of theirbranches. Consequently, embolization of the majority of the thyroid glands wasachieved, and evidence of chemical inflammation, necrosis and fibrosis wereobserved in embolized thyroid tissue.Conclusion: TAE is an effective, minimally invasive and safe method for the treat-ment of Graves' disease patients who cannot, or choose not to, accept currenttherapies.

B-398 15:03

Transarterial embolization using absolute ethanol for the management ofrenal angiomyolipomaB.-S. Tan1, J. Reidy2, L.-R. Chong1, K.-H. Tay1, A.M.M. Htoo1, C. Cheng1;1Singapore/SG, 2London/UK ([email protected])

Purpose: To assess the efficacy of transarterial embolization using absolute eth-anol for the management of renal angiomyolipomas.Methods and Materials: From February 1991 to October 2004, 22 patients fromtwo institutions presenting with renal angiomyolipomas were treated by transcath-eter embolization. There were 5 males and 17 females, ranging in age from 12 to62 years (mean 38.9 years). Eight patients presented with bleeding and 4 withloin pain. The remaining 10 patients were asymptomatic at presentation but weretreated either for having tumours that were large (more than 4 cm diameter) orincreasing in size. The mean largest diameter of the tumours was 10 cm (range2.5 to 24.4 cm). Super selective embolization of the feeding arteries was per-formed using a mixture of absolute ethanol and lipiodol with microcatheter tech-niques, to achieve capillary level ablation of the tumours.Results: The procedures were all technically successful with no significant pro-cedural complications. On a mean follow-up of 35.7 months (range 2 to 65 months),no bleeding episodes were encountered. Thirteen of the patients had tumoursthat decreased in size by a mean of 41%. Eight patients had tumours that re-mained stable in size while an increase in tumour size by 17% was documentedin one patient.Conclusion: Transarterial embolization using absolute ethanol is an effectiveand safe alternative to surgery in the management of renal angiomyolipoma.

B-399 15:12

Transcatheter embolization of male varicocele with Fibro-Vein mousseinfusion: Experience with 312 patientsR. Gandini, M. Chiocchi, C. Reale, L. Maresca, M. Stefanini, G. Simonetti;Rome/IT ([email protected])

Purpose: The aim of the study was to evaluate efficacy and safety of Fibro-Veinmousse in the treatment of male varicocele.Methods and Materials: From January 2000 to March 2005, 312 patients (meanage of 28.2 years, range 15-44) were treated by sclerotization of the pampiniformplexus venous dilatation. All patients had infertility, pain or both; the diagnosiswas routinely obtained by physical examination and Doppler ultrasound. Phle-bography was performed under local anesthesia with access through the basilicvein using a 4 Fr Simmons II catheter. In all patients embolization was perfomedwith Fibro-Vein mousse 3% and air (ratio 1:4). Follow-up was routinely performedby Doppler ultrasound and clinical examinations.Results: A technical success was obtained in all but 31 patients (10%). Of the 88patients (28.1%) with infertility 41 (46.6%) regained fertility. Of the 184 (58.8%)patients with pain 158 (86.3%) had resolution of symptoms. Of the 41 (13.1%)patients with pain and infertility, 32 (79.4%) experienced relief of pain and im-provement in sperm count, while 8 (20.6%) patients had relief of pain but re-tained a reduced sperm count.Conclusion: Male varicocele treatment by Fibro-Vein mousse and air is a safeand very effective technique that allows easy pampiniform plexus sclerotization.

B-400 15:21

What happens with the arterial wall after embolization with HepaSphere™?E. de Luis1, J.J. Noguera1, A. García-Rodríguez2, M.D. Lozano1,A. Martínez-Cuesta1, J.I. Bilbao1; 1Pamplona/ES, 2Zaragoza/ES([email protected])

Purpose: HepaSphere™ is a new permanent embolic material that adapts per-fectly to the vessel wall and reaches small diameter arteries (> 60 microns). Thereis little information about the wall damage that these particles induce. The aim ofthis study is to elucidate the mechanism and timing of the vessel wall alterationand its possible influence if HepaSphere™ is used as a drug carrier.Methods and Materials: Kidneys and left hepatic lobes of nineteen pigs werepartially embolized using two different dry sized particles (50 and 150 microns).Animals were sacrificed right after the embolization (4) or four weeks later (15).Histologic studies were performed with hematoxilin-eosin and orcein staining.Results: There was little or no destruction of the arterial wall structures in ani-mals sacrificed immediately after the embolization. Four weeks after the proce-dure, the internal elastic lamina was disrupted in all cases. Hypertrophy of themuscular layer and damage of the adventitia depended on the amount of theinflammatory reaction and the extension of the infarct (mainly in kidney). All thesefindings were only observed with orcein staining. Perfect definition of the portalspace structures and preservation of biliary tracts were also observed in all cas-es.Conclusion: HepaSphere™ is a deformable particle that perfectly attaches tothe arterial wall. Four weeks after embolization damage of the arterial wall layersis found. This can be studied with special staining techniques such as orcein. Theknowledge of arterial disruptions is crucial if the particle is to be used as a drugcarrier.

14:00 - 15:30 Room I

Genitourinary

SS 707ProstateModerators:M. Baramia; Tbilisi/GEB. Hamm; Berlin/DE

B-401 14:00

MR spectroscopic imaging for the early diagnosis of prostate cancer: InitialresultsJ.C. Vilanova1, J. Comet1, J. Barceló1, A. Maroto1, J. Areal2, C. Barceló1;1Girona/ES, 2Barcelona/ES ([email protected])

Purpose: To determine the diagnostic assessment of prostate cancer on MRimaging and MR Spectroscopy (MRSI) in an asymptomatic population.Methods: Prospective study with 51 patients, with PSA > 4 ng/ml < 15 ng/ml.Endorectal MR imaging and 3D-MRSI were performed, followed by biopsies. MRimaging and MRSI findings were analyzed and compared with biopsy and orhistopathological tumor maps as a reference standard, on a sextant distribution.306 sextant MR images and spectroscopic values were analyzed and correlatedindividually. Diagnosis of prostate cancer was also compared with PSA, free PSAratio, MR imaging and MRSI results for each patient.Results: Of 306 sextants, 78 contained cancer (25%) corresponding to 23 pa-tients (45%) with cancer. Sensitivity, specificity, and accuracy were respectively42%, 94% and 81% for MR imaging and 80%, 98% and 93% for MRSI. Ratio CC/Ci was significantly higher (mean 1.05 ± 0.41) for sextant neoplasms than fornon cancerous ones (mean 0.51 ± 0.23). Mean free PSA ratio was significantlylower (11.35%) for patients with cancer than for the ones without (16.55%). Totalmean PSA was not significantly different in patients with (7.72 ng/ml) and withoutcancer (6.42 ng/ml). When analysis was performed based on the prostate as awhole unit, it showed a sensitivity, specificity and accuracy of 70%, 75% and72% for MR imaging and 87%, 93% and 90% for MRSI respectively.Conclusion: MR spectroscopic imaging of the prostate is a non-invasive methodwith a significantly high degree of accurate prediction of prostate cancer in anasymptomatic population, as the metabolic ratio of CC/Ci shows a better discrim-ination value for prostate cancer than the PSA levels.

Scientific Sessions

226 C D E FBA G

B-402 14:09

The feasibility and initial results of 3 T body array coil proton MRspectroscopy to determine metabolite ratios for discrimination betweencancer and healthy tissueS.W.T.P. Heijmink1, T.W.J. Scheenen1, S.A. Roell2, A. Heerschap1,J.O. Barentsz1; 1Nijmegen/NL, 2Erlangen/DE ([email protected])

Purpose: To evaluate the feasibility of 3 T proton MR spectroscopic imaging (1H-MRSI) with solely a body-array coil and determine optimal metabolite ratios dis-criminating between healthy and prostate cancer tissue.Methods and Materials: Twenty-two patients underwent 1H-MRSI at 3 T (Sie-mens TRIO) before radical prostatectomy. After acquiring conventional T2-weight-ed MR images, a PRESS sequence with optimized 180° pulses and 2 dual-bandfrequency selective excitation pulses to suppress water and lipids was applied.Up to 8 outer volume saturation slabs were used. Voxel size was 7x7x7 mm3. Inconsensus, a radiologist and MR spectroscopist reviewed whole-mount sectionhistopathology on which tumours were outlined. Blinded to metabolite spectra,healthy peripheral zone (PZ), periurethral (U), central gland (CG) tissue and tu-mor voxels were selected by matching the T2-weighted images with histopathol-ogy. Using PRISMA software (University of Bremen and Siemens MedicalSolutions, Erlangen, Germany) (Choline+Creatine)/Citrate ratios were calculat-ed and areas under the receiver operating curve (AUC) were determined.Results: Adequate spectral quality was obtained in all patients. Mean and medi-an (Choline+Creatine)/Citrate ratios for tumour were 2.1 and 0.42, respectively.The AUCs for discriminating cancer from healthy tissue in the PZ and combinedU+CG were 0.89 and 0.75, respectively. The optimal ratio cut-off point for detect-ing PZ cancer was 0.27 (sensitivity: 87%, specificity: 81%). For CG cancer thiswas 0.33 (73% and 69%, respectively).Conclusion: 1H-MRSI with body-array coil only is feasible and can discriminatebetween cancer and healthy t issue. Optimal cut-off points of the(Choline+Creatine)/Citrate ratios for detecting PZ and U+CG tumor were 0.27and 0.33, respectively.

B-403 14:18

Metabolic and T2-relaxation time mapping of prostate: 1H MRS studyV. Rogozhyn, Z. Rozhkova, S. Vozyanov, V. Korobko; Kiev/UA([email protected])

Purpose: We propose a quantitative description of metabolic and T2- alterationof relaxation time for prostate tissue differential diagnosis prostate carcinoma(PC) and benign hyperplasia (BHP).Methods and Materials: Two groups of patients were examined by 1.5 T Magne-tom Vision System (SIEMENS). The 1st group (PG) included 53 patients (34-76y), the 2nd group (VG) consisted of 15 healthy volunteers (30-48y). To build themetabolic map the 1H spectra are recorded with 2DCSI:TR/TE = 1500/60 ms. InT2 measurements the 1H spectra are recorded in the central zone (CZ) and in theperipheral zone (PZ) of prostate with SVSSTEAM: TR/TE = 1500/60, 80, 100,140, 180 ms.Results: From 1H MR spectra the peak areas of main metabolite signals: citrate(Cit, 2.6 ppm), creatine (Cr, 3.0 ppm) and choline (Cho, 3.2 ppm) and the meanpeak height ratios of R = Cit/(Cr+Cho) were obtained. R values R > 1.12 areindicative for PC, R > 2.30 for BHP, R > 4.14 for normal regions of prostate tis-sue. In the VG the mean values of T2 for water (Wat), and for Cit, Cr, and Cho arethe following: T2

CZ (Wat)=(71.3+-0.56)ms, T2CZ (Cit)=(170.2+-0.23)ms, T2

CZ (Cr)=(134.5+-0.3)ms, and T2

CZ (Cho)=(254.13+-0.81)ms; T2PZ (Wat)=(123.5+-0.74)ms,

T2PZ (Cit) =(216.8+-0.79)ms, T2

PZ (Cr)=(241.2+-0.15)ms, T2PZ (Cho)=(323.7+-0.2)ms.

In the PG the values of T2 (from 56.4 to 93.2 ms) are indicative for water signal in

PC, and (from 98.7 to 117.4 ms) for BHP.Conclusion: Metabolic and T2-relaxation time mapping can be helpful for diag-nosis and characterization of clinically localized PC and also to estimate the effi-ciency of therapy.

B-404 14:27

Prostate dynamic contrast-enhanced MR imaging with low temporalresolution: Is it reasonable?O. Rouvière, N. Girouin, F. Mège-Lechevallier, A. Tonina Senes,J.-M. Maréchal, M. Colombel, D. Lyonnet; Lyon/FR([email protected])

Purpose: To evaluate a dynamic contrast enhanced (DCE) sequence with lowtemporal resolution (and thus good spatial resolution and slice coverage) for thedetection and localization of prostate cancer foci.Methods and Materials: 46 patients underwent prostate MR imaging before rad-

ical prostatectomy. In addition to axial, sagittal and coronal T2-weighted sequenc-es, an axial DCE sequence (14 slices, 30 seconds) was repeated 3 times afterinjection of gadolinium (20 cc at 2 cc/s). For image analysis, the prostate wasdivided into 20 segments. DCE images were assessed only visually, all earlyenhancing lesions being considered malignant in the peripheral zone. In the cen-tral gland, only early enhancing lesions with a homogeneous appearance and orinvading the peripheral zone were considered malignant. Three independent read-ers, blinded to clinical data and patients names, evaluated T2 and DCE imagesseparately. They had to specify cancer involvement within each segment and torecord tumor nodule number and location. Results were compared with whole-mount step-sections of the surgical specimens.Results: Of the 920 segments examined histologically, 227 were invaded (25%).Sensitivity and specificity for detection of segments involvement were respec-tively, depending on the readers, 0.13-0.23 and 0.93-0.98 for T2 images and 0.46-0.56 and 0.94-0.99 for DCE images. Forty-nine malignant nodules (> 0.3 cc) werepresent at histology: 49-60% were detected and correctly located on T2 sequencewith 19-50 false positive; 77-81% were detected and correctly located on DCE-MR imaging with 19-45 false positive.Conclusion: DCE-MR imaging allows better detection and localization of pros-tate cancer foci than T2-weighted MR imaging, even if a low-temporal resolutionis used.

B-405 14:36

Detection of prostate cancer using a combined approach of contrastenhanced color Doppler, targeted and systematic biopsy with a prostatespecific antigen of 4 to 10 ng/mlJ. Gradl, L. Pallwein, D. zur Nedden, F. Frauscher; Innsbruck/AT([email protected])

Purpose: Transrectal gray scale ultrasound guided biopsy is the standard meth-od for diagnosing prostate cancer (PC). Improved cancer detection with ultra-sound contrast agents is related to a better detection of tumor vascularity. Weevaluated the impact of a combined approach of contrast enhanced, color Dop-pler targeted biopsy (CECD) and systematic biopsy (SB) for the PC detectionrate in men with prostate specific antigen (PSA) 4.0 to 10 ng/ml.Methods and Materials: We examined 380 screening volunteers with a totalPSA of 4.0 to 10 ng/ml (percent free PSA less than 18). CECD was always per-formed before SB. Another investigator blinded to the contrast enhanced findingsperformed 10 SBs. The cancer detection rate for the CECD, SB and combinedapproaches was assessed.Results: PC was detected in 143 of 380 patients (37.6%, mean total PSA 6.2 ng/ml). The PC detection rate for CECD and for SB was 27.4% and 27.6%, respec-tively. The overall cancer detection rate with the 2 methods combined was 37.6%.For targeted biopsy cores the detection rate was significantly better than for SBcores (32.6% vs 17.9%, p < 0.01). CECD in a patient with cancer was 3.1 timesmore likely to detect PC than SB.Conclusion: CECD allows for the detection of lesions that cannot be found ongray scale ultrasound or SB. CECD allows for the assessment of neovascularityassociated with PC. However, the combined use of CECD and SB allows formaximal detection of PC with a detection rate of 37.6% in our patients with a PSAof 4 to 10 ng/ml.

B-406 14:45

Is there a relation between pain and prostate volume during transrectalultrasound guided prostate biopsy? A retrospective linear regressionmodelT. Yun, H. Lee, S. Kim; Seoul/KR ([email protected])

Purpose: To determine the relationship between pain and prostate volume dur-ing transrectal ultrasound guided prostate biopsy.Methods and Materials: Between March 2004 and September 2004, the totalprostate volume, peripheral zone volume, and transition zone volume of 71 pa-tients, who had prostate biopsies, were retrospectively analysed. Pain was as-sessed with a visual analogue scale, during periprostatic nerve block (VAS 1),during the biopsy procedure (VAS 2) and 20 minutes afterwards (VAS 3). Associ-ations between prostate volume and pain during transrectal prostate biopsy wereretrospectively evaluated using linear regression models.Results: Total prostate volume was positively associated with VAS 2 (p = 0.0220)and VAS 3 (p = 0.0012). Peripheral zone volume was also positively associatedwith VAS 2 (p = 0.0231) and VAS 3 (p = 0.0001). However no significant associ-ation could be demonstrated between total prostate volume or peripheral zonevolume and VAS 1. There was not any significant relationship between transitionzone volume and VAS values.

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 227C D E FA G

Conclusion: Higher volume of total prostate or peripheral zone is related to ahigher level of pain during and after transrectal prostate biopsy. We suggest thatan additional analgesic strategy is necessary for a larger prostate or peripheralzone biopsies.

B-407 14:54

Prostate cancer detection in the transitional-zone by biopsy: Is it stillnecessary?J. Gradl, L. Pallwein, D. zur Nedden, F. Frauscher; Innsbruck/AT([email protected])

Purpose: This retrospective study was designed to investigate the value of tran-sition zone (TZ) biopsies for prostate cancer (PC) detection rate in a combinedcontrast enhanced color Doppler targeted (CECD) and gray-scale systematicbiopsy (SB) approach.Methods and Materials: PSA screening participants totalling 1475 with tPSA of> 1.25 ng/ml (fPSA </=18%) were assessed. Ten SB and additionally 5 or fewerCECD were performed. The impact of TZ biopsies on the PC detection rate andthe biological significance of the detected TZ cancers were analyzed.Results: Out of 1475 biopsied patients, 395 (26.8%) were identified as PC pa-tients; 5925 biopsy cores from these patients were analyzed. In 86 patients (21.8%of PC), we found 102 PC positive cores in the TZ, and only in 9 of them solitaryTZ cancers without any other PC location (2.3% of PC or 0.6% of all investigatedpatients). Pathological findings after retropubic prostatectomy (RPE) revealedmultifocal adenocarcinoma including involved peripheral zone (PZ) in eight ofthese nine patients, and solitary TZ cancer in one patient. There was no positivecorrelation between prostate volume and TZ detection rate and no patient withsolitary TZ PC after rebiopsy.Conclusion: Biopsy revealed 9 solitary TZ cancers (1.8%) and RPE revealedonly one of them to be truly TZ confined cancer (0.6%). Furthermore PC detec-tion did not improve, even in patients with rebiopsy and there was no correlationbetween detection of TZ cancers and prostate volume. A combined use of CECDand SB to investigate participants of a PSA screening program suggests that TZbiopsies do not improve PC detection rate and are therefore unnecessary.

B-408 15:03

Localization of prostate cancer using unenhanced and contrast-enhancedtransrectal ultrasound with whole-mount section histopathology as areference standardS.W.T.P. Heijmink, H. van Moerkerk, J.J. Fütterer, C.A. Hulsbergen-v.d. Kaa,J.A. Witjes, J.O. Barentsz; Nijmegen/NL ([email protected])

Purpose: To evaluate the additional effect of contrast administration on prostatecancer localization performance by means of transrectal ultrasound (TRUS) us-ing whole-mount section histopathology as a reference standard.Methods and Materials: Forty-five consecutive patients with biopsy proven pros-tate cancer underwent TRUS, on average, five days prior to radical prostatecto-my. Mean age: 59 years, mean PSA: 6.70 ng/ml, median Gleason score: 6. Allexaminations were performed on the Viking® ultrasound machine using the side-fire probe type 8808. Axial images from apex to base were obtained in gray-scale, colour Doppler and power Doppler mode. A second set of power Dopplerimages was obtained during continuous manual infusion of 2.4 ml of sulphur hex-afluoride (SonoVue®) at a rate of 1 ml/min with a mechanical index < 0.3. Tworadiologists independently scored the presence of prostate cancer in a 14-seg-ment model of the prostate on a five-point probability scale. Whole-mount sectionhistopathology was the reference standard. Diagnostic performance parameterswere calculated for the whole prostate as well as the dorsal and ventral partseparately.Results: Contrast enhanced power Doppler had the highest sensitivity (37%)while unenhanced power Doppler had the highest specificity (90%). All modeshad lower sensitivity in the ventral half of the prostate compared to the dorsalhalf. Contrast administration significantly improved the Doppler localization ofventrally located tumors.Conclusion: Administration of contrast agent improved the localization sensitiv-ity for prostate cancer. This increase was statistically significant for ventral tu-mors.

B-409 15:12 !Prostate gland signal changes on T2 weighted MR imaging after dose-escalated intensity modulated radiotherapy (IMRT): Follow-up of 56 patientsC.M. Zechmann, K. Aftab, B. Didinger, K. Baudendistel, S. Delorme,H.-U. Kauczor; Heidelberg/DE ([email protected])

Purpose: The prostate gland becomes hypointense on T2W images after radio-therapy, the change being most apparent in the peripheral zone (PZ) which isnormally hyperintense. These signal changes are not well documented in theliterature. The purpose of our study was to evaluate signal change of the PZ andcentral gland (CG) of the prostate before and after radiotherapy and determinewhether PZ prostate cancer could still be distinguished after radiotherapy.Methods and Materials: Prostate gland PZ and CG signal between the first andlast available MR imaging follow-up was assessed in 56 prostate cancer patientswho received dose-escalated stereotactic intensity modulated radiotherapy (IMRT).Mean interval between the first and last MR imaging was 10.2 months. T2W sig-nal changes of the PZ and CG before and after radiotherapy were noted. T2W-hypointense areas were regarded as suspicious for cancer. Prostate cancers onpostradiation scans were assumed to be in the location where pre-treatment scanshad clearly shown them.Results: Mean duration of radiotherapy was 2.5 months. PZ and CG distinctionon the first MR imaging was present in 48/56 prostates (85.7%). Distinction of thePZ and CG was lost in 19 of these 48 postradiotherapy prostates (39.6%). PZcancers were visible in 35/56 patients (62.5%) on the first MR imaging. Follow-upMR imaging did not pick up 25 of these 35 cancers (71.4%).Conclusion: For patients who undergo dose-escalated stereotactic IMRT, thereis a definite prostate gland signal change. T2W signal changes render differenti-ation of the PZ and CG difficult. With the PZ turning hypointense after radiother-apy, it is hardly possible to distinguish hypointense cancer in the PZ on T2Wimages.

B-410 15:21

Ferumoxtran-10 enhanced MR imaging in the detection of lymph nodemetastases outside the normal surgical area in prostate cancerR.A.M. Heesakkers1, B. De Hoop1, H. van denBosch2, F. Raat3, F. Witjes1,J. Barentsz1; 1Nijmegen/NL, 2Eindhoven/NL, 3Terneuzen/NL([email protected])

Purpose: To evaluate the feasibility of finding lymph node metastases outside ofthe normal (= limited) surgical dissection area, using Ferumoxtran-10 enhancedMR imaging, in patients with prostate cancer. This is in order to select whichpatients are best suited for an 'extended' lymphadenectomy or CT-guided biopsy.Methods and Material: 300 consecutive patients with biopsy proven prostatecancer (PSA > 10, or Gleason > 6 or DRE stage T3) were enrolled in this study.T1-weighted TSE and T2*-weighted GRE MR images of the pelvis were obtained,24 hours after Ferumoxtran-10 administration. Positive nodes on MR examina-tions were called inside, or outside the normal surgical area. Based on MR imag-ing results, a lymph node dissection, a MRL guided minimal extended lymphnode dissection, or CT-guided lymph node biopsy was performed.Results: Forty-nine (16%) patients had malignant nodes on histological exami-nation. In 21 patients (43 %), malignant nodes were found exclusively outside thelimited surgical area. In 3 (6%) patients malignant nodes were found both insidethe limited dissection field and outside this area. In the remaining 25 (51%) pa-tients malignant nodes were found in the area inside the limited dissection field.Conclusion: in prostate cancer patients with intermediate to high risk for lymphnode metastases Ferumoxtran-10 enhanced MR imaging. Additional nodal me-tastases are found outside the field of limited nodal dissection in 43%. Ferumox-tran-10 enhanced MR imaging may help to select which patients are best suitedfor an "extended" lymphadenectomy or CT-guided biopsy.

Scientific Sessions

228 C D E FBA G

14:00 - 15:30 Room K

Pediatric

SS 712Neuroradiology and tumor imagingModerators:L. MacDonald; London/UKM. Raissaki; Iraklion/GR

B-411 14:00

Myelination process abnormalities in preterm subjects with periventricularleukomalacia assessed by magnetization transfer ratioV. Xydis, L.G. Astrakas, A.K. Zikou, D. Gasias, K. Pantou, S. Andronikou,M.I. Argyropoulou; Ioannina/GR

Purpose: To study in preterm subjects with periventricular leukomalacia (PVL)the progress of myelination by measuring the magnetization ratio (MTR) and tocompare with normal values.Methods and Material: Brain MTR of 26 PVL subjects (16 males, 10 females, ofgestational age 30.6 ± 2.5 weeks, corrected age 37.6 ± 36.1 months) was meas-ured using a three-dimensional gradient-echo sequence (TR=32/TE=8/flip an-gle = 60 4 mm/2 mm overlapping sections) without and with magnetization transferprepulse and compared with normal values for preterm subjects. The magnetiza-tion transfer ratio was calculated as: MTR = (SIo-SIm)/SIo X 100%, where SImrefers to signal intensity from an image acquired with a MT prepulse and SIo thesignal intensity from the image acquired without a MT prepulse.Results: MTR followed a monoexponetial function model (y=A-B*exp (-x/C)). MTRof frontal and occipital lobes, of the genu of the corpus callosum, of the thalamusand caudate nucleus reached a normal final value (parameter A) within normalage. The splenium of the corpus callosun reached a lower than normal final value(mean; 66.9%, stderr; 0.7, mean; 70%, stderr; 0.6, respectively), p < 0.05, fur-thermore it attained at a younger age 95% of the final value of total MTR change(parameter B) (6.5months versus 17.7months) and 99% of the final MTR value(parameter A) (10.3months versus 18.9months)Conclusion: In preterm subjects with PVL, myelination process of the spleniumis characterised by an early arrest resulting in a deficient degree of maturation.

B-412 14:09

Identification of different embryologic cortex types in children. Validity ofthe FLAIR sequenceK. Vergesslich, J.F. Schneider; Basle/CH ([email protected])

Purpose: MR imaging and especially the Fluid Attenuated Inversion Recovery(FLAIR) sequence is known to show higher signal intensity of cortices belongingto limbic structures. Our purpose was to evaluate the possibility of further differ-entiation between embryologic areas of the limbic and neocortical system.Methods and Materials: We investigated 30 children without neurological dis-ease and measured signal intensity in 14 different cortical areas with a region ofinterest (ROI) function, grouped into 7 categories: (1) cingular cortex, (2) insularcortex, (3) olfactory cortex, (4) hippocampal cortex, (5) parahippocampal cortex,(6) frontal neocortex, (7) parietal neocortex. Paired sampled t - test was per-formed.Results: Significant signal intensity differences were found with high intensity ingroups 4 and 5, intermediate intensity in groups 1 and 3 and low intensity ingroups 6 and 7.Conclusion: FLAIR sequence allows differentiation of cytoarchitectonically andchronoarchitectonically distinct cortical areas.

B-413 14:18

Quantitative T2* measurements of fetal brain oxygenation during hypoxiawith MR imaging at 3 T: Correlation with blood gas analysesU. Wedegärtner, H. Kooijman, A.N. Priest, H. Schröder, G. Adam; Hamburg/DE([email protected])

Purpose: Determination of oxygen saturation in the fetal brain during hypoxia byT2* measurements and comparison with fetal blood gas analysis.Methods and Material: 11 sheep fetuses were examined during a control andhypoxic period on a 3 T MR imaging scanner (Philips, Intera). Multiecho gradient-echo (39 echoes, spacing 3.2 ms) and turbo-spin-echo-sequences (30 echoes,spacing 5 ms) were performed on the fetal brain. Quantitative T2* and T2 mapswere calculated. MR oxygen saturation (MRSO2calc) was calculated as follows:

MRSO2calc = 100% * [1 - 1/T2'ox*delta*4/3*pi*Deltachi0*Hct*gamma*B0], where 1/T2'ox =1/T2-1/T2*-1/T2'FE (delta=0.045 blood volume fraction; T2'FE =250 ms con-tributions from brain iron; Deltachi0 =0.27 susceptibility difference fully oxy/deoxy-genated blood; gamma=gyromagnetic ratio; B0=field strength).Fetal blood samples were taken from a carotid catheter at control and hypoxicperiods to determine blood oxygen saturation. Linear regression analysis com-pared the calculated MRSO2calc and blood samples oxygenation.Results: During the control period (n = 17 measurements) mean values were71% [95% confidence interval: 62-79] for MRSO2calc and 60% SO2 [50-69] forblood oxygenation. During hypoxia MRSO2calc decreased to 37% [23-50] and bloodoxygenation to 21% [15-27]. MRSO2calc and blood oxygenation correlated signifi-cantly (r = 0.67; p = 0.00003). For MR-calculated oxygen saturations below 40%SO2, the value determined from blood samples was always less than 50% SO2.Conclusion: The feasibility of quantitative T2* measurement has been demon-strated in the fetal brain. Although there are some differences between SO

2 val-

ues determined by MR imaging and blood samples, there is reasonable correlationbetween these two methods. Therefore MR imaging has potential to non-inva-sively assess oxygen saturation and consequent fetal risk.

B-414 14:27

The role of fetal MR imaging in the detection of gyration and migrationanomalies before 24 weeks gestationD. Horák1, J. Hyjánek1, P. Vlašín2, M. Herman1; 1Olomouc/CZ, 2Brno/CZ([email protected])

Purpose: Most prenatal MR imaging in the Czech Republic is done before 24weeks gestation (WG) for medicolegal reasons (possibility to terminate pregnan-cy up to 24 WG for medical causes).The fetal brain during this period is undergoing early developmental changes,especially in terms of gyration and sulcation.The authors want to assess the benefit of fetal MR imaging for detection of migra-tion and gyration anomalies of fetuses up to 24 WG.Methods and Material: 48 CNS examinations of fetuses up to 24 WG were per-formed on a 1.5 T scanner.Heavily T2W images of the brain were acquired in three orthogonal planes and atleast in two planes for the spine. T1W images were routinely used for the brainand spine in one plane.We acquired all images without fetal or maternal sedation.MR imaging results were compared to autopsy or postnatal imaging, clinicalfollow-up and the US study.Results: We found 8 (16.7%) cases of migration and gyration anomalies: 3 com-plete callosal agenesis, 1 callosal dysgenesis, 1 schizencephaly, 2 complex mi-gration anomaly and 1 parenchymal atrophy. US found mostly indirect changessuch ventriculomegaly, dilation of subarachnoid space or thin cortex.We found 25 (52.1%) other anomalies (ventriculomegaly, posterior fossa anoma-lies, spinal dysrhaphisms etc), 15 (31.3%) examinations were normal.Autopsy/follow-up confirmed the MR diagnosis in 87.5%.Conclusion: Despite the presence of only early developmental changes in theexamined brain, fetal MR imaging before 24 WG is capable of detecting CNSmigration and gyration anomalies and provides valuable diagnostic informationcompared to US.

B-415 14:36

Ultrasound and MR imaging of paediatric neuropathies (brachial plexusexcluded)M. Valle1, P. Toma'1, G. Bodner2, S. Bianchi3, E. Silvestri1, C. Martinoli1;1Genova/IT, 2Innsbruck/AT, 3Geneva/CH ([email protected])

Purpose: To describe and explain the anatomical basis, the clinical characteris-tics as well as the US appearances for a variety of neuropathies (brachial plexusexcluded) which occur in children.Methods and Materials: Thirty consecutive patients with either dysfunction orabnormal nerve conduction studies or a palpable mass over the course of a nerveof the upper or lower limb were retrospectively evaluated. US was performed with12-5 MHz and 17-5 MHz US transducers. All patients had correlative 1.5 T MRimaging.Results: US allowed detection of a spectrum of nerve abnormalities. Traumaticnerve lesions due to penetrating injuries included: partial tear of the ulnar nerveat wrist (n = 2) and elbow (n = 1); combined tear of ulnar, median and radial nervesat the forearm (n = 1); complete tear of the median nerve at the arm (n = 1). Closedstretching injuries selectively involved the peroneal nerve in the popliteal fossawith formation of a fusiform neuroma (n = 4). Other lesions included: entrapmentof the peroneal nerve at the fibular neck by bone exostoses (n = 3) or superior

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 229C D E FA G

tibiofibular joint ganglia (n = 2); carpal tunnel syndrome in hemophilia (n = 1);radial nerve entrapment at the spiral groove by fibrous band (n = 1) and postop-erative scar encasement of the sural nerve (n = 1). Nerve tumors included: fibrol-ipomatous hamartomas of the median (n = 3) and plantar (n = 1) nerves, plexiformschwannoma (n = 1), diffuse neurofibroma (n = 2) and type-1 neurofibromatosis(n = 6).Conclusion: US is a promising means for evaluating a wide range of nerve ab-normalities in children.

B-416 14:45 !Intraperitoneal administration of Ferumoxtran: A new MRI tracer forreticulo-endothelial system (RES) cells in new born ratsR. Azoulay1, P. Olivier1, F. Guérin1, O. Baud1, M. Schaefer2, C. Verney1,P. Gressens1, G. Sebag1; 1Paris/FR, 2Roissy/FR ([email protected])

Purpose: Ferumoxtran is administrated for in-vivo MRI monitoring of macrophagiccells on experimental adult animals. Several new born experimental models con-tain macrophagic activated cells (periventricular leukomalacia for example). In-travenous injection is impossible in these animals (< 10 grams). Thus, we test anew type of administration for Ferumoxtran on healthy new born rats: the intra-peritoneal administration (IP).Methods and Materials: Ferumoxtran IP (2400 micromole/kg) was evaluatedafter a delay of 0 to 105 hours in 61 rats (age 1-7 days) with:

1/ in-vivo 1.5-T abdominal MRI (TR=2000, TE=50, liver, spleen, and bone mar-row signal)

2/ plasma and peritoneal iron concentrations3/ liver, spleen and bone marrow histology (Perls coloration for iron intracellular

deposit).5 rats were control rats (NaCl injection).Results: Ferumoxtran IP was resorbed from the peritoneum before 11 hours,and detected in plasma between 0 to 75 hours (concentration peak at 8 h, 660 mg/l). A significant (P < 0.05) decrease of liver, spleen and bone marrow signal wasobserved from 2 hours (liver, spleen) and 4 hours (bone marrow) after IP andbeyond that.Intracellular iron deposits were significantly observed (P < 0.05) in liver and spleenfrom 3 hours and beyond that and at 96 hours in bone marrow.Conclusion: Ferumoxtran IP allows systemic diffusion and in vivo MRI targetingof RES cells in healthy new born rats. An application on a periventricular leu-komalacia experimental model will be shown.

B-417 14:54

Multimodality imaging in paediatric tumour patientsP. Kovacs, M. Knoflach, C. Hinterleithner, T. Lang, R. Crazzolara, B. Meister,W.R. Jaschke, R.J. Bale; Innsbruck/AT ([email protected])

Purpose: To determine the feasibility of multimodality imaging in paediatric pa-tients suffering from neuroblastoma, hepatoblastoma and Ewing sarcoma.Methods and Materials: In the last three years 16 paediatric patients underwentmultimodality imaging in our institution. Among them 4 patients with neuroblasto-ma (aged 9 months to 2 years) were imaged 8 times, 2 patients with hepatoblas-toma (1 and 4 years) 3 times, and 2 patients with Ewing sarcoma (11 and 14years) 3 times for staging and restaging. Overall 7 superimpositions of CT andSPECT, 4 of CT and PET, 2 of MR and SPECT and 2 of MR and PET wereperformed on a TREON StealthStation® Navigation System (Medtronic, USA).Up to 4 years of age, all datasets were obtained one by one in one session duringgeneral anaesthesia. Movement was additionally inhibited using a Bluebag® vac-uum mattress (Medical Intelligence, Germany).Results: Superimposition of datasets was successful in all examinations (n = 15).Follow-up and restaging after therapy was possible in measuring tumour extentand functional activity of tumour tissue.Conclusion: Multimodality imaging is possible even in young patients using gen-eral anaesthesia and immobilization devices. Tumour extent and function can bedetermined in one dataset leading to appropriate adjustment following surgical,chemo or radiation therapy.

B-418 15:03

Post transplant lymphoproliferative disorder (PTLD) in pediatric patients:A possible role for [18F]-FDG-PET/CT during initial staging and therapymonitoringC. von Falck, B. Maecker, A. Boerner, E. Schirg, C. Klein, M. Galanski;Hannover/DE ([email protected])

Purpose: To evaluate the potential role of [18F]-FDG PET (/CT) imaging in theinitial staging and therapy monitoring of pediatric patients suffering from posttransplant lymphoproliferative disorder (PTLD) after solid organ transplantation.Methods and Material: PET (/CT) (n = 12), CT (n = 14) and MRI (n = 11) exam-inations were performed in six pediatric patients with histopathologically provenPTLD in the initial staging and during the course of therapy with an anti-CD20chimeric antibody (Rituximab). One (n = 3), two (n = 2) or three (n = 1) follow-upstudies were performed in all patients. Images were retrospectively correlated ona lesion by lesion basis.Results: In five patients PTLD occurred 12 (± 7) months after transplantation ofthe heart and lungs (n = 1) or the kidney (n = 4). In one patient PTLD was diag-nosed 84 months after transplantation of the heart (n = 1). PTLD showed organinvolvement of the lung (n = 2), the liver (n = 1), the kidneys (n = 2), the digestivetract (n = 2), the lymphatic nodes (n = 6) and the skeletal system (n = 1). All sitesof organ involvement as demonstrated by CT or MRI showed markedly increasedtracer uptake in FDG-PET in the initial staging. During therapy, FDG-PET wasable to demonstrate response early whereas in CT and MRI the lesions showedonly delayed changes.Conclusion: [18F]-FDG-PET (/CT) aids with the accurate initial staging of pediat-ric patients suffering from PTLD. It allows response to therapy to be demonstrat-ed earlier than CT or MR imaging.

B-419 15:12

Reduction of radiotherapy in children with early stages of Hodgkin'slymphoma, influenced by a new imaging and FDG-PET based strategyI. Sorge, C. Mauz-Koerholz, D. Koerholz, R. Kluge, A. Krausse, T. Kahn,W. Hirsch; Leipzig/DE ([email protected])

Purpose: Although Hodgkin's disease (HD) in children has an event free survival(EFS) of about 95%, the rate of secondary malignancies is about 25% after 30years.Radiation therapy seems to be the most important risk factor.Therefore the GPOH-HD 2002 study group attempts to reduce the indication forradiation therapy.Methods: Until now only the patients with complete remission on CT/MRI afterchemotherapy get no radiation therapy.This strategy had no influence on the EFS. The next study generation is intendedto further reduce the indication for radiation therapy using FDG-PET for responseevaluation.In our study in 41 patients with early stage HD CT/MRI was paralleled by FDG-PET in initial staging and in the follow-up.Results: 7 of the 41 patients were in CR based on PET and CT/MRI criteria, onepatient with CT/MRI-based CR had still a positive PET result. In 17 patients PETwas completely negative after chemotherapy whereas residual masses were foundby CT/MRI. In 2 cases PET was not able to be evaluated in its entirety.In summary, using a PET-based strategy, 27/41 patients (66%) with early stageHD would not have to be irradiated in the future in contrast to 9/41 patients (21%)according to the current strategy.Conclusion: By using FDG-PET-based stratification in addition to the establishedCT/MRI, the number of non-irradiated patients will increase dramatically. We es-timate that about 60 - 70% of all patients with early stage HD will no longer re-ceive radiation therapy without reduction of EFS rates.

B-420 15:21

Evaluation of airway obstructive lesions in complex congenital heartdisease: Using composite volume rendered images from MSCTK.S. Choo, H. Cho, H. Lee, S. Sung, C. Kim; Pusan/KR([email protected])

Purpose: To investigate clinical usefulness of composite volume rendered imag-es in the evaluation of the relationship between cardiovascular structure and air-ways in children with complex congenital heart disease.Methods and Materials: 4-slice or 16-slice MSCT scanning was performed con-secutively in 77 children (age range 5 days to 5 years, mean age: 6.4 months)with congenital heart disease and respiratory symptoms, chest radiograph ab-normality or abnormal course of the pulmonary artery on echocardiography. MSCT

Scientific Sessions

230 C D E FBA G

scanning was performed during breathing and after sedation with chloral hydratein all patients. Contrast medium was administered through a pedal or arm venousroute in all patients. The CT volume data was transferred to commercially availa-ble software (voxelplus2, Seoul, Korea). The VR technique was use to recon-struct the cardiovascular structures and airways, and then both volume-renderedimages were composed using the software.Results: Stenoses were seen in the trachea in 1 patient and in bronchuses in 14patients (19%). Their causes were abnormal course of the pulmonary artery in 2patients, circumflex retro-esophageal aortic arch in 2, pulmonary artery dilationin 1, anterior displacement of the descending aorta or posterior displacement ofthe ascending aorta following operation in 5, posterior displacement of the pul-monary artery in 4, and sub-pulmonary sling in 1 patient. Other patients withcomplex congenital heart disease did not have significant airway stenoses.Conclusion: Composition of both volume-rendered images with MSCT can pro-vide more exact airway images in relationship to surrounding cardiovascular struc-tures and can thus offer optimal management strategies in treating congenitalheart disease.

14:00 - 15:30 Room N/O

Physics in Radiology

SS 713Computed tomographyModerators:C. Leidecker; Erlangen/DEA. Noel; Vandoeuvre-les-Nancy/FR

B-421 14:00

Evaluation of kymogram-based phase-correlated in comparison to ECG-based cardiac CT image reconstructionD. Ertel, M. Kachelriess, S. Achenbach, W.A. Kalender; Erlangen/DE([email protected])

Purpose: To compare the performance of a rawdata-based synchronization sig-nal to the commonly used ECG for a 64-slice CT scanner.Methods and Materials: The kymogram represents the temporal variation ofmass and provides a motion function which is correlated to the heart rate. It isgenerated by a center-of-mass tracking of the scanned object [Med. Phys., 29(7):1489-1503, 2002]. We evaluated this purely rawdata-based synchronizationsignal as an alternative to the ECG-signal for phase-correlated image recon-struction using a 64-slice CT scanner (Sensation 64, Siemens, Forchheim). Forevaluation data of 90 consecutive patients were used scanned with 2•32*0.6 mmcollimation, 0.33 s rotation time and concurrent ECG recording. As a quality meas-ure the correlation of the ECG with the kymogram was used for validation. Addi-tionally, visual assessment of the kymogram-based and the ECG-basedreconstructed images was performed by a double-blind comparative rating.Results: In 70 out of 90 patients a high correlation between the kymogram andthe ECG was observed with a mean difference of the detected heart rate below4 bpm. In 6 out of 90 patients a strong improvement in image quality was achievedby using a kymogram-based reconstruction. This was due to bad ECG record-ings.Conclusion: Synchronization signals on the basis of a kymogram correlated wellwith the ECG for most patients and improved image quality in some cases. Weconclude that a kymogram-based phase-correlated image reconstruction in car-diac CT is practical and may provide advantages.

B-422 14:09

Image quality improvements in cardiac imaging with step-and-shootvolumetric CTJ. Hsieh, J. Londt, M. Vass, J. Li, X. Tang, D. Okerlund; Waukesha, WI/US([email protected])

Purpose: Although nearly all cardiac CT protocols utilize low-pitch helical dataacquisition, previous studies have shown advantages of using step-and-shootmode acquisition to significantly improve dose efficiency. We propose new recon-struction approaches to effectively improve the temporal resolution of the scan-ner by reducing cardiac motion artifacts and to suppress cone beam artifacts.Methods and Materials: In helical acquisition, cardiac phases and scan loca-tions are closely coupled, and the projection view range is selected primarily bythe EKG signal. Since these parameters are completely independent in the step-and-shoot case, we utilize a set of projection consistency conditions to further

adjust the view selection to minimize the banding artifacts caused by the residualphase misalignment. Consistency conditions include both H-L condition and cor-relation pattern. In addition, a complimentary cone beam reconstruction algo-rithm is used to compensate for the reduced FOV due to cone beam effect in z.Two adjacent scans are used in combination to produce images.Results: Both phantom and animal experiments were conducted. Phantom stud-ies with high contrast objects have shown that the complimentary reconstructionalgorithm is effective in combating cone beam artifacts due to extrapolation. An-imal and balloon experiments have shown that optimization of view selection bythe proposed algorithm can significantly reduce banding artifacts and improvethe effective temporal resolution.Conclusion: We proposed new approaches for the step-and-shoot cardiac im-aging. Both animal and phantom studies have shown significant improvements inimage quality in terms of motion and cone beam artifacts reduction. Clinical ex-periments are currently under way.

B-423 14:18

CEnPiT: Helical cardiac CT reconstructionC. Bontus, P. Koken, M. Grass, T. Koehler, R. Proksa; Hamburg/DE([email protected])

Purpose: Evaluation of a retrospectively gated helical cardiac CT reconstructionmethod, which is based on an exact algorithm. Reconstruction results from a CTsystem with 64 rows as well as simulated data for systems with 128 rows areused.Methods and Materials: For cone-beam CT scanners with a large number ofdetector rows, exact reconstruction algorithms are necessary. Otherwise cone-beam artifacts in the images reach an unacceptable level. When applying ECGbased gating in a naïve way to the acquired helical projection data, the acquisi-tion becomes incomplete. Considering cardiac CT, the question, how an exactalgorithm can be combined with a gating approach is addressed. The used ap-proach splits the data into contributions of different kinds of Radon-planes. Inparticular, we combined different gating-functions with different kinds of Radon-plane contributions. Radon-planes intersecting only once with the helix are meas-ured without redundancy. Therefore, the corresponding contributions areback-projected without gating. For Radon-planes with the maximum number ofintersections, the full gating-scheme is applied.Results: The reconstruction results are very promising. The images show thatthe algorithm is capable of delivering diagnostic image quality for large coneangle CT systems, though gating is applied. Neither the inclusion of data fromthe complete scan, nor the variability of the heart rate introduce additional arti-facts. The reconstructed images were analyzed regarding image quality and arti-facts.Conclusion: An exact filtered back-projection algorithm can be combined with acardiac gating approach. Thus the method is a good candidate for future systemswith a large area detector.

B-424 14:27 !Streaking reduction in reconstruction of CT using waveletJ.J. Liu, S.R. Watt-Smith, S.J. Golding, D.B. Dobson, C.M. Alvey; Oxford/UK([email protected])

Purpose: When metal implants present, low signal-to-noise ratio is caused inprojection and the noises are with high energy. Conventional methods reducedthe noises by smoothing or interpolating the affected region in the projection,which inevitably cause errors and blur reconstruction. This study provides a newtechnique to amend the wavelet coefficients of the projection by developing ashrink algorithm based on statistical evaluation to isolate the noises from theprojection.Methods and Materials: Two dimensional phase shift wavelet decomposition isadapted to decompose the projection into several components. The energy ofeach component in the affected region is compared with that in the rest part ofthe projection, and a shrink algorithm is derived to reduce the correspondence ofthe noises in the projection according to statistical evaluation. A repaired brokenbone of a leg of pork is scanned with a stainless steel plate and screws to dem-onstrate the method.Results: A pig's leg with a lead nail inside the bone was scanned with the param-eters of 140.0 kV and 40 mA. FBP with interpolation causes low frequent distur-bance, normal wavelet method cannot reduce streaking completely. The methodin this study reduces the streaking effectively without significant reconstructionblurring.Conclusion: The decomposition of projection using phase shift wavelet describesthe noises with high energy effectively, and a shrink algorithm is derived accord-ing to the distribution of the noises in wavelet domain to isolate the correspond-

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 231C D E FA G

ence of the noises. The streaking is removed completely from the reconstructionwithout significant blurring.

B-425 14:36

Dosimetry and image quality assessment of different acquisition protocolsof a 64 slice computerised tomographic (CT) scannerF. Tanzi, R. Novario, C. Vite, M. Mangini, S. Strocchi, G. Carrafiello, L. Conte,C. Fugazzola; Varese/IT ([email protected])

Purpose: Dose and image quality assessment in CT are affected by the varietyof CT scanners and imaging protocols in use. Our aim is to assess image qualityrelated to patient dose indexes.Methods and Materials: Dose and image quality were all measured with a stand-ard CT phantom for the main protocols (head, chest, abdomen and pelvis).Results: CT dose indexes (CTDIw, DLP) for each protocol were compared withvalues displayed by the scanner. The differences were always below 7%. All in-dexes were below the DRL defined by the European Council directive 97/42.Effective doses were measured for each protocol with thermoluminescent dosim-eters inserted in an anthropomorphic phantom and compared with the same val-ues computed by the ImPaCT CT patient dosimetry calculator software andcorrected by a factor taking in account the number of slices (16 to 64). The differ-ences were always below 20%. The effective doses range from 1.5 mSv (head)to 12.4 mSv (abdomen). The dose reduction algorithm of the scanner was as-sessed comparing the effective dose measured in a phantom to the dose com-puted on 46 patients. The reduction factors range from 5% to 32%, depending onthe protocol and patient size. The possibility of a further dose reduction was in-vestigated measuring spatial resolution, contrast and noise as a function of CT-DIw. This curve shows a quite flat trend decreasing the dose approximately to80% and a sharp fall below that value.Conclusion: All the information allowed optimizing the protocols obtaining anoverall mean dose reduction up to 35%.

B-426 14:45

Visualization of 3D low contrast objects by cone beam reconstruction witha dynamic large flat panel detector in rotational angiography (CT-mode)R.W.R. Loose1, M. Wucherer1, R. Adamus1, T. Brunner2; 1Nuremberg/DE,2Forchheim/DE ([email protected])

Purpose: To evaluate the low-contrast resolution of a dynamic CsI-aSi flat paneldetector for 3D-reconstructions from rotational angiography (CT-mode) for phan-toms and clinical examinations.Methods and Materials: 3D datasets of low contrast objects and patients wereacquired with a dynamic digital 40x30 cm FD angiography system (Axiom ArtisdTA, Siemens) with rotation speeds of 60°/sec and 30 frames/sec image fre-quency. The detector matrix of the FD is 2480x1920 with 154 µm pixel size. Theprojection data were reconstructed with a CT cone beam algorithm by summingof 2 contiguous detector lines and columns (binning). A spatial resolution of 0.3 mmat the detector level and 0.2 mm at the object level was obtained with up to 1000reconstructed slices from one 220° rotation.Results: With detector entrance doses of 0.36 µGy and 1.2 µGy we obtaineddoses at the 16 cm CTDI phantom of approx. 13 mGy and 43 mGy which arebelow the diagnostic reference values for CT head examinations. At the 16 cmCATPHAN phantom objects with 10 mm diameter at 5 mm slice thickness with≤ 5 HU are visible. Cerebral structures with liquor spaces and dilated bile ductswere visualized for the first time in head and abdominal examinations withoutcontrast and with high image quality.Conclusion: The significant higher image quality of a dynamic FD in comparisonto an image intensifier enables for the first time 3D acquisitions with excellent lowcontrast resolution in clinically acceptable acquisitions times without exceedingthe CT dose reference values.

B-427 14:54

Quantitative analysis of coronary arteries with computer tomographycoronary angiography: Methodology and preliminary results usingintracoronary ultrasound as reference standard using a software prototype(QCT)F. Cademartiri, G. Runza, L. La Grutta, A. Palumbo, N.R. Mollet, N. Bruining,R. Hamers; Rotterdam/NL ([email protected])

Purpose: To compare a newly developed software (QCT) for the quantitativeanalysis of coronary lumen and plaque with computer tomography coronary an-giography (CT-CA) with intracoronary ultrasound (ICUS).Methods and Materials: 15 patients with suspected coronary artery disease

underwent retrospectively ECG-gated CT-CA (Sensation 16, Siemens, Germa-ny) and conventional coronary angiography (CAG) with retrospectively ECG-gat-ed ICUS (Intelligate, CURAD, The Netherlands). A dedicated software (QCT,CURAD, The Netherlands) was used to extract the CT-CA dataset for the evalu-ation of the coronary arteries. The CT-CA dataset and ICUS dataset were loadedinto a software platform for quantification purposes. Then, a target vessel regionwas identified for both CT-CA and ICUS. An operator performed all QCT quantifi-cation and another one all ICUS quantification. The parameters analysed were:target vessel length, lumen volume, vessel volume, plaque volume. The differ-ence between the measurements were tested with a paired t-test and a p value of< 0.05 was considered significant.Results: Overall 15 coronary vessels were studied (12 left anterior descending,2 left circumflex and 1 right coronary artery). The mean target vessel length was24.7 ± 12.9 and 25.7 ± 14.4, the mean lumen volume was 242.7 ± 121.5 and255.6 ± 157.3, the mean vessel volume was 428.7 ± 212.7 and 476.1 ± 292.8,the mean plaque volume was 189.4 ± 94.6 and 220.5 ± 149.9, for ICUS and QCT,respectively (p > 0.05). There were no significant differences when the repeatedQCT measurements were compared.Conclusion: Quantitative parameters can be extracted from CT-CA dataset. Thenew software allows to obtain quantitative assessment of coronary artery param-eters without significant differences with ICUS.

B-428 15:03 !Beam hardening approximation in CT using neural networksJ.J. Liu, S.R. Watt-Smith, S.J. Golding, D.B. Dobson, C.M. Alvey; Oxford/UK([email protected])

Purpose: Beam hardening is regarded as one of the key factors causing artefactand impeding a properly post-operation evaluation using computed tomography(CT), especially when a metal implant presents. This research provides an algo-rithm to approximate beam hardening in the projection of CT to deal with the ill-posted problems, amend the projection, and obtain a clearer reconstruction.Methods and Materials: Based on the initial reconstruction, ideal correspond-ence of the metal implants and human tissues is generated separately by forwardmodelling, which is used as reference of beam hardening approximation. A groupof curves are selected for beam hardening adjustment and the parameters areidentified using neural networks under the supervised by consistency test, toobtain an optimal beam hardening approximation and computational efficiency.The projection is amended according to the beam hardening approximation priorto reconstruction. This method is demonstrated by an experimental scanning.Results: A repaired broken bone of a leg of pork is selected as the scanningobject with a stainless steel plate and scews, with the parameters of 140.0 kVand 40 mA, using a GE CT scanner. In the original reconstruction, serious arte-fact was caused. Using the method in this research, the information lost in theoriginal reconstruction was recovered and the reconstruction was significantlyclearer.Conclusion: Beam hardening approximation is transferred into parameters iden-tification, which is performed automatically using neural networks under the su-pervision of consistency test. The amended projection provides clearerreconstruction where the lost information caused by beam hardening is success-fully recovered.

B-429 15:12 !Heartbeat-operating phantom designed for CT scanning and nuclearmedicine examinationsK. Tsujioka, Y. Uebayashi, M. Uchida, T. Takaki, S. Enomoto, Y. Okubo,M. Anzui; Toyoake-city Aichi/JP ([email protected])

Purpose: Although examinations of heart function are widely conducted in CTscanning and nuclear medicine, there is no existing dynamic type of phantomwhich can be used for both purposes. In this experiment we created a new phan-tom with heartbeat movement, which can be used in CT scanning as well asnuclear medicine examinations.Methods and Materials: We created a completely new type of phantom, whichexactly imitates heart action. We injected a contrast agent into the heart lumenand radioactive isotope into the heart muscle. The aortic valve began to move,triggered by the heart movement, thus causing the ventricle lumen to distort atcontraction. With this phantom we studied the heart function by both CT scanningand nuclear medicine examination and compared the results.Results: We measured the cardiac output rate by both CT scanning and nuclearmedicine examination using the phantom, operating under the same conditions.Compared with the primary cardiac output rate of 45%, we observed a 43% car-diac output rate in the CT scan and 35% in the nuclear medicine.

Scientific Sessions

232 C D E FBA G

Conclusion: The heartbeat-operating phantom we used in this experiment pre-cisely duplicated actual heart function. In the CT scan, we were able to obtainaccurate data of variations in the position of the aortic valve as well as the condi-tion of the ventricle lumen. However, the nuclear medicine examination failed toproduce accurate data of cardiac output rate. These results, therefore, have prov-en that CT scanning is capable of accurate volume measurement.

B-430 15:21

Influence of automatic exposure control and body habitus on effective dosein MDCT of the abdomen: A phantom studyG.M. Weiner, D. Hellermann, M. Gergeleit, G. Agla, K.-J. Wolf, T. Albrecht;Berlin/DE ([email protected])

Purpose: This phantom study was conducted to compare the ability of two differ-ent versions of automatic exposure control (AEC) systems to reduce effectivedose in abdominal CT at different patient sizes.Methods and Materials: Effective dose was measured in an Alderson-Rando-Phantom with 100 thermoluminescence dosimeters, representing the 23 organsrequired by ICRP. Abdominal CT-scans were performed using a 16-slice multide-tector-CT (Somatom Sensation16, Siemens). The original AP-diameter of thephantom (22 cm) was increased to 28 and 34 cm using two additional layers offat-equivalent material to simulate slim, normal and obese patients. The threephantom sizes were scanned without AEC, with mAs-modulation only in the axialplane (CareDose2D®, Siemens) and using mAs-modulation in the axial plane aswell as along the Z-axi (CareDose4D®). Each measurement was performed threetimes.Results: The mean effective dose for the slim phantom was 12.5mSv withoutAEC, 11.4mSv with CareDose2D and 6.3mSv with CareDose4D. For the normalsized phantom these measurements were 18.2, 16.4 and 12.9mSv and for theobese phantom 19.8, 17.1 and 15.9mSv respectively. The mean dose reductionfrom no AEC to CareDose2D for the slim, normal and obese phantom were 9.5%,10.0% and 13.4% and from no AEC to CareDose4D 49.7%, 29.4% and 19.5%.Conclusion: AEC reduced effective dose in all phantom sizes. CareDose4D wasmore effective than CareDose2D in all experiments. Most substantial dose re-ductions can be achieved in slim patients (children) with CareDose4D.

14:00 - 15:30 Room P

Vascular

SS 715Peripheral vascular systemModerators:A. Basile; Catania/ITM. Maynar; Santa Cruz de Tenerife/ES

B-431 14:00

State-of-the-art cardiovascular imaging with parallel acquisition techniqueson a whole body MR scanner: Experience in more than 200 individualsH. Kramer, H.J. Michaely, K. Nikolaou, M.F. Reiser, S.O. Schönberg;Munich/DE ([email protected])

Purpose: Optimizing a magnetic resonance (MR) protocol using parallel acquisi-tion techniques (PAT) on a dedicated whole body scanner to image the entirecardiovascular system without compromises in image quality.Methods and Material: 200 individuals (age 55 ± 8 years) participating in a health-care program underwent a whole-body MR exam. Real-time functional imagingof the heart and dynamic cardiac perfusion at rest and, with special indications,at stress, was performed with high temporal (48 ms) and spatial resolution. MRangiography (MRA) with PAT was performed within 88 seconds to cover the bodyat a spatial resolution of < 1.4×1.0×1.5 mm3. MRA was divided into 26 segmentsand judged in terms of vessel conspicuity, venous overlay and artifacts. Imagequality and pathologic findings were evaluated by two blinded readers (whetherdiagnostic/nondiagnostic and present/absent).Results: We detected 19 myocardial pathologies (5 functional abnormalities, 12perfusion defects, 2 infarcted areas). MRA showed 42 pathologies, e.g. peripher-al vessel occlusions or high grade stenoses. More than 80% of all segmentswere rated good in terms of vessel conspicuity, > 90% showed no artifacts, > 95%had no venous overlay. There was good agreement between both readers (kappavalues between 0.69 and 0.88). Examination time could be reduced to less than80 minutes.Conclusion: PAT offers fast cardiovascular imaging with high image quality. This

enables the combination of different morphologic and functional techniques in awhole body protocol within less than 80 minutes without any compromises interms of spatial or temporal resolution.

B-432 14:09

High resolution peripheral MRA using a keyhole hybrid technique incombination with SENSEB. Tombach1, T. Allkemper1, B. Buerke1, R. Hoogeveen2, H. Kooijman3,W. Heindel1; 1Münster/DE, 2Best/NL, 3Hamburg/DE([email protected])

Purpose: Since venous enhancement is a significant pitfall of bolus-chase tech-niques, for peripheral contrast-enhanced MR angiography, the feasibility of a key-hole hybrid technique was evaluated.Methods and Materials: On a clinical 1.5 T scanner (Intera, Philips, Eindhoven,The Netherlands) a keyhole-hybrid technique was investigated in a total of 30patients suspected of having peripheral arterial occlusive disease, using a dedi-cated peripheral coil. Following a manual bolus timing in the abdominal aorta(1 cc of 1 M Gadobutrol (Schering, Berlin, Germany, flow rate 2 cc/s, followed by20 ml saline) the run-off vessels were imaged first with 0.1 mmol/kg bw 1 M Gado-butrol using 8 dynamic keyhole scans (TA 10.3 s) followed by a 44.3 s acquisitionof peripheral k-space lines. A second injection of 0.1 mmol/kg bw 1 M Gadobutrolwas used for a two station bolus-chase MRA of the abdominal aorta and upperlegs. Qualitative evaluation was performed by two experienced radiologists inconsensus with regard to arterial contrast, venous overlay and overall imagequality.Results: The keyhole technique allowed discrimination of the arterial phases onboth sides of run-off vessels separately with excellent image quality in each in-vestigated patient. Using a SENSE factor of 1.8 this technique enabled meas-ured (reconstructed) voxel sizes of 1.22 x 1.22 x 1.22 (0.88 x 0.88 x 0.61 mm) inthe run-off vessel system.Conclusion: The keyhole hybrid technique is a robust method for clinical routineto overcome venous contamination in patients suspected of having peripheralarterial occlusive disease.

B-433 14:18

Intravascular optical coherence tomography: Differentiation ofatherosclerotic plaques and quantification of vessel dimensions in cruralarterial specimensO.A. Meissner, J. Rieber, G. Babaryka, M. Oswald, S. Reim, M.F. Reiser,U. Mueller-Lisse; Munich/DE ([email protected])

Purpose: Intravascular optical coherence tomography (OCT) is a new imagingmodality providing micro-structural information of atherosclerotic plaques with aresolution of 10-20 µm. Intravascular ultrasound (IVUS), current reference stand-ard in vivo, has limited resolution of 150-300 µm. Purpose of the present studywas to compare OCT, histopathology and IVUS in differentiating atheroscleroticplaques and quantifying vascular dimensions in crural arteries ex vivo.Methods and Materials: 50 atherosclerotic segments were derived from 5 am-putated lower limbs. Different plaque types (fibrous, lipid-rich, calcified) wereassigned by 2 examiners and sensitivity, specificity as well as intra- and interob-server agreement were calculated. Quantitative measurements comprised lumi-nal area (LA) and vessel wall area (VA).Results: 50/50 segments were available for comparison. Sensitivity and specifi-city for OCT criteria were 81% and 89% for fibrous plaques, 100% and 93% forlipid-rich plaques, and 80% to 89% for calcified plaques (overall agreement 83%).Inter- and intraobserver agreement was high (κ ϖαλυεσ οφ 0.86 ανδ 0.89,π < 0.001). There was also a high correlation between quantitative measurements(Bland-Altman plot [LA]: mean bias, 0.1 mm²; accuracy ± 1.8 mm² [p = 0.9],r = 0.95 [p < 0.001]; Bland-Altman plot [VA]: mean bias, 0.3 mm²;accuracy ± 2.3 mm² [p = 0.9], r = 0.94 [p < 0.001].Conclusion: OCT provides histology-like information of different atheroscleroticplaque types in human crural arteries ex vivo. A high degree of agreement wasfound between OCT and histopathology. Quantitative measurements showed ahigh correlation with IVUS, the current standard of reference.

Sa

Sa

Sa

Sa

Sa

tur

tur

tur

tur

turd

ad

ad

ad

ad

ayy yyy

Scientific Sessions

B 233C D E FA G

B-434 14:27 !Thrombin augments arteriogenesis and ischemic tissue reperfusion in therabbit: A functional-CT and angiographic comparative studyK. Katsanos, C. Kalogeropoulou, G.C. Kagadis, J. Maroulis,A. Diamantopoulos, N. Tsopanoglou, T. Petsas, D. Siablis; Patras/GR([email protected])

Purpose: To evaluate and compare the angiogenic and arteriogenic capacity ofthrombin with other well-established growth factors (bFGF, VEGF) in the NewZealand rabbit, in hindlimb ischemia.Methods and Materials: Bilateral hindlimb ischemia was invoked by surgicalexcision of the femoral artery. On day 20, after quiescence of endogenous angio-genesis 5,000 IU of Thrombin or 10 µg of VEGF or 10 µg of bFGF was intramus-cularly injected in the medial thigh of one ischemic hindlimb per rabbit (n = 4-5 ineach group), while the contralateral limbs were injected with equal volumes ofnormal saline and served as the control groups. On day 40, functional CT (fCT) atthe level of the saphenous artery below-the-knee and intraarterial DSA followedby computerized quantitative analysis of the depicted vasculature were performed.Results: fCT studies calculated significantly increased peak enhancement val-ues and shorter time-to-peak periods in all groups of agents as compared tocontrol hindlimbs. Quantitative DSA detected a significantly increased total areaof large collaterals only in the Thrombin group (p < 0.05). Histopathology of theadductor muscle in the medial thigh documented a significantly increased arteri-ogenic outcome in the thrombin group (p < 0.05) and only a mild non-significantelevation of small tunica-covered vessels in the bFGF and VEGF groups.Conclusion: Thrombin augments arteriogenesis and ischemic tissue reperfusionand seems to outperform bFGF and VEGF in establishing new vessel conduitswith a developed tunica media in the ischemic rabbit model.

B-435 14:36

Radial artery harvest in coronary artery bypass grafting: Use of ultrasoundof forearm arteries in selection of patients for the procedureB. Brkljacic, R. Huzjan Korunic, I. Cikara, D. Delic Brkljacic, G. Ivanac;Zagreb/HR ([email protected])

Purpose: To ultrasonographically evaluate radial (RA) and ulnar arteries (UA)and collateral forearm circulation prior to RA harvest for coronary artery bypassgrafting (CABG).Methods and Materials: B-mode and CDUS of RA and UA were performed in512 candidates for CABG, using a 14 MHz transducer. 409 were male, 103 fe-male, age range 35-85 (mean 63) years. Vessel diameters were measured, athero-sclerotic changes noted. Tests to evaluate collateral forearm circulation after RAharvest were performed: RA compression test (RACT) with PSV measurementsin UA before/during RA compression and superficial palmar RA branch flow testduring RA compression. RA harvesting was not recommended in abnormal func-tional tests, hypoplastic RA, diffuse and severe RA atherosclerotic changes.Results: Abnormal RACT was observed in 7.6%, abnormal superficial RA branchtest in 8.0%, RA hypoplasia in 4.9% of patients. Severe atherosclerotic changeswere noted in 7.5% UA and 3.2% RA (p < 0.001). Stenosis > 50% of RA/UA wasobserved in 10 pts (2%), abnormally high brachial artery bifurcation in 3 pts, du-plicated RA in 1 patient. RA and UA diameters were larger in males (p < 0.01).RA harvest on one/both sides was not recommended in 11.7% of pts. After RAharvesting (148 pts) one acute brachial artery thrombosis was observed. In oth-ers, no acute hand ischemia, motor and/or sensory deficits were noted in thefollow-up period of 1-34 months.Conclusion: US provides information about collateral forearm circulation afterRA harvesting, about RA/UA anatomy and atherosclerotic changes. US identifiesRAs not suitable for CABG and alters approach to CABG.

B-436 14:45

Diagnostic value of high resolution MR imaging in giant cell arteritisT.A. Bley, M. Uhl, M. Markl, H.-H. Peter, D. Schmidt, M. Langer; Freiburg/DE([email protected])

Purpose: Giant cell arteritis remains a diagnostic challenge. The purpose of thisstudy was to determine the diagnostic value of high resolution MR imaging of thetemporal arteries in giant cell arteritis.Methods: We performed high resolution MR imaging of the superficial temporalartery on 64 consecutive patients with suspected giant cell arteritis. Mural thick-ness and lumen diameter were assessed with contrast enhanced, T1-weightedspin echo images with sub-millimeter in-plane spatial resolution (195 µm x260 µm). MR imaging results were compared to the final rheumatologist's diag-nosis according to the clinical criteria of the American College of Rheumatologyincluding results of temporal artery biopsies from 32 patients.

Results: MR imaging sharply demonstrated the superficial temporal arteries al-lowing for an evaluation of their lumen and vessel wall. 31 of the 64 patients werefinally diagnosed as having giant cell arteritis. Scoring of the mural inflammationwith MR resulted in a sensitivity of 80.6 % and a specificity of 97.0 % as com-pared to a sensitivity of 77.8 % and specificity of 100 % based on histology. Themean wall thickness increased significantly from 0.39 (+-0.18) mm to 0.74 (+-0.32) mm (p < 0.001) and the lumen diameter decreased significantly from 0.84(+-0.29) mm to 0.65 (+-0.38) mm (p < 0.05) for the patients with giant cell arteritis.Conclusion: Contrast enhanced high resolution MR imaging is feasible for non-invasive assessment of mural inflammation in giant cell arteritis and assisting inthe diagnosis of the disease, the selection of a biopsy site, and long-term moni-toring of therapeutic intervention.

B-437 14:54

Multicenter randomized controlled trial of the costs and effects of non-invasive diagnostic imaging in patients with peripheral arterial disease: TheDIPAD trialR. Ouwendijk1, M. de Vries2, T. Stijnen1, P.M.T. Pattynama1,M.R.H.M. van Sambeek1, A.V. Tielbeek3, L. Schutze Kool4, M.W. de Haan2,J.M.A. van Engelshoven2; 1Rotterdam/NL, 2Maastricht/NL, 3Eindhoven/NL,4Nijmegen/NL ([email protected])

Purpose: To prospectively compare the costs and effects of three non-invasiveimaging tests as the initial imaging test in the diagnostic workup of patients withperipheral arterial disease.Method and Materials: A total of 514 patients with peripheral arterial diseasewere randomized to magnetic resonance angiography (MRA) or duplex ultra-sound (DUS) or computed tomographic angiography (CTA) in four hospitals. Theoutcome measures included the clinical utility, functional patient outcomes, qual-ity of life, and costs related to the initial imaging test during 6 months follow-up.Clinical utility was assessed with therapeutic confidence (0-10) in the initial im-aging test. Functional patient outcomes included change in ankle-brachial index,maximum walking distance, and clinical status. Quality of life was assessed us-ing 3 questionnaires. The diagnostic cost included the initial imaging test and alladditional vascular imaging. The therapeutic cost included percutaneous vascu-lar interventions and vascular surgery. We analyzed the differences between groupmeans adjusted for clinically important baseline characteristics with multivaria-ble linear and logistic regression.Results: A significant higher clinical utility was found for MRA and CTA com-pared to DUS. There were no statistically significant differences in improvementin functional patient outcomes and quality of life between the groups. The totalcosts were significantly lower in the CTA group compared to the MRA group andthe DUS group.Conclusion: The results suggest that both CTA and MRA are clinically moreuseful than DUS and that CTA leads to cost-savings compared to both MRA andDUS in the initial imaging evaluation of peripheral arterial disease.

B-438 15:03

Peripheral MR angiography with MultiHance: Results of large-scalemulticenter experienceS. Thurnher1, G. Schneider2, S. Miller3, C. Ballarati4, G. Bongartz5,S. Schoenberg6, C. Herborn7, G. Morana8, G. Pirovano9; 1Vienna/AT,2Homburg/Saar/DE, 3Tuebingen/DE, 4Como/IT, 5Basel/CH, 6Munich/DE,7Essen/DE, 8Verona/IT, 9Princeton, NJ/US ([email protected])

Purpose: To assess the accuracy of MultiHance®-enhanced (CE)-MR angiogra-phy in detecting peripheral artery disease using DSA as the gold standard.Methods and Materials: Of the 287 recruited patients with Fontain's stage pe-ripheral disease IIa to IV, 272 underwent both MRA and DSA. MRA was acquiredat 1.5 T using a 2D-TOF before and a 3D-SPGRE sequence after the administra-tion of 0.1 mmol/kg MultiHance® covering from the aortic bifurcation to the trifur-cation. Three experienced radiologists for MRA and one for DSA, blinded to allpatient information, evaluated the images. Sensitivity, specificity, accuracy, tech-nical failure rate and inter-reader agreement for detection of significant disease(≥ 50%) using DSA as gold standard were calculated for both TOF and CE-MRAand compared using McNemar's and Chi-square tests.Results: A total of 983 vessels with significant disease (597 stenoses and 386occlusions) were identified by the DSA reader. Sensitivity for detection of signif-icant disease ranged from 33.2% to 62.8% and from 54% to 80.9% for TOF-MRAand CE-MRA, respectively. Specificity increased from 74.3-88.9% to 89.7-95.3%and accuracy from 68-77.3% to 85-87.5. All the increases were statistically sig-nificant (p < 0.001). Significant decreases (p < 0.001) in technical failure rate from6.2%-18% for TOF-MRA to 2.5%-3.4% for CE-MRA approaching the values ob-

Scientific Sessions

234 C D E FBA G

tained for DSA (1.4%). Significantly better reproducibility (p < 0.00001) was ob-tained for CE-MRA (82% agreement, k-value: 0.66) compared to TOF MRA (65.2%agreement; k-value: 0.47).Conclusion: Significant increases in diagnostic performance and reproducibilitywere demonstrated for MultiHance-enhanced MRA of peripheral vessels in com-parison with TOF MRA.

B-439 15:12

Standard bolus chase and time-resolved MR angiography for imaging theperipheral arteries in diabetic patientsG. Andreisek1, T. Pfammatter1, K. Goepfert1, D. Nanz1, P. Hervo2,R. Koppensteiner1, D. Weishaupt1; 1Zurich/CH, 2Buc/FR ([email protected])

Purpose: To determine the diagnostic performance of a combination of standardbolus-chase MRA and time-resolved imaging with contrast kinetics (TRICKS) formapping peripheral arteries of the lower extremity in diabetic patients.Methods and Material: Standard three-station single bolus-chase MRA and ded-icated TRICKS MRA of the calf and foot was performed within one imaging ses-sion in 31 consecutive diabetic patients. Two readers separately assessed allarterial segments from the renal arteries to the foot as diagnostic or non-diag-nostic and analyzed the grade of stenosis. The results were compared to DSA,which served as standard of reference when the corresponding arterial segmentswere considered diagnostic on DSA. Wilcoxon signed rank test was used to de-termine if a significant difference between the imaging techniques existed. Kap-pa statistics were used to determine interobserver agreement.Results: In the thigh, the differences between bolus-chase MRA and DSA withregard to diagnostic segments were not significant. A significantly higher numberof segments was considered diagnostic in TRICKS MRA of the calf and foot thanin standard bolus-chase MRA (p 50 % narrowing) arterial stenosis were 84/83%and 97/97% in the thigh (reader 1/2), 79/80% and 90/90% in the calf and notapplicable in the foot. TRICKS MRA improved sensitivity and specificity in thecalf (83/86% and 91/93%) and foot (58/66% and 90/88%).Conclusion: Standard bolus-chase MRA combined with time-resolved MRA us-ing TRICKS technique is a reliable alternative to DSA for assessing the arteriesof the lower extremity, including the pedal arteries, in diabetic patients.

B-440 15:21

Dysfunctional hemodialysis access fistulas and grafts: Prevalence andsignificance of arterial inflow stenosesF.J. Nobrega1, L.E.M. Duijm1, Y.S. Liem2, R.H.H. van der Rijt1, P.W.M. Cuypers1,A.V. Tielbeek1, P. Douwes-Draaijer1, H.C.M. van denBosch1; 1Eindhoven/NL,2Rotterdam/NL ([email protected])

Purpose: To determine the prevalence and significance of arterial inflow sten-oses in dysfunctional hemodialysis access fistulas (AVFs) and grafts (AVGs).Methods and Materials: Contrast-enhanced magnetic resonance angiography(CE-MRA) was performed of 60 failing AVFs and 33 AVGs in 54 men and 39women (mean age 62 years; age range 31-86 years). Complete arterial inflow(from the subclavian artery), shunt region and complete venous outflow (includ-ing subclavian vein) were depicted at CE-MRA. In addition to standard digitalsubtraction angiography (DSA) of the shunt region and outflow, DSA of the com-plete inflow was obtained if CE-MRA had depicted a significant (> 50%) arterialstenosis. Endovascular intervention of significant lesions was performed afterDSA.Results: CE-MRA demonstrated 16 significant arterial lesions in 11 (12%) pa-tients. DSA confirmed 15 of these lesions in 10 patients and depicted no addi-tional inflow lesions. Of these 10 patients, seven had an isolated arterial stenosisand three had accompanying stenoses in the shunt region and/or outflow. Refer-ral criteria for the 10 patients to undergo access evaluation had been decreasedflow rates (7 cases), steal symptoms (2) and insufficient access maturation (1).Access flow of the seven patients with low flow access improved from 496 ± 73 ml/min to 860 ± 212 ml/min after angioplasty. One patient with steal symptoms be-came symptom free after angioplasty. Endovascular intervention of two patientsproved to be unsuccessful.Conclusion: Flow-limiting inflow stenoses are frequently encountered in dys-functional hemodialysis access shunts. We suggest that assessment of the com-plete arterial inflow is comprised of CE-MRA or DSA.

B 235C D E FA

Su

nd

aS

un

da

Su

nd

aS

un

da

Su

nd

ayy yyy

Scientific Sessions

G

SundaSundaSundaSundaSundayyyyy, Mar, Mar, Mar, Mar, Marccccch 5h 5h 5h 5h 5

B

Scientific Sessions

236 A C D E F G

SY 11Satellite

SymposiumPACS/RIS andEPR - The truth

behind it all(p. 553)

IISImage

InterpretationSession(p. 72)

HL 2(p. 72)

Gold MedalAwards


Non traumaticabdominal



Non traumaticthoracic



Shoulder joint(p. 60)

RC 907GenitourinaryImaging andinterventional

urogenitalradiology

(p. 61)


applications:Coronary

radiology: Riskstratification and

stenosisdetection

(p. 61)

RC 901Abdominal andGastrointestinalCrohn's diseaseof the intestinaltract: Advances

in imaging(p. 62)

NH 9New Horizons

SessionImaging of

13C molecules:A new functional

diagnosticmethod inradiology

(p. 63)

RC 911Neuro

Spine imaging(p. 64)

SF 9Special Focus

SessionHistory ofradiology

(p. 64)


RadiologyHepatic


NH 10New Horizons

SessionPlaque imaging

(p. 69)

EM 2"ECR meets"

PolandAdvances in

Neuroradiology(p. 70)

SS 1010Musculo-skeletalTrauma(p. 232)

SY 8Satellite

SymposiumAdvances incontrast and

scanningprotocols for

MDCTapplications

(p. 551)

SS 1002BreastClinical

issues (2)(p. 234)

SY 9Satellite

SymposiumManaging

patients at risk ofCIN - An update

(p. 552)

SS 1004Chest

Pulmonarythromboem-bolism and

vasculardiseases(p. 236)

SS 1001GI Tract

CT and MRvirtual

colonoscopy:Examination

protocols(p. 238)


RadiologyBiopsy anddrainage(p. 240)


Sports injuries(p. 73)

SF 12Special Focus

SessionPET-CT inoncology

(p. 74)

RC 1202Breast

Contrast agentsin non

mammographicimaging(p. 75)

RC 1204Chest

Smoking-relateddiseases(p. 75)

RC 1211Neuro

Brain ageing(p. 76)

RC 1208Head and NeckCPA and internal

auditory canal(p. 77)

RC 1205Computer

ApplicationsWorkflow

re-engineeringfor a digital

hospital(p. 78)

EP

OS


22:00

registration


room A2nd level

room B2nd level

room C2nd level

room E1entr. level

room E2entr. level

room F1entr. level

room F2entr. level

room Hlower level

room Glower level

08:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:00

14:00

14:30

15:00

15:30

16:00

16:30

17:00

17:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:30

14:00

14:30

15:00

15:30

16:00

16:30

17:00

12:3012:30

13:00

18:00

18:30

19:00

18:30

19:00

18:00

17:30

13:30

08:30

08:00

07:30

07:00

07:30

08:00

07:00

SY 10Satellite

SymposiumBreakthroughapplications in

CT/MR imaging(p. 552)

Su

nd

aS

un

da

Su

nd

aS

un

da

Su

nd

ayy yyy

Scientific Sessions

B 237A C D E F G

WS 1224Virtual

InterventionsPeripheral"Hands-on"Workshop

WS 1024Virtual

InterventionsRenal

"Hands-on"Workshop

WS 1021Musculo-skeletal

Ultrasound"Hands-on"Workshop

E3 1220b IHE (Integratingthe Healthcare

Enterprise)(p. 81)

E3 1020bHow to write agrant proposal

(p. 72)

SS 1214Radiographers

Futurechallenges for

radiography andradiographers

(p. 252)

E3 1020aHow to submit anelectronic poster:

EPOSTM

(p. 72)

SS 1015Vascular

Imaging ofabdominal

vessels(p. 250)

E3 920Foundation


Large bowel:CT/MR imaging -

plain film nonbowel structures

(p. 69)


RadiologyAdvanced

arterialinterventions

(p. 78)


RadiologyThe indicationsfor stent-grafts

(p. 66)

CC 919Pediatric

RadiologyLung diseases

(p. 67)


Adult TodayInfection in theface/Infection in

the pelvis(p. 67)


PET: A rapidlygrowing imagingtechnique of the

next decade?(p. 68)



SS 1011NeuroBrain:

Degenerativeand MS(p. 242)

SS 1012Pediatric

Cardiac/Chest/Radiationexposure(p. 243)


CT technicaldevelopments

(p. 245)

SS 1007GenitourinaryFemale pelvis

(p. 248)

RC 1212Pediatric

Problem solvingin pediatricradiology

(p. 79)

RC 1207GenitourinaryFemale imaging

(p. 80)


Image qualityand dose in

multislice CT:Current practice,

new develop-ments and

optimisation(p. 80)

E3 1220aTips and tricks tosearch for and to

download...(p. 81)

WS 23C3Screening



WS 23C2Screening



WS 23C1Screening




"Hands-on"Workshop

WS 924Virtual


"Hands-on"Workshop

room Klower level

room L/M1st level

room N/O1st level

room Plower level

room Wbasement

room Y1st level

The Globe2nd level

08:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:00

14:00

14:30

15:00

15:30

16:00

16:30

17:00

17:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:30

14:00

14:30

15:00

15:30

16:00

16:30

17:00

12:3012:30

13:00

18:00

18:30

19:00

18:30

19:00

18:00

17:30

13:30

08:30

08:00

07:30

07:00

07:30

08:00

07:00

room Ilower level

room Xentr. level

room Z1st level

WS 23C4Screening



238 C D E FBA

Scientific Sessions

G

10:30 - 12:00 Room B

Musculoskeletal

SS 1010Articular cartilageModerators:A.Z. Ginai; Rotterdam/NLA. Grainger; Leeds/UK

B-441 10:30

3D water-excitation trueFISP for the diagnosis of articular cartilageabnormalitiesS.R. Duc, P. Koch, M.R. Schmid, J. Hodler, C.W.A. Pfirrmann; Zürich/CH([email protected])

Purpose: To evaluate the diagnostic performance of water-excitation trueFISP incartilage abnormalities of the knee.Methods and Materials: 29 patients (30 knees, 13 men and 17 women) with amean age of 56 years (range 18-86 years) were prospectively evaluated with apreoperative MR examination including a sagittal 3D trueFISP sequence and astandardized evaluation of cartilage surfaces during surgery. The mean time in-terval between MR imaging and surgery was 1 day (range 0-9 days). Duringarthroscopy the medial and lateral femoral condyles and tibial plateaux, the fem-oral groove and the patella were separately evaluated using a modified Outer-bridge score. The MR images were evaluated separately and blinded by tworeaders on two separate occasions, using the same score.Results: If the cutoff between normal and abnormal was set between grades 1and 2, sensitivity, specificity and accuracy were 54%-64%, 76%-92% and 70%-73% respectively. For a cutoff between grades 2 and 3, sensitivity, specificity andaccuracy were 53%-64%, 91%-95% and 80%-83% respectively. Mean interob-server agreement was 0.73 and 0.65 for the first and second evaluations respec-tively at the first cutoff, and 0.78/0.79 at the second cutoff. Intraobserver agreementwas 0.84/0.78 for Reader 1 and 0.60/0.83 for Reader 2.Conclusion: 3D water-excitation trueFISP represents a viable alternative to im-aging sequences currently used for detection of cartilage abnormalities.

B-442 10:39

Evaluation of patellar cartilage volume and thickness at 3.0 T: 3D-trueFISPversus 3D-FLASHS. Weckbach1, T. Mendlik1, S. Wagner1, W. Horger2, M.F. Reiser1, C. Glaser1;1Munich/DE, 2Erlangen/DE ([email protected])

Purpose: To compare the reproducibility of patellar cartilage volume and thick-ness measurements with a validated 3D-FLASH WE and an optimized 3D-true-FISP WE sequence at 3.0 T.Methods and Materials: The patellar cartilage of 6 healthy volunteers was ex-amined. The measurements were performed on a 3.0 T whole body imager (Mag-netom Trio, Siemens Medical Solutions) using a transmit-receive extremity-coil.The patella was covered by 40 axial partitions (0.31x0.31x1.5 mm). Image datawere acquired with a 3D-FLASH WE and a 3D-trueFISP WE sequence. To as-sess the reproducibility, 3 consecutive data sets of each volunteer were acquiredfor both sequences, the knee joints being repositioned. Cartilage was delineatedby a segmentation routine. Volume, mean and maximum thickness were calculat-ed, reproducibility and interindividual variability were determined.Results: Patellar cartilage volume and thickness calculated from the trueFISPimages were smaller than from the FLASH images (p > 0.05). Reproducibility ofcartilage volume and thickness showed slightly lower values for the trueFISPsequence, the inter-individual variability was comparable.Conclusion: In healthy volunteers, patellar cartilage volume and thickness from3D-trueFISP images were smaller than from the 3D-FLASH data. The lower val-ues resulting from the trueFISP sequence might be attributable to high SNR/CNR of synovial fluid that might lead to underestimation of cartilage. Reproduci-bility was slightly better for the FLASH sequence, reflecting a complex signalbehaviour of the trueFISP-sequence with a slightly lower reliability of cartilagedelineation. SSFP techniques however, may improve cartilage segmentation inOA patients where cartilage-joint differentiation is difficult.

B-443 10:48

Imaging articular cartilage of the ankle at 3.0 and 1.5 TeslaJ.S. Bauer, C. Barr, L. Steinbach, D. Malfair, R. Krug, C. Ma, T.M. Link;San Francisco, CA/US ([email protected])

Purpose: To optimize MR imaging of the ankle cartilage at 1.5 Tesla (T) and3.0 T in volunteers and to compare these optimized sequences concerning im-age quality and in assessing cartilage pathology in fresh human cadaver speci-mens.Methods and Materials: Bandwidth, repetition time (TR), echo time (TE) as wellas flip angle were optimized for (i) a fat-saturated (fs) fast spin echo (FSE), (ii) afs spoiled gradient echo (SPGR), and (iii) a balanced free precession steadystate (bSSFP) sequence in three volunteers at 1.5 T and 3.0 T. Optimization out-come measures were effective signal- and contrast-to-noise-ratios (SNR, CNR),artifacts and image quality assessed by two radiologists in consensus. Using theoptimized sequences 74 regions of interest in thirteen fresh human cadaver an-kle joints with (n = 33) and without (n = 41) cartilage pathology were imaged.Four radiologists independently assessed cartilage pathology, separately for everysequence according to a five-level-confidence-scale. Macroscopic findings afterdissection served as standard of reference. Receiver-operator-characteristic(ROC) curves were calculated to investigate diagnostic performance for the dif-ferent sequences at 1.5 T and 3.0 T.Results: The optimized sequences at 3.0 T were characterized by higher BWand TR as well as SNR and CNR. The FSE sequence at 3.0 T showed the high-est area under the curve (AZ

value 0.89), significantly better than all other se-

quences (at 3.0 T: AZ bSSFP=0.83, AZ

SPGR=0.80; at 1.5 T: AZ

FSE=0.77, AZ

bSSFP=0.75, AZ SPGR=0.67, p < 0.05).

Conclusion: Among all protocols used, the 3.0 T-FSE sequence was suited bestto assess cartilage pathology at the ankle. Comparing 3.0 T and 1.5 T, all se-quences consistently showed a better diagnostic performance at 3.0 T.

B-444 10:57

Reproducibility of patellar cartilage T2 relaxation times at 1.5 T and 3 TC. Glaser1, T. Mendlik1, S. Wagner1, W. Horger2, M.F. Reiser1; 1Munich/DE,2Erlangen/DE ([email protected])

Purpose: To compare reproducibility of T2 relaxation time quantitation in articu-lar cartilage at 1.5 and 3.0 T.Methods and Materials: Three consecutive Multiecho (3000 at 1.5 T; 4500 at3 T/13.2-105.6 ms) and Flash datasets (0.6² x 3 mm³) were acquired from patel-lar cartilage of the right knees of 6 healthy volunteers with repositioning of theknees between the measurements. Segmentation of cartilage (from Flash) wassuperimposed on the Multiecho sequence and T2 values were calculated on apixel by pixel basis. Reproducibility was determined as coefficient of variationfrom the 3 consecutive measurements for the total patella, superficial, middleand deep layers and for both facets and ridge.Results: The intra-individual reproducibility for the lower, middle and upper layerranged from 0.9 to 7.5% (1.5 T)/0.7 to 5.0% (3.0 T), from 1.3 to 5.1% (1.5 T)/2.5to 7.5% (3.0 T) and from 3.0 to 9.6% (1.5 T)/1.4 to 9.5% (3.0 T) respectively. Theaverage reproducibility was 4.1, 3.8 and 6.3% (1.5 T)/3.7, 4.6 and 5.4% (3.0 T)respectively. Inter-individual variability was 2.7, 4.7 and 7.2% (1.5 T)/7.2, 4.5 and5.6% (3.0 T) respectively. Cartilage T2 values at 1.5 T data were significantlyhigher than at 3.0 T data for all three cartilage layers (p1 = 0.00002, p2 = 0.0009,p3 = 0.03).Conclusion: Reproducibility of regional patellar cartilage T2 times varies be-tween 1 and 9% and is comparable between 1.5 T and 3 T. It is small comparedto described changes in OA (up to 100% increase). Data give a basis for follow-upexams in OA.

B-445 11:06

Quantification of T1 and T2 in the presence of Gd-DTPA of proteoglycan andcollagen depleted cartilageE. Wiener, K. Wörtler, M. Settles, E.J. Rummeny; Munich/DE([email protected])

Purpose: To compare the tissue distributions of Gd-DTPA (Magnevist®) in selec-tively proteoglycan and collagen depleted cartilage. To quantify and compare T1and T2 in the presence of Gd-DTPA throughout cartilage thickness of untreated,papain-treated and collagenase-treated patellae.Methods and Materials: MR-studies were performed with nine bovine patellae.Three specimens were exposed to papain-solution for proteoglycan depletion, tocollagenase-solution for collagen depletion and to buffer solution without enzy-matic treatment. Then, patellae were placed in 2.5 mM of Gd-DTPA solution and

B 239C D E FA

Su

nd

aS

un

da

Su

nd

aS

un

da

Su

nd

ayy yyy

Scientific Sessions

G

high resolution (in plane resolution 300 µm) T1-weighted SE, T1- and T2-param-eter images were acquired on a 1.5 T scanner every 30 minutes for 11hrs. T1and T2-changes were calculated throughout cartilage thickness and comparedfor intact, proteoglycan- and collagen-depleted cartilage. Furthermore matrix-degradation was correlated with histological findings.Results: Histological findings demonstrate that these enzymatic treatments suc-cessfully altered collagen and proteoglycan concentration selectively.For proteoglycan and collagen degraded cartilage a significant decrease in T1and T2 in the presence of Gd-DTPA was observed in comparison to intact carti-lage. The T1- and T2-effects were of the same magnitude and depended on thedepth of the cartilage layer, being most pronounced in superficial layers (1 mm)and less pronounced in the deeper calcified layers. There was no difference,neither in T1 nor in T2 in the presence of Gd-DTPA between proteoglycan andcollagen degraded cartilage.Conclusion: T1 as well as T2 can be used for Gd-DTPA enhanced cartilageimaging to demonstrate early matrix-degradation, but it is not possible to differ-entiate between proteoglycan and collagen loss.

B-446 11:15

Delayed gadolinium-enhanced magnetic resonance imaging (dGEMRIC) ofhip joint cartilage. Better cartilage delineation after intraarticular thanintravenous gadolinium injectionM. Boesen1, K.E. Jensen2, E. Qvistgaard1, B. Danneskiold-Samsoe1,C. Thomsen2, M. Østergaard3, H. Bliddal1; 1Frederiksberg, Copenhagen/DK,2Copenhagen/DK, 3Hvidovre and Herlev, Copenhagen/DK([email protected])

Purpose: To investigate and compare delayed gadolinium (Gd-DTPA)enhancedMR imaging of cartilage (dGEMRIC) in the hip joint, using intravenous (i.v). orultrasound-guided intraarticular (i.a.). Gd-DTPA injection.Methods and Materials: In 10 patients (50% males, mean age 58y), with clinicaland radiographic hip osteoarthritis (OA; Kellgreen score II-III), MR imaging of thehip was performed twice on a clinical 1.5 Tesla MR-scanner: on day 1, before and90-180 minutes after 0.3 mmol/kg body weight i.v. Gd-DTPA; and on day 8, 90-180 minutes after ultrasound-guided i.a. injection of a 4 mmol/l Gd-DTPA solu-tion. Coronal STIR, coronal T1 fat-saturated spin-echo and a cartilage sensitivegradient-echo sequence (3D T1 SPGR) in the sagittal plane were applied.Results: Both the post-i.v. and post-i.a. Gd-DTPA images showed significantlyhigher signal to noise (SNR) and contrast to noise (CNR) in the joint cartilage,compared to the non-enhanced images (p < 0.002). I.a. Gd-DTPA provided sig-nificantly higher SNR and CNR compared to i.v. Gd-DTPA (p < 0.01). Further-more, a better delineation of the cartilage in the synovial/cartilage zone and ofthe chondral/subchondral border was observed.Conclusion: The dGEMRIC MR imaging method markedly improved delineationof hip joint cartilage, compared to non-enhanced MR imaging. The i.a. Gd-DTPAprovided the best cartilage delineation. The dGEMRIC method is a clinically ap-plicable MR imaging method, which may improve identification of early subtlecartilage damage and the accuracy of volume measurements of hip joint carti-lage.

B-447 11:24

Diffusion imaging in the femoral head for cartilage assessment: Correlationwith macroscopic and histological findingsT.C. Mamisch1, G. Muhr1, G.H. Welsch2, W. Horger2, M. Deimling2, W. Bautz2,A. Cavallaro2, B. v. Rechenberg3, M.I. Menzel2; 1Essen/DE, 2Erlangen/DE,3Zurich/CH ([email protected])

Purpose: The aim of the study was to asses the diagnostic quality of steady-state diffusion weighted imaging (SS-DWI) in the hip by comparison with intraop-erative findings and histology of the femoral head in patients with osteoarthritis(OA).Methods and Materials: Five patients with OA of the hip underwent in vivo 3 TMR imaging with cartilage sensitive isotropic 3D-DESS (TR=15.6, TE=4.5, reso-lution 0.6 mm3) and a diffusion weighted 3D-PSIF sequence (diffusion moment45[mT/m*ms], 75[mT/m*ms]) in the contact pressure area of the affected hip. Formacroscopic classification the resected femoral head was divided in differentregions and analysed, then pin-marked and imaged with an isotropic MR imag-ing protocol. Afterwards the femoral head was fixed and semi thick slices of 0.3 mmwith toluidin blue staining were histopathologically registered by the pin locationsto the MR data.Results: We found correlation of macroscopic changes to MR imaging in 3D-DESS and SS-DWI for normal, moderate and severe grades of OA. For histolog-ical changes in the articular cartilage we found specified alterations in signal

intensity for 3D-DESS and the trace weighted diffusion. For the 3D-DESS wefound correlation with the amount of proteoglycans represented by the toluidinblue staining, but a high percentage of false negative for mild changes. The SS-DWI did not correlate directly with the staining and the cartilage thickness, butshowed significantly better results for matrix structure changes in mild grades ofOA.Conclusion: Our results show the potential for SS-DWI in the detection of struc-tural defects in articular cartilage. It was possible to correlate macroscopic andhistological findings to SS-DWI and use a not quantitative approach for validationof the method. Especially in clinically important, mild grades of OA, SS-DWI couldbe beneficial.

B-448 11:33

Parameter selection in MR imaging of cartilage repair with use of a collagenmatrix (AMIC)J. Gellissen, N.O. Wendler, B.M. Stoeckelhuber, M. Schelzel, P. Behrens,T. Helmberger; Lübeck/DE ([email protected])

Purpose: To compare the performance of different MR pulse sequences for high-contrast and high-spatial-resolution imaging of collagen matrices.Methods and Materials: Chondral resection 2 cm² in size was carried out re-sembling a grade- IV defect of a porcine femoral condyle. A bilayer matrix con-sisting of collagen I/III (Chondro-Gide®, Geistlich Biomaterials, Switzerland) wasdissected and fixed with fibrin glue. MR-imaging was conducted using a 1.5 Tdevice (Siemens, Germany). Twenty-seven pulse sequences were varied withregard to their main properties (T1- and T2-weighted spin-echo or gradient echo),mode of volumetric data acquisition (2D, 3D), image matrix, slice thickness (1.3-2 mm) and the choice of fat saturation techniques. Signal intensities were meas-ured in the subchondral layer, collagen matrix, adjacent cartilage and fluid.Contrast-ratios were computed.Results: Contrast-ratios of the matrix vs. adjacent fluid/subchondral layer werecalculated with values 0.05-0.88 ± 0.04 and 0.03-0.99 ± 0.05 respectively. Whilehighest contrast was achieved using PD-weighted TSE-sequences with low echo-time, high image matrix and thin sections without fat saturation, T1-weighted IR-sequences yielded an intermediate contrast-ratio with optimal depiction of thedefect depth. Five sequences were unable to differentiate between subchondrallayer and matrix.Conclusion: Appropriate MR imaging parameter selection is crucial for optimizedvisualization of the matrix and has the potential to provide information about matrixintegrity and success of cartilage repair.

B-449 11:42

Evaluation of the MR signal modifications induced by hyaluronic acidtherapy in chondromalacia patellae: A preliminary studyN. Magarelli1, A. Ierardi1, P.A. Mattei2, A. Leone1, G. Guglielmi3, L. Bonomo1;1Roma/IT, 2Chieti/IT, 3San Giovanni Rotondo/IT ([email protected])

Purpose: To evaluate the MR signal modifications induced by one cycle of hy-aluronic acid therapy in young adult patients positive for intermediate grade chon-dromalacia patellae.Methods and Materials: Sixteen consecutive patients within the age range of18-35 years old with mono or bilateral anterior gonalgia but without trauma anddiagnosed as having early-intermediate grade chondromalacia patellae (grade I-II-III according to the Outerbridge-Insall classification) were enrolled. All MR ex-aminations were performed with the same instrument, a 1.5 Tesla (Philips-INTERA)and a dedicated coil (150 mm FOV phased array (SYN-flex-M)). Images wereacquired on the axial oblique plane perpendicular to the patella using a turbo-spin-echo proton density (TSE-PD)SPIR. Each patient underwent one cycle oftreatment with hyaluronic acid which consisted of five weekly intra-articular injec-tions of 20 mg/2 ml. MR examinations were performed prior to initiating treat-ment, immediately after completion of the treatment cycle, three and six monthspost-treatment. Modification of the lesion was assessed by placing a region-of-interest (ROI) of 2-3 mm2 within the lesion. Variations were assessed using ANOVAanalysis.Results: Fifteen of the sixteen patients (5 men and 11 women; 4 with grade I and12 with grades II-III) treated reported a remission of pain and return to normalactivity. Initial results indicate that in intermediate grade (II-III) lesions the signalintensity increased immediately after and decrease three-six months after treat-ment. Significant modifications were not observed in patients with closed chon-dromalacia.Conclusions: Initial results indicate that MR imaging is an useful instrument forevaluating modifications induced by hyaluronic acid therapy.

240 C D E FBA

Scientific Sessions

G

B-450 11:51

Articular cartilage lesions of the glenohumeral joint: Diagnosticeffectiveness of multidetector spiral CT arthrography and correlation witharthroscopyF. Lecouvet1, B. Dorzee1, J. Jamart2, B. Vande Berg1, J. Malghem1,J.-E. Dubuc1; 1Brussels/BE, 2Yvoir/BE ([email protected])

Purpose: To assess the diagnostic effectiveness of multidetector spiral CT ar-thrography (MDSCTA) in detecting cartilage abnormalities in the gleno-humeraljoint.Methods and Materials: MDSCTA images obtained in 22 consecutive patients(mean age 50 years, range 23-74 years, 12 female and 10 male) were evaluatedfor glenohumeral cartilage lesions. Two musculoskeletal radiologists who wereblinded to the arthroscopy report independently analysed the humeral head andglenoid cartilage (9 areas for each surface). The sensitivity and specificity ofMDSCTA for grade 2 (substance loss < 50 % of cartilage thickness) or higher,and grade 3 or higher cartilage lesion (substance loss > 50 % of cartilage thick-ness), the Spearman correlation coefficient between MDSCTA and arthroscopicgrading, and K statistics for assessing intra- and interobserver reproducibilitywere determined.Results: Sensitivities and specificities of MDSCTA ranged between 90 and 95 %for the detection of grade 2 or higher cartilage lesions and ranged between 80and 98 % for the detection of grade 3 or higher lesions. There was moderate tosubstantial agreement between grading at arthroscopic examination and gradingat MDSCTA (Spearman correlation coefficients from 0.532 to 0.651). Interobserveragreement was substantial for detecting grade 2 or higher, and almost perfect fordetecting grade 3 or higher cartilage lesions (K of 0.618 and 0.876). Intraobserv-er agreement was almost perfect for detecting grade 2 or higher and grade 3 orhigher lesions (K of 0.845 and 0.851).Conclusion: MDSCTA is an accurate tool for the study of the articular cartilagesurface in the shoulder joint.

10:30 - 12:00 Room C


SS 1001Liver metastasesModerators:E.J. Rummeny; Munich/DEG. Tsivtsivadze; Tbilisi/GE

B-451 10:30

Liver regeneration: Long term results of a qualitative and quantitativecrossectional evaluation for liver surgery, percutaneous ablation (LITT) andloco regional chemotherapy (TACE)T. Lehnert, A. Gazis, A. Thalhammer, M.G. Mack, J.O. Balzer, T.J. Vogl;Frankfurt/Main/DE ([email protected])

Purpose: To evaluate and compare the complexity of liver regeneration post liverresection, MR-guided laser induced thermotherapy (LITT) and transarterial chem-oembolization (TACE) in correlation with the regeneration factors TNF-a and HGF.Methods and Materials: 81 patients were evaluated, 31 patients post surgicalliver intervention, 32 patients post ablation (LITT), and 18 patients post localregional chemotherapy (TACE). During a period of 12 months the effects of thetherapy on normal liver parenchyma and possible liver regeneration after surgi-cal and interventional radiological procedure were evaluated using unenhancedand contrast enhanced MR imaging and CT. Volumetric examination (CT-volume-try) of non-involved liver parenchyma and tumor/metastasis pre and post inter-vention was examined.Results: A significant influence on the HGF process could be confirmed on lossof liver parenchyma resulting form surgery or LITT induction. Pre-interventionallow HGF-values correlate with increased liver regeneration. Patients younger than55 years show stronger dynamics than 55+ year old patients. After 6 months,original liver volume can be observed in 58.1 % (n = 18) of all patients post sur-gical intervention, and in 65.6 % (n = 21) of LITT patients. Patients post chem-oembolization presented an increase in liver volume of 13.1 % (3.2 % - 17.8 %)after 6 months by reduction of tumor volume.Conclusion: Interventional radiologic therapy methods such as MR-guided laserinduced thermotherapy (LITT) and chemoembolization (TACE) basically stimu-late liver regeneration and cause a significant increase of cytokine HGF, on aver-age 1.7 weeks post intervention. No variation of TNF-a value can be observed.

B-452 10:39

Detection of hepatic metastases from gastrointestinal tumors: Comparisonof US vs low MI real time contrast-enhanced USV. Cantisani, P. Ricci, R. Iannaccone, F. Arduini, F.M. Drudi, R. Passariello;Rome/IT ([email protected])

Purpose: To compare low mechanical index (MI) real time contrast-enhancedUS (CEUS) with conventional B-mode US (US) for the detection of hepatic me-tastases from gastrointestinal tumors.Methods and Materials: From February to September 2005, 88 patients (48males, 40 females; mean age 62 years; range 39-78) with known hepatic lesionsfrom gastrointestinal tumors were evaluated. All patients underwent conventionalUS followed by CEUS after administration of SonoVue (Bracco, Milan, Italy).Number, location and size of metastases on baseline and CEUS were comparedwith contrast-enhanced 64-row spiral CT (Siemens Cardiac Sensation. Erlan-gen, Germany). US and CT images were assessed by two blinded readers.Results: 190 metastases were identified by 64-spiral CT in 88 patients, with 90lesions lesser than 1 cm. US identified 114 lesions and contrast-enhanced US180. The addition of CEUS improved sensitivity for the detection of individualmetastases from 60% to 95% (P <.001). Contrast-enhanced US was significantlymore accurate than baseline US, especially for small metastases (p < 0.001) withno statistical difference if compared with 64-row spiral CT. Contrast-enhancedUS showed more metastases than CT in seven patients, and CT showed morelesions than contrast-enhanced US in two of 22 patients with available independ-ent reference (magnetic resonance imaging, intraoperative US, or pathologic find-ings).Conclusion: Detection of hepatic metastases is significantly improved by CEUS.CEUS seems to be a reliable tool, alternative to spiral CT in the detection ofhepatic metastases. In the evaluation of patients with suspected hepatic metas-tases, US examination must be performed after contrast administration.

B-453 10:48

Value of variable IV bolus injection-to-scan delay determined by automaticmonitoring liver parenchyma enhancement in oncological patients using16-row and 64-row MDCTG. Zamboni, V. Raptopoulos, J.B. Kruskal, B. Siewert, E. Zeikus; Boston,MA/US ([email protected])

Purpose: To assess the value of triggering abdominal and torso CT scanningbased on liver parenchyma enhancement in patients with potential hypodenseliver lesions.Methods and Materials: 58 oncological patients underwent CT torso on 16-rowand 64-row multidetector (MDCT) scanners. 21 patients were scanned with afixed 60 second delay for the abdomen, while in 37 patients the abdominal scanwas initiated when liver parenchyma enhancement reached 50 HU. In each pa-tient attenuation values from liver, portal and hepatic veins were obtained fromboth the non-contrast and contrast-enhanced scans, as for 37 hypodense, pre-sumed metastatic, liver lesions.Results: Mean delay for scanning in the triggered group was 56.9 seconds (range43-70). Mean post contrast liver attenuation was 127.6 ± 21.3 HU (range 93.2-155.3) in the triggered and 101.7 ± 24 HU (range 63.6-132.8) in the non-trig-gered group. Mean parenchyma enhancement was 26 HU greater in the triggeredgroup (p < 0.0001). The mean portal and hepatic vein enhancement was 31.5 HUand 30 HU greater in the triggered group (p = 0.0026 and p = 0.017, respective-ly). The mean value of hypodense lesions enhancement was 8.7 HU greater inthe triggered group (p = 0.3; not significant) with a higher conspicuity (normalliver/lesion HU) in the triggered group (mean ratio 3.8) than in the non-triggeredgroup (mean ratio 2.8), with a significant difference (p = 0.0346).Conclusion: Liver triggering is a simple technique which produces consistentand optimal liver parenchymal and vascular enhancement in oncological patientsindependent of their cardiovascular status while increasing the conspicuity ofhypodense liver metastases.

B-454 10:57

Assessment of size of colorectal liver metastases at quadruple-phasecontrast-enhanced MDCT: Which phase is most accurate to measure truesize of liver metastases?H. Langenberger, T. Gruenberger, C. Plank, F. Laengle, A. Ba-Ssalamah,M. Schindl, W. Schima; Vienna/AT ([email protected])

Purpose: Assessment of tumor response in oncologic therapy is largely basedon CT measurements of tumor size. Purpose of this study was to investigate (1)if the size of hepatic metastases at different CT phases are significantly different

B 241C D E FA

Su

nd

aS

un

da

Su

nd

aS

un

da

Su

nd

ayy yyy

Scientific Sessions

G

from macroscopic measurements of the actual specimen and (2) which CT phaseis most accurate to assess the true size of liver metastases, as measured in thesurgical specimen.Methods and Materials: 15 patients with a total of 24 hepatic metastases fromcolorectal carcinoma underwent 16-row MDCT before and after 150 ml IV con-trast agent administration in the arterial, venous, and 5 minutes delayed phase.Following surgery within 5 days the metastases in surgical specimens were dig-itally photographed and measured on magnified images.Results: The mean metastasis size was 3.4 cm as measured on the specimens.Four lesions were not identified on CT. Mean size of metastases was 3.2 cm inthe unenhanced, arterial, and venous phase and 3.3 cm in the delayed phase.Mean difference in size measurement between CT phase and specimen was28%, 20%, 17%, and 22% for the unenhanced, arterial, venous, and delayedphase, respectively.Conclusion: Neither unenhanced nor contrast-enhanced CT are very accurateto measure the size of liver metastases in comparison to the surgical specimen.CT measurement obtained in the venous phase were most accurate. Due to thehigh variability in size measurements between different CT phases progressionor regression of metastatic disease should only be judged by comparing meta-static lesions in the same CT phase and using the same windowing.

B-455 11:06

Detection of hepatic colorectal metastases: Effect of different sectionthicknesses using 64-slice CTD. Marin, R. Iannaccone, C. Catalano, A. Guerrisi, M. Di Martino,G. De Filippis, M. Baski, R. Passariello; Rome/IT ([email protected])

Purpose: To evaluate diagnostic impact of different section thicknesses using64-slice CT to detect hepatic colorectal metastases.Methods and Materials: 43 patients with known colorectal metastases wereexamined with a 64-slice scanner (Sensation 64, Siemens). CT parameters were:collimation 0.6 x 64 mm, mAs, 250, and kVp, 120. All patients received non-ioniccontrast medium (400 mgI/mL; Iomeron 400, Bracco) at a rate of 5 mL/sec. Atriple-phase protocol was acquired with hepatic arterial (delay time determinedwith bolus tracking technique), portal-venous (60 sec), and delayed (180 sec)phases. CT datasets were reconstructed with four different section thicknesses-0.6, 1, 3, and 5 mm. Each dataset was interpreted separately by three observersin four reading sessions. Sensitivity, positive predictive value (PPV), and areaunder the receiver operating characteristic curve (A (z)) were calculated.Results: Mean sensitivity and PPV were 79% and 77% on 0.6-mm images, 78%and 77% on 1-mm images, 74% and 78% on 3-mm images, and 72% and 80%on 5-mm images, respectively. No significant difference in sensitivity and PPVamong images was detected. Mean A (z) values were 0.79, 0.78, 0.77, and 0.75for 0.6, 1, 3, and 5 mm images, respectively (no significant difference). With re-gard to small (≤ 1 cm) metastases, mean sensitivities of 0.6 (72%), 1 (71%), and3 mm (71%) images were statistically significantly superior to that of 5 mm (62%)images (p < 0.05).Conclusion: When using 64-slice CT, a section thickness ≤ 3 mm should be cho-sen to maximize detection of small (≤ 1 cm) hepatic colorectal metastases.

B-456 11:15

Oral manganese as contrast medium in detecting liver metastases with MR-imaging at 1.5 and 3 TH.M. Dekker1, C. van Herpen1, Y. Hoogeveen1, J.G. Blickman1, H. Thomsen2,T. Ruers1, J.O. Barentsz1; 1Nijmegen/NL, 2Copenhagen/DK([email protected])

Purpose: Evaluation of the diagnostic performance of oral manganese as a newcontrast medium in liver MR-imaging in patients with liver metastases.Methods and Materials: 18 patients with known liver metastases were exam-ined with MR imaging at 1.5 T and 3 T before and 3 hours after oral administra-tion of Mn-contrast diluted in 400 ml water. MR imaging included T1-w FLASHbreathold sequences in coronal and in transversal planes. At 1.5 T contiguous5 mm slices and at 3 T 3 mm were made. Additionally, a T2w true-FISP sequencewas performed to recognize liver cysts and haemangiomas. Contrast betweenliver tissue and metastases was determined on the pre- and post Mn-contrastscans. The homogeneity of liver enhancement was evaluated. In addition, thenumber of detected liver metastasis was evaluated.Results: There were no side-effects after the intake of oral Mn-contrast. Themean liver-metastases contrast improved at 1.5 and 3 T respectively with a fac-tor 2.1 and 1.5. Higher liver-metastases contrast increased the number of liverdetected metastasis with more than 50% at both 1.5 and 3 T. In patients with ahistory of chemotherapy, liver enhancement was inhomogeneous, probably due

to disturbance of the portal circulation, nonetheless this did not influence theimproved metastases detection.Conclusion: This pilot study shows oral Mn-contrast to be a simple and promis-ing contrast agent which results in improved visualization of liver metastases byselective increase of liver signal.

B-457 11:24

MR imaging finding of focal hepatic eosinophilic infiltrations in patientswith gastric cancerJ. Sun, J. Kim, J. Lee, M. Kim, H. Noh; Suwon/KR ([email protected])

Purpose: The purpose of this study is to evaluate the MR imaging findings ofhepatic eosinophilic infiltration in patients with gastric cancer.Methods and Materials: Ten gastric cancer patients who had focal hepatic eosi-nophilic infiltration were included in this study (pathologic confirmation in 9 cas-es, clinical confirmation in 1 case). There is no any evidence of cause provokingperipheral blood eosinophilia in these patients. Seven patients of gastric cancerwere undergoing staging workup and 3 gastric cancer patients undergoing post-operative follow-up. We evaluated the size, number and distribution of these he-patic lesions on MR scan. We also evaluated the signal intensity of focal lesionsin T1-weighted, T2- weighted images and the pattern of enhancement in a dy-namic contrast study.Results: A total 22 focal hepatic lesions were observed. The mean size of hepat-ic lesions is 1.2 cm (0.86-1.9 cm). Nineteen lesions (86%) showed subcapsulardistribution. The lesions were isointense (50%) or hypointense (50%) on T1-weight-ed images and isointense (19%) or hyperintense (81%) on T2-weighted images.The arterial phase of contrast study revealed 11 (50%) hyperintense lesions, 10isointense lesions, and 1 hypointense lesion. During the portal and delayed phasescans, 18 (82%) and 21 (95%) lesions were hyperintense with fuzzy margins,respectively.Conclusion: The focal eosinophilic infiltrations showed homogenous enhance-ment with fuzzy margin in the portal and delayed phases in the dynamic contrastMR study. These findings should help to distinguish focal eosinophilic infiltrationfrom metastasis in patients with gastric cancer patient undergoing pre-operativestaging or post-operative flow up.

B-458 11:33

Detection of solid liver lesions: Quantitative and qualitative comparison ofbreath-hold and navigator sequencesC. Hillerer, J. Gaa, K. Woertler, M. Dobritz, E.J. Rummeny; Munich/DE([email protected])

Purpose: Comparison of seven breath-hold sequences (BHSE) with three navi-gator sequences (NSE) for detection of hepatic lesions.Methods and Materials: Thirty patients were examinated with a 1.5 T scanner,seven BHSE (T1-GRE, T2-TSE, HASTE, all with and without fat suppression[FS], and STIR) and three NSE (T2-TSE PACE with and without FS, STIR PACE)were obtained. Images were analysed quantitatively (lesion-liver contrast-to-noiseratio [CNR], liver-spleen CNR and liver signal-to-noise ratio [SNR]) and qualita-tively (detectability of hepatic structures and lesions, presence of artefacts).Results: Highest liver SNR was found using T1-GRE sequences. Best liver-spleenCNR was obtained with T2-TSE-FS and with both HASTE-sequences, best le-sion-liver CNR with T2-TSE-FS. There was no significant difference between BHSEand NSE in lesion-liver CNR. Comparing T2 weighted BHSE with the correspond-ing NSE, BHSE showed a better liver-spleen CNR and liver SNR. However, in thequalitative analysis T2-TSE PACE displayed better lesion clarity especially forsmall liver lesions, probably due to a higher acquisition matrix. Qualitative com-parison of all sequences showed an advantage of HASTE, followed by the T1-GRE sequences.Conclusion: Although no significant quantitative difference in lesion-liver CNRwas found, qualitatively the HASTE, the T1-GRE and the T2-PACE sequencesshowed the best results. Particularly in detection of lesions smaller than 1 cm,T2-PACE can be advantageous. Thus, for clinical practice, BHSE should be pre-ferred due to reduced acquisition time and inferior artefact susceptibility. NSEmay be helpful examinating respiratory insufficient patients who cannot sustainsufficiently breathing.

242 C D E FBA

Scientific Sessions

G

B-459 11:42

Contrast-enhanced-FDG-PET/CT vs CT vs SPIO (superparamagnetic ironoxide)-MR imaging in patients with metastases from colorectal cancer: Aprospective blinded study with intraoperative and histological confirmationE.D. Rappeport, A. Loft, A.K. Berthelsen, P. von der Recke,P. Noergaard Larsen, A. Mellon Mogensen, A. Wettergren, A. Rasmussen,C. Thomsen; Copenhagen/DK ([email protected])

Purpose: Compare contrast-enhanced-FDG-PET/CT with PET, CT and SPIO-MRI in patients with liver-metastases (LM) from colorectal-cancer (CRC) usingintraoperative findings as gold-standard.Methods and Materials: 36 candidates for liver surgery had PET/CT with a bi-phasic contrast-enhanced CT-protocol and SPIO-MRI. 31 patients finally hadsurgery. Prospective blinded interpretation of SPIO-MRI, PET/CT, CT-part andPET-part of the PET/CT-study was performed. Liver-lesions and suspected extra-hepatic tumor were registered. For LM surgical findings including intraoperativeultrasound and histology were the gold standard.Results: 71 LM were identified at surgery. Lesion-sensitivity for detection of LMwas 54 %, 66 %, 82 % and 89 % for PET, PET/CT, SPIO-MRI and CT respective-ly. CT alone was significantly (p < 0.05; McNemar-test) more sensitive than PETand PET/CT. Particularly LM < 1 cm were PET-negative. Four patients had chem-otherapy within one month prior to PET/CT; with these patients excluded sensi-tivity was 61 %, 75 %, 84 % and 86 % for PET, PET/CT, SPIO-MRI and CTrespectively. Twelve patients had extra-hepatic disease. PET/CT and CT werepositive in ten and seven of these patients respectively. In three patients CT wasfalse-positive for extra-hepatic tumor and PET/CT in one patient.Conclusion: CT alone and SPIO-MRI has higher sensitivity than PET and PET/CT in detection of individual LM especially for smaller LM. PET and PET/CT candetect some patients with extrahepatic tumor not detected by conventional imag-ing.

B-460 11:51

Utility of diffusion weighted MR in patients undergoing loco-regionaltreatment for focal liver lesions: An ongoing studyS.K. Venkatesh, G.B.H. Lau, F. Leung, S.-c. Wang, C. Au; Singapore/SG([email protected])

Purpose: To describe the utility of diffusion weighted MR imaging in patientsundergoing loco-regional treatment of liver tumours.Methods and Materials: Diffusion-weighted MR imaging (DWI) was performedbefore and after transarterial chemoembolization (TACE) (n = 22) or radio-fre-quency ablation (RFA) (n = 18) for hepatocellular carcinomas (n = 28) and meta-static lesions of liver (n = 12). DWI was performed with single shot echo planarsequence and unidirectional gradient sequences with two b values (200, 500)and apparent diffusion coefficient (ADC) maps generated. The lesion to liver sig-nal intensity ratio and ADC values were calculated. Routine T2-W and T1-W se-quences were also obtained. One month follow-up DWI were also performed for6 lesions.Results: All hepatic tumours were hyperintense to liver on DWI before TACE andRFA (lesion: liver = 2.2 ± 0.43). Post TACE and RFA, the lesions demonstrated ageneral decrease in signal intensity (1.5 ± 0.32) on DWI. RFA lesions were eitherhypointense with a variable hyperintense rim or mixed-intensity. The post-TACEand RFA lesions demonstrated higher mean ADC values (190 ± 30; 155 ± 25) ascompared to the tumours prior to treatment (160 ± 20; 120 ± 22) (p < 0.02). TheADC values increased after 1-month follow-up suggestive of further necrotic chang-es. In three patients recurrence was noted which appeared hyperintense to thetreated lesion with low ADC.Conclusion: There is a trend for ADC of the liver lesions to increase followingloco-regional treatment signifying necrotic changes. DWI with ADC maps may beuseful for monitoring response to treatment and detect recurrence.

10:30 - 12:00 Room F2

Cardiac

SS 1003Evaluation of cardiac functionModerators:G. Napoli; Bologna/ITS. Soimakallio; Tampere/FI

B-461 10:30

Multidetector row computed tomography and magnetic resonance imagingcan accurately estimate aortic valve area as compared to transthoracic andtransesophageal echocardiographyA.-C. Pouleur, J.-B. le Polain de Waroux, A. Pasquet, E. Coche,J.-L. Vanoverschelde, B. Gerber; Brussels/BE ([email protected])

Background: Assessment of aortic stenosis severity usually involves the meas-urement of aortic valve area (AVA) by use of transthoracic echocardiography(TTE). The aim of this study was to assess the feasibility and accuracy of meas-uring AVA with MDCT and MR imaging as compared to TTE and transesopha-geal echocardiography (TEE).Methods and Materials: We studied 30 pts (18 men, 59 ± 14 yrs) who under-went MDCT, MR imaging, TTE and TEE. Contrast enhanced MDCT images wereacquired using a 40-slice system (Philips Brillance 40) and reconstructed intoserial short-axis slices through the aortic valve using retrospective ECG gating ateach 10% of the cardiac cycle. MR imaging was performed on a 1.5 T magnet(Intera CV, Philips). AVA was measured by planimetry (short-axis cine MR imag-ing and MDCT) and compared to AVA calculated by the continuity equation (TTE)and by planimetry (TEE).Results: AVAs measured by MDCT and MR imaging were similar to those plan-imetred by TEE (3.3 ± 1.7 cm2 by MDCT, 3.2 ± 1.7 cm² by MR imaging vs3.2 ± 1.7 cm² by TEE, p=ns). Yet, measurements by MDCT and MR imaging werelarger than AVAs calculated by TTE with continuity equation (2.9 ± 1.5 cm2,p < .001). Planimetric AVAs measured by MDCT and MR imaging correlatedstrongly with those measured by TTE (r = 0.96 and r = 0.97 respectively, p < .001)and TEE (r = 0.98 and r = 0.97 respectively, p < .001). MDCT and MR imagingalso allowed identified all 7 bicuspid valves, 3 of which had been missed by TTE.Conclusion: MDCT and MR imaging accurately estimate AVA. This suggeststhat these techniques might be clinically useful in the non-invasive assessmentof aortic valve disease.

B-462 10:39 !Diagnostic accuracy of 16-multidetector computed tomography to identifypatients with aortic stenosis and aortic regurgitationG.M. Feuchtner, W. Dichtl, G.J. Friedrich, T. Schachner, A. Mallouhi, S. Mueller,J. Bonatti, O. Pachinger, D. zur Nedden; Innsbruck/AT([email protected])

Purpose: To evaluate whether 16-multidetector computed tomography (16-MDCT)accurately identifies [1] patients with asymptomatic aortic stenosis (AS) basedon the quantification of the aortic valve area (AVA), [2] patients with aortic regur-gitation (AR) in comparison to transthoracic echocardiography (TTE), and [3] toassess the predictive value of 16-MDCT aortic valve calcium (AVC) score.Methods and Materials: [1] 46 patients and [2] 72 consecutive patients whounderwent 16-MDCT coronary angiography (Sensation 16™, Siemens)(16x0.75 mm; 0.42 s; 120 kV, 500 mAs; 100 ml Iodixanol (Visipaque™, Amer-sham); flow 3-4.5 ml/s; inc.0.6, eff.sl.1 mm; retrospective ECG-gating at mid-to-late systole[1] and mid-late diastole[2]) were examined prospectively. TTEmeasurements included [1] Doppler VTI for calculation of the AVA and [2]semi-quantitative Doppler regurgitation jet analysis (severity:grade0-3+). [3] AVCs werequantified with 16-MDCT (16x1.5 mm, 0.5 s; inc.3; eff.sl.3 mm; 120 kV; 130 mAs)and compared with other risk factors (e.g. AVA, NT-proBNP, CRP) in 25 patients.Results: [1] Sensitivity of 16-MDCT to identify patients with AS was 100% (30/30), specificity 98% (15/16). Quantification of the AVA with 16-MDCT(0.90 cm² ± 0.24) showed a good correlation to TTE (r = 0.86; p < 0.001), inter-observer variability was 4.6%. [2] 16-MDCT correctly detected 39/48 patientswith AR (sensitivity 81%, specificity 91%, NPV 70 %, PPV 95 %). Sensitivity of16-MDCT to identify moderate/severe AR (≥ grade1.5+) was 95%(20/21), specif-icity 96%, PPV 90%, NPV 97%. Sensitivity of 16-MDCT to detect patients withmild AR (≤ grade1+) was 70%, specificity 92%(19/27). [3] Within a 1-year-follow-up,6 of 25 patients developed a major adverse clinical event (MACE). Agatston AVC

B 243C D E FA

Su

nd

aS

un

da

Su

nd

aS

un

da

Su

nd

ayy yyy

Scientific Sessions

G

score was the strongest predictor of MACE (p < 0.01) among all risk factors(6377 ± 2282vs. 2323 ± 983).Conclusion: [1] 16-MDCT accurately identifies patients with asymptomatic aor-tic stenosis and patients with moderate/severe aortic regurgitation. Patients whoundergo coronary 16-MDCT angiography can be evaluated whether concomitantaortic stenosis/regurgitation is present. [2] AVCs were the strongest predictor ofMACE.

B-463 10:48

Assessment of ventricular volume determination on MDCT and EBTM.J.W. Greuter, J. Overbosch, P. Van der Vleuten, P.M.A. Van Ooijen,M. Oudkerk; Groningen/NL ([email protected])

Purpose: Assessment of the variation of ventricular volume measurements in amoving heart phantom on a 64-detector MDCT and EBT-system compared toabsolute values.Methods and Materials: A respiration device operated moving heart phantomwas used on a 64-detector MDCT and a 100 ms EBT unit. The heart rate (HR)was varied between 50 and 100 bpm. End diastole and end systole ventricularvolume of the heart phantom were determined by two independent readers usingCardIQ (GE) with data of both modalities, from which ejection fraction was de-rived as EF=(EDV-ESV)/EDV. Stroke volume (SV=EDV-ESV) of the heart wasdetermined with a gas-flow analyzer and EDV was calibrated by filling the phan-tom with water.Results: An EDV = 258 ± 1 ml was obtained from the water calibration. From thegas-flow analysis, a decrease of SV=76 ml at 50 bpm to 60 ml at 100 bpm wasobtained, corresponding to a decrease of EF from 30%-23% (∆=7%). From theMDCT-data a decrease in EF was obtained from 20%-15% (∆=5%) at 50 and100 bpm, respectively. From the EBT-data a decrease in EF was obtained from23%-17% (∆=6%) at 50 and 100 bpm, respectively. The average underestimationof EF using CardIQ with MDCT-data yields 39 ± 9% and with EBT-data 27 ± 5%.Conclusion: EBT is more sensitive than MDCT for decrease in EF with increaseof HR. Determination of ventricular volume in the heart phantom, as derived fromMDCT and EBT, shows a significant underestimation at HR between 50 and100 bpm, for EF between 23-30%. EBT more accurately determines the absoluteventricular volume as compared to MDCT.

B-464 10:57

64-slice computed tomography assessment of left ventricular volumetricand functional parametersG. Runza1, F. Cademartiri2, A. Palumbo2, L. La Grutta2, T. Baks2, N.R. Mollet2,W.B. Meijboom2, M. Midiri1, G.P. Krestin2; 1Palermo/IT, 2Rotterdam/NL([email protected])

Purpose: The aim of this study was to assess left ventricular (LV) volumetric andfunctional parameters by using 64-slice Computed Tomography (CT) comparedwith cardiac-Magnetic Resonance (MR).Methods and Materials: Thirty-four patients (30 male; mean age 59.6± 10.4 years) with a previous acute coronary syndrome and scheduled for leftventricular function assessment with MR were retrospectively enrolled to per-form within 48h a 64-slice CT (Sensation 64, Siemens, Germany) by using thestandard protocol for coronary angiography. Only patients in sinus rhythm, spon-taneous HR < 75 bpm, and able to breath-hold for 15 sec, were included. Pa-tients with contra-indications to intravenous iodinated contrast material wereexcluded. A dedicate semi-automated software (Argus, Siemens, Germany) wasused to assess end-diastolic, end-systolic LV volumes (LVEDV, LVESV) LV strokevolume, and ejection fraction (LVSV, LVEF) by using LV short axis images. Theresults were compared with those of cardiac MR using a standard protocol.Results: No patient was excluded from the study. LVEDV was mean (± SD)144.16 mL (± 37.79) and LVESV mean (± SD) 55.94 mL (± 21.78) as determinedwith CT correlated well with MR results 163.83 mL ± 48.14) [r = 0.93; p < 0.05]and 56.96 mL ± 25.49) [r = 0.85; p > 0.05], respectively). LVEF (61.13% ± 9.93for CT and 66.32% ± 6.35 for MR) and LVSV (88.22 mL ± 28.82 for CT and106.87 mL ± 25.6 for MR) also showed good correlation, r = 0.76 and r = 0.77,respectively (p < 0.001). A statistical analysis paired T-test was performed and asignificant underestimation was found when compared the overall CT measure-ments for LV to those obtained with MR (p < 0.001).Conclusion: 64-slice CT shows a good correlation with cardiac MR regardingthe assessment of LV volumes and functional parameters, however it significant-ly underestimates both parameters.

B-465 11:06

Left ventricular blood flow and coronary flow reserve with MR imaging inchronic heart failureA. Aras, Y. Anik, A. Demirci, C. Balci, G. Kozdag, D. Ural, B. Komsuoglu;Kocaeli/TR ([email protected])

Purpose: To quantify and compare global left ventricular (LV) perfusion and cor-onary flow reserve (CFR) in patients with chronic heart failure and healthy con-trol group by measuring coronary sinus flow with velocity-encoded cine (VEC)magnetic resonance (MR) imaging.Methods and Materials: MR measurements were performed in 20 consecutivepatients with chronic heart failure due to coronary artery disease and in the con-trol group (n = 11). Global LV perfusion was quantified by measuring coronarysinus flow in an oblique imaging plane perpendicular to the coronary sinus withnon-breath hold VEC MR imaging. LV mass was measured by means of cineimaging that encompassed the heart. CFR was measured at rest and after infu-sion of dipyridamole.Results: At rest, LV perfusion was not significantly different in patients with chronicheart failure and control group (0.85 ± 0.3 ml/min/gr vs 0.83 ± 0.26 ml/min/gr,p < 0.001). After administration of dipyridamole, LV perfusion and CFR were sig-nificantly decreased in patients with chronic heart failure compared with that ofthe control group (LV perfusion 1.19 ± 0.4 vs 2.39 ± 0.69; p < 0.001, and CFR1.45 ± 0.32 vs 3.00 ± 0.99; p < 0.001).Conclusion: Combined cine and VEC MR imaging revealed that patients withchronic heart failure have normal LV perfusion at rest but severely depressed LVperfusion after vasodilatation. Impaired CFR may contribute to progressive de-cline in LV function in patients with chronic heart failure.

B-466 11:15

Dual breath-hold evaluation of cardiac function with high temporal andspatial resolution: Accuracy of accelerated multi-slice CINE MR imagingB.J. Wintersperger1, S.B. Reeder2, O. Dietrich1, A. Greiser3, D. Theisen1,M.F. Reiser1, S.O. Schönberg1; 1Munich/DE, 2Madison, WI/US, 3Erlangen/DE([email protected])

Purpose: To evaluate the accuracy of a 4-fold accelerated multi-slice CINE SSFPtechnique compared with standard single-slice CINE SSFP.Methods and Materials: 40 individuals (24 patients, 16 volunteers) underwentCINE imaging on a 32-channel 1.5 T scanner (MAGNETOM Avanto) equippedwith multi-element coils. All data was acquired in short axis orientation usingtriggered SSFP (2.8/1.4; 60°) technique with a spatial resolution of 2.5x1.9 mm2

and a temporal resolution of 48 ms. Multi-slice imaging was accelerated usingTSENSE (R=4) acquiring 2 stacks of 5 slices each for ventricular coverage. Sin-gle-slice non-accelerated CINE SSFP (nonTSENSE) was performed at identicallocations in consecutive breath-holds. Image data was analyzed regarding signalparameters and overall scan time for ventricular coverage. LV volumetric param-eters (EF, EDV, ESV) and intraobserver variability were calculated for all datasets.Results: 4-fold acceleration comes along with an estimated loss of CNR of ~74%(P < 0.0001). Data acquisition for LV coverage was 75% shorter for TSENSEmulti-slice CINE compared to nonTSENSE single-slice acquisition (64 ± 24secvs. 258 ± 44 sec; P < 0.0001). ESV evaluation did not show significant differenc-es between TSENSE (67.4 ± 48.8 ml) and nonTSENSE (68.3 ± 50.2 ml) meas-urements (MD: -0.9 ± 3.3 ml; P=0.08). EDV evaluation revealed minor differencesfor TSENSE CINE vs. nonTSENSE CINE (144.9 ± 46.7 ml vs. 148.4 ± 48.2 ml;P=0.0001). EF results of TSENSE and nonTSENSE CINE imaging showed nosignificant differences (MD: -0.4 ± 1.3%; P=0.06). EF intraobserver variability was2.9% for nonTSENSE and 3.9% for TSENSE (P > 0.05).Conclusion: Accelerated CINE SSFP imaging using a standard coil system al-lows to substantially shortening data acquisition for cardiac volumetric analysis.Accuracy of EF measurements is maintained at 4-fold acceleration.

B-467 11:24

Assessment of left ventricular function and mass in patients undergoing CTcoronary angiography using 64-detector row CT: Comparison to MRimagingT. Schlosser1, O.K. Mohrs2, A. Magedanz2, T. Voigtländer2, A. Schmermund2,J. Barkhausen1; 1Essen/DE, 2Frankfurt/DE ([email protected])

Purpose: We hypothesized that the administration of beta blockers prior to MDCTcoronary angiography has a significant impact on left ventricular functional pa-rameters.Methods and Materials: Multiplanar reformations in the short axis orientation

244 C D E FBA

Scientific Sessions

G

were calculated from axial contrast enhanced 64-detector row CT images in 21patients. Patients whose heart rates exceeded 60 bpm received 5 mg bisoprololorally. In case of insufficient heart rate reduction, up to 20 mg of metoprolol wereinjected intravenously. The end-diastolic volume (EDV), end-systolic volume (ESV),stroke volume (SV), ejection fraction (EF) and cardiac output (CO) were ana-lyzed and compared to MR measurements based on continuous short axis imag-es.Results: Mean heart rate was 56 ± 5 bpm during the MDCT and 73 ± 9 bpm dur-ing the MR imaging examination (p < 0.05). The mean EDV and ESV measuredby MDCT were significantly higher compared to MR (MDCT vs. MR: EDV =164 ± 53 ml vs. 144 ± 47 ml, ESV = 77 ± 47 vs. 64 ± 47 ml, p < 0.05). The meanEF and CO derived from MDCT images were significantly lower compared to MR(MDCT vs. MR: EF = 55 ± 12 % vs. 59 ± 15 %, CO = 4.8 ± 0.8 l vs. 5.8 ± 1.3 l,p < 0.05). The mean SV was not significantly different between both methods(MDCT vs. MR: SV = 87 ± 18 ml vs. 80 ± 16 ml, p = 0.44).Conclusion: Left ventricular volumes assessed by MDCT are significantly high-er compared with MR imaging, whereas ejection fraction and cardiac output aresignificantly lower in MDCT. We believe that these differences are caused by theapplication of beta blockers prior to MDCT examinations.

B-468 11:33

Value of functional multi-detector computed tomography (MDCT) of theheart in comparison with cardiopulmonary exercise testing (CPET) inassessment of patients with type A and B (AHA) heart failureM. Zagrodzka, S. Szmit, B. Jaron, M. Cholewa, P. Twarkowski; Warsaw/PL

Purpose: The aim of the study was to assess the value of left atrium volume(LAV) and left ventricle (LV) function in MDCT as an independent marker of heartfailure in patients with atypical chest pain or with low exercise tolerance.Methods and Materials: 69 patients, 48 male, 21 female, age 56 ± 8 years withlow exercise tolerance and atypical chest pain were examined in both MDCT andCPET. In CPET parameters:VO2peak, VO2%N, VO2 AT, VO2 AT%N, VO2pulse,RER=VCO2/VO2, dVO2/dV, EQO2, EQO2 AT, EQCO2, EQCO2 AT, EQCO2 peak/rest, VE, VE-CO2slope, FetO2, FetCO2, VD/VT, VO2/VT, VCO2 were obtained.MDCT functional examination provided data regarding left atrium volume (LAV),total pulmonary vein area (PVA), ejection fraction (EF), stroke volume (SV), car-diac output (CO), left ventricle end diastolic and systolic mass and volumes, peakejection rate (PER), peak filling rate (PFR), time of PER and PFR, PER/EDV,PFR/EDV. Multivariant regression analysis and Student's t-test were performed.Results: Patients with LAV > 120 ml (median of dispersion) had significantly higherPVA (p = 0.025), PFR (p = 0.047), pathological ventilation (VE-VCO2slopep = 0.01) and lower VO2 peak (p = 0.046), VO2%N (p = 0.018). LAV correlatedstrongly with parameters of systolic and diastolic function: CO (p = 0.026, r = 0.46),PFR (p = 0.0016, r = 0.62). LAV strongly correlated with oxygen consumption(VO2 peak p = 0.024, r= - 0.47, VO2 AT p = 0.045, r= - 0.42, VO2%N relative toage and sex p = 0.008, r= - 0.54) and pathologic hyperventilation parameter VE-VCO2slope (p = 0.038, v=0.43). No other of LV function parameters correlatedwith CPET results.Conclusion: MDCT of the heart together with CPET allow early diagnosis andrisk stratification in patients with type A and B (AHA)heart failure. Volume of leftatrium strongly correlates with parameters of diastolic and systolic left heart func-tion. Patients with LAV > 120 ml are especially endangered by heart failure. LAvolume is an accurate and precise morphological parameter of physiologicalpathologies in patients with low exercise tolerance. LA volume can be consideredas an independent non-invasive imaging marker of heart failure in patients withatypical chest pain and low exercise tolerance.

B-469 11:42

Cine-MR imaging demonstrates mechanical ventricular asynchrony inpatients with left bundle branch blockK. Muellerleile, G.K. Lund, A. Stork, P.M. Bansmann, T. Meinertz,H. Reichenspurner, G. Adam; Hamburg/DE ([email protected])

Purpose: Conduction delay in patients with left bundle branch block (LBBB) mayresult in reduced LV performance due to an asynchronous contraction pattern. Inthese patients cardiac resynchronisation therapy (CRT) by biventricular pacingcan be beneficial. Several approaches by echocardiography and MR imaging areused to identify these patients. In the present study we evaluated the ability ofCine-MR imaging to assess mechanical inter- and intraventricular asynchrony.Methods and Materials: 28 patients with reduced LV function (16 patients with,and 12 patients without LBBB ≥ 120 ms) were imaged at 1.5 Tesla. A shared-phases trueFISP sequence using parallel imaging enabling a reconstructed tem-poral resolution of 6.3 ms was used. Electromechanical delay (EMD) was defined

as the interval between R-wave and the peak of systolic shortening. EMD wasassessed for the septal and lateral LV- and lateral RV-wall. Inter- and intraven-tricular delay was calculated as the difference in EMD between these regions.Patterns of contraction in patients with and without LBBB were compared. Inter-and intraventricular delay was correlated with QRS width.Results: In patients with LBBB, inter- and intraventricular delay (130.6 ± 17.3 msand 322 ± 66.4 ms, resp) was longer than in patients without LBBB (18.8 ± 12.4 ms(p < 0.0001) and 51.1 ± 84.3 ms (p < 0.0001), resp). Correlation of inter- and in-traventricular delay with QRS width was only moderate (r = 0.47; p < 0.0001 andr = 0.67; p < 0.0001, resp).Conclusion: Cine-MR imaging identifies mechanical inter- and intraventricularasynchrony in patients with LBBB and may be a valuable tool for the identifica-tion of patients eligible for CRT.

B-470 11:51

Impaired right ventricular systolic function in patients with acute pulmonaryembolism assessed by a combined pulmonary and cardiac MDCT protocolH. Dogan, L.J.M. Kroft, M. Nijkeuter, K. Geleijns, R. van der Geest, H.J. Lamb,M. Huisman, A. de Roos; Leiden/NL ([email protected])

Purpose: To determine right ventricular (RV) systolic function in patients withacute pulmonary embolism (PE) classified by location by multidetector row CT(MDCT).Methods and Materials: Pulmonary and ECG-gated cardiac MDCT scans (Aq-uilion 16 or 64) from 66 patients (29 male, 37 female; mean age ± SD:58 ± 15 years) with suspected PE were evaluated. Twenty cardiac phases werereconstructed retrospectively for each investigation. The end-diastolic and end-systolic endocardial ventricular borders of the RV and left ventricle (LV) weretraced. Ventricular volumetrics and ejection fractions (EF) were calculated. RVvolumetrics and RV/LV volume ratios between patients with PE and without PE(controls) were compared, using independent Samples T-tests. ANOVA post hocTukey test was used to investigate whether the RVEF, RV volumetrics and RV/LVvolumetric ratio differs within subgroups of PE.Results: There were 29 patients with and 37 without PE. Significant difference(p < 0.05) in RVEF (44.4 ± 2.6% vs 52% ± 1.2) and RVESV (105.9 ± 10.2 vs.77.6 ± 3.5 ml) was found between patients and controls. Significant differences(p < 0.05) in RVESV between central PE (136.2 ± 19.4 ml) and controls(77.6 ± 3.5 ml), subsegmental (54.8 ± 3.4 ml) and segmental PE (87.3 ± 10.1 ml)were found. Significant differences (p < 0.05) in RV/LV EDV ratio between centralPE (1.7 ± 0.20) and controls (1.04 ± 0.02), subsegmental (0.93 ± 0.08), segmental(1.05 ± 0.04) and lobar PE (1.11 ± 0.07) were found.Conclusion: RV systolic function assessed with MDCT is variably affected inpatients with acute PE. RV/LV EDV seems to correlate with the PE location. Pro-spective studies are needed to evaluate whether these parameters predict clini-cal outcomes.

10:30 - 12:00 Room G

Head and Neck

SS 1008New approaches in head and neck tumorsModerators:n.n.A. Trojanowska; Lublin/PL

B-471 10:30

Thin-slice spiral CT of the neck: A must for cancer staging in routine clinicalpractice?M. Lell, C. Gmelin, H. Greess, W. Bautz; Erlangen/DE([email protected])

Purpose: To evaluate the benefit of thin-slice multiplanar evaluation of head andneck spiral CT in tumor staging.Methods and Materials: 100 patients with histologically proven squamous cellcarcinoma of the head and neck were examined with 16 x 0.75 mm collimation.Identical CT data sets, reconstructed with 3 mm and 1 mm were used for evalu-ation. The examinations were evaluated with a) 3 mm thick axial images; b) 1 mminteractive MPR. T-staging as well as infiltration into designated anatomical struc-tures were assessed with both methods. Two examiners blinded to clinical dataassessed the data-sets.Results: No statistical difference was found for T-staging of the tumors. There

B 245C D E FA

Su

nd

aS

un

da

Su

nd

aS

un

da

Su

nd

ayy yyy

Scientific Sessions

G

was a trend towards a better performance in identifying infiltrated structures atthe larynx, but not at a statistically significant level.Conclusion: Thin slice (1 mm) reconstruction seems to be unnecessary for stagingpurposes, but thin slice collimation for data collection is advantageous to clarifyselected cases with MPR.

B-472 10:39

Functional CT imaging of extracranial head and neck tumours with adistributed-parameter tracer kinetic model analysis: Initial resultsS. Bisdas1, M. Baghi1, T.S. Koh2, A. Smolarcz1, N. Chambron Pinho1, V. Khan1,D.L. Cheong2, W. Gstoettner1, T.J. Vogl1; 1Frankfurt/DE, 2Singapore/SG([email protected])

Purpose: To evaluate the feasibility of head and neck tumor perfusion CT meas-urements using a two-compartment distributed-parameter (DP) physiologic model.Methods and Materials: 26 patients with extracranial head and neck neoplasmsand metastatic nodes underwent dynamic contrast-enhanced CT scans obtainedfrom a 16-detector row multislice CT scanner. After normalization of the areaunder the arterial curve to the concentration curve at a vein voxel, perfusion par-ametric maps that corresponded to blood flow (F), fractional intravascular (V1)and extracellular extravascular blood volume (V2), transit time in vascular com-partment (T), first pass extraction ratio (E) and permeability-surface area product(PS) were generated. A time delay or time lag parameter (Tlag) was introducedso as to denote the time difference between the initial rise of the arterial andtissue density curves.Results: After optimization of resolution and speed, low-inherent noise paramet-ric maps provided adequate anatomical and functional information of the slicethat was scanned in vivo. The average values of F, V1, V2, E, and PS were signif-icantly higher in primary tumors and metastatic nodes than in adjacent normalmuscle structures. The mean values of T and Tlag were significantly reduced inthe neoplasms than in normal tissue. The V1 and V2 maps showed wide inter-and intratumoral variability. F, V1, Tlag, E, and PS maps provided the best tumordelineation.Conclusion: Perfusion CT parameters obtained from a DP physiologic tracerkinetic model can be used to differentiate extracranial head and neck neoplasmsand metastatic lymph nodes from adjacent normal tissue.

B-473 10:48

Whole-body dual modality PET/CT for primary tumor detection in patientswith lymph node metastases in the head and neck regionK. Nassenstein1, P. Veit1, H. Stergar1, A. Gutzeit2, J. Barkhausen1, G. Antoch1;1Essen/DE, 2Arau/CH ([email protected])

Purpose: To assess the diagnostic accuracy of combined positron emission tom-ography (PET)/computer tomography (CT) for primary tumor detection in cervi-cal metastases of unknown origin compared to PET, CT, and PET+CT side-by-sideevaluation.Methods and Materials: Between December 2001 and April 2005 39 consecu-tive patients (8 women, 31 men, mean age 59.9 ± 11.2 years, ranging from 39 to89 years) with cervical lymph node metastases of an unknown primary (CUP)were enrolled in this study. PET/CT images were obtained 1 hour after injectionof 350 MBq of fluorodeoxyglucose. Oral and intravenous contrast agents wereadministered in all patients to ensure diagnostic CT data. Fused PET/CT datawere evaluated for primary tumor detection. The diagnostic accuracy was calcu-lated and compared with CT, PET, and side-by-side PET+CT evaluation. Statisti-cal analysis of obtained differences in diagnostic performance was based on theMcNemar test.Results: Fused PET/CT depicted the primary tumor in 11 of 39 (28%) patients.In 28 (72%) patients the primary tumor remained occult. CT revealed the primaryin 5 (12.8%), PET alone in 10 (25.6%), and side-by-side evaluation of PET+CT in10 (25.6%) of 39 patients. Statistical analysis showed no significant differencesbetween the imaging modalities.Conclusion: PET, side-by-side PET+CT, and PET/CT revealed similar detectionrates for primary tumors in cervical CUP patients. Therefore, cervical metastasesof an unknown primary may be assessed with either of these imaging modalities.However, detection rates of CT were substantially lower. Thus, inclusion of func-tional data for assessment of cervical CUP patients must be recommended.

B-474 10:57

Comparison of FDG PET-CT with contrast enhanced MR imaging for theevaluation of head and neck tumorsJ. Gasser, P. Reinprecht, K. Hausegger, P. Lind, F. Chiari, P. Eckl;Klagenfurt/AT ([email protected])

Purpose: To compare the diagnostic potential of PET-CT with contrast enhancedMR imaging for the evaluation of head and neck tumors.Methods and Materials: 23 patients with known malignancies of the head andneck region were examined on a commercial PET-CT scanner, and they had anadditional contrast enhanced MR imaging scan on the same day. The tumorswere staged based on the results of the examinations according to the TNM clas-sification. All patients proceeded to surgery, including neck dissection, the refer-ence standard being intraoperative and histological findings.Results: PET-CT and MR imaging correctly staged the local spread of 20 tumors(87%). MR imaging overestimated the tumor extent in 2 cases and underestimat-ed 1 tumor. PET-CT understaged 2 tumors and overstaged one. In nodal stagingMR imaging correctly classified 19 of 23 patients (sensitivity 77%, specificity 85%),whereas PET-CT correctly staged 21 Patients (sensitivity 88%, specificity 92%).Conclusion: In this prospective study PET-CT and contrast enhanced MR imag-ing showed similar results in predicting local tumor spread in head and necktumors. PET-CT had a better performance in lymph node staging then MR imag-ing. However, false positive and false negative findings occurred in both meth-ods.

B-475 11:06

Accuracy of FDG-PET/CT in head and neck cancer staging: Initial resultswith histopathological correlationP. Veit, I. Wanke, C. Luczak, T. Egelhof, A. Bockisch, S. Rosenbaum,G. Antoch; Essen/DE ([email protected])

Purpose: To compare the accuracy of combined PET/CT, PET+CT viewed sideby side and CT alone for TNM staging in head and neck cancer using histopatho-logical findings as the standard of reference.Methods and Materials: 30 patients (mean age:59 years) with suspicion of headand neck cancer underwent whole-body [18 F]-2-fluoro-2-deoxy-D-glucose (FDG)PET/CT. PET/CT was evaluated by a nuclear medicine physician and a radiolo-gist in consensus, PET+CT was evaluated by a different team and the CT wasevaluated by two radiologists. The readers were aware of the clinical background.Histopathological results served as the standard of reference for all lesions foundin all patients. The accuracy of TNM staging was assessed and compared withthe reference standard. Differences between the staging procedures were testedfor statistical significance by McNemar's test.Results: Overall TNM-staging proved more accurate (p < 0.05) with combinedPET/CT compared to PET+CT and CT alone. This result was based on a moreaccurate assessment of the T-stage with PET/CT (p < 0.05). No statistically sig-nificant difference was detected for N- and M-staging. Tumors were overstagedwith combined PET/CT in 1, with PET+CT in 6, and with CT alone in 4 cases,whereas PET/CT understaged 8, PET+CT 10, and CT alone 15 tumors.Conclusion: Combined PET/CT proved to be significantly more accurate in as-sessing the overall TNM-stage than the compared modalities. These results arebased on a higher accuracy concerning the T-stage. Since these findings areinitial results, larger patient cohorts are needed to define role of PET/CT in thispatient population.

B-476 11:15

To establish the normal range of FDG uptake in palatine tonsils and tocompare these values with uptake in known tonsillar primaries using PET/CTD.J. Gibson; London/UK

Purpose: 1. To establish normative data for FDG PET-CT uptake in the palatinetonsil and to evaluate this against tonsillar malignancy. 2. To define normal rangeof uptake between left and right tonsil for control group and those with tonsillarcarcinomas.Methods and Materials: We retrospectively analysed 49 consecutive, non aero-digestive tract oncology patients with whole body PET-CT. This population con-sisted of 24 females and 25 males. This patient group had a low risk of head andneck cancer. A second group of six patients with tonsillar carcinomas was simi-larly analysed. Patients who had chemotherapy were excluded. FDG PET wasobtained using a Discovery ST PET-CT. Images were acquired 1 hour ± 5 min-utes post injection using an average of 381 MBq FDG. Two experienced observ-ers identified palatine tonsil using axial CT images from PET/CT data. Localisation

246 C D E FBA

Scientific Sessions

G

of tonsillar FDG uptake was refined with fused SUV maps. Regions-of-interestwere placed over selected slices to determine the maximum uptake for eachtonsil.Results: In the control group studied, the SUV right tonsil = 4.44 ± 1.95 (M ± SD),left tonsil = 4.33 ± 2.06. Average percentage difference SUV between left andright tonsil uptake was 13% ± 9.3 (M ± SD), which did not reach significance.Five of six tonsillar primaries had SUV values outside physiological range (5th,95th centiles).Conclusion: Within the control group, no significant difference in FDG uptakebetween left and right tonsil was observed. We quantitatively discerned 5 of 6tonsillar carcinomas.

B-477 11:24

Evaluation of thyroid malignancy using real-time tissue elastographyK. Tanaka1, N. Fukunari2, T. Igarashi1, H. Akasu1, W. Kitagawa1, K. Shimizu1,K. Ito1; 1Tokyo/JP, 2Yokohama/JP ([email protected])

Purpose: The aim of this study was to investigate the improvement of differenti-ation between benign and malignant thyroid tumors by real-time tissue elastogra-phy (RTE).Methods and Materials: Two hundred thyroid nodular lesions found on B-modeultrasonography were examined and evaluated using RTE (EUB-8500, HITACHI).All cases were also compared with the cytological and pathological results.Results: Unique and characteristic images were obtained by RTE in every typeof thyroid nodular disease. Images were mainly divided into four visual types;Pattern 1: colored light green throughout nodule, Pattern 2: colored light green inthe center and blue peripherally, Pattern 3: blue zone at base with mixed coloringof light green and sometimes with red, Pattern 4: blue coloring throughout thenodule.130 tumors were histologically diagnosed as malignancy and consisted of 80tumors (61.5%) of pattern 4, 17 (13.1%) of pattern 3, 21 (16.2%) of pattern 2 and12 (9.2%) of pattern 1. In papillary thyroid cancer, 87 of the 100 tumors (87.5%)showed pattern 3 or 4. Especially, lymph node metastasis demonstrated pattern4. Eight of the 15 follicular cancer cases (53.3%) showed pattern 2.Conclusion: In the differential diagnosis of thyroid nodules, RTE with simultane-ous gray scale imaging provides useful information, especially in follicular tumorand papillary cancer diagnosis.

B-478 11:33

Clinical evaluation of real-time tissue elastography (RTE) for the differentialdiagnosis of thyroid follicular lesionsN. Fukunari1, K. Tanaka2, K. Arai1, A. Nakamura1, K. Yamazaki1, K. Ito2,T. Mitake2; 1Yokohama/JP, 2Tokyo/JP ([email protected])

Purpose: The diagnosis and management of follicular carcinoma of the thyroidremains a controversial topic. The aim of this study was to evaluate the clinicalusefulness of real-time tissue elastography (RTE) imaging for the differential di-agnosis of thyroid follicular lesions.Methods and Materials: The 56 follicular tumors obtained from 49 patients wereexamined by gray-scale US imaging, colour Doppler imaging and RTE. A com-parative study of these images was made and final diagnosis was determined byhistological diagnosis. The images of thyroid follicular lesions obtained by RTEwere recognized as four typical patterns as follows: 1) nodule is relatively ho-mogenous and colored with light green, 2) the centre of nodule is colored withgreen, and its periphery is colored with blue, 3) light green and red mixed insideblue zone and 4) the whole tumor is hard and displayed in blue.Results: All the follicular tumors were cytologically and histopathologically con-firmed. In this series, 30 adenomatous nodules, 21 follicular adenomas and 5follicular cancers were found. Of the 51 benign thyroid nodules, 48 cases (94.1%)demonstrated pattern 1, and 3 cases demonstrated pattern 3. On the other hand,all the follicular cancer cases manifested as pattern 2.Conclusion: In follicular cancer, the difference of elasticity from the core and theperiphery of the tumor is supposed to be caused by the difference of hypercellu-larity. RTE can provide new useful information for the differential diagnosis ofthyroid follicular lesions.

B-479 11:42 !Fine needle aspiration of thyroid gland: A retrospective analysis of thesafety and efficacy of 1468 consecutive ultrasonography (US) guidedproceduresP.D. Domínguez, A. Alonso, J.J. Noguera, E. de Luis, A. Villanueva, A. Benito,J. Larrache, M.D. Lozano; Pamplona/ES ([email protected])

Purpose: To review our experience with US-guided fine-needle aspiration biop-sies (FNAB) of the thyroid gland to determine the frequency of complications andunsatisfactory smears and the correlation with surgical excision results.Methods and Materials: Thyroid US-guided FNAB from January 2000 to De-cember 2004 were reviewed. Radiological and pathological reports were ana-lyzed. In benign lesions, repeat FNAB in 1 year was recommended if asymptomaticand surgery if symptomatic. When results were malignant or indeterminate (fol-licular proliferation) surgery was indicated.Results: 1468 lesions in 1112 patients (189 men, 923 women, mean age 51y,range 12-87) were punctured. 8.9% were diffuse disease. Average nodule sizewas 20.0 mm (SD 11.52, range 2-78 mm). Cytological results included 351 inad-equate specimens (23.9%), 1002 benign (68.63%), 60 indeterminate (4.1%), and53 malignant (3.6%). There was no statistically significant difference in meannodule size of adequate and inadequate specimens (p 0.07). There were statisti-cally significant differences in the proportion of inadequate smears between per-formers (range 12.9% to 32.0%) yet not related only to experience, suggestingimportance of puncture technique or other factors. In 355 (24.2%) lesions whichunderwent surgery, pathology demonstrated benign change in 269 (18.3%) andmalignancy in 86 (5.9%). For the diagnosis of malignancy sensitivity was 46.5%,specificity 98.5%, PPV value 90.9% and NPV value 85.2%.No severe complications and only 6 mild complications occurred.Conclusion: FNAB of thyroid gland is a safe procedure, with a 23.9% rate ofinadequate cytological specimens. For the diagnosis of malignancy sensitivity isintermediate and specificity high. Repeated FNAB controls do not exclude malig-nancy.

B-480 11:51

Imaging features of cytology/histologically-proven secondary tumors to thethyroid (STT). A retrospective multicenter studyP.-Y.R. Marcy1, M.-M. Plantet†2, D. Dequanter3, A. Bozec1, N. Amoretti1,K. Marsot-Dupuch†4; 1Nice/FR, 2Saint Cloud/FR, 3Bruxelles/BE,4Le Kremlin-Bicêtre/FR ([email protected])

Purpose: To describe the radio-pathologic features of STT.Methods and Materials: Over an eleven year period, imaging features of STTwere retrospectively reviewed in 34 cancer patients (Nice, Paris, Brussels). AllSTT were cytologically or histologically proven.Results: Sex ratio M/F was 0.6; mean age was 61.2 years. Origin of the primarytumour included breast (8/34), bronchus (7/34), kidney (6/34), head and neck (8/34), sarcoma (2/34), and genitourinary carcinoma (1/34). At initial presentation(mean 42.4 months after diagnosis of the primary), patients presented with soli-tary or multiple thyroid nodules (11/34), a heterogeneous infiltrated lobe (15 /34),or an ill defined hypoechogenic/hypodense goitre (8/34). T staging of the primarytumor was T3 or T4 in 70% of cases. When present (60%), metastatic adenopa-thy was located in levels III, IV, VI, paratracheal, and VII [breast and head andneck cancers], and may be indicative of lymphatic retrograde dissemination tothe thyroid. Mean survival was 10.2 months (range, 10-26). Prognosis was betterfor renal metastases despite frequent jugular invasion (3/6). Calcifications wererare, and due to invasion of preexisting benign nodules. Firm cystic-like lesions(n = 2) recurred rapidly after Fine Needle Aspiration (FNA).Conclusion: STT are the result of direct extension of malignancies from contig-uous structures or the result of hematogenous/lymphatic spread from distant sites.Prognosis is poor, imaging features are various. Metastatic renal cell carcinomamay mimic primary thyroid carcinoma. Cystic-like lesion are seen in case of he-matogeneous spread. Sensitivity of FNA for diagnosis is 89.5%.

B 247C D E FA

Su

nd

aS

un

da

Su

nd

aS

un

da

Su

nd

ayy yyy

Scientific Sessions

G

10:30 - 12:00 Room H


SS 1009Spinal interventionModerators:P. Lohle; Tilburg/NLO. Tervonen; Oulu/FI

B-481 10:30

Vertebroplasty of osteoporotic vertebral fractures: New adjacent anddistant fractures during follow-upM.B. Pitton, S. Herber, N. Morgen, J. Schneider, P. Drees, B. Böhm, C. Düber;Mainz/DE ([email protected])

Purpose: To evaluate the number and time-interval of new vertebral fracturesadjacent and distant to pretreated vertebrae, and to evaluate the impact of in-tradisc cement leaks.Methods and Materials: Osteoporotic fractures were treated with CT-guidedvertebroplasty and were followed-up with CT. Clinical results, numbers of extra-osseous cement leakages, and numbers and time-intervals of new fractures wereanalysed. New fractures were treated again with vertebroplasty and were includ-ed in the follow-up protocol.Results: Three hundred and fifteen vertebral fractures were treated in 160 pa-tients with osteoporotic vertebral fractures (age 71.1 ± 9.9 years). During follow-up(16.7 ± 10.3 months, range1-36), a total of 55 new fractures occurred in 48 pa-tients (26.7%) and were treated again with vertebroplasty. These were distribut-ed with 29 adjacent to pretreated vertebrae at 2.6 months (± 2.5, range 0.5-11),and 26 in distant vertebrae at 5.8 months (± 5.1, range 0.5-22). New adjacentfractures were located inferior to pretreated vertebrae in 15/29, superior in 10/29and between two pretreated fractures in 4/29 (sandwich fracture). There werecement leakages from the pretreated vertebrae into adjacent disc spaces in 11/29 fractures, 18/29 occurred without intradisc cement leaks. However, 3/4 sand-wich fractures occurred in cases with pre-existing intradisc cement leaks in adja-cent disc spaces. The distribution of new fractures represented the fracturedistribution of the whole collective.Conclusion: Adjacent fractures occur earlier than distant fractures during follow-upand might be caused by several biomechanical factors, e.g. changes in the stat-ics of the spinal column and hardening of the cemented vertebral bodies. In-tradisc leaks might be of minor influence.

B-482 10:39

Sacroiliac joint disruptions: CT-guided fixationN. Amoretti1, I. Hovorka1, P. Eude2, P.-Y. Marcy1, P. Brunner3, M. Fonquerne1,P. Chevallier1, J. Bruneton1, P. Boileau1; 1Nice/FR, 2Aurillac/FR, 3Monaco/FR([email protected])

Purpose: We propose describing the procedure performed under CT-guidanceof trans-articular sacroiliac application of screws on patients with posttraumaticsacroiliac disruption in 12 patients.Methods and Materials: We performed this procedure on patients with unstablebut reducible sacroiliac disruption: classified as Tile B (vertically stable but rota-tionally unstable) and Tile C (vertically and rotationally unstable). With CT-guid-ance each patient then underwent percutaneous placement of sacral and posterioriliac screws. An 8.0 mm diameter cannulated screw (Stryker, Asnis) was used.Depth and angulation measurements for placement of the materials were deter-mined from an axial image. A small skin incision was made at the entry sites. Theexact length of the screw was calculated by the CT-scan.Results: The mean time for the procedure was 2 hours. No complications, inparticular infectious, occurred during the follow-up of the patients. The hemor-rhage during the procedure was negligible, less than 25 cc. All patients had asatisfactory sacroiliac stabilization. Less than 20% had residual pain with sacro-iliac arthrosis.Conclusion: CT-guided fixation of reducible sacroiliac disruptions is a safe tech-nique and a good alternative to classic surgery.

B-483 10:48

Percutaneous laser diode disc decompression (PL3D)P.P. Menchetti1, E. Mazza1, G. Canero2, W. Bini3, L. Sali1, S. Lucarini1,K.S. Ternovoy4, L. Longo1, N. Villari1; 1Firenze/IT, 2Roma/IT, 3Seligenstadt/DE,4Moscow/RU ([email protected])

Purpose: To evaluate the effectiveness of percutaneous laser diode disc decom-pression (PL3D) in the treatment of lumbosacral disc herniation.Methods and Materials: From September 2002 to July 2005 we performed 800percutaneous laser diode disc decompressions in 680 patients (age range 16-82, mean age 42) suffering with relevant symptoms resistant to medical therapy.We used fluoroscopic guidance in 550 patients and CT guidance in 130. Thelevel of disc decompression was L2-L3 in 42 cases, L3-L4 in 92, L4-L5 in 379,L5-S1 in 287. In 48 cases the PL3D was performed after an unsuccessful micro-surgical approach. Access to the intervertebral disc was performed with a 21GChiba needle of 100 or 150 mm. The needle was introduced either with an ob-lique or a paramedian transdural approach. In all cases a multidiode laser with awavelenght of 980nm was used. The clinical success was evaluated by the Mac-Nab criteria.Results: After a mean follow-up of 14 months 527 patients (77.5%) had a goodor excellent response according to MacNab criteria. We had complications in 21patients: Aseptic spondylitis in 5 cases and headache related to CSF leakage in16 cases.Conclusion: Percutaneous disc decompression, performed with CT or fluoro-scopic guidance using a multidiode laser inserted through a 21G needle, ap-pears to be an effective and practically uncomplicated treatment for herniatedintervertebral discs.

B-484 10:57

Outcome of disk nucleolysis by oxygen-ozone injection in patients withdegenerative disease of the lumbar spine: Impact of morphologic disorder,age and genderB. Oder, M. Loewe, W. Lang, W. Ilias, S. Thurnher; Vienna/AT([email protected])

Purpose: Oxygen-ozone injection for disc nucleolysis is a minimal invasive ther-apy for patients with chronic low back pain. This study assessed potential de-pendencies of the outcome regarding different underlying morphologic disorders,multilevel degenerative disease, age and gender.Methods and Materials: A total of 612 patients were treated with an intradiscal(4 mL) and periganglionic (6 mL) injection of an oxygen-ozone mixture at an ozoneconcentration of 30 µg/mL, followed by a periganglionic injection of long lastingcorticosteroid and anesthetic. Patients were divided into 5 groups according tounderlying disease. Questionnaires regarding the Oswestry disability score (ODS)and a visual pain scale (VPS) were evaluated 6 months after treatment.Results: At 6 months follow-up, improvement in ODS was statistically significant,in particular in patients with disc protrusion (mean value before and after treat-ment 46% and 31% respectively, Wilcoxon test, Kruskal-Wallis test p < 0.0001).Improvement in VPS was also statistically significant (mean value before andafter treatment 8.5 and 5.4 respectively, Wilcoxon Test p < 0.0001). Patients withonly one affected segment showed a significantly better outcome compared topatients with multilevel disease, and patients under 50 years had more benefitfrom treatment.Conclusion: Disc nucleolysis by ozone injection is an effective, minimally inva-sive treatment in patients with chronic low back pain who fail to respond to non-invasive means, especially if they have single level disease, disc protrusion, andare under 50 years of age.

B-485 11:06

CT-guided vertebro- and kyphoplasty: Comparing technical success andcomplications in 101 casesC.H. Weber, M. Kroetz, R.T. Hoffmann, E. Euler, S. Heining, K.J. Pfeifer,M.F. Reiser, U. Linsenmaier; Munich/DE([email protected])

Purpose: To compare technical success and complications in CT-guided verte-bro- and kyphoplasty.Methods and Materials: From 2002 to 2005, 69 patients (101 vertebrae) weretreated with vertebroplasty (n = 82) or kyphoplasty (n = 19) using only CT-fluor-oscopy of a 4-slice MSCT as guidance. 78 vertebrae were osteoporotic compres-sion fractures or diffuse plasmocytoma mimicking osteoporosis and 23 werecircumscript metastatic osteolysis treated with radiofrequency ablation (RFA) andvertebroplasty.

248 C D E FBA

Scientific Sessions

G

Results: Technical success was achieved in all 101 cases. Clinical outcome iscurrently under investigation. Post-interventional CT demonstrated asymptomat-ic cement leakage in 51/101 vertebrae. Kyphoplasty showed leakage in 11/19cases (57.9%) versus 40/82 cases (48.8%) of vertebroplasty. There was no sig-nificant difference between kyphoplasty and vertebroplasty (p = 0.41, Mann-Whitney Test). There was one minor complication consisting of L5 root irritationafter RFA plus vertebroplasty of a sarcoma metastasis, which subsided withouttreatment after 8 weeks. There was one major complication involving intraspinalcement leakage during tumour vertebroplasty, causing T5 root compression andrequiring laminectomy for cement removal. The overall rate of major complica-tions requiring treatment was 0.99%.Conclusion: Vertebro- and kyphoplasty can be safely performed under MSCT-fluoroscopy guidance alone. The rate of major complications is very low. We sawa high rate of small asymptomatic cement leakages which would likely be over-looked on conventional fluoroscopy (CF). This raises the question of whether CFis indeed the standard of guidance. Further studies comparing CF and MSCT-fluoroscopy are therefore required. There was no statistically significant advan-tage for kyphoplasty concerning cement leakage and technical success, leavingthe increased costs without an obvious benefit.

B-486 11:15

Lumbar corticoid epidural infiltrations in sciaticas and cruralgias under CTguidance (500 patients): A long term analysisJ. Mazaud, D. Ben Salem, R. Loffroy, B. Guiu, N. Vignon, M. Dranssart,A. Kousmenko, D. Krause; Dijon/FR ([email protected])

Purpose: To assess the efficacy, accuracy and safety of foraminal and directintracanalar epidural steroid injections under CT guidance for sciaticas and cru-ralgias. The long term analysis of the clinical results is reported and discussed.Methods and Materials: Five-hundred patients were treated for sciatica and/orcruralgia due to a disk herniation, without sensory-motor deficit, resisting con-ventional medical treatment of 3-4 weeks duration, between January 2000 andNovember 2004. The patient is in the prone position; the incriminated level (diskherniation) is first controlled with a lateral scout-view (CT). After local anesthe-sia, the extremity of a thin needle (22G) is positioned under CT guidance asfollows: in the proximal part of the foramen in the case of lateral herniations (cru-ralgias,1/4 of the patients series) and directly in the lumbar canal, through theligamentum flavum, in a strict epidural location, near the root. A contrast CT con-trol verifies the epidural diffusion space (even with a foraminal approach) beforethe corticoid infiltration (Cortivazol 7.5 mg).Results: The clinical follow-up is realized at 1, 3, 6 and 18 months. GP question-ing, professional activity recovery and medical treatment stoppage confirm 65 %of good results after a mean delay of 26 months. The treatment failed in 35% ofcases (175/500) but less than half of these patients underwent surgery (85/500 =17%).Conclusion: Delayed corticoid injection under CT guidance allows, in all ana-tomic situations, marked infiltration of the root, the ganglion and also of the intra-canalar epidural space. This rapid, safe and accurate procedure now representsthe treatment of choice in the case of resisting radiculalgia particularly in ambu-latory patients. The operated failure rate (17%) after procedure, mainly during thefirst month, is acceptable.

B-487 11:24

Vertebral augmentation through vertebroplastyJ. Hierholzer, A. Strassenburg, H. Fuchs, K. Westphalen; Potsdam/DE([email protected])

Purpose: The ability of vertebroplasty (VTP) to enhance vertebral body heighthas been retrospectively evaluated in our patient population.Methods and Materials: Vertebral bodies treated (n = 102) in 70 patients withsymptomatic osteoporotic collapse were evaluated by 3D-measurements on pre-op MR imaging and post-op CT respectively. Nine point measurement and as-sessment of kyphosis angle were performed and results correlated with clinicalfeatures.Results: The measurement method proved to be reliable in control vertebrae.Mean height increased significantly after VTP (5-30% of initial height, p < 0.001).The best elevation was obtained in the center of the vertebrae, while the posteri-or wall, being less compressed initially, displayed only mild augmentation.Conclusion: VTP is able to increase the height of compressed vertebral bodiestreated for symptomatic osteoporotic fractures.

B-488 11:33

Percutaneous nucleoplasty in the treatment of discal lumbar painS. Masala, A. Ursone, R. Fiori, F. Massari, G. Simonetti; Rome/IT([email protected])

Purpose: To assess the effectiveness of percutaneous nucleoplasty in the treat-ment of lumbar pain due to contained discal herniation.Background: Ninety-two patients with positive visual analogue scale (VAS) (meanscore: 8) and lumbar pain were selected for treatment. In all patients lumbar MRimaging revealed a contained discal herniation (3 pts: L1-L2; 15 pts: L2-L3; 26 pts:L3-L4; 42 pts: L4-L5; 6 pts: L5-S1). All patients were refractory to medical andrehabilitative therapy for more than three months. The main physical principle ofthis procedure is a bipolar radiofrequency energy emitted by an electrode-nee-dle, with consequent volumetric reduction of the nucleus polposus and elasticreturn of the anulus fibrosus. The treatment is performed under biplanar fluoro-scopic guidance.Results: Postprocedural evaluation confirmed reduction of the VAS (mean score:3). MR imaging performed 6 and 12 months after the procedure demonstratedcomplete resolution of the treated discal pathology in all patients. No post-proce-dural complications or disease recurrences were observed.Conclusion: In our experience percutaneous nucleoplasty demonstrated to bean easy, effective and feasible technique. This minimally invasive procedure isideal for treatment of patients with painful contained discal herniation, refractoryto conservative medical therapy, not candidates for surgical treatment.

B-489 11:42

CT-guided ozone-oxygen injection into lumbar disc herniation:A prospective study for lower back pain due to disc herniationT. Lehnert, S. Mundackatharappel, A. Wetter, W. Schwarz, M.G. Mack,T.J. Vogl; Frankfurt/Main/DE ([email protected])

Purpose: To evaluate the therapeutic benefit and morphologic changes after CT-guided intradiscal and intraforaminal ozone-oxygen injection combined with anintraforaminal drug application.Methods and Materials: 69 patients with lumbar radiculopathy were treated witha combined intradiscal and intraforaminal injection of ozone-ozygen, Betametha-sone-21-dihydrogenphosphate, Betamethasone-21-acetate and Bupivacainehy-drochloride under CT-guidance. During a period of 3 months the effects of thetherapy on lumbar radiculopathy were evaluated. Volumetric examination (MRimaging-volumetry) of disc herniation was accomplished, changes on lumbar ra-diculopathy were examined with a 0-10 numerical pain rating scale.Results: At 7 days after injection 15 patients (21.7%) showed a complete reliefof pain, in 12 patients (17.4%) we had a reduction of pain of 1 point, in 10 patients(14.5%) a reduction of 2 points, in 8 patients (11.6%) a reduction of 3 points andin 20 patients (29%) a reduction of ≥ 4 points. In 4 patients (5.8%) no significantchange in lumbar radiculopathy was demonstrated. Three months after interven-tion the follow-up revealed in 17 patients (24.6%) a complete relief of pain, astrong relief (reduction of > 4 points) in 23 patients (33.3%), in 22 patients (31.9%)a mild relief (reduction of 1-3 points) and in 7 patients (10.1%) no relief of painwas demonstrated. Volumetric examination showed a reduction of disc hernia-tion, an average of 13.4% after three months.Conclusion: CT-guided combined intradiscal and intraforaminal injection of ozone-ozygen represents a therapeutic alternative for lumbar radiculopathy with prom-ising results. Its ease of execution permits the successful outpatient treatment oflumbar sciatic pain.

B-490 11:51

Posterior arthrodesis of the spine by percutaneous CT-guided application ofscrewsN. Amoretti1, I. Hovorka1, M.E. Fonquerne1, N. Bronssard1, P.-Y. Marcy1,P. Brunner2, P. Boileau1, J. Bruneton1; 1Nice/FR, 2Monaco/FR([email protected])

Purpose: To benefit from a mechanical fixation after an anterior osteosynthesisof the spine without the inconveniences of the classic posterior surgical interven-tion, we have adopted a technique, performed under CT guidance, that permits atrans-articular application of screws (CT lumbar interbody-fusion). We proposethe evaluation of this new radio-surgical procedure.Methods and Materials: From September to August 2005 we have performedunder CT guidance fifteen spinal arthrodesis by a posterior approach. This CT-lumbar interbody fusion is an adaptation of the procedure of percutaneous oste-osynthesis performed on the pelvis. It is performed by a radiological-surgical teamand takes place in the CT room.

B 249C D E FA

Su

nd

aS

un

da

Su

nd

aS

un

da

Su

nd

ayy yyy

Scientific Sessions

G

Results: No complications occurred during the follow-up of the patients.Fourteen patients showed a stabilization of the immobilized lumbar segment.Three months from the intervention one patient showed clinical evidence of nar-rowing of the lumbar canal. The CT examination did not show an abnormal posi-tion of the trans-articular screws.Conclusion: Our results of CT-lumbar interbody fusion are encouraging, we dem-onstrated its technical feasibility and its efficacy. A larger series of cases shouldconfirm the value of this technique.

10:30 - 12:00 Room I


SS 1005Web-based support for radiological practice andteachingModerators:W. Hruby; Vienna/ATE. Neri; Pisa/IT

B-491 10:30 !Comparison of different teaching and learning methods by standardisedinformation. Classroom teaching, print media, and digital media -exemplified through the subject of X-ray anatomyD. Ketelsen1; F. Schrödl1, I. Knickenberg1, R.A. Heckemann2, W.L. Neuhuber1,W.A. Bautz1, M. Grunewald1,1Erlangen/DE, 2London/UK

Purpose: The intention of the trial was to compare differing teaching and learn-ing methods (classroom teaching, print media, and digital media) conducted instandardized conditions. This was carried out in order to achieve an increase inefficiency by utilising the best teaching method possible.Methods and Materials: In order to standardise the information that was to beimparted, the digital media was compiled on the same basis as that of the printmedia. The lecturers adhered to these media during their classroom instructions.Thus, medical students in the second semester, who possessed some anatomi-cal but no radiological qualifications, were lectured by this method in the subjectof X-ray anatomy. These teaching methods were evaluated objectively with mul-tiple choice tests and subjectively with a student survey.Results: The evaluation of the objective data showed no detectable differencesconcerning the teaching methods. The subjective data, however, displayed thatthe students' preferred a combination of digital media and classroom teaching bydedicated lecturers. Learning by print media was mostly depreciated.Conclusion: The lack of experienced and committed lecturers, who are superiorto digital media, necessitates compromises. The trial indicates, therefore, thatthe combination of digital media and highly qualified, dedicated lecturers wouldprovide an ideal solution to the current problems.

B-492 10:39 !eHealth initiatives and action plans around the world: Implications on thebusiness of radiologyH.K. Pohjonen1, J.G. Blickman2; 1Espoo/FI, 2Nijmegen/NL([email protected])

Purpose: The purpose of the paper was to review eHealth initiatives to improvehealthcare and lower costs with information and communication technology inEurope and North America, and to define implications on the business of radiol-ogy.Methods and Materials: eHealth initiatives have been launched both in Europeand North America to augment provider-based systems with regional and nation-al solutions to improve accessibility to imaging information. This work includescomparing approaches in different geographic regions including technologies andbusiness impacts. In 2004 an eHealth initiative accompanied by an action planwas launched by the European Commission. In Canada, there is an initiativecalled Canada Health Infoway. The US government has called for a national healthrecord to be available in a decade and has begun assisting states and communi-ties to implement regional solutions linking together.Results: eHealth programs seem to be the key drivers to change the business ofradiology. In summary the implications are 1) there will be regional and nationalprojects, 2) projects will be driven by high level governmental decision makers, 3)purchasing will be focused on solutions and services rather than software andhardware components, 4) PACS will be integrated to national level healthcarerecord summaries, 5) images will be utilized across boundaries and 6) long-termarchiving will include all clinical data, not just images.

Conclusion: There are different technical approaches, but similar business im-plications, of the process to improve health and lower costs in the eHealth pro-grams in Europe and North America.

B-493 10:48

The European Radiological Digital Data System project (ERDDS)G. Pellicanò1, J. Albillos2, J. Neuwirth3, J. Alarcón4, G. Frijà5, Y. Maratos5,P. Rinaldi6, L. Bartoli1, C. Conti7; 1Florence/IT, 2Alcorcon/ES, 3Prague/CZ,4Madrid/ES, 5Paris/FR, 6Rome/IT, 7Milan/IT ([email protected])

Purpose: ERDDS project, funded by the eContent programme of the EuropeanCommission, is an initiative of a pool of partners from four different Europeancountries: Radiological Societies of Italy (SIRM), Spain (SERAM), France (SFR)and Czech Republic, European hospitals (AUOC, FHA, HEGP, FNMotol), privatesoftware firms (Softeco and Sadiel) and a consulting company (IMS). ERDDSaims to integrate the national radiological nomenclatures through a software plat-form. It will represent a strong pillar in achieving a European shared classificationof radiological services.Methods and Materials: The system requirements have been defined throughthe analysis and modelling of real data (clinical and administrative nomenclature,etc). collected from content providers. The project bottom-to-top approach is as-sured thanks to the identification of the content providers in the final users, and tothe IT modelling directly built upon their realities.Results: Through best practice principle, a European Radiological TransNomen-clature (ERTN) has been defined. A platform prototype has been developed. It iscomposed of two modules, the first transcodes the different nomenclatures andallows the second to perform benchmarks (on prices, volumes of activities, tech-nologies, etc). among the healthcare providers belonging to the network. Theprototype offers to MD and patients the radiological appropriateness guidelines.Conclusion: The platform will be able to communicate directly with the hospitalinformation system. These tools will allow health professionals and patients toaccess, through web technology, comparable information in the radiological sec-tor, and so be of use for clinical studies, department management and for steer-ing the patient to the appropriate health services.

B-494 10:57

Evaluation of formal quality of radiological educational resources on thewebT. Adla, P. Kasal, M. Hladíková, A. Janda; Prague 5/CZ ([email protected])

Purpose: Expansion of use of the internet has brought genesis of many webeducational resources, including those directed at radiologists. The quality of eachindividual website varies considerably. The aim of the study was to investigateformal qualities among radiological educational resources.Methods and Materials: The main inclusion criterion was radiological educa-tional content of a website (whole or a considerable part). Secondly, resourceswere selected with respect to covering a whole spectrum of online CME resourc-es (from big portals to one-man-sites). 10 resources were selected. Rankmed.czwas used for the evaluation, which is based on internationally accepted stand-ards. Websites were evaluated by 20 criteria, which are divided into 4 groupsdescribing the main formal qualities of website: Presentation, navigation, func-tions and credibility.Results: Overall average score was 59.8 points from 100 available (min. 38.0,max. 76.5). Regarding the groups of criteria, presentation contributed to the aver-age score with 20%, navigation 21%, function 23% and credibility 36%. Averageresults for particular types of resources were: portals 65.4, databases of cases63.8, online books 59.5 and one-man-sites 47.7 points.Conclusion: Hypothesis of the presence of measurable quality differences amongtested websites was confirmed. Surprisingly the average credibility of websiteswas relatively high (making up more than one third of the average score). Difficul-ty in creating websites and keeping them up-to-date is evident - the best scorewas achieved by websites which are provided by professional teams, as opposedto one-man-sites and websites provided by small groups.

B-495 11:06

e-Learning is a well-accepted tool in supplementary training among medicaldoctors. Experience of obligatory radiation protection training in healthcare in FinlandT.H. Autti, H.J. Autti, T. Vehmas, V. Laitalainen, L. Kivisaari; Helsinki/FI([email protected])

Purpose: To evaluate the possibilities of internet-based supplementary radiationprotection training fulfilling the requirements of the directive (MED/97/43/Euratom).

250 C D E FBA

Scientific Sessions

G

Methods and Materials: All 324 medical doctors participating in the internet-based radiation protection course between 1.4.2004 and 31.5.2005 were includ-ed. The course included open questions and simulated patient cases(www.prewise.fi/radiationsafetytraining). 210 participants (96 female, 114 male,aged 25-64 years) fulfilled the questionnaire (a Likert-type format with five poten-tial responses) including questions regarding their expectations, the scope, sched-ule and the benefit of the course. In addition, we asked if they found using theinternet to be a difficult learning method, whether e-learning saved time and wheth-er they felt that they learned better compared to conventional lectures.Results: 19% of the respondents had previous experience in e-learning. Theyfound e-learning to be easy (74%) and timesaving (61%). 57% stated that theylearned better using e-learning compared to conventional lectures. 91% foundthat the course fulfilled their expectations. 84% stated that they benefited fromthe course and 94% were willing to study using the internet in the future. Whenusing multiple ordinal regression no sex or age differences in answers were found.A significant positive Spearman correlation was found between the participants'expectations and the questions regarding the scope (p < 0.001) of the course,the benefit derived from the course (p < 0.001), the time-saving aspect (p < 0.001)and better learning (p < 0.002).Conclusion: e-Learning seems to be a well accepted and practical learningmethod for radiation protection training.

B-496 11:15 !eLearning content creation and distribution: A scenario using web services,standard office applications and SCORMJ. Hohmann, T. Schaaf, H. Kunz, T. Tolxdorff, K.J. Wolf; Berlin/DE([email protected])

Purpose: Wide distribution and re-use of standardised eLearning content is es-sential for today's education strategies. The aim of our scenario is to show thatwhile using open international standards and standard interfaces eLearning con-tent creation can be almost as easy as writing a text document and therefore canbe done by all knowledge providers who are familiar with standard office applica-tions. Nevertheless for installation and maintenance of the system IT profession-als are needed.Methods and Materials: The Sharable-Content-Object-Reference-Model(SCORM) defines a standard for eLearning objects and is strongly required forthe mentioned purpose. The workflow implies a teaching file database using XML-WebServices, Microsoft's PowerPoint®-application with a Simple-Object-Access-Protocol (SOAP)-adapter and the Producer-for-PowerPoint extension to generateSCORM-compatible eLearning objects. Alternatively we used the Thesis® Pro-fessional 2.5 Software which has additional features. To evaluate the resultingdocuments we used a freeware SCORM player and the OpenSource LMS "moo-dle".Results: In our scenario the knowledge provider of eLearning content just needsto be familiar with the commonly used PowerPoint application. The teaching filedatabase is seamlessly integrated in the user workflow via WebServices. Theoutput of the Producer/Thesis Software is usable in SCORM-ready learning man-agement systems (LMSs).Conclusion: While using de-facto standard office applications together with theinherent interoperability features of international standardised interface proto-cols and data formats eLearning documents can be easily created by most of theknowledge providers themselves and not only by IT professionals. Therefore thescenario we present might be a solution to provide the emerging LMS with itsnecessary content.

B-497 11:24 !A remote access training system for the detection of colorectal polyps atcomputed tomography colonographyV. Luauté; Dublin/IE ([email protected])

Purpose: Computed tomography colonography (CTC) is emerging as an alter-native to conventional colonoscopy (CC). However CTC is not yet in widespreaduse due in part to the lack of suitably trained radiologists. We have developed anovel remote access system to train radiologists for colorectal cancer screeningusing CTC.Methods and Materials: In order to ensure that radiologists gain the relevantexperience without the need for any specialist equipment or software, we optedfor designing a system that is accessible via the internet using a standard brows-er. The interface lets the user locate and characterise polyps with the help ofappropriate tools such as windowing, polyp measurement and zooming. Eachuser has an account in order to allow monitoring of their training. They can alsorun an automatic evaluation of their work based on gold standard informationpreviously gathered from specialists.

Results: Our system has been developed using a combination of Java servletsand applets. The interface allows the user to start working on the initial imageswhile the rest of the dataset is being transferred. The system is able to offer fastresponse time. An iteration of lumen tracking using the system takes approxi-mately 40 seconds.Conclusion: We have developed a system which allows a trainee radiologist togain proficiency in an emerging colorectal cancer screening technique. This re-search has yielded an operational system that meets the needs of remote ac-cess users.

B-498 11:33

Performance of an integrated web-based RIS/PACS system to facilitateeffective mobile diagnostic imaging services in the UKC. Stavrianou, P. Tually, J. Walker; Wimbledon/UK([email protected])

Purpose: Faster delivery of medical services and an increase in capacity hasbeen the focus of a recent NHS scheme which has been implemented to addressthe excessive waiting lists for diagnostic services in the UK. The mobile opera-tions will enable delivery of tests to regions of greatest need but require highbandwidth networks and sophisticated software applications to distribute largedata volumes, interpret results and integrate with existing hospital systems. Toaccomplish this, a comprehensive, digital, web-based patient information systemhas been installed to manage patient care. The technology allows on-line sched-uling, exam requisition, patient tracking and off site sub-specialist opinion in real-time. We describe the predominate factors that affected the implementation,acceptance, technical parameters to attain optimal system functionality andpresent results from a formal evaluation on user satisfaction.Methods and Materials: Four interconnected methods categories of user ac-ceptance were examined approximately 6 months after use. User acceptanceresponses from four user groups were measured on a five point Likert scale, withdata input on a web browser terminal.Results: Frequency, Likert mean and standard deviation (SD) items were testedusing SPSS software. A quantitative chart of response frequency (%) for eachcategory was generated. System Quality, Information Quality, Service Quality,and User Satisfaction all demonstrated significant user acceptance.Conclusion: The patient information systems were well received by all users,reducing the administrative task load and improving clinical workflow efficien-cies. Despite novel techniques to maximise the use of current services, the lackof a modern network has impeded the full potential of RIS/PACS.

10:30 - 12:00 Room K

Pediatric

SS 1012Musculoskeletal imagingModerators:H. Ringertz; Stockholm/SER.R. van Rijn; Amsterdam/NL

B-500 10:30

Remotely accessible standard database for skeletal maturityK. Witko, A. Gertych, S. Pospiech-Kurkowska, E. Pietka; Gliwice/PL([email protected])

Purpose: Assessment of skeletal development is frequently performed in paedi-atric radiology. The most commonly applied clinical method is the atlas matchingmethod which compares a hand radiograph with a film-based hand atlas. A com-puter assisted bone age assessment method has been developed. It matchesautomatically a patient image with the digital hand atlas.Methods and Materials: Four components are integrated in order to obtain acomputer aided diagnosis system. First, a set of standard images has been se-lected from 500 clinical cases. A comparison of each image with hand patterns,included in an atlas of radiologically approved standard, has pointed the closestmatch in each age group. Then, the application software extracts radiologicalfeatures describing the size, shape, and density of epiphyses and metaphyses,as well as parameters describing the stage of epiphyseal fusion. Next, standardimages in diagnostic and compressed resolution along with their extracted fea-tures are organized into a hand e-atlas. At the radiological image analysis stagefeatures are extracted and compared with the e-atlas to find the closest match.Finally, a graphical user interface has been designed. It views the image analysis

B 251C D E FA

Su

nd

aS

un

da

Su

nd

aS

un

da

Su

nd

ayy yyy

Scientific Sessions

G

results and displays 2 closest matches from the e-atlas for the radiological ap-proval.Results: A set of 250 clinical hand images has been tested. The agreement be-tween radiological and computerized assessment has reached 90%.Conclusion: The CAD system offers a fast and objective tool able to assist radi-ologists in performing the bone age assessment. In case of failure, a manual e-atlas-based assessment is available.

B-501 10:39

Assessment of skeletal age in children and adolescents using radiussonography in comparison to standardised radiographyH.-J. Mentzel1, C. Vilser1, M. Eulenstein1, T. Schwartz2, J. Böttcher1, L. Tsoref2,E. Kauf1, W.A. Kaiser1; 1Jena/DE, 2Tel Aviv/IL([email protected])

Purpose: Determination of skeletal development in children is important. Themost used evaluation method is to study left hand X-ray based on the standardsdepicted by Greulich & Pyle (G&P). The aim of our study was to compare theaccuracy of a new sonographically based method with the standard method.Methods and Materials: 160 consecutively evaluated children and adolescents(77m, 83f) who received a X-ray of the left hand were evaluated. Ultrasound ex-amination of the same hand was performed on the same day using the Bon-Age™ system (Sunlight Medical Ltd., Israel).This system evaluates the relationship between the velocity of the wave and growthusing gender and ethnicity based algorithms.Three investigators (U1-U3) analysed the X-ray and assigned bone age basedon the G&P atlas.Results: In 152 patients BonAge™ measurement was successfully. The correla-tion between the three investigators using the G&P method was between 0.977and 0.980; correlation between BonAge™ and the investigators (U1-3) was 0.902and 0.920. The paired t-test showed no significant differences between BonAge™and the investigators and also for the comparison between U1 and U2. Therewere significant differences between U1 vs. U3 and U2 vs. U3 (p < 0.05).Conclusion: The BonAge™ device demonstrates the ability to produce a suffi-cient assessment of bone age using an ultrasound method. Advantages of theultrasound device are objectivity, lack of ionizing radiation, and easy accessibili-ty. In the case of first investigations, X-ray is necessary to look for dissociatedskeletal age, dysplasia, and mineralisation.

B-502 10:48

Sonographic evaluation of nasal bone fractures in childrenS.W. Lee, J.Y. Whang; Seoul/KR ([email protected])

Purpose: To describe the sonographic findings of nasal bone fractures and eval-uate the diagnostic value of sonography (US).Methods and Materials: Nineteen children who had sustained nasal traumaunderwent plain radiography, US and CT and the CT findings were used as thegold standard. The US findings were retrospectively reviewed and compared withthese of other modalities.Results: Sonographic findings of nasal bone fractures were irregularity (n = 3),discontinuity (n = 12), displacement with discontinuity (n = 14)of the thin smoothechogenic line formed by ventral cortex of the nasal bone. Twenty-nine nasalbone fractures (19 patients) were identified by CT and US but only 14 by plainradiographs. US and CT demonstrated unilateral involvement of nasal bone frac-ture in seven bones and bilateral involvement in 11, while plain radiograph re-vealed unilateral involvement in four and bilateral involvement in five.Conclusion: The US findings of nasal bone fractures were similar to those of CTin detecting the location and degree of depression or displacement of nasal bonefractures. We believe that US is very useful for the evaluation of such fractures,especially in children with equivocal abnormalities on plain radiographs.

B-503 10:57

Fluroscopy in paediatric fractures: Setting a local diagnostic reference levelM. Jain1, A. Pillai2; 1Liverpool/UK, 2Wishaw/UK ([email protected])

Background: The Ionizing Radiations (Medical Exposure) Regulation 2000 hasmade it mandatory to establish Diagnostic Reference Levels (DRLs) for all radio-logical examinations. Currently available National and European guidelines haveno data on orthopaedic trauma.Purpose: We attempt to propose a provisional local Diagnostic Reference Levelfor common orthopaedic trauma screening procedures in paediatric patients aged5 -10 years.Methods and Materials: The Dose Area Product (DAP) in 865 paediatric trauma

examinations was analyzed. Median DAP values and the median screening timefor each examination type along with quartile values for each range are presented.Results: In the upper limb, screening of the elbow had maximum exposure witha median DAP value of 1.21 cGycm2. Median DAP values for forearm and wristexaminations were 0.708 cGycm2 and 0.538 cGycm2 respectively.In lower limb, tibia & fibula examinations had a median DAP value of 3.23 cGy-cm2 followed by ankle examinations with a median DAP of 3.10 cGycm2. Screen-ing times were longer in the lower limb compared to the upper limb. Dosagevalues between the first and third quartile for each distribution can be used as aprovisional DRL range for the specific examination type.Conclusion: A DRL is a measure of patient dose and serves as a quantitativeguide to optimize patient protection. The dose data presented in the study can beused to establish a provisional local DRL range.

B-504 11:06 !Evaluation of low dose multidetector CT (MDCT) of the sinuses in childrenand comparison of effective dose with standard X-rayT.H.E. Mulkens1, C. Broers1, S. Fieuws2, P. Bellinck1, J.-L. Termote1; 1Lier/BE,2Leuven/BE ([email protected])

Purpose: Evaluate the use of low dose MDCT of sinuses in children and com-pare effective dose (E) with standard X-ray.Methods and Materials: In standard X-rays of sinuses with 75 kV and 3 mm Alfiltration in 100 children (mean: 4.5 years) the Dose-Area-Product (DAP, mGy.cm²)was converted to E according to coefficients of the British National RadiologicalProtection Board. 200 children (mean: 5.9 years) underwent low dose MDCT ofsinuses with 6- and 16-MDCT (Siemens, Germany):6- MDCT (n = 100): 80 kV, 6 x 0.5 mm collimation, 0.6 sec rotation, pitch 1.5 and15 to 25 mAs (according to age);16- MDCT (n = 100): 80 kV, 16 x 0.75 mm colli-mation, 0.5 sec. rotation, pitch 1.5 and 17 to 25 mAs. E for MDCT was calculatedfrom conversion from Dose-Length-Product (mGy.cm).Results: 15% of children had acute sinus disease and 85% recurrent or chronicsinusitis. Wilcoxon test (p = 0.3) showed no statistically significant difference be-tween E for X-ray (0.0525 mSv) and low dose CT (0.0524 mSv). Low dose CTgives diagnostic image quality at very low dose. Interobserver agreement wasexcellent between three readers: 0.89 - 0.96. All negative scans (16%) were scorednegative by three readers. Mean scan time was only 2.1 seconds and 9.5 sec-onds for 6- and 16-MDCT, whereby sedation becomes unnecessary. Evaluationof adenoids and middle ears is possible in the same examination.Conclusion: Low dose MDCT of sinuses in children can be used as a alternativefor standard X-ray, because of high accuracy at a comparably low dose.

B-505 11:15

US and MR imaging of chondral and osteochondral erosions and pannusformation in juvenile chronic arthritisA.K. Karpenko, V. Purin, T. Panova; St. Petersburg/RU ([email protected])

Purpose: Therapeutic strategy in various forms of JChA attracts interest to effec-tive evaluation of pannus formation and osteochondral erosions in immature skel-eton. US and MR imaging are preferable imaging modalities. The purpose was toassess the most effective imaging modality in evaluation of chondral injuries inchildren with JChA.Methods and Materials: 27 children (68 joints) aged 3-14 years with JChA wereexamined. Symmetrical US with power Doppler in affected hips and knees wasobligatory for evaluation of arthritis activity, pannus and chondral and osteochon-dral formation. MR imaging was performed in 23 patients in order to confirm find-ings also symmetrically using 1.5 T MR imaging with standard Pd, T1, T2 SE (FatSat) and GRE. Contrast enhancement and 3D SPGR were used for pannus andchondral changes.Results: In 23 joints there was immature epiphyseal irregularity, especially inchildren with more than 1 year of disease activity. In small persons with JChAreactive epiphyseal chondritis was hardly differentiated from chondral defects. Insuspicion of erosions 3D SPGR MR imaging technique with contrast enhance-ment showed real chondral defects in bare areas only in 11 cases. Other epiphy-seal irregularities evaluated by US we interpreted as chondral reactive hyperemiachanges and their irregular overgrowth.Conclusion: US, MR imaging are necessary methods for evaluation of chondrallesions. 3D SPGR with contrast enhancement shows registration of erosions dueto pannus formation and is necessary method for their assessment.

252 C D E FBA

Scientific Sessions

G

B-506 11:24

MR imaging in measuring the thickness of hyaline cartilage cap inosteochondroma with histological correlation: New balanced FFE sequencecompared to cartilage-sensitive 3D gradient echo sequencesT. Schmidt, G. Staatz, B. Hermanns, C. Hohl, P. Haage, D. Honnef,R.W. Günther; Aachen/DE ([email protected])

Purpose: MR imaging measurement of the thickness of cartilage cap of osteo-chondroma - a sign of malignant transformation - is regarded as the most accu-rate non-invasive modality. We evaluated the new BalancedFFE sequence(balFFE) in comparison to an established double-echo gradient echo sequence(3D-GE) in measuring cap thickness with histopathological correlation.Methods and Materials: In 17 patients with solitary or multiple osteochondroma,we performed either 3D-GE (n = 14), balFFE (n = 9) or both MR imaging se-quences (n = 6) - in case of multiple lesions, a single lesion of each patient wasmeasured. Lesions were benign, proven by histopathology. Thickness of carti-lage was measured perpendicular to cap surface. MR imaging was correlated tohistopathological measurements. Statistical analysis was done using 97.5% con-fidence intervals for mean of differences between MR imaging and histopatholo-gy.Results: Histologically determined cap thicknesses ranged from 2 - 4 mm (mean3 mm). BalFFE measurements (n = 9) could be compared to histology with meanof differences 0.9 mm (CI -0.7 mm through 2.5 mm). Mean of differences betweenfirst echo of 3D-GE and histology (n = 14) was -0.11 mm (CI -1.1 mm through0.9 mm), between second echo and histology (n = 14) it was -0.3 mm (CI -1.2 mmthrough 0.7 mm). All confidence intervals included zero (0 mm) indicating an ex-cellent correlation to histology.Conclusion: The new balFFE MR sequence corresponds well to histopatholog-ical thickness of cartilage caps of osteochondromata. The visualization of carti-lage caps with balFFE is reliable so that this new sequence can be used as a fastand high resolution alternative to established cartilage sensitive sequences.

B-507 11:33

MR imaging of the temporomandibular joint in children with juvenileidiopathic arthritisP.N. Margariti, A. Karali, A. Karamanos, F. Papadopoulou, A. Siamopoulou,M.I. Argyropoulou; Ioannina/GR ([email protected])

Purpose: Temporomandibular joint (TMJ) is commonly affected in juvenile idio-pathic arthritis (JIA) patients. Lacking in early symptoms and signs, TMJ arthritisresults in craniofacial growth disturbances. Our purpose was to evaluate by MRimaging the TMJ of JIA patients.Methods and Materials: 36 patients (13.1 ± 7.2 years) with JIA (oligoarthritis15, polyarthritis 11, systemic 10), were enrolled in the study. The MR imagingprotocol consisted of: sagittal T2-weighted turbo spin-echo (2019/90) scans, sag-ittal and coronal, plain and contrast-enhanced, fat-suppressed, T1-weighted spin-echo (425/15) scans and gradient-echo (65/14/20°) scans with incremental mouthopening. A grading score was used for the evaluation of the condyle (1=normal,2=bifidity, 3=absence of the anterior half of the condyle), articular eminence (AE)(1=normal, 2=shallow), articular disc (AD) (1 = "bow-tie" shape, 2 = abnormal),acute pannus (1=normal, 2= pannus), intraarticular fluid (1=normal, 2=fluid) andthe range of condylar motion (RCM) (1=90º, 2=75º, 3=45º, 4=0°).Results: 70 joints were examined. Higher grades were more frequent in system-ic arthritis (condyle3=55%, AE2=60%, AD2=65%, RCM4=5%) than in oligoar-thritis (condyle3=17.9%, AE2=0%, AD2=0%, RCM4=0%) and polyarthritis(condyle3=27.3%, AE2=13.6%, AD2=18.2% RCM4=0%) (p < 0.001). There wasno significant difference between systemic arthritis, oligoarthritis and polyarthri-tis in the frequency of pannus (25%, 52.4%, 45.5% respectively) and intraarticu-lar fluid (14.3%, 10.7%, 13.6% respectively). Patients with abnormal condyleshad longer disease duration (11.8 ± 5.1 years) than patients with normal con-dyles (5.5 ± 6.12 years) (p < 0.001).Conclusion: MR imaging reveals TMJ involvement of different severity and fre-quency in the 3 subtypes of JIA. MR imaging may be useful for early detection ofthe inflammatory process before irreversible lesions occur.

B-508 11:42

Role of whole body MR imaging in diagnosis of osteonecrosis in childrenwith intensive chemotherapyM. Beer, E. Hoppe, H. Girschick, K. Darge; Würzburg/DE([email protected])

Purpose: Osteonecrosis (ON) in children is a severe complication of combina-tion chemotherapy including high doses of steroids. Whole body MR imaging

(MRI) allows detection of ON, however, studies are missing which determine thebenefit of single sequences or of looking at the changes during follow-up.Methods and Materials: Up to now, 5 children who complained of sudden onsetof bone pain were examined. Imaging included an X-ray of the painful region andadditionally an MR imaging scan of the whole body (T2w TIRM and T1w GEsequences -pre and post-contrast (the latter with fat sat)). 4 children had at leastone follow-up examination during 6 months.Results: None of the baseline X-ray images showed pathological changes. MRimaging revealed typical ON bone marrow lesions in 3/5 children. Additional tothe sites with bone pain, MR imaging revealed other ON locations. All of thelesions had bone edema and contrast enhancement. Follow-up examinationsshowed slight reduction of bone edema during 6 months.Conclusion: Whole body MR imaging allows rapid detection of ON, much earlierthan plain X-ray diagnostics. MR imaging reveals additional ON even before clin-ical symptoms occur. T2w TIRM seems to be the most sensitive sequence.

B-509 11:51

Development of fetal adipose tissues as seen with fetal magnetic resonanceimagingP.C. Brugger, D. Prayer; Vienna/AT ([email protected])

Purpose: Adult fat is characterized by T1-weighted hyperintensity in magneticresonance imaging (MRI). The present study investigates the development ofthese MR imaging signal characteristics in fetal adipose tissues.Methods and Materials: 150 fetuses [20-38 gestational weeks (GW)] were im-aged on a 1.5 T magnet using T2-, T1-weighted and steady-state free-preces-sion (SSFSE) sequences. Signal intensities of adipose tissues on differentsequences were assessed with respect to topographical regions.Results: T1-weighted hyperintensity of fetal adipose tissues was not evident untilthe beginning of the 23rd GW when the buccal fat pad was the first to display it. By25 GW moderate hyperintensity appeared in the ischio-rectal fossa and face, theretropatellar fat pad and the deep extensions of buccal fat pad could be recog-nized. By 27 GW subcutaneous tissues showed hyperintense signals, followedby popliteal (28 GW), intraorbital (30 GW) and perirenal fat (32 GW). SSFSEsequences showed marked hypointensity of buccal fat between 25-32 GW. FromGW 32 onwards, chemical shift artefacts delineated subcutaneous fat from der-mis and underlying structures.Conclusion: Fetal adipose tissues gain their T1-weighted hyperintense signalproperties in a regular temporary sequence depending on their composition andtopography. Though abundant, perirenal fat displays no T1-weighted hyperinten-sity until 32 GW, since it is made of brown adipose fat in the fetus which hasdifferent signal properties. These findings may have a bearing on the prenatalMR imaging detection of (intracranial) lipomas, and assessment of fetal macro-somia or intrauterine growth restriction. SSFSE sequences may be an alterna-tive in detecting lipoma.

10:30 - 12:00 Room L/M

Neuro

SS 1011Functional MRIModerators:D. Bilecen; Basle/CHP.H. Nakstad; Oslo/NO

B-510 10:30

Evidence of the tonotopical organization within the human primary auditorycortex: fMRI studyM. Orzechowski, T. Wolak, E. Piatkowska-Janko, P. Bogorodzki, R. Kurjata,W. Szeszkowski, L. Sliwa, K. Kochanek, O. Rowinski; Warsaw/PL([email protected])

Purpose: The aim of the study was to investigate the tonotopical organizationwithin primary auditory cortex using fMRI technique.Methods and Materials: For each of 22 healthy subjects with good hearing (20males, 2 females, all right handed), aged 23.6 ± 0.7, one functional session wasacquired. The auditory stimuli paradigm was designed as 4 simultaneous orthog-onal pseudorandom sequences of 12 sec blocks and applied in one run. Each ofthe tasks was a modulated sinusoidal signal with base frequencies:250 Hz ± 200 Hz, 1000 Hz ± 500 Hz, 2000 Hz ± 500 Hz and 4000 Hz ± 1500 Hz.Statistical parametric maps were generated using SPM2. Single subject (mean

B 253C D E FA

Su

nd

aS

un

da

Su

nd

aS

un

da

Su

nd

ayy yyy

Scientific Sessions

G

and variance) and group analysis (one sample T-test, ANOVA) were calculated.Regions of maximum activations for each frequency range were extracted andmapped on a flattened primary auditory cortex image.Results: The random effect analysis showed strong activations in different partsof auditory cortex for each frequency range respectively (250 Hz: max. t-valueleft hemisphere (LH) 6.5, right hemisphere (RH) 6.3, 1000 Hz LH 9.1, RH 7.6,2000 Hz LH 5.4, RH 3.8, 4000 Hz LH 8.2, RH 7.9, all values with p < 0.001). Theactivations from lower to higher frequencies were found in lateral to medial direc-tion. Maximum probability map was created for localization of frequency band-widths on primary auditory cortex.Conclusion: The experimental data seems to show a tonotopical structure ofprimary auditory cortex. Lower values of T-statistic for range 1.5-2.5 kHz mightbe caused by stimuli masking with scanner noise.

B-511 10:39

Tonotopic plasticity effects in the auditory cortex in unilateral tinnituspatients: An fMRI studyS. Kovacs1, M. Smits2, D. De Ridder3, R. Peeters1, P. Van Hecke1, S. Sunaert1;1Leuven/BE, 2Rotterdam/NL, 3Edegem/BE ([email protected])

Purpose: To study the tonotopic organization of plasticity effects in the auditorycortex (AC) of unilateral tinnitus patients when stimulated with their experiencedtinnitus frequency or another frequency by means of fMRI.Methods and Materials: Fifteen Patients were scanned on a 3 T MR imagingscanner. A T2*w EPI silent gap sequence (TR = 5s) was used during the stimula-tion paradigm. The stimulation paradigm consisted of a blocked design of 18 ep-ochs in which white noise filtered through a narrow (freq ± 500 Hz) or widebandpass (freq ± 1000 Hz) and tones were presented 3 times binaurally each.The stimuli were either the tinnitus frequency experienced by the patients e.g.8000 Hz (TF-session) or a different frequency e.g. 1000 Hz (OF-session). Datawere analyzed using SPM99 software and activation patterns were visualized.Quantitative analysis consisted of calculating an activation ratio of activated vox-els in the AC and assessing differences between the left and the right side with apaired samples t-test.Results: Significant fMR signal change was found bilaterally in the AC in bothsessions. Signal change was symmetrical in patients when exposed to the OF-session. When exposed to the TF-session, a decrease of signal change (plastic-ity effect) was visualized in the contralateral AC. Comparing the activation ratio inthe left and right AC showed a significant difference during the TF-session, butnot during the OF-session.Conclusion: By means of 3 T fMRI we were able to demonstrate tonotopic plas-ticity effects in unilaterally affected continuous tinnitus patients. Acknowledge-ment: Supported by the ECR-Research Fund.

B-512 10:48

Brain activity in the visual pathways using fMRI in correlation withvolumetric MR imaging in MCI patients and healthy elderly subjectsC.M. Born, T. Meindl, S. Britsch, A. Bokde, M.F. Reiser; Munich/DE([email protected])

Purpose: During a face matching task healthy subjects selectively activate corti-cal areas along the ventral visual stream, whereas patients with amnestic mildcognitive impairment (MCI) show activation along the ventral and the dorsal vis-ual stream. We studied the morphological substrate of this differential activationof visual areas during face matching.Methods and Materials: Sixteen subjects with amnestic mild cognitive impair-ment (MCI) and 17 healthy subjects underwent structural and functional MR im-aging examinations. The object matching task consisted of two faces presentedsimultaneously and participants were asked to decide on each trial if a pair offaces were identical or not. The data were analyzed using AFNI (Cox, 1996) andFSL (FMRIB Software Library). For morphometric analysis a high resolution T1w3D-MPRAGE sequence was performed. The processing of the 3D sequencesfollowed a protocol implemented through Statistical Parametric Mapping (SPM).Results: Both in MCI and healthy subjects greater fusiform activation in fMRIwas correlated with better preserved grey matter volume along the ventral visualpathway. Smaller fusiform activation was correlated with better preserved corti-cal grey matter along the dorsal visual pathway in both groups.Conclusion: The morphological separation between ventral and dorsal visualstreams may result from system-specific neurodegeneration in MCI. The inter-pretation of activation pattern in fMRI should take into account the degree ofcortical atrophy along the entire functional system underlying the task under study.

B-513 10:57

Diffusion tensor imaging and diffusion tensor fiber tractography study oflanguage function areas and fiber pathology changes of Broca's aphasia by3 T MR imagingX. Sun1, J. Dai2, S. Li2, R. Pang1, J. Ma2, H. Chen2; 1Kunming/CN, 2Beijing/CN([email protected])

Purpose: To present the findings in Broca area and fiber under pathological con-ditions of Broca aphasia using 3 Tesla diffusion tensor imaging (DTI) and diffu-sion tensor fiber tracking (DT-FT).Methods and Materials: Axial diffusion tensor images covering the entire brainvolume were obtained in thirty Broca aphasia patients. Images were acquiredwith two b factors (0, 1500 s/mm2). The fractional anisotropy (FA) map of Broca,arcuate fasciculus were used to measured FA values. Siemens DTI and DT-FTsoftware was used for post-processing to measure the FA value and display thecourse of Broca 45 area and 44 area on both sides.Results: Broca aphasias result from the damage of Broca function area andpathway of Broca function fibers. Broca area FA values were significantly differ-ent between left and right (t-test p < 0.05). Broca area fibers were also shown tobe damaged in Broca aphasias. The fibers of Broca 45 and 44 areas and FAvalues between the left and right side were significantly different (p < 0.05). Bro-ca's down-through fibers were usually broke or translocated, and the relationbetween Broca and arcuate fasciculus became loosened or dissociated from eachother. Broca fibers up to the precentral gyrus and the lateral foreside frontal lobedecreased. Broca aphasia's left arcuate fasciculus was damaged.Conclusion: Not only the Broca area was shown to be damaged, also the relat-ed structures related to Broca aphasia were damaged. The DTI and DT-FT re-sults might be used to evaluate the language dysfunction prognosis.

B-514 11:06

Covert language production could not touch the nature of stuttering: AnfMRI studyY. Yang, K. Li, D. Peng, C. Lu; Beijing/CN ([email protected])

Purpose: Our previous fMRI studies showed that the striatal-thalamus-corticalneural circuit contributes exclusively to stuttering. While it was responsible forsequencing and timing of movement in the traditional view, could speech taskswithout movement touch the nature of stuttering? We wanted to test this hypoth-esis by fMRI.Methods and Materials: Nine normal controls and ten stuttering subjects werescanned in a 1.5 T MR imaging scanner. While the subjects named pictures cov-ertly or overtly in two separate runs (block design), fMRI images were acquired.The data was processed using the Analysis of Functional NeuroImages (AFNI)programs.Results: In agreement with previous results, we detected the abnormal activityof the striatal-thalamus-cortical circuit in the stuttering brains in the overt namingcondition. There were also some areas in the stutterers' brains exhibiting failingdeactivation, including some superior and middle frontal areas. In the covert nam-ing condition, the areas which manifested differences between the two groupswere bilateral superior and middle frontal cortex, left superior temporal cortexand right insula. Although, to some extent, it indicated different activity patternsbetween the two groups related to pre-articulatory procedures, it failed to revealthe important aspects of dysfunction of the striatal-thalamus-cortical circuit in thestuttering brain.Conclusion: Because of the complexity of stuttering pathology and its close re-lationship with language production, it was necessary to involve almost everycritical aspect in the testing tasks so as to touch the nature of stuttering. Taskssuch as covert naming might fail this criterion.

B-515 11:15

Cerebral blood flow (CBF) changes during continuous hand motion usingcontinuous arterial spin labelling (CASL) technique: Attempts at mapping ofthe brain motor cortexJ. Jeong, S. Kim, J. Lee, E. Kim; Seoul/KR ([email protected])

Purpose: The CBF changes during continuous hand motion were measured us-ing CASL MR imaging to evaluate whether mapping of the motor cortex is possi-ble using it.Methods and Materials: Multislice CASL MR imaging was performed at rest andduring continuous hand motion in a 3.0 T scanner. Six subjects of twelve healthyindividuals performed handgrip and the others performed finger tapping of thedominant hand with an average frequency of 1 motion/sec. The CBF changes ofactivated motor cortex, contralateral motor cortex and ipsilateral sensory cortexwere calculated at rest and during hand motion.

254 C D E FBA

Scientific Sessions

G

Results: The mean percent CBF changes at activated motor cortex were signif-icant during handgrip (46 ± 23%, p = 0.006) and finger tapping (48 ± 17%,p = 0.001). The mean percent CBF changes at the contralateral motor cortexwere 9 ± 6% during handgrip and 7 ± 23% during finger tapping. The mean per-cent CBF changes at the ipsilateral sensory cortex were 0 ± 17% during hand-grip and 21 ± 23% during finger tapping. The CBF change at activated motorcortex showed a negative correlation during handgrip motion (r = -0.92) and apositive correlation (r = 0.56) during finger tapping to the contralateral motor cor-tex. The CBF change at activated motor cortex showed a positive correlation tothe ipsilateral sensory cortex during finger tapping (r = 0.70).Conclusion: The significant CBF change (> 40%) during continuous hand mo-tion detected by CASL may suffice the mapping of motor cortex with this tech-nique. The negative correlation of the contralateral motor cortex during handgripmotion as opposed to a positive correlation during finger tapping suggests thathandgrip exercise should be a better motor task than finger tapping in motorcortex mapping.

B-516 11:24

The cortical representation of taste in the human brain: An MR imagingstudyG. Polonara1, M. Fabri1, G. Mascioli1, G. Tassinari2, G. Berlucchi2, T. Manzoni1,U. Salvolini1; 1Ancona/IT, 2Verona/IT ([email protected])

Purpose: To identify the human primary gustatory cortex using fMRI and to es-tablish whether this cortical representation is mainly ipsilateral, as in non-humanprimates, or bilateral, as recent neuropsychological studies of patients with cal-losal resection seem to indicate.Methods and Materials: Data were obtained from 9 healthy subjects. Functionalimages were acquired with a 1.5 Tesla machine. The stimulation protocol lasted5 min and consisted of 60 sec rest, 30 sec stimulation, 90 sec rest, 30 sec stim-ulation, and 90 sec rest. Gustatory stimulation was applied by placing a smallcotton pad soaked in a 1M NaCl solution (the salty stimulus), distilled water (theneutral stimulus) or 10% sucrose (the sweet stimulus) on the side of either hem-itongue. Data was analyzed with BrainVoyager and SPM2 software.Results: Salty stimulation of the left hemitongue evoked bilateral activation ofarea GI in all subjects, more pronounced in the ipsilateral hemisphere. Salty stim-ulation of the right hemitongue also evoked bilateral activation of area GI in allsubjects, but without appreciable side differences. Sucrose stimulation evokedcortical activation in both hemispheres in areas similar to those activated by thesalty stimulus. Neutral stimulation activated more posterior and more dorsal cor-tical regions likely corresponding to the tactile cortical representation of the tonguein the first somatic sensory area.Conclusion: These findings show that the organization of taste afferents in mandiffers from the organization described in non-human primates, and are in agree-ment with previous neuropsychological studies performed in control subjects andcallosotomized patients.

B-517 11:33

Differences in cerebral activation between females in distinct menstrualphases during viewing of erotic stimuli: An fMRI studyE.R. Gizewski, A. Baars, M. Schlamann, M. Forsting; Essen/DE([email protected])

Purpose: There is evidence that men experience more sexual arousal than wom-en, but also that women in mid-luteal phase experience more sexual arousalthan women outside this phase. Recently, a few functional brain imaging studieshave tackled the issue of gender differences of reactions to erotica. The questionof possible differences in reactions to erotica with women in distinct phases hasnot yet been answered from a functional brain imaging perspective.Methods and Materials: fMRI was performed in 22 female volunteers measuredtwice (mid-luteal and menstrual phase). Subjects viewed erotic film excerpts al-ternating with emotionally neutral excerpts in a standard block-design. Arousal toerotic stimuli were evaluated using standard rating scales after scanning.Results: Two-sample t-Test with uncorrected p < 0.001 values for a priori deter-mined ROIs involved in processing of erotic stimuli revealed differences betweenboth scanning times: Comparing women in mid-luteal phase and during theirmenses, superior activation was revealed for women in mid-luteal phase in theanterior cingulate, left insula, and orbitofrontal cortex. However, only the superioractivation in the orbitofrontal cortex and left insula was maintained using the sex-ual arousal as a covariate. No significant superior activation was revealed forwomen scanned during their menses when compared with women in mid-lutealphase.Conclusion: Our results indicate that there are few differences in cerebral activ-

ity during viewing of erotic stimuli with women in mid-luteal phases and outsidethe mid-luteal phase. Women in mid-luteal phase revealed a superior activationin some brain areas specific for emotional processing.

B-518 11:42

Functional MR imaging (fMRI) of acupuncture: Observation of stimulatingthe acupoint ST36 (Zusanli)F. Miao; Shanghai/CN ([email protected])

Purpose: To explore the experimental method of obtaining functional magneticresonance imaging (fMRI) evidence of changes in cerebral function by punctur-ing the stomach 36 [ST36 (Zusanli)] acupoint, and to discuss the mechanisms ofacupuncture to treat diseases.Methods and Materials: The experiments were performed on seventeen volun-teers. Using fMRI to obtain the experimental evidence when acupuncturing ST36.Image acquisition was performed on a 1.5 T superconductive clinical scanner.The fMRI acquisitions were obtained using gradient echo-EPI sequence;TR=3000 ms, TE=60 ms. The data was analyzed by using statistical parametricmapping (SPM). Cerebral blood flow was analyzed.Results: There are certain increases of cerebral blood flow in ipsilateral hypoth-alamus and bilateral temporal lobe, when acupuncturing ST36.Conclusion: Acupuncturing ST36 can lead to functional changes in vegetativenever center and temporal lobe, which is correlated with the therapeutic effectsof acupoint ST36.

B-519 11:51

Functional MR imaging studies of cognitive processing in patients withacute and remitted bipolar-I disorder: Analysis of treatment effectsT. Wolak, W. Szeszkowski, B. Loza, R. Kuriata, M. Orzechowski, P. Bogorodzki,K. Papierski, L. Krolicki, R. Andrysiak; Warsaw/PL ([email protected])

Purpose: The purpose of our work was to assess the usefulness of fMRI tech-nique for comparative study of bipolar-I patients before and after treatment.Methods and Materials: In a double-blind manner, after one-week washout, 30treated bipolar-I patients were compared before and after an 8-week trial. Norefractory or recently lithium-treated patients were included. 22 patients were indepressive, 4 in mixed, and 4 in manic phases (DSM-IV-TR). MADRS, YMRS,MDQ, and CGI ratings were compared with functional MR imaging results. Thecontrol group consisted of 15 healthy volunteers. The Stroop test was implement-ed as the fMRI stimulus. Final and initial fMRI brain activations were 'subtracted'and used for group analysis. Two sample T-test analysis as well as ANOVA anal-ysis were used to compare means and variations of activations.Results: Patients improved significantly in nearly all clinical ratings. The com-mon fMRI findings in the patient group were improvement in dorsolateral prefron-tal (T=5.5, p-value < 0.001) and dorsal cingulate gyrus (T=6.4, p-value< 0.001)activations. Signal increased both at cluster-level and voxel-level analyses, how-ever remission and positive cognitive response were also correlated with somereductions in global activation (subcortical, temporal and rear frontal areas). Therewas statistically significant correlation between the values of activation in regionsof interest and clinical rating tests (p-value < 0.001).Conclusion: (1) Clinical and cognitive improvements of bipolar-I patients wererelated to increases in dorsal cingulate gyrus and dorsolateral prefrontal activity,mostly in the right hemisphere. (2) Increase in activation level correlated withclinical ratings. (3) The fMRI results seem to be helpful in analysis of treatmenteffects.

B 255C D E FA

Su

nd

aS

un

da

Su

nd

aS

un

da

Su

nd

ayy yyy

Scientific Sessions

G

10:30 - 12:00 Room N/O


SS 1013Optimisation of digital systemModerators:H.H. Hjemly; Tolvsrød/NOF.R. Verdun; Lausanne/CH

B-520 10:30

Supplementary recommendations for applying the "European protocol forthe quality control of the physical and technical aspects of mammographyscreening" to digital systemsR. VanMetter1, L. Fletcher-Heath2, M. Heath2; 1Washington, DC/US,2Rochester, NY/US ([email protected])

Purpose: Image quality assurance for digital mammography screening systemsis now widely recognized. The European protocol for the quality control of thephysical and technical aspects of mammography screening (EPQCM) prescribescriteria for image quality, including "threshold contrast visibility." This study quan-tifies the effects of key variables, including the choice of CDMAM phantom andmethodological choices unspecified in EPQCM.Methods and Materials: FFDM and CR images of the CDMAM phantom wereacquired and scored by trained observers. The presentation context was modi-fied to a display protocol that automatically isolates targets, optimizes window/level, and records visual scores. Visual scores were supplemented with syntheticscores to compare nearest-neighbor correction and maximum-likelihood-estima-tion analyses.Results: Contrast sensitivity depends upon the viewing environment and pres-entation context. Optimized presentation software lowers threshold contrast ofsmall diameter targets by as much as 26%, while reducing the scoring time. How-ever, substantial inter- and intra-observer variability remain important limitations.A strict 4-AFC paradigm lowers threshold contrasts, in conjunction with reducingthe subjectivity of the test. These results for human visual scoring were support-ed by studies with synthetic data. Finally, preliminary studies reveal systematicdifferences between two CDMAM phantoms, with one phantom showing 20%lower threshold contrast visibility when averaged over all target diameters from0.1 to 2.0 mm.Conclusion: The EPQCM guidelines omit critical experimental details and there-by suffer several substantial sources of variability that can be mitigated by follow-ing optimized techniques. However, other important sources of variability cannotbe mitigated. Supplementary recommendations to reduce test variability are pro-vided.

B-521 10:39

Digital mammography: An evaluation of five test objects for assessment ofimage qualityA. Pascoal, C. Lawinski; London/UK ([email protected])

Purpose: In recent years the use of digital imaging systems for mammographyhas become widespread. The assessment of image quality continues to be ofparamount importance as part of the routine testing of equipment performanceand images produced with test objects are commonly used for this purpose. Inthis study the performance of 5 test objects was evaluated on a range of digitalimaging systems covering most technologies currently available for digital mam-mography.Methods and Materials: The 5 test objects evaluated were TOR (MAM),TOR (MAX), Dmam, CDMAM and DIGIMAM. Images were produced on 3 digitalmammographic systems (GE, Lorad, Fischer) and one CR system (Konica). Ad-ditionally, screen film images were also acquired to obtain reference data. Eachtest object was imaged over a range of dose values, aiming at variations of imagequality. All images were evaluated as softcopy by 4 experienced observers fol-lowing recommended methodology.Results: The results were analysed using "contrast-detail curves" and singleimage quality figures. The various test objects were compared in terms of abso-lute and relative performance. Additionally, the image quality performance ratingof the various systems was investigated with each test object. The results shownnoticeable variations in the performance of the various test objects and indicatewhich may be most useful.Conclusion: Contrast detail analysis using test objects is straightforward meth-od for the assessment of technical image quality. The understanding of the suita-

bility and performance of various tools available provides useful information tobuyer and facilitates the comparison of results obtained with different test ob-jects.

B-522 10:48

Discussion of fog standards in mammographyR.E. Dickerson; Rochester, NY/US ([email protected])

Purpose: This paper discusses the rationale for existing fog standards in mam-mography and suggests that these standards are based on performance fromolder mammography films using outdated emulsion technology. New mammo-graphic films using modern AgX microcrystals no longer exhibit the same per-formance as older mammographic films. A new proposal for fog standards issuggested that is more relevant to existing mammographic films.Methods and Materials: This paper describes details of older and more modernmammography films and their response to film aging and resulting fog levels.Results: Increased fog values in mammographic films are due to continuing chem-ical sensitization and cosmic energy radiation. Older pre-1950's mammographicfilms by virtue of several factors result in higher film fog, lower toe contrast andhigher film granularity. These include halide composition, grain morphology anddistribution of grain diameter and volume. Modern mammographic films do nothave the same response to film aging as older films systems. As a result, in-creases in fog values no longer have the same consequence. Further, increasedDmin values from higher amounts of blue colorant in the film support result inimproved image quality by enhancing scotopic vision.Conclusion: Existing fog standards for mammography are out-dated and no longerreflect the performance of new mammography films. Further, limits on fog valuesprevent further improvements in mammography image quality by limiting additionof blue colorant in the support enhancing psychovisual response. A new propos-al for fog standards is presented which is more relevant to modern mammographicfilms.

B-523 10:57

Optimisation of digital mammography systemsK.C. Young1, J.M. Oduko1, H. Bosmans2; 1Guildford/UK, 2Leuven/BE([email protected])

Purpose: To determine the optimal exposure factors for different types of CR andDR digital mammography system.Methods and Materials: Mean glandular dose (MGD) and contrast-to-noise ra-tio (CNR) were measured using PMMA phantoms simulating breasts 20-90 mmthick, with a 0.2 mm thick square of aluminium for contrast. For each thickness, arange of mAs, kV and target/filter combinations were used, for various DR andCR systems. The optimal exposure factors for each thickness were determinedas those giving a target CNR for minimum MGD.Results: Aiming for a target CNR resulted in the selection of different exposurefactors from those which give constant detector dose. The use of higher energyspectra generally resulted in the lowest patient dose, but with a higher detectordose to compensate for lower contrast. For smaller breast thicknesses (21 or32 mm), the optimal beam quality was usually 25 or 28 kV and Mo/Mo target/filtercombination. For thicker breasts a target/filter combination such as Rh/Rh or W/Rh were usually optimal, leading to a dose saving of about 30% compared to28 kV Mo/Mo. Systems with a relatively large amount of structure noise wereoptimal using lower energy spectra.Conclusion: A simple method of optimizing digital mammography systems isdescribed, and results indicate that current systems aiming for a fixed detectordose are not optimal. Maintaining a target level of CNR usually results in theselection of higher energy spectra (Rh/Rh or W/Rh) for breast thicknesses above45 mm.

B-524 11:06

MoniQA, monitor quality assurance: Initial experience with a dedicatedsoftware tool to perform general quality control of medical screen deviceswith daily renewed test imagesJ. Jacobs, T. Deprez, G. Marchal, H. Bosmans; Leuven/BE([email protected])

Purpose: To develop a complete environment (MoniQA) for quality control ofradiological screen devices. We report on the practical work-out of the softwaretools and on the first validation on a large series of different monitors in a digitalradiology department.Methods and Materials: MoniQA (developed in JAVA) uses a modular approachin which patterns are packed into plug-ins. Therefore MoniQA is able to present

256 C D E FBA

Scientific Sessions

G

any available test pattern, including a newly created pattern for daily quality con-trol. This pattern allows the evaluation of luminance, gradient, resolution and ge-ometric distortion. Certain parts in this image are randomly created to avoidmemorization. The validation of the software tool was performed in our radiologydepartment on 68 medical screen devices. Radiologists scored the patterns on adaily basis. All scores were automatically sent to a central computer for analysisas a function of time by another tool, QAMPR. We compared the different scoresand we assessed the learning curve and the reproducibility of the scores. Anevaluation of the MoniQA pattern against the AAPMtg18 QC pattern is on-going.Results: We tested our tools successfully in a medical environment. The person-nel accepted the evaluation procedure well (learning curve is less than one week).It takes roughly 300 s to evaluate two monitors. The results were different fordifferent monitor types and known artefacts were detected.Conclusion: Our preliminary results demonstrate that we can appropriately mon-itor the quality of medical screens on a daily basis for a large series of monitorsspread over different locations.

B-525 11:15

Optimization of image quality and radiation dose in digital radiographyD. Lavin1, B. Tuohy1, E. Keavey1, P. Woulfe1, J. Gallet2, W. Van Der Putten1;1Galway/IE, 2New York, NY/US ([email protected])

Purpose: Conventional film screen (FS) imaging over recent years is graduallybeing replaced with digital imaging alternatives such as computed radiography(CR) and direct digital radiography (DR). One of the inherent benefits of this ap-proach is the ability to optimise image acquisition and image display independ-ently. CR/DR and FS have inherently different imaging characteristics and thereforeimages should not be acquired using the same exposure factors. Insufficient at-tention however, has been directed in the scientific literature, at optimisation studiesin diagnostic radiography using CR and DR systems.Methods and Materials: Phantoms to simulate various body parts were con-structed. CDRAD contrast detail and Huttner resolution phantoms were placed inthe middle of a patient phantom to simulate realistic clinical conditions. Startingpoints for each X-ray examination series were the kVp/mAs, filtration and SIDvalues which had previously been used with the FS technique. Entrance doses tothe phantoms were also measured. For each X-ray examination type, an imagequality factor (IQF) was obtained. Each image series consisted of; the original FSimage, and subsequent images acquired using CR/DR technology with exposureparameters/filtration/SID settings changed. Image optimization was achieved bydetermining the IQF as a function of entrance dose.Results: Dose reduction factors of between 1.3 and 4 have been found betweenthe FS and CR/DR technology over a range of X-ray examination types.Conclusion: It is clear from our findings that, with CR/DR technology, there ispotential for significant reduction in patient entrance dose with no negative im-pact on image quality.

B-526 11:24

Radiation dose reduction by efficient 3D image restorationB. Svensson, M. Andersson, Ö. Smedby, H. Knutsson; Linköping/SE([email protected])

Purpose: The trade-off between radiation dose and image quality in CT moti-vates development of image restoration methods for maintaining image qualitywhile reducing the radiation dose. This study concerns 3D restoration applied tolow-dose CT images, using filter networks.Methods and Materials: Post-mortal volume CT images of the orbit were ac-quired at 80, 160 and 320 mAs. By post-processing, the volumes were restoredby searching local 3D neighborhoods for planar structures. Although CT datasetsare 3D, post-processing is usually carried out in 2D due to the computationalload of 3D filtering. The load is however significantly reduced by applying smallerfilters in sequence and exploiting similarities of the filters, i.e. using filter net-works. Original and restored images acquired at 80 mAs and 160 mAs as well asoriginal images with 320 mAs, were shown to 4 radiologists who assessed thesubjective degree of noise and overall image quality.Results: A significant improvement of the noise score by the restoration wasseen at 80 mAs (the mean score increased from 2.5 to 3.3) but not at 160 mAs(from 3.6 to 3.8). For general image quality, none of the datasets differed signifi-cantly from the 320 mAs reference.Conclusion: Filter networks make 3D post-processing feasible and seem to im-prove the subjective image quality with respect to noise in low-dose CT datasets.To further evaluate the dose reduction potential, larger studies are needed.

B-527 11:33 !Image repeat analysis for a digital radiography systemA. Steffen, U. Neitzel, K. Förger; Hamburg/DE ([email protected])

Purpose: Image reject/repeat analysis is an important quality control tool withincreasing importance for digital radiography systems. We report on methods forand results of analysing repeat rates with clinically used digital radiography sys-tems.Methods and Materials: Two different methods were applied to record and ana-lyse the rejected and repeated images on digital flat-panel detector radiographysystems in routine clinical use. Method I used the logfile entries generated foreach X-ray examination. For method II additionally all images (even rejected)were stored on a separate storage device during the study period for later analy-sis. In total, method I was applied to 162,000 image acquisitions in 27 hospitals,located in different parts of the world; method II was applied to 13,500 imageacquisitions in 3 hospitals located in Germany.Results: The overall repeat rate determined with method I was 6.3 ± 3.1%, how-ever the rate varied from 0.9 to 12.2% for individual hospitals. For the three hos-pitals analyzed according to method II the repeat rate ranged from 4.6 to 7.4%.The mean repeat rate per examination type reached from values below 2% forthorax to almost 25% for sacrum. The main reasons for repeating image acquisi-tions were positioning and collimation errors in 65-75% of the cases; technicalproblems or malfunctions accounted for about 5%.Conclusion: Examination log files can be used for repeat analysis in clinicalroutine. Repeat rates for the digital radiography system evaluated are mainlydetermined by positioning/collimation errors, whereas technical reasons form onlya minor part.

B-528 11:42

Is it time to re-introduce formal radiation protection teaching for allspecialties?M. Nadarajah1, N. Jayasekera2; 1Manchester/UK, 2London/UK([email protected])

Purpose: To assess level of knowledge, training and safety in the use of theimage intensifier by orthopaedic trainees in England.Methods and Materials: A questionnaire survey was conducted among ortho-paedic specialist registrars with operating responsibility for the image intensifier.Results: 75 trainees responded. 72% had received some tuition in ionising radi-ation protection. Only 15 correctly located X-rays on an energy spectrum depict-ing microwaves to gamma rays. 21% were unable to correctly identify the X-raytube end of the c-arm on a picture of an image intensifier. 71% admitted to inad-vertent screening of their hands at least once a month. 85% routinely held theirperiphery within 60 cm of the flare of the greater trochanter during screening.16% routinely wore insufficient radiation protection. The original ionising radia-tion regulations made it compulsory for all staff working with X-ray equipment toreceive formal instruction on the hazards of ionising radiation and the safe use ofX-ray equipment. In 2000 with the advent of IRMER the radiation protection coursewas disbanded without alternate compulsory instruction. Therefore there is a lackof consensus regarding the level and content of knowledge in ionising radiationprotection issues required by a trainee.Conclusion: The majority of orthopaedic trainees do not use the image intensifi-er safely. This may result in unjustified ionising radiation exposure of patients,assistants, other theatre staff and themselves to unnecessary ionising radiation.Instruction in ionising radiation protection should be reintroduced as a prerequi-site for higher training in orthopaedics in order to stem this dangerous trend.

B-529 11:51

Patient doses and diagnostic reference levels in paediatric radiology inFinland: Phase I: The methodT. Kiljunen, H. Järvinen, T. Parvianen, T. Komppa, S. Savolainen; Helsinki/FI([email protected])

Purpose: Based on the MED-directive, 97/43/Euratom, STUK has the responsi-bility to set the national Diagnostic Reference Levels (DRLs) for the most com-mon radiological examinations in Finland. Paediatric patients deserve a specialattention due to the higher radiation risk compared with adults. The purpose wasto present a method which takes into account patient size when setting DRLs inpaediatrics.Methods and Materials: Patient doses collected from six hospitals during theyears 1994-2001 were reviewed, and new measurements were done in two hos-pitals at 2004. For the overall data, the total number of examinations was 700chest examinations, 100 micturating cystourethrography (MCU) fluoroscopy ex-

B 257C D E FA

Su

nd

aS

un

da

Su

nd

aS

un

da

Su

nd

ayy yyy

Scientific Sessions

G

aminations and 10 - 30 other conventional or fluoroscopy examinations. Differentmethods for taking the patient size into account were compared by statisticalmethods. Both dose quantities, entrance surface dose (ESD) and dose area prod-uct (DAP) were used.Results: The method established by the National Radiation Protection Board(UK) for setting DRLs was found to produce extra uncertainty in the procedureand was not easy to apply to Finnish practise. Patient doses correlated exponen-tially (R2 = 0.86 - 0.99) with both the projection thickness, which is measureddirectly from the patient, and the estimated cylindrical diameter (ECD) of thepatient.Conclusion: For the specification of the size of patient, a projection thickness ismore practical than the calculated ECD. Reference levels for paediatric radiolog-ical examinations could be given as a function of patient projection thickness,presented as a linear curve on half logarithmic scale.

10:30 - 12:00 Room P

Vascular

SS 1015Magnetic resonance angiographyModerators:G.M. Bongartz; Basle/CHM.I. Furmanek; Warsaw/PL

B-530 10:30

Free breathing steady state free precession (SSFP) MR angiography of therenal arteries without contrast agent: A prospective intraindividualcomparison with high spatial resolution contrast-enhanced MRangiographyR. Wyttenbach1, A. Braghetti1, M. Alerci1, M. Wyss2, M. Di Valentino1, L. Cozzi1,R.M. Hoogeveen3, M. Katoh4, A. Gallino1; 1Bellinzona/CH, 2Zurich/CH,3Best/NL, 4Aachen/DE ([email protected])

Purpose: To prospectively assess diagnostic accuracy of non contrast renalmagnetic resonance angiography (MRA) using a navigator gated free breathing3D SSFP sequence, compared to contrast-enhanced (CE) MRA.Methods and Materials: Thirty hypertensive patients (mean age 57 years) withsuspected renal artery stenosis underwent navigator-gated 3D SSFP MRA dur-ing free breathing (acquired voxel 0.9x1.2x2.0 mm) at 1.5 T, followed by a highresolution (acquired voxel 1.1x1.3x2.5 mm), breath-hold CE-MRA (0.15 mmol/kg Gd-BOPTA) using parallel imaging technique. Renal artery stenosis was gradedon a 5 point scale. Maximal visible vessel length was measured for both tech-niques. Consensus reading of CE-MRA and DSA (n = 8) were used as standardof reference. Two blinded, independent observers evaluated non-contrast SSFPversus CE-MRA.Results: Eleven stenoses of renal arteries > 50% were diagnosed. Sensitivity ofnon-contrast SSFP MRA for detection of significant renal artery stenosis (> 50%)was 100% for both observers (CI 0.82-1.15 and 0.82-1.18). Specificity was 90%(CI 0.85-0.95) for reader 1 and 95 % (CI 0.91-1.0) for reader 2, and accuracy was92% (CI 0.88-0.97) and 96% (CI 0.92-1.01), respectively. Maximal visible renalartery length was significantly greater in CE-MRA (66.7 ± 24.5 mm) compared toSSFP MRA (51.6 ± 21.9 mm) (p < 0.001).Conclusion: Compared to CE-MRA, free breathing navigator gated SSFP MRangiography has a high diagnostic accuracy for assessment of renal artery sten-osis without the need for contrast media, therefore reducing the cost of MR imag-ing for screening of renovascular disease. However, severity of renal arterystenosis may be overestimated by SSFP MR angiography due to signal loss dis-tal to the stenotic lesion.

B-531 10:39

Feasibility of contrast-enhanced MR angiography during preoperativeevaluation of donor kidney for transplantationJ.J. Lee, Y. Kim, T. Kwon; Daegu/KR ([email protected])

Purpose: During renal transplantation, the feasibility of CE-MRA as a modalityfor preoperative donor kidney exam was evaluated.Methods and Materials: Prospectively, 69 kidney donors (mean age of 41) wereexamined using both CE-MRA and DSA. CE-MRA preceded DSA without delay.The interpretation of CE-MRA was blinded from DSA. In 55 subjects, hand as-sisted laparoscopic nephrectomy was performed.As check points, four facts of early branching of renal artery, accessory artery,

accessory vein, and renal arterial stenosis were selected. The check points ofCE-MRA were compared with DSA and surgery. Statistical analysis of CE-MRAbased on either DSA or surgery was performed.Results: Based on surgery, sensitivity, specificity and accuracy of CE-MRA todepict "early branching of renal artery" were 50%, 98% and 95% (p < 0.05, k=0.54).The score for "accessory renal artery" was 80%, 96% and 95% (p < 0.001, k=0.70),respectively. The depiction of "accessory renal vein" using CE-MRA was statisti-cally insignificant.Based on DSA, sensitivity, specificity and accuracy of CE-MRA to depict "earlybranching of renal artery" were 79%, 100% and 98% (p < 0.001, k=0.87). Thedepiction of "accessory renal artery" showed 97%, 100% and 99% (p < 0.001,k=0.98), respectively. The depiction of "accessory renal vein" was also statisti-cally significant (100%, 99% and 99%, p < 0.05, k=0.49). In evaluation of renalarterial stenosis, Spearman' rho (rs) between CE-MRA and DSA was 0.95(p < 0.001).Conclusion: Renal CE-MRA showed significant correlation with DSA. Althoughthe depiction of accessory renal vein was limited, CE-MRA is feasible as a mo-dality for informing one about the arterial morphology of the donor kidney preop-eratively.

B-532 10:48

Contrast-enhanced MR angiography of the renal arteries at 3.0 T: InitialresultsU. Kramer1, K. Nael2, M. Lichy1, M. Fenchel1, P.J. Finn2, C.D. Claussen1,S. Miller1; 1Tübingen/DE, 2Los Angeles, CA/US([email protected])

Purpose: Purpose of this study was the evaluation of a high-resolution MRA ofthe kidneys at 3.0 T in healthy volunteers and patients with suspected renal ar-tery stenosis (RAS).Methods and Materials: 20 healthy volunteers and 12 consecutive patients sus-pected of having RAS (mean age 47.3 years) were scanned with a 3D-GRE Flash-sequence (TR/TE 2.9/1.2 ms, flip 23°, matrix 576, voxel size 0.6x0.9x1.2 mm³,GRAPPA 2, acquisition time 22 sec). Using a phased-array surface coil on a3.0 T whole body scanner (Siemens) a high-resolution MRA was implementedafter injection of 0.1 mmol/kg BW gadodiamide. Images were assessed subjec-tively based on visibility and delineation of vessel wall and image quality (0-3points). Quantitative evaluation was done by measuring the contrast-to-noise ra-tio (CNR) and signal-to-noise ratio (SNR).Results: Our technique was able to visualize renal arteries (RA) in all subjects(100%) with mean visibility score of 2.3 ± 0.3. RA were identified up to the sec-ond order branches in all subjects. In 4/12 patients a haemodynamic relevantnarrowing of the RA was found and proven by conventional angiogram. No mo-tion degradation of the RA manifest by blurring and reduction of arterial signalintensity (SI) were found. No significant difference concerning mean CNR andSNR were found (CNR abdominal aorta 45.4 ± 19.2, right RA 44.3 ± 16.9, leftRA 44.4 ± 17.1).Conclusion: Using parallel imaging techniques renal artery MRA at 3.0 T is fea-sible and initial results are promising. Signal gain at 3.0 T imaging is an importantfactor to further increase spatial resolution and improves the visualization of smallvessel segments.

B-533 10:57

Renal perfusion: Comparison of SR-turboFLASH measurements at 1.5 Twith SR-turboFLASH and TREAT (time-resolved echo-shared angiographictechnique) measurements at 3.0 TH.J. Michaely1, K. Nael2, S.O. Schönberg1, K.-P. Lodemann3, M.F. Reiser1,S.G. Ruehm2; 1Munich/DE, 2Los Angeles, CA/US, 3Konstanz/DE([email protected])

Purpose: The aim of this study was to assess the influence of the acquisitiontechnique and field strength on the variability of semiquantitative renal perfusionparameters.Methods and Materials: 8 healthy volunteers were examined at 1.5 T (SiemensAVANTO) using a 2D-SR-turboFLASH sequence (voxel size 2.3x2.1x8 mm³, tem-poral resolution 1s/4 slices) for the assessment of renal perfusion after the bolusadministration of 7 ml 0.5-molar Gd-chelates. At 3.0 T (Siemens TRIO), another8 healthy volunteers were examined using the above 2D-SR-turboFLASH afterthe bolus administration of 7 ml 0.5-molar Gd-chelates. Another 8 healthy volun-teers were examined at 3.0 T using TREAT (voxel size 1.7x1.2x5 mm³, temporalresolution 1.4s/3D data set) after bolus administration of 6 ml of 0.5-molar Gd-chelates. Postprocessing based on a gamma-variate fit, yielded the renal func-tional parameters mean transit time (MTT), time to peak (TTP), maximal upslope(MUS) and maximal signal intensity (MSI).

258 C D E FBA

Scientific Sessions

G

Results: MTT and TTP did not show significant differences between the differenttechniques and field strengths. MSI and MUS were significantly (p = 0.002) high-er with TREAT (range of means (ROM) 591 A.U. and 103.5 A.U./s) than with SR-turboFLASH at both field strengths (MSI ROM 400.5 A.U. - 362 A.U. and MUSROM 65.4 A.U./s - 68.7 A.U./s). Image quality was rated better for the 3 T images.Conclusion: Semiquantitative renal perfusion measurements are feasible withTREAT at 3.0 T and turboFLASH techniques at 1.5 T and 3.0 T. MTT and TTPare independent of the technique and field strengths applied. Temporal samplinggreater than 1.4 seconds does not seem to be required.

B-534 11:06

Contrast enhanced magnetic resonance angiography of renal arteries andaorta: Comparison of peak signal intensity using 1.0M and 0.5M gadoliniumchelatesB.C. Anil Kumar1, T.A. Sudarshan2, W. Milne2, G. Houston2; 1Nottingham/UK,2Dundee/UK ([email protected])

Purpose: To compare the effect of 1M Gadolinium chelate to that of three 0.5Mchelates on peak signal intensity (PSI) in contrast enhanced magnetic resonancerenal angiography (CE-MRA).Methods and Materials: 72 consecutive patients referred for CE-MRA with sus-pected renal artery stenosis were randomised into four groups. Group A (n = 18)received gadobutrol (0.1 ml/kg of 1M solution), Group B (n = 15) received dimeg-lumine gadopentate (0.2 ml/kg of 0.5 M solution), Group C (n = 23) received ga-doteridol (0.2 ml/kg of 0.5 M solution) and Group D (n = 16) received gadobendatedimeglumine (0.2 ml/kg of 0.5 M solution). PSI in the abdominal aorta and therenal arteries (3 measurements on each side) were measured on angle correct-ed, subtracted maximum intensity projection (MIP) images by two radiologistsindependently, blinded to the type of contrast used. Kidneys with > 50% stenosesof major renal artery were excluded from the study.Results: 24 normal kidneys were found in Group A, 25 in group B, 30 in group Cand 23 in group D. Mean PSI values for aorta were: group A -65.3, group B-63,group C-64.3 and group D-64.7: for the renal arteries were-group A-42, group B-40.65, group C-41.15 and group D-40.8. There was no statistically significantdifference in the PSI values between the four groups for the aorta and renal arter-ies using one-way ANOVA test.Conclusion: In our routine practice of CE-MRA, using recommended dosage,there was no demonstrable or statistically significant difference in PSI of the aor-ta or renal arteries between 1M and 0.5M gadolinium chelates.

B-535 11:15

Diagnostic accuracy of contrast-enhanced MR angiography with differentdoses of gadobenate dimeglumineR. Manfredi1, L. Grazioli2, S. Thurnher3, C. Ballarati4, M. Kirchin5, J. Parker6,G. Pirovano6, A. Spinazzi6; 1Rome/IT, 2Brescia/IT, 3Vienna/AT, 4Como/IT,5Milan/IT, 6Princeton, NJ/US ([email protected])

Purpose: To determine the accuracy of contrast-enhanced MR angiography (CE-MRA) with four doses of gadobenate dimeglumine (Gd-BOPTA) for detection ofclinically significant steno-occlusive disease versus digital subtraction angiogra-phy (DSA).Methods and Materials: 84 patients with suspected disease of the renal (n = 16),pelvic (n = 41) or carotid (n = 27) arteries underwent CE-MRA (3D-spoiled gradi-ent-echo sequences at 1.5 T) and DSA. CE-MRA was performed with Gd-BOPTAat 0.025, 0.05, 0.1 or 0.2 mmol/kg (21, 25, 17 and 21 patients, respectively).Sensitivity, specificity, and accuracy of CE-MRA for detection of significant dis-ease (> 50% stenosis/occlusion for renal and pelvic arteries; > 70% stenosis/occlusion for carotid arteries) was determined by three fully blinded, independentradiologists versus DSA. All comparisons were tested statistically (McNemar andChi-square tests).Results: Values for sensitivity for the 4 doses and the 3 readers were 70.4%,63.%, 44.4% following 0.025/mmol/kg, 67.6%, 67.6%, 55.9% following 0.05/mmol/kg, 84.2%, 84.2%, 84.2% following 0.1/mmol/kg, 65.2%, 69.6%, 65.2% following0.2/mmol/kg. Specificity was 79.8%, 78.2%, 90.3% following 0.025/mmol/kg,91.3%, 83.3%, 90.7% following 0.05/mmol/kg, 96.9%, 96.9%, 99.2% following0.1/mmol/kg, 84.7%, 83.8%, 86.5% following 0.2/mmol/kg. Accuracy was 78.1%,75.5%, 82.1% following 0.025/mmol/kg, 87.0%, 80.4%, 84.2% following 0.05/mmol/kg, 95.2%, 95.2%, 97.3% following 0.1/mmol/kg, 81.3%, 81.3%, 82.8% fol-lowing 0.2/mmol/kg. The differences in accuracy between the 0.1 mmol/kg doseand the other doses was statistically significant (p < 0.01, all readers).Conclusion: Significantly better diagnostic performance on CE-MRA of the re-nal, pelvic and carotid arteries is achieved with Gd-BOPTA at a dose of 0.1 mmol/kg compared to doses of 0.025, 0.05 and 0.2 mmol/kg.

B-536 11:24

USPIO-enhanced MR imaging of high-risk atherosclerotic plaques in WHHLrabbits at 3 T: Injection-dose dependent detectionH. Ittrich1, A.N. Priest1, P.G. Begemann1, C. Jahntz1, C. Weber1, B. Misselwitz2,G. Adam1; 1Hamburg/DE, 2Berlin/DE ([email protected])

Purpose: To evaluate the potential of 3 T MR imaging for detection of high-riskatherosclerotic plaques after injection of different USPIO dosages.Methods and Materials: High resolution MR imaging of the aorta was performedon 15 WHHL (ages 27.2 ± 5.9 months) and 6 NZW rabbits (control, 22.7 ± 3.2months) before and 24-120h after USPIO administration (DDM 43/34,Schering)in a 3 T scanner (Philips Intera) using an 8-channel head coil. 3 groups withdosages of 1.0, 0.25 and 0.1 Fe/kg were used (each 5 WHHL/2 NZW). T2*w3DGRE (TR/TE=30/2.7-11.0 ms, FA20°, res.0.35×0.35×3 mm) and 3D MRA (TR/TE 5.6/2.1 ms, FA25°, res.0.7×0.7×0.7 mm, USPIO/Gd-DTPA-enhanced) wereused to investigate areas of signal intensity (SI) reduction within the aortic walldemonstrating the uptake of USPIO. MR Images were matched with histology(HE, Prussian blue, RAM11).Results: Areas of SI reduction within the aortic wall were detected after injectionof 1.0 mmol/kg (4/5 animals) correlated with USPIO uptake in inflammatoryplaques (double-immunostaining). No significant SI reduction was detected afterinjection of 0.25 and 0.1 mmol/kg in WHHL animals with negligible USPIO up-take by macrophages despite the high burden of atherosclerotic plaques. NZWanimals showed no aortic SI changes corresponding to the absence of plaques.Increasing TE (2.7-11.0 ms) decreased the detection threshold for aortic sus-ceptibility artefacts, but increased artificial image distortion from outside the aor-ta (USPIO-uptake in bone marrow, lymph nodes).Conclusion: Contrast enhanced MR imaging techniques at 3 T using experi-mental dosages of USPIO (1.0 mmol/kg) allow identification of vulnerable athero-sclerotic plaques. The absence of aortic SI reduction after lower dose USPIOadministration does not correspond to an absence of plaque formation((0.25),0.1 mmol/kg).

B-537 11:33

Molecular imaging of atherosclerosis-associated matrix metalloproteinases(MMPs) using a novel and specific MR imaging contrast agent - P947S. Amirbekian1, V. Amirbekian1, J.S. Aguinaldo1, F. Hyafil1, E. Vucic1,E. Lancelot2, C. Corot2, Z.A. Fayad1; 1New York, NY/US, 2Paris/FR([email protected])

Purpose: Matrix Metalloproteinases (MMPs) play an important role in athero-sclerotic plaque progression, rupture, and thrombosis. Targeting MMPs with mo-lecular MR imaging Probes may facilitate detection of atherosclerosis andassessment of plaque stability. P947 is a short peptide ligand for MMPs. P947has a Gadolinium attached to it turning it into an MR imaging contrast agent. Theobject of the study is to test the ability of P947 to detect and quantify atheroscle-rosis.Methods and Materials: 15 month-old Apolipoprotein-E-Knockout mice (n = 10)underwent in vivo MR imaging of the abdominal aorta using a 9.4T MR system.Pre-contrast enhanced (CE) and post-CE MR imaging was performed at 1, 2, 3,and 24 hrs post-injection using a T1W black blood sequence. P947 (100 µmol/kg) was injected via the tail vein. For control, a group ApoE-KO mice (n = 4) wereinjected with Gd-DOTA (Dotarem-standard Gadolinium agent) using an equiva-lent dose of Gd (100 µmol/kg). After MR imaging, the aortas were isolated andfixed; pathology, MMP zymography and immunohistochemistry was performedfor MMPs. MRI images of the matched (pre and post) slices were used for analy-sis.Results: In comparison to controls, in ApoE-KO mice that were injected withP947 there was heterogeneous enhancement seen on MR imaging, with signifi-cant increase in contrast-to-noise ratio (CNR) for wall/lumen&wall/muscle in thepost-CE images (1, 2 & 3 hrs post-injection) with no significant increase at 24 hrsvs. pre-CE images [ANOVA, p < 0.05]. The ratio of the post to pre-contrast signalintensity of the wall (normalized to muscle) with P947 was 1.92 ± 0.19 (91% en-hancement) in KO group at 1 hr, 1.51 ± 0.14 (52%) at 2 hrs, 1.23 ± 0.13 (22%) at3 hrs, and 1.13 ± 0.06 (14%) at 24 hrs.Conclusion: Noninvasively, P947 provided good contrast enhancement of aorticatherosclerotic plaque in ApoE-KO mice. P947 offers potential for novel noninva-sive detection of atherosclerosis.

B 259C D E FA

Su

nd

aS

un

da

Su

nd

aS

un

da

Su

nd

ayy yyy

Scientific Sessions

G

B-538 11:42

High-resolution whole-body MRA using parallel imaging on a 32-channelMR system and a blood pool contrast agent: Clinical implementationK. Nikolaou1, H. Kramer1, C. Grosse1, D. Clevert1, O. Dietrich1, M. Hartmann2,P. Chamberlin2, M.F. Reiser1, S.O. Schoenberg1; 1Munich/DE, 2Cambridge,MA/US ([email protected])

Purpose: The aim was to assess the clinical utility of a whole-body MRA protocolon a whole-body MR scanner in combination with a blood pool contrast agent.Methods and Materials: After protocol optimization in ten healthy volunteers,ten patients with proven atherosclerotic disease underwent whole-body MRA ona 32-channel 1.5 T MRI imager with matrix coils and parallel imaging (MagnetomAvanto, Siemens Medical Solutions, Erlangen, Germany). For contrast-enhancedMRA, an albumin-binding gadolinium chelate (MS325, EPIX Pharmaceuticals/Vasovist, Schering AG) was used. Angiograms of the carotid arteries (or renalarteries) and lower legs were obtained during first-pass imaging. In the steady-state phase, a whole-body MRA was performed, with a spatial resolution of up to0.074 mm3 voxel size. Image quality of all MRA data was rated subjectively. Inpatients, study results were compared to Doppler ultrasound, invasive digital sub-traction angiography, or state-of-the-art MRA protocols.Results: In the first-pass images (n = 280 vessels), vessel conspicuity of thecarotid arteries, renal arteries, and lower leg arteries was rated excellent in 87%,83%, and 99%, respectively. Regarding steady-state (n = 887 vessels), vesselconspicuity was rated as being excellent, good, moderate or poor in 54%, 35%,10% and 1%, respectively. In patients, MS325-MRA data showed a diagnosticagreement in assessing stenosis of the carotid arteries, the renal arteries and ofthe peripheral run-off vessels in 93%, 90%, and 91%, respectively.Conclusion: Combining a 32-channel MRI imager and a blood pool imaging agentallows for whole-body MRA without compromises in spatial resolution or anatom-ic coverage, providing a high diagnostic accuracy.

B-539 11:51

Comprehensive diabetes imaging with whole body MR imaging at 1.5 and3.0 T in patients with longstanding diabetesS. Weckbach, S.O. Schoenberg, H. Kramer, K.G. Parhofer, C. Spitzweg,S. Kessler, M.F. Reiser; Munich/DE([email protected])

Purpose: To implement and evaluate an MR protocol for comprehensive diabe-tes imaging on a 1.5 and 3.0 T whole body scanner using iPAT.Methods and Materials: 30 patients with type 1 or 2 diabetes (> 10 years) wereexamined on a 1.5 T (12 pts) or 3.0 T (18 pts) whole body MR system (Magne-tom Avanto/Trio, Siemens Medical Solutions). Imaging of the brain with T2-w-,diffusion weighted and FLAIR-images was performed. Cardiac function was as-sessed with a single breath-hold multi-slice real time trueFISP technique (TSENSEfactor 2-4). For DCE imaging, segmented trueFISP PSIR images were performed.3D-Gd-MR-Angiography of the carotids, abdominal aorta (iPAT factor 3), thighs,calves and pedal arteries (iPAT factor 2) was obtained. Images of the feet wereacquired. Correlation with a group of 200 healthy adults who received whole bodyMR imaging for cardiovascular screening was performed.Results and Discussion: Vascular pathologies were substantially more oftenfound in diabetics than in healthy controls. Carotid artery stenosis was observedin 21% (6% of healthy adults). Renal artery stenosis was seen in 11 % of diabet-ics (controls 0.25%). Stenoses of peripheral arteries were seen in 47% -58%(healthy controls 14%). In four diabetes patients (21%) silent myocardial infarc-tions were detected (control group 1%). Chronic ischemic brain lesions were moreoften found in diabetics. Osteomyelitis was detected in 1 patient, neuropathic footfound in 3 patients.Conclusion: Whole body MR imaging is a highly promising method for compre-hensive disease specific imaging in diabetes since typical pathologies can beidentified and treated earlier.

260 C D E FBA

Scientific Sessions

G

Mo

nd

aM

on

da

Mo

nd

aM

on

da

Mo

nd

ayy yyy

Scientific Sessions

B 261C D E FA G

MondaMondaMondaMondaMondayyyyy, Mar, Mar, Mar, Mar, Marccccch 6h 6h 6h 6h 6

B

Scientific Sessions

262 A C D E F G

EARGeneral

Assembly

HL 3(p. 98)

Scient.Exhib./The Matrix

Awards

SY 13Satellite

SymposiumContrast

enhanced US inclinical practice:Liver, prostate,

pancreas, kidneyand lymph nodes

(p. 554)


RadiologySupra-aorticinterventions

(p. 280)





RadiologyNon vascularinterventions

(p. 288)

SS 1510Musculo-skeletalJoints

(p. 282)

EM 3"ECR meets"

KoreaCholangio-carcinoma:

Imaging andintervention

(p. 98)

SS 1502Breast

Magneticresonance (1)

(p. 284)

SS 1504Chest

CT pulmonarynodule detection

and diagnosis(p. 286)

SS 1508Head and Neck

Temporalbone and

maxillo-facialimaging(p. 290)

SS 1503Cardiac

Myocardialviability andwall-motion

(p. 292)

SF 16Special Focus

SessionPediatric(p. 101)

RC 1602Breast

Radiopathologicalcorrelation

(p. 102)

HospitalAdministratorSymposium

RC 1604Chest

CT angiographyof the chest

(p. 103)

RC 1608Head and NeckNasal cavity and

paranasalsinuses(p. 103)

RC 1605Computer

ApplicationsIntegrating ofdigital medical

imaging inelectronic health

care records(p. 104)

SS 1403Cardiac

Evaluation ofcardiac function

(p. 270)

SS 1411aNeuro

Functional MRimaging(p. 268)

SS 1401bGI TractVirtual

endoscopy:Clinical use and

impact(p. 266)





Wrist(p. 86)

SF 13Special Focus

SessionTumor

angiogenesis(p. 87)


applications:Pediatric cardiac

imaging(p. 88)


Imaging ofcolorectal cancer

(p. 89)

RC 1304Chest

Thoracicmanifestations

of extra-thoracicdisease(p. 89)

RC 1311Neuro

Epilepsy(p. 90)

RC 1308Head and NeckHow to image

the neck:A multimodality

approach(p. 91)


RadiologyUterine arteryembolization

(p. 91)

SA 14State-of-the-Art

SymposiumMolecularimaging(p. 96)

SS 1410Musculo-skeletalCartilage(p. 258)

SS 1401aAbdominal

Viscera(Solid Organs)Hepatocellular

carcinoma:Multimodality

studies(p. 260)

SS 1402Breast

Interventionalprocedures

(p. 262)

SS 1404Chest

MR imaging ofthe thorax(p. 264)

EP

OS


20:00

registration


room A2nd level

room B2nd level

room C2nd level

room E1entr. level

room E2entr. level

room F1entr. level

room F2entr. level

room Hlower level

room Glower level

08:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:00

14:00

14:30

15:00

15:30

16:00

16:30

17:00

17:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:30

14:00

14:30

15:00

15:30

16:00

16:30

17:00

12:3012:30

13:00

18:00

18:30

19:00

18:30

19:00

18:00

17:30

13:30

08:30

08:00

07:30

07:00

07:30

08:00

07:00

ECRGeneral

Assembly

Mo

nd

aM

on

da

Mo

nd

aM

on

da

Mo

nd

ayy yyy

Scientific Sessions

B 263A C D E F G

WS 1424Virtual

InterventionsRenal

"Hands-on"Workshop

E3 1320Kindergarten

biology forradiologists

(p. 94)


MR technicaldevelopments

(p. 298)

RC 1612Pediatric

Neuro(p. 106)

SS 1506Contrast Media

MR imagingagents(p. 300)

SS 1515Vascular

Multidector CTangiography

(p. 302)

E3 1520How to prepareand present ascientific paper

and exhibit(p. 100)

SS 1511Neuro

Spine andspinal cord

(p. 294)

SS 1512Pediatric

Musculoskeletal(p. 296)

E3 1420Added value of3D acquisitions

(p. 98)

SS 1415Vascular

Experimentalvascularimaging(p. 278)


RadiologyVenous accessinterventions

(p. 105)

WS 1315VascularVascularradiology:

Practical issues(p. 92)

RC 1312Pediatric

Musculoskeletal(p. 93)

RC 1314Radiographers

CT(p. 93)


Methods andapplications ofmedical image

registration(p. 94)



SS 1411bNeuro

Interventions(p. 272)

RC 1412Pediatric

Genitourinaryimaging(p. 97)

SS 1413Physics inRadiologyCT dose(p. 274)


RadiologyAortic



MR imaging(p. 106)


Safetyconsiderations

in MR(p. 107)

E3 1620Error

in radiology(p. 107)

WS 1624Virtual

InterventionsRenal

"Hands-on"Workshop

WS 23E3Screening



WS 23E2Screening



WS 23E1Screening



WS 23D3Screening



WS 23D2Screening



WS 23D1Screening



WS 1524Virtual


"Hands-on"Workshop



WS 1324Virtual


"Hands-on"Workshop

room Klower level

room L/M1st level

room N/O1st level

room Plower level

room Wbasement

room Y1st level

The Globe2nd level

08:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:00

14:00

14:30

15:00

15:30

16:00

17:00

17:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:30

14:00

14:30

15:00

15:30

16:00

16:30

17:00

12:3012:30

13:00

18:00

18:30

19:00

18:30

19:00

18:00

17:30

13:30

08:30

08:00

07:30

07:00

07:30

08:00

07:00

room Ilower level

16:30

room Xentr. level

roomZ1st level

WS 23D4Screening


test"Hands-on"WorkshopSY 12

SatelliteSymposium

Open PET/CT -Setting newstandards...

(p. 554)

Scientific Sessions

264 C D E FBA G

10:30 - 12:00 Room A

Breast

SS 1402Full-field digital mammography/CADModerators:S. Barter; Bedford/UKC. Van Ongeval; Leuven/BE

B-540 10:30

Grid removal and dose reduction in full-field digital mammographyG. Gennaro1, H. Souchay2, L. Katz2, C. Alberelli2, R. Klausz2, C. di Maggio1,V. Borile1; 1Padua/IT, 2Buc/FR ([email protected])

Purpose: To estimate the dose savings which could result from removing theanti-scatter grid in full-field digital mammography, without degrading image qual-ity, as a preliminary step for a clinical experiment.Methods and Materials: Contrast-to-noise ratio (CNR) provided by 100 mmsquare aluminium foils, 0.1 and 0.2 mm thick, positioned on a PMMA phantom,was measured for PMMA thickness in the range 20-70 mm using a clinical digitalmammography unit. For each thickness, one image was acquired in automaticmode with the grid present and the Average Glandular Dose determined; then siximages were acquired without grid in manual mode, retaining the same tech-nique factors and varying the mAs. A flat-fielding correction was applied to thegrid-less images to correct the non-uniformity of the scattered radiation. For eachthickness, the mAs number providing the same CNR as with the grid was com-puted. Using a distribution of PMMA-equivalent breast thickness measured pre-viously, the average dose to the population was computed in both standard andgrid-less cases.Results: For phantom thickness between 20 and 40 mm, the mAs were reducedby 30% to 15%, and by less than 10% between 40 and 50 mm. The breakevenoccurred around 60 mm. The dose to the population at constant CNR would bereduced by 6%.Conclusion: Dose reduction achievable by removing the anti-scatter grid at con-stant CNR is small for most of the phantom thickness range; the resulting dosereduction to population does not justify a clinical application.

B-541 10:39

Low dose digital breast tomosynthesisA.P. Smith, L. Niklason, Z. Jing, C. Ruth, B. Ren, J. Stein; Bedford, MA/US([email protected])

Purpose: Tomosynthesis, through the reduction of structure noise caused byoverlying breast tissue, may improve mammographic sensitivity and specificity.In particular, tomosynthesis provides improved contrast and conspicuity of masses.We have studied the visibility of low contrast objects using tomosynthesis imag-es acquired at reduced radiation doses compared to full field digital mammogra-phy (FFDM).Methods and Materials: Mass visibility as a function of radiation dose was stud-ied using phantoms. The mammography contrast detail (CD-MAM) and otherphantoms were imaged in the presence of complex backgrounds and breast tis-sue. FFDM of the objects was performed at conventional doses and tomosynthe-sis images were taken at varying doses ranging down to 45% of the FFDMexposures. Contrast detail and object visibility performance for the FFDM andtomosynthesis images were compared.The prototype tomosynthesis system used an amorphous-selenium image re-ceptor, and the FFDM system was the commercially available Hologic Selenia.Both image receptors have an imaging area of 24 x 29 cm, and a native resolu-tion of 70 microns.Results: Object detection for the tomosynthesis images at 45% dose was statis-tically superior to the equivalent detection task for the FFDM images at 100%dose. Tomosynthesis detection performance was increasingly superior as doseincreased to match the FFDM dose.Conclusion: Tomosynthesis is superior to 2-D mammography in the detection oflow contrast objects in the presence of structure noise, even at lower dose. Opti-mal mammography performance may indicate the use of tomosynthesis. Theseresults are being used to design our clinical trials.

B-542 10:48

Phantom experiments with a digital breast tomosynthesis prototypeR. Schulz-Wendtland, E. Wenkel, M. Lell, C. Böhner, W. Bautz, T. Mertelmeier;Erlangen/DE ([email protected])

Purpose: To compare the image quality of digital breast tomosynthesis prototypewith digital projection images employing tungsten and molybdenum X-ray spectra.Methods and Materials: The research tomosynthesis system is based on a full-field digital mammography system (NovationDR, Siemens) modified for a wideangular range for tomosynthetic image acquisition. The large-area a-Se detectorenables fast readout and can be switched between 2-D and tomosynthesis mode.Images of a mammographic random phantom (RMI 152 A) were acquired in both2-D and tomosynthesis mode. The phantom structures simulate masses, calcifi-cation, and fibres. From the measured projection data slice images were recon-structed with filtered backprojection. The slice images were assessed andcompared to the 2-D images in a observer study with five observers. For the usedX-ray spectra, the dose was estimated with Dance's method.Results: The dose estimation resulted in 2.1 mGy for the anode/filter combina-tion Mo/Mo (63 mAs), 3.3 mGy for Mo/Mo (100 mAs), 1.4 mGy for W/Rh(100 mAs), and 1.9 mGy for W/Rh (140 mAs). In the projection images 60 (Mo/Mo, 2.1 mGy), 65 (Mo/Mo, 3.3 mGy), 60 (W/Rh, 1.4 mGy), and 70 (W/Rh, 1.9 mGy)structures out of 75 possible findings were detected. In the reconstructed tomo-synthesis slices all 75 structures could be detected for all filter/anode and dosecombinations. When the phantom structures are obscured by superimposed tis-sue, tomosynthesis reconstruction still allows the detection of structures, whiledetection is strongly impaired with 2-D imaging.Conclusion: The diagnostic performance of tomosynthesis is higher than that ofprojection imaging for all cases investigated. In addition, tomosynthesis provides3-D localization of the detected features.

B-543 10:57

Hard-copy reading of screen-film mammography versus soft-copy readingof full-field digital mammography: Inter- and intraobserver variationP. Skaane1, C. Balleyguier2, F. Diekmann3, S. Diekmann3, J.-C. Piguet4,K. Young1, M. Abdelnoor1, L. Niklason5; 1Oslo/NO, 2Villejuif/FR, 3Berlin/DE,4Geneva/CH, 5Hillsborough, NC/US ([email protected])

Purpose: To investigate inter- and intraobserver variability in hard-copy readingof screen-film mammography (SFM) versus soft-copy reading of full-field digitalmammography (FFDM).Methods and Materials: 232 women attending a screening program were exam-ined with both modalities. The study included 46 cancers, 88 benign lesions, and98 normal findings. Images were interpreted by six radiologists experienced inmammography screening. Standard two-view mammograms of one breast ob-tained with SFM and FFDM were interpreted by the six readers in two sessions 5weeks apart, the same case not being seen twice in any session. The BI-RADS5-point rating scale was used for image interpretation. Analysis included ROC(Az) and kappa statistics. Kappa values were compared with u-test.Results: Five of six radiologists demonstrated a higher performance (Az-value)with FFDM (one with statistical significance). Mean kappa (quadratic weightingkappaqw) value for all lesions was nearly equal for SFM and FFDM (kappaqw

0.74

and 0.71, respectively), the range between reader pairs being 0.68 - 0.81 forSFM and 0.61 - 0.82 for FFDM. Mean kappaqw

values for microcalcifications only

were nearly equal for the two modalities (0.54 and 0.51, respectively), while rangebetween reader pairs was 0.40 - 0.68 for SFM and 0.40 - 0.75 for FFDM. Intraob-server agreement was lower for microcalcifications only, than for all lesions (meankappaqw 0.45 and 0.67, respectively).Conclusion: Overall, there is no difference in observer agreement between FFDMsoft-copy reading and SFM hard-copy reading. For both modalities, the main chal-lenge in screening mammography is the variability between readers.

B-544 11:06

Physical characterisation of a scanning multi-slit full field photon countingmammography systemM. Åslund1, B. Cederström1, M. Lundqvist2, M. Danielsson2; 1Stockholm/SE,2Kista/SE ([email protected])

Purpose: The physical performance of the Sectra MDM (Sectra Mamea AB,Sweden) full field digital mammography (FFDM) system was determined usingbasic image quality parameters. This scanning multi-slit system employs a directdetection photon counting detector comprised of linear silicon strip detectors inan edge-on geometry. The pixel size is 50 microns and the field of view is 24 by26 centimetres.Methods and Materials: The performance was quantified using the pre-sampled

Mo

nd

aM

on

da

Mo

nd

aM

on

da

Mo

nd

ayy yyy

Scientific Sessions

B 265C D E FA G

modulation transfer function (MTF), the normalized noise power spectrum (NNPS)and the detective quantum efficiency (DQE).The MTF, NNPS and DQE were determined as a function of spatial frequencyand detector exposure. The measurements were performed using the availableW/Al anode/filter combination at various tube voltages with PMMA or Aluminiumadded in the X-ray beam.Results: The system was found to have a relatively constant DQE > 0.60 at zerospatial frequency for detector exposures ranging from 15 to 150 µGy with a half-value layer in the 0.70-1.1 mm Al range. The 10% MTF was 10 mm-1 in the scan-ning direction and 12 mm-1 in the slit direction measured 60 mm from the chestwall edge and 45 mm above the breast support table, which resulted in a lowerDQE in the scanning than the slit direction for spatial frequencies above zero.Conclusion: The system has a wide dynamic range (15 - 150 µGy) in which theDQE at zero spatial frequency is above 0.60.

B-545 11:15 !Threshold studies for a new compression algorithm for mammographicimagesR.W. Grant1, G. Schaefer2, M. Alsewan1, R. Given-Wilson1, G.R. Kaplan1,C. Lawinski1; 1London/UK, 2Nottingham/UK ([email protected])

Purpose: To establish limits for visually lossless compression of a selection ofmammogram images containing interesting or challenging features.Methods and Materials: An image compression algorithm based on a combina-tion of discrete cosine (DCT) and wavelet transforms has been assessed at dif-ferent compression ratios. Compressed images from lossless to 120:1 weresubjected to blind analysis by a group of experienced readers who evaluated andranked the images in order of quality.Results: Visually lossless format levels of up to 80:1 compression were judgedas indistinguishable from the original; higher ratios such as 120:1 proved identifi-able although pixilation became apparent when the image was subjected to x3magnification.Microcalcification and architectural changes were among the features highlight-ed and compression at 65:1 was suggested as a suitable limit for transmission toother centres.Reduction in interval cancers could follow wider dissemination of difficult caseswhich provide up to 40% of those missed cancers.The same algorithm has also been applied to European Quality Control test phan-toms (CDMAM version 3.4, Nijmegan University) with results indicating satisfac-tory ratios of 40/60:1 for these very exacting analyses.Conclusion: Mammographic images can be safely compressed to take up only1/65 of the original storage space/bandwidth. The resulting reduction in terms oftransferring the images over electronic media proves useful in online reportingapplications.Future work will investigate all the factors that influence the visual/imaging path-way to form conclusions on a variety of influences, from electronic to ambientlighting and relevant to the decision pathway by reporting radiologist.

B-546 11:24

Computer-aided detection in digital mammography: Comparison ofcraniocaudal, mediolateral oblique and mediolateral viewsW. Moon, S. Kim, N. Cho; Seoul/KR ([email protected])

Purpose: To compare the sensitivity of a CAD system for the detection of breastcancer in three digital mammographic views.Methods and Materials: A commercially available CAD system was applied tothe craniocaudal (CC), mediolateral oblique (MLO), and mediolateral (ML) digitalmammographic views of 83 women with 83 histologically proven breast cancers.Findings were 59 masses and 41 microcalcifications (17 lesions showed bothfindings; 42 lesions, mass only; and 24 lesions, microcalcification only).Results: The sensitivities of the CAD system in the three views were 92% (76 of83) for CC, 83% (69 of 83) for MLO, and 86% (71 of 83) in ML, which were notstatistically significant (P =.07). Its sensitivity increased to 96% (80 of 83) in CCplus MLO views, and to 99% (82 of 83) in CC, MLO plus ML view. For masses,the sensitivity of the CAD system was 76% (45 of 59) in CC view and 75% (44 of59) in MLO view and this increased to 93% (55 of 59) when in combination withMLO and CC views (P =.0003). For microcalcifications, sensitivity was 98% (40of 41) in CC view and 95% (39 of 41) in MLO view, and this increased to 100%(41 of 41) when in combination with MLO and CC views (P =.31).Conclusion: The sensitivities of the CAD system were not significantly differentin these three digital mammographic views. The sensitivity for masses was in-creased when the craniocaudal view was added to the mediolateral oblique view.

B-547 11:33

Computer-aided detection (CAD) of breast cancer in screeningmammography: Early signs of breast cancer detected on priormammogramsZ. Huo1, J. Lewin2, W. Menhardt1; 1Rochester, NY/US, 2Denver, CO/US([email protected])

Purpose: To recognize early signs of breast cancer detected by CAD and tolearn to work effectively with CAD to improve the efficacy of screening mammog-raphy.Methods and Materials: While clinical use of CAD has shown to be helpful,greater benefit can be achieved if radiologists could recognize more early can-cers that had been detected by CAD. We retrospectively identified 47 visible priorcancers (28 masses, 17 Ca++, and 2 with both) on mammograms that were takenan average of 15 months before examinations with biopsy was confirmed can-cers. These priors were randomly mixed with 150 normal and 29 positive cancerexaminations. Four readers, in two sessions, were asked to mark suspicious ar-eas and provide a BIRADS rating for each breast examination with and withoutCAD.Results: The CAD system correctly marked 30 (15 masses, 13 Ca++, and 2 withboth) of the 47 prior cancers. Of the 120 readings (30 prior cancers read by fourreaders), 38% were not considered suspicious enough for biopsy and were sim-ply dismissed when the CAD correctly prompted the locations. Of the 30 priorscorrectly marked by CAD, 29 were considered suspicious enough for biopsy, byat least one reader, when the CAD correctly prompted the locations.Conclusion: It is important to recognize early signs of breast cancer detected bya CAD system to improve the benefits of mammography screening. In the pres-entation, we review the characteristics of the prior cancers and the cancers thatwere not detected, even when the cancer locations were correctly prompted.

B-548 11:42

The use of computer aided classification (CAC) in BI-RADS category 4 AlesionsF. Sperber1, R. Lederman2, S. Fields2, S. Buchbinder3, P. Bamberger2,I. Leichter2; 1Tel Aviv/IL, 2Jerusalem/IL, 3Staten Island, NY/US([email protected])

Purpose: To explore the performance of a CAD device with computer aided clas-sification (CAC) capabilities, on mammographic lesions that were assigned BI-RADS category 4 A by conventional interpretation.Methods and Materials: BI-RADS category 4 A is used for findings requiringbiopsy, but with a low suspicion of malignancy. A CAC scheme was used to retro-spectively analyze 76 cases forwarded for stereotactic biopsy (6 masses, 70 clus-ters), which were prospectively assigned BI-RADS Category 4 A. Of these cases,5 masses and 64 clusters proved to be benign at pathology. The CAC schemeautomatically extracted from the digitized mammograms, quantitative featuresthat characterize the lesions. Based on these features, the system used a classi-fication scheme to score the lesions by their likelihood of malignancy, and strati-fied the lesions according to their score into low, intermediate and high levels ofsuspicion.Results: For the 7 malignant cases included in the study, the computer assigneda high level of suspicion to all lesions and of the 69 (5 masses, 64 clusters)benign cases the computer assigned a low level of suspicion to 17 cases (3masses, 14 clusters). The CAD analysis, thus, correctly classified all malignantcases, yielding a sensitivity of 100%. The computerized analysis could correctlyclassify 17 benign cases, improving the specificity for BI-RADS Category 4 Alesions by 24.6%.Conclusion: In a case series consisting predominantly of clusters of microcalci-fications, the use of computerized classification could have correctly character-ized all the malignant lesions assigned BI-RADS Category 4 A, and could possiblyeliminate 24.6% of unnecessary biopsies.

B-549 11:51

Do CAD-systems pass the official mammography tests required forscreening and do they substantially support the radiologists to pass thesetests?A. Malich1, A. Beier1, P. Bank2; 1Nordhausen/DE, 2Jena/DE([email protected])

Purpose: The role of CAD-systems as primer diagnostic tools and as secondreader is discussed controversially. This study evaluated whether CAD-systemsare able to pass official tests offered to radiologists (using similar criteria) andwhether CAD-results assists the radiologists in a significant way to pass the tests.

Scientific Sessions

266 C D E FBA G

Methods and Materials: An official test collection including 200 images (bothsides; two views; 50 patients) was scanned (Second look, v.6.0, iCAD, USA) andanalyzed (each side separately) blinded to the histological results by 3 radiolo-gists (two experienced, one assistant radiologist) with and without CAD-print-outs. Test-collection included (officially given number to all participatingradiologists) 20-27 malignancies, at least one bilateral cancer and a few benignlesions but no BI-RADs3 cases. Test was failed in case of either > 2 false nega-tive or > 5 false positive interpretations.Results: CAD-system highlighted 26/27 cancers and not marked (correctly) 14/73 unsuspicious mammograms (Sen: 96.3%, test passed; Spec: 19.2%, test failed).Both experienced radiologists passed the test with and without CAD, the assist-ant doctor failed twice. In mean, sensitivity increased by 1%, specificity increasedby 0.6% with CAD. Per image CAD revealed 0.43 and 0.44 FP-markers (MA/MC).Conclusion: Due to the overall high number of false positives, CAD-system failedthe official mammographic test despite the highly sensitive detection of malig-nancies. A significant decrease of FP-markers for masses using version 6.0 wasverified. The effect of CAD on the analyzing radiologist, however, did not influ-ence the test results substantially.

10:30 - 12:00 Room B

Musculoskeletal

SS 1410Shoulder and hip: MRI/CT arthrographyModerators:P.W.P. Bearcroft; Cambridge/UKJ. Kramer; Linz/AT

B-550 10:30

Non-outlet impingement of the shoulder in athletes: Arthro MR imagingevaluation versus arthroscopyC. Bultrini, N. Limbucci, A. Barile, V. Calvisi, C. Masciocchi; L'Aquila/IT([email protected])

Purpose: To assess the diagnostic accuracy of arthro-MR imaging versus ar-throscopy in 13 athletes with clinical suspicion of non-outlet impingement.Methods and Materials: We enrolled 13 patients (19-34 years, mean age 28), allathletes. All patients presented with clinical evidence of non-outlet impingementof the shoulder. Before surgery, all patients underwent arthro-MR imaging andarthroscopy. We acquired SE T1-w and FSE T2-w sequences on standard exam-ination and SE T1-w and fat saturated SE T1-w after intra-articular injection of10-15 ml of contrast medium (dilution of 0.6 ml of Gd-DTPA in 250 ml of saline).We compared arthro-MR imaging with arthroscopic findings.Results: In 11 patients (84.6%) MR imaging showed shoulder abnormalities.Abnormalities were: 5 type-II SLAP lesions, 2 posterosuperior impingements, 2supraequatorial capsular enlargement, 1 SGHL, 1 Buford complex. Two patientshad non-pathologic findings. Arthroscopy confirmed MR imaging findings in 10cases (76.9%). In one patient with MR imaging findings suggestive for type-IISLAP lesion, arthroscopy showed a type-III bicipital anchor associated to subla-bral hole. At arthroscopy the two patients with normal MR imaging had post-traumatic enlargement of the rotator interval.Conclusion: Arthro-MR imaging appears to be the method of choice in the pre-operative study of articular lesions associated with non-outlet impingement inathletes.

B-551 10:39

Rotator cuff tears: Assessment with MR arthrography in 250 patients witharthroscopic correlationS. Waldt, K. Woertler, M. Bruegel, A. Burkart, E.J. Rummeny; Munich/DE([email protected])

Purpose: To assess the diagnostic accuracy of MR arthrography in the diagno-sis of rotator cuff tears of the shoulder.Methods and Materials: MR arthrograms obtained in 250 patients including astudy group of 120 patients with arthroscopically proved rotator cuff tears and acontrol group of 130 patients with intact rotator cuff tendons were reviewed inrandom order. MR imaging was performed on a 1.0 T system (Magnetom Expert,Siemens). MR arthrograms were analyzed by two radiologists in consensus forpartial- and full-thickness tears of the supraspinatus, infraspinatus and subscapu-laris tendons.

Results: At arthroscopy 189 rotator cuff tears were diagnosed, including 67 par-tial- (58 supraspinatus and 9 infraspinatus) and 122 full-thickness (62 suprasp-inatus, 36 infraspinatus, 24 subscapularis) tendon tears. For full-thickness tearsthe sensitivity, specificity, and accuracy were 92%, 98%, and 97% respectively,and for partial tears 73%, 97%, and 88% respectively. False negative assess-ments in the diagnosis of partial-thickness tears were predominantly (85%) ob-served with very small articular-sided tendon tears.Conclusion: MR arthrography is highly accurate in the diagnosis of full-thick-ness rotator cuff tendon tears, but is less accurate in the diagnosis of partial-thickness tears, mainly due to limitations in the differentiation between very smallarticular-sided partial-thickness tears and frayed or friable tendon margins and/or synovitic changes at the tendon margin.

B-552 10:48

The influence of rotator cuff abnormalities on the acromiohumeral distanceN. Saupe, C.W.A. Pfirrmann, M.R. Schmid, B. Jost, C.M.L. Werner, M. Zanetti;Zuerich/CH ([email protected])

Purpose: The purpose of this study was to evaluate the association of rotatorcuff abnormalities on the acromiohumeral distance (AHD).Methods and Materials: The AHD was measured on conventional radiographsand MR images. Three age and sex matched patient groups each including 21patients were stratified by the AHD on conventional radiographs (group I ≤ 7 mm,group II 8-10 mm, group III > 10 mm). The AHD was related to the presence,location and size of a rotator cuff tear, and fatty degeneration of the muscle as-sessed on MR arthrography. The relative influence of the different MR arthrogra-phy findings on the AHD was assessed. Spearman rank correlation and stepwiseregression was used for statistical analysis.Results: In group I (AHD ≤ 7 mm) full thickness tear in supraspinatus tendonwas present in 90% (19/21), in infraspinatus tendon in 66% (14/21), and in sub-scapularis tendon in 42% (9/21), respectively. The size of rotator cuff tendon tearsand the degree of fatty degeneration in all rotator cuff muscles showed a signifi-cant negative correlation with the AHD (P < 0.05). After stepwise regression asignificant relative influence on the AHD remained for the size of rotator cuff tear(P < 0.0001) and for the degree of fatty degeneration of the infraspinatus muscle(P = 0.013).Conclusion: Tendon tears and fatty muscle degeneration of the rotator cuff cor-relate with a reduced AHD. The size of the rotator cuff tear and fatty degenerationof the infraspinatus muscle have the most pronounced influence on AHD.

B-553 10:57

Correlation between size of rotator cuff tear and degree of tendon retractionA.M. Venkatesan1, B. Fritz2, A. Kassarjian1, W.E. Palmer1; 1Boston, MA/US,2Duesseldorf/DE ([email protected])

Purpose: To determine the relationship between anteroposterior extent of rota-tor cuff tear (RCT) and measurement of tendon retraction, and to test if tendondelamination causes a disproportionate degree of retraction of bursal or articularcuff fibers.Methods and Materials: In 59 patients (32 male, 27 female, ages 18-79) withfull-thickness RCTs at surgery, MR images were retrospectively reviewed to meas-ure the anteroposterior length of tendon tear. Two reviewers reached consensus.After 4 months the cuff retraction from the osseous attachment site to the proxi-mal tendon margin was measured. Whenever tendon delamination resulted indifferent degrees of bursal or articular retraction, the more retracted margin wasmeasured. MR images were acquired at 1.5 T using a dedicated shoulder coiland routine protocol. IRB approval was obtained.Results: Anteroposterior length of RCT ranged from 2.0-5.0 cm (mean = 2.8 cm),and retraction of tendon margin ranged from 3.0-5.5 cm (mean = 3.2 cm). Theanteroposterior length of tear showed a statistically significant positive associa-tion with tendon retraction (P <.0001). There were 49 tendons without cuff delam-ination and 10 tendons with surgically confirmed delamination. Comparing theratio of RCT length to retraction, tendons without delamination (1.2:1) showedsignificantly less retraction (P=.04) than tendons with delamination (1:1.1).Conclusion: Anteroposterior length and tendon retraction in full-thickness RCTscorrelate closely approximating a 1:1 relationship. An anteroposterior length lessthan the greatest amount of tendon retraction (ratio less than 1:1) indicates pos-sible cuff delamination of the supraspinatus tendon. Cuff delamination is impor-tant to recognize because it is associated with poorer surgical outcome.

Mo

nd

aM

on

da

Mo

nd

aM

on

da

Mo

nd

ayy yyy

Scientific Sessions

B 267C D E FA G

B-554 11:06

Assessment of labral lesions of the shoulder: Diagnostic value of MRarthrographyH. Hong, J.-Y. Park, Y.-C. Yoon, S.-H. Choi, H.-W. Chung; Seoul/KR([email protected])

Purpose: To evaluate the diagnostic value of MR arthrography (MRA) for as-sessment of labral lesions of the shoulder according to the locations.Methods and Materials: From January 2003 to June 2004, 221 shoulders of 220patients underwent both preoperative MRA and arthroscopic surgery. An experi-enced orthopedic surgeon performed arthroscopic surgeries and recorded thelabral status according to the location (superior, anterior, posterior). Each medi-cal record for labral status was reviewed retrospectively and classified into threegrades (A: normal, B1: tear, B2: non-tear abnormalities such as degeneration,fraying, cracking, concealed avulsion, etc). Musculoskeletal radiologists retro-spectively reviewed the MRA and recorded the presence of labral lesions, re-gardless of differentiation between B1 and B2, according to the location.Results: Arthroscopy revealed 146 B1 lesions (37 superior, 90 anterior, 19 pos-terior) and 78 B2 lesions (31 superior, 11 anterior, 36 posterior). Relative inci-dence of B2 lesion in the posterior labrum was significantly higher than in the ofsuperior (p = 0.001) or anterior (p = 0.012) labrum. Sensitivity, specificity, andaccuracy according to the location were 54%/90%/79% for superior, 80%/93%/87% for anterior, and 44%/90%/79% for posterior labrum. The sensitivities for B1and B2 lesions according to location were 77%/27% for superior, 84%/41% foranterior, and 61%/35% for posterior labrum. Sensitivity for B1 lesion of the poste-rior labrum was significantly lower than that of the superior (p = 0.2384) and an-terior (p = 0.003) labrum.Conclusion: The diagnostic values of MRA for posterior labral lesions were rel-atively poor compared to the anterior labrum, which may be caused by a higherproportion of non-tear abnormalities and location of the posterior labrum itself.

B-555 11:15

CT arthrography of the shoulder: Influence of iodinated contrastconcentrationJ.-Y. Choi, S. Hong, J. Lee, J.-A. Choi, H. Kang; Seoul/KR([email protected])

Purpose: To present the optimal concentration of iodinated contrast for CT ar-thrography of the shoulder comparing four different concentrations of iodinatedcontrast.Methods and Materials: Forty consecutive CT arthrograms of the shoulder wereperformed were randomly assigned to either 25% (n = 10), 50% (n = 10), 75%(n = 10), or 100% (n = 10) iodinated contrast. CT arthrograms were assessedquantitatively (CT number ratio of intraarticular fluid to pectoralis major muscle)and qualitatively (visual grading scale of 1-5 for overall image quality, image con-trast, degree of joint distension, and beam hardening artifacts).Results: The CT number ratio (mean ± SD) of intraarticular fluid to pectoralismajor muscle was 18.9 ± 3.2 for 25% iodinated contrast, 28.6 ± 2.2 for 50% iodi-nated contrast, 37.9 ± 6.1 for 75% iodinated contrast, and 41.0 ± 5.7 for 100%iodinated contrast. CT arthrograms with 75% and 100% groups showed signifi-cantly higher CT number ratio than other groups (p < 0.05, one-way ANOVA). Inqualitative assessment, 50% iodinated contrast was significantly better than oth-er concentrations (p < 0.05, one-way ANOVA).Conclusion: CT arthrograms of the shoulder obtained with more than 75% iodi-nated contrast provided higher CT number ratio. However, we preferred the im-age quality of CT arthrograms with 50% iodinated contrast.

B-556 11:24

Shoulder CT arthrography in ABER position in rotator cuff tearO. Silbermann-Hoffman, A. Feydy, H. Guerini, F. Clarençon, J.-L. Drapé,E. Schouman-Claeys; Paris/FR ([email protected])

Purpose: To determine if CT arthrography obtained in ABER (abduction andexternal rotation) position improves the detection of an horizontal component inpartial thickness tears of the rotator cuff.Methods and Materials: Fourteen patients underwent CT arthrography includ-ing a first study with the shoulder in neutral position and a second study with theshoulder in the ABER position. Two musculoskeletal radiologists independentlyreviewed the images obtained in both positions. The characteristics of the hori-zontal component of partial tears on ABER images were compared with those inneutral position. Classification of the horizontal component was done accordingto the appearance of the articular surface of the tendon. Type A lesion appears asan horizontal or intra substance component with no abnormality of the articular

surface. Type B lesion exhibits horizontal component with irregularity of the artic-ular surface. Type C lesion is associated with a flap lesion along the articularsurface. Type D consists of irregular articular surface.Results: The horizontal component of partial thickness tears was detected morefrequently on ABER CT arthrography; eight were visible only in the ABER posi-tion. The type of lesion was: type A= 2, type B= 6, type C= 4 and type D= 3. Fullthickness tears were diagnosed in neutral and ABER position. ABER imagesallowed the visualisation of unstretched tendon floating inside the articulation.Conclusion: ABER CT arthrography allows a better detection and depiction ofhorizontal partial tears. CT arthrography with ABER and neutral position is easilyrealized without loss of time during the same examination.

B-557 11:33

MR imaging findings of femoroacetabular impingementS.L.J. James1, D.A. Connell1, K. Ali1, F. Malara2, D. Young2, J. O'Donnell2;1Stanmore/UK, 2Melbourne/AU ([email protected])

Purpose: To evaluate the diagnostic accuracy of non-contrast MR imaging in theidentification of labral and articular cartilage lesions in patients with a clinicalsuspicion of femoroacetabular impingement.Methods and Materials: Pre-operative MR imaging was performed in 46 con-secutive patients (26 males, 20 females, age range 21-45 years, mean age 32.3years) with a clinical suspicion of femoroacetabular impingement. Two muscu-loskeletal radiologists independently assessed the MR images for the presenceand anatomical site of labral pathology, labral-chondral transitional zone pathol-ogy, femoral cartilage lesions and acetabular cartilage lesions. Surgical correla-tion was obtained in all cases by two surgeons experienced in hip arthroscopy.Results: Seven patients demonstrated labral tears on MR imaging and this wasconfirmed surgically in all cases. Thirty-seven patients (97%) had labral-chondraltransitional zone lesions on MR imaging compared with 38 surgically confirmedcases. The site of labral-chondral transitional zone pathology at arthroscopy wasas follows: 50% anterosuperior, 36% anterosuperior/superolateral, 11% supero-lateral and 3% superolateral/posterosuperior. The site was identified correctly inbetween 92% and 95% of cases on MR imaging. Acetabular cartilage abnormal-ity was surgically identified in 39% and femoral cartilage lesions were found in20% of cases. The acetabular lesions were correctly identified in between 89% to94% and femoral lesions in 100% of cases.Conclusion: Non-contrast MR imaging provides an accurate method of assess-ment of patients with a clinical suspicion of femoroacetabular impingement. It isnon invasive and provides useful preoperative information regarding the pres-ence and anatomical site of labral and cartilage abnormality.

B-558 11:42

MR imaging findings of acetabular dysplasiaS.L.J. James1, M. Miocevic2, F. Malara2, J. Pike2, D. Young2, D. Connell1;1London/UK, 2Melbourne/AU ([email protected])

Purpose: To evaluate the diagnostic accuracy of MR imaging in the identificationof labral and articular cartilage lesions in patients with acetabular dysplasia.Methods and Materials: Pre-operative MR imaging was performed in 25 con-secutive patients (27 hips, 16 males, 9 females, age range 19-52 years, meanage 31.2 years) with radiographic evidence of acetabular dysplasia (centre edgeangle of Wiberg < 20 degrees). The average duration of symptoms was 16.2months. Two musculoskeletal radiologists assessed MR images for the presenceof abnormality involving the acetabular labrum and adjacent acetabular articularcartilage. A high resolution technique was used to assess the labrum and labralchondral transitional zone. Surgical correlation was obtained in all cases by asingle surgeon experienced in hip arthroscopy.Results: The acetabular labra in the dysplastic hips demonstrated abnormal sig-nal intensity and had an elongated appearance when compared with the controlgroup (mean length 10.9 mm v 6.4 mm). Morphological appearances in the labraincluded surface irregularity, fissures and cleft formation. MR imaging correctlyidentified the severity of chondral abnormality in 24 of 27 hips (89%) when com-pared with arthroscopic findings.Conclusion: MR imaging demonstrates an elongated labrum, focal intra-sub-stance signal change and irregularity and fissuring of the margins in patients withacetabular dysplasia. Abnormality is also identified at the labral chondral transi-tional zone where fissuring, focal clefts, chondral deficiency and subchondralcyst formation may be apparent. A high resolution, non-arthrographic techniquecan provide an accurate preoperative assessment and evaluate the presence ofpremature osteoarthritis.

Scientific Sessions

268 C D E FBA G

B-559 11:51

Assessment of acetabular labrum morphology and corresponding cartilagequality in patients with hip dysplasia based on MR imagingG.H. Welsch1, F.F. Hennig1, W. Horger1, M.I. Menzel1, A. Cavallaro1, W. Bautz1,G. Muhr2, T.C. Mamisch2; 1Erlangen/DE, 2Bochum/DE ([email protected])

Purpose: Labral lesions in patients with hip dysplasia could indicate biomechan-ical decompensation of the hip joint and often cause early clinical symptoms. Ourobjective was to assess the condition of the acetabular labrum and its associa-tion with corresponding cartilage quality based on MR imaging.Methods and Materials:15 patients in an age range from 18 to 37 years withclinical (groin pain, impingement) and radiological (CE-angle) signs of hip dys-plasia but no signs of moderate or severe osteoarthritis underwent MR imaging.As a control group we assessed MR imaging of 15 age and sex matched healthyvolunteers. For MR imaging a 3D-Dess sequence (resolution 0.6x0.6x0.6 mm,Siemens Trio) and a radial sectioned PD-TSE sequence perpendicular to theacetabular labrum were performed. For double-oblique planning of the 12 radialplanes through the acetabular opening, isotropic images were used. The acetab-ular labrum morphology was compared to the corresponding femoral and acetab-ular cartilage quality in the radial portions.Results: The acetabular labrum was larger and the acetabular cartilage thickerin dysplastic hips. All patients had portions with signal intensity changes withinthe labrum. In 12 patients MR imaging suggested degeneration or a tear of thelabrum or both. Labrum lesions and corresponding cartilage inhomogeneity wereobserved mainly in the anterior-superior sections.Conclusion: It was possible to compare cartilage morphology to acetabular ab-normalities by using an isotropic cartilage sequence for positioning of the acetab-ular opening and the radial section. The presented findings are important for earlydetection and accurate diagnosis, and may impact optimal treatment planningand prognosis. Follow-up examinations in these young patients can help to clar-ify the natural course of labral disorders.

10:30 - 12:00 Room C


SS 1401aFocal liver lesionsModerators:N. Karabulut; Denizli/TRI.M. Ramos; Porto/PT

B-560 10:30

Free-breathing T2-weighted magnetic resonance imaging of liver:Evaluation of new MR sequencesB. Kim1, J. Kim2, K. Choi1, J. Park1, B.-C. Song1; 1Jeju/KR, 2Seoul/KR

Purpose: To compare the usefulness of turbo spin-echo with navigator triggeredprospective acquisition correction (PACE-TSE), half-Fourier single-shot fast spin-echo with navigator triggered prospective acquisition correction (PACE-HASTE),and respiratory-triggered turbo spin-echo (RT-TSE) sequences in detecting andcharacterizing hepatic lesions.Methods and Materials: Two reviewers retrospectively and independently ana-lyzed all T2-weighted images with the three sequences in 49 patients with 54malignant and 32 benign lesions. Reviewers identified and characterized all focallesions by using a five-point scale. Receiver operating characteristic (ROC) curveanalysis (area under the ROC curve [Az]) was used to compare the confidence ofdetection and characterization for the reference standard. The images were com-pared quantitatively by calculating the liver signal-to-noise ratio (SNR) and le-sion-to-liver contrast-to-noise ratio (CNR).Results: The PACE-TSE had significantly higher Az values in lesion detection(P < 0.05 for two reviewers) and lesion characterization (P < 0.05 for two review-ers) compared with PACE-HASTE. For two reviewers, the Az values in lesiondetection and characterization of PACE-TSE were superior to those of RT-TSE,but there was no significant difference (P > 0.05). Mean SNR for liver and CNRfor hepatic lesions were highest on PACE-HASTE than those of other sequenc-es. Mean acquisition times of the PACE-TSE, PACE-HASTE, and RT-TSE se-quences were 3.18 minutes, 3.17 minutes, and 6.17 minutes, respectively.Conclusion: The PACE-TSE sequence could provide the highest tumor detec-tion and characterization, although the mean lesion-to-liver CNR was inferior tothose of PACE-HASTE and RT-TSE.

B-561 10:39

Value of contrast-enhanced CT in combined PET/CT protocols foroncological and non-oncological imagingC. Pfannenberg, S.M. Eschmann, P. Aschoff, K. Brechtel, R. Bares,C.D. Claussen; Tuebingen/DE ([email protected])

Purpose: To determine the value of contrast-enhanced CT in comparison to low-dose non-contrast CT in combined PET/CT protocols for oncological and non-oncological imaging.Methods and Materials: PET/CT scans of 154 consecutive patients, 138 pa-tients with malignancy and 16 patients with vasculitis or infection were evaluatedretrospectively. PET/CT was performed with a high resolution LSO detector and16-row CT. The CT protocol varied depending on the specific clinical questionand was multi-phasic in 101 patients and single-phasic in 53 patients. PET/CTstudies were analyzed by different categories to determine the value of contrast-enhanced CT on the PET/CT interpretation. The results were validated either byhistopathology or clinico-radiological follow-up. The clinical impact of contrast-enhanced PET/CT was evaluated with respect to changes in patient manage-ment.Results: Diagnostic CT was of additional value in 93/154 patients altering thePET/CT interpretation, in the remaining 51 patient no additional benefit was de-rived from the diagnostic CT. 10 patients were excluded due to inconclusive diag-nosis in both methods. The greatest benefit of diagnostic CT was found in thefollowing categories: improved localization of pathologic FDG uptake; improvedlocal tumor staging; and detection of additional relevant findings. PET/CT inter-pretation was altered in 25% of patients in each of these groups. The benefit ofdiagnostic CT was influenced by the tumor type and changed clinical manage-ment in 35% of these patients.Conclusion: Diagnostic CT as part of the combined PET/CT examination pro-vides additional value in specific clinical conditions with resultant change of ther-apy in a substantial proportion of patients.

B-562 10:48

Multiphase multidetector computer tomography (MDCT) in evaluatingfollow-up of the regenerative nodules in Budd-Chiari syndrome (BCS)N. Flor, F. Brovelli, A. Tentori, R. Arcaini, M. Maggioni, M. Zuin, F. Sardanelli,G. Cornalba; Milano/IT ([email protected])

Purpose: A longitudinal liver evaluation in BCS.Methods and Materials: Five patients with chronic BCS underwent 3 serial MDCT:collimation 4x2.5 mm; delay 25 s (arterial), 70 (portal), 300 (delayed). Noduleswere measured: < 14 mm (size I), > 14 but < 30 mm (II), > 30 mm (III) and classedfor enhancement: hyperdense, isodense/hyperdense, isodense (pattern A); hy-perdense with central hypodensity, hyperdense with central hypodensity, isodense/hypodense with central hyperdensity (pattern B FNH similar); hyperdense, hy-perdense, hypodense (pattern C). In all patients biopsy of at least one nodule (Bor C) was performed.Results: At first CT we detected 61 nodules: all of them of size I; pattern A (n = 57);pattern B (n = 1), pattern C (n = 3). At second CT, 66: size I (n = 27), size II (n = 30),size III (n = 9); pattern A (n = 46); pattern B (n = 14), pattern C (n = 6); newlydiagnosed 12/66. Third CT, 84: size I (n = 40), size II (n = 41), size III (n = 3);pattern A (n = 42); pattern B (n = 30), pattern C (n = 12); newly 28/84. The newlydiagnosed were 101 (82%); the sum of all nodules observations was 211. Forbiopsied nodules, a diagnosis of "benign non inflammatory nodule" was obtained;for pattern B indirect FNH signs were described.Conclusion: A 10-20% annual increase of the number of nodules was observed.I-A nodules were largely prevalent (82%) at first detection and a size progressionseems to be associated with the development of B pattern (36% at the third CT).This trend is in agreement with the hypothesis of BCS-related lesions similar toFNH.

B-563 10:57

Characterisation of focal liver lesions by contrast-enhanced ultrasound,multi-phasic MDCT and Mn-DPDP MR imagingD. Rouw, D.M. Blom, E.J. Van der Jagt, H.J. Van der Zaag, M. Oudkerk;Groningen/NL ([email protected])

Purpose: To evaluate the diagnostic value of contrast enhanced ultrasound(CEUS), multi-phasic MDCT and Mn-DPDP-enhanced MR imaging in the char-acterisation of focal liver lesions.Methods and Materials: 112 patients were included in this study, data from 65patients were available for analysis. Patients underwent MDCT, CEUS and MRimaging within 4-6 weeks. Gray-scale US was performed to identify one target

Mo

nd

aM

on

da

Mo

nd

aM

on

da

Mo

nd

ayy yyy

Scientific Sessions

B 269C D E FA G

lesion per patient. When more than one lesion was seen, the performing radiolo-gist chose the lesion with the best possible acoustic window. Ultrasound wasperformed continuously (5 minutes) at a low mechanical index after a 2.4-4.8 mlbolus injection with microbubbles containing sulphurhexafluroride. CT exams (4phases) with injection of iodixanol 270 mg/mL and MR exams with hepatocellularspecific 0.01 mmol/mL mangafodipir. The reference standard was defined as ei-ther pathology or a combination of clinical follow-up of at least 6 months, com-prising any imaging and clinical features.Results: Sensitivity for a malignant versus benign lesion was 97% for CEUS (31/32), 94% for MDCT (30/32) and 72% for MR imaging (23/32); specificities were70% (23/33), 75% (24/33) and 91% (30/33), respectively. Accuracy with respectto lesion type was highest for MR imaging (75%; 49/65), CT correctly character-ised 48 (74%) of all lesions, whereas CEUS had an accuracy of 68% (44/65).Conclusion: CEUS and MDCT are equally diagnostic with respect to ruling out alesion, whereas MR imaging is more specific for ruling in a malignant lesion.However, CEUS does not perform as well as CT and MR imaging in specificlesion type characterisation.

B-565 11:06

Study of a diffusion-weighted single-shot spin echo echo planar imagingsequence using parallel imaging in the detection of focal liver lesions:Comparison with a T2-weighted turbo spin echo techniqueK. Coenegrachts1, J. Delanote1, L. ter Beek2, M. Haspeslagh1,F. Van Kerkhove1, L. Steyaert1, H. Rigauts1, J.W. Casselman1; 1Bruges/BE,2Best/NL ([email protected])

Purpose: To compare diffusion-weighted Single-Shot Spin Echo Echo PlanarImaging (SS SE-EPI DWI) with T2-weighted Turbo Spin Echo (T2w TSE) in thedetection of focal liver lesions, with focus on detection of small (< 10 mm) le-sions.Methods and Materials: Twenty-four consecutive patients with suspected malig-nant liver lesions underwent routine MR imaging examinations. A respiratory-triggered T2w TSE sequence and a respiratory-triggered SS SE-EPI DWI withb-values of 0, 20, 300 and 800 s/mm2 were compared quantitatively by double-blind measurements and calculation of the lesion-to-liver Contrast-to-Noise Ra-tio (CNR). All sequences were compared qualitatively by grading image quality,lesion conspicuity, and artifacts. Rank order analysis (RIDIT-analysis) and non-parametric tests were used to compare the ordinal data.Results: The lesion-to-liver CNRs were highest on T2w TSE when evaluatingbiliary cysts and on SS SE-EPI DWI with b=20 s/mm2 when evaluating hemangi-omas and metastatic lesions (p < 0.05). Concerning qualitative lesion detectabil-ity, biliary cysts were best detected with T2w TSE and hemangiomas andmetastases were best detected with SS SE-EPI DWI with b=20 s/mm2. The bestimage quality was achieved with T2w TSE. The best lesion conspicuity wasachieved with T2w TSE for biliary cysts and with SS SE-EPI DWI using b=20 s/mm2 for hemangiomas and metastases. Image artifacts were lowest with T2wTSE (p < 0.05).Conclusion: The application of SS SE-EPI DWI with low b-values can provideuseful information regarding the detection of focal liver lesions in general andsmall lesions in particular when compared with T2w TSE.

B-566 11:15

Superparamagnetic iron oxide (SPIO)-enhanced liver MR imaging withferucarbotran: Efficacy for characterization of focal liver lesions with T2-weighted TSE and T2*-weighted GRE imagesC.J. Zech1, S. Namkung1, T. Helmberger2, M.F. Reiser1, S.O. Schönberg1;1Munich/DE, 2Lübeck/DE ([email protected])

Purpose: To evaluate the efficacy of ferucarbotran in T2w-and T2*w sequencesfor the characterization of focal liver lesions.Methods and Materials: In 68 patients pre-contrast T2w-TSE and contrast-en-hanced T2w-TSE and T2*w-GRE images after i.v.-injection of 1.4 ml ferucarbotran(Resovist, Schering AG) were obtained on a 1.5 T MR-system. The final diag-noses included 46 malignant and 22 benign liver lesions established by histology(35/68) and long-term follow-up (33/68). SNR and CNR was calculated for allsequences. The percentage signal-intensity loss (PSIL) of focal lesions was cal-culated for the T2w-TSE sequence. Qualitative analysis was performed for imagequality and lesion conspicuity.Results: The mean SNR of solid benign lesions showed a decrease from 34.1 to21.0 (p < 0.05). In malignant lesions, the mean SNR showed only a minor de-crease from 33.3 to 32.5. The mean CNR of the malignant lesions was the high-est in the contrast-enhanced-T2*w sequence as compared to the post- and thepre-contrast T2w-TSE sequence (p < 0.01). With a threshold PSIL of 25 %, the

sensitivity and specificity for the characterization of malignant lesions was 97.8%respectively 92.9%. Contrast-enhanced-T2*w images showed a superior imagequality and lesion conspicuity (p < 0.05).Conclusion: Solid benign lesions showed SPIO uptake with a significant signalloss in contrast to malignant lesions. When a threshold of 25% is used, PSIL is anaccurate tool for the characterisation of benign and malignant lesions. For thedetection of malignant lesions, the add-on of a contrast-enhanced-T2*w-GREsequence is helpful due to the higher CNR, a superior lesion conspicuity andbetter image quality.

B-567 11:24

The use of combined quantitative and qualitative analysis for thedifferentiation of focal liver lesions with magnetic resonance imagingA. Cieszanowski, W. Szeszkowski, M. Golebiowski, M. Grodzicki, M. Krawczyk,O. Rowinski; Warsaw/PL ([email protected])

Purpose: To evaluate the efficacy of the protocol of combined quantitative andqualitative analysis for the differentiation of focal liver lesions.Methods and Materials: The study group included 168 patients with 292 liverlesions confirmed by histology (n = 138) or follow-up (n = 154). Lesions were di-vided into: benign lesions treated conservatively (group-A:120 lesions), malig-nant tumors and benign lesions treated surgically (group-B:172 lesions). MRimaging (1.5-T) consisted of sequences: T2-double-echo-TSE, T2-STIR, T1-GREand dynamic study. The first part of differentiation protocol-quantitative analysis,based on lesions T2-relaxation times calculated from T2-double-echo TSE se-quence, was performed in order to discriminate non-solid (n = 88) from solid tu-mors (n = 204). Subsequently, all tumors defined as solid underwent qualitativeevaluation (based on visual assessment of lesions SI), in order to discriminatebenign lesions (FNH, focal fatty infiltration) from other solid tumors. Remainingtumors were defined as group-B lesions.Results: Statistically significant difference between mean T2-relaxation time ofsolid (84.1 ms) and non-solid lesions (250.5 ms) was noted, allowing diagnosisof solid tumors with sensitivity of 96% and specificity of 93% (at the threshold of116 ms). 202 lesions were defined as solid (196 true positive,8 false negative,6false positive results). Qualitative analysis allowed correct characterization of all7 focal fatty infiltrations and 21 of 24 FNH. Remaining 168 lesions were definedas group-B lesions with sensitivity and specificity of 92% (158 true positive, 14false negative and 10 false positive diagnoses were noted).Conclusion: Combined protocol of quantitative and qualitative analysis enableddiscrimination of group-B lesions with high sensitivity and specificity of 92%.

B-568 11:33

Double contrast liver MR imaging with administration of gadobenate-dimeglumine followed by ferucarbotran injection: Value for detection andcharacterization of focal hepatic lesionsS. Kim, J. Lee, J. Lee, S. Kim, J. Han, B. Choi; Seoul/KR([email protected])

Purpose: To evaluate the diagnostic performance of double contrast MR imagingwith sequential administration of gadobenate-dimeglumine (Gd-BOPTA) and fer-ucarbotran (Resovist) for detection and characterization of focal hepatic lesions.Methods and Materials: Unenhanced Gd-BOPTA-enhanced dynamic MR imag-ing using 3D-gradient echo sequence, and then Resovist-enhanced MR imagingusing T2-weighted turbo spin-echo and T2*-weighted gradient echo sequenceswere obtained in 57 patients with focal hepatic lesions. Among total of 121 le-sions, 89 lesions were confirmed by pathology. Thirty two benign lesions werediagnosed by a consensus panel of the all-clinical and imaging data, and at least6 months follow-up. Two radiologists independently reviewed the dynamic MRimaging set and the double contrast set, respectively. In both sets, sensitivity andaccuracy of detection and characterization were compared by alternative ROCanalysis.Results: Overall detection accuracy (Az) of double contrast set (0.902 for reader1, 0.870 for reader 2) was better than that of dynamic set (0.836 for reader 1,0.828 for reader 2) (P <.05 for reader 1). Both readers were more accurate indifferentiating benign from malignant lesions with double contrast set (Az: 0.882for reader 1, 0.929 for reader 2) than with dynamic set (Az: 0.832 for reader 1,0.804 for reader 2) (P <.05). The sensitivities of detection and characterizationfor the double contrast set were higher than dynamic set.Conclusion: Double contrast liver MR imaging with sequential use of Gd-BOPTAand Resovist in single MR examination may improve the accuracy of lesion de-tection and characterization compared to Gd-BOPTA-enhanced dynamic MRimaging.

Scientific Sessions

270 C D E FBA G

B-569 11:42 !Clinical significance of the diffusion-weighted imaging of focal malignanthepatic lesions at 3 Tesla magnetic resonanceD. Yu, C. Li, X. Zhang, J. Xiu, Q. Wang; Jinan/CN ([email protected])

Purpose: To investigate the clinical significance of the diffusion-weighted imag-ing of focal malignant hepatic lesions with 3.0 Tesla magnetic resonance.Methods and Materials: Magnetic resonance diffusion-weighted imaging (DWI)was obtained in a total of 47 cases with focal malignant hepatic lesions including:hepatocellular carcinoma (HCC), cholangiocellular carcinoma (CCC) and sec-ondary tumors of liver proved pathologically. The characteristics of the DWI andthe apparent diffusion coefficient (ADC) values of these diseases were analyzedwhen the b value was 600s/mm2.Results: The lesions with hyperintensity, heterogeneity or indistinct rim were in46, 24, or 21 cases on DWI images respectively. The ADC value of the HCC,CCC or secondary tumors of l iver was (145.86 ± 33.97)×10-3 mm2/s,(185 ± 53.09)×10-3 mm2/s or (212.62 ± 81.72)×10-3 mm2/s respectively. There wasa significant difference of the ADC values between in liver tissue and in second-ary tumors (t = 3.202, P=0.011), and the ADC values were higher in HCC thanthat in CCC or in the secondary tumors (P < 0.05).Conclusion: DWI at 3.0 Tesla MR may be used to improve the accuracy in diag-nosis of malignant lesions.

10:30 - 12:00 Room E2

GI Tract

SS 1401bEsophageal and gastric cancerModerators:F.-T. Fork; Malmö/SEC. Triantopoulou; Athens/GR

B-570 10:30

Staging of oesophageal carcinoma with integrated positron emissiontomography: Computed tomography (PET-CT) versus PET with diagnosticCTW.-. Wong1, V. Goh1, J. Chambers1, N. Wakeham1, K. Goodchild1, M. Harrison1,E. Townsend2, R.G. Berrisford3; 1Northwood/UK, 2Harefield/UK, 3Exeter/UK([email protected])

Purpose: To identify the benefit of FDG PET-CT over FDG PET with diagnosticCT in staging thoracic oesophageal and gastro-oesophageal junction cancer.Methods and Materials: Following ethical approval, 43 consecutive patients (15male, 6 female; mean 68 years) with proven oesophageal cancer were examinedprospectively with PET-CT (Discovery ST, GE Healthcare Technologies). Follow-ing IV injection (4.5 MBq/kg FDG), scan acquisition (5 minutes emission scan/bed position) was commenced 1 hour post injection; iterative reconstruction wasperformed to view images. All studies were read in consensus by two radiologistswith PET experience. Initially the PET alone was read with the diagnostic CT;then the combined PET-CT was read; overall stage was noted at each instance.Comparison was made with the original clinical stage and any change in man-agement was subsequently recorded.Results: There were 34 adenocarcinoma, 11 squamous carcinoma; 29 thoracicoesophagus, and 14 gastro-oesophageal junction. In 12 (23%) patients PET-CTchanged stage from M0 to M1; in 5 this was due to unexpected nodal involvement(M1a) and in 7 due to distant metastases (M1b- liver (4), lung (2), bone (1)). In 4M1 disease was shown only on PET-CT; this included 1 M1a, and 3 M1b disease(liver (2), bone (1)). In 2 further patients PET-CT showed a asymptomatic color-ectal cancer, subsequently confirmed histologically. Thus PET-CT specifically in-fluenced management in 5 cases (12%).Conclusion: In thoracic oesophageal and GOJ cancer, FDG PET-CT conferssignificant additional benefit to PET read with diagnostic CT, providing valuableincremental information which results in change of clinical management.

B-571 10:39

The CT patterns of disease recurrence following oesophagectomy forcarcinoma of the gastro-oesophageal junctionA.M. Riddell, G. Brown, A.C. Wotherspoon, J.N. Thompson, W.H. Allum;London/UK ([email protected])

Purpose: This study describes the CT patterns of disease recurrence for tu-mours of the gastro-oesophageal junction (GOJ) in patients post neo-adjuvant/

adjuvant chemotherapy and relates this to tumour site (Type I, II or III) and stage,to establish whether tumour site can help predict recurrence patterns.Methods and Materials: Imaging and clinical data were collected for patientswho underwent oesophagogastrectomy between February 1997 and March 2004with respect to tumour location, UICC stage, the time and recurrence location(locoregional or metastatic).Results: There were 82 patients; F ratio 3.5:1. The median age was 59 years(range 29-78), 72 patients had adenocarcinoma and 10 squamous cell carcino-ma. The median follow-up was 21.5 (2-89) months.There were 30 Type I, 44 type II and 8 Type III tumours. 9/30 Type I tumours(30.0%) were Stage III or above, compared with 20/44 (45.4%) Type II and 4/8(50.0%) of Type III tumours.Forty four patients had CT and/or histologically confirmed recurrence (54.0%).Median time to recurrence was 9.0 (3-42) months. The timing and location ofrecurrence (locoregional or metastatic) was similar for all GOJ tumour subtypes.However significantly more, 86.5% (7/8) of Type III tumours recurred comparedwith only 46.7% (14/30) of Type I (p = 0.05). Sites of metastatic recurrence in-cluded; the liver (7/26, 26.9%), peritoneum (5/26, 19.2%) and brain (5/26, 19.2%).The median survival post recurrence was 6 (0-23) months.Conclusion: GOJ tumour subtypes have similar CT patterns of recurrence, butour study indicates Type III tumours are significantly more likely to recur.

B-572 10:48

Optimal oral preparation for gastric cancer stagingI. Britton1, S. Chang2, D. Wright2; 1Stoke-upon-Trent/UK, 2Boston/UK([email protected])

Purpose: Optimal gastric distension is essential to ensure normal mucosal foldsare not mistaken for tumour growth. Whole milk contains 4% fat, providing nega-tive contrast and delaying gastric emptying to produce gastric distension. Thisstudy compares different gastric preparations, to determine which oral prepara-tion provides optimal distension in gastric cancer staging.Methods and Materials: All post-surgical gastric cancer pathology reports be-tween April 2002 to April 2004 were retrospectively reviewed (n75). Of these, 50patients had preoperative staging CT scans.Three different oral preparations were in use over this time period; (1) 500-1000 mlsof water (n 14), (2) 1L of full fat milk (n 24) or (3) carbex with prone scanning (n 8).All images were acquired in 3 mm increments. The 8 mm hardcopy images werereviewed retrospectively. For each patient, maximum normal wall thickness wasmeasured in each slice and graded; grade 1 (1 - 4 mm), grade 2 (5 - 10 mm) andgrade 3 (> 10 mm). The percentage of images per patient demonstrating optimaldistension (grade 1) was calculated. Four patients with linitus plastica were ex-cluded for the purposes of statistical analysis.Results: Milk achieved the greatest percentage of images with grade 1 disten-sion (74%) compared with water (38%) and carbex/prone (41%) P < 0.04.Conclusion: Milk provides optimal reproducible distension as a negative con-trast agent, is palatable to patients and delivers cost benefits when compared tobarium or calogen preparations.

B-573 10:57

Utility of multi-detector helical CT (MDCT) in the diagnosis of proximalgastric carcinoma (PGC) with gastric bare area (GBA) invasionB. Wu, P. Min, Y. Pan; Chengdu, Sichuan/CN ([email protected])

Purpose: Patients with PGC of the posterior wall are more difficult to cure andhave poorer prognosis than those with PGC of the anterior wall. The cause mustconsist in the fact that the posterior wall of the fundus and subcardial portion (wecalled it GBA) is not covered by the visceral peritoneum. To date there has beenvery little radiologic literature assessing GBA invasion in PGC with MDCT.Methods and Materials: 148 consecutive patients with biopsy-proven PGC un-derwent MDCT scan prior to gastrectomy. We evaluated the CT images sepa-rately for the site, size, depth, lymph node and enhancement characteristic ofeach case. Each post-surgical stomach specimen was axially sectioned and com-parison was made to determine the correlation between the CT findings and thepathological examination of each tumor bearing slice.Results: MDCT correctly predicted 72 of the 94 tumors with obliteration of the fatplane between the gastric tumor and an adjacent organ or invasion of an adja-cent organ. MDCT correctly identified 65/76 patients with GBA invasion and 31/40 patients with metastatic lymph node greater than 1 cm in GBA. Patients withPGC of the posterior wall (case group, n = 90) showed significant CT features ofGBA invasion as compared to the patients with PGC of the anterior wall (controlgroup, n = 58); increased tumor infiltration into subphrenic fat, marked gastro-phrenic ligament invasion, and greater lymph nodes along the GBA or subphren-ic retroperitoneal space (p < 0.05).

Mo

nd

aM

on

da

Mo

nd

aM

on

da

Mo

nd

ayy yyy

Scientific Sessions

B 271C D E FA G

Conclusion: MDCT may be of value in assessing the important radiological char-acteristics of PGC with GBA invasion.

B-574 11:06

In vivo measurement of the surface distance from the pylorus to the earlygastric carcinoma using CT gastrographyJ. Choi1, H. Chung2, H. Ryeom2, H. Choi2, W. Yu2, J. Lee2, K. Cleary1,S.K. Mun1; 1Washington, DC/US, 2Daegu/KR ([email protected])

Purpose: Early gastric carcinomas (EGCs) are often not perceptible by surgicalpalpation or inspection. Even gastroscopy does not provide the accurate positionfor carcinomas, so determining the extent of surgical resection is extremely diffi-cult. The aim of this study is 1) to develop software measuring in vivo surfacedistance from anatomic landmarks to carcinomas using CT gastrography, 2) tocompare the distances measured by the developed software and pathologic find-ings.Methods and Materials: Twenty-one consecutive patients (male = 19, female = 2,mean age: 59) with surgically proven EGC were included in this study. Preopera-tive CT scans (Slice thickness/Reconstruction interval: 0.625 mm, kVp: 120, mAs:200) were performed after insufflations of room air via a 5-F nasogastric catheter.Three-dimensional volume rendering and surface distance measurement frompylorus to EGC were performed with the developed software. We assessed thecorrelation between the distances measured by pathologic specimen and thedeveloped software.Results: The difference between the lesion to pylorus distances measured frompathological specimens and CT gastrography was 5.04 ± 2.97 mm (range, 0-11 mm). CT gastrography-based measurement overestimated the lesion to py-lorus distance in 18 patients and underestimated in three patients. The lesion topylorus distance measured from CT gastrography was well correlated with thatmeasured from the pathological specimens (r = 0.995, p < 0.01).Conclusion: These results suggest that the surface distance from an anatomiclandmark to the EGC can be measured accurately by CT gastrography. This tech-nique could be used for preoperative localization of early gastric carcinomas todetermine the optimal extent of surgical resection.

B-575 11:15

Hydro-multidetector row CT and endoscopic ultrasound in patients withgastric and duodenal lymphoma: Comparison with histopathological resultsM. Uffmann1, A. Püspök1, M. Raderer1, C. Kölblinger1, C.M. Schaefer-Prokop2,A. Ba-Ssalamah1; 1Vienna/AT, 2Amsterdam/NL([email protected])

Purpose: To evaluate whether indolent and aggressive gastric and duodenal lym-phoma can be differentiated by hydro-multidetector CT (CT) or endoscopic ultra-sound (EUS) and to assess their correlation of findings.Methods and Materials: The study group consisted of 14 patients with histopatho-logically proven lymphoma: 9 had an aggressive lymphoma and 5 an indolentlymphoma. Lesion thickness was measured in mm and invasion of the gastric orduodenal wall was morphologically described by a 5-point-scale according to themethodological capacities (CT: 1=focal, 2=diffuse, 3=blurring of fat plane; EUS:1=mucosa, 2=muscularis, 3=serosa, for both; 0=normal, 4=infiltration of adja-cent organs).Results: With both imaging methods, wall thickness and infiltration was larger inaggressive lymphoma than in indolent lymphoma, but differences were not sig-nificant. In aggressive lymphoma, thickness of diseased wall was 0.8-12 cm withCT but only 0.3-1.3 cm with EUS. In indolent lymphoma the wall thickness was0.7-1.3 cm with CT and 0.3-0.9 cm with EUS. The mean scores for wall invasionwere 2.9 and 2.4 for aggressive lymphoma, measured with CT and EUS, respec-tively, and 1.5 and 0.8 for indolent lymphoma. For wall invasion, EUS and CTshowed a high correlation (rs = 0.8, p < .05) although the criteria used differed.For wall thickness no linear correlation was found (r = 0.1).Conclusion: Although aggressive lymphoma tends to cause higher wall thick-ness and more advanced infiltration, no clear distinction can be made based onmorphology provided by EUS or CT. Both methods showed a high correlation inassessment of wall infiltration but did not find a correlation in wall thickness.

B-576 11:24

In vitro imaging of gastric cancer: Phase contrast imaging with synchrotronradiationX.P. Zhang, L. Tang, Y.S. Sun, J. Li, G. Li; Beijing/CN ([email protected])

Purpose: To investigate the ability of phase contrast imaging with synchrotronradiation (SR-PCI) in demonstrating gastric wall layers and invasion depth ofgastric cancer in vitro.

Methods and Materials: Six specimens of gastric cancer were fixed in formalinfor 2-7 days. Then the specimens were cut into strips with a thickness of 1 mmand performed SR-PCI examination. The synchrotron radiation came from 4 W1 Abeamline of Beijing Synchrotron Radiation Facility (BSRF). Double-crystal mon-ochrometer Si (111) was employed to obtain the coherence and monochromaticX-ray. The X-ray energy was 9 keV, and the distance from the X-ray source to thespecimens was 45m. A X-ray CCD with 10 µm spatial resolution was used as thedetector. The results were compared with histopathology.Results: It could differentiate 4-5 layers of normal gastric wall with SR-PCI, whichwere mucosa, submucosa, muscularis propria and subserosa/serosa. All of thesix specimens could be correctly judged about the cancer invasion depth. It couldalso show clearly the foveolae gastricae and vessels in the submucosa whosediameter exceeding 20 µm.Conclusion: SR-PCI can demonstrate the tiny architecture of normal gastric walland gastric cancer in vitro.

B-577 11:33

Gastric cancer: Diffusion-weighted MR imaging with sensitivity-encodingtechniqueX.P. Zhang, L. Tang, Y.S. Sun, J. Li, Y. Li; Beijing/CN ([email protected])

Purpose: To determine the feasibility of diffusion-weighted MR imaging usingsensitivity-encoding technique (ASSET-DWI) in depicting gastric cancer.Methods and Materials: Twenty-three patients with gastric cancer were exam-ined at 1.5 T. One DWI sequence and three ASSET-DWI sequences with differentNEXs (Se1: 1NEX; Se2: 2NEXs; Se3: 4NEXs) were employed to generate diffu-sion-weighted images. When the acquisition time exceeded 15 seconds, the se-quence will be separated to 2-3 breath-holds using a "pause scan" function key.The b values were 0 and 1000 sec/mm2. Two radiologists evaluated the obtainedimages from the viewpoints of tumor detectability, gradation of artifacts and con-trast-to-noise ratio (CNR). The apparent diffusion coefficients (ADCs) of tumorsand water in the gastric lumen were measured and compared.Results: The EPI-related artifacts and image distortion of ASSET-DWI was re-duced compared to that without ASSET (p < 0.01). There was no influence ofseparate breath-holds on the ADC measurement (p > 0.05) of both water andtumors. Separate breath-holds with multi-NEX technique could reduce the AS-SET-related artifacts (p < 0.01) and achieve higher CNR (Se1: 12.16 ± 3.75, Se3:17.65 ± 7.59, p < 0.01). The mean ADC of gastric cancer was 1.17 mm2/s×10-

3 ± 0.22.Conclusion: High quality ASSET-DWI imaging of gastric cancer can be achievedthrough the combination of separate breath-holds and multi-NEX technique. Thediffusion-weighted MR imaging may provide functional parameters for the diag-nosis of gastric cancer.

B-578 11:42

Gastric cancer: Preoperative evaluation with MDCT and EUSR.F. Grasso, M. Sammarra, F.M. Di Matteo, R. Coppola, A. Mazzei, I. Baeli,C. Gaudino, B. Beomonte Zobel; Rome/IT ([email protected])

Purpose: We considered the role of MDCT (Somatom Sensation, Siemens) andendoscopic ultrasound (EUS)(FG3830UT Pentax Europe GmbH)in the detection,localization and staging of gastric cancer and we compare the accuracy of bothtechnique.Methods and Materials: Between September 2003 and September 2005, 53patients, 33 male and 20 female, mean age 65.3, were evaluated with conven-tional endoscopy with biopsies, EUS and MDC. We considered only patients withgastric adenocarcinoma. All the CT examinations were performed after the oraladministration of 700/900 ml of water for gastric wall distension; each examina-tion was performed with triphasic technique (basal, arterial and portal phase).The images were post-processed using Multi-Planar Reformatting (MPR). Thelesion morphology and thickness, the degree of gastric wall infiltration, locore-gional lymph nodes and systemic metastases were evaluated by two radiologistsblinded to the EUS findings. All the patients enrolled in the study underwent sur-gery; then the preoperative EUS and MDCT staging was evaluated against thepathology findings.Results: Gastric cancer was confirmed in all patients. Overall accuracy, sensitiv-ity, and specificity in the evaluation of T stage for EUS and three-dimensionalMDCT were 86%, 87%, 90% and 85%, 77.5% and 92.8% respectively. For theevaluation of N factor for EUS and three-dimensional MDCT were respectively,77%, 67.6%, 80% and 72.8%, 57% and 88%.Conclusion: Although EUS seem to have a better overall accuracy than MDCTin the evaluation of T1 and T3 tumors, three-dimensional MDCT still presentshigher specifity and remains the cross sectional imaging technique of choice inthe preoperative evaluation of gastric cancer.

Scientific Sessions

272 C D E FBA G

B-579 11:51

Diagnostic performance of 18F-FDG PET scan in detecting postoperativerecurrence of gastric cancer: Comparison with computed tomographyJ. Sun, J. Lee, J. Kim, M. Kim, H. Noh; Suwon/KR ([email protected])

Purpose: We retrospectively assessed the diagnostic performance of 18F-FDGPET and CT scan in detecting recurrent gastric cancer.Methods and Materials: Thirty-four patients who had underwent curative surgi-cal resection for gastric cancer and had subsequently undergone 18F-FDG PETfor suspected recurrent gastric cancer on prior CT scans, were enrolled in thisstudy. The interval between the CT and 18F-FDG PET scans ranged from 1 to 128days (mean 23 days). The result of CT and 18F-FDG PET scans were comparedwith either histological confirmation (11 cases) or with radiological/clinical follow-up(23 cases) if histology was negative or unavailable.Results: Disease recurrence was established in 30/35 patients (86%). The 14cases of recurred metastatic lymphadenopathy were suggested on CT scansand twelve cases of recurred lymphadenopathy were confirmed. 18F-FDG PETscans were concordant with final results in 7 patients (58%) and not in 5 patients(32%). Seven cases of hepatic metastases were depicted on CT scans (all ofthem were proved to true hepatic metastasis) and 18F-FDG PET correctly diag-nosed hepatic metastases in 4 cases (57%). CT scan suggested 4 cases of loco-regional recurrence (remnant stomach or anastomosis site) and 18F-FDG PETscans did not correctly diagnose in 3 cases (75%). Peritoneal carcinomatosiswas confirmed in 12 cases; CT scan correctly diagnosed in 9 cases (75%) and18F-FDG PET correctly diagnosed in 3 cases (25%).Conclusion: Our results indicate that 18F-FDG PET scan is not suitable modalityfor confirming or detecting recurrent gastric cancer.

10:30 - 12:00 Room F1

Chest

SS 1404Pulmonary embolism: CT diagnosis, CAD andoutcomeModerators:L. Bakon; Warsaw/PLG. Mostbeck; Vienna/AT

B-580 10:30

Integrated cardio-thoracic imaging with ECG-gated 64-slice multidetector-row CT: Preliminary experience in 133 patientsM. Rémy-Jardin, R. Salem, D. Delhaye, C. Khalil, A. Teisseire,V. Delannoy-Deken, A. Duhamel, A. Duhamel, J. Rémy; Lille/FR([email protected])

Purpose: To investigate the possibility of assessing the underlying respiratorydisease as well as cardiac function during ECG-gated CT angiography of thechest with 64-slice multidetector-row CT (MDCT).Methods and Materials: 133 consecutive patients underwent an ECG-gated CTangiographic examination of the chest without ß-blockers using the following pa-rameters: (a) collimation: 64x 0.6 mm; pitch: 0.3; (b) 120 mL of a 35% contrastagent. Data was reconstructed: (a) to evaluate the underlying respiratory disease(1-mm thick lung and mediastinal scans reconstructed at 55% of the R-R inter-val); and (b) to determine right and left ventricular ejection fractions (short-axissystolic and diastolic images).Results: The mean heart rate was 73 bpm and the mean scan time was 18.11± 2.67 sec. A total of 123 examinations (92%) had both lung and mediastinalimages rated as diagnostic whereas 10 examinations (8%) had nondiagnosticimages altered by respiratory-motion artefacts (n = 4) or cyclic artefacts relatedto the use of a pitch value of 0.3 in patients with a very low heart rate (n = 6).Assessment of right and left ventricular function was achievable in 124 patients(93%, 95% CI: 88%-97%). In the remaining 9 patients, an imprecise segmenta-tion of right and left ventricular cavities was considered as a limiting factor forcalculation of end-systolic and end-diastolic ventricular volumes. The mean (± SD)DLP value of the examinations was 279.86 (± 117.50) mGy.cm.Conclusion: Assessment of underlying respiratory disease and cardiac functionfrom the same data set was achievable in 92% of patients with ECG-gated 64-slice MDCT.

B-581 10:39

Low-dose ECG-gated 64-slice CT angiography of the chest: Preliminaryexperience in 105 patientsA.-G. D'Agostino1, C. Khalil1, M. Rémy-Jardin1, T. Flohr2, A. Duhamel1,J. Rémy1; 1Lille/FR, 2Forchheim/DE ([email protected])

Purpose: To evaluate image quality of low-dose ECG-gated CT angiograms ofthe chest obtained with 64-slice CT technology.Methods and Materials: 105 consecutive patients underwent an ECG-gated CTangiographic examination of the chest using (a) collimation 64x 0.6 mm; rotationtime 0.33 sec; pitch 0.3; (b) 120 kV; 200 mAs; automatic milliamperage modula-tion. Subjective and objective image quality was evaluated on 3-mm thick scansby consensus between two radiologists.Results: Subjective image quality analysis showed that: (a) graininess was presenton 8 examinations (8%), always rated as mild; (b) artefacts due to the pitch valueselected in patients with a low heart rate were observed on 22 examinations(20%), with a severity significantly linked to the patient's bradycardia (p < 0.0001).Objective image quality was assessed by measuring the standard deviation ofthe attenuation of the pixels in a homogeneous region-of-interest within the tra-chea (15.91 HU) and descending aorta (22.16 HU). Among the studied popula-tion, (a) 99 scans (94%) were either completely devoid of artefacts (n = 93) orshowing mild artefacts which did not alter their diagnostic value (n = 6); (b) 6scans (6%) were degraded by artefacts, compromising the reading of the study.The mean DLP of the examinations was 274.34 mGy. cm which is lower thanEuropean reference levels for chest CT (500 - 650 mGy.cm).Conclusion: Low-dose ECG-gated CT angiography of the chest was obtained inroutine clinical practice with 64-slice CT technology without altering the diagnos-tic value of CT scans in 94 % of patients.

B-582 10:48

Cardiogenic artefacts on multidetector CT angiograms of the pulmonarycirculation: Comparison of ECG-gated and nongated acquisitionsM. Rémy-Jardin, A. Teisseire, J. Bruzzi, J. Kirsch, C. Khalil,V. Delannoy-Deken, J. Rémy; Lille/FR ([email protected])

Purpose: To evaluate the impact of temporal resolution on the frequency of car-diogenic motion artefacts.Methods and Materials: This study included 135 patients who underwent amultidetector CT angiogram of the pulmonary circulation with a temporal resolu-tion of 500 ms (Group 1; n = 37; 16-slice MDCT), 375 ms (Group 2; n = 32; 16-slice MDCT), or 165 ms (Group 3; n = 66; retrospective ECG-gated 64-sliceMDCT). The presence, location and severity of cardiac motion artifacts at thelevel of pulmonary arteries were analyzed on 1 mm-thick CT angiograms ob-tained in strict apnea.Results: Considering central pulmonary arteries, the frequency of cardiogenicartefacts in Group 3 was significantly lower than that observed in Group 1 andGroup 2 at the level of : (a) the pulmonary trunk: n = 6 (9%; Group 3) vs n = 4(12.5%; Group 2), n = 25 (68%; Group 1) (p < 0.001); (b) right main pulmonaryartery: n = 0 (0%; Group 3) vs n = 5 (16%; Group 2), n = 18 (49%; Group 1)(p < 0.001). Considering intraparenchymal pulmonary arteries, the frequency ofcardiogenic artefacts was significantly lower in Group 3 (n = 9; 14%) comparedto Group 2 (n = 22; 69%) and Group 1 (n = 34; 92%) (p < 0.001). Whereas car-diogenic motion artefacts were responsible for pseudo-filling defects in 20 pa-tients in Group 1 (54%) and in 8 patients in Group 2 (25%), only 1 patient showedpseudo-filling defects in Group 3 (1.5%) (p < 0.0001).Conclusion: Motion artefacts in the pulmonary arteries causing interpretive dif-ficulties on multislice CT scans can be significantly reduced by ECG gating.

B-583 10:57

Multidetector-row CT of suspected and unsuspected acute pulmonaryembolism: Prognostic value of morphological cardiac parametersC. Engelke, E.J. Rummeny, K. Marten; Munich/DE ([email protected])

Purpose: Evaluation of the prognostic value of morphological cardiac parame-ters in patients with suspected and incidental acute pulmonary embolism (PE)using multidetector-row chest CT (MSCT).Methods and Materials: 2335 consecutive MSCT scans were evaluated for pres-ence of PE. Arterial enhancement, artefacts, and ability to analyze pulmonaryarteries and the heart were assessed as parameters of scan quality. Diastolicright and left ventricular short axes (RVD, LVD) and the interventricular septal de-viation (ISD) were measured in PE-positive patients and echocardiography re-ports reviewed. Clinical data assessment included cardio-respiratory and otherco-morbidities, systemic anticoagulant therapy (ACT) and 30-day outcome. Pre-

Mo

nd

aM

on

da

Mo

nd

aM

on

da

Mo

nd

ayy yyy

Scientific Sessions

B 273C D E FA G

dictors of acute cor pulmonale (CP) and short-term outcome were calculated bylogistic regressions and ROC analyses.Results: 90 patients with acute PE were included (36 with clinically suspectedPE, 54 with incidental PE). 26 patients had cardio-respiratory co-morbidities. Fourpatients underwent systemic thrombolysis, 43 anticoagulation in therapeutic doses,19 in prophylactic doses, and 24 patients no ACT. 15 of 41 patients had echocar-diographic evidence of acute CP. 8 patients died within 30 days. The RVD was thebest independent predictor of acute CP (p = 0.002, OddsRatio (OR)=9.16, posi-tive-predictive-value (PPV)=0.68, negative-predictive-value (NPV)=1 at 4.49 cmcutoff) and short-term outcome (p = 0.0005, OR=2.82, PPV=0.23, NPV=0.98 at4.75 cm cutoff). The RVD/LVD

ratio had a PPV of 0.85 for CP.

Conclusion: RVD and RVD/LVD ratio were suitable to identify patients with acute

CP and benign short-term prognosis. Further studies should prospectively ad-dress the combined use of CT-morphological and clinical parameters for predic-tion of patient outcome.

B-584 11:06

To assess the added value of a prototype computer-aided detection (CAD)system for pulmonary embolism (PE) in contrast-enhanced multi-detectorcomputed tomography (CT) imagesC. Schaefer-Prokop1, O. van Delden1, H. Bouma2, J.J. Sonnemans3,F.A. Gerritsen3, J.S. Lameris1; 1Amsterdam/NL, 2Eindhoven/NL, 3Best/NL([email protected])

Purpose: To assess the added value of a prototype computer-aided detection(CAD) system for pulmonary embolism (PE) in contrast-enhanced multi-detectorcomputed tomography (CT) images.Methods and Materials: Datasets used for training of the CAD system (n = 20)and evaluation (n = 13) were obtained with a 4-slice scanner with a pixel size of0.6 mm, a spacing between slices of 0.6 mm and a slice thickness of 1.3 mm.The study group consisted of 13 CTPA studies with a mean number of 4 thrombi(range 1-10). Datasets were selected to demonstrate a variety of thrombus load,considerable breathing artifacts (n = 2), sub-optimal contrast (n = 1) and addi-tional parenchymal disease (n = 6). One experienced radiologist was asked toannotate all thrombi. The most proximal location of the thrombus defined its ana-tomic classification (main, lobar, segmental or sub-segmental), the thrombus sizedefined its conspicuity (obvious, medium or subtle).Results: Without CAD, the radiologist annotated 37 emboli located in the main(n = 7), lobar (n = 6), segmental (n = 14) and sub-segmental arteries (n = 10).With CAD, 12 additional thrombi were detected (increment of 32%), located insegmental (n = 2/16, 12%) and sub-segmental arteries (n = 10/20, 50%). Throm-bi additionally detected by CAD were classified as being of subtle conspicuity (9/12, 75%) or medium conspicuity (3/12, 25%). Per dataset the CAD system showedon average 8.7 false positive calls (median = 7).Conclusion: Compared to the radiologist, the prototype CAD system found addi-tional emboli in segmental and sub-segmental arteries. These preliminary resultssuggest that CAD may become a helpful tool for diagnosing low grade acutepulmonary embolism in peripheral arteries.

B-585 11:15

Clinical evaluation of a CAD prototype for the detection of pulmonaryembolismS. Buhmann1, P. Herzog1, J. Stoeckel2, M. Salganicoff2, M. Wolf2, M.F. Reiser1,C.R. Becker1; 1Munich/DE, 2Malvern, PA/US([email protected])

Purpose: To evaluate the performance of a prototype CAD-tool using artificialintelligence for the detection of pulmonary embolism (PE) and the benefit forradiologists.Methods and Materials: 40 MDCT datasets using 100 kVp, 100 mAs effectiveper slice and 1 mm thick axial reformats in a low frequency reconstruction kernelwere evaluated. 80 cc iodinated contrast agent was injected with a 5 cc/sec flowrate.Primarily, 6 community level radiologists marked any PE using dedicated lungevaluation software for visualization with consecutive processing with the PE-CAD- prototype. An expert panel consisting of two dedicated chest radiologistshad to confirm any PE marked by the readers and/or the CAD-prototype and tomark any additional PE primarily missed by both, creating the reference stand-ard.Results: The 6 community level readers detected 166 emboli without CAD, 38 %central and 62% peripheral, with an average of 4.15 findings/case whereas theCAD-tool detected 322 candidate lesions, 87 (27%) central and 235 (73%) pe-ripheral. The experts confirmed 168 (52%, 4.2 findings/case) lesions detected by

CAD, 48 (29%) central and 120 (71%) peripheral emboli, while 154 (48%, 3.85findings/case) were dismissed as false positives. The experts confirmed 352 emboliresulting in 8.8 findings/case, including the lesions correctly detected by the CAD-tool.Conclusion: The investigated CAD-tool increased the number of PE found bycommunity level radiologists by a factor of 2, but requires a lesion-by-lesion dis-missal of false positives. Future CAD-tools have the potential to improve the as-sessment of emboli while today's tool still suffers from a considerable false positiverate.

B-586 11:24

Visualization of candidate filling defects deriving from a CAD program forthe detection of pulmonary embolism using automatic vessel aligned,rotating MPRsS. Buhmann1, P. Herzog1, M. Salganicoff2, J. Stoeckel2, M.F. Reiser1,C.R. Becker1; 1Munich/DE, 2Malvern, PA/US([email protected])

Purpose: To evaluate vessel aligned, rotating MPRs for visualization/evaluationof candidate lesions deriving from a prototype CAD-tool for the detection of pul-monary embolism (PE) on MDCT datasets.Methods and Materials: 40 MDCT datasets using 100 kVp, 100 mAs effectiveper slice and 1.0 mm thick axial reformats were evaluated. 80 ml iodinated con-trast material was administered at a 5 ml/sec flow rate.A prototype CAD tool marked all PE candidate lesions. 2 experienced radiolo-gists evaluated them randomly, with and without using vessel aligned MPR pro-vided by dedicated lung evaluation software. The readers confirmed each markdepending on real present PE. The level of confidence of the readers and theevaluation time was recorded for each mark.Results: The CAD-tool detected 322 candidates, 87 (27%) central and 235 (73%)peripheral. The readers confirmed 168 (52%) lesions, 48 (29%) central and 120(71%) peripheral emboli. 154 (48%) detected lesions were dismissed as falsepositives. The rate of confirmed lesions did not significantly depend on the use ofMPRs (168 vs. 163). The readers' confidence level was significantly higher usingMPRs (8.7 vs. 5.2) while the reading time/lesion was only marginally increased(44 vs. 39 seconds, 12%). The interreader agreement was significantly increasedusing rotating MPRs (0.81 vs. 0.62)Conclusion: Rotating, automatic vessel aligned MPRs showed their potential tofacilitate the reading even in experienced readers. They could help to cope withthe abundance of candidate lesions generated by today's and tomorrow's CAD-programs especially when reading chest CT scans on a day-by-day basis.

B-587 11:33

Mean pulmonary arterial pressure estimation by means of high temporalresolution phase-contrast MR imaging in patients with chronicthromboembolic pulmonary hypertensionR.P. Kunz, F. Gössler, S. Herber, T. Kramm, R.S. von Bardeleben, K. Koch,M.B. Pitton, C. Düber, K.-F. Kreitner; Mainz/DE([email protected])

Purpose: To estimate mean pulmonary arterial pressure (mPAP) non-invasivelyin patients with chronic thromboembolic pulmonary hypertension (CTEPH) byanalysing flow parameters of the pulmonary trunk derived from high temporalresolution phase-contrast MR imaging (PC-MRI).Methods and Materials: Retrospective analysis of ten CTEPH patients (medianage: 60 years) who underwent both selective pulmonary DSA with haemody-namic catheter measurements as well as a comprehensive MR imaging protocol(Sonata syngo MR 2002B, Siemens) including high temporal resolution PC-MRIof the pulmonary trunk (WIP FlowM2: TR 12.0 ms, TE 3.1 ms, Venc 150 cm/s,BW 698 Hz/pixel) in the diagnostic work-up for pulmonary endarterectomy (PEA)within the period from 07/2003 to 12/2004. The parameters absolute and relativeacceleration time ATa [ms] and ATr [%], peak of mean velocities MFV [cm/s] anddistensibility DIS [%] were evaluated according to Abolmaali et al. (Fortschr Rönt-genstr 2004; 176:837-845 and Proc Intl Soc Mag Reson Med 2003; 11:407) andused to compute mPAP with the following regression equation: mPAP-MRI = 88.9- 0.29 x ATa - 1.1 x ATr - 0.38 x MFV - 0.16 x DIS.Results: mPAP-MRI correlated very well with invasive pressure recordings in 8of 10 patients (linear coefficient of determination R2 = 0.983; median difference1.8 mm Hg, range -3.0 - +6.2 mm Hg). PC-MRI underestimated mPAP-DSA intwo patients by 20 mm Hg and 25 mm Hg (outliers).Conclusion: Non-invasive mPAP-estimations are possible in CTEPH patients bymeans of high temporal resolution PC-MRI and correlate very well with pulmo-nary artery catheterization.

Scientific Sessions

274 C D E FBA G

B-588 11:42

Pulmonary embolism on multislice CT of the chest: One-year survival oftreated and untreated patientsC. Engelke, E.J. Rummeny, K. Marten; Munich/DE ([email protected])

Purpose: To retrospectively assess the outcome of clinically unsuspected pul-monary embolism (PE) on chest multidetector-row CT (MDCT).Methods and Materials: 2536 consecutive chest MDCT scans of 1966 patients(mean age 60 years (15-96 y); male/female = 1.79) within a one-year period werereviewed for PE. PE was assessed by a CT severity scoring system (SSC). Cor-rect-positive and false-negative studies were determined. Co-existing morbidity,anticoagulant therapy (ACT), complications of PE and therapy and one-year out-comes were reviewed. Statistics included Mann-Whitney-U, Chi2, Poisson-regres-sion and Kaplan-Meier statistics.Results: 117 patients were PE-positive. File review was complete in 96 patients,(active malignancy=63/96, unsuspected PE=58/96). 38/58 patients with unsus-pected PE had false-negative CT-reports (mean SSC of 20.21 ± 17.88 and9.55 ± 7.12 for correct-positives and false-negatives, respectively; p=.012). 49patients received therapeutic ACT, 21 had prophylactic ACT and 26 had no treat-ment. PE severity was significantly higher in patients with therapeutic vs. thosewithout therapeutic ACT (p <.0002). Bleeding complications were more frequentwith therapeutic ACT (2 early deaths, 5 major nonfatal hemorrhages), than with-out (1 minor hemorrhage; p=.037). There were 8 early deaths (7 with therapeuticACT, 1 without ACT due to contraindications; p=.037). Positive predictors of earlydeath included SSC > 28, systemic thrombolysis, major hemorrhage and newonset cardiac/renal failure (p=.001-.043). Negative predictors were a false-nega-tive report and no therapeutic ACT (p=.007-.037). Predictors of late death (n = 25)were old age, malignancy and renal failure (p=.001-.043).Conclusion: Unsuspected PE may be undetected on routine MDCT of the chest.These patients may have a favorable short-term outcome even without therapeu-tic ACT.

B-589 11:51

Acute pulmonary embolism on multidetector-row CT of the chest:Prediction of cor pulmonale and short-term patient survival frommorphological embolus burdenC. Engelke, E.J. Rummeny, K. Marten; Munich/DE ([email protected])

Purpose: We retrospectively investigated three morphology-based multidetec-tor-row CT (MDCT) pulmonary arterial (PA) embolic (PE) obstruction scoringsystems to predict cor pulmonale (CP) and short-term outcome.Methods and Materials: Eighty-nine consecutive patients (median age 66 years[23-83] m/f=51/38) with acute PE on MDCT were included. 64 patients had co-existing malignancy. PE-severity was assessed by two blinded observers usingthree PA obstruction indices (two scores after Miller et al. (excluding and includ-ing arterial branch obstruction grading: scores A and B, respectively) and onescore after Mastora et al. (index C)). Echocardiographic reports were reviewedfor right ventricular dysfunction. Obstruction index results were analyzed for pre-diction of PA pressures, CP and 30-day survival. Statistics included Kappa, ANO-VA, linear correlation, Chi2 and logistic regressions.Results: Interobserver agreement of PE-severity was κ=0.89, 0.86 and 0.78 forindices A, B and C, respectively. PE-severity varied significantly between thescores (median = 25.0% [6.3-100], 12.5% [3.1-65.6] and 7.1% [0.65-65.8] for in-dices A, B and C, respectively; ANOVA p < 0.0001). Index C correlated best withPA pressures (r = 0.69, p < 0.0016) and presence of CP (p = 0.0051; odds ratio(OR)=1.20 [95%CI=1.10-1.36]; for index C cutoff=21.3%: p < 0.001; positive pre-dictive value (PPV)=1, negative predictive value (NPV)=0.87). Eight patients diedwithin 30 days after CT. Using score C, PE-severity was a significant predictor ofearly death (p = 0.018; OR=1.03 [95%CI=1.00-1.06]; for index C cutoff=21.3%:p = 0.018; overall OR=6.77; PPV=0.24; NPV=0.96).Conclusion: The "Mastora" obstruction index was a significant predictor of corpulmonale and short-term outcome and may therefore allow for therapy and riskstratification in patients with acute PE.

10:30 - 12:00 Room F2

Cardiac

SS 1403Non-ischemic cardiomyopathiesModerators:A.J.B.S. Madureira; Porto/PTL. Natale; Rome/IT

B-590 10:30

3T MR imaging in acute myocarditisS. Ferraresi, G. Ligabue, F. Fiocchi, C. Ricci, A. Barbieri, M. Modena,R. Romagnoli, P. Torricelli; Modena/IT ([email protected])

Purpose: Describe imaging features of 3T MR imaging in patients with clinicaldiagnosis of myocarditis.Methods and Materials: Twelve symptomatic males (mean age 30.2 ± 10) withaltered electrocardiogram and echocardiography, increased troponin and CK-MBvalues, negative coronary angiography underwent cardiac MR imaging within 2days after admission. Cardiac 3T MR imaging was performed including the fol-lowing breath-hold ECG-triggered sequences: T2w, Cine-TFE, IR-GE T1w preand post gadolinium. Follow-up was performed within two months. We evaluatedsignal intensity changes in T2 images, delayed enhancement (DE), Cine MR al-teration and their evolution during follow-up.Results: At baseline, focal or diffuse T2 increased intensity of left ventricle waspresent in 8 patients: 6 in posterolateral and 2 in antero-lateral wall. DE was focalor diffuse but always related to epicardium and evident in 8 cases in postero-lateral and in 2 in antero-lateral wall. Two patients had no T2 hyperintensity andno DE. Pericardial effusion was present in 6 patients. EF was depressed in 4patients and normal in the remaining. At follow-up MR imaging, T2 alterationsdisappeared, DE was reduced in 7 and unmodified in 3 cases, the remaining 2were still negative. Mean EF improvement was 8%. One patient evolved in dilata-tive cardiopathy one year later.Conclusion: Gadolinium enhancement is a more accurate finding respect T2alteration in acute myocarditis and has a typical localization and distribution. Itmay therefore serve as a powerful noninvasive tool to assess the presumptiveclinical diagnosis and help in the management of this disease.

B-591 10:39

Delayed enhanced cardiac magnetic resonance in patients with biopsyproved chronic myocarditisA. Esposito1, F. De Cobelli1, M. Pieroni2, E. Belloni1, R. Mellone1, F. Gentinetta1,A. Frustaci2, A. Maseri1, A. Del Maschio1; 1Milan/IT, 2Rome/IT([email protected])

Purpose: Chronic myocarditis (CM) is a common evolution of acute myocarditissustained by viral persistence and/or autoimmune self-perpetration. Detection ofchronic myocardial inflammation has relevant therapeutic implication. Delayed-enhanced cardiac magnetic resonance (DE-CMR) has been revealed useful inidentifying areas of myocardial damage in acute myocarditis. Our aim was toassess the effectiveness of DE-CMR in evaluation of CM.Methods and Materials: 23 patients with CM underwent CMR and endomyocar-dial-biopsy (EB). Diagnostic criteria for CM were: 1)chronic (> 6 months) heartfailure and/or ventricular arrhythmias, 2)no recent flu-like symptoms or infections,3)histological evidence of active (AM) or borderline myocarditis (BM) accordingto Dallas criteria at EB. CMR included T2-weighted (BB-T2w) images also withfat-sat, cine-SSFP images and DE-CMR study performed at 1.5 T (Philips). DE-CMR images were acquired 10-15 minutes after gadolinium injection with T1w-TFE-IR sequence.Results: Histology showed AM in 14 patients and BM in 9 patients. BB-T2w fat-sat images showed the presence of oedema in only 5 out of 14 (36%) patientswith AM. In the same group, DE-CMR revealed delayed-enhanced (DE) areas in12 patients (84%) with midwall or subepicardial pattern of distribution. No evi-dence of oedema was found in the patients with BM, and only 4 of them (44%)presented DE areas associated with replacement fibrosis at EB.Conclusion: Areas of DE are frequently present in patients with CM. DE areasare associated with histological evidence of active inflammation or fibrosis. DE-CMR may be a useful non-invasive diagnostic tool to evaluate CM patients and toindicate and guide EB.

Mo

nd

aM

on

da

Mo

nd

aM

on

da

Mo

nd

ayy yyy

Scientific Sessions

B 275C D E FA G

B-592 10:48

Comparison of different cardiac magnetic resonance imaging (MRI)sequences in patients with persistent virus myocarditisB. Spors, S. Mehl, T. Thoma, H. Bertram, R. Felix, H.-P. Schultheiss, U. Kühl,M. Gutberlet; Berlin/DE ([email protected])

Purpose: To evaluate different cardiac MR imaging sequences and analyse glo-bal relative myocardial enhancement, late enhancement and function analysis inpatients with myocarditis.Methods and Materials: Cardiac MR imaging was performed in 62 patients (meanage 49 ± 13) with positive endomyocardial biopsies (EMB) acquired more than 3months after the clinical onset of acute myocarditis. To assess different CMRapproaches for myocarditis evaluation we acquired: STIR-, T2- and T1-weightedFast Spin Echo (FSE) images before and after contrast agent administration todetect myocardial edema and inflammation as well as inversion recovery gradi-ent-echo sequences (IR-GRE) and steady-state free-precession (SSFP) sequenc-es to assess myocardial fibrosis and function.Results: 45 patients had a simple virus infection, 17 double infections (parvovi-rus B19, n = 59; human herpes type 6 [HHV 6], n = 16; enterovirus, n = 4. Edemawas present in 4%, increased global relative enhancement in 40% a indicatorsfor active inflammation, which was histopathologically (EBM) obvious in 61%,and late enhancement (LE) in 22 of the patients as an indicator for irreversiblyinjured myocardium. LE was slightly more common in patients with double infec-tion (37 vs. 24%).Conclusion: In patients with chronic myocarditis and cardiotropic viral persist-ence LE is a common finding but less common as in acute myocarditis, further-more, it is more frequent in patients with double infections. Not only one sequencemay be useful to detect myocarditis.

B-593 10:57

Late enhancement and Fabry cardiomyopathy: Impact of enzymereplacement therapy on cardiac morphology and functionM. Beer, F. Weidemann, F. Breunig, W. Machann, C. Wanner, D. Hahn;Würzburg/DE ([email protected])

Purpose: Enzyme replacement therapy (ERT) potentially reduces left ventricularhypertrophy and improves cardiac dysfunction in most patients with Fabry dis-ease. Whether those with longstanding Fabry disease with evidence of late en-hancement in left ventricular myocardium respond to ERT is unknown.Methods and Materials: We conducted a prospective observational study in thirty-five consecutive patients (20 male, 39 ± 10years) with genetically proven Fabrydisease and assessed the occurrence of late enhancement by magnetic reso-nance imaging. Seventeen patients were eligible to receive ERT with agalsidaseß and were followed by magnetic resonance imaging and coloured Doppler im-aging over a period of 12 month.Results: Late enhancement of left ventricular myocardium was detected in 11/35patients (31%; two female). Left ventricular mass was higher in late enhance-ment positive as compared to late enhancement negative patients (p = 0.003).During ERT (n = 17) left ventricular mass significantly decreased in patients with-out late enhancement (n = 9; 160 ± 23 g vs. 145 ± 27 g; p = 0.008) but not in thosewith late enhancement (n = 8; 211 ± 58 vs. 195 ± 64 g; p = 0.161). In parallel,improvement of cardiac function, as demonstrated by coloured Doppler imaging,was restricted to segments without late enhancement.Conclusion: Late enhancement is prominent in patients with Fabry disease, cor-relates with increased left ventricular hypertrophy and may predict the effect ofERT on left ventricular mass and cardiac function.

B-594 11:06

Enzyme replacement therapy in patients with Fabry diseaseW. Machann, J.J. Sandstede, F. Weidemann, F. Breunig, D. Hahn, M. Beer;Wuerzburg/DE ([email protected])

Purpose: The aim of this study was to find out whether late enhancement (LE) inmagnetic resonance imaging (MRI) as a marker of myocardial fibrosis is a goodparameter to establish a success of enzyme replacement therapy (ERT) in pa-tients with Fabry disease.Methods and Materials: 17 patients (13 men/4 women; age 41 ± 8) with genet-ically proven Fabry disease received ERT (agalsidase ß) and were followed for12 months. At baseline and after a year all 17 patients underwent cardiac MRimaging and additionally colour Doppler myocardial imaging to confirm an im-provement of the regional function under ERT.Results: 8/17 (47%) patients showed LE at baseline (1.3 ± 1.1% of left ventricu-lar (LV) mass). The regional function was reduced in segments with LE. The LV

mass of patients without LE at baseline was higher than in patients with LE(160 ± 23 g vs. 211 ± 58 g; p = 0.046). The LV mass at the follow-up after a yearof ERT only decreased in patients without LE (145 ± 27 g; p = 0.008). An im-provement of the regional function was limited o segments without LE. The shareof LE did not decrease under ERT.Conclusion: Patients without LE showed a decreased LV mass and an improve-ment of the regional function, as a sign for the benefit of ERT. LE cannot bereduced by 12 months of ERT as well as the regional function in segments withLE did not recover.

B-595 11:15

Delayed gadolinium enhanced cardiac magnetic resonance in Anderson-Fabry diseaseA. Esposito1, F. De Cobelli1, M. Pieroni2, E. Belloni1, C. Chimenti2,F. Gentinetta1, A. Frustaci2, A. Maseri1, A. Del Maschio1; 1Milan/IT, 2Rome/IT([email protected])

Purpose: Anderson-Fabry disease (AFD) is rare X-linked disorder of glycosphin-golipid metabolism. Cardiac involvement represents one of the severest aspectsof this disease. Our aim was to investigate the occurrence and distribution of late-enhancement (LE) in the myocardium of these patients and to correlate LE pres-ence with histological features.Methods and Materials: 10 AFD patients underwent cardiac magnetic resonance(CMR) on a 1.5-T system (Philips). CMR included the acquisition of delayed im-ages 10-15 minutes after gadolinium injection, with 3D-T1w-TFE sequence, ad-justing inversion time to null normal myocardium. LE was expressed as percentageof left ventricle (LV) myocardium. All patients underwent LV endomyocardial bi-opsy with automated morphometric quantification (Nikon) of fibrosis on sectionstained with Masson's trichrome.Results: 7 patients (70%) showed the presence of LE areas ranging from 1 to17.5% (mean 6.5 ± 5.8%) of LV myocardium. The fibrosis measured on histolog-ical samples ranged between 4 and 18% (mean 9.1 ± 4.8%). LE percentage atCMR was strongly related to the percentage of fibrosis calculated at histology(R=0.93; P < 0.0005). LE areas were located in the medium-basal segments oflateral LV wall in all patients. Only two patients with severe LV hypertrophy showedadditional LE areas in apical segments.Conclusion: LE located in the medium-basal lateral wall of LV might be consid-ered as a specific sign of cardiac involvement in AFD patients. The strong rela-tion between LE extension and histological fibrosis quantification suggests thatLE percentage could represent an important non-invasive marker of the severityof cardiac involvement in Fabry disease.

B-596 11:24

Therapy follow-up via cardiac MR imaging in patients with cardiacsarcoidosisM. Heller1, M. Schiemann1, N.D. Abolmaali2, T.J. Vogl1; 1Frankfurt/DE,2Dresden/DE ([email protected])

Purpose: To assess the follow-up in patients with cardiac sarcoidosis under ther-apy with corticoids using cardiac MR imaging.Methods and Materials: 18 patients with cardiac sarcoidosis underwent an ini-tial and 3, 6 and 12 month follow-up MR imaging under therapy with corticoids.Short axis views covering the entire heart and 4-chamber views were done withT2-w images using a breath hold turbo inversion recovery sequence. Additionallydelayed enhancement imaging was done with a 3d turbo flash sequence usingan optimized TI-time of 210 ms in the short and long axis. Dynamic left ventricu-lar analysis was done with contiguous short axis views covering both ventriclesusing a True-FISP cine sequence. Additionally 3- and 4-chamber views wereperformed. MR imaging findings were categorized into stable (± 50 cardiac infil-tration volume), worsening (> 50 % more infiltration volume) or improving situa-tions (> 50 % less cardiac infiltration volume).Results: 4 patients showed an improving situation with 3 patients presented acomplete clearing of patchy granulomatous infiltrations with no areas of late en-hancement while in 9 patients the MR imaging showed a stable situation withpersistence of cardiac infiltrations in the same extension as in the initial imaging.Worsening situations of cardiac involvement was seen in 5 patients who all dieddue to therapy refractory ventricular tachyarrhythmias within 2 years.Conclusion: Cardiac MR imaging offers a rapid and non-invasive screeningmethod with high spatial resolution for the detection of myocardial involvementand fibrosis occurring in cardiac sarcoidosis and is a useful method in follow-upunder treatment.

Scientific Sessions

276 C D E FBA G

B-597 11:33

MR characterization of cardiac abnormalities in asymptomatic HIV+individuals with increased BNP levelsK. Nassenstein, F. Breuckmann, T. Neumann, J. Kondratieva, R. Erbel,J. Barkhausen; Essen/DE ([email protected])

Purpose: Cardiac failure in HIV+ patients may be caused by AIDS associatedcardiomyopathy or treatment related. Our study aimed to characterize cardiacabnormalities in HIV+ patients with increased serum brain-derived natiuretic petide(BNP).Methods and Materials: 15 of 131 screened HIV+ patients meet the inclusioncriteria of an elevated BNP level (> 100pg/ml). 3 patients refused to participate,thus 12 patients (mean-age 57 ± 19 years) were enrolled into the MR study. Cinesteady state free precession sequences (TR 3.0 ms, TE 1.5 ms, FA 60°, slicethickness 8 mm) of long and contiguous short axis slices were performed, fol-lowed by T2-weighted images to assess myocardial edema. After Gd-DTPA(0.2 mmol/kg), short axes were scanned using a segmented IR-TurboFLASHsequence (TR 8 ms, TE 4 ms, TI 200-260 ms) to detect "late enhancement" (LE).Results: Adequate image quality could be achieved in all but one patient. MRimaging analysis revealed hypertrophic cardiomyopathy (n = 3), residual myo-carditis (n = 2), myocardial infarction (n = 2), dilated cardiomyopathy (n = 1), rightventricular dilatation (n = 1) and focal dyskinesia (n = 1). Delayed enhancementwith spotted, streaky, subepicardial or transmural distribution could be found in 3of our 12 patients. A reduction of the ejection fraction could be detected in 4patients resulting in a mean shortening fraction of 45% and hypo- to akineticareas in up to 3 patients. Only two patients showed no pathologic abnormalities.Conclusion: Although no specific MR pattern was found in our study cohort, MRallowed characterizing the underlying cardiac pathologies in 82% of HIV patientswith elevated BNP levels.

B-598 11:42

Constrictive pericarditis: Assessment of diastolic function by velocityencoded flow measurementsK.U. Bauner, A. Huber, M. Schmoeckel, M.F. Reiser, S.O. Schönberg;Munich/DE ([email protected])

Purpose: To analyze, whether assessment of transmitral and transtricuspid flowcurves using velocity encoded flow measurements allows to identify pathologicaldiastolic function in patients with constrictive pericarditis.Methods and Materials: 18 patients with constrictive pericarditis (CP) were ex-amined with MR imaging. 10 healthy subjects served as controls. Cine TrueFISP,HASTE and spin-echo sequences were performed to evaluate cardiac morphol-ogy and function. Pericardial thickness, septal wall motion and systolic and di-astolic right ventricular diameters were measured by two experienced radiologists.Transmitral and transtricuspid velocity-time curves were examined with velocityencoded flow measurements with retrospective ECG-gating (TR/TE/Bandwidth/Matrix/VENC: 41 msec/3.2/391/256 x 100/100-150 cm/sec). The reconstructedimages had a temporal resolution of 20 msec. The amplitude of E and A waveand the ratio of the amplitudes of the E and A-wave were calculated. Student's t-test was used to differentiate patients with CP from healthy volunteers.Results: The E/A-ratio of transmitral flow velocity was significantly smaller in CP(1.39 ± 0.45%) than in healthy subjects (2.8 ± 1.5%) p < 0.01, as was the E/Aratio of transtricuspid flow velocity in CP (1.06 ± 0.24%) in comparison to normalvolunteers (2.7 ± 1.7%) p < 0.01.Conclusion: In addition to the assessment of morphology and cine function, MRimaging is able to identify pathological diastolic function by flow measurementsof the atrioventricular valves. Our results indicate, that MR imaging flow meas-urements may support to distinguish systolic from diastolic causes for heart fail-ure in patients with CP.

B-599 11:51

MR imaging in the differential diagnosis of left ventricular non-compaction:Our experienceR. Dore, A. Valentini, D.M. Savulescu, V. Vespro, C. Raineri, L. Monti,A. Raisaro; Pavia/IT ([email protected])

Purpose: For the diagnosis of Left ventricular non-compaction (LVNC) four maincriteria have been proposed, both in echocardiography and more recently in car-diac MR imaging. In this study we specifically evaluated the "non-compacted tocompacted myocardium" ratio (NC/C) in distinguishing LVNC from other cardio-myopathies.Methods and Materials: Nine patients with suspected LVNC were enrolled. Weperformed the following MR imaging sequences (Siemens Symphony 1.5 T): 1)

TSE T1 black blood and cine TrueFISP sequences for morphological and func-tional analysis in both ventricles; 2) a perfusion study with first pass sequenceduring gadolinium injection (0.05 mmol/kg) on short axis at the basis, mediumand apical level of the left ventricle; 3) delayed enhancement at10 minutes wasevaluated on short axis, two chambers long axis and four chambers long axis todetect myocardial fibrosis.Results: In all patients we identified the typical pattern of thin-layered epicar-dium and hypertrabeculation of the endocardium in the antero-lateral, medial-distal wall and in the apex. The NC/C ratio measured during diastolic phase was> 2.3 in all the cases, but in 3 patients this value was not confirmed from themeasurement obtained in the systolic phase. Systolic function of left ventriclewas reduced in 5 patients. First pass study revealed delayed perfusion both inthe non compacted area and the compacted one. Only in one case myocardiacfibrosis was detected.Conclusion: In our experience the NC/C ratio may be not a sufficient index in thedifferential diagnosis of LVNC. Other MR imaging criteria should be prospectivelyevaluated.

10:30 - 12:00 Room G

Head and Neck

SS 1408Cervicofacial MRModerators:A.S. Dimitriadis; Thessaloniki/GRD. Farina; Brescia/IT

B-600 10:30

Proton MR spectroscopy of the extracranial head and neck: A pilot in vivostudyS. Bisdas, M. Baghi, F. Huebner, C. Mueller, A. Smolarcz, R. Sader,W. Gstoettner, T.J. Vogl; Frankfurt/DE ([email protected])

Purpose: To determine the feasibility of performing in vivo proton MR Spectros-copy (MRS) of the extracranial head and neck tissue and differentiating malig-nant and benign extracranial head and neck neoplasms from normal tissue.Methods and Materials: Benign and malignant neoplasms of the extracranialhead and neck as well as metastatic lymph nodes of 38 patients were examinedusing localized single voxel MRS (TR 1500, TE 135) at 1.5 T. New techniqueswith simultaneous correction of motion artefacts during the acquisition, three di-mensional saturation pulses, respiratory triggering, and smaller volume of inter-est (VOI) size were applied. Ratios of peak areas under the choline (Cho) andcreatine (Cr) resonances were estimated in all cases and compared with sam-ples of normal tissue.Results: Interpretable spectra were acquired in all patients. The smallest VOIwas 0.40 cm3. The Cho/Cr ratios in all neoplasms were significantly elevated inrelation to the normal muscle structures. The average Cho/Cr ratio for the malig-nant tumors was 3.02. The benign neoplasms showed an elevated Cho/Cr ratioin relation to the malignant ones. The Cho/Cr ratio in the metastatic lymph nodeswas 5.42, whereas for benign lymphoid hyperplasia was 2.74. MRS measure-ments could differentiate recurrent disease from post-therapeutic tissue chang-es.Conclusion: In vivo proton MRS is feasible for the evaluation of benign and ma-lignant tumors arising in the extracranial head and neck. MRS is also a potentialtool in differentiation of lymph nodes' enlargement and therapeutic monitoring ofthe neoplasms.

B-601 10:39

Diffusion weighted magnetic resonance imaging of head and necksquamous cell carcinomasK.M. Friedrich, W. Matzek, S. Gentzsch, I. Sulzbacher, C. Czerny,A.M. Herneth; Vienna/AT ([email protected])

Purpose: Diffusion-weighted imaging (DWI) is a powerful tool in oncologic radi-ology, with the potential to evaluate the microstructure of a biological tissue. Micro-structural differences between vital tumor tissue and tumor free soft tissue canbe appraised by quantitative analysis of DWI, which can be achieved by calculat-ing the apparent diffusion coefficient. The objective of this study was to evaluatewhether DWI is a reliable technique to quantify microstructural differences be-tween head and neck squamous cell carcinomas and tumor free soft tissue.Methods and Materials: We prospectively investigated 24 patients (10 women

Mo

nd

aM

on

da

Mo

nd

aM

on

da

Mo

nd

ayy yyy

Scientific Sessions

B 277C D E FA G

and 14 men), who were referred for MR imaging of known or suspected head andneck cancer. In these 24 patients 20 squamous cell cancers were found. For DWIwe used diffusion weighted, navigated echo-planar imaging (1.0-Tesla MR; max-imum b-value: 800). For objective assessment of image quality, the signal-to-noise ratio was calculated. Microstructural differences between vital tumor tissueand tumor free soft tissue were quantified by calculating the apparent diffusioncoefficients on a pixel by pixel method.Results: Diffusion-weighted echo-planar imaging provided good image quality inall patients (mean signal-to-noise ratio 18.4). The mean apparent diffusion coef-ficient of squamous cell carcinoma (0.64 ± 0.28 x 10-³ mm²/sec), was significant-ly (P < 0.0001) lower than that of the tumor free soft tissue, (2.51 ± 0.82 x 10-³ mm²/sec).Conclusion: DWI is a reliable diagnostic tool to quantify the microstructural dif-ferences between vital tumor tissue and tumor free soft tissue in patients withhead and neck squamous cell carcinoma.

B-602 10:48

Diffusion-weighted echo-planar MR imaging: A valuable tool fordifferentiating primary parotid gland tumors?C.R. Habermann, P. Gossrau, J. Graessner, C. Arndt, M.C. Cramer,F. Reitmeier, M. Jaehne, G. Adam; Hamburg/DE([email protected])

Purpose: Investigate the feasibility of using diffusion-weighted (DW) echo-pla-nar imaging (EPI) for differentiating primary parotid gland tumors.Methods and Materials: Fifty consecutive patients with suspected primary tu-mor of the parotid gland were examined with DW EPI sequences. The b factorsused were 0, 500, and 1,000 sec/mm2. ADC maps were evaluated with a manu-ally placed region of interest (ROI) containing the entire tumor. Additionally, acircular ROI containing 100 - 200 pixels was placed in the cerebrospinal fluid(CSF) next to the spinal cord in every patient. For comparison of the results, thepaired Student's t-test was used, based on the median ADC values for eachpatient, and a p value of less than 0.05 was determined to indicate statisticalsignificance.Results: In 45 patients a primary neoplasm of parotid gland could be histologi-cally verified. Seven types of parotid gland tumor were histologically discriminat-ed. Pleomorphic adenomas (2.14 x 10-3 mm2/sec ± 0.11 x 10-3 mm2/sec) (mean± standard deviation) and Warthin tumors (0.85 x 10-3 mm2/sec ± 0.1 x 10-3 mm2/sec) showed significantly different ADC values in comparison to all other evaluat-ed tumors (p < 0.001), and also between each other (p < 0.001). ADC valuespresented by lipomas were statistically significant compared to all other entities(p < 0.001 to 0.015). Among primary malignant parotid gland tumors no statisti-cally significant ADC values could be observed (p = 0.18 to 1).Conclusion: Diffusion-weighted echo-planar MR imaging seems to be a valua-ble tool for differentiating benign from malignant primary parotid gland tumors.

B-603 10:57

Functional imaging of parotid glands using diffusion-weighted echo-planarMR imaging: Comparison of 1.5 T and 3 TM.C. Cramer1, P. Gossrau1, H. Kooijman2, J. Graessner1, M.G. Kaul1,F. Reitmeier1, M. Jaehne1, G. Adam1, C.R. Habermann1; 1Hamburg/DE,2Best/NL ([email protected])

Purpose: To investigate the feasibility of using diffusion-weighted (DW) echo-planar imaging (EPI) for measuring different functional conditions of the parotidgland and to compare different field strengths.Methods and Materials: Parotid glands of 27 healthy volunteers were examinedwith a DW EPI sequence at 1.5 T and 3 T before and after oral stimulation withcommercially available lemon-juice. The b-factors used were 0, 500, and 1,000sec/mm2. Apparent diffusion coefficient (ADC) maps were evaluated with a man-ually placed region of interest (ROI) including the entire parotid gland. For com-parison of the results, Student's t-test was used, based on the mean of thevolunteer median values. To compare both field strengths, the Pearson's correla-tion coefficient (r) was calculated.Results: DW EPI MR imaging successfully visualized the parotid glands of allvolunteers. Using 1.5 T, the mean ADC prior to stimulation was 1.12 x 10-3 mm2/sec ± 0.08 x 10-3 mm2/sec, whereas after stimulation with lemon juice, the ADCincreased to 1.18 x 10-3 mm2/sec ± 0.09 x 10-3 mm2/sec. For 3 T, the ADC prior tostimulation was 1.14 x 10-3 mm2/sec ± 0.04 x 10-3 mm2/sec, with an increase to1.17 x 10-3 mm2/sec ± 0.05 x 10-3 mm2/sec after stimulation. For both field strengths,the increase in ADC after stimulation was significant (P < 0.001). High correla-tions between both field strengths were found prior and post stimulation (r = 0.955,and 0.936 respectively).

Conclusion: Diffusion-weighted echo-planar MR imaging allows quantificationof functional changes in the parotid glands with high correlation between 1.5 Tand 3 T.

B-604 11:06

MR-elastography of the parotid gland: Feasibility and first experiencesC.R. Habermann1, R. Sinkus2, A. Albrecht1, M.C. Cramer1, F. Weiss1, G. Adam1;1Hamburg/DE, 2Paris/FR ([email protected])

Purpose: Investigate the feasibility of MR-elastography of the parotid gland.Methods and Materials: Twenty-seven healthy volunteers (17 women, ten men;mean age, 32 yrs; age range 27 to 38) were examined in a supine position with a3 T whole body scanner (3 T Intera; Philips Medical Systems, Best, The Nether-lands) using a 6-channel SENSE head coil. The mechanical transducer was placedon the left parotid gland pulsing in a left-right direction at a frequency of 100 Hz.MR-elastography data were acquired using a motion sensitized spin echo se-quence with EPI readout. Total scan time for all three spatial displacement direc-tions was about 9 minutes. Ten volunteers were additionally evaluated poststimulation of the parotid gland with commercially available lemon juice.Results: Very good penetration of the mechanical waves throughout the entireparotid gland was observed in 25 of 27 volunteers. Mode-conversion of the longi-tudinal wave at interfaces leads to shear waves generated everywhere inside theorgan. The amplitude of the penetrated wave close to the transducer was 80 to100 µm. The mean shear modulus for the whole parotid gland of all 25 volunteerswas 1.37 ± 0.35 kPa. Shear viscosity was calculated with 0.45 ± 0.14 Pa.s. Oralstimulation leads to an increase of shear modulus and shear viscosity.Conclusion: This work demonstrates the technical feasibility of in-vivo parotidgland MR-elastography. Small standard deviation of the shear modulus of healthyparotid glands indicates that a differentiation of tumorous lesions within the pa-rotid gland should be possible.

B-605 11:15

The efficacy of MR imaging with USPIO in N-staging in oral cavity cancerF. Mrakic-Sposta, G. Villa, F. Chiesa, G. D'Andrea, F. Maffini, M. Bellomi;Milan/IT ([email protected])

Purpose: To compare the diagnostic accuracy of MR imaging with USPIO (ul-trasmall superparamagnetic iron oxide particles) and plain MR imaging in necklymphnode staging of patients with oral cavity cancer using histology as a goldstandard.Methods and Materials: 24 patients with spinocellular carcinoma of the oralcavity have been examined by plain and USPIO MR imaging with T1-weightedspin echo (SE), T2-weighted fast spin echo (FSE) and T2 weighted gradient echo(GE) sequences in axial orientation. Nodes were defined metastatic 1)when 10 mmat plain MR imaging, and 2) when there is no decrease of signal intensity atUSPIO-MR imaging. 23 patients underwent surgery with monolateral (11) or bi-lateral (12) lymphnode dissection within 15 days from MR imaging.Results: Correlation between histology and imaging features was evaluated forall 7 lymphnode stations. Pathology demonstrated 37 metastases in the 1137lymphnodes examined. USPIO MR imaging diagnosed 31 metastases (sensitiv-ity 83%, specificity 1.7%), plain MR imaging diagnosed 30 metastases (sensitiv-ity 81%, specificity 5%). Planning surgery on the basis of USPIO MR imagingstaging would have spared 7 controlateral nodal dissections.Two patients suffered mild allergic reaction to injection of the contrast agent.Conclusion: This study shows the MR imaging with USPIO has a higher accura-cy than plain MR imaging in diagnosis of neck nodal metastases, thus offeringbetter therapeutic planning.

B-606 11:24

Diagnosis of skull base invasion of head and neck tumor by fusion imagingof MR imaging and CTS. Kandatsu1, R. Kishimoto1, H. Ojiri2, S. Komatsu1, J. Mizoe1, H. Tsujii1,K. Fukuda2; 1Chiba-ken Chiba-shi/JP, 2Minato-ku Tokyo/JP([email protected])

Purpose: To evaluate the efficacy of fusion imaging of MR imaging and CT indiagnosis of skull base invasion of head and neck tumors.Methods and Materials: From April 2004 to March 2005, fifty patients, including24 men and 26 women, with head and neck tumors which were suspected toinvade the skull base, underwent MR imaging and 16-slice MDCT at the sametime. Fusion images of color mapped MR images and 1 mm thick CT data wereobtained using a 3D workstation. Two independent experienced radiologists com-pared the fusion images with CT and MR imaging, respectively, regarding

Scientific Sessions

278 C D E FBA G

perineural invasion, intracranial invasion, intra-orbital invasion, and paranasalspace invasion.Results: Of the 50 patients, bone destruction of the skull was found in over 90%of the cases. Invasion of the tumors into the skull bones was well visualized bythe fusion images. Perineural invasion was also well visualized. Certainty for theseissues of evaluation using fusion imaging was superior to single CT and MRimages, respectively.Conclusion: Fusion imaging of MR imaging and CT is effective for diagnosis ofskull base invasion of head and neck tumors.

B-608 11:33

Perineural extension along the infraorbital nerve evaluated on coronal MRimages: Its frequency and MR findingsT. Otani, N. Tomura, S. Takahashi, I. Sakuma, J. Watarai; Akita/JP([email protected])

Purpose: To evaluate MR findings of perineural extension along the infraorbitalnerve and study its frequency in patients with head and neck tumors.Methods and Materials: MR studies (n = 362) including coronal sections wereperformed in 187 patients with head and neck tumors. Coronal MR images werescanned from Meckel's cave to the infraorbital foramen. MR imaging included T2-weighted images and post-contrast fat-saturated T1-weighted images. Perineuralextension along the infraorbital nerve was defined as swelling or thickening of theinfraorbital nerve with contrast enhancement. The frequency of perineural exten-sion along the infraorbital nerve was retrospectively studied. Its coexistence withtumors in the pterygopalatine fossa and superficial musculoaponeurotic systemwas also evaluated.Results: Perineural extension along the infraorbital nerve was observed in 16patients (8.6%). The primary sites of the tumors were the maxilla (n = 8), orbit(n = 2), hard palate (n = 2), nasal cavity (n = 2), and upper gingival (n = 2). Thepathological diagnoses were squamous cell carcinoma (n = 10), adenoid cysticcarcinoma (n = 3), malignant melanoma (n = 2), and malignant lymphoma (n = 1).All cases except one, an orbital adenoid cystic carcinoma, showed coexistencewith tumors in the pterygopalatine fossa and superficial musculoaponeurotic sys-tem.Conclusion: The present study demonstrated that perineural extension alongthe infraorbital nerve was not rare in patients with head and neck tumors. Perineuralextension usually coexists with tumors in the pterygopalatine fossa and superfi-cial musculoaponeurotic system. Coronal MR images are excellent for demon-strating perineural extension along the infraorbital nerve.

B-609 11:42

Inspiratory upper airway obstruction during partial neuromuscularblockade: Evaluation with realtime MR imagingF.M. Vogt1, M. Eikermann2, M.O. Zenge1, V. Dastgerdi1, A. deGreiff1,J. Barkhausen1; 1Essen/DE, 2Boston, MA/US ([email protected])

Purpose: To assess influence of partial neuromuscular blockade on the upperairway system using realtime TrueFISP MR imaging and volumetry of the upperairways.Methods and Materials: 10 healthy volunteers were studied before, during (train-of-four [TOF]-ratio: 0.5 and 0.8) and after (TOF-ratio: 1.0 and 15 minutes later)rocuronium induced partial neuromuscular transmission failure which did not evokedyspnea or oxygen desaturation. Imaging was performed on a 1.5 T scannerusing a head/neck surface-coil for signal reception. Functional exploration of theupper airway system was performed using sagittal real-time TrueFISP sequenc-es (temporal resolution 250 ms). Datasets were acquired at rest and forced inspi-ration before, during and twice after relaxation. Minimal distance from back oftongue to posterior wall of pharyngeal space was evaluated. Additionally, se-quential T1w images were collected in transverse orientation from the top of thehard palate to the vocal cords for the volumetric airway measurements.Results: During partial neuromuscular blockade, upper airways partially collapsedat end-inspiration. Measurements of minimal distances during forced inspirationshowed a mean reduction from 20 mm before relaxation to 14 mm during maxi-mal relaxation. Comparing pre-relaxation datasets to TOF-ratio 0.5 datasets, meaninspiratory volumes correspondingly decreased to 78 ± 11 % of baseline (p < 0.01).Complete recovery was not found until 10 minutes after termination of relaxation.Conclusion: Real time MR imaging demonstrated that even a minor degree ofpartial neuromuscular blockade is associated with upper airway obstruction. Up-per airway obstruction should be considered in patients at risk for neuromusculartransmission failure presenting with dysphagia.

10:30 - 12:00 Room H


SS 1409Vascular and non-vascular stentingModerators:J.W. Chung; Seoul/KRS. Müller-Hülsbeck; Kiel/DE

B-610 10:30

Stenting of mesenteric arteries in monorail-technique in patients withchronic mesenteric ischemiaP.J. Schaefer, F.K.W. Schaefer, T. Jahnke, N. Charalambous, M. Heller,S. Mueller-Huelsbeck; Kiel/DE ([email protected])

Purpose: To analyze immediate, midterm success and durability of stenting ofmesenteric arteries with a monorail-technique in patients with chronic mesenter-ic ischemia.Methods and Materials: The prospective case series was conducted accordingto the Declaration of Helsinki and with informed consent of patients. 19 patients(11 male, 8 female; mean-age 62.9 ± 10.4 years, range 36-82) with 23 sympto-matic mesenteric artery stenoses were treated with stent implantation with amonorail-technique in a radiological intervention center over 4.5 years. Clinicalexaminations and duplex sonography were used to evaluate the stents´ patencyand clinical success. Kaplan-Meier graphs were calculated to analyze the paten-cy and freedom of symptoms rate.Results: Initial technical success rate was 22/23 (96%). Two patients died ofcardiac failure in hospital within 30 days of intervention; deaths were not relatedto intervention. No peri-interventional complications occurred. Mean follow-up was17 months (range 1-58). Primary patency and primary clinical success rates were82% and 78%, respectively. According to Kaplan-Meier tables patency rate was96/87/76/61% at 0/1/15/24 months, and freedom of symptoms rate was 95/90/72/54% at 0/1/24/30 months.Conclusion: Stent implantation with a monorail-technique in mesenteric arteriesis an efficient and very safe treatment of symptomatic stenoses in patients withchronic mesenteric ischemia.

B-611 10:39

Effects of global renal artery stenting on chronic renal failureP. Stradiotti, A. Nicolini, R. Vercelli, A. Zagarella, C. Bazzi, R. Rivolta; Milan/IT([email protected])

Purpose: To examine prospectively the clinical effects of renal artery stenting(RAS) in patients with global renal artery stenosis and chronic renal insufficien-cy.Methods and Materials: A successful RAS was performed in 30 patients withchronic renal impairment (serum creatinine > 1.5 mg/dl) and global atheroscle-rotic obstruction (bilateral renal artery stenosis or unilateral stenosis in a pres-ence of a solitary or single functional kidney). To assess a significant change ofRF, we managed to compare the slopes of the regression lines derived from thereciprocal of creatinine versus time, plotted before and after stent deployment.We also evaluated changes in renal length and interlobar resistive index. Medianfollow-up was 24 months (range 9-52).Results: Before stenting all the patients showed a negative slope, indicating pro-gressive renal failure. After stenting the mean of slopes became positive (meanpre -0.0249 ± 0.0527 vs post 0.0048 ± 0.0196, p < 0.008). Ten patients (33.3 %)showed a significant positive slope, indicating improvement of RF; in 13 patients(43.3 %) the slopes were not significantly different from 0 and were associatedwith a stable RF. Seven patients (23.3 %) presented negative slopes and showeda continuous decrease of RF. A low previous serum creatinine are associated topositive output (p = 0.05) while patients with negative slopes showed a signifi-cant decrease in kidney size (p < 0.02).Conclusion: In the majority of patients with chronic renal failure and global ob-structive renovascular disease, RAS appears to be associated with an improve-ment or stabilization of impaired RF.

Mo

nd

aM

on

da

Mo

nd

aM

on

da

Mo

nd

ayy yyy

Scientific Sessions

B 279C D E FA G

B-612 10:48

Stent placement in pediatric patients with hepatic artery stenosis orthrombosis after liver transplantationF. Boyvat, C. Aytekin, H. Karakayali, U. Ozyer, R. Emiroglu, M. Haberal;Ankara/TR ([email protected])

Purpose: Hepatic artery stenosis (HAS) or thrombosis after orthotopic liver trans-plantation remains a significant cause of graft loss. In the past, repeat transplan-tation was considered the first therapeutic choice. But there is no data related tostent placement in the thrombosed hepatic artery in the early postoperative peri-od.Methods and Materials: Between March 2000 and March 2005, percutaneousendoluminal stent placement was performed in 6 pediatric liver transplant pa-tients (2 girls, 4 boys; mean 11.2 years, range: 5-16). Initially the HAT or HAS wasdiagnosed in all cases with Doppler ultrasound, then confirmed angiographically.4 hepatic artery stenoses and 2 hepatic artery occlusions were found. Wall-stentsand coronary stents were used during the procedure. Stents were placed; 1st day(3 patients), 17th day (1 patient), 18th day (1 patient) and 180th day (1 patient) ofthe postoperative period.Results: Technical success was 100 %. Two ruptures were seen during PTA in-tervention of the hepatic artery in patients on the 1st and 17th postoperative dayand graft covered coronary stents were placed. In one of the patients, dissectionof the origin of the common hepatic artery developed due to the guiding sheathand a second stent was placed. One of the stents was occluded three monthsafter intervention without any clinical findings. The follow-up was changed be-tween 4 months to 2.5 years.Conclusion: Early and late postoperative stent placement to the graft hepaticartery is technically feasible. This technique needs more development but showsgreat promise for the future.

B-613 10:57

PTFE covered stents in malignant biliary strictures: Are they useful?M. Corona, F. Fanelli, M. Bezzi, A. Bruni, V. Dominelli, A. Pucci, G. Luppi,R. Passariello; Roma/IT ([email protected])

Purpose: To evaluate the efficacy of e-PTFE covered stent (VIABIL - W.L. Gore)in the treatment of malignant biliary strictures.Methods and Materials: From 2001, 65 patients (mean age 69.4y) with malig-nant common bile strictures were treated with an e-PTFE covered stent-graft(Viabil - W.l. Gore). The Viabil is a self-expandable stent-graft with an inner ePT-FE/FEP lining and an outer supporting structure of nitinol wire. Biliary obstructionwas secondary to pancreatic carcinoma (n = 38), cholangiocarcinoma (n = 10),gallbladder cancer (n = 6) and metastatic lymphadenopathy (n = 11). The stric-ture was located in the upper third of the common bile duct in eight patients, inthe middle portion in 43 and in the lower third in 14. Sixty-nine stent-grafts wereimplanted: 49 with side holes and 20 without holes. In 62/65 patients the distalend of the device was deployed in the duodenum. Clinical evaluation and assess-ment of serum bilirubin and liver enzyme levels were done before stent place-ment and at 1, 3, 6, 9 and 12 months.Results: Stent-graft placement was successful in all cases. The 30-day mortalityrate was 11%. The survival rates were 40% at 6 months and 20.2% at 12 months.The stent patency rates were 91%, 78% and 78% at 3, 6 and 12 months respec-tively. After a mean follow-up of 37.2 months we reported acute cholecystitis in 6cases (9.2%). Complications other than stent occlusion occurred in 12 patients(18.4%).Conclusion: e-PTFE covered stent-graft seems to be an effective tool in casesof malignant strictures of the common bile duct.

B-614 11:06 !ePTFE/FEP-covered metallic stents for palliation of malignant biliarydisease. Clinical results in 35 patientsM. Krokidis1, A. Hatzidakis1, E. Charonitakis1, C. Papadopoulos1, N. Filippakis1,G. Chlouverakis2, N. Gourtsoyiannis1; 1Heraklion/GR, 2Rethymnon/GR([email protected])

Purpose: To study applications and clinical effectiveness of ePTFE/FEP coveredmetallic stents for palliation of malignant biliary disease.Methods and Materials: Thirty-five patients (37-88 years old, mean 67.4), 22men and 13 women, comprised and were followed-up in this study. We used the10 mm wide Viabil© stent (WL GORE, USA) of 6 or 8 cm length, with/withoutproximal side holes. Biliary obstruction was found in the lower CBD (in 17), upperCBD including cystic duct (in 9), lower CHD including cystic duct (in 4), CHD (in2), upper CHD and left main bile duct (in 2), and left main biliary duct (in 1).

Results: One stent was placed in each patient, while in 2 patients a secondViabil was inserted after occlusion of the first one. In another 2 cases uncoveredstents were placed after Viabil occlusion. All stents were successfully placedexcept one, which migrated distally (technical success rate 97%). Thirty daymortality was 24% and not procedure related. Eleven out of 35 patients are stillalive (mean survival 105 days). Mean primary and secondary patency was 82%and 99%, respectively. No serious complications were noted despite a relativelylarge delivering catheter diameter. No pancreatitis occurred in cases when thepapilla was fully covered.Conclusion: Patency, survival and complication rates of the Viabil stents arecomparable with those of the known uncovered stents. Larger patient numbersand longer follow-up, as well as correlation studies with patient groups with un-covered metallic stents are needed for more conclusive results.

B-615 11:15

The development of a retrievable stent for the recanalization of thrombus-occluded vesselsJ. Koizumi1, A. Kitamura2, K. Tokuyasu3, R. Hatomi3, Y. Imai1; 1Isehara-city/JP,2Matto-city/JP, 3Sagamihara-city/JP ([email protected])

Purpose: For massive thrombus-occluded vessels such as pulmonary embolismwith cor pulmonale or deep venous thrombosis, direct stenting alongside thepulmonary emboli may be effective for immediate decompression. However, thelack of long-term durability and possible thrombogenic effect are theoretical dis-advantages. Thus we developed a retrievable stent to obtain an immediate buttemporary recanalization of the thrombus-occluded vessels by compressing theclot.Methods and Materials: The conical stent was developed using one string ofnitinol wire including a hook at one end for loop snare retrieval. The 15 x 65 mmstents composed of three different diameters (0.20, 0.25, 0.30 mm) of wire fila-ments were deployed in clot-filled test tubes (inner diameter; 13 mm). For variousage clot-filled test tubes from day 0 to 7, the minimal recanalization ratios (in-stent diameter/total lumen diameter) were measured using CT. After measure-ment the stents were retrieved and the existence of in-stent thrombus wasobserved.Results: All stents' recanalization ratios decreased with clot age and the strong-er two stents (0.25, 0.30 mm) had significantly higher ratios (70-80% vs. 50-60%)than the weakest one (0.20 mm). In-stent thrombus was not observed.Conclusion: A conical shaped retrievable stent was feasible for immediate reca-nalization by compressing thrombus in vitro.

B-616 11:24

Endobronchial dilatation and stent placement for bronchial stenosis inpatients after lung transplantationN. Abi-Jaoudeh1, G. Soulez2, V.L. Oliva2, M.-F. Giroux2, E. Therasse2,R. Jean-Francois2, C. Poirier2; 1Quebec, PQ/CA, 2Montreal, PQ/CA([email protected])

Purpose: To evaluate the outcome of balloon dilation and/or stent placement inpatients with bronchial strictures after lung transplantation.Methods and Materials: Forty-one lung recipients underwent balloon dilationand/or stent placement for bronchial strictures between January 1997 and Janu-ary 2005. Clinical assessment, pulmonary function testing and bronchoscopicevaluation were obtained before and after each procedure. Clinical and broncho-scopic follow-up ranged from 1 to 87 months (mean 17 months, median 11months).Results: Of the forty-one patients treated, twenty-four were alive on the lastfollow-up (58%). All patients had primary balloon dilatation; endobronchial stentplacement was done if balloon dilation had failed to re-expand the bronchus dur-ing the procedure or at a later time. The average number of procedures per pa-tient was 2.8 (median 2). During a procedure an average of 1.75 lesions weretreated (median 1). Twenty-three patients (56%) received a stent because of fail-ure of balloon dilatation. Among those, the mean number of stents per patientwas 1 (median 1). Primary patency (after the first procedure) was noted in thir-teen patients (32%). Secondary patency (after the last procedure) was noted inthirty-eight patients (93%). There was significant improvement in clinical and meanpulmonary function testing after each procedure, with an amelioration of 8% ofthe VEMS (P < 0.00003).Conclusion: Balloon dilation and endobronchial stent placement improve thesymptoms and pulmonary function tests in patients with bronchial stenosis fol-lowing lung transplantation. However, it should be noted that repeat interventionsmight be necessary to maintain patency.

Scientific Sessions

280 C D E FBA G

B-617 11:33 !Follow-up of temporary partially-covered metal stent insertion in achalasiaY.S. Cheng, M.H. Li, K.Z. Shang; Shanghai/CN ([email protected])

Purpose: To study the long-term therapeutic efficacy of temporary partially-cov-ered metal stent insertion by follow-up in achalasia.Methods and Materials: We inserted 103 temporary partially-covered metal stentsin 103 patients with achalasia, which were removed 3-5 days later. The place-ment and removal of stents were successful in 100 % of cases. Immediately afterplacing the stent, we performed an esophageal contrast examination, which con-firmed good stricture patency. The mean follow-up time was 12-120 months.Results: Twelve patients of 103 exhibited dysphagia relapse during a 12-monthfollow-up, 13 patients out of 75 exhibited dysphagia relapse during a 36-monthfollow-up and 8 patients out of 38 exhibited dysphagia relapse during a 60-monthfollow-up.Conclusion: Temporary partially-covered metal stent dilation is the best inter-ventional procedure for achalasia in terms of follow-up.

B-618 11:42

Stenting of the alimentary tractS. Profili, G.B. Meloni, C. Urigo, A. Manca, G.C. Canalis; Sassari/IT([email protected])

Purpose: To evaluate the clinical effectiveness of recanalization by insertion ofself-expandable metal stents in the treatment of benign and inoperable neoplas-tic symptomatic strictures of the alimentary tract.Methods and Materials: From 1992 to 2005 we treated 247 patients (33-85 yearsold), 235 with neoplastic and 12 with benign strictures by insertion of 284 stents.Different kind of prostheses, depending on commercial and technological devel-opments, were employed. The procedures were performed in 147 cases foroesophageal strictures or anastomotic leakage (3 pts), for gastroduodenal andjejunal strictures in 45 cases and for large bowel stenosis in 55 patients. Proce-dures were all performed under fluoroscopic guidance; endoscopy was used some-times in duodenal and colonic strictures for guidewire insertion. Oesophagealand tracheal combined recanalization and duodenal and biliary tree combinedrecanalization was performed respectively in 3 and 11 patients.Results: Technical and clinical success was achieved in 98 and 94%; 97 and92%; 98 and 94 % of cases respectively for, oesophageal, gastrojejunal and largebowel strictures. Better results occurred when dedicated stents with specific fea-tures for each anatomical site were used. The overall complication rate was 21%including, occlusion (12%), migration (3%), jaundice (1.2%), perforation (0.3%),stent rupture (0.3%) and others. The mean follow-up time for patients with neo-plastic strictures was 4.5 months.Conclusion: Stenting of the alimentary tract has a high procedural and clinicalsuccess rate. In inoperable patients it provides a prompt reduction or disappear-ance of symptoms, good quality of life and survival improvement. It can be con-sidered a safe and effective procedure and represents a less expensive treatmentcompared to palliative surgery which is affected by a higher mortality and mor-bidity rate.

B-619 11:51

De novo TIPS creation with the VIATORR® endoprosthesis: Long-termfollow-upM.T. Schmook, B. Angermayr, M. Peck-Radosavljevic, M. Schoder, T. Rand,J. Lammer, M. Cejna; Vienna/AT ([email protected])

Purpose: To evaluate the long-term outcome of TIPS with VIATORR® endopros-theses in patients at least 2 years after de novo TIPS creation.Methods and Materials: In 68 patients with portal hypertension (52m/16f; meanage 55, range 14-77) a transjugular intrahepatic portosystemic shunt (TIPS) wascreated with a GORE VIATORR® endoprosthesis. Etiology of liver cirrhosis waschronic alcohol abuse (58%), hepatitis C (9%), hepatitis B (3%), a combination ofalcohol and viral hepatitis (18%) and miscellaneous (12%). Child-Pugh stage Awas present in 13%, B in 60% and C in 27%. Indications for TIPS implantationwere refractory ascites in 60% and recurrent variceal bleeding in 40%. Proce-dure-related and patient survival data were collected. Venographic and combined(venographic + ultrasound + clinical data) follow-up data were used to calculatepatency rates.Results: Technical success was obtained in 100%. Major complications occurredin 7%. The portosystemic pressure gradient could be reduced from 22 mm Hg to9 mm Hg. Thirty day mortality was 3.4%. Overall survival rates were 93/88/76/26% after 6/12/24/48 months. Venographic primary and secondary patency rateswere 76/55/36/23% and 100/97/88/68% respectively after 6/12/24/48 months.

Combined evaluation primary and secondary patency rates after 6/12/24/48months were 81/64/48/33% and 98/96/90/72%, respectively. TIPS dysfunctionlead to re-intervention in 42% of the patients. Due to persistent hepatic enceph-alopathy TIPS reduction was necessary in 10% of patients. 4 patients underwentOLT.Conclusion: TIPS creation with the VIATORR® endoprosthesis is safe and ef-fective. Long-term success rates are not as promising as early and mid-termrates presented in previous studies.

10:30 - 12:00 Room I

Neuro

SS 1411Neurodegenerative brain diseasesModerators:M. Essig; Heidelberg/DES. Karampekios; Iraklion/GR

B-620 10:30

Vascular versus neuronal defects in ALS: An fMRI studyC. Sage, R. Peeters, Y. Zhou, W. Robberecht, S. Sunaert; Leuven/BE([email protected])

Purpose: To evaluate whether the motor impairment seen in ALS patients can beattributed to vascular and/or neuronal defects in the brain using BOLD fMRI.Methods and Materials: 33 ALS patients and 21 age- and sex-matched controlsunderwent fMRI examination using a 3 T MR imaging scanner. In the first ses-sion, a motor task was performed with the right hand to assess the activationpattern in the brain when executing voluntary movements. In the second session,an alternating hyperventilation - breath-hold task was performed to assess thevascular reactivity in the brain without interference of task-induced activations.Data was analyzed using SPM2 and activation patterns of both tasks were visu-alized. Additionally, a ROI analysis of different areas was performed on the imag-es of both sessions.Results: In the motor task both groups showed activation of areas typical forvoluntary movement. In ALS patients, there was a recruitment of both motor andextramotor areas. The ROI analysis showed significant increase of percentagesignal change of motor areas (except M1) in ALS patients.Vascular reactivity seemed to be impaired in ALS patients in the motor cortexand frontal cortex, whereas there was no difference in vascular reactivity in thevisual cortex.Conclusion: ALS patients recruit additional areas for the generation of voluntarymovement when compared to healthy controls. Vascular compliance is reducedin frontal and especially motor cortex of ALS patients when compared to healthycontrols. Thus, there are both vascular and neuronal defects leading to the phys-iopathology seen in ALS.

B-621 10:39

White matter changes in amyotrophic lateral sclerosis studied by DT-MRimaging at 3 TC. Sage, R. Peeters, Y. Zhou, W. Robberecht, S. Sunaert; Leuven/BE([email protected])

Purpose: To evaluate white matter integrity in ALS patients compared to healthyvolunteers using DT-MR imaging.Methods and Materials: 33 ALS patients and 21 controls were scanned using astandard DTI-scanning protocol (SE-EPI; 16 directions; b= 800 mm2/s; 2iso mmresolution) on a 3 T MR imaging scanner. The DTI images were coregistered anddiffusion-encoded fractional anisotropy (FA) maps were then calculated. Fibertracking analysis of the corticospinal tract (CST) was performed on DTI basedcolor maps, providing a visualisation of left and right CST. Furthermore, FA andADC (apparent diffusion coefficient) maps were generated and normalized to theMNI template. These images were entered in a voxel-based analysis.Results: Fiber tracking of CST showed a significant decrease of FA and a signif-icant increase of MD in ALS patients compared to healthy controls. When divid-ing the CST into motor and sensory parts, the above mentioned changes werepresent in the motor part of the CST, whereas no significant changes of FA/MDwere found in the sensory part. In the voxel-based analysis, a distinct pattern ofFA/MD changes was demonstrated, showing a significant increase of FA in ALSpatients throughout the CST as well as in the corpus callosum, frontal and pari-etal areas. A significant increase of MD in ALS patients was demonstrated in thecaudal part of the CST.

Mo

nd

aM

on

da

Mo

nd

aM

on

da

Mo

nd

ayy yyy

Scientific Sessions

B 281C D E FA G

Conclusion: ALS patients suffer from significant white matter loss mainly in themotor part of the corticospinal tract. Degeneration also occurs within the projec-tion fibers from/to the primary motor cortex: Corpus callosum, frontal and parietalareas.

B-622 10:48

Radiological features, clinical significance and utility of DaTSCAN's in themanagement of Parkinson's diseaseC. Cronin, D. Lohan, N. Ellis, T. Counihan, D. O'Keeffe; Galway/IE([email protected])

Purpose: Parkinson's disease is diagnosed in more than 1/200 worldwide; how-ever 20-25% of these are estimated to be misdiagnosed. As the average popula-tion age is increasing, the need for accurate diagnosis and treatment is highlighted.We retrospectively reviewed the radiological features, and in correlation with clin-ical background, determined the clinical and radiological utility of DaTSCAN's indiagnosis and management of 40 diverse patients with atypical presentation's,suspicious for Parkinson's.Methods and Materials: 40 patients' (22/40-M, 18/40-F, aged 28-80, mean 60.5years, Hoehn/Yahr Stage 1-2) images were reviewed independently by 2 radiolo-gists, discrepancies resolved by consensus. Striatial radiotracer uptake was as-sessed visually, abnormal images graded on a 1-4 scale, and correlated clinically.Results: Number/severity of symptoms and sides involved, correlated with de-gree of deficit at visual assessment. Of 40 atypical presentations, 12/40 suspi-cious for essential tremor (ET), 10/40 for drug/toxin induced Parkinson's (D/TIP),6/40 for Parkinson-plus syndrome (PPS) and 12/40 for early PD (without classi-cal triad). DaTSCAN confirmed diagnosis of ET in 5/12 (7/12 features of IPD), D/TID in 5/10 (5/10 features of IPD); PPS confirmed in 2/6 PPS group, IPD con-firmed in 12/12 suspicious for IPD. Imaging changed clinical diagnosis in 12/40(30%), medication in 17/40 (42.5%) (12/40 (30%) receiving a new medication, 5/40 (12.5%) medications deemed unnecessary).Conclusion: Although the cost of a DaTSCAN is not inconsiderable (approxi-mately 5-10 months of anti-Parkinson's medications), long-term cost: benefit is infavour of scanning, reducing patient anxiety, limiting further investigations/inap-propriate medications and ensuring best patient management. Our institutionalexperience demonstrates DaTSCAN's have a significant role in the diagnosisand management of unusual presentations for Parkinson's.

B-623 10:57

Value of volumetric and diffusion tensor MR imaging in differentiation ofearly-stage MSA-P and idiopathic Parkinson's diseaseJ.-J. Lu, F. Feng, Z.-Y. Jin; Beijing/CN ([email protected])

Purpose: The differential diagnosis between typical idiopathic Parkinson's dis-ease (PD) and the striatonigral variant of multiple system atrophy (MSA-P) isoften difficult because of the presence of signs and symptoms common to bothforms of Parkinsonism, particularly at symptom onset. To find the correct diagno-sis early in the course of the disorders is, however, of major importance withrespect to prognostic aspects and therapeutic strategies.This study investigated striatal, midbrain and cerebellar findings in MSA-P andPD patients in comparison with normal controls with three dimensional magneticresonance imaging (3D MR imaging) based volumetry, together with DTI quan-tification in the nigrostriatal projection to increase the differential diagnostic ac-curacy between both disease entities.Methods and Materials: 11 patients with MSA-P within 5 years from symptomonset, and 12 patients with PD of similar duration were studied by 3D based MRimaging volumetrics and diffusion tensor imaging (DTI) using a GE excite II MRmachine. 10 age-matched healthy controls were also studied. Direct measure-ment of the striatum, midbrain and cerebrellum were made and data were nor-malized by 3D brain volumetry. The fractional anisotropy (FA) values of magneticresonance diffusion tensor imaging in ROI of the extrapyramidal system includ-ing globus pallidus, putamen, head of caudate and substantia nigra, were com-pared among the groups using one-way ANOVA.Results: MSA-P differed significantly from PD patients in terms of decreasedputaminal and cerebellar volume. The FA significantly decreased in substantianigra and globus pallidus in both MSA-P and PD groups, compared with control,and the decrease was more prominent in MSA-P patients than in PD patients(Table 1). MR imaging volumetry of the midbrain region did not contribute to thedifferential diagnosis between PD and MSA-P.Conclusion: The extent and spatial distribution of morphological changes in thestriatum and cerebellum permit the differentiation of MSA-P from PD. Assumingthat the loss of FA parallels the neuronal change in the brain, the results indicatethat neuronal loss in nigrostriatal projection is more severe in MSA-P patients

than idiopathic Parkinson's disease in the early stage. Close comparison of stri-atal and cerebellar volume in addition to FA in the nigrostriatal projection maycontribute to the early differential diagnosis of MSA-P and Idiopathic Parkinson'sdisease.

B-624 11:06

Determination of regional differences in content and in T2 relaxation timesof the main cerebral metabolites in patients with Parkinson's disease (PD):In vivo 1H MRS studyV. Rogozhyn, Z. Rozhkova, I. Karaban', N. Karaban'; Kiev/UA([email protected])

Purpose: From the analysis of 1H MRS data and their echo-time dependence wepropose the quantitative indicators for the characteristics of the brain state inpatients with PD.Methods and Materials: Two groups of patients are studied by 1H MRS with1.5 T Magnetom Vision (Siemens). The 1st group (PG) includes 70 patients withPD (48-74y). The 2nd group (VG) consists of 20 healthy volunteers (50-73y). Spectraare recorded in the temporoparietal cortex (TC), lentiform nucleus (LN), putamen(P) and substantia nigra (SN) with the STEAM method: TR/TE=1500/135, 155,175, 200, 270 ms.Results: We introduce two indicators: the metabolite content AM as the peak areaand the concentration CM=AM/S. We describe the metabolic state in VOI by thetriad T*={ACho, ACr, ANAA}, where ACho, ACr, and ANAA are the peak areas of thesignals. For each of the areas we assign three values: 1.2,3, to obtain six spectralconfigurations: 1* ={1.2,3}, 2* = {2.1,3}, 3* = {1.3,2}, 4* = {3.2,1}, 5* = {3.1,2}, 6*= {2.3,1}. In the VG triads 1* and 2* dominate in all VOI. In the PG the mostfrequent configurations in TC are 2* and 5* and in LN, P and SN are 5* and 6*. InPG the mean T2

M values are: T2Cr = 220-230 ms, T2

Cho, T2NAA = 380-400 ms. From

comparison of CM and T2M in the both groups the shortening of T2

M and decreas-ing of CNAA in the PG in LN and in SN are obtained. In the TC the T2

M values aresimilar in PG and VG.Conclusion: This study gives a new insight into brain biochemistry in patientswith PD.

B-625 11:15

Hypertension-related and drug-induced PRES: Is there a difference in theMR appearance?C. Mueller-Mang, T.G. Mang, D. Prayer; Vienna/AT([email protected])

Purpose: Posterior reversible encephalopathy syndrome (PRES) is a recentlyrecognized entity, characterized by specific neurological deficits, a characteristicimaging pattern, and usually a benign course. Primarily related to hypertensiveconditions, many other etiologies, notably neurotoxic agents, have been associ-ated with PRES. The aim of this study was to determine differences in the MRfeatures of PRES based on the underlying etiology.Methods and Materials: Fifteen patients with clinical and imaging findings con-sistent with PRES were included. All exams were performed on a 1.5 T super-conducting system (Gyroscan NT-Intera, Philips). Based on their underlyingetiology, we classified the patients into 2 groups: one with exposure to neurotoxicsubstances such as immunosuppressants (n = 6) or chemotherapeutic agents(n = 3), and the other with hypertension-related conditions (n = 6), including 4patients with eclampsia.Results: In all but one patient, FLAIR imaging revealed hyperintense lesions inthe subcortical white matter of the parieto-occipital regions bilaterally, partially inan asymmetrical distribution (n = 7). All patients with hypertension-related PREShad additional lesions in the basal ganglia (n = 9), whereas this region was notaffected in the other patients. Brainstem involvement was seen only in the neuro-toxic group (n = 3). There was no predominance of involvement of the corpuscallosum (n = 2) and the cerebellum (n = 4) in either group. In one patient witheclampsia, diffusion-weighted imaging (DWI) revealed high signal lesions with alow apparent diffusion coefficient (ADC), indicating progression to infarction.Conclusion: Our preliminary results suggest that there are differences in the MRappearance of hypertension-related and drug-induced PRES, possibly reflectingdifferences in underlying pathophysiology.

Scientific Sessions

282 C D E FBA G

B-626 11:24

Visual attention deficits in Alzheimer's disease: An fMRI studyJ. Hao1, K. Li2, K. Chen1, K. Li2; 1Shanghai/CN, 2Beijing/CN([email protected])

Purpose: To define the anatomical areas of activation associated with visual at-tention processing and to define deficits or changes that may occur in AD pa-tients compared with control group.Methods and Materials: Thirteen AD patients and thirteen age- and education-matched normal controls were tested in two visual search tasks (one was a con-junction task, where feature binding is required, the other was a subset task,where group stimuli are needed without feature binding) using fMRI techniques.Data analysis was performed by Statistical Parametric Mapping software (SPM99). Voxel-by-voxel t statistics were used to compare activated brain areas be-tween patients and controls.Results: Our findings suggest that both search tasks are controlled by partiallyoverlapping cerebral networks, including parietal, frontal and occipito-temporalcortical regions. The AD patient group showed less activation in both parietallobes and the left frontal regions, while increased activation was found in theright frontal lobes and the right occipito-temporal cortical regions with the con-junction task. In the subset task, decreased activation in AD patients was seen inthe left parietal lobe and bilateral frontal lobes, while increased activation wasseen in both medial temporal lobes. In addition, for the comparison between tasks,the difference is very small for AD patients. Control group showed a higher ampli-tude in the right prefrontal region, temporal cortical regions and parietal lobe.Conclusion: AD patients have a particular impairment in both search tasks. Wesuggest attention deficits in AD patients may be attributed to both binding prob-lem and grouping inefficiency.

B-627 11:33

A voxel-based morphometric study of Alzheimer's diseaseJ. Hao1, K. Chen1, K. Li2; 1Shanghai/CN, 2Beijing/CN([email protected])

Purpose: Up till now, the study of regional gray matter atrophy in Alzheimer'sdisease (AD) has been assessed with regions of interest (ROI), but this methodhas many shortcomings. In this study we used the fully automated voxel-basedmorphometry (VBM) approach in order to provide a full-brain assessment of ADmorphology.Methods and Materials: 19 patients with AD and 15 healthy subjects of similarage and gender ratio were included in the study. Magnetic resonance imagingwas performed on a 1.5-T Siemens Sonata MR imaging system (Siemens, Erlan-gen, Germany). Whole brain T1-weighted 3D MPRAGE (magnetization-preparedrapid-acquisition gradient echo) data sets were acquired in the sagittal plane(TR, 1900 ms; TE, 3.93 ms; slice thickness, 1.7 mm; skip, 0.85 mm; FOV, 250 mm;Matrix, 448×512; Flip Angle, 15°). Data analysis was performed by StatisticalParametric Mapping software (SPM 99) in conjunction with MATLAB version 6.5to generate gray matter density map.Results: Our results confirmed earlier findings with traditional MR imaging meas-urement, but additionally we demonstrate global asymmetrical cortical atrophywith sparing of the sensorimotor cortex, occipital lobe and cerebellum. Moreover,atrophy of right caudate head and left medial thalamus were showed.Conclusion: Our results offer a comprehensive view of the atrophy pattern inAD. The consistency of these findings with histopathological descriptions dem-onstrates the utility of the VBM technique for investigating the neuroanatomy ofAD.

B-628 11:42

MR proton spectroscopy of the brain in Wilson's disease: Analysis oftreatment effectsW. Szeszkowski, M. Golebiowski, B. Tarnacka, A. Czlonkowska; Warsaw/PL([email protected])

Purpose: To assess treatment effects in patients with Wilson's disease using anMR proton spectroscopy method.Methods and Materials: Fourteen patients diagnosed with WD and presentingonly with neurological symptoms were studied before and after fourteen-monthchelation therapy. A control group consisted of 12 healthy volunteers (age andsex matched). Single-voxel spectra were acquired on a 1.5 T scanner, usingPRESS sequences with parameters of 22 and 2000 ms (TE/TR). Voxels(15x15x15 mm) were located in the center of the left and right globi pallidi andleft and right thalami separately for each examination. All results were expressedas metabolite/creatine ratios. Data was analyzed using nonpaired t-test for signif-icance testing; p < 0.05 indicated a statistically significant difference.

Results: The most significant metabolic changes in WD patients before treat-ment were located in the globi pallidi. There were notably increased Lip/Cr(p < 0.004), Glx/Cr (p < 0.025), Cho/Cr (p < 0.03) and decreased NAA/Cr(p < 0.008) ratios as compared to the control group. In post-treatment follow-upstudies, decreased Lip/Cr ratio (p < 0.04) for globi pallidi and increased NAA/Crratio (p < 0.04) for thalami were observed when compared to pre-treatment re-sults. The differences in findings between the left and right globi pallidi and thala-mi for the same patient were found to be insignificant.Conclusion: The reversible changes in NAA/Cr and Lip/Cr ratios may be due tofunctional and probably anatomical neuronal disturbances as an end result ofcopper deposition. The chelation therapy of WD reduces copper deposition. Thepresence of decreased NAA/Cr in WD patients does not necessarily reflect celldeath. The study showed new potential value of MRS in the assessment of treat-ment effects.

B-629 11:51

Progressive multifocal leukoencephalopathy (PML): Value of diffusion-weighted and contrast enhanced MR imaging for diagnosis and treatmentcontrolW. Küker1, I. Mader2, U. Herrlinger3, G. Quaghebeur1, T. Nägele3, P. Pretorius1;1Oxford/UK, 2Freiburg/DE, 3Tübingen/DE ([email protected])

Purpose: Progressive multifocal leukoencephalopathy (PML) is caused by thereplication of JC virus in oligodendrocytes of immunocompromised patients. Di-agnosis usually relies on the PCR-based demonstration of JC virus DNA in thecerebrospinal fluid. Since previous reports have suggested beneficial effects ofantiviral therapy, and because effective treatment is possible but expensive, non-invasive early diagnosis and treatment control are highly desirable.Methods and Materials: Repetitive MR imaging examinations (2 to 9) were ob-tained in eight immuno-compromised patients (age 40 to 67 years, 6 male, 2female) with classical clinical and imaging findings of PML. Five patients hadunderlying hematological disorders and three AIDS. PCR of the CSF specimenwas positive for JC virus DNA in seven patients. CSF samples were obtained inall cases. MR imaging-sequences included T2-, T1- and diffusion-weighted (DW)images in all patients as well as diffusion-tensor imaging (DTI) in four cases. DTIwas once performed at 3 T, in the remaining patients at 1.5 T.Results: PCR of the CSF specimen was positive for JC-virus DNA in five pa-tients. MR imaging showed areas of T2-hyperintensity with involvement of thesubcortical U-fibers and restricted diffusion in all patients. Areas of diffusion ab-normality correlated with disease progress. Contrast enhancement was encoun-tered once after successful treatment and heralded clinical remission with viruselimination from the CSF.Conclusion: MR imaging including DW and contrast- enhanced images may beused to evaluate disease activity in PML. Contrast enhancement may indicate aninflammatory response and thus herald immunological virus elimination.

10:30 - 12:00 Room K

Pediatric

SS 1412Body imagingModerators:K. Aristidou; Nicosia/CYS. Vegar-Zubovic; Sarajevo/BA

B-630 10:30

Prospective evaluation of image quality by means of a patient image-gallery(IG) for dose reduction in pediatric 16-slice multidetector-row spiralcomputed tomography (MDCT)D. Honnef1, J.E. Wildberger1, G. Haras2, G. Staatz1, M. Das1, C. Hohl1,M. Barker1, R.W. Günther1, A.H. Mahnken1; 1Aachen/DE, 2Forchheim/DE([email protected])

Purpose: To prospectively adjust dose settings in pediatric 16-MDCT with a newlydeveloped patient-IG for dose reduction.Methods and Materials: 53 examinations (chest n = 26, abdomen n = 14, pelvisn = 13) in 37 children (mean 8.1 ± 6.6 years) were planned by means of a pa-tient-IG prior to contrast-enhanced 16-MDCT. The IG consists of weight depend-ent reference examinations with simulated dose reduction. Examinations wereplanned by selecting scan parameters of the IG according to the required imagequality. Objective image quality was assessed by measuring CT-attenuation inthe autochthonal and psoas muscle with the standard deviation being assigned

Mo

nd

aM

on

da

Mo

nd

aM

on

da

Mo

nd

ayy yyy

Scientific Sessions

B 283C D E FA G

to image noise. 3 radiologists rated the image quality using a 4-point gradingscale not knowing which images belonged to the IG and the actual MDCT-scans.Results: Statistical analysis did not show any significant difference in image noisebetween the scans of the IG and the real examinations (thorax p > 0.1617, abdo-men p > 0.1042, pelvis p > 0.1984). There was also no significant difference invisually assessed image quality (thorax-soft tissue p > 0.17, thorax-lung p > 0.16,abdomen p > 0.17, pelvis p > 0.36). Compared to weight-adapted pediatric stand-ard protocols, using the IG resulted in further mean dose reduction in abdominal(mean 41.7% ± 11.3%) and pelvic (mean 44.1% ± 11.3%), but not in chest CT(mean 0.7% ± 18%).Conclusion: The IG is a feasible tool for routine 16-MDCT, which allows for apre-exam overview of the expected image quality. Radiation exposure can beoptimized with regard to patient weight. It allows for problem-oriented examina-tion planning and may be used as a guideline for pediatric MDCT-examinations.

B-631 10:39

A prospective study of radiation dose in paediatric interventionalcardiology: Preliminary resultsS.L. Mc Fadden1, C. Dhelft2, C. Hughes1, A. Mcgee2, L. Rainford2, P. Brennan2,I. Bradbury1, J. Winder1; 1Belfast ,/UK, 2Dublin/IE ([email protected])

Purpose: This study aims to determine current dose levels in paediatric Inter-ventional Cardiology (IC) in Ireland. IC procedures are responsible for some ofthe highest radiation exposures. Studies have indicated that the radiation risk is4-8 times higher for infants than for adults. A recent United Nations ScientificCommittee on the Effects of Atomic Radiation (UNSCEAR) report emphasisesthat exposure to ionising radiation during childhood may result in a 2-3 fold in-crease in certain detrimental effects (including solid cancers) compared to adults.Methods and Materials: A prospective study has been carried out recordingdata on 179 paediatric patients in 2 specialised centres in Ireland. Both centresused identical bi-plane equipment. Radiation doses were recorded with the aid ofDose Area Product meters (DAP) along with procedural technical details.Results: The mean patient age was 5.47 years (range 0.003-15.89) and weight21.49 kgs (range 2.4-77.4). The mean fluoroscopy time and Dose Area Product(DAP) to patients were 14.3 minutes and 787.8 cGycm², with a maximum of 55 min-utes and 15,351 cGycm², respectively. These data were not normally distributedaround a mean but had a skewed distribution with a significant number of pa-tients receiving large doses due to the range of procedural complexity.Conclusion: At present, no Diagnostic Reference Levels (DRLs) exist for paedi-atric IC. National or local DRLs need to be established and implemented in thishigh-risk group of patients. If DRLs are exceeded, optimisation of the imagingprotocol should be carried out with a view to reducing the radiation burden.

B-632 10:48

Structural lung abnormalities in bronchopulmonary dysplasia: A newscoring system for standardised lung-CT scan evaluationV.V. Casotti1, L. Rubin1, M.H. Lequin2, W. Hop2, H. Tiddens2; 1Verona/IT,2Rotterdam/NL ([email protected])

Purpose: Frequently in BPD patients lung function impairment and respiratorysymptoms persist in childhood. Little is known about the structural substrate re-sponsible for these abnormalities.To describe long term structural changes on CT scans of BPD patients, develop-ing a scoring system to define these abnormalities in a systematic and reproduc-ible fashion.Methods and Materials: Retrospective analysis of patients with a diagnosis ofBPD and who had a thorax CT scan was performed. CT's were evaluated toidentify and record all structural abnormalities. Next, a scoring system was devel-oped and 3 independent readers scored all scans to determine inter-observerand intra-observer reproducibility.Results: 37 BPD patients were identified. Age at CT was 2 months to 17 years. Ascore was developed evaluating the 5 lobes, that included the following lesions:decreased density of parenchyma; interstitial fibrosis, thickening of the interlobu-lar septa, triangular sub pleural opacities; bullae; air trapping; atelectasis andconsolidation; ground glass; bronchiectasis and airway wall thickening.Mean CT score was 23 (3 to 45). Interobserver agreement assessed by intra-class correlation coefficient was good (ICC from 0.83 to 0.89); intra-observeragreement was excellent (ICC 0.933). A significant negative correlation betweenthe mean score and the FEV1 was observed (R=0.612, P < 0.002).Conclusion: CT scans of a selection of BPD patients show considerable struc-tural abnormalities. A CT-scoring system was developed that was highly repro-ducible. A correlation between structural and functional abnormalities was found.

B-633 10:57

Magnetic resonance imaging of the fetal oesophagusP.C. Brugger, D. Prayer; Vienna/AT ([email protected])

Purpose: To establish the normal appearance of the fetal oesophagus as seenwith magnetic resonance imaging (MRI).Methods and Materials: 150 fetal MR imaging examinations of fetuses [19-38gestational weeks (GW)] without thoracic pathologies were studied. Four parts ofthe oesophagus were assessed on T2-weighted axial, sagittal and frontal sectionplanes and scored as being evident, fluid-filled or not: O1: cervical, O2: thoracicinlet to tracheal bifurcation, O3: tracheal bifurcation to diaphragm, O4: abdominalportion. In addition, 30 axial or sagittal dynamic studies were analysed.Results: Before 20 GW the fetal oesophagus could not be demonstrated. After-wards, O3 could be visualized in 95% of cases on axial, in 31% on frontal and in20% on sagittal scans. It was fluid filled in 64%. Parts of O2 were visualized in5.9% (axial), 8% (frontal) and 10.5% (sagittal). Delineation of O1 was possible in1.4% (axial), 5.6% (frontal) and 15.5% (sagittal). O4 was never visualized. Thewhole thoracic part could be visualized distended and fluid-filled on sagittal (4.9%)or frontal scans (4%). On dynamic scans, fluid-filled oesophagus was seen forseveral seconds.Conclusion: Visualization of the fetal oesophagus depends on section planeand oesophageal segment. The O3 portion can be usually demonstrated inde-pendent of fluid-filling, since the T2-weighted hyperintense lungs supply contrast.T2-visualization of O1-2 requires fluid-filling (during swallowing). The distendedcondition of long parts of the oesophagus may reflect the immature nature of fetaloesophageal neuromuscular activity, since intrathoracic pressure cannot be re-sponsible and the duration is too long for peristalsis.

B-634 11:06

Magnetic resonance imaging of the fetal gallbladder and bileP.C. Brugger, D. Prayer; Vienna/AT ([email protected])

Purpose: The appearance and development of the fetal gallbladder was studiedwith magnetic resonance imaging (MRI) with special reference to fetal cholelith-iasis.Methods and Materials: 200 fetuses [20th -40th gestational weeks (GW)] wereimaged on a 1.5 T magnet using T2-, T1-weighted (w) and steady-state free-precession sequences in at least two section-planes. Visibility was assessed onall sequences and gallbladder dimensions (maximum length and width) were re-corded. Signal intensities of bile were scored and measured on T1w images (com-pared to liver) and a bile/liver ratio was computed.Results: The gallbladder could be visualized in 98% of cases. Both dimensionsincreased in a linear manner, though widths showed more variation. Until 30 GW,T2w hyperintensity of bile corresponded to T1w hypointensity. Afterwards an in-crease of the bile/liver ratio was found, with bile being iso- or hyperintense to liveror even isointense to meconium, which corresponded to T2w hypointensity. Thisincrease was more pronounced in males.Conclusion: The fetal gallbladder can be visualized with MR imaging by virtue ofthe signal properties of fetal bile, which shows age-dependent signal character-istics. An increase in signal intensity is a normal finding after 30 GW and morepronounced in male fetuses. This may cause non-visualization of the gallbladderon T2w sequences. It is proposed to consider bile iso- or hyperintense to hepatictissue as sludge and refer to bile isointense to meconium as gallstones. Thesenew findings may offer new insights into the physiology of fetal bile and aid inter-pretation of fetal and neonatal abdominal MR imaging scans.

B-636 11:15

Role of genitography in the era of US and MR imaging in the evaluation ofintersex childrenC.J. Das, A.K. Haloi; Delhi/IN ([email protected])

Purpose: To evaluate the genital anatomy in intersex children using genitogra-phy, US and MR imaging with surgical correlation.Methods and Materials: Genitography, trans-abdominal US and MR imagingwere performed in 33 children of 13 days-17 years age group after proper clinicalevaluation, karyotyping and hormonal analysis. The ability to detect a commonurogenital sinus (UGS), vagina, uterus and gonads by each modality were as-sessed.Results: Common UGS was properly evaluated in 16 (48%) cases by genitogra-phy, whereas US and MR imaging were not sensitive enough to detect it. Vaginawas detected by genitography, US and MR imaging in 100%, 39.4% and 72.4%cases respectively. Uterus was detected in genitography, US and MR imaging in69%, 75% and 100% cases respectively. US detected 21 inguinal and 8 intra-

Scientific Sessions

284 C D E FBA G

abdominal gonads (total 44%), whereas MR imaging detected 40 inguinal and 11intra-abdominal gonads (total 77.3%). Histological examination was consideredas gold standard for gonadal characterization. The detection rate of each internalorgan by the 3 modalities was different for different intersex disorder types.Conclusion: MR imaging is the single most useful investigation to be performedin all intersex patients as it can give detailed anatomical information of the pelvisincluding information about Mullerian remnants and gonads. But MR imaging isnot sensitive enough to evaluate the length of common UGS. This information isrequired when corrective surgery of the genitalia is planned especially in patientswith conditions such as congenital adrenal hyperplasia. Genitography is the bestinvestigation to delineate the length of UGS.

B-637 11:24

The significance of missed reflux on VCUG, demonstrated only by VUSA. Anthopoulou, F. Papadopoulou, A. Fotopoulos, E. Siomou, J.A. Bokhrhli,F. Katzioti, K. Vlachos, S.C. Efremidis; Ioannina/GR ([email protected])

Purpose: In recent years, voiding-urosonography (VUS) has been found to havehigher sensitivity compared to fluoroscopic voiding-cysteourethrography (VCUG).The purpose of our study is to evaluate the clinical significance of vesicoureteralreflux (VUR) missed on VCUG and demonstrated only by VUS as an expressionof pathological findings on renal cortical scintigraphy (DMSA).Methods and Materials: A total of 135 children with reflux on either VCUG orVUS (54 males, 81 females, median age: 1.23 years) were also studied with DMSAwithin 6 months following the first proven UTI. Ninety-six of them were VUR pos-itive on both examinations while the remaining 39 were positive only on VUSusing harmonic imaging and a 2nd generation U/S contrast-agent (SonoVue®,Bracco, Italy). Kidney abnormalities associated with VUR were excluded fromour study. A result of reflux on either examination was studied in association withDMSA findings using McNemar's, chi-square, t-Student's and Fisher's statisticaltests. Results were analyzed according to age (< 2 yrs vs > 2 yrs).Results: Overall, 190 refluxing and 80 nonrefluxing KUU were evaluated. Renaldamage was detected in 50/190 refluxing KUU (26%) and in 4/80 (5%) nonre-fluxing ones (p = 10-6).Scar formation was found in 9 out of 78 KUU (11.5%) with reflux missed on VCUGand demonstrated only on VUS, compared to 41 from 192 KUU (21%) refluxingon both methods (p = 0.09). The difference of scar formation was also not signif-icant between children < 2 yrs vs > 2 yrs (36/190KUU:19% vs 18/80:22.5%,p = 0.6).Conclusion: Reflux missed on VCUG and demonstrated only on VUS correlatespositively with renal scar formation just as reflux depicted on VCUG does.

B-638 11:33

Virtual cystoscopy in children based on three-dimensional ultrasound(3DUS) data: Preliminary resultsM. Riccabona, A. Haberlik, E. Ring; Graz/AT([email protected])

Purpose: To evaluate the feasibility and diagnostic potential of 3DUS based vir-tual cystoscopy in infants and children, presenting our initial and preliminary ex-periences.Methods and Materials: 10 patients (age: 1 to 14 years) who presented withbladder pathology on conventional 2DUS (Sequoia 512, Acuson/Siemens, Moun-tain View, USA) underwent a subsequent 3DUS study with reconstruction of a"virtual cystoscopy" from the 3DUS data set using gradient light shaded surfacerendering (Voluson Expert Pro, Kretztechnik/GE, Zipf Austria). The results werethen evaluated for artifacts and diagnostic resolution and compared to routinetransurethral cystoscopy findings.Results: 3DUS acquisition with sufficient image quality for virtual cystoscopywas feasible in all patients without sedation. Patient diagnoses included intra-vesical polyps or ureteroceles, ureter ostium pathology, bladder wall trabecula-tion and bladder diverticula. In all patients the virtual cystoscopy images andfindings matched the result from transurethral cystoscopy; however, a significantchange in diagnosis of the initial 2DUS report was achieved only in 1 patient,though 3DUS conspicuity was higher and presentation of the findings was moreconvincing using 3DUS based virtual cystoscopy display.Conclusion: 3DUS based virtual of the urinary bladder is feasible in children. Itoffers a great visualization tool that matched transurethral cystoscopy findings inall patients and has the potential to reduce diagnostic cystoscopies in those pedi-atric patients who do not need biopsy or endoscopic therapy.

B-639 11:42

Sonographically guided reduction of meconium ileus in preterm neonatesM. Riccabona, M. Haim, J. Kutschera; Graz/AT([email protected])

Purpose: To report our preliminary experience with ultrasound (US) guided re-duction of meconium ileus in preterm neonates.Methods and Materials: 5 preterm neonates with bowel obstruction due to meco-nium ileus underwent bedside enema reduction using US (Sequoia 512, Acuson/Mountain View, USA) for guidance of the prewarmed and saline diluted Jopami-ro® (Schering, Berlin, Germany) enema, as they were too ill to be brought to thefluoroscopy suite for the conventional fluoroscopic procedure.Results: In all 5 patients US managed to guide the transrectal catheter into thesigmoid or descending left colon for effective and gentle instillation of the enema.With repeated enema instillation up to 40 mls, all 5 meconium plugs were dis-solved and obstruction was relieved, even in a child that later demonstrated adistal small bowel stenosis and on the initial US was thought to show an unusedmicrocolon. No complications were observed during and after these procedures.Conclusion: US guided reduction of meconium plugs appears to be a safe andeffective method that reduces radiation to these preterm babies and can easilybe performed at the patient's bed side in the NICU.

10:30 - 12:00 Room L/M

Radiographers

SS 1414Radiography: Enthralling and multifacetedModerators:S. Huber; Munich/DEC. Roche; Galway/IE

B-640 10:30

Autonomous, sonographer provided, symptomatic breast clinics for womenunder the age of 35 yearsM. Holland1, C. Theodorakou2, G. Massey1, R. Jones1, I. Britton3; 1Boston/UK,2Hertford/UK, 3Stoke-upon-trent/UK ([email protected])

Purpose: To establish the feasibility for autonomous sonographer practice, asthe radiology lead, in a fast track symptomatic breast clinic, specifically for wom-en under 35 years.Methods and Materials: The level of agreement between sonographer and radi-ologist was compared in a prospective, double blind, comparative study with in-ter-observer and intra-observer reliability. Trust ethical and R&D approval wereobtained. Data collection was categorical, R1-normal/benign, R2-probably be-nign, R3-indeterminate, R4-probably malignant and R5-malignant. Comparisonwas made with the radiologist opinion or histology, where available. Agreementwas assessed by descriptive kappa statistical analysis (SPSS).Results: 78 patients were examined from December 2004 to May 2005, agerange 14-34 (median = 29) years. 47 were normal, 18 demonstrated aberranceof normal development. 13 biopsies were performed, with pathological correla-tion. There were 2 disagreements (1) lymphnode R2/fibroadenoma R3 and (2)incidental cyst R2/normal R1. Both of these represented disagreements of cate-gorisation, neither influenced subsequent management. A kappa value of 0.9was demonstrated between radiologist and sonographer, inter-observer and in-tra-observer, representing excellent agreement.Conclusion: A sonographer's opinion in a diagnostic breast clinic is equivalentto that of a radiologist, providing an accurate, reliable and effective clinic for pa-tients under 35 years.

B-641 10:39

The study of nosocomial infection and transmission by way of diagnosticimaging facilities in TaiwanC. Lin, Y. Huang; Taichung/TW ([email protected])

Purpose: Taiwan was impacted by severe acute respiratory syndrome (SARS) in2003. The high percentage of nosocomial infections in Taiwan suggests that in-fection from hospitalized patients is a major factor in the spread of disease. Thestudy was designed to investigate the epidemiology of nosocomial infections in-volving radiographic equipment.Methods and Materials: The study was in four parts. The first one was a ques-tionnaire to investigate the cleaning behaviour in different sizes of hospital. The

Mo

nd

aM

on

da

Mo

nd

aM

on

da

Mo

nd

ayy yyy

Scientific Sessions

B 285C D E FA G

second part was to spray three commonly found bacteria, staphylococcus au-reus, Escherichia coli, Pseudomonas aeruginosa, onto a radiographic table, cas-sette, and chest board to investigate the survival rate of each bacterium. Thesampling work for the radiographic table, chest board, and cassettes from the outpatient department (OPD) and accident and emergency room (A&E) were de-signed in the third part. The last part was to evaluate the effectiveness of threedifferent cleaning solutions, water, 70% alcohol and 0.05% bleach to bacteria.Results: Cassettes should be labelled for intensive care units use to decreasethe potential bacteria transmission rate. Bacteria could be survived between onehour and two weeks, it is recommended that radiographic equipment should becleaned after use. No significant difference was found between 70% alcohol and0.05% bleaches (p > 0.05).Conclusion: No evidence suggests radiographic equipment in OPD and A&E isa potential nosocomial transmission source. Despite this, cleaning should be madeafter each use. In light of convenience, the author recommended that 70% alco-hol is useful to clean radiographic equipment surfaces.

B-642 10:48

Dose reduction in skeletal radiography using a flat-panel detector onamorphous selenium: Long term results of radiation dose and imagequalityT. Lehnert, J. Wohlers, A. Wetter, K. Manegold, V. Jacobi, T.J. Vogl;Frankfurt/Main/DE ([email protected])

Purpose: Evaluation of a flat-panel X-ray detector, based on amorphous seleni-um with respect to skeletal radiography. Conventional images were comparedwith digital radiographs using identical and reduced radiation doses.Methods and Materials: 84 consecutive patients were studied prospectively usingconventional screen-film radiography (detector dose 2.5 microGy). Digital imag-es were taken from the same patients with detector doses of 2.5, 1.25, 0.625 and0.3125 microGy, respectively, using a flat panel detector with implemented P-tone/EVP image processing software. The active-matrix detector had a pixel sizeof 35 x 43 cm, a matrix of 2.5 x 3K, and a pixel size of 139microm. All hard copieswere presented in a random order to five independent observers, who rated im-age quality according to subjective and objective quality criteria. Results wereassessed for significance.Results: A statistically significant preference for digital over conventional imageswas revealed for all quality criteria, expect for over-exposure (detector dose 2.5microGy). Digital images with a dose reduction of 25% showed a statisticallysignificant supremacy compared to screen-film technique. The screen-film radi-ography was significantly inferior to flat-panel detector-technique at 50% dosereduction regarding bone cortex and trabecula, contrast and overall impression.Conclusion: Due to its high detective quantum efficiency and dynamic rangecompared with traditional screen-film systems, a dose reduction of up to 50%and more is possible by using a flat panel detector with implemented P-tone/EVPimage processing software without loss of image quality.

B-643 10:57

Radiographers' professional knowledge related to good patient careB.T. Andersson1, Å. Axelsson2; 1Lund/SE, 2Göteborg/SE([email protected])

Purpose: To define and describe radiographers' professional knowledge relatedto good patient care in the light of critical incidents in connection to radiologicalexaminations and treatments.Methods and Materials: A qualitative descriptive approach and design was cho-sen for the study. The method used for collection of and analysis was CriticalIncident Technique (CIT). A strategic selection was performed and consisted of14 radiographers working at hospitals in Sweden.Results: Data analysis resulted in two main categories, Professional Knowledgedirectly related to patients and Professional Knowledge indirectly related to pa-tients, respectively. Knowledge directly related to patients was subdivided intofour categories: to prepare, to support, to show respect and to be vigilant. Knowl-edge indirectly related to patients was subdivided into four categories; to plan, toassure quality, to handle the radiological image and to collaborate. In both maincategories examples of behaviour of positive as well as negative for good patientcare occurred. 279 critical incidents were identified.Conclusion: By the results of this study the professional skills and unique knowl-edge of the radiographer were elucidated, and the importance of solid healthcare knowledge was emphasized. The study has given insight to the radiogra-pher's nursing role and partially into the role as a medical technician. It alsoemphasises and clarifies a "tension" between the two.

B-644 11:06

Mobile, digital radiography for nursing homes' residentsF. Laerum, E. Sager, S. Oswold, J.C. Hellund; Oslo/NO([email protected])

Purpose: To improve radiology services to nursing home residents. There arethree times as many beds in nursing homes as in all Norwegian hospitals com-bined, each resident has in average more than three chronic diseases and 65%have dementia.Methods and Materials: A prototype combination of Sedecal X-ray equipmentwith Canon full-format digital detector (total weight 94 kg) was carried around tonursing homes on demand, in a Renault Kangoo car equipped with wheel-chairramp. The digital radiographic (DR) examinations were performed in the resi-dents' rooms, and cryptated image data either WI-FI uploaded via broad bandinternet, or brought by CD-ROM, to the radiology department's PACS-network forinterpretation and reporting. Questionnaires without patients' identification werereturned by the referring physician in 125 out of the first 195 examinations.Results: The DR image quality was assessed to be of satisfactory diagnosticquality. The questionnaires revealed that tentative diagnosis was verified in 58%and invalidated in 42 % of the cases, with consequences to therapy in 85% andnursing in 71%. Because of the residents' serious impairments, 10% had beengiven no alternative to DR bed-side in their nursing home. 8% had to be hospital-ized after the examination, while expensive patient transportation for out-patientradiology examinations, or hospitalization, could be avoided in 90% of cases.Conclusion: After more than 400 examinations to date, DR services have provedbetter for the residents, less expensive for society, and of good diagnostic quality.

B-645 11:15 !CT scanning technique for patients who cannot raise their armsK. Tsujioka, Y. Uebayashi, M. Anzui, A. Hattori, S. Narita, K. Shibata;Toyoake-city Aichi/JP ([email protected])

Purpose: Many patients cannot raise their arms when undergoing CT examina-tion. In these cases, the correct diagnosis cannot be obscured by the resultingartifact. We introduced an easy scanning method to solve this problem.Methods and Materials: We arranged the patient's CT examination, their armsare not put on the body side but both arms are arranged in front of the body. Theartifact was evaluated with standard deviation (SD) of CT number.Results: When the arms arranged in side of the body, SD of CT number was19.40. The arms arranged in front of the body, SD was 9.57. The artifact from theboth arms was able to be decreased by new patient positioning.Conclusion: It was a cause of the artifacts that the both arms and the spineexisted straight. This originates in a specific direction the fact that the amount ofthe projection data decreases. The both arms and the spine become not straightby putting the both arms on the front of the body. As a result, the artifact hasdecreased.

B-646 11:24

What should radiographers know about contrast agent inducednephrotoxicityG.A. Marshall; Lancaster/UK ([email protected])

Purpose: Radiographers work with contrast agents routinely, and although muchhas been written on this topic, little has been focussed particularly at radiogra-phers. This is becoming increasingly important as, via role extension, radiogra-phers especially in the U.K., are increasingly performing, under written schemesof work, that were previously designated medical tasks e.g. intravenous injectionof contrast agents. Consequently there is a need for radiographers to receivefurther education regarding these agents, and it is incumbent on them to under-stand the possible deleterious effects of these drugs.Methods and Materials: An extensive literature review was performed to estab-lish key facts re contrast agent induced nephrotoxicity (CIN); i.e. incidence, patho-genesis, cost implication, protocols to reduce CIN, as this information would bevaluable to injecting radiographers.Results: The reduction of CIN via the identification of patients at high risk, con-sideration of the dose and osmolality of contrast agent injected, water and sodi-um intake regimens peri-procedurally, the role of pharmacological agents andthe use of alternative modalities will be evaluated.Conclusion: Protocols that reduce the incidence of CIN will be presented.

Scientific Sessions

286 C D E FBA G

B-647 11:33

Evaluation of Crohn disease activity using perfusion MR imagingM. Duriez, S. Gagnayre, S. Aufort, B. Gallix, J.-M. Bruel; Montpellier/FR([email protected])

Purpose: To determine the feasibility of morphological sequences and dynamiccontrasted enhanced MR imaging (DCE-MRI) to study Crohn's disease activity.Methods and Materials: 23 patients were included in this prospective study.Each patient underwent 3 different MR examinations, before treatment, 2 weeksand 6 weeks after. An amount of 57 MR was finally performed. Polyethylene gly-col (PEG) was administrated orally 45 minutes before performing exam. Using a1.5 Tesla Philips Gyroscan Intera®, morphological sequences including T2 SS,BTFE, and T1 WATS after gadolinium injection, was performed in axial and coro-nal planes. DCE-MRI images were added to these classical sequences withmultislice 3D TFE perfusion sequences (TR/TE/FA 3.3, 1.66.10. SL 5 mm, acqui-sition time: 180 ms. Flow rate 6 cc/s).Results: With the evolution of the MR imaging technology it's now possible toperform in the same exam morphological and DCE-MRI sequences for perfusionevaluation of the bowel wall. The role of the technician in radiology is crucial forpatient information and preparation and is directly related to image quality. Fillingof the small bowel by oral contrast agent is correlated to the volume of PEGabsorbed by the patient.Knowledge in GI tract anatomy and Crohn disease is mandatory for the techni-cian to be able to perform perfusion post treatment analysis.Conclusion: Abdominal MR imaging (morphological and DCE-MRI) is a feasi-ble, reproductive, radiation less and simple technique, very useful in the evalua-tion and prognosis of Crohn's disease.

10:30 - 12:00 Room N/O


SS 1413MRI and advanced technologyModerators:E. Atalar; Ankara/TRR. Wirestam; Lund/SE

B-648 10:30

Human imaging using small-size high temperature superconducting (HTS)tape coil in low-field MRI systemY.W. Wong, E.S. Yang; Hong Kong/HK ([email protected])

Purpose: High temperature superconductor (HTS) coil is effective in reducingreceiver-coil noise, resulting in substantial SNR improvement for small-size coil/sample and low-field system (< 1T). A single-turn 60 mm HTS tape receiver coil(8.92 MHz at 0.21T) is first demonstrated.Methods and Materials: A commercially available silver-alloy sheathed Bi(2-x)PbxSr2Ca2Cu3O10 (Ag/Bi-2223) HTS tape is adopted to fabricate our receivercoil, which possesses the advantages of lower cost and easier fabrication com-pared to traditional thin film. A new coil holder is designed to resist relatively largerecoil force experienced by small-size HTS coil. Inductive coupling is used forfrequency tuning and signal pick-up. A custom-made cryogenic system is em-ployed to accommodate our coil in LN2. Images are acquired using conventionalspin-echo (SE) pulse sequence (TR = 400 ms, TE = 30 ms, NEX = 2, slicethickness = 5 mm, matrix size = 256x256).Results: Images including phantom of saline solution, human hand and wrist arecaptured in our home-built 0.21T MR imaging system. Results show that the Q-factor of phantom-loaded (0.21T iso-centre) HTS coil at 77K (325.38) is 5 timesbetter than that of an equivalent copper coil at room temperature (64.25). SNRimprovement of about 350% is achieved with HTS at 77K (83.1) over convention-al copper coil at room temperature (22.9). Previous best result at 0.21T was again of 256%.Conclusion: Considering Q-factor of a coil with unchanged dimension is propor-tional to its resonant frequency, higher SNR is expected in high-field MR imagingsystem. Potential applications include body parts which are too small for largesurface coil.

B-649 10:39

New post-processing strategies for dynamic contrast enhanced MRimaging by combining in-phase and opposed-phase dataK.-H. Herrmann, A. Rauscher, W.A. Kaiser, J.R. Reichenbach; Jena/DE([email protected])

Purpose: To explore new image contrasts which result from different post-process-ing combinations of simultaneously acquired complex in-phase (IP) and opposed-phase (OP) data in contrast enhanced (CE) dynamic MR imaging by performingsimulation and phantom experiments.Methods and Materials: A simple 2-compartment model (water, fat) was ana-lyzed for a single voxel. A phantom containing several compartments with differ-ent fractions of water/fat, was simulated and visualized. The effects of differentsubtractions of the complex data on contrast were investigated. Additionally, aphantom experiment was performed at 1.5 T using a T1-weighted 3D dual-echogradient-echo sequence. (TR/TE1/TE2/FA=7 ms/2.38 ms/4.76 ms/12°). A ROIbased analysis was used to verify the simulation results.Results: The theoretically expected contrast amplification due to complex dataprocessing, compared to the conventional IP magnitude subtraction (CE minusnative), was confirmed in the experiment. Different processing of the dual-echodata makes it possible to directly map the fat content. The phantom measure-ments verified the simulated contrast gain by a factor of 2 and the possibility ofdirect fat imaging. A first proof of principle of contrast gain in partially volumedlesions was successfully demonstrated with dynamic breast MR imaging.Conclusion: Acquiring IP and OP data allows new post-processing combina-tions, which range from generating new image contrasts to significant contrastamplifications of partially volumed lesions. This proposed method can be usedfor better delineation of lesion boundaries and improved detection of small le-sions without relinquishing the conventional IP magnitude subtraction.

B-650 10:48 !Towards rapid analysis of diffusion tensor MR imagingD. Goksel, M. Ozkan; Istanbul/TR ([email protected])

Purpose: In diffusion tensor MR imaging (DTI), each voxel is assigned a tensorthat describes local water diffusion. In this study, an algorithm in Matlab is devel-oped to analyse DT images.Methods and Materials: The b factor in our study, which gives information aboutthe direction and amplitude of the diffusion gradient, is a 1x7 matrix in form B =[B0, Bxx, Byy, Bzz, Bxy, Bxz, Byz]. In our sample, 10 diffusion weighted imagesand a null image namely the T2 image creating a set of intensity images of size256x256x11 is used for the analysis. The relation can be expressed generally;B (n,:) has the information of the intensity image S (:,:,n). Regarding to that andthe Stejskal Tanner equation D = [Dxx, Dyy, Dzz, Dxy, Dxz, Dyz] is calculated inthe algorithm. For the validation of the algorithm a set of human cardiac diffusionMR images are used.Results: In this study, algorithm for analyzing the diffusion characteristics is de-veloped. The eigensystem D is calculated in every pixel, apparent diffusion coef-ficient ADC is represented with respect to D. The other characteristic values ofdiffusivity namely fractional (FA) and relative (RA) anisotropy, linear (cl), planar(cp), and spherical (cs) anisotropy values are calculated.Conclusion: There are still many drawbacks in the literature on the analysis ofDTI especially in neurological disorders. The classification of brain white matterfiber tracking is of great importance. Developing a reliable and rapid tractographytool for the clinical use is the future study of the work in progress.

B-651 10:57

Phase map based simulation of signal dephasing in gradient echo MRimagingA. Rauscher1, M. Barth2, S. Witoszynskyj3, J. Sedlacik1, J.R. Reichenbach1;1Jena/DE, 2Nijmegen/NL, 3Vienna/AT ([email protected])

Purpose: MR gradient-echo images (e.g. EPI) suffer from signal loss due to in-travoxel dephasing caused by field inhomogeneities. We present a phase-imagebased method to simulate and to investigate intravoxel signal dephasing.Methods and Materials: High-resolution, 3D gradient-echo images were acquiredat 1.5 T (TR=42 ms, TE=25 ms, alpha=25°, FoV=25.6 cm x 19.2 cm x 10.8 cm,matrix=512 x256x72). Additionally, EPI data were acquired for comparison (TR=4s,TE=50 ms, alpha=90°, FoV=25.6 cm x 25.6 cm, matrix=64x64, 45 slices,thickness = 4 mm). The 3D-phase was unwrapped and divided into large (L) vox-els, each consisting of several high resolution (H) voxels (e.g. 8x8x8=512 voxelsfor a simulated resolution of 4x4x4 mm). Each of the H-voxels within an L-voxelwas assigned a magnetization vector (M) with length 1 and precession frequency

Mo

nd

aM

on

da

Mo

nd

aM

on

da

Mo

nd

ayy yyy

Scientific Sessions

B 287C D E FA G

as defined by the local phase. The M-vectors were allowed to precess and thecomplex sum over all M was computed and assigned to L. Simulated imagesL (x, y, z) were computed for different imaging geometries (in-plane resolution,slice thickness, slice orientation), correction methods (z-shimming) and echo-times.Results: Simulated and measured EPI images showed very similar patterns ofsignal loss. Acquisition schemes could be optimized in terms of SNR and BOLDsensitivity= TE*Signal, based on a single phase map.Conclusion: The method is scalable to other field strengths and echo times thanthose used for the acquisition of the 3D phase-map. Signal behavior can be in-vestigated in the presence of static magnetic field inhomogeneities and as a func-tion of echo time and other acquisition parameters. This provides a tool for off-lineoptimization of scan parameters for, e.g., BOLD-fMRI.

B-652 11:06 !The sensitive and accurate de-convolution technique to correct for the"penumbral-spread" effect in CT, MR and DFK.M. Chui1, S.L. Chui2, D.B. Stanfield3; 1Uxbridge/UK, 2Redditch/UK,3St. Albans/UK ([email protected])

Purpose: Due to the sizes of the energy source and detector (s), a line-spread-function (LSF) or blurred edge-response-function (ERF) in the image domain isobtained from a Delta-function or Step-function in the object domain. The meanblurring in CT & MR images is 2.4 pixels. A patented de-convolution techniqueuses a Running filter to pin-point the true-edge position to sub-pixel accuracy.The original high-resolution feature of the edge-profile may be restored via anAlgorithm to correct empirically for the "penumbral-spread" under strict spatialbandwidth controls to avoid ambiguity.Methods and Materials: A New York Catphan 500 was used for CT, MR and DFcalibrations.Results: Distance measurements: mean accuracy to 1/50th of a pixel (dr/r) at 1%low-contrast edge. Area/volume measurements: mean accuracies to 1/30th and1/10th of a pixel (dr in 2prdr/pr2) in area and volume (volume=area*slice-thick-ness) at high and 1% low-contrast edges respectively. MTF at high-contrast edgeimproves from 5.83 lp/cm to 13.90 and 15.03 lp/cm at 10% normalised modula-tion level after 5x and 7x magnifications and enhancement. Low-contrast resolu-tion: detected 0.3% low-contrast to 2 mm rod at 300 mA, 120 kVp, 10 mm slice;random noise effect on the enhanced linear edge-profile was reduced as indicat-ed by da1 and da0 in Y=(a2 ± da2)X

2+(a1 ± da1)X+(a0 ± da0) from 57.50% to 12.40%and 1.47% to 0.38% respectively by a dynamic filter.Conclusion: The ERF-blurring/"penumbral-spread" cannot be overcome by cur-rent digital enhancement. This technique permits sensitive and accurate edge-profile detection and enhancement. Potential applications are in radiologicaldiagnosis, oncological response, microsurgery, and radiotherapy treatment plan-ning. Various field-trial examples are illustrated by slides.

B-653 11:15

Diffraction enhanced imaging: A new tool for cartilage and bone studiesP. Coan1, A. Wagner2, N. Sieber2, K. Schmuck2, J. Mollenhauer2,C. Muehleman3, M. Lohmann4, A. Bravin1; 1Grenoble/FR, 2Jena/DE,3Chicago 60612, IL/US, 4Hamburg/DE ([email protected])

Purpose: Early visualization of damage in cartilage disease is still a challengefor present clinical techniques (MR imaging, CT, and ultrasound). A new tech-nique allowing non-invasive 3D visualization has been successfully applied invitro on excised human joint specimens.Methods and Materials: Diffraction enhanced imaging (DEI) has been used forthis study. DEI can provide dramatic gains in contrast over conventional radiogra-phy by utilizing X-ray refraction and scatter rejection in addition to the absorptionof conventional technique. Pathological and healthy human ankles, toes, and fem-oral joints have been imaged in projection and computed tomographic DEI modeat high resolution (pixel size: 46 micron) and clinical compatible doses (< 10 mGy).Volume rendering and segmentation permitted to visualize cartilage from volu-metric CT-scans. Images of the same samples were compared with conventionalCT scanners and MR imaging, used as golden standard; histology provided thefinal diagnosis.Results: In conventional absorptive X-ray examination, cartilage was either in-visible in the 5 examined samples, or the cartilage height was not clearly defined.The thickness of the cartilage in histology matched data extracted from DEI im-ages. In one sample, MR imaging did not show the structural damage of carti-lage; in all other cases, the cartilage damage was always underestimated.Conclusion: DEI generates images that most closely resemble the appearanceof anatomical and histological tissue structures than conventional CT and MR

imaging examinations. DEI can provide an unequivocal non-invasive diagnosisof the state of disease of the joint and be considered a new tool in the osteoarthri-tis research.

B-654 11:24

Dear-mama a photon counting X-ray machineM. Chmeissani1, A. Diaz2, F. Kainberger3, M. Lozano1, J. Montagne4,M. Sentís5; 1Bellaterra/ES, 2Algete/ES, 3Vienna/AT, 4Paris/FR, 5Sabadell/ES([email protected])

Purpose: Dear-Mama is a project funded by EU FP5 programme for a period of4 years to develop digital imaging X-ray machine, for general radiology and mam-mography, that reduces the dose, and increases the image contrast.Methods and Materials: Based on direct capture of X-ray photons with roomtemperature solid-state pixel detectors coupled to photon counting readout elec-tronics, Medipix2 chip, via bump bonding. Medipix2 chip is a matrix of 256x256pixels (55 microns pixel size). The effective area of one Medipix2 assembly is14.08x14.08 mm. Medipix2 assemblies are mounted on small PCBs, later mountedon large PCB that we call the "Motherboard". Two different high voltage powersupplies and two different X-ray tubes will be used accordingly for each type ofclinical evaluation. The clinical trials will be carried out for mammography at UDI-AT (Barcelona), chest radiology at Hôpital d'Efant Armand Trousseau (Paris),and osteoporosis at radiology department of Vienna medical school.Results: Allows discrimination of signal after amplification and pulse shaping,from the intrinsic detector noise with the electronics. This leads to superb signalto noise and a high contrast image. This implies that with little X-ray dose one canachieve a very good image with practically zero electronic noise. One frame iscaptured and displayed in less than 10 seconds and then sent to the PACS. Thetotal active area of the sensor is 18x24 cm.Conclusion: In the presentation we will discuss the overall performance of theX-ray machine and the outlook for an industrialization plan.

B-655 11:33

Evaluation of the Medipix2 read-out system capabilities for applications inX-ray computed tomographyM. Quattrocchi1, R. Avitabile2, N. Belcari1, M.G. Bisogni1, C. Carpentieri1,E. Cicalini1, A. Del Guerra1, P. Delogu1, M.E. Fantacci1, G. Mettivier2,M.C. Montesi2, D. Panetta1, V. Rosso1, P. Russo2, A. Stefanini1; 1Pisa/IT,2Napoli/IT ([email protected])

Purpose: The aim of this work is to evaluate the capability of the acquisitionsystem based on the Medipix2 read-out chip for future application in computertomography (CT).Methods and Materials: The X-ray source has a W anode and a 30 micrometersfocus size. The imaging system is based on the Medipix2 read-out chip, bump-bonded to a 300 um and 700 um silicon pixel detector. The read-out chip geome-try is a matrix of 256 x 256 cells, 55 x 55 um2. The detection system operates insingle photon counting mode and is able to detect low contrast object (less than1%).Results: It has been initially investigated how to optimize the equalization proc-ess of a planar image for a uniform and a variable thickness phantom by varyingthe flat field acquisition conditions. Due to the possibility of the detection systemto apply a variable energy windowing, the contrast resolution as a function of theenergy has been studied. For a given thickness and material, the contrast resolu-tion has a minimum at a different energy.This device is now being tested in CT acquisition. The improvements of the pla-nar image quality should be reflected in a better quality of the reconstructed CTimages.Conclusion: The CT image quality will be investigated as a function of differentenergy windows. A comparison with existing CT systems will be made. In order toevaluate the system performances at low doses, a comparison between analyti-cal and iterative reconstruction algorithms will be also presented.

B-656 11:42

Expanded FOV cone beam reconstruction algorithm for PET-VCTJ. Hsieh, A.H. Lonn, M. Nyka; Waukesha, WI/US ([email protected])

Purpose: CT reconstruction FOV is 50 cm, which is much smaller than the 70 cmfor PET. When CT images are used to perform attenuation correction, significantartifacts result when scanned objects are outside 50 cm. Although H-L consist-ency is used to perform corrections for multi-slice CT, it cannot be applied to VCTdue to large cone angles.Methods and Materials: The total missing attenuation is first calculated based

Scientific Sessions

288 C D E FBA G

its continuity over views. The projection distribution is estimated using propertiesthat the projection profile does not change abruptly as a function of angle. Whencombined with the boundary projection values, the size and location of a watercylinder best fitting these parameters are obtained. To ensure the existence ofthe projection first derivative, the estimated projection is blended with the mirrorimage of the measured projection at the boundary.Results: For evaluation, various phantom experiments were conducted. Phan-toms include oval cylinder with angled Teflon rods, GE Performance, 48 cm poly,and shoulder phantom. They were positioned near the edge of 70 cm FOV. Bothaxial and helical scans were collected on a 64-slice scanner. Results have shownthat the proposed algorithm produces nearly perfect reconstruction for portionsof the object inside 50 cm, while the shape of the object outside are nicely re-stored.Conclusion: A new FOV expansion algorithm is proposed. Unlike previously pro-posed algorithms, it makes use of the continuity of the total attenuation and grad-ual change of the projection profile. Extensive phantom experiments havedemonstrated the efficacy of the proposed algorithm for PET-VCT applications.

B-657 11:51

Inter- and intra-radiograph reproducibility of bone mineral density (BMD)and metacarpal index (MCI) calculated by digital X-ray radiogrammetry(DXR)M.L. Schäfer1, J. Böttcher1, A. Pfeil1, A. Kramer1, A. Petrovitch1, A. Malich2,H.J. Mentzel1, W. Linss1, W.A. Kaiser1; 1Jena/DE, 2Nordhausen/DE([email protected])

Purpose: The potential of DXR for estimating cortical bone loss seems to bepromising, in particular because of the high precision. Aim of this study is toverify the inter- and intra-radiograph reproducibility of DXR under considerationof conventional and digital performed radiographs.Methods and Materials: Radiographs of the non-dominant hand from two de-ceased males were performed according to a predefined study protocol and un-der constant image capturing conditions (FFD 100 cm, 6 mAs, 42 kV). BMD andMCI were measured by Pronosco-X posure System (Version 2, Sectra, Sweden).For preparing conventional radiographs a Philips Super 80 CP System were usedwith films of Kodak T-MAT-Plus 200 and Agfa-Curix 200. Digital print-outs wereperformed using a Diagnost-Philips-Optimus and Siemens-Polydoros SX 80 Sys-tem. Inter- and intra-radiograph reproducibility of BMD and MCI were calculated.Results: The inter-radiograph reproducibility of BMD using conventional imagesunder standardised conditions showed data with CV=0.49% (Agfa) and CV=0.33%(Kodak), whereas reproducibility using digital images ranged from CV=0.57%(Philips) to CV=1.18% (Siemens). The intra-radiograph reproducibility was ob-served to be CV=0.13% and CV=0.26% (conventional, Kodak and Agfa) versusCV=0.27% and CV= 0.05% (digital Philips and Siemens).Conclusion: DXR shows a high inter- and intra-radiograph reproducibility forconventional as well as digital images and seems to be a reliable osteogeometri-cal technique for evaluation of the peripheral cortical bone which approve thequantification of minor disease-related alterations of the cortical bone partition.

10:30 - 12:00 Room P

Molecular Imaging

SS 1406PET/CT: New trendsModerators:C.D. Claussen; Tübingen/DEJ. Stojanovska; Skopje/MK

B-658 10:30

FDG-PET scanning correlates with tissue markers of poor prognosis andpredicts mortality for patients subjected to liver resection for colorectalmetastasesC.C. Riedl, T. Akhurst, S. Larson, S.F. Stanziale, M. Gonen, S. Tuorto,H. Hricak, Y. Fong, D.S. Klimstra; New York, NY/US ([email protected])

Purpose: To determine whether tumor uptake of FDG correlates with tumor cellexpression of Ki67, Glut1, P53 and patient survival in patients with metastaticcolorectal cancer.Methods and Materials: 90 patients with resectable metastatic colorectal can-cer in the liver were subjected to FDG-PET prior to hepatectomy. At surgery,tumors were harvested and prepared for assessment by histology and immuno-histochemistry. Expression of Ki67 (proliferation marker), Glut1 (glucose trans-

porter) and P53 (cell cycle control protein) were assessed by pathologists blind-ed to the PET results and the clinical outcome. Patients were followed for threeyears to determine survival.Results: The 53 patients with standard uptake values (SUV) less than 7 on PETwere compared to the 37 patients with SUV greater than 7. SUV correlated withGlut1 (p = 0.03), Ki67 (p = 0.026) and p53 overexpression (p = 0.02). The three-year survival rate was 42% in patients with SUV > 7 and 74% for those with SUV< 7(p = 0.007 by log-rank).Conclusion: In resectable colorectal metastases, the strong correlation of SUVwith selected tissue biomarkers suggests a biologic link between increasing glu-cose metabolism, rapid proliferation, and P53 mutation. FDG-PET is a functionalimaging technique that can predict prognosis in patients with metastatic colorec-tal cancer.

B-659 10:39

11C-metomidate PET of adrenocortical tumours in correlation withhistopathological findingsA. Sundin, J. Hennings, O. Lindhe, M. Bergstrom, B. Langstrom, P. Hellman;Uppsala/SE ([email protected])

Purpose: To evaluate 11C labelled metomidate (MTO) as a tracer for positron-emission tomography (PET) imaging and characterization of adrenal tumours incorrelation to histopathology.Methods and Materials: MTO is a PET tracer which accumulates in normal ad-renal cortex and adrenocortical tumours. MTO-PET examinations in 73 patients(mean age 56 years; 41 women) were analyzed and retrospectively compared tothe findings of 75 histopathological examinations (65 surgical specimens and 10biopsies). The histopathological diagnoses were adrenocortical adenoma (n = 26),adrenocortical cancer (n = 13), adrenocortical hyperplasia (n = 8), pheochromo-cytoma (n = 6), metastasis (n = 3) and tumours of non-adrenal origin (n = 19).Results: The sensitivity and specificity for MTO-PET in proving adrenocorticalorigin of the lesions were 89% and 96%, respectively. Non-adrenocortical lesionswere all MTO-negative. PET measurements using standardized uptake values(SUV) could distinguish adrenocortical tumours larger than 1 - 1.5 cm from nor-mal adrenocortical tissue. Five PET examinations proved false negative - twoprofusely necrotic adrenocortical cancers and three benign adrenocortical le-sions with a size (1 cm) near the resolution limit of the PET camera. One falsepositive result of distal metastasis to a groin lymph node was noted in a patientwith a leiomyosarcoma, which by contrast was MTO-negative.Conclusion: MTO-PET is a specific and sensitive method for diagnosing adren-ocortical tumours. MTO-PET is useful in the imaging work-up of adrenal inciden-talomas and may be beneficial for examination of patients with primaryaldosteronism. In adrenocortical cancer extensive necroses may produce falsenegative results which hamper the imaging results in these patients.

B-660 10:48

Accuracy of PET for detection of lymph node metastasisM.K. Jantsch, M.A. Saksena, M. Braschi, M.G. Harisinghani, R. Weissleder;Boston, MA/US ([email protected])

Purpose: PET scanning is increasingly been used to stage various primary ma-lignancies. The purpose of this study was to compare established CT size criteri-on, used to diagnose metastatic nodal disease, to FDG uptake within nodes inpatients undergoing combined PET/CT with 18 F-Fluorodeoxyglucose for stag-ing of primary malignancies.Methods and Materials: 37 patients (M:F; 10:27) with a mean age of 54 yearsundergoing combined PET-CT studies were included in the study. Pathologicalexamination of the detected nodes was performed after surgical dissection orbiopsy. Of the sampled nodes, 112 could be unequivocally correlated to patholo-gy and were included in the study. Sensitivity, specificity and accuracy were cal-culated for size criteria and FDG uptake.Results: 54 of 112 nodes were benign and 58 were malignant on pathology. Themean short axis diameter of malignant nodes was 10.5 mm (5- 41 mm) and formalignant nodes showing increased FDG uptake 14.1 mm (6-41 mm). Size crite-ria demonstrated a sensitivity of 62%, a specificity of 78% and accuracy of 69.9%.Quantitative evaluation of FDG uptake within the same nodes demonstrated asensitivity of 78%, specificity of 63% and accuracy of 70.3%. All false negativenodes on PET were less than 10 mm in short axis.Conclusion: PET is limited in detecting malignant lymph nodes smaller than7 mm. PET imaging demonstrates a lower specificity than size criteria on CT, asnon malignant conditions can also cause an increase in nodal FDG uptake.PET has a higher sensitivity than CT as PET is able to detect metastatic diseasein borderline sized lymph nodes.

Mo

nd

aM

on

da

Mo

nd

aM

on

da

Mo

nd

ayy yyy

Scientific Sessions

B 289C D E FA G

B-661 10:57

Whole-body MR imaging vs. whole-body PET/CT in staging of newlydiagnosed malignant melanoma: Initial resultsF.M. Vogt, P. Veit, R. Jablonka, J. Barkhausen, G. Antoch; Essen/DE([email protected])

Purpose: To assess the accuracy of whole-body MR imaging in comparison withwhole-body [18F]-2-Fluoro-2-deoxy-D-glucose (FDG) PET/CT for staging newlydiagnosed malignant melanoma of the skin.Methods and Materials: 31 patients underwent whole-body MR imaging andwhole-body PET/CT for staging of lymphnode (N) and distant (M) metastasesafter resection of the primary tumor. Evaluation was done according to the Amer-ican Joint Committee on Cancer Staging classification. MR images were evaluat-ed by two radiologists while PET/CT images were read by one radiologist andone nuclear physician. Histology and a mean clinical follow-up of 295 days servedas standards of reference. Differences between imaging procedures were testedby McNemar's test.Results: Overall NM stage was correctly determined in 21/31 patients (68%)with MR imaging and in 23/31 patients (74%) with PET/CT (p > 0.05). Assess-ment of N-stage revealed MR imaging to be accurate in 25/31 (81%) and PET/CTin 28/31 (90%) of the patients (p > 0.05). M stage was correctly diagnosed in 23/31 (74%) patients based on MR imaging in comparison with 24/31 (77%) usingPET/CT (p > 0.05). Both modalities showed high specificity values of 100% forN-staging and 95% for M-staging. However, sensitivity for N-staging was 50 %using PET/CT, while MR imaging showed a lower sensitivity of 33%. For M-stag-ing sensitivity values were even lower, PET/CT 37.5% and MR imaging 12.5%.Conclusion: A significant number of patients initially staged as N-negative andM-negative with whole-body MR imaging and PET/CT proved to have undetectedregional or distant metastases. Thus, short-term follow-up examinations either bywhole-body MR imaging or PET/CT seem to be mandatory.

B-662 11:06

Whole-body PET/CT-colonography: Technical feasibility and first clinicalexperiences of a new method in colorectal cancer stagingP. Veit, C. Kuehle, H. Kuehl, C. Herborn, H. Stergar, T. Beyer, A. Bockisch,J. Barkhausen, G. Antoch; Essen/DE ([email protected])

Purpose: Colorectal cancer is one of the leading cancers in western countries.To provide whole-body tumor TNM-staging in a single session, PET/CT protocolsfor whole-body tumor staging and CT-colonography were integrated into one ex-amination.Methods and Materials: 14 prospective patients underwent whole-body PET/CT after water-based bowel distension and pharmacological bowel relaxation.PET/CT data was acquired using oral and intravenous contrast agents. PET/CTdata sets were read by an experienced radiologist and a nuclear physician inconsensus, and histopathology from conventional colonoscopy as well as surgi-cal histopathology served as the standard of reference.Results: Integration of CT-colonography into whole-body PET/CT staging provedfeasible and was performed without technical errors in all patients. PET/CT-colonography revealed 18/19 colonic lesions detected by conventional colonos-copy, only one flat lesion was missed. Additionally, 6 extracolonic tumor siteswere detected with the integrated PET/CT exam. Based on the imaging findings11 patients were treated surgically, whereas three received chemotherapy. Forthe surgically treated patients PET/CT revealed correct T- stage in 8/11 patientsand a correct N-stage in 9 patients.Conclusion: Whole-body PET/CT with integrated CT-colonography is feasibleand provides accurate results for TNM staging in patients with colorectal cancer.Therefore it might be used as an "all-in-one" staging procedure, especially inpatients with incomplete colonoscopy.

B-663 11:15 !Evaluation of side-by-side reading of PET and MDCT images: A nuclearphysician and a radiologist's perspectiveH.M.E. Quarles van Ufford, P.F.G.M. van Waes, L.G.B.A. Quekel,P.M.J. de Haas, J.M.H. de Klerk; Amersfoort/NL ([email protected])

Purpose: Does PET-MDCT provide more accurate PET lesion localisation, andadditional clinical information to MDCT?Methods and Materials: 285 consecutive patients (164 males, 121 females; agerange 17-84 years; 153 lung cancer, 112 lymphoma, 20 miscellaneous) referredfor PET and separate recent MDCT were included. Patients were retrospectivelyclassified into 4 categories: (I) No pathological uptake on PET; (II) uptake on PETlocalises lesion; (III) side-by-side evaluation is sufficient to assess localisation of

lesions; (IV) side-by-side evaluation of PET and MDCT is not sufficient to assesslocalisation of lesions. For MDCT evaluation both benign and malignant lesionswere analysed and classified. An additional category was set up for unsuspectedlesions detected by PET.Results: In 58% lesion localisation was correctly assessed by either convention-al PET alone (Group II: 76/208) or side-by-side evaluation of PET and MDCT(Group III: 44/208). In 42% of the patients with pathological FDG-uptake (GroupIV), integrated PET-MDCT or software fusion might be beneficial. For 15% of thepatients clinical relevant additional information was provided by the MDCT (dis-seminated small pulmonary lesions, AAA > 5 cm, thrombi, pulmonary emboli). In7% of the patients unsuspected pathological lesions were detected by PET (ma-jority bone metastasis).Conclusion: MDCT not only improves localisation of PET lesions, but also addsrelevant clinical information. PET visualised for the radiologists unsuspected le-sions. PET-MDCT images improve diagnostic imaging and a synergy of radiolo-gists and nuclear physicians is becoming indispensable.

B-664 11:24

Recurrent breast cancer: Is there an advantage in re-staging with PET/CT?P. Veit, H. Stergar, T. Beyer, E. Hauth, A. Bockisch, J. Barkhausen, G. Antoch;Essen/DE ([email protected])

Purpose: To compare the value of combined PET/CT, PET+CT viewed side byside, CT alone and PET alone concerning the rTNM stage and influence on ther-apy in patients with recurrent breast cancer.Methods and Materials: 58 patients with suspicion of recurrent breast cancerunderwent whole-body [18 F]-2-fluoro-2-deoxy-D-glucose (FDG) PET/CT. Imag-es of combined PET/CT, PET+CT, PET alone and CT alone were evaluated byfour different reader teams and diagnostic accuracies concerning the rTNM stageand their influence on therapy were compared. Histology and a mean clinicalfollow-up of 456 days served as the standard of reference. Differences betweenthe staging procedures were tested for statistical significance by McNemar's test.Results: The overall TNM tumor stage was correctly determined in 93% withPET/CT, in 90% with PET+CT, in 79% with PET and in 84% patients with CTalone. PET/CT was more accurate in rTNM staging than PET alone (p < 0.05).No significant differences were detected compared to CT (p = 0.06) or PET+CT.PET/CT changed the therapy in 2 patients compared to PET+CT, in 8 patientscompared to PET (statistically significant difference, p < 0.05) and in 5 patientscompared to CT (no significant difference).Conclusion: Combined PET/CT proved to be more accurate assessing the rTNMthan PET alone with an impact on patient management in patients with recurrentbreast cancer. There was a trend to more accurate staging compared to CT alone,though this did not prove to be of statistical significance. Thus, further studiesevaluating larger patient populations are needed.

B-665 11:33 !18PET/CT has influence on therapeutic strategy in pancreatic lesionsL.J. Delrue, V. Casneuf, M. Kelles, G. De Meerleer, M. Peeters, P. Duyck;Ghent/BE ([email protected])

Purpose: To study the role of PET/CT in changing therapy in patients with prima-ry or recurrent pancreatic cancer.Methods and Materials: A retrospective analysis of prospectively collected18 FDG-PET/CT scans in 24 patients was performed. All patients underwent PET/CT for staging primary pancreatic tumours or for recurrent disease. All PET/CTscans were assessed by a radiologist in cooperation with a nuclear medicinephysician. Pancreatic lesions were interpreted based on visual assessment andby calculation of a semi-quantitative evaluation (SUV). These results were corre-lated with pathologic diagnosis and/or clinical course. In each patient therapeuticchange was assessed.Results: In the staging group (n = 13), PET/CT changed therapy in 6 patients(46%). One patient was correctly down-staged by PET/CT and underwent sur-gery with curative intent. Five patients were correctly up-staged (liver metastas-es) and surgery was cancelled. In 7 patients (54%) PET/CT didn't change therapy.In 2 patients PET/CT was unable to discriminate between pancreatitis and malig-nancy. Four patients underwent surgery because PET/CT showed operability,but vascular in-growth (n = 1), liver (n = 2) or peritoneal (n = 1) metastases ap-peared intra-operatively. In 1 patient PET/CT suggested liver metastasis but per-operatively no lesion was found.In the recurrence group (n = 11), PET/CT changed therapy in 10 patients (91%).PET/CT proved recurrence or distant metastases in 8 patients, where other im-aging modalities remained uncertain. In 2 patients, recurrence was excluded. In1 patient (9%) recurrence was missed, possibly because of hyperglycaemia(160 mg/dl) at the time of injection.

y)

Scientific Sessions

290 C D E FBA G

Conclusion: PET/CT is useful in directing therapy for primary or recurrent pan-creatic cancer. Overall, therapy changed in 67% of the cases.

B-666 11:42

Early detection of histologically proven peritoneal carcinomatosis with PET/CTA. Dirisamer1, M. Heinisch2, F. Wolf1, W. Langsteger2, W. Schima1; 1Vienna/AT,2Linz/AT ([email protected])

Purpose: Peritoneal carcinomatosis is a strong prognostic indicator in severalprimary malignancies. An early detection is of crucial importance regarding fur-ther treatment and patient outcome. In respect to recognized limitations in detec-tion of small lesions with CT and misinterpretation of normal physiologicalabdominal activity at FDG-PET, our aim was to assess if fused PET/CT offers thecombined benefits of anatomical and functional imaging.Methods and Materials: 40 patients with ovarian (n = 28) and colorectal (n = 12)cancer with suspected peritoneal carcinomatosis were reviewed for the pres-ence of peritoneal lesions on FDG-PET/CT scans (Discovery LS, GE MedicalSystems). The results were compared with the histological findings. 27 patientshad peritoneal metastasis while 13 patients had negative histological findings atlaparotomy.Results: FDG-PET was positive in 23/27 patients with 1 false positive (sensitiv-ity, 85%; specificity, 93%). CT was positive in 24 patients without false positiveresults (sensitivity, 89%, specificity, 100%). PET/CT was positive in 28 patientswith 1 false positive (sensitivity, 100%; specificity, 93%).Conclusion: Fused PET/CT improves the sensitivity for detection of peritonealcarcinomatosis, especially in small lesions, and offers exact anatomical informa-tion regarding surgical treatment.

TT TTTu

esd

au

esd

au

esd

au

esd

au

esd

ayy yyy

Scientific Sessions

B 291C D E FA G

TTTTTuesdauesdauesdauesdauesdayyyyy, Mar, Mar, Mar, Mar, Marccccch 7h 7h 7h 7h 7

B

Scientific Sessions

292 A C D E F G

EAR Award

JIISJunior ImageInterpretation

Session(p. 120)

ClosingCeremony

SS 1804Chest

Pulmonaryinfection andobstructivepulmonary

disease(p. 318)

SS 1803Cardiac

Coronary arterystents and by-

pass grafts(p. 324)


RadiologyPeripheralvascular



Spine(p. 310)

SS 1801aGI Tract

Imaging of theacute abdomen

(p. 312)

SS 1802Breast

Magneticresonance (2)

(p. 314)

SS 1801bGI Tract

Cross-sectionalimaging ofthe GI tract

(p. 316)


RadiologyEmbolization

(p. 320)

WS 1709Interventional

RadiologyPercutaneoustumor therapy

(p. 117)


Non traumaticvascular

emergencies(p. 112)


Knee joint(p. 112)

SF 17Special Focus

SessionNew develop-

ments in imageanalysis(p. 113)

RC 1702Breast

Imaging focusedon therapy

(p. 114)


Non-invasivevisceral

angiography inthe abdomen

(p. 115)

RC 1704Chest

Multislice CTof the thorax

(p. 115)

RC 1711Neuro

Orbit andvisual system

(p. 116)

SS 1807Genitourinary

Bladder/Urethra/Pelvic floor

(p. 322)

EP

OS

™ - scientific exhibition

registration

room B2nd level

room C2nd level

room E1entr. level

room E2entr. level

room F1entr. level

room F2entr. level

room Hlower level

room Glower level

08:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:00

14:00

14:30

15:00

15:30

16:00

16:30

17:00

17:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:30

14:00

14:30

15:00

15:30

16:00

16:30

17:00

12:3012:30

13:00

18:00

18:30

19:00

18:30

19:00

18:00

17:30

13:30

08:30

08:00

07:30

07:00

07:30

08:00

07:00

room A2nd level

TT TTTu

esd

au

esd

au

esd

au

esd

au

esd

ayy yyy

Scientific Sessions

B 293A C D E F G





WS 1724Virtual


"Hands-on"Workshop

SS 1811Neuro

Imaging cerebralblood vessels

(p. 326)

SS 1806aMolecularImaging

Near infraredfluorescence -

FDG-PET(p. 330)

SS 1806bContrast Media

Iodinated andUS agents

(p. 332)

SS 1815VascularPeripheralmagnetic

resonanceangiography

(p. 334)

WS 1824Virtual


"Hands-on"Workshop

RC 1715Vascular

Venous imagingand intervention

(p. 117)

RC 1712PediatricNeonatalproblems(p. 118)


Qualityassurance/

Professionalmatters(p. 119)

SS 1805Computer

ApplicationsAdvanced image

processing(p. 328)

E3 1720Foundation

Course:AbdominalRadiologyPerforationand trauma

(p. 119)

room Klower level

room L/M1st level

room N/O1st level

room Plower level

room Wbasement

room Y1st level

The Globe2nd level

08:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:00

14:00

14:30

15:00

15:30

16:00

16:30

17:00

17:30

09:00

09:30

10:00

10:30

11:00

11:30

12:00

13:30

14:00

14:30

15:00

15:30

16:00

16:30

17:00

12:3012:30

13:00

18:00

18:30

19:00

18:30

19:00

18:00

17:30

13:30

08:30

08:00

07:30

07:00

07:30

08:00

07:00

room Ilower level

room Xentr. level

room Z1st level

Scientific Sessions

294 C D E FBA G

10:30 - 12:00 Room B

Musculoskeletal

SS 1810Muscles and ligamentsModerators:S. Anderson; Berne/CHA.R. Mester; Budapest/HU

B-667 10:30

Are intramuscular injections into the buttocks truly intramuscular?V.O. Chan, J.A.C. Colville, T. Persaud, O. Buckley, S. Hamilton,W.C. Torreggiani; Dublin/IE ([email protected])

Purpose: To determine the percentage of patients who actually receive an intra-muscular dose in the gluteal region.Methods and Materials: Between October 2004 to March 2005, 50 patients wereprospectively recruited. They were already receiving intramuscular medicationand were scheduled for a CT abdomen and pelvis scan. There were near equalnumbers of both genders (age ranged from 21-87 years). Prior to the CT scan,the intramuscular medication and 1 mL of air was administered into the upperouter quadrant of the buttocks using a standard 23 gauge, 30 mm needle. The CTscans were analyzed by 2 independent radiologists establishing the location ofthe air bubble. Body mass index (BMI), gender preponderance, distance to injec-tion site, and thickness of subcutaneous fat and muscle were also calculated.Results: Only 12 (25%) received an intramuscular injection. There was not alinear relationship with BMI. However, most of the patients who received subcu-taneous injections were female, had an average fat thickness of more than 2 cmand an average muscle thickness of less than 3 cm. In contrast, those who re-ceived an intramuscular injection were all male, had an average fat thickness ofless than 2 cm and an average muscle thickness of more than 3 cm.Conclusion: 75% of patients did not actually receive an intramuscular dose inthe gluteal region. The efficacy of intramuscular injections is related to gender aswell as subcutaneous fat and muscle thickness. The results of this study mayhave an impact on the administration of intramuscular injections.

B-668 10:39

Analysis of pennation angles of muscles with high resolution ultrasound(HR-US) in track and field athletesL.M. Sconfienza, S. Parodi, M. Giglio, E. Silvestri, G. Garlaschi; Genova/IT([email protected])

Purpose: To evaluate the pennation angles of lower limb muscles (vastus latera-lis and medial head of gastrocnemius) in trained track and field athletes with HR-US to investigate whether the arrangement of muscle fascicles is related to differentmuscle activities (aerobic vs. anaerobic).Methods and Materials: Thirty-six track and field athletes trained in anaerobic(100-meters runners) and aerobic (marathon runners) muscle activities wereevaluated at the end of their training session. In each subject, HR-US evaluationof the vastus lateralis (VL) and the medial head of the gastrocnemius (MHG) wasperformed on both limbs by means of a high-resolution broadband linear arraytransducer (12-5 MHz). Muscle pennation angles were calculated on scanningplanes oriented along the long axis of the muscle.Results: Athletes performing anaerobic training showed greater fascicle angles(VL, mean 23°; MHG, mean 24°) than the aerobic trained ones (VL, mean 16°;MHG, mean 18°).Conclusion: HR-US evaluation of pennation angles is an easy technique toachieve information on the architecture of muscles. Our data indicate a signifi-cant difference of orientation of muscle fascicles between trained aerobic andanaerobic athletes.

B-669 10:48

Quantitative muscle perfusion imaging at 3 Tesla using a FAIR-trueFISPtechniqueA. Boss, P. Martirosian, C.D. Claussen, F. Schick; Tübingen/DE([email protected])

Purpose: Many chronic diseases such as diabetes mellitus type 2, systemic lu-pus erythematosus, and scleroderma are correlated with diminished skeletalmuscle perfusion. The aim of the present study was to test and demonstrate thefeasibility of quantitative muscle perfusion imaging on a clinical 3 Tesla whole-body scanner using a new FAIR-trueFISP arterial spin labeling approach.

Methods and Materials: Perfusion maps of the forearm musculature were ac-quired in eight healthy volunteers (24 to 42 years old) before and after intenseexercise. All measurements were performed in a 3.0 Tesla whole-body MR unit incombination with an 8-channel head coil. Pulsed arterial spin labeling and signaldetection were performed with an adapted FAIR-trueFISP technique. Quantita-tive perfusion maps were calculated on a pixel-by-pixel basis by means of theextended Bloch equations.Results: The acquired perfusion images with an in-plane resolution of 1 mmshowed no significant distortions or blurring. Perfusion time curves could be ac-quired with a temporal resolution of 6.4 seconds. Maximum perfusion in the mus-culature was found approximately two minutes after exercise, reaching values ofup to 220 ml/min/100 g tissue with good delineation between the active musclesand the musculature not involved in the exercise.Conclusion: The trueFISP pulsed arterial spin labeling technique allows for pa-tient-friendly assessment of muscular perfusion in a clinical whole-body scanner.

B-670 10:57

Diffusion tensor imaging and fiber tractography of the human calf:Comparison between 1.5 T and 3.0 T (preliminary experience)N. Saupe1, L.M. White1, M.S. Sussman1, M.D. Noseworthy2; 1Toronto, ON/CA,2Hamilton, ON/CA ([email protected])

Purpose: The aim of this study was to evaluate the imaging parameters neces-sary to perform DT-MRI and fibertractography at 1.5 T and 3 T in normal humanskeletal muscle using different b-values and to calculate and to determine a numberof quantitative parameters to characterize diffusion anisotropy in organized tis-sue.Methods and Materials: Five healthy adult volunteers were included in this study.All volunteers underwent MR imaging at 3 T and 1.5 T (GE Healthcare, Milwau-kee, WI) within five days. Imaging on both MR systems was performed utilizing astandard knee coil and using the same imaging protocol. Three routine diffusion-weighted spin-echo EPI pulse sequences (with different b values) were appliedto collect a series of 2D images through the calf muscles of each subject. Theimage analysis and tensor calculations were performed using the DTIStudiopro-gram. The eigenvalues (Λ1,Λ2,Λ3), the mean diffusivity ⟨D⟩, the fractional anisot-ropy (FA), the relative anisotropy (RA) and the volume ratio (VR) were calculated.Results: The mean values for (Λ1,Λ2,Λ3) and D were 1.47 ± 0.08, 1.30 ± 0.09,1.06 ± 0.08 and 1.86 ± 0.09 at 3 T and 1.77 ± 0.07, 1.90 ± 0.09, 1.83 ± 0.08 and2.23 ± 0.09 at 1.5 T. (All values are reported in 10-5 cm2·s-1; mean ± SD). Therewas no significant difference between calculated parameters at 3 T and 1.5 T(P < .001).Conclusion: The study demonstrates useful parameters to perform DT-MRI andfibertractography at 1.5 T and 3 T. DT-MRI at 1.5 and 3 T provides in vivo valida-tion of quantitative structural analysis of human skeletal muscle. Fiber tracking at3 T MR provides information about muscle fiber orientation and architecture ofthe diffusion pathways.

B-671 11:06

MR imaging in patients with peripheral neuropathies of the upper extremity:Spectrum of imaging findings, clinical correlation and impact on therapyG. Andreisek, D. Burg, A. Studer, V. Meyer, B. Marincek, D. Weishaupt;Zurich/CH ([email protected])

Purpose: To investigate the role of MR imaging in patients with peripheral neu-ropathies of the upper extremity in whom clinical evaluation (including electrodi-agnostic studies) revealed unclear or ambiguous findings and to demonstrate theimpact of MR imaging on therapy.Methods and Materials: Fifty patients with peripheral median, ulnar and/or radi-al neuropathies were referred to MR imaging at 1.5 T. In these patients, either theetiology (32/50) or the site and localization (18/50) of the lesion was unclear. Atotal of 50 MR imaging examinations were performed (1 shoulder/axilla, 2 upperarms, 26 elbows including the forearm, 21 wrists including the hand). Two radiol-ogists reviewed all MR images with regard to nerve, muscle, and other soft-tis-sues abnormalities. A panel of a radiologist, a hand surgeon and a neurologistreviewed all clinical records. The correlation of clinical and radiological findings,and the impact of MR imaging on therapy were assessed. Cohen's kappa andSpearman's (p) rank correlation test were calculated.Results: A total of 17 neurogenic tumors, 20 compressive nerve lesions, and 10cases of focal or multifocal neuritis were detected. In 13 of the 18 patients (72%),MR imaging demonstrated exactly the site and location of the neuropathy. MRimaging revealed the etiology in 27 of the 32 patients (84%). Cohen's kappa was0.73 to 1.0 and p was 0.78 to 0.94. The impact on therapy was "major" in 29,"remarkable" in 18, and "no impact" in 3 patients.

TT TTTu

esd

au

esd

au

esd

au

esd

au

esd

ayy yyy

Scientific Sessions

B 295C D E FA G

Conclusion: MR imaging in patients with peripheral neuropathies of the upperextremity is useful and has significant impact on therapy.

B-672 11:15

High resolution ultrasound of small motor and sensory peripheral nervesG. Bodner, B. Glodny, H. Gruber, K. Galiano; Innsbruck/AT([email protected])

Purpose: New soft ware and equipment have contributed to a marked qualityincrease of ultrasound imaging of the soft tissue. The aim of this cadaveric studywas to assess feasibility of detecting small sensory peripheral nerves using highfrequency probes.Methods and Materials: We performed ultrasound examinations on two pre-served cadaveric specimens (age 75/84 years). We used a 17 MHz broad bandlinear array with a silicon stand off pad. Small motor and sensory peripheral nerves(accessory nerve, intercostal nerve, superficial radial nerve, digital nerves, cuta-neous femoris lateralis, saphen nerve, superficial peroneal nerve, infrapatellarnerve, sural nerve) were sonographically detected and, under ultrasound guid-ance, a small amount of dye (0.05 ml) was injected close to each nerve and thearea was marked on the skin. Following this, anatomical dissection was performedat the area where the dye was injected.Results: All small motor and sensory peripheral nerves have been correctly lo-cated by ultrasound.Conclusion: High frequency ultrasound allows visualization of small motor andsensory peripheral nerves.

B-673 11:24 !Ultrasound-guided percutaneous needle therapy for epicondylitismanagementE. LLopis, V. Higueras, P. Ferrer, P. Aparisi; Valencia/ES([email protected])

Purpose: Lateral epicondylitis is an overuse syndrome, caused by repeated mi-crotrauma to the origin of the common extensor tendon, and the most frequentsoft tissue abnormality affecting the elbow joint. The purpose of this study is tointroduce and to evaluate the efficacy of a simple ultrasound-guided percutane-ous therapy technique.Methods and Materials: Thirty-seven patients (20 men, 17 women, mean age42 years) were treated with needle tenotomy and peritendinous steroid injectionunder US-guidance. All patients had previously failed conservative management,10 patients had undergone blind steroid injections. Preprocedural diagnostic sono-graphic evaluation was performed and only those with hypoechoic areas or calci-fications were included. After subcutaneous anesthesia, a 22G needle wasinserted towards the areas that contain scar tissue (hypoechoic and calcifiedareas) with local anesthesia, using the needle as a tiny surgical instrument withrepeated movements, ending with a peritendon steroid injection (40 mg triamci-nolone). The entire procedure was performed under sonographic guidance. Twopatients required repeated treatment.Results: Thirty-five patients had at least 6 months follow-up, 2 patients were lostto follow-up. Complete symptom relief was reported in 70% of patients, moderatein 19% and 5% failed, requiring surgical treatment. Side effects were: 1 local skinatrophy and 1 vasovagal reaction. No complications from bleeding or infectionwere observed. The average procedure time was 25 min.Conclusion: Percutaneous needle debridement with ultrasound guidance is asafe, cost-efficient procedure with high rates of therapeutic success.It is an alternative to blind steroid injections, and surgical options remain open.

B-674 11:33

Collateral ligaments of the ankle: Value of multiplanar reconstruction (MPR)of 2D turbo spin-echo (TSE) imagesS.R. Duc, B. Mengiardi, C.W.A. Pfirrmann, J. Hodler, M. Zanetti; Zürich/CH([email protected])

Purpose: To assess the value of multiplanar reconstructions (MPRs) of 2D turbospin-echo (TSE) images for the evaluation of the collateral ligaments of the an-kle.Methods and Materials: Coronal and axial T2-weighted TSE and MPRs of 15asymptomatic ankles were analyzed by two musculoskeletal radiologists. Overallvisibility, visualization of the entire ligament on one single image, signal intensityand delineation of the ankle ligaments were determined for TSE and MPRs. Read-ing time for TSE and MPRs was measured. Image quality of the MPRs was grad-ed and correlated with the reformation angle.Results: Visualization of the entire ligament on one single image was significant-

ly superior with MPRs when compared to the original 2D TSE images for thetibionavicular (Reader1: p = 0.002, Reader2: p = 0.001) and calcaneofibular(Reader1: p = 0.001, Reader2: p < 0.001) ligaments. The tibionavicular ligamentwas considered to be better delineated on MPRs by Reader1 (p = 0.031). Nostatistically significant difference was found for the evaluation of the remainingligaments and criteria. Image quality of the MPRs was considered to be identicalor minimally blurred when compared to the original TSE images in 98% of allevaluated images. Grading of image quality of the MPRs did not significantlycorrelate with reformation angle (p > 0.05, Spearman's rho). Reading times were237 (Reader1) and 168 seconds (Reader2) for the TSE images and 362 (Read-er1) and 249 seconds (Reader2) for the MPRs.Conclusion: MPR improves visualization of the collateral ankle ligaments withina reasonable time. The advantages are most pronounced for the tibionavicularand calcaneofibular ligaments with their oblique course.

B-675 11:42

Medial collateral ligament complex of the ankle: MR appearance inasymptomatic subjects on standard ankle MR examinationB. Mengiardi, C.W.A. Pfirrmann, J. Hodler, M. Zanetti; Zurich/CH([email protected])

Purpose: To characterize the MR appearance of the medial collateral ligament inasymptomatic subjects.Methods and Materials: Standard ankle MR examinations of 56 asymptomaticsubjects (23-60 years) were analyzed by two musculoskeletal radiologists. For allfive components of the medial collateral ligament complex the visibility, the opti-mal imaging plane and the signal characteristics were analyzed. The thickness ofall ligaments was measured.Results: The posterior tibiotalar and tibiospring ligament were visible in all asymp-tomatic subjects, the tibiocalcaneal ligament was seen in 49 (88%) subjects,whereas the anterior tibiotalar and tibionavicular ligaments were only visible in31 (55%) subjects. The coronal imaging plane was considered to be always thebest imaging plane for all components except for the tibionavicular ligament (55%coronal, 45% transverse). The anterior and posterior tibiotalar ligament had avariable signal on T2-weighted images and mean thickness of 1.4 mm and 8.2 mmrespectively. On T2-weighted images the posterior tibiotalar ligament had a typi-cally striated appearance in all subjects under 45 years (89% of the whole studypopulation). The tibionavicular ligament had a variable signal on T2-weightedimages and a mean thickness of 1.6 mm. The tibiospring and the tibiocalcanealligament had a low signal on T2-weighted images (75% and 78% respectively)and a mean thickness of 2.0 mm and 1.2 mm respectively.Conclusion: The posterior tibiotalar ligament was always visible with a typicallystriated appearance on T2-weighted images in all subjects under 45 years. Thetibiospring and tibiocalcaneal ligaments are the consistently visible superficialcomponents with a typically low signal on T2-weighted images.

B-676 11:51

Correlation between X-ray and MR imaging of ankle fractures with regard tosyndesmotic injuryJ.J. Hermans, N. Wentink, P.G.H. Mulder, A. Beumer; Rotterdam/NL([email protected])

Purpose: To validate three fracture classification systems with respect to syn-desmotic injury.Methods and Materials: In 33 patients X-rays of the ankle fractures were classi-fied according to Weber, AO-Müller and Lauge-Hansen. Both fracture type andadditional measurements of the ratio of the medial and superior clear space (MCS/SCS) and tibiofibular overlap (TFO) were used to assess syndesmotic injury. AnMCS/SCS > 1 mm and a TFO < 0 mm are indicative for syndesmotic injury. DualTSE images in three orthogonal planes and an additional oblique plane wereacquired with a 1.5 T magnet. The syndesmotic ligaments were classified as nor-mal, thickened, possibly or definitely ruptured. Syndesmotic injury was definedas a definite rupture or avulsion of the anterior tibiofibular ligament or a completerupture of the entire syndesmosis (anterior and posterior). The agreement score(κ) concerning the presence of syndesmotic injury between the X-ray and MRimaging was determined.Results: With Lauge-Hansen, syndesmotic injury could be assessed in all pa-tients. Using Weber resp AO-Müller in 30% resp 9% of patients syndesmoticinjury could not be assessed either because the fracture could not be classifiedand/or the measurements were normal. The agreement score (κ) between thethree fracture classifications and MR imaging with respect to syndesmotic injurywas: Weber (23/33) κ=1.00 (p = 0.000); AO-Müller (30/33) κ=0.68 (p = 0.000);Lauge-Hansen (33/33) κ=0.84 (p = 0.000).

Scientific Sessions

296 C D E FBA G

Conclusion: With Lauge-Hansen classification evaluation of syndesmotic injurycorrelates well with MR imaging findings. Therefore Lauge-Hansen is the methodof choice when assessing syndesmotic injury in ankle fractures. With MR imag-ing syndesmotic injury can be more accurately assessed than with X-ray.

10:30 - 12:00 Room C


SS 1801aDisorders of the biliary systemModerators:B.I. Choi; Seoul/KRA. Lewszuk; Bialystok/PL

B-677 10:30

Contrast-enhanced multidetector-CT cholangiography after living donorliver transplantationT. Schroeder1, A. Radtke1, M. Malagó1, J.F. Debatin2, S. Kinner1, S.G. Ruehm3;1Essen/DE, 2Hamburg/DE, 3Los Angeles, CA/US([email protected])

Purpose: To evaluate the feasibility and efficacy of contrast enhanced multide-tector-CT cholangiography (ceMDCT-CA) for the postoperative assessment ofbiliary complications in living liver donors and living donated liver transplant re-cipients.Methods and Materials: 15 living donors of and 11 recipients of a right hepaticlobe underwent meglumine-iotraxate (Biliscopin®) enhanced MDCT cholangiog-raphy. Six donors were referred to exclude biliary leakage. Nine donors and all 11recipients were examined to exclude postoperative biliary obstruction. The ex-amination protocol included the intravenous short-infusion of 100 ml of a biliarycontrast agent. CT cholangiography data was acquired with a slice thickness of1 mm. Data sets were evaluated quantitatively by measurement of the biliaryopacification, and qualitatively on the basis of a scale ranging from 1 (non-diag-nostic) to 4 (excellent). The opacification was correlated with the post-operativeserum bilirubin level.Results: CeMDCT-CA provided diagnostic delineation of the biliary tree in all 15donors and seven of 11 recipients; in four recipients the degree of biliary opacifi-cation was rated non-diagnostic. The biliary opacification was generally higher inthe donor collective. Four donors and three recipients presented a moderate fo-cal biliary constriction without elevation of laboratory values indicating biliaryobstruction. Six patients presented with postoperative fluid collections sugges-tive of perihepatic biloma, however no biliary fistula could be visualized.Conclusion: CeMDCT-CA allows for non-invasive assessment of postoperativebiliary anatomy and complications in living liver donors and in the majority of livertransplant recipients. Direct visualization of biliary leakage appears to be limitedto fistulae of bigger size.

B-678 10:39

Preoperative T-staging of the carcinoma of the gallbladder using MDCT:Radiologic-pathologic correlationS. Kim, J. Lee, J. Lee, S. Kim, J. Han, B. Choi; Seoul/KR([email protected])

Purpose: To evaluate the performance of MDCT in the preoperative T- staging ofgallbladder cancer.Methods and Materials: Two abdominal radiologists retrospectively reviewedaxial images of two phase MDCT scans (3-mm collimation and 3-mm reconstruc-tion) of 83 patients who had undergone surgical resection or biopsy for gallblad-der cancers; and then reviewed axial and multiplanar reconstruction (MPR) imagesin 31 patients who had MPR images by consensus. The local spread of the dis-ease was evaluated according to the TNM system, and the results were correlat-ed to the pathologic findings. The statistical analysis was performed using Fisher'sexact test. Results of the consensus interpretations were used to calculate sensi-tivity, specificity, and accuracy of MDCT.Results: The sensitivities of the consensus interpretations of T1, T2, T3, and T4lesions were 75%, 77.8%, 81.5%, and 100%; specificities were 98.4%, 94.6%,80.4%, and 98.7%, respectively (P <.05). Overall accuracy was 80.7%. In the 31patients with MPR images, the combined reading of the axial and MPR imagesincreased diagnostic accuracy of T-staging compared with axial image only. Thesensitivities of the consensus interpretations of T1, T2, T3, and T4 lesions onMPR images were 33.3%, 83.3%, 100%, and 100%; specificities were 100%,79.6%, 89.5%, and 100%, respectively (P <.05).

Conclusion: MDCT provided 80.7% accuracy in the diagnosis of the local extentof carcinomas of the gallbladder, showing acceptable sensitivity and specificity.Addition of the MPR images to axial CT data was helpful for the differentiation ofT3 from T2 lesion.

B-679 10:48

Homogeneous signal intensity of bile in the gallbladder: A new sign fordiagnosis of chronic cholecystitis on MR imagingA.E. Mahfouz, H. Sherif; Doha/QA ([email protected])

Purpose: To evaluate difference in signal intensity of the most and least depend-ent bile within the gallbladder on MR imaging as sign of chronic cholecystitis.Methods and Materials: MR images of 40 patients with pathologically provenchronic cholecystitis (28 calcular and 12 non-calcular) and 50 patients with nohistory of gallbladder disease and normal ultrasonography were reviewed. Pa-tients were examined by identical MR imaging protocol in supine position afterfasting for four hours. The protocol included transverse T2-weighted half-Fourier(HASTE) images (TR/TE 1100 ms/87 ms), T1-weighted GRE images (TR/TE/flipangle of 170 ms/4.8 ms/70º) before and after fat saturation. The signal intensityof bile was measured at the most and least dependent parts of bile within thegallbladder.Results: Ratio between signal intensity of the most and least dependent bilewas1.8 ± 0.46 on fat-saturated T1-weighted images and 0.64 ± 0.07 on T2-weight-ed images in patients with normal gall bladder versus 1.06 ± 0.07 and 0.96 ± 0.08on the respective pulse sequences in patients with chronic cholecystitis. Differ-ence between the two groups was statistically significant for both sequences(p <.001). Setting the ratio threshold of 1.15 on fat-saturated T1-weighted imag-es, the sign is diagnostic of chronic cholecystitis with positive predictive value of88%, negative predictive value of 94%, specificity of 93%, sensitivity of 90%, andaccuracy of 91%.Conclusion: Absence of difference in signal intensity between most and leastdependent parts of bile within the gallbladder may be useful diagnostic sign ofchronic cholecystitis. The sign may reflect poor concentration of bile by the dis-eased gallbladder wall.

B-680 10:57

Choledocholithiasis: Repetitive thick-slab single-shot projection magneticresonance cholangiopancreatography (MRCP) versus endoscopicultrasonography (EUS)S. Schmidt1, E. Gelsie2, F. Demarquay2, P. Schnyder1, P. Chevallier2;1Lausanne/CH, 2Nice/FR

Purpose: Prospective comparison of repetitive thick-slab single-shot projectionmagnetic resonance cholangiopancreatography (MRCP) with endoscopic ultra-sonography (EUS) for the detection of choledocholithiasis.Methods and Materials: Fifty-seven consecutive patients (36 women, 21 men,mean age 61) referred for suspected choledocholithiasis underwent MRCP, fol-lowed by EUS within 46 hours. Each procedure was performed by different oper-ators unaware of results of the other investigation. MR technique included a turbospin-echo T2-weighted axial sequence with selective fat saturation and respira-tory compensation (SPIR/TSE, TE=70 msec, TR=1600 msecs), followed by twelvecoronal dynamic slices of MRCP acquired as respiratory-triggered 40 mm thick-slab single shot repetitive projection imaging centered on the common bile duct,especially the sphincterian complex (SSh, TE=900 msec, TE=8000 msec).Results: Common bile duct stones were demonstrated in 18 out of 57 patients(31.6%), confirmed by endoscopic retrograde cholangiography (n = 17) or intra-operative cholangiography (n = 1). Clinical follow-up served as gold standard inpatients with negative results and without following invasive procedure (n = 28).Sensitivity, specificity, accuracy, positive and negative predictive value for MRCPwere 94.9%, 94.4%, 94.7%, 97.4% and 89.5%, respectively. Corresponding val-ues of EUS were 97.4%, 94.4%, 96.5%, 97.4% and 94.4%, not significantly dif-ferent from MRCP results.Conclusion: Repetitive thick-slab single-shot projection MRCP is an accuratenoninvasive imaging modality for suspected choledocholithiasis and should beused increasingly to select those patients who require a subsequent therapeuticprocedure, namely ERCP.

B-681 11:06 !Area and caliber of the bile ducts and main pancreatic duct in maturesubjects using magnetic resonance cholangiopancreatographyY. Kitagawa; Obu/JP ([email protected])

Purpose: Magnetic resonance cholangiography (MRC) provides 3D images andcross sections of viscus, however, limited studies have been focused on the actu-

TT TTTu

esd

au

esd

au

esd

au

esd

au

esd

ayy yyy

Scientific Sessions

B 297C D E FA G

al size of the duct. Aim of this study is to elucidate the standard size of the bileduct in mature population (age 45+) using this modality.Methods and Materials: In 90 cases with normal bile duct, area and caliber ofthe bile duct were prospectively measured by 1.5-T MR system using 3D T2image of HASTE sequence. The provided image voxel was 1.0x0.9x1.0 mm. Thefollowing sites were measured; upper boarder of pancreas (A), common hepaticduct (B), left hepatic duct (C) and right hepatic duct (D). Among the 90 subjects,10 cases were measured for main pancreatic duct using the same procedures.The pancreatic duct was measured in head (E) and neck (F) of the pancreas.Results: The measured areas (mm2) were 40.4 ± 20.1 in A, 43.2 ± 24.0 in B,27.9 ± 14.9 in C and 27.4 ± 16.0 in D, respectively. The diameters (mm) were6.9 ± 1.9 in A, 7.1 ± 2.2 in B, 5.7 ± 1.6 in C and 5.6 ± 1.8 in D, respectively. Therewas no correlation between age and size of the bile ducts. As for the pancreaticduct, the areas were 10.8 ± 2.9 in E and 8.8 ± 2.8 in F. The calibers were 4.4 ± 0.8in E and 3.8 ± 0.7 in F.Conclusion: MRC provided useful parameters to define standard size of onlybile duct. According to this results, upper limit of normal value of intrahepatic bileducts were 8.9 mm (left) and 9.2 mm (right).

B-682 11:15

Differentiation of malignant from benign bile duct stricture using MRCP,conventional MR and Gd-BOPTA-enhanced 3D dynamic imaging: Results ofunivariate and multivariate statistical analysisj. Kim, J. Lee, S. Kim, J. Lee; Seoul/KR

Purpose: To retrospectively evaluate criteria for differentiating malignant frombenign stricture in magnetic resonance cholangiopancreaticography (MRCP),conventional T1- and T2-weighted images and gadolinium-dimeglumine (Gd-BOPTA)-enhanced three-dimensional dynamic imaging.Methods and Materials: In last three years, 37 patients (19 men and 18 women;mean age, 60.8 years) with biliary stricture in MR were identified from our radio-logic and surgical database. All 18 patients with cholangiocarcinoma had patho-logic proof after surgery. 19 patients with benign stricture were diagnosed bystrict clinical criteria. Two radiologists analyzed MR features by consensus forlocation, wall thickness, length, asymmetricity, luminal irregularity, gradual or abruptnarrowing, outer margin, signal intensity at T2WI and hyperenhancement relativeto liver parenchyma during portal phase at Gd-BOPTA-enhanced dynamic T1WI.MR findings relevant as predictors were identified with Chi-square or Fisher'sexact test. Multivariate stepwise logistic regression analysis was used to deter-mine the most predictable factors.Results: Malignant stricture of bile duct were significantly thicker (6.7 ± 2.8 mm)(P < 0.0001) and longer (27.7 ± 8.1 mm) (P = 0.003) than benign stricture. Hy-per-enhancement relative to liver was more frequently seen in malignant cas-es (17/18, 94%) (P < 0.0001). Luminal irregularity, asymmetricity, indistinct outermargin and high signal intensity at T2WI image relative to liver were significantfactors for malignant stricture (P < 0.005). Multivariate stepwise logistic regres-sion analysis showed that hyper-enhancement relative to liver parenchyma andluminal irregularity were the only two variables that independently differentiatedmalignant from benign strictures.Conclusion: Hyper-enhancement during portal phase at Gd-BOPTA-enhanceddynamic T1-weighted image and luminal irregularity of involved bile duct stronglysuggest cholangiocarcinoma.

B-683 11:24

Biliary tract complications following orthotopic liver transplantation:Evaluation with mangafodipir trisodium (Mn-DPDP)-enhanced MRcholangiography (MRC)P. Boraschi, F. Donati, R. Gigoni, A. Volpi, L. Urbani, F. Filipponi, F. Falaschi,C. Bartolozzi; Pisa/IT ([email protected])

Purpose: To determine whether Mn-DPDP-enhanced MRC could improve thediagnostic confidence provided by conventional T2w MRC in patients with sus-pected biliary complications after liver transplantation.Methods and Materials: Thirty liver transplant subjects with suspected biliarycomplications (abnormal liver function tests and/or biliary ductal dilation at ultra-sound) underwent MRC at 1.5 T (Signa, GE Healthcare). After acquisition of ax-ial T1w and T2w sequences, conventional MRC was performed through arespiratory-triggered, thin-slab (2 mm), heavily T2w FSE sequence and a breath-hold, thick-slab (10/50 mm), single-shot T2w sequence in the coronal plane. Ineach patient coronal T1w fat-suppressed breath-hold 3D gradient-echo imaging(FAME) was also obtained 30-40 minutes after mangafodipir trisodium (Mn-DPDP,Teslascan®) intravenous administration. All source images and 3D MIP and vol-ume rendered reconstructions were reviewed in conference by two observers

that were asked to determine presence and site of biliary tract complications.Imaging results were correlated with direct cholangiography (n = 21), surgery(n = 4) and/or follow-up (n = 5).Results: In 5 patients no biliary tract abnormality was identified. The remaining25 subjects were affected by ischemic-type biliary lesions (n = 6), non-anasto-motic strictures (n = 6), anastomotic strictures (n = 8), inflammatory ampullarystenosis (n = 2), intrahepatic biloma (n = 1) and stones (n = 11). Mn-DPDP-en-hanced MRC tended to out-perform conventional T2w MRC in delineating 3 slightanastomotic stenoses and 8 out of 12 strictures involving the biliary tree of thegraft. Additionally, it allowed the correct diagnosis of an intrahepatic biloma.Conclusion: Mn-DPDP-enhanced MRC may improve the diagnostic perform-ance of conventional T2w MRC in the evaluation of biliary complications in livertransplant recipients.

B-684 11:33

MR cholangiography at 3 Tesla in patients post cholecystectomy: Dosusceptibility artifacts from surgical clips impair image quality?E.M. Merkle, B.M. Dale, J. Thomas, E.K. Paulson; Durham, NC/US([email protected])

Purpose: To evaluate whether surgical clips from prior cholecystectomy impairimage quality during MR cholangiography (MRC) at 3 T.Methods and Materials: Seven surgical clips were embedded in a gel phantomand positioned at predefined distances from a fluid-filled tube designed to simu-late the common bile duct. In-vitro MR imaging protocol included T1-w GRE,RARE, HASTE, and 3D TSE T2-w sequences on a 1.5 T and a 3 T MR system.Within each magnet class, the maximum clip distance was noted where suscep-tibility artifacts obscured the fluid-filled tube. Susceptibility artifact size was alsocalculated for each sequence within each magnet class. In-vivo analysis includ-ed 21 patients with a history of cholecystectomy who underwent MRC at 3 T.Presence of susceptibility artifacts was noted for each sequence. MRC imagequality was also graded with specific emphasis on whether image quality wasimpaired by susceptibility artifacts.Results: In-vitro, mean area of susceptibility artifacts was 104 mm2 on the 3 TMR system and 75 mm2 on the 1.5 T MR system. While surgical clips within a2 mm range impaired visualization of the fluid-filled tube on 1.5 T imaging, thisrange increased to 4 mm on 3 T MR imaging. In-vivo, MRC image quality wasimpaired by susceptibility artifacts in three of 21 cases by obscuring adjacentbiliary ductal branches.Conclusion: Susceptibility artifacts from surgical clips are larger with ultra-highfield 3 T MR imaging than on standard 1.5 T MR imaging. Adjacent biliary ductalbranches will be obscured by these susceptibility artifacts in selected cases.

B-685 11:42

Performance and value of adding dynamic MR imaging to MRcholangiopancreatography for the evaluation of tumor extent of hilarcholangiocarcinoma and CBD cancerH.H.J. Kim, J.J.M. Lee; Seoul/KR (jung @radiol.snu.ac.kr)

Purpose: To determine the value and performance of dynamic MR imaging as asupplement to MRCP imaging for the evaluation of tumor extent of hilar cholangi-ocarcinoma and CBD cancer.Methods and Materials: Twenty four patients with Klatskin tumor or CBD cancerwho had undergone operation were included in this study. MR studies includedT1- and T2-weighted images, as well as gadolinium-enhanced dynamic imagesand MRCP images. Two radiologist reviewed MRCP images with source imagesand MRCP images with non-enhanced T1- and T2-weighted images and gadolin-ium-enhanced dynamic images. At the all steps, the readers evaluated longitudi-nal extent of the tumor according to the Bismuth-Corlette type and anatomiclocation and used a five-point scale to assign a confidence level to their evalua-tion of involving second branch of both IHD and distal CBD. At the second steps,gadolinium-enhanced dynamic images, the possibility of unresectability wereadditionally evaluated.Results: The accuracy of longitudinal extent of the tumor was increased at MRCPwith enhanced dynamic images (87.5% for MRCP with source images, and 95.8%for MRCP with dynamic images). For the evaluation of involving second branch ofboth IHD and distal CBD the area under receiver operation characteristic curvewas not significant difference between MRCP images interpreted with dynamicimages (0.987 for reader 1, 0.926 for reader 2) and MRCP images alone (0.926for reader1, 0.905 for reader 2).Conclusion: Use of enhanced dynamic images with MRCP images significantlyimproved the diagnostic accuracy of MR examinations of longitudinal extent ofthe Klatskin tumor and CBD cancer.

Scientific Sessions

298 C D E FBA G

B-686 11:51

Value of MR imaging VIBE sequence in preoperative evaluation of hilarcholangiocarcinomaB. Song, L. Yin, Y. Wu, J. Xu; Chengdu/CN ([email protected])

Purpose: To investigate the diagnostic value of a spoiled gradient-echo T1-weight-ed, isotropic 3D MR imaging sequence (volumetric interpolated breath-hold ex-amination, VIBE) in the preoperative evaluation of patients with hilarcholangiocarcinoma.Methods and Materials: 27 surgically confirmed patients with hilar cholangi-ocarcinoma (11 infiltrating type and 16 nodular type) underwent preoperative MRexamination of upper abdomen with a 1.5 T MR unit. Imaging sequences includ-ing FSE T2W, 2D GE T1W plain scan, 2D MRCP, Gadolinium-enhanced VIBE(TR/TE=4.2/1.8 msec, 12° flip angle, interpolation in three directions, pixel size≤ 2.0 mm in all dimensions) tri-phasic acquisitions (early, late arterial and portalvenous phases), followed by 2D GE T1W scanning at equilibrium phase with6 mm slice thickness. Imaging data were interpreted by 2 abdominal radiologiststhrough consensus reading.Results: VIBE alone was correct in classifying the morphological types in 26patients (96.3%, 11 infiltrating and 16 nodular) while combined T2W, 2D T1Wplain and enhanced imaging were correct in 17 patients (63%, 7 infiltrating and10 nodular) and MRCP alone in 14 patients (51.9%, 3 infiltrating, 11 nodular).VIBE was also statistically more accurate in depicting the location, size, extentand relationship with neighboring structures of hilar cholangiocarcinoma thanother MR sequences (ρ < 0.05).Conclusion: VIBE is a valuable MR imaging sequence for the preoperative as-sessment of hilar cholangiocarcinoma and should be included into the MR imag-ing protocol for such patients.

10:30 - 12:00 Room E2

GI Tract

SS 1801bFunction and motility disordersModerators:L. Dagdilelis; Kalamaria/GRP. Pokieser; Vienna/AT

B-687 10:30

Comparison of dynamic MR swallowing with esophageal manometry and24h-pH-monitoring in dysphagic patients: First experienceC. Kulinna-Cosentini, W. Schima, J. Lenglinger, G. Bischof, A. Ba-Ssalamah,E.P. Cosentini; Vienna/AT ([email protected])

Purpose: To evaluate the accuracy of dynamic MR swallowing of dysphagic pa-tients compared to esophageal manometry and 24h-pH-monitoring as gold stand-ard.Methods and Materials: 27 patients (12 m, 15 f; mean age 47 a) with dysphagiaunderwent MR swallowing, manometry and 24h-ph-monitoring of the esophagus.MR imaging was performed in supine position with dynamic gradient echo se-quences (B-FFE) in three planes in a 1.5 T unit, using buttermilk spiked withgadolinium chelate (40:1) as oral contrast agent. At MR imaging, esophagealbolus transit, gastro-esophageal junction and the presence of reflux during Val-salva maneuver were assessed. 24-hour ambulatory esophageal pH-monitoringand manometry with a pneumohydraulic low-compliance system were performedat a median 24 d interval (range 1-32 d) to MR swallowing. Manometric, ph-mon-itoring and MR imaging recordings were analyzed without information from theother examination.Results: All 27 patients (100%) had abnormal findings in manometry/ph-moni-toring. 23 patients (85%) showed gastroesophageal reflux disease (GERD) andin 4 patients (15%) a motility disorder was detected. The accuracy of MR imagingin detecting GERD was 78% (18 of 23 patients). MR imaging detected 75% (3 of4 patients) of motility disorders diagnosed by manometry. The overall accuracy ofMR swallowing to make a correct diagnosis was 78%.Conclusion: Dynamic swallowing MR imaging is considered to be a useful com-plementary tool for noninvasive diagnosis of GERD or motility disorders and war-rants further investigations.

B-688 10:39

Swallowing study of gastroesophageal junction with dynamic MR imaging:Preliminary results in healthy volunteersC. Kulinna-Cosentini, W. Schima, E.P. Cosentini; Vienna/AT([email protected])

Purpose: To evaluate the feasibility of dynamic MR imaging for swallowing studyof gastroesophageal junction and to assess normal function in healthy volun-teers.Methods and Materials: 20 healthy volunteers (mean age 34, 10f:10m) under-went MR swallowing in supine position with dynamic gradient echo sequences(B-FFE) in three planes at 1.5 T, using a phased-array body coil. As oral contrastagent buttermilk spiked with gadolinium chelate (40:1) for bolus passage wasused. We evaluated visualization of esophageal bolus transit, bolus transit time,peristalsis, identification of gastroesophageal junction and Valsalva manoeuvreto provoke potential reflux.Results: Sagittal plane was superior in evaluation of bolus transit, bolus transittime, and potential reflux. Mean esophageal transit time was 8 s. Mean length ofvisualized esophagus in sagittal plane was 16.2 cm. Axial and coronal view sur-passed sagittal view in identification of gastro-esophageal junction. 5 volunteersshowed abnormalities like gastroesophageal reflux, axial hernias or non-peri-staltic contractions.Conclusion: Dynamic swallowing MR imaging is simple, rapid to perform andwell tolerated. It warrants further investigations in patients with gastroesopha-geal reflux or motility disorders.

B-689 10:48 !MR fluoroscopy in the evaluation of esophageal motility disorders: Acomparison with barium fluoroscopy and manometry findingsV. Panebianco, M. Anzidei, E. Tomei, F.I. Habib, E. Corazziari, R. Passariello;Rome/IT ([email protected])

Purpose: To evaluate subjects affected by esophageal motility disorders usingDynamic TURBO-FLASH sequence during positive oral contrast agent adminis-tration and to compare MR findings with videofluoroscopy and manometry re-sults.Methods and Materials: Thirty patients who previously underwent videofluoros-copy and manometry for suspected esophageal motility disorders were studiedusing a 1.5 T magnet (Magnetom Vision: Siemens, Erlangen, Germany) equippedwith phased-array coil. Imaging protocol was based on the employment of Dy-namic TURBO-FLASH sequence on sagittal and axial planes to achieve motilityevaluation (TR 416 mS; TE 1.2 mS; Flip Angle 8°; Thickness 20 mm; FoV 350;Matrix 90x128; N. acquisition 45; TA=25 sec; Position: prone) during oral admin-istration of positive contrast agent (yoghurt + Gd-DTPA 0.5 M, 1:100 boluses).Images were reviewed by two radiologists with long experience in reporting bar-ium fluoroscopy. MR imaging findings were compared with manometry and vide-ofluoroscopy results.Results: Good quality images were obtained in all patients, with adequate lumencontrast. Frame rate was 4 fps, allowing a near real-time evaluation of esopha-geal motility. Comparing with videofluoroscopy and manometry, achalasia wascorrectly diagnosed in ten cases, "corkscrew esophagus" in six and diffuse es-ophageal spasm in nine; in three cases esophageal spasm was incorrectly diag-nosed as reflux; in two cases there was were evidence of normal findings.Conclusion: MR fluoroscopy approach represents a promising radiation-freemodality in the evaluation of esophageal motility disorders. Exists elevate agree-ment between MR imaging findings and standard modality findings.

B-690 10:57 !Magnetic resonance imaging (MRI) of swallowing process: Real-timetrueFISP versus cine turboFLASH sequences in anatomical and functionalevaluationI. Sansoni, M. Sammarra, R.M. Del Vescovo, P. Cimini, F.M. Martina, P. Simoni,R.F. Grasso, B. Beomonte Zobel; Rome/IT ([email protected])

Purpose: To evaluate the ability of different real-time available high-speed MRsequences (trueFISP, cine turboFLASH) to dynamically image the swallowingprocess. To determine which sequence provided the best dynamic informationand anatomical resolution of the oropharyngeal soft tissues involved in each ofthe swallowing phases.Methods and Materials: Ten healthy volunteers with no history of dysphagiawere studied on a 1.5 T scanner (Symphony, Siemens) equipped with high-per-formance gradients (amplitude 40 mT/m). All examinations were performed insupine position, obtaining sagittal and coronal views during swallowing semolina

TT TTTu

esd

au

esd

au

esd

au

esd

au

esd

ayy yyy

Scientific Sessions

B 299C D E FA G

pudding with/without paramagnetic contrast medium. Thirty serial images wereobtained using four real-time trueFISP (AcqT/slice 0.6 s) and three cine turbo-FLASH (AcqT/slice 0.2 s) sequences with varying parameters (field-of-view,matrix, slice thickness, etc) to assess the best spatial and temporal resolution.Functional and anatomical information obtained with the different acquisitionsand the clarity of swallowing-related motion were evaluated. Additionally, signal-to-noise ratios (SNRs) were measured in each used sequence.Results: Among turboFLASH sequences, the 5 mm acquisition showed the bestcompromise between SNR and spatial resolution, whereas 8 mm sequence re-vealed to be the optimal trueFISP acquisition. Considering the most performingtrueFISP and turboFLASH sequences, the last provided the best temporal andcontrast resolution, with a sufficient spatial resolution during motion.Conclusion: High-speed kinetic MR imaging is a very promising tool for evaluat-ing deglutition. Real-time trueFISP is well suited for this purpose, providing goodimage contrast and sufficient temporal and spatial resolution, but turboFLASHacquisition currently provides the best anatomical and functional information forthe evaluation of the swallowing process.

B-691 11:06 !Digital cineradiographic study of swallowing in patients with amyotrophiclateral sclerosisG. Lo Re, M. Galia, I. Spoto, G. Runza, T. D'Agostino, V. Lo Greco,T. Bartolotta, M. Midiri; Palermo/IT ([email protected])

Purpose: To evaluate the usefulness of digital cineradiography in detecting swal-lowing disorders of patients affected by Amyotrophic Lateral Sclerosis (ALS) inorder to plan an adequate therapeutic approach.Methods and Materials: From March 2004 to July 2005, 20 patients (7 malesand 13 females; mean age 35) affected by ALS were evaluated with digital cine-radiography. The main aim of the study was the presence of dysphagic disorderswith weight loss. All the examinations were performed with radio-controlled equip-ment provided with a digital C arm.Results: Using cineradiographic technique, it was possible to differentiate pa-tients with disorders of the oral and/or pharyngeal swallowing phase (16/20) fromthose without swallowing dysfunction (4/20). In 11/16 patients a passage of con-trast medium into the upper airways was observed during swallowing, while in 5/16 cases an aspiration of contrast medium into the lower airways was recorded.In 12 patients the grade of aspiration was decreased in patients with hyperflexionof the head.Conclusion: The dynamic swallowing study with digital technique allows a clear-cut evaluation of the swallowing and to plan a suitable therapeutic approach.

B-692 11:15

Transient small bowel intussusception: An incidental finding on ultrasoundM.A. Mateen, S. Saleem, C. Rao, V. Gangadhar, D.N. Reddy; Hyderabad/IN([email protected])

Purpose: To study the ultrasound findings and clinical significance of transientsmall bowel intussusceptions in adults and children.Methods and Materials: The US findings and clinical outcome of 108 patients ofintestinal intussusception diagnosed on US between August 1995 and August2004 were reviewed. US examinations were performed with a convex transducerof frequency 3.5 to 5 MHz, a linear transducer of frequency 7 to 12 MHz and atransvaginal transducer of frequency 5 to 7.5 MHz. In all cases the length, diam-eter, wall thickness and color Doppler study of the bowel segment involved inintussusception were evaluated. Subsequent follow-up scans were performed at30 minutes, three days and two weeks. Patients were clinically followed up for sixmonths.Results: 41 patients were diagnosed as transient small bowel intussusception.36 intussusceptions were incidentally detected during US performed for someunrelated disease or vague abdominal symptom. Two patients had two adjacentintussusceptions. All the 36 intussusceptions showed spontaneous reduction andwould have gone undetected had it not been for the US performed at that partic-ular time. Five patients presented with signs of obstruction at the time of the initialUS diagnosis; however the intussusceptions resolved without any treatment andwere not detected on follow-up scans. 67 symptomatic patients required surgicalintervention.Conclusion: Incidentally detected, small bowel intussusceptions without an iden-tifiable pathological lead point, with a normal wall thickness, a length of less than3.5 cm, normal non-dilated proximal bowel and normal vascularity on color Dop-pler reduce spontaneously and are of no clinical significance.

B-693 11:24

Functional cine MR imaging for the assessment of the colonic transit time:A pilot studyS. Buhmann1, C. Kirchhoff1, R. Ladurner1, T. Mussack1, M.F. Reiser1,A. Lienemann2; 1Munich/DE, 2Schwerin/DE([email protected])

Purpose: The application of radiopaque markers with conventional abdominaloverview and nuclear medical procedures are easy methods assessing colonictransit time, but both methods repeatedly expose patients to radiation providingonly insufficient transit characterisation. Therefore, the aim was to develop anassessment for the determination of colonic transit time using functional cine MRimaging.Methods and Materials: In a pilot study 8 healthy volunteers (4 women, 4 men,mean age 37.5 ± SD14.4y) without anamnesis of gastrointestinal diseases wereexamined using a 1.5 Tesla system with a protocol for respiratory gated 3D T1and T2w- coronal oriented sequences with 12 minute duration after a 3-day dietfor the standardization of the gastric-colonic-content. The volunteers swallowed5 cups filled with Gd-NaCl-ratio prior to the first exam which was repeated after 3,6, 12, 24 and 48 hours.Results: The resulting MR-images were searched for the appliqued capsules.Similar results were obtained to those described for conventional methods incurrent literature. The male transit time was significantly shortened compared towomen (28 ± 2h vs. 48 ± 3h).Conclusion: Functional cine MR imaging is an adequate method for the assess-ment of colonic transit time offering the advantages of non-existing radiation,minor expenditure of time and the possibility of 3D image reconstruction. In termsof future adoption in clinical routine and for surgical planning an automatic soft-ware tool for evaluation is in development.

B-694 11:33

Endosonographic prediction of external anal sphincter function in faecallyincontinent womenM. Titi, J.T. Jenkins, A. Urie, R.G. Molloy; Glasgow/UK([email protected])

Purpose: Female faecal incontinence (FI), in many cases, is the result of exter-nal anal sphincter (EAS) injury secondary to childbirth. Assessment of the analsphincter aims to confirm the presence of EAS defect as this may be surgicallycorrectable, however, few studies attempt to predict EAS function by global as-sessment of endosonography features, in addition to visible defects.Methods and Materials: Prospective study of 100 consecutive women with FI.All underwent endosonography (EUS) and manometry studies. Symptom sever-ity assessed using Wexner score. Multiple endosonographic features of the EASwere assessed and recorded and correlated with EAS function, [determined byMaximum Squeeze Pressure (MSP)]. Features recorded were sphincter bulk,quality, definition, thickness (single/multi-quadrantic), perineal body thickness(PBT) and defect characteristics (angle, edges, scarring, length). A p-value of0.05 or less was considered statistically significant.Results: Median Wexner score was 18 (IQR 15-21). All defects were anterior. 46had low MSP (< 100 cmH20), 19 had low normal MSP (100-120). A mid-to-distaldefect had a reduced MSP [median 85 (IQR 65-103)] compared mid-canal defect[119 (IQR 75-155)](p = 0.006). EAS quality: uniform echogenicity [median 125(IQR 123-155)]; mixed echogenicity [93 (IQR 69-121)] (p = 0.006): EAS thick-ness was correlated with MSP with one quadrant assessment (p = 0.019), multi-quadrant mean (2, 3 or 4) assessment was not significant. Definition, bulk, PBT,number of visible defects and defect angle were not significant.Conclusion: Anal endosonography can potentially be more useful when supple-mentary features are assessed. In addition to confirmation of the presence of adefect, endosonography potentially predicts functional status of the externalsphincter.

B-695 11:42

A prospective study of male faecal incontinence: Anal endosonography isof limited valueM. Titi, A. McCallum, J.T. Jenkins, R.G. Molloy; Glasgow/UK([email protected])

Purpose: Faecal incontinence (FI) is a significant problem that can affect a per-son's life-style. True prevalence of FI is probably underestimated owing to poten-tial embarrassment of the affected individual. FI is believed to predominantly affectwomen with child-birth related trauma. However, some studies suggest that menare equally affected, particularly with advancing age. However, male FI has re-

Scientific Sessions

300 C D E FBA G

ceived little attention. We prospectively assessed male FI according to aetiology,symptoms, and diagnostic approach.Methods and Materials: 59 consecutive males with FI were recruited. Symp-toms assessed with Wexner Score. All patients followed standard investigationpathway starting with manometry, rectal sensation studies then anal endosonog-raphy (EAUS). Co-morbidities and anorectal histories were recorded.Results: Median age was 60 yrs (IQR 46-71); median Wexner was 18/30 (IQR 4-20). Manometry was normal in 48/59 (81.4%). 12 had anal surgery (8 haemor-rhoidectomy, 2 anal fissure, 2 fistula-in-ano). 5 had rectal surgery. No significantdifference in sphincter function with surgery vs no surgery. No significant differ-ence in function was found when co-morbidities were assessed. Age > 60 yrswas associated with reduced Maximum Resting Pressure (MRP) [52 (IQR 42-76)v 77 (IQR 57-94); p = 0.007]. EAUS found entirely normal sphincters in 38/59(64.4%). 16 had thinning of IAS or EAS or both, but no defects. 5/59 (8.5%) de-fects were recorded, 4/5 post surgery. All rectal surgery patients had abnormalrectal sensation.Conclusion: Male incontinence is an existing problem that deserves more atten-tion. It is infrequently attributed to sphincter defects, even post-surgery, and othercauses should be investigated. EAUS may be of limited value.

B-696 11:51

Posterior pelvic floor disorders: A prospective comparison withcolpocistodefecography and dynamic vaginal endosonographyS. Piciucchi, R.F. Grasso, M. Sammarra, C.C. Quattrocchi, E. Leo, C. Gaudino,D. Vaccaro, B. Beomonte Zobel; Rome/IT

Purpose: To evaluate the accuracy of dynamic vaginal endosonography in de-tecting diseases of posterior pelvic floor as a possible alternative technique forearly diagnosis, particularly in young patients. A comparison of ultrasound andradiological findings is presented.Methods and Materials: Thirty-seven consecutive females (age, range) with func-tionality impairment of posterior pelvic floor were enrolled after clinical evalua-tion. Exclusion criteria were age under 18 and previous history of vulvectomy.They underwent colpocistodefecography (DF) and dynamic endovaginal US ex-amination (Siemens Antares). Recto-anal angle was measured during "squeeze"maneuver to evaluate the strength of voluntary musculature contraction and dur-ing straining to assess pelvic floor descent. Images were blindly evaluated by tworadiologists experienced in pelvic disorders and the interobserver agreement wasalso calculated. The same measuring reference lines were obtained for each groupof examinations.Results: No statistical differences were observed between colpocistodefecogra-phy and dynamic ultrasound measurement during the "squeeze" maneuver(p = 0.09) and during straining (p = 0.59).Presence of intussusception was blindly evaluated with a 94.5% agreement be-tween ultrasound and defecography. Moreover US study allowed measurementsof internal and external sphincters: sphincter thickness and degree of clinicallyevaluated faecal incontinence showed good correlation.Conclusion: We propose dynamic endovaginal ultrasound as a simple, non-in-vasive way to assess recto-anal angle functionality, allowing in the same time theevaluation of the thickness and integrity of the sphinterial complex whereas DFcan provide only functional information. together with colpocistodefecography,anatomic evaluation of the perineal body in diagnostic evaluation of pelvic floordiseases.

10:30 - 12:00 Room F1

Chest

SS 1804Advanced MR imaging of lung vessels andparenchymaModerators:L. Bonomo; Rome/ITN.J. Screaton; Cambridge/UK

B-697 10:30

Half-Fourier single-shot turbo spin-echo (HASTE) of the lung at 3 T:Feasibility study using parallel imaging with 32-receiver channel technologyC. Fink1, O. Dietrich1, R. Eibel1, P. Herzog1, D. Theisen1, T. Lanz2, M.F. Reiser1,S.O. Schönberg1; 1Munich/DE, 2Rimpar/DE([email protected])

Purpose: To assess the feasibility of half-Fourier single-shot turbo spin-echo(HASTE) of the lung at 3 T using parallel imaging with 32-receiver channel tech-nology.Methods and Materials: In an ongoing feasibility study 5 healthy volunteers wereexamined with HASTE on a new 32-channel 3 T MR imaging system (MagnetomTIM Trio, Siemens Medical Solutions) using a prototype 32-channel torso arraycoil (RAPID Biomedical). HASTE MR imaging (TR=800 ms, FOV=420x420 mm,matrix=256x256, slice thickness = 5 mm) was acquired at both end-inspiratoryand end-expiratory breathhold using parallel acquisition techniques (GRAPPA,64 reference lines) with acceleration factors ranging between 2 and 4. Depend-ing on the acceleration factor, a minimal TE between 26 ms (PAT=2) and 16 ms(PAT=4) was achieved. The image quality of pulmonary and extrapulmonary ana-tomical structures and noise level was evaluated.Results: All images were of diagnostic quality. Using an acceleration factor of 4image blurring was substantially reduced compared to lower acceleration factorsresulting in sharp delineation of intrapulmonary and extrapulmonary structures.Expiratory scans provided a higher intrapulmonary signal intensity than inspira-tory scans. As visually assessed, loss of signal to noise did not impair imagequality even at acceleration factors of 4.Conclusion: HASTE MR imaging of the lung is feasible at 3 T. The use of highparallel imaging acceleration factors allows shorter echo times and reduces im-age blurring. Due to optimised multi-element coil geometry, sufficient signal-to-noise can be achieved despite high acceleration factors. Expiratory scanningmay be favourable to compensate for susceptibility associated signal loss of thelung at 3 T.

B-698 10:39

Detection of small pulmonary nodules in high-field MR at 3 T: Experimentalevaluation of different pulse sequences using porcine lung explantsM. Regier1, S. Kandel2, M. Kaul1, J. Kemper1, B. Hoffmann2,C.C.A. Nolte-Ernsting1, G. Adam1, M. Heller2, J. Biederer2; 1Hamburg/DE,2Kiel/DE ([email protected])

Purpose: To compare and evaluate different MR imaging sequences for the de-tection of injected pulmonary nodules in porcine lungs.Methods and Materials: 84 agarose nodules of different diameter (range 2 mm-20 mm) were injected into 11 porcine lungs within a chest phantom. Three-di-mensional (3D) gradient echo (GRE) and HASTE sequences were performed inaxial and coronal orientations, followed by multislice CT (MSCT) as gold-stand-ard in the diagnosis of pulmonary nodules. Images were read by 3 blinded ob-servers with different diagnostic experience. Sizes and locations of the detectednodules were assessed. Sensitivity, specificity, positive and negative predictive-value (PPV, NPV) for all lesions were calculated for each sequence.Results: The 3D-GRE sequence showed a sensitivity greater than 90% for allreaders in nodules greater than 5 mm in diameter which was comparable to CT.In lesions smaller than 5 mm the sensitivity and PPV declined below 70%. TheHASTE sequence showed a sensitivity greater than 86% in nodules over 5 mm.A specificity, sensitivity, PPV and NPV of nearly 100% was recorded for all se-quences in nodules greater than 10 mm. The rate of false positive findings de-pended on the reader's experience, but was lower for the 3D-GRE than the HASTEsequence. The 3D-GRE sequence showed a higher overall accuracy than theHASTE sequence.Conclusion: The study confirms the feasibility of 3 T-MR for the detection of smallsolid lung lesions. In lesions greater than 5 mm, the sensitivity of the 3D-GRE

TT TTTu

esd

au

esd

au

esd

au

esd

au

esd

ayy yyy

Scientific Sessions

B 301C D E FA G

sequence was comparable to MSCT (> 90%) while sensitivity and PPV with theHASTE sequence were slightly inferior.

B-699 10:48 !Morphological features and clinical feasibility of thoracic duct detectionwith 3.0 Tesla magnetic resonanceD. Yu, C. Li, X. Zhang, J. Xiu, Q. Wang; Jinan/CN ([email protected])

Purpose: To evaluate the position, configuration, dimensions of the thoracic ductwith fat-suppressed T2-weighted sequences and MR lymphography (MRL) andthe clinical significance.Methods and Materials: MRL by heavily T2-weighted techniques and fat-sup-pressed T2-weighted sequences was performed in 139 patients. Forty-two pa-tients were excluded because of poor reconstructional images. On T2-weightedimages, the position of thoracic duct with reference to the vertebrae was defined,and the coronal and sagittal MIP reconstructional images were obtained for meas-urement of the dimension. The differences of the parameters in the groups in-cluding a control group, portal hypertension group and choledochal obstructiongroup were compared.Results: The cisterna chyli was shown in 91.37% of patients on fat-suppressedT2-weighted images. Most were located at the vertebral level between L1 and L2in 80.31%, and in the midline of the vertebrae in 75.59%. On the MIP images, thethoracic duct was shown in 69.78%, and the most common configuration of thecisterna chyli was tubular or saccular. The mean longitudinal, anteroposteriorand transverse diameters of the cisterna chyli were 26.04 (± 10.45)mm,5.08 (± 2.08)mm, and 5.02 (± 2.13)mm respectively. The anteroposterior and trans-verse diameters of the thoracic duct were 3.08 (± 0.44)mm and 3.58 (± 0.13)mmin control group. There was a significant difference of the transverse diameter ofthe thoracic duct between control and disease groups (F=5.638, P=0.005).Conclusion: Noncontrast-enhanced MR techniques may be used to detect thethoracic duct, and portal hypertension may influence its dimension.

B-700 10:57

T1 mapping of the entire lung: Influence of respiratory phase in healthyindividualsA. Stadler, P.M. Jakob, M. Barth, E. Eisenhuber, A.A. Bankier; Vienna/AT([email protected])

Purpose: To test the feasibility of in-vivo T1 measurements of the entire lungparenchyma, to quantify T1 measured in the entire lung, and to determine theeffect of inspiration and expiration on the measured T1 values.Methods: Ten healthy volunteers underwent MR imaging investigation of the lungusing a 1.5 T scanner. Image acquisition was based on a SNAPSHOT-FLASHsequence. The acquired series of images were used for calculation of T1 relaxa-tion time on a pixel-by-pixel basis. For each volunteer image acquisition was per-formed both in inspiration and expiration.Results: The average of the mean T1 values was in inspiration 1199 ± 117 msand in expiration 1333 ± 167 ms. T1 values were higher in expiration than in in-spiration in all 10 volunteers. The difference was statistically significant (P=0.005).In inspiration, the difference between the average dorsal T1 value and the aver-age ventral T1 value was 81 ms with higher ventral values. This difference wasstatistically significant (P=0.022). In expiration, the difference between the aver-age dorsal T1 value and the average ventral T1 value was 33 ms, with higherdorsal values. This difference was not statistically significant.Conclusion: Our study documents the feasibility of T1 measurements of the en-tire lung parenchyma. These measurements resulted in significantly lower T1values in inspiration, and in significant differences between inspiratory T1 valuesfrom the ventral and the dorsal parts of the lung. Our findings underline the needfor a close monitoring of the inspirational status during MR examinations of thethorax, notably when T1 measurements of the lung are performed.

B-701 11:06

Oxygen-enhanced MR imaging of the lung using a pneumotachograph forrespiratory monitoringF. Molinari, G. Corbo, S. Valente, T. Pirronti, L. Bonomo; Rome/IT([email protected])

Purpose: To demonstrate the benefits of respiratory monitoring with a pneumo-tachograph in oxygen-enhanced MR imaging of the lung.Methods and Materials: 7 healthy-volunteers (age-range: 26-32) were imagedon a 1.5 T MR-scanner. To evaluate the impact of monitoring respiration with apneumotachograph, 5 series of IR-SSFSE images (TE=28.2 ms; TI=1200 ms)were obtained in each subject during the ventilation-paradigm room air-100%

oxygen-room air: a) Series no.1, respiratory-triggered [R-tr]; b) Series no.2, car-diac-triggered [C-tr]; c) Series no.3, cardiac-triggered with respiratory-monitoringbased on a pneumatic-belt [C-tr/BELT-m]; d) Series no.4, cardiac-triggered andrespiratory-monitoring with a pneumotachograph [C-tr/PNTC-m]; e) Series no.5,respiratory-monitoring with a pneumotachograph [PNTC-m]. Standard deviationsof the lung (SIvar) and diaphragm mismatch (DM) were measured on room-air-images. The relative SI-change (∆SI) was computed from room-air and oxygen-enhanced-images. Parametric maps were obtained from cross-correlation analysisof the ventilation-paradigm. Mean correlation-coefficients (cc) and the percent-age of activated pixels over the lung (Act%) were calculated from these maps. Allfive parameters were compared among the five series (Friedman-ANOVA, Dunn'spost-hoc test).Results: In series no.4 [C-tr/PNTC-m], DM and SIvar were significantly lowerthan in respiratory and cardiac-triggered series (DM= 4.7% vs 12.4% and 14.4%;SIvar = 7.6% vs 14.1% and 12.8%). In the same series cc and Act% were alsosignificantly higher than in series 1 and 2 (cc= 0.86 vs 0.70 and 0.68; Act%=71.3% vs 44.7% and 41.2%). ∆SI was not significantly different among all series.Conclusion: Continuous monitoring of breathing has potential for: a) reducingsignal variability and diaphragm mismatch in IR-SSFSE images; b) improving thequality of oxygen-enhanced ventilation-maps of the lung.

B-702 11:15

Intrapulmonary ³He gas distribution depending on bolus positionK.K. Gast1, K. Hawig1, A.E. Morbach1, A. Herweling1, M. Windirsch1, C. Düber1,H.-U. Kauczor2, W.G. Schreiber1, C.P. Heussel1; 1Mainz/DE, 2Heidelberg/DE([email protected])

Purpose: Dynamic ³He-MRI of ventilation is a new method for quantitative as-sessment of intrapulmonary gas distribution. This study intends to investigate thetemporal and regional differences in intrapulmonary distribution of a gas bolusdepending on its volume and application delay.Methods and Materials: 12 healthy domestic pigs were anaesthetized, intubat-ed and mechanically ventilated. ³He boli of 100 ml, 150 ml and 200 ml were ap-plied via a gas application device followed by room air. The 150 ml bolus wasadditionally applied with delays of 0, 200 … 1200 ms after initial inspiration.Results: The mean global tracheoparenchymal delay (td90) was 120 ms in the100 ml bolus. It was prolonged to 190 ms with increasing bolus size (150, 200 ml).Introducing a 200 ms delay, td90 was 250 ms decreasing to 140 ms with increas-ing delay (up to 1200 ms). In the euthanized pig the td90 was 160 ms withoutapplication delay.Conclusion: Dynamic ventilation ³He-MRI shows differences in inspiratory gasdistribution depending on volume and delay of the contrast-gas. The increase oftd90 with bolus size is probably caused by the longer application duration of alarger bolus. The euthanized pig shows accelerated gas inflow with post mortemloss of compliance of the lung. Gas application with increasing delay after initialinspiration leads to an accelerated gas inflow due to facilitated convection anddiffusion by widening of airways during later inspiratory phases.Acknowledgements: The ³He project is supported by the German Research Coun-cil (DFG, FOR474) and the research program of Mainz University (MAIFOR).This work contains data from K. Hawig's doctoral thesis.

B-703 11:24

Pulmonary imaging using hexafluorethane (C2 F6): Initial experienceU. Wolf, A. Scholz, C. Düber, K. Markstaller, C.P. Heußel, W.G. Schreiber;Mainz/DE

Purpose: The actual strategy to improve SNR in 19F-MRI is a high number ofaverages using gases with short T1 such as SF6 (1.65 ms), necessitating shortecho times and high bandwidths. The aim was to improve SNR through loweringreceiver bandwidths using C2

F6

(T1 5x SF6).

Methods and Materials: 3 pigs were ventilated with 70% C2 F6 (or SF6) and 30%

O2. FLASH projection was done during and after gas wash-in. Dynamic imagingwith 1 frame per breath every 5s (TA=2s) and static imaging after completion ofthe wash-in (TA=10s) at three different bandwidths 200, 488, 700 Hz/pixel duringan inspiratory hold was done. Furthermore, triples of 1, 3 and 100 acquisitionswere taken after wash-in at 488 Hz/pixel, exclusively.Results: C2

F6 was clearly superior to SF6 at all tested bandwidths. The SNR for

C2 F6 was maximal at 200 Hz/pixel (488 Hz/pixel for SF6). For the first time, sub-

second 19F-MRI of the lung proved feasible using C2 F6 at 488 Hz/pixel. A peak

SNR of 11 was observed for a scan time of 780 ms.Conclusion: MR imaging of C2

F6

yielded good image quality in healthy pig lungs.

In the medium bandwidth range tested here, C2 F6 was clearly superior to SF6.

This is a result of the relatively low bandwidths applied here which are well scaled

Scientific Sessions

302 C D E FBA G

to the T1 of C2 F6. Therefore, SNR improvement through lowering of the receiver

bandwidth proved to be efficient using C2 F6. In summary, subsecond 19F-MRI of

the lung proved feasible using C2 F6 at the medium bandwidth of 488 Hz/pixel.

B-704 11:33

Navigator triggered oxygen-enhanced MR imaging in patients withinterstitial lung diseaseF. Molinari1, M. Eichinger2, F. Risse2, L. Bonomo1, H.-U. Kauczor2, C. Fink3;1Rome/IT, 2Heidelberg/DE, 3Munich/DE ([email protected])

Purpose: To demonstrate a correlation between navigator-triggered oxygen-en-hanced MR imaging of the lung and pulmonary function tests (PFT) in patientswith interstitial lung disease (ILD).Methods and Materials: 12 healthy-volunteers (age range: 20-32y) and 10 pa-tients (age range: 37-87y) with ILD underwent pulmonary function tests (PFTs)and oxygen-enhanced MR imaging. Static lung volumes, forced expiratory flowsand diffusion capacity were obtained in all subjects. Arterial blood gases wereobtained only in the patient-group. Oxygen-enhanced MR imaging was performedon a 1.5 T MR-scanner. The paradigm room-air-O2-room-air was acquired with anonselective-IR-HASTE (TE=12 msec; TI=1200 msec). Navigator and cardiac-trigger were combined in a double-triggering-method to control simultaneouslycardio-respiratory motion. Image acquisition was performed in end-expiration.Inhouse-developed software was used to further reduce motion artifacts. Cross-correlation analysis was performed in ROIs encompassing both lungs (r-val-ue = 0.5; p-value = 0.01 as thresholds). The percentage of activated pixels overthe total number of pixels in the ROIs (OAP%) was calculated from both lungs.Mean OAP%s calculated from ILD-patients and volunteers were compared (Mann-Whitney test). The correlation between OAP%s and PFTs was assessed withSpearman-test.Results: The double-triggering method reduced artifacts from cardio-respiratorymotion. Free-breathing end-expiratory acquisition was time-effective and not tir-ing for all subjects. Image post-processing further reduced the signal variability.The mean OAP% of patients was significantly lower than that of volunteers (61.3vs 84.0;p = 0.04). OAP%s correlated signif icantly with ar ter ial-pO2(r = 0.67;p = 0.02), and with arterial-O2-saturation (r = 0.72;p = 0.01).Conclusion: Oxygen-enhanced MR imaging of the lung using navigator and car-diac-triggering is an efficient imaging method and has a potential role in the func-tional assessment of ILD.

B-705 11:42

Free-breathing non-contrast VIBE MR-angiography of the lung using ECGtriggering and respiratory gating: Image quality assessmentJ. Zaporozhan1, S. Ley1, C. Fink2, A. Stemmer3, M. Puderbach1, R. Eberhardt1,E. Grünig1, H.-U. Kauczor1; 1Heidelberg/DE, 2Munich/DE, 3Erlangen/DE([email protected])

Purpose: Aim of this study was to evaluate the feasibility and image quality of a"work in progress" ECG triggered and respiratory gated volume interpolated bodyexamination sequence (VIBE) for non-contrast MR-angiography of the pulmo-nary arteries.Methods and Materials: The whole lung was imaged in 36 subjects (10 volun-teers, 26 patients) during quiet respiration on a 1.5 T MR scanner. The sequencewas applied in a coronal orientation (TR/TE=9/3.32 msec; flip angle 8°;1.3x1.2x2 mm3; GRAPPA 2; 88 slices). In each heartbeat, the navigator for respi-ratory gating is executed after an appropriate trigger delay time, followed by a fatsuppression pulse and the data acquisition bloc. Images were assessed for sharp-ness and quality of pulmonary arterial branches and thoracic structures and pres-ence of artifacts using a 4-point-score (4-excellent to 1-insufficient/severe artifacts).Results: The mean acquisition time was 10 ± 3 min. 86% of the images were ofgood to excellent quality with no or minor artifacts. In 8% some parts of the lungwere not diagnostic. In 6% the images could not be evaluated due to severeartifacts. The mean score for sharpness of pulmonary arteries was 3.2 and 3.6for thoracic structures. Overall the image quality was fine, mean score 3.3. Mostartifacts were found in the lung parenchyma or close to the diaphragm, overallmean score was 3.6.Conclusion: The new ECG triggered and respiratory gated VIBE sequence wassuccessfully used in 86% of patients during free-breathing with a high imagequality and high level of sharpness of vascular and thoracic structures.

B-706 11:51

Quantitative whole tumour perfusion assessment using MDCT:Reproducibility and observer variability in advanced lung cancerV.J. Goh1, Q. Ng1, E. Klotz2, H. Fichte2, M. Saunders1, P.J. Hoskin1,A. Padhani1; 1Northwood/UK, 2Forchheim/DE ([email protected])

Purpose: Perfusion techniques have been confined to a small tumour volume todate, which may not be representative of the entire tumour. We aimed to deter-mine the reproducibility and observer variability of a quantitative whole tumourperfusion CT technique.Methods and Materials: Following IRB approval and informed consent, 10 pa-tients with advanced lung cancer (mean size 8.3 cm) were scanned prospective-ly using a 16-detector row scanner. The perfusion studies, consisting of sequentialvolumetric acquisitions encompassing the entire tumor, were acquired followingIV contrast infusion. Studies were repeated within 24 hours to assess reproduc-ibility. All 20 studies were analysed by two independent observers. The wholetumour was analysed pixel-by-pixel using perfusion software based on Patlakanalysis; median values of permeability and blood volume were obtained. Meas-urement reproducibility and observer agreement were assessed using Bland-Altman statistics.Results: The mean difference (95% limits of agreement) for repeated studieswas 0.07 ml/100 ml/min (-3.73 to +3.87) and 1.30 ml/100 ml (-2.63 to +5.23) forpermeability and blood volume respectively; the coefficient of variation was 9.49%and 26.31% respectively indicating good reproducibility. For inter-observer agree-ment, the mean difference (95% limits of agreement) was 0.07 ml/100 ml/min (-1.01 to +1.08) for permeability, and 0.44 ml/100 ml (-0.59 to +0.68) for bloodvolume. The coefficient of variation between observers, and within observers was3.47%, 6.34% and 3.50% respectively for permeability, and 3.64%, 3.30% and3.95% respectively for blood volume, indicating good agreement.Conclusion: Good measurement reproducibility and inter- and intra-observeragreement can be achieved with whole tumour perfusion CT, boding well for clin-ical application.

10:30 - 12:00 Room F2

Cardiac

SS 1803Advances in coronary imagingModerators:A. Huber; Munich/DES. Katsilouli; Athens/GR

B-707 10:30

64-slice CT coronary angiography in patients with acute coronarysyndromes: First experienceF. Pugliese, N.R.A. Mollet, W.B. Meijboom, A.A. Palumbo, L. La Grutta,F. Cademartiri, G.P. Krestin; Rotterdam/NL ([email protected])

Purpose: To test the feasibility and accuracy of 64-slice CT coronary angiogra-phy (gantry rotation time 330 msec) in patients with acute coronary syndromes,with invasive coronary angiography and intravascular ultrasound (IVUS) as refer-ence standard.Methods and Materials: Twenty-five patients with unstable angina (UA) or sus-pected acute myocardial infarction (AMI) underwent CT coronary angiographywithin 24 hours prior to cardiac catheterization. IVUS studies were conducted in16 patients. Subjects in sinus heart rhythm were included, provided they wereable to hold their breath for 12-14s. Patients with heart/renal failure or allergy tocontrast agent were excluded. Beta-blockers were administered to patients withheart rates above 65 beats per minute (bpm).Results: 11/25 patients presented with UA and 14/25 with AMI. Beta-blockerswere administered to 15 patients. Average heart rate dropped from 67 to 58 bpm.All subjects were correctly identified as normal (n = 1), single-vessel (n = 8) ormulti-vessel (n = 16) disease patients. Five/25 patients were overdiagnosed with3-vessel instead of 2-vessel disease. In the analysis of lesion characteristicsperformed blinded of the clinical context and IVUS results, the culprit lesion couldbe defined in all patients with 1-vessel disease (n = 8), in 6 patients with 2-vesseldisease (n = 8) and in 5 patients with 3-vessel disease (n = 8).Conclusion: 64-slice CT in acute coronary syndromes may be helpful in identify-ing single-vessel versus multi-vessel disease. Assessment of the culprit lesionseems feasible based on plaque characteristics and could be enhanced by knowl-edge of the clinical context and clinical tests results.

TT TTTu

esd

au

esd

au

esd

au

esd

au

esd

ayy yyy

Scientific Sessions

B 303C D E FA G

B-708 10:39

Non-invasive coronary angiography using 64-detector row computedtomographyT. Schlosser1, O.K. Mohrs2, A. Magedanz2, B. Nowak2, T. Voigtländer2,A. Schmermund2, J. Barkhausen1; 1Essen/DE, 2Frankfurt/DE([email protected])

Purpose: The aim of this study was to evaluate the value of 64-detector rowcomputed tomography for ruling out high-grade coronary stenoses in patientswith a low to moderate pre-test probability of significant coronary artery disease.Methods and Materials: The study included 179 patients with a suspicion ofcoronary artery disease on the basis of atypical angina or ambiguous findings innon-invasive stress testing. All patients were examined using 64-detector rowcomputed tomography angiography (CTA). All Patients with stenoses or uncer-tain lesions at CTA (n = 61) underwent invasive coronary angiography (ICA). Ona coronary segmental level, the presence of high-grade (≥ 50% diameter) sten-oses was compared.Results: In a total of 915 segments, CTA detected 62 significant stenoses. Thir-ty-four significant stenoses were confirmed by ICA, whereas 28 stenoses couldnot be confirmed by ICA. 22 of those 18 lesions showed stenoses < 50% at ICAwhereas 6 segments were angiographically normal. Accordingly, on a coronarysegmental basis, 28 false positive and 0 false negative findings resulted in asensitivity of 100%, a specificity of 96.8%, a positive predictive value of 54.8%,and a negative predictive value of 100%. The diagnostic accuracy was 96.9%.Conclusion: Sixty-four-detector row computed tomography reliably detects sig-nificant coronary stenoses in patients with suspected coronary artery diseaseand appears helpful in the selection of patients who need to undergo ICA. How-ever, grading of stenoses is still challenging and stenoses are frequently overes-timated by 64-detector row computed tomography.

B-709 10:48

64-slice CT angiography reliably detects and excludes significant coronaryartery stenosis on a per-patient but not on a per-vessel or per-segmentbasisG. Savino1, J.U. Schoepf1, P. Costello1, P. Zwerner1, T.J. Vogl2, C. Herzog2;1Charleston, SC/US, 2Frankfurt/DE ([email protected])

Purpose: To investigate the diagnostic accuracy of 64-slice multi detector-rowCT (MDCT) in the assessment of coronary artery disease (CAD).Methods and Materials: 55 patients with suspected CAD underwent both, con-ventional coronary catheter angiography and contrast enhanced, retrospectivelyECG-gated 64-slice MDCT of the coronary arteries. Two experienced observersanalyzed all MDCT-scans for signs of CAD (degree of stenosis = 0%; < 50%;< 70%; < 99%; 100%). Diagnostic accuracy compared to cardiac catheter wasdetermined on a per-segment, per-vessel and per-patient basis.Results: Globally, all CT coronary angiograms (55/55) were considered of diag-nostic quality. On a per-segment basis, 92.2% (762/826) of all segments could beclearly evaluated. In 7.8% of segments, image quality was compromised eitherby misregistration (15.6%), motion artefacts (31.2%) or small vessel size (53.1%).The correlation coefficients for the detection and grading of stenosis were r = 0.65on a per-segment, r = 0.83 on a per-vessel and r = 0.88 on a per-patient basis.Stenoses > 50% were detected with 96.0% (802/825) accuracy, 81.9% (50/61)sensitivity and 97.1% (743/765) specificity on a per-segment, 90.9% (150/165)accuracy, 88.9% (32/36) sensitivity and 91.5% (118/129) specificity on a per-vessel, and 89.1% (49/55) accuracy 100% (19/19) sensitivity and 83.3% (30/36)specificity on a per-patient basis. Stenoses > 70% were detected with higheraccuracy than stenoses > 50% without reaching a level of significance (p = 0.82).Conclusion: On a per-patient basis, 64-slice MDCT coronary angiography al-lows for reliable diagnosis and exclusion of significant CAD. On a per-segmentand per-vessel basis, diagnostic accuracy is still impaired, primarily by limitedspatial resolution.

B-710 10:57

Flat panel computed tomography of human intramyocardial coronarycirculationF. Knollmann, T. Kertesz, F. Haller, L. Füsezi, H.-J. Radzun, E. Grabbe;Göttingen/DE ([email protected])

Background: Recent improvements in computed tomography have expanded itsapplication to coronary artery imaging. Using a flat panel CT research system(GE Global Research), a spatial resolution of 0.2 mm can be achieved, whichraises the possibility of depicting the intramyocardial coronary circulation of hu-man hearts.

Methods and Materials: The post mortem flat panel CT examination of excisedhuman hearts displays the coronary arteries lumen throughout all branches. Thecoronary arteries were filled with a barium sulphate contrast agent to achieve anintraarterial opacification that is equivalent to results obtained with venous injec-tion of iodinated contrast agents in humans.Results: The sixth generation coronary artery branches could consistently bediscerned, and intramyocardial vessel segments were clearly visualized.Discussion: The consistent visualization of intramyocardial coronary artery seg-ments opens new avenues for the noninvasive diagnosis of cardiac disease. Al-though the method is presently limited to a static situation, it may serve as proofof principle that new CT techniques with superior spatial resolution are capableof investigating the intramyocardial coronary circulation. Also, research on therole of vascular growth factors for arterial myocardial architecture would poten-tially benefit from the new technique.

B-711 11:06

Multi-slice CT-angiography of the coronary arteries in patients with atrialfibrillationA. Kovacs, C. Probst, T. Sommer, P. Naehle, H. Schild, S. Flacke; Bonn/DE([email protected])

Purpose: To determine the optimal reconstruction window during the cardiaccycle in patients with atrial fibrillation (AF) for each of the three major coronaryarteries using a new frequency adapted reconstruction algorithm.Methods and Materials: 20 patients with permanent AF were imaged on a 16slice scanner (slice coll.16x0.75 mm; rotation time 0.42 s; 140 kV; 380 mAs). Thepatients had not received previously any drugs for heart frequency regulation.Acquisition was started after bolus tracking of a biphasic bolus of 120 ml Ultravist370Â® injected intravenously. Each coronary segment was reconstructed at 0%-90% of the cardiac cycle in increments of 10%. Two blinded independent readersassessed the image quality of the coronary segments in terms of visibility andartifacts (five point rating scale 1=very poor up to 5=excellent) and the degree ofstenosis (five point rating scale 1=0%, 2=1%-49%, 3=50%-74%, 4=75%-99%,5=100%) on axial slices, multi-planar reconstructions and three-dimensional vol-ume rendered images.Results: The heart rate during the examination ranged between 42 and 156 beatsper minute, the average heart rate was 78 bpm ± 23. Each of the two readersevaluated 300 segments in 20 patients. Visualisation of all coronary artery seg-ments was superior (p < 0.01) at 40% (mean score of the image quality 2.79) ascompared to the standard diastolic reconstruction window at 80% (image quality2.33). The second best image quality (2.57) was acquired at 0% of the cardiaccycle.Conclusion: Using a frequency adapted delay algorithm with the choice of anendsystolic reconstruction window provided diagnostically valuable images inpatients with AF.

B-712 11:15

Non-calcifying coronary plaques and acute coronary syndromes:A retrospective study using 16-multidetector computed tomographycoronary angiographyG.M. Feuchtner, G.J. Friedrich, F. Weidinger, T. Postel, A. Mallouhi, M. Frick,H. Alber, O. Pachinger, D. zur Nedden; Innsbruck/AT([email protected])

Purpose: To evaluate whether different coronary plaque types as assessed with16-multidetector computed tomography (16-MDCT) are correlated with acutecoronary syndromes (ACS).Methods and Materials: 63 consecutive unselected patients who underwent 16-MDCT coronary angiography (Sensation 16™, Siemens) (16 x 0.75 mm, 0.42 s,TF=6.7 mm/s; 120 kV, 500 mAs; 100 ml Iodixanol (Visipaque™, Amersham), flow3-4.5 ml/s; eff.sl.1 mm, inc.0.5;retrospective ECG-gating) were retrospectivelyanalysed. Coronary plaques were defined as type 1=calcifying (C); type 2=mixed(C > non-calcifying (N)); type 3=mixed (N > C); type 4=non-calcifying (prox. seg-ments). All patients underwent cardiac catheterization (CAG).17 patients had in-tracoronary ultrasound (ICUS). ACS was defined as Troponin T elevation,ECG-abnormalities, chest pain.Results: A total of 58 patients had coronary plaques as detected with 16-MDCT(predominant plaque type 1 [n = 12]; type 2 [n = 10]; type 3 [n = 16]; type 4 [n = 14]).The presence of a non-calcifying plaque component (type 2-4)(40/63) was stronglycorrelated with an ACS [n = 17](p < 0.0001; Fisher's-exact-test). Mixed plaquetype 3 was significantly correlated with ACS (p = 0.01). Mixed plaque type 2 (p = 1)and plaque type 4 (p = 0.09) were not related to ACS. 16-MDCT correctly classi-fied plaque types in 15 of 17 patients when compared with ICUS. 45 of 63 pa-

Scientific Sessions

304 C D E FBA G

tients had at least one coronary stenosis > 50% on CAG (Sensitivity of 16-MDCT:98% (43/45); specificity 90%; NPV 97%; PPV 97%). 15 of 17 patients with ACShad coronary artery disease (coronary anomaly [n = 1]; acute myocarditis [n = 1]).Conclusion: Non-calcifying coronary plaques as detected with 16-MDCT arecorrelated to ACS suggesting higher plaque vulnerability. 16-MDCT may havethe potential to identify high-risk patients who may be treated more aggressivelyand in whom close clinical follow-up is mandatory.

B-713 11:24

Influence of cardiac hemodynamic parameters on coronary arteryopacification with 64-slice CTL. Husmann, T. Boehm, S. Leschka, T. Schepis, P. Koepfli, B. Marincek,P.A. Kaufmann, S. Wildermuth, H. Alkadhi; Zurich/CH ([email protected])

Purpose: To evaluate the influence of ejection fraction (EF), stroke volume (SV),heart rate, and cardiac output (CO) on coronary artery opacification with 64-slicecomputed tomography (CT).Methods and Materials: 60 patients (22 women, 38 men; mean age 61.9 ± 10.5years) underwent retrospectively electrocardiography-gated 64-slice CT coronaryangiography. Ten data sets were reconstructed in 10% steps of the R-R intervalfor quantification of left ventricular EF, SV, and CO using software capable ofsemi-automated contour detection. Attenuation values were measured and con-trast-to-noise ratios (CNR) were calculated in the proximal right coronary artery(RCA) and left main artery (LMA). CNR was defined as difference of mean atten-uation in the vessel from the mean attenuation in perivascular tissue divided byimage noise in the ascending aorta.Results: Mean EF during scanning was 61.5 ± 12.4%, SV was 63.2 ± 15.6 ml,heart rate was 62.5 ± 11.8 bpm, and CO was 3.88 ± 1.06L/min. No significantcorrelation was present between the EF and heart rate and the attenuation andCNR in both coronary arteries. A significant negative correlation was found inboth arteries between SV and attenuation (RCA: r=-0.26, P < 0.05; LMA: r=-0.34,P < 0.01) and between SV and CNR (RCA: r=-0.26, P < 0.05; LMA: r=-0.26,P < 0.05). Similarly, a significant negative correlation was found between the COand attenuation (RCA: r=-0.42, P < 0.05; LMA: r=-0.56, P < 0.001) and betweenthe CO and CNR (RCA: r=-0.39, P < 0.05; LMA: r=-0.44, P < 0.001).Conclusion: The actual hemodynamic status of the patient influences the coro-nary artery opacification with 64-slice CT, in that vessel opacification decreasesas SV and CO increases.

B-714 11:33

Prospective ECG-pulsing and retrospective ECG-gating in 64-slice cardiacMDCT: Evaluation of image quality of coronary arteries in 37 patientsJ. Schmitt1, K. Koenig1, N. Abolmaali2, H. Ackermann3, G. Görge1, G. Brill1;1Saarbrücken/DE, 2Dresden/DE, 3Frankfurt am Main/DE ([email protected])

Purpose: To evaluate the quality and dose differences between prospective pulsingand retrospective gating in coronary MDCT.Methods and Materials: 37 consecutive patients in a stable clinical conditionand sinus rhythm were examined. Patients with heart rates above 60 b.p.m. re-ceived 100 mg atenolol p.o. and up to 50 mg esmolol i.v. Expiratory and contrastenhanced (80 ml) MDCT was performed using 64 slices at 0.75 mm collimation,120 kV, and ECG-gated tube current modulation. The images were visually ana-lysed by two readers in consensus using the 16-segment coronary artery modelof the American Heart Association (AHA). The quality of the visualization of eachcoronary segment was rated (1= excellent, 2= evaluable 3= not evaluable). 21patients (group A) were examined with prospective ECG-pulsing, 16 patients(group B) were examined with retrospective ECG-gating. The difference in imagequality was done by the Mantel-Haenszel-test on a 95%-level. The radiation ex-posure was compared using the DLP.Results: Mean quality of all coronary segments in group A was 1.71 and 1.89 ingroup B, respectively. In all coronary segments the difference in quality betweengroups A and B did not reach significance. The p-values varied between 0.1 and0.9. The DLP was 447 mSvcm in group A and 854 mSvcm in group B, respective-ly. Mean heart rate in both groups was 59 b.p.m. during MDCT.Conclusion: ECG-pulsing is a valuable tool for 64 slice coronary MDCT thatallows reduction of radiation exposure by about 50% preserving complete imagequality.

B-715 11:42

Multislice computed tomography of the coronaries: Intraindividualcomparison of radiation dose exposure with conventional coronaryangiographyC. Weber1, A. Deetjen2, O. Ekinci2, J. Hansel2, A. Kluge2, M. Weber2,G. Bachmann2, C. Hamm2, T. Dill2; 1Hamburg/DE, 2Bad Nauheim/DE([email protected])

Purpose: The aim of this study was the intra-individual comparison of the effec-tive radiation dose applied for Multislice spiral computed tomography (MSCT)and conventional coronary angiography (CXA).Methods and Materials: 56 patients (68 ± 8 years, 49 male) underwent MSCTand CXA. For MSCT, the standard protocol consisted of pre-monitoring, monitor-ing, calcium scoring and MSCT coronary angiography using a 16 slice CT scan-ner. CXA was performed on a current system with automatic selection of X-raybeam filtration following a protocol with 8 standard projections (LCA 6, RCA 2).To compare the radiation doses of both investigations, the effective dose waschosen as a parameter. A conversion factor was used for the equation of theeffective dose: E=k x DLP. For conventional angiography, the effective dose wascalculated using the dose-area product and a conversion factor according to datapublished by Le Heron.Results: The mean effective dose for MSCT including pre-monitoring and moni-toring scans was 10.36 ± 2.18 mSv (n = 56); if calcium scoring was added toMSCT, the dose was 12.67 ± 2.58 mSv (n = 45). In comparison, the mean effec-tive dose of CXA was 3.33 ± 2.42 mSv (n = 56). A subgroup of patients (n = 10)with bypass grafts were analyzed in addition. The effective dose for MSCT was12.95 ± 1.75 mSv. In comparison, the mean effective dose of CXA was6.27 ± 4.04 mSv.Conclusion: This intraindividual comparison demonstrates that MSCT is stillassociated with a relatively high radiation dose exposure. This study emphasizesthat indications for MSCT have to be clearly defined.

B-716 11:51

Pediatric cardiovascular 64-slice CT angiography: Radiation dose reductionby automated anatomical tube current modulationG. Savino1, J.U. Schoepf1, P. Costello1, T.J. Vogl2, D. Mulvihil1, C. Herzog2;1Charleston, SC/US, 2Frankfurt/DE ([email protected])

Purpose: To assess the effect of weight-based scan protocols and automatictube current modulation on radiation dose and image quality at pediatric 64-sliceCTA.Methods and Materials: Our pediatric cardiovascular 64-slice CT protocols usea weight based algorithm to determine nominal tube current settings with 80-120 kV. Automated anatomic tube current modulation was used for each case.mAs, CTDIvol and DLP values were recorded and the effective dose calculated.Based on the selected nominal tube current, CTDIvol, DLP and effective dose,which would have been delivered without tube current modulation, were alsocalculated. Two radiologists independently rated image quality on a 5-point scale.Image noise was objectively assessed. Findings at CT were clinically correlatedwith cardiac ultrasound, angiography or surgery.Results: 38 CT scans were evaluated. Mean diagnostic quality was rated at 3.7(range 3-4.4), with mean image noise of 9.1 HU STD (range 3.3-15.9). Com-pared to scanning without tube current modulation, mAs, CTDIvol, DLP and meaneffective dose values were significantly (p < 0.01) reduced by 57.8% (54.1/128 mAs), 57.1% (2.5/5.9 mGy), 55.4% (70.1/156.9 mGy*cm) and 59.9% (2.38/5.9mSv), respectively. Body weight and mAs-values were significantly (p < 0.01)lower for 80 kVp than for 100 kVp and 120 kVp scans and radiation dose wassignificantly (p < 0.01) higher for 120 kVp than for 100 kVp and 80 kVp scans.Image quality between 120 kVp, 100 kVp and 80 kVp scans was not significantlydifferent. CT diagnoses correlated well with clinical findings.Conclusion: Automated tube current modulation combined with low tube voltagesettings significantly reduce radiation exposure and are generally recommendedin pediatric cardiovascular 64-slice CT.

TT TTTu

esd

au

esd

au

esd

au

esd

au

esd

ayy yyy

Scientific Sessions

B 305C D E FA G

10:30 - 12:00 Room G

Head and Neck

SS 1808Old problems: New answersModerators:B.-J. de Bondt; Maastricht/NLS.J. Golding; Oxford/UK

B-717 10:30

First results of flat-panel based volume-CT (fpVCT) of the whole humanskull base and maxillo-facial region and comparison with MSCTS.H. Bartling1, O. Majdani1, I. Madisch1, A. Schlesinger1, C. Dullin2, E. Grabbe2,H. Becker1; 1Hannover/DE, 2Goettingen/DE ([email protected])

Purpose: FpVCT offers a higher contrast resolution then MSCT.Until now, all published comparisons were performed by scanning small, exiledspecimens such as cut temporal bones. However, with the recent availability ofclinical scale fpVCT scanners, it was possible to test the performance on wholehuman skull bases and faces.Methods and Materials: Three embalmed human cadaver heads were scannedin a state of the art MSCT (GE Lightspeed 16) and an experimental fpVCT (GE).MSCT scanning was performed to optimize the spatial resolution (120 kV, 80 mA,0.625 mm collimation) as were reconstruction (9.6 cm rFOV, 512x512 matrix,0.3 mm z-spacing, "boneplus"). The fpVCT scanner consisted of two flat-paneldetectors (10242 elements at 2002µm2) that were mounted side by side on a stand-ard CT gantry. Scan FOV was 33 cm with a z-coverage per rotation of 3.4 cm.Step and shoot scanning (1000 projections, 8 s rotation time, 140 kV, 25 mA) andmodified FDK reconstruction were performed. The image appearance of 21 struc-tures of the skull base and maxillo-facial region were rated on a 4 point scale bytwo independent raters.Results: The comparison revealed a significantly (p=< 0.05) higher average sumscore for fpVCT (60.8) then for MSCT (41.5). Structures such as the stapedialmuscle and the bony walls of the facial nerve canal were much more crisplydisplayed in fpVCT. FpVCT showed more streak artifacts than MSCT.Conclusion: Using fpVCT imaging of the whole human skull base and maxillo-facial region is possible and superior to MSCT.

B-718 10:39

Preclinical evaluation of root-canal-anatomy of human teeth with cone-beam flat-panel volume-CT (FP-VCT)G. Heidrich, F. Hassepass, C. Dullin, T. Attin, E. Grabbe, C. Hannig;Göttingen/DE

Purpose: For successful endodontic diagnostics and therapy, adequate depic-tion of root canal anatomy with multimodality diagnostic imaging is necessary.The aim of the present study was the evaluation of the presentability of the endo-dont with a flat-panel volume-CT device (FD-VCT).Methods and Materials: 13 human teeth were examined with a prototype of aFD-VCT. After acquisition of the data and generation of volume data sets in Vol-ume Rendering Technology (VRT) the findings were compared to X-rays andcross-sectional preparations of the teeth.Results: Anatomical structures of the endodont such as root-, side-canals andcommunications between different root canals as well as denticles could be de-tected precisely with FD-VCT. Also the length of curved root canals was deter-mined accurately. The spatial resolution of the system amounted to 10µm. Only73% of the main root canals, detected with FD-VCT and 87% of the roots couldbe visualized with a conventional dental X-ray. None of the side canals, shownwith FD-VCT were detectable in conventional X-rays. In all cases enamel anddentin of the teeth could be differentiated well. Stored and freshly extracted teethas well as primary and adult teeth showed no differences in image quality.Conclusion: The FD-VCT is an innovative diagnostic procedure in preclinicaland experimental use for non-destructive three-dimensional analysis of teeth.The smallest structures such as side canals can be detected and evaluated dueto the high isotropic spatial resolution, in contrast to conventional X-rays.

B-719 10:48

3D reconstruction-flat-detector dacryocystography: A new view of thelacrimal drainage systemK.E. Wilhelm, S. Lohmeier, H. Strunk, H. Schild; Bonn/DE([email protected])

Purpose: To evaluate the utility of a new diagnostic radiographic technique formorphologic and functional evaluation in the lacrimal outflow system of patientswith suspected nasolacrimal duct obstructions.Methods and Materials: Dacryocystography (DCG) was performed by gentlemanual injection of nonionic water-soluble contrast media (Ultravist 300, Scher-ing, Berlin, Germany) through a micro catheter inserted into the inferior cannalic-ulus. Imaging included digital flat-detector DCG (2 frames/sec) in frontal and lateralview (Alura Xper FD 20 flat detector System, Philips, The Netherlands). Addition-ally, 3D rotation scan was performed using C-arm roll movement at 55 degrees/sec acquiring 120 images. Images were analyzed using functional and morpho-logical criteria based on 3D surfaced shaded and multiplanar reconstructed 2Dimages. Additional radiation exposures were measured and analyzed.Results: In two cases no tear duct obstruction was found. In ten cases a stenosisand in four cases occlusion was found. Soft tissue swelling within the nasolac-rimal duct was the most common cause of obstruction. In three cases additionaldacryocele, respectively nasal pathology was depicted. The radiation dose ap-plied was lower compared to conventional post-DCG computer tomography, pro-viding comparable diagnostic information.Conclusion: 3D- rotation- flat- detector dacryocystography provides detailedimaging of the soft tissue lacrimal outflow system and surrounding bony struc-tures within one diagnostic tour. This new technique combines morphologic andfunctional diagnostic information that improves planning of surgical intervention-al procedures.

B-720 10:57

Three dimensional, non-destructive visualization of vertical root fracturesusing flat-panel volume-CT (FP-VCT)G. Heidrich, C. Dullin, M. Hülsmann, C. Hannig; Göttingen/DE

Purpose: To detect and to visualize radiographically vertical root fractures inextracted teeth with a prototype of a novel, high resolution, three dimensionalflat-panel volume-CT-system (FP-VCT).Methods and Materials: Five teeth with root fillings and clinical symptoms suchas fistulas and isolated periodontal pockets of 8 mm or more were extracted afterdental X-rays indicating lateral or periapical lesions. Vertical root fractures or crackssuspected due to the symptoms and clinical findings were evident after extrac-tion in all cases but fracture lines were not visible on routine dental radiographsacquired before extraction. The extracted teeth were explored with a prototype ofa flat-panel volume-CT (FP-VCT).Results: Using the FP-VCT, in all cases vertical root fractures or crack linescould be detected clearly in different views, depiction-modes and cross-sectionsat a spatial resolution of 140 µm. The evaluation of the fracture lines and teethcould be performed in three dimensional views. The FP-VCT findings were con-firmed by detailed inspection of the extracted teeth.Conclusion: The FP-VCT is an innovative, diagnostic tool for non-destructive,three dimensional evaluation of extracted teeth in pre-clinical and experimentalstudies. The FP-VCT allows precise visualization and evaluation of vertical rootfractures or cracks in extracted teeth. Clinical application of the system may bepossible if technical modifications reduce the exposure dose: the high resolutiondetector-systems of the FP-VCT should be combined with radiation-systems thatcollimate the beam to the area of interest.

B-721 11:06

Flat-panel based volume-CT (fpVCT) improves intraoperative navigationaccuracy: Innovative surgery concepts such as navigated, minimallyinvasive cochleostomy might become possibleS.H. Bartling1, O. Majdani1, J. Graute1, C. Dullin2, T. Lenarz1, H. Becker1;1Hannover/DE, 2Goettingen/DE ([email protected])

Purpose: Multislice-CT based intraoperative navigation is a well establishedmethod in ENT surgery. However, currently the best possible navigation preci-sion (0.7-1 mm) is limited compared with small structures of the lateral skull base.The resolution of the underlying CT dataset could be one significant influentialfactor; therefore we tested whether high resolution imaging provided by flat-pan-el based Volume-CT translates to higher navigation precision.Methods and Materials: An acrylic glass phantom with a rigidly attached refer-ence star providing 40 navigation markers in the form of inverted cones was

Scientific Sessions

306 C D E FBA G

designed. It was scanned in a state of the art MSCT (GE Lightspeed 16) and anexperimental fpVCT (GE). MSCT scanning (~350 µm feature size in plane, ~500 µm in z) and fpVCT scanning (~ 200 µm feature size, isotropic) were opti-mized for high resolution. The navigation system VectorVision2 (BrainLAB) wasused. Six cones were used for registration, seven for measuring the averagetarget registration error (TRE) (5 reiterations).Furthermore, a stereotactic, minimally invasive cochleostomy was tried on fourexcised temporal bones that were scanned in fpVCT.Results: Average TRE using fpVCT was 0.46 (SD 0.22), significantly lower(p < 0.1, two-sided t-test) than using MSCT: 0.82 mm (SD 0.35). Stereotacticcochleostomy was successful in all four temporal bones.Conclusion: High resolution fpVCT imaging improves intraoperative navigationaccuracy. As shown here in ex-vivo experiments, this could be highly beneficialfor skull base surgery, where precise image guided interventions such as cochle-ostomy or stereotactic drilling biopsies could become possible.

B-722 11:15

Contrast-enhanced computed tomography examination may provide non-invasive differentiation between cholesteatoma and granulation tissueA. Trojanowska, P. Trojanowski, E. Czekajska-Chehab, W. Olszanski,W. Golabek, A. Drop; Lublin/PL ([email protected])

Purpose: To determine the role of delayed post-contrast computed tomographyimaging in the differentiation between residual or recurrent cholesteatoma andgranulation tissue in patients who have undergone middle ear surgery.Methods and Materials: Sixteen patients who underwent resection of cholest-eatoma were included in the study. In each case, CT examination of the temporalbone was performed in the following protocol: slice thickness 0.6 mm, overlap0.3 mm, FOV 9.8 mm, matrix 512 x 512, reconstruction kernel - soft tissue, 80 kV,160 mAs. Examination was performed before and after intravenous injection ofiodinated non-ionic contrast medium: 60 ml, flow rate 1.5 ml/s, post-contrast de-lay 70 s. During post processing, the density of middle ear masses was meas-ured in Hounsfield units (HU) with the use of specially designed color-codedmaps. An experienced radiologist evaluated images searching for post-contrastenhancement. Imaging findings were correlated with surgical findings.Results: Based on computer created maps coding HU in colours, it was possibleto observe non-homogenous enhancement of a middle ear mass in twelve cas-es, which were proven to be granulation tissue at surgery. In four cases the mid-dle ear mass showed no significant post-contrast enhancement; these were provento be cholesteatomas.Conclusion: Evaluation and comparison of pre-and post-contrast CT imagescan confirm residual or recurrent cholesteatoma in patients who underwent mid-dle ear surgery by showing non-contrast enhancement. In cases with granulationtissue, delayed post-contrast enhancement of more than 15 HU is typical. Thiswill help in non-invasive differentiation between post-inflammatory and cholest-eatoma lesions.

B-723 11:24

High resolution multislice CT (MSCT) for predicting basal cochlear lengthfor electric-acoustic stimulationM.G. Mack1, O. Adunka2, M. unkelbach1, A. Wetter1, T. Lehnert1, J.O. Balzer1,T.J. Vogl1; 1Frankfurt a. Main/DE, 2Chapel Hill, NC 27599-7070, NC/US([email protected])

Purpose: To assess the feasibility and accuracy of predicting electrode insertiondepth necessary in cochlear implantations (CI) for electric acoustic stimulationby means of pre-operative MSCT.Methods and Materials: 8 fresh human temporal bones were examined withMSCT with a 0.5 mm collimation (increment 0.2 mm). The basal cochlear struc-tures were reconstructed and measured using a 3D-program (vessel view, Sie-mens). Standard cochlear implantations with a free fitting array were performed.A special technique, which allows sectioning of undecalcified bones was used.After embedding, all bones underwent conventional X-ray analysis with furtherinsertion measurements. Pre-implantation HRCT data was compared to X-rayand histology.Results: Pre-implantation MSCT reconstruction was feasible and easy-to-use inevery bone. Linear reconstruction revealed all relevant anatomical details of thecochlea and adjacent structures. Using the described linear reconstruction meth-od, an average basal cochlear length for the first 360° of 20.3 mm (range 18.8 to22.0 mm) was measured. For the 8 specimens, values for the implanted lengthuntil the 360° point ranged from 19.2 to 21.9 mm with an average of 20.3 mm.Pre-implantation CT-based measurements correlated very well with post-implan-tation X-ray data. Comparing the results of the pre-, and post-implantation meas-

urements, a mean failure of 0.3 ± 0.28 mm in predicting the correct length of thebasal turn of the cochlea was observed. The maximum failure was 0.7 mm.Conclusion: MSCT allows the good prediction of the cochlea lengths before CI.This is especially useful when aiming for hearing preservation, where insertiondepths of 360° are necessary.

B-724 11:33

Vascular mapping for reconstructive surgery: Comparison of CTA and DSAM. Lell1, U. Baum1, B. Tomandl2, E. Nkenke1, W. Bautz1; 1Erlangen/DE,2Bremen/DE ([email protected])

Purpose: To compare the potential of computed tomography angiography (CTA)with digital subtraction angiography (DSA) in vascular mapping of the externalcarotid artery (ECA) branches for planning of microvascular reconstructions ofthe mandible with osteomyocutaneous flaps.Methods and Materials: In 35 patients CTA and DSA was performed prior tosurgery. 23 patients had prior surgery or radiation therapy because of cancer ofthe floor of the mouth/oral cavity. Selective common carotid angiograms wereacquired in 2 projection for both sides of the neck. 16- or 64-slice-spiral-CT wasperformed with a dual phase protocol, using the arterial phase images for 3D-CTA reconstruction. Volume rendering (VR) was employed for postprocessing ofCTA data. The detectability of the different ECA branches in CTA and DSA wasevaluated by two examiners.Results: No statistically significant differences between CTA and DSA were foundfor identifying branches relevant for microsurgery. DSA was superior to CTA ifmore peripheral ECA branches were included. All patent parent vessels of thetransplants could be visualized with CTA.Conclusion: CTA proved to be a promising non-invasive alternative to DSA invascular mapping for planning microvascular reconstruction of the mandible.

B-725 11:42 !Reference values of Doppler ultrasound parameters of the thyroid in ahealthy iodine-replete populationT.A.A. Macedo1, M.C. Chammas1, L.P. Souza2, L. Farage2, M.A.P. Castro2,P.T. Jorge2; 1São Paulo/BR, 2Uberlândia/BR ([email protected])

Purpose: In spite the importance, few studies are reported about the normalvalues of Doppler parameters of superior and inferior thyroid arteries. The aim ofthis study was to describe the normal values of Doppler parameters of thyroidarteries in an iodine replete region.Methods and Materials: One hundred and sixty-five individuals were randomlyselected in a community located in southeast Brazil. We took a clinical history ineach subject and determined serum thyrotropin, antiperoxidase antibodies, TRAband thyroid volume by US. Subjects with thyroid disease, and those under 20years of age were excluded. 84 representative subjects (30 men and 54 female)remained. Systolic peak velocity (SPV), resistive index (RI) and pulsatility index(PI) of superior and inferior thyroid arteries were measured by a 5-12 MHz linearprobe.Results: Except for RI, the distribution of all Doppler parameters was not gaus-sian. The median and mean references for SPV, RI and PI were 24.80 cm/s and25.85 cm/s, 0.60 and 0.62, and 0.98 and 1.04, respectively for superior thyroidarteries; these reference values for the inferior thyroid artery were 20.92 cm/sand 21.50 cm/s, 0.57 and 0.57, and 0.84 and 0.88, respectively (p = 0.0004).Females had greater SPV values (p < 0.007). There was a linear correlation be-tween SPV and age (r = 0.41 and p < 0.0001).Conclusion: We have determined reference values of thyroid Doppler parame-ters in our iodine non-deficient population and prepared tables by sex and age.

B-726 11:51

Long-term survival of autologous fat injected in the vocal fold: A clinicaland radiological studyP.R. Biondetti, L.V. Forzenigo, G. Cantarella, R.F. Mazzola, B. Maraschi;Milano/IT ([email protected])

Purpose: To restore glottic competence after paralysis and soft tissue defects,endoscopic vocal fold injection or open surgery are available. Fat is the idealimplant: it is soft, easily available and biocompatible. Histological evaluation andimaging studies show contradictory results about survival rate of fat grafts.Methods and Materials: We prospectively evaluated long term survival of inject-ed fat with CT and MR imaging.10 patients (aged 40-76 years, mean 62) withbreathy dysphonia secondary to laryngeal hemiplegia (7), bilateral scarring (2),and right cordectomy (1), who underwent vocal cord lipoinjection were evaluatedwith CT and MR imaging, 10-28 months post-operatively. A 4-slice multirow CT

TT TTTu

esd

au

esd

au

esd

au

esd

au

esd

ayy yyy

Scientific Sessions

B 307C D E FA G

and a 1.5 Tesla MR imaging unit were used. Imaging was performed without con-trast, with contiguous 2.5 mm thick transaxial CT slices and spin-echo T1 and T2weighted 1.5-4 mm thick transaxial and coronal MR imaging slices. All patientsunderwent clinical evaluation with maximum phonation time (MPT) measurements,GRBAS perceptual evaluations, and Voice Handicap Index (VHI) self-assess-ments.Results: CT/MRI showed persistence of the fat implants in all examined cases.In 6/7 cases CT and MR imaging equally well displayed the injected fat. In 1 caseof bilateral injection the fat was displayed only in one vocal cord. There was goodcorrelation between the amount of cordal fat demonstrated by CT/MRI and clini-cal evaluation.Conclusion: Lipoinjection proved to be a valuable technique for voice rehabilita-tion in glottic incompetence. Imaging studies showed that fat cells survive trans-plantation into the vocal muscle for up to 28 months with good correlation betweenclinical outcome and imaging results.

10:30 - 12:00 Room H


SS 1809aRF ablationModerators:T. de Baère; Villejuif/FRJ. Tacke; Passau/DE

B-727 10:30

Radiofrequency ablation of pulmonary tumors response evaluation(RAPTURE) trial: Final reportR. Lencioni1, L. Crocetti1, R. Cioni1, D.W. Glenn2, W. Lees3, D. Regge4, R. Suh5,T. Helmberger6, C. Bartolozzi1; 1Pisa/IT, 2Sidney/AU, 3London/UK, 4Candiolo/IT,5Los Angeles, CA/US, 6Lübeck/DE ([email protected])

Purpose: To determine the survival outcomes of patients with non-small cell lungcancer (NSCLC) or pulmonary metastases who underwent percutaneous CT-guided radiofrequency (RF) ablation.Methods and Materials: One-hundred-six patients with 186 malignant lung tu-mors ≤ 3.5 cm (mean, 1.7 cm ± 0.8) were enrolled in a prospective, multicenterclinical trial. Diagnoses included NSCLC in 33 patients, colorectal cancer (CRC)metastasis in 53, and metastasis from other primary malignancy in 20. Surgicaloption was excluded in all patients. CT-guided RF ablation was performed byusing 150-200-W generators and multitined expandable electrodes (RITA Medi-cal Systems, Mountain View, CA). Follow-up period ranged 1-27 months (mean,9 months ± 6).Results: Major complications consisted of pneumothorax (n = 27) or pleural ef-fusion (n = 4) requiring treatment, pneumonia (n = 2), and atelectasis (n = 1).Complete ablation of macroscopic tumor as shown by 3-month CT was achievedin 173 of 186 tumors (primary effectiveness rate, 93%). Overall survival was 69%and 49% at 1 and 2 years respectively in patients with NSCLC; 86% and 62% at1 and 2 years in patients with CRC metastases. Cancer-specific survival was91% and 91% at 1 and 2 years in patients with NSCLC; 88% and 72% at 1 and 2years in patients with CRC metastases.Conclusion: CT-guided RF ablation yields high local tumor control rates in pa-tients with pulmonary malignancies, and is associated with acceptable morbidity.Cancer-specific survival outcomes support the use of RF ablation for patientswith NSCLC or CRC metastases who are not candidates for surgery.

B-728 10:39

CT-guided radiofrequency ablation (RFA) for renal cell carcinoma in patientswith solitary kidney after nephrectomyR.-T. Hoffmann1, T. Jakobs1, C. Trumm1, T. Helmberger2, M.F. Reiser1;1München/DE, 2Luebeck/DE ([email protected])

Purpose: The aim of this study was to evaluate the feasibility and effectivenessof RFA in treatment of patients with a solitary kidney suffering from inoperableRCC.Methods and Materials: Within 30 months 14 patients (9 male, 5 female; age63 ± 8 years) with 16 tumors (diameter 1.3 to 4.2 cm, mean 2.9 cm) were treated.All patients had a history of nephrectomy of contralateral kidney. RFA was indi-cated due to inoperability or due to high probability of complete renal failure aftersurgical enucleation of the tumor. If tumor-diameter was larger than 3 cm arterialembolization was performed prior to RFA to reduce heat sink effect and risk of

bleeding. RFA was performed under CT-fluoroscopy using analgosedation andlocal anaesthesia. Follow-up examinations were after 6 and 12 weeks, then eve-ry three months using CT or MR imaging. Definition of success was lack of con-trast enhancement and shrinking of the treated area.Results: Targeting of RCC using CT fluoroscopy was feasible in all patients. Withinthe follow-up period no major complication or tumor relapse occurred. In onepatient a second RCC after was detected and treated with second RF ablation.None of the patients developed renal failure needing hemodialysis. In one of thepatients bleeding into pararenal tissue was noticed, which stopped spontane-ously.Conclusion: RFA for inoperable RCC in a solitary kidney is a very effective ther-apeutic option due to the small complication rate, no deterioration of the renalfunction and excellent tumor control.

B-729 10:48

Follow-up of liver tumors treated with RF ablation: Can a volumetricnecrosis/tumor quotient predict local tumor control?H. Kuehl, B. Elo, T. Boes, J. Stattaus, J. Barkhausen, M. Forsting; Essen/DE([email protected])

Purpose: Safety margins are a prerequisite to avoid local tumor recurrence afterRF ablation of liver tumors. The aim was to determine the prognostic value of avolumetric necrosis/tumor quotient.Methods and Materials: Twenty-nine colorectal cancer (CRC) metastases and24 hepatocellular cancer (HCC) nodules were treated by radiofrequency abla-tion. Volumetric measurements of the tumor volumes were performed prior toablation and the necrosis volume was measured after ablation based on con-trast-enhanced CT data sets. The quotient of necrosis and tumor volumes wascalculated for all lesions. Follow-up exams were performed up to 1 year afterablation.Results: The CRC metastases and HCC nodules showed a mean tumor volumeof 12.8 ml and 11.5 ml. The mean ablation volumes were 37.6 ml and 29.5 ml,respectively. This resulted in post-/pre-ablation quotient of 5.08 ± 3.8 for metsand 4.16 ± 2.7 for HCC. The follow-up revealed local tumor recurrence in 16 of 29(55%) metastases and 11 of 24 HCC (46%). The necrosis/tumor quotient wassignificantly smaller in local recurrence with 3.64 compared to 5.71 in lesionsfree of tumor (p < 0.002). CRC metastases, but not HCC nodules showed a cor-relation between initial tumor volume and local outcome.Conclusion: Although small lesions were treated and the area of necrosis wasmore than four times larger than the tumor volumes, local tumor control could beachieved only in the half of patients. However, the volumetric necrosis/tumor quo-tient can identify patients at risk for local tumor recurrence.

B-730 10:57

Percutaneous CT-guided radiofrequency ablation of osteoid osteoma withmultitined expandable electrodeR. Cioni, L. Crocetti, R. Lencioni, V. Zampa, C. Cappelli, I. Bargellini,C. Della Pina, V. Consoli, C. Bartolozzi; Pisa/IT ([email protected])

Purpose: To determine the feasibility, safety and effectiveness of percutaneousCT-guided radiofrequency (RF) ablation of osteoid osteoma (OO) with a multi-tined expandable electrode.Methods and Materials: Eleven patients (male, age 14-58 years, median age 22years) with single OO were enrolled in a prospective, single center, single-armclinical trial. Diagnosis was obtained by clinical and imaging findings in 10 pa-tients and by CT-guided biopsy in 1 patient. CT-guided RF ablation was performedunder conscious sedation by using a 150-200-W generator and expandable mul-titined electrode (RITA Medical Systems, Mountain View, CA). Follow-up periodranged 3-22 months (mean 12 months ± 7) and included MR examinations andclinical visits performed 6 months after the procedure and at 12-month intervalsthereafter.Results: RF ablation was technically feasible in all 11 patients (technical suc-cess 100%) without need of surgical access. No major complications occurred -in particular, no skin burns were observed. Prompt pain relief and complete abla-tion of the nidus, as shown by 6-month MR, was achieved in 10 of 11 osteomas(primary effectiveness rate 91%). In one patient, recurrence of pain and persist-ent contrast uptake of the nidus at MR were present. The OO was effectively re-treated by means of RF ablation. No recurrences were observed on follow-up.Conclusion: Percutaneous CT-guided RF ablation of OO performed with a mul-titined expandable electrode yields high effectiveness rates and has a favourablesafety profile.

Scientific Sessions

308 C D E FBA G

B-731 11:06

Radiofrequency ablation (RFA) in the treatment of osteoid osteoma: Mid-term resultsR.-T. Hoffmann1, T. Jakobs1, C. Trumm1, T. Helmberger2, M.F. Reiser1;1München/DE, 2Lübeck/DE ([email protected])

Purpose: The aim of our study was to determine the efficacy of RFA in the treat-ment of osteoid osteoma regarding complication rate and duration of pain relief.Methods and Materials: Within 42 months, 22 patients (12 male, 10 female, age9 to 45 years) suffering from osteoid osteoma (OO) were treated. In 8 of 22 pa-tients the OO were localized in the lumbar (5 of 8) or thoracic spine (3 of 8),whereas the other OO were located in the femur (5 of 14), the acetabulum (2 of14) or other long bones (radius, tibia). In children RFA was performed using gen-eral anaesthesia, in adults analgo-sedation and local anaesthesia was preferred.Ablation periods of 5 to 8 minutes at a temperature of 90oC were used. Cooling ofthe skin was performed in osteoid osteoma next to the skin (tibia, radius) and anadditional needle was inserted to do saline flushing if the OO was next to neuralstructures (spine). Primary success rate, complications, disease-free interval andfollow-up time were evaluated.Results: Within the observation period (3-42 months) all of our patients weresuccessfully treated and had no more complaints. After the first ablation 19 of 22patients were free of pain, while in 3 patients the ablation was repeated to obtaincomplete response. No major complications occurred. In 2 patients minor compli-cations (1 hematoma, 1 first degree skin burn) were observed.Conclusion: Radiofrequency ablation is a highly effective, efficient, minimallyinvasive and safe method of treating osteoid osteoma.

B-732 11:15 !Radiofrequency ablation of hepatomas: The Mount Elizabeth (Singapore)experienceP.Y.T. Goh, T.-N. Lau, R.K. Kwok; Singapore/SG ([email protected])

Purpose: The aim of this paper is to review our treatment protocols for hepato-mas and to highlight the efficacy of combined trans-arterial chemo-embolization(TACE) followed by RFA in the treatment of large lesions beyond 10 cm in diam-eter.Methods and Materials: Over 250 sessions of RF ablation have been performedfor hepatomas in our centre since 2001. Our protocol is as follows: < 2 cm - RFAalone; 2-3 cm - RFA with overlapping ablation; 3-10 cm - TACE followed by RFA(± repeat sessions); > 10 cm: - multiple sessions of TACE followed by RFA; im-paired liver function - bland embolization with lipiodol ± gelfoam followed by RFA.For lesions above 10 cm, treatment protocols in many centres advocate TACEalone. However, we are finding that the combination of TACE followed immediate-ly by RFA for lesions above 10 cm has favorable outcomes. Repeat sessions ofcombined TACE and RFA are usually performed for these lesions at 3-6 weekintervals, depending on the appearance of the lesion at follow-up, correlated withtumour markers.Results: Overall, results are extremely favorable. For large lesions, a regime ofmultiple sessions of chemoembolization and RFA yields good results with com-plete tumour eradication in many lesions, even those > 13 cm in diameter. Com-plication rates are low and post-procedure recovery is usually uneventful.Conclusion: Multiple sessions of TACE followed by RFA yield encouraging re-sults for large hepatomas. Further studies are strongly recommended to estab-lish the limits of this regime and its survival rates.

B-733 11:24

Assessment of liver tissue following radiofrequency ablation: Anexperimental study with different imaging proceduresG. Antoch, F.M. Vogt, P. Veit, N. Blechschmid, A. Bockisch, H. Kühl; Essen/DE([email protected])

Purpose: Rim-like contrast enhancement on morphologic imaging and increasedtracer uptake on FDG-PET may hamper differentiation of residual tumor fromregenerative tissue following radiofrequency ablation (RF-ablation). The purposeof this study is to assess the typical appearance of normal liver tissue followingRF-ablation when imaged with FDG-PET/CT and contrast-enhanced CT and MRimaging.Methods and Materials: Two lesions of 3 cm each were created by RF-ablationin normal liver parenchyma of 21 fully anesthesized mini-pigs. Post-intervention-al follow-up was performed by FDG-PET/CT as well as contrast-enhanced (CE)MR imaging and CE CT. Imaging was performed 1 day, 3 days, 10 days, 1, 2, 3and 6 months post-intervention. All imaging procedures were evaluated qualita-tively and quantitatively for areas of increased contrast enhancement and re-

gions of elevated tracer uptake. Imaging findings were compared with histopa-thology.Results: One day following RF-ablation no increase in FDG uptake was visiblewhile elevated contrast enhancement was noticed in the periphery of the necro-sis with all morphologic imaging procedures. However, on further follow-up, anarea of rim-like increase of FDG-uptake was detected surrounding the necrosiswith PET/CT. Both, increased glucose metabolism as well as elevated contrastenhancement were detectable qualitatively and quantitatively for 6 months post-intervention. Histology showed an hemorrhagic border converting into a regener-ation capsule.Conclusion: FDG-PET and PET/CT may be of benefit over morphologic imagingprocedures when assessing patients following RF-ablation for residual tumor, ifperformed within one day of intervention. Follow-up at a later time will be limitedby visualization of peripheral hyperperfusion and tissue regeneration on func-tional and morphologic data sets.

B-734 11:33

Comparative study of wet electrode-mediated and hypertonic saline-enhanced radiofrequency ablation in in vivo porcine liverJ. Lee, J. Han, J. Lee, S. Kim, K. Sohn, J. Chang, B. Choi; Seoul/KR([email protected])

Purpose: The aim of this study was to compare the in-vivo efficiency of wet radi-ofrequency ablation (RFA) with that of hypertonic saline (HS)-enhanced RFA forcreating larger coagulation necrosis.Methods and Materials: Twenty six ablation zones were created in seven pigswith five different regimens: Group A, RFA with 6 mL of 36% HS preinjection;group B, wet RFA with infusion of 14.6% HS at 1 mL/min; group C, dry RFA;group D, 6 mL of 36% HS injection; group E, 12 mL of 14.6% HS infusion. RFenergy was applied to the liver for 12 minutes using a 200 W generator and acooled-wet electrode. The dimensions of the ablation zones and delivered ener-gy during RFA were then compared between the groups.Results: The mean delivered energy was higher in group B (104.6 kJ) than ingroup A (97.2 kJ), which was higher than in group C (49.5 kJ) (p < 0.05). WetRFA (55.0 ± 18.1 cm3) created greater coagulation necrosis than HS-enhancedRFA (31.7 ± 8.8 cm3) or dry RFA (11.1 ± 3.5 cm3)(p < 0.05). An infusion of 36%or 14.6% HS into the liver induced coagulation necrosis measuring 2.3 ± 2.6 cm3

and 0.9 ± 0.9 cm3, respectively.Conclusion: Wet RFA is more efficient in creating a large coagulation necrosisthan either the dry RFA or HS-enhanced RFA.

B-735 11:42

MR-guided galvano-therapy for the treatment of localized prostate cancer:Results of a phase-III trialT.J. Vogl1, F. Mayer2, H. Mayer2, S. Zangos1; 1Frankfurt a. Main/DE,2Regensburg/DE ([email protected])

Purpose: To evaluate a novel technique of MR-guided galvano-therapy (MR-GG)(electro-chemical therapy (ECT)) for the treatment of prostate cancer in a phase-III trial.Methods and Materials: Thirty three patients with histologically proven prostatecancer were prospectively treated with MR-guided galvano-therapy. After trans-gluteal puncturing of the prostate, under local anesthesia two MR compatibleelectrodes were positioned under MR guidance in the periphery of the right andleft lobes of the prostate. The patients were treated three times at a weekly timeinterval. A total of 350 Coulomb was applied to the localized cancer and theprostate gland. Follow-up was performed after three, six and twelve months, re-spectively, using laboratory data (PSA) and endorectal MR imaging.Results: All patients tolerated the MR-GG well without any major side effects.Twelve of the thirty three patients had some reversible urinating problems. In thefollow-up MR evaluation 12 months post therapy twenty four (72.7%) of the pa-tients were classified as "stable" with a constant lesion size. In five patients (15.2%)MR imaging showed a homogenization of the neoplastic area, in two patients(6%) we found a size reduction of 30%, and two patients (6%) presented anincrease in size of 20% and 30%, respectively. The PSA values showed a de-crease in the twelve-month control from an average PSA of 13.5 ng/ml to 5.3 ng/ml, which is equivalent to a reduction of 60.8% and statistically significant(p > 0.01).Conclusion: MR-guided galvano-therapy possibly results in local control of lim-ited prostatic carcinoma with a concomitant reduction in the PSA value.

TT TTTu

esd

au

esd

au

esd

au

esd

au

esd

ayy yyy

Scientific Sessions

B 309C D E FA G

B-736 11:51

Can the response to radiofrequency ablation of hepatocarcinoma beadequately measured by means of contrast-enhanced US and dynamic MRimaging?A. Lüttich, J. Puig, J. Martín, A. Darnell, A. Malet, M. Veintemillas; Sabadell/ES([email protected])

Purpose: To evaluate the reliability of contrast-enhanced US and dynamic MRimaging in determining the response to percutaneous radiofrequency ablation ofhepatocarcinoma.Methods and Materials: Fifteen patients with hepatocarcinoma treated by radi-ofrequency ablation underwent contrast-enhanced (Sulfur Hexafluoride (SF6)) andgadolinium-enhanced MR imaging four weeks after treatment. Any focus of in-tense contrast uptake within the tumor in the arterial phase was considered aspersistence of the tumor. All patients were biopsied after contrast-enhanced US.Results: Biopsy showed persistence of the tumor in ten hepatocarcinomas andcomplete ablation in five. Both contrast-enhanced US and MR imaging foundcomplete ablation in thirteen hepatocarcinomas and persistent tumor in two, withone false positive in each technique. Sensitivity, specificity, PPV and NPV were10%, 80%, 50% and 30% for both techniques.Conclusion: Biopsy to measure the response of hepatocarcinoma to radiofre-quency ablation is rarely performed and histopathological criteria are uncertain.There is a discrepancy between the response to radiofrequency ablation of hepa-tocarcinoma measured by imaging techniques and histopathological study.

10:30 - 12:00 Room I


SS 1809bTreatment of aortic diseaseModerators:S. Fabiano; Rome/ITK.A. Hausegger; Klagenfurt/AT

B-737 10:30

Expanding the indications for thoracic aortic stent grafting by means ofextra-anatomic bypass. Our experience in thirteen patientsG.J. Munneke, J. Raja, A.-M. Belli, M. Thompson, R. Morgan; London/UK([email protected])

Purpose: Endovascular stent grafting has a proven role in treating diseases ofthe thoracic aorta. The mortality and morbidity associated with the procedure isfar lower than open surgery. However many patients are unsuitable for treatmentdue to the absence of an adequate proximal or distal landing zone. We performextra anatomic bypass grafting of the arch or mesenteric vessels to facilitatestenting in such cases.Methods and Materials: Thirteen patients were treated with bypass prior to tho-racic stent grafting. Aortic pathologies included aneurysms affecting the arch anddescending aorta, aneurysmal type B dissections, thoraco-abdominal aneurysmand one aneurysm of an aberrant right subclavian artery. Patients underwent acombination of grafting from the ascending aorta, carotid-carotid artery bypass,carotid-subclavian artery bypass and mesenteric artery bypass. Four patientswere treated with Medtronic Talent stents, one patient with a Gore TAG stent andeight patients with the new Medtronic Valiant stent graft.Results: Initial technical success was achieved in all patients. One procedurewas complicated by rupture of the iliac artery. One patient died in the periopera-tive period giving a 30 day mortality of 7.7%. There were two small endoleaks(15%) on the immediate post operative CT. However there were no endoleakspresent at three months (0%). At a mean follow-up of eight months there havebeen no further deaths, ruptures or complications related to the bypass grafts.Conclusion: Extra anatomic bypass is a valid and safe technique to expand theproportion of patients suitable for endovascular stenting of the thoracic aorta.

B-738 10:39

MR-guided endovascular aortic stent-graft placement in an animal model ofaortic dissectionH.H. Quick, H. Eggebrecht, H. Kühl, G.M. Kaiser, S. Aker, M.O. Zenge,F. Breuckmann, M.E. Ladd; Essen/DE ([email protected])

Purpose: To evaluate real-time MR-guided aortic stent-graft placement in a pigmodel of experimental aortic dissection.

Methods and Materials: Descending thoracic aorta dissections were experimen-tally created with a catheter-based technique under X-ray in 8 domestic pigsweighing 63-98 kg. All subsequent MR imaging was performed on a 1.5-T whole-body MR imaging scanner (Avanto, Siemens, Erlangen, Germany). Pre-interven-tional dissection evaluation was commenced with ECG-gated TrueFISP imagingand with contrast enhanced 3D FLASH MRA. A self-expandable, nitinol thoracicaortic stent-graft device (GoreTAG, Gore Inc., Flagstaff, AZ) was chosen for theexperiments. For implantation, the delivery system with mounted stent-graft wasadvanced from the iliac artery to the thoracic aorta under real-time MR imagingfluoroscopy featuring an interactive radial TrueFISP sequence providing 7fps.Images were projected onto an in-room 60" high-contrast projection screen (MR-Iscreen, MR-Innovation GmbH, Essen, Germany). Following stent-graft place-ment, the pre-interventional MR imaging protocol was repeated to evaluateprocedure success.Results: Descending aortic dissections were successfully created in all 8 pigs.Real-time MR imaging permitted: 1) successful and safe device navigation withinthe true lumen from the iliac arteries to the thoracic aorta, 2) precise positioningof the stent-graft, 3) successful stent-graft deployment, and 4) safe catheter with-drawal. Post-interventional MR imaging demonstrated correct position of the stent-graft and complete obliteration of the false lumen which was confirmed by necropsy.Conclusion: Endovascular stent-graft placement for experimental aortic dissec-tion using commercially available catheter devices is feasible and safe underreal-time MR imaging-guidance. Beyond detailed pre-interventional evaluationand procedure guidance, MR imaging permits immediate post-interventional an-atomic and functional evaluation of procedure success.

B-739 10:48

Midterm results of transcaval approach in treatment of type II endoleak:One year follow-upG. Mansueto, D. Cenzi, M. D'Onofrio, G. Carbognin, R. Pozzi Mucelli; Verona/IT([email protected])

Purpose: To assess feasibility and clinical success in treatment of type II en-doleaks using transcaval approach.Methods and Materials: From September 2004 to March 2005 we treated 11patients with type II endoleaks after EVAR. Via a venous access, a curved guid-ing cannula was advanced into the caval vein to the site of the aneurysm sac. Aflexible needle in the cannula was oriented towards the sac and embolizationinside the sac was performed with transcaval puncture. Systemic and intrasacpressures were recorded before and after the embolization. CT scans were per-formed at 24 hours, 30 days and 1 year after the procedure: Both the presence ofleaks and the changes in sac diameter were evaluated. Technical success wasidentified by feasibility of the procedure and clinical success with no evidence ofleaks at follow-up.Results: Transcaval puncture was feasible in 10/12 procedures (technical suc-cess 83%). The mean systemic pressure was 150/86 mmHg (range 117/69-176/100). The mean intrasac pressure before embolization was 79 mmHg (range 140-40). After embolization, the mean intrasac pressure was 10 mmHg (range 5-39)in 10 patients, while in one it increased. At the 24 hours CT scan, gas bubblesand stable contrast medium with coils inside the sac were seen in 10 Patients.10/11 patients were successfully treated; at 1 year CT scan, a reduction in diam-eter of the sac and no leak recurrence were observed (clinical success 91%).Conclusion: Transcaval approach is feasible for type II endoleak treatment. Highclinical success rate makes it an alternative to direct translumbar puncture.

B-740 10:57

A percutaneously implantable telemetric pressure sensor to detectendoleaks after endovascular abdominal aortic aneurysm repair (EVAR)F. Springer, R. Schlierf, J. Pfeffer, J. Tomko, A.H. Mahnken, U. Schnakenberg,T. Schmitz-Rode; Aachen/DE ([email protected])

Purpose: To evaluate whether endoleaks, simulated in an in-vitro model of anabdominal aortic aneurysm (AAA), can be detected by a percutaneously implant-able telemetric pressure sensor (TPS).Methods and Materials: An elastic silicone AAA model including a deployedendograft was connected to a pulsatile circulation with physiologic flow and pres-sure rates. Different types (I-III) and sizes (3-11 Fr) of endoleaks were created.Before and after generating endoleaks aneurysm sac pressure (ASP) was tele-metrically obtained with our newly designed TPS and simultaneously measuredat several positions with 3 Fr. wire mounted pressure sensors (WPS) as a refer-ence. Measurements were performed in fluid filled and in partly/fully thrombosedaneurysm sacs. ASP readings detected by TPS and WPS were compared toeach other and to the intra-aortic pressure (IAP).

Scientific Sessions

310 C D E FBA G

Results: Both telemetrically and wire based obtained ASP measurements showeda significant increase (p < 0.05) after generating a type I endoleak. For a fluidfilled aneurysm sac the exemplary results are averaged over all investigated siz-es and positions. With WPS systolic ASP shown as the percentage of IAP in-creased from 10.1 ± 1.1 to 72.4 ± 1.5 (TPS: 14.6 ± 4.2 to 77.0 ± 2.2), diastolicASP from 13.6 ± 2.2 to 111.7 ± 3.0 (TPS: 17.3 ± 7.5 to 113.8 ± 4.6), mean ASPfrom 11.6 ± 1.5 to 93.6 ± 0.7 (TPS: 16.2 ± 5.8 to 98.8 ± 4.5) respectively.Conclusion: Type I endoleaks of all investigated sizes cause an increase in ASP,high enough to be telemetrically detected at all investigated positions even in athrombosed aneurysm sac. Therefore a telemetric pressure sensor might be in-tegrated into the EVAR procedure and become a routinely used device forfollow-up.

B-741 11:06

Eight-year experience in the percutaneous treatment of abdominal aorticaneurysms: Results and technical considerationsA. Bruni, F. Fanelli, F.M. Salvatori, M. Corona, G. Luppi, E. Boatta, V. Dominelli,A. Pucci, R. Passariello; Rome/IT ([email protected])

Purpose: To report our eight-year experience using bifurcated endoprosthesis inthe treatment of abdominal aortic aneurysms (AAA).Methods and Materials: From September 1998, 182 patients with AAA (mean50.2 mm) were treated using different type of stent-grafts: 25 Vanguard II (Bos-ton-Scientific), 119 Excluder (WL-Gore), 24 Talent (World-Medical), 2 Anaconda(Sulzer-Medical), 1 Quantum (Cordis), 2 Endologix (Endologix), and 9 Zenith(Cook). Procedures were performed in the angio-suite (81) and operating room(101) with patients under general, epidural or local anesthesia.Results: Technical success with complete exclusion of the aneurysm sac wasachieved in 181/182 patients, while in 1 case an immediate conversion was nec-essary. After a follow-up ranging from 1 to 96 months, 163 patients (90.1 %) arealive while 19 died. A reduction of the aneurysm sac (mean value 2.3 mm) wasobserved in 68 patients (37.5%). The sac diameter remained stable in 87 pa-tients (48.1%), and increased in 26 patients (14.3%). Endoleak was detected in28 patients (15.4 %): 3 type I, 22 type II, 1 type III, 2 type IV. Two late conversionswere performed after 22 and 31 months respectively for multiple fractures of thestent-graft and for a persistent type II endoleak. In one case (0.5 %) an occlusionof the right renal artery was observed due to stent-graft migration. Limb obstruc-tion occurred in 5 patients (2.7 %) solved with a fem-fem by-pass.Conclusion: Percutaneous treatment of AAA seems to be a feasible techniquein selective patients only. An accurate follow-up is mandatory to prevent andmanage complications.

B-742 11:15

Endovascular abdominal aortic aneurysm repair and renal dysfunction:Comparison between suprarenal and infrarenal fixation of endograftsR. Iezzi, A.R. Cotroneo, A. Pierro, D. Giancristofaro, M. Santoro, M.L. Storto;Chieti/IT ([email protected])

Purpose: To compare the impact of suprarenal and infrarenal fixation of endog-rafts on renal artery complications and late postoperative renal function.Methods and Materials: A retrospective analysis of the data of 102 patients withunruptured infrarenal abdominal aortic aneurysms who underwent endovascularrepair (EVAR) was performed: 60 patients were treated with a suprarenal fixationendograft (Cook Zenith, Medtronic Talent) (group A) whereas in 42 patients aninfrarenal fixation device (Gore Excluder, Medtronic AneuRX) (Group B) wasimplanted. Baseline patient demographic data, pre- and post-procedural renalfunction (CrCl: creatinine clearance), incidence of renal artery (steno-obstructivedisease) and procedure-related (proximal type-I endoleak and migration rates)complications, detected on post-operative surveillance performed with CT-angi-ography (1, 3, 6 and 12 months) were collected and compared.Results: No significant differences were found between the two groups in termsof demographic data and pre-operative risk factors. Proximal neck length wassignificantly shorter in group A (2.14 ± 0.84) than in group B (3.41 ± 1.21). CrCldecreased during the follow-up period in both groups (Group A: 64.24 ml/min to60.66 ml/min; Group B: 72.41 ml/min to 68.73 ml/min) without statistically signif-icant differences in terms of rate of CrCl decrease during the follow-up period aswell as renal artery and procedure-related complications.Conclusion: In our experience, suprarenal fixation, with proximal extent of theproximal anchoring zone, allowed us to expand the inclusion criteria for EVARwithout increasing renal artery and procedure-related complication rates as wellas the likelihood of postoperative renal impairment.

B-743 11:24

Endovascular treatment of thoracic aorta aneurysms and dissections aftertransposition of supra-aortic trunksR. Rostagno, R. Llorens, T. Zander, I. Zerolo, S. Baldi, E. Sanabria, M. Maynar;Santa Cruz de Tenerife/ES ([email protected])

Purpose: To report our experience with a combined repair of aortic arch aneu-rysms or dissections by sequential transposition of the supra-aortic trunks andendovascular stent-graft placement in cases without a proximal neck.Methods and Materials: Nine patients were selected for transposition of supra-aortic trunks previous to endovascular treatment. The indications were: thoracicaneurysms in 6 cases and dissections in 3. TST was performed using minithora-cotomy without extracorporeal circulation. Endovascular procedures were per-formed in a second operative time in 8 cases. Long term follow-up was plannedat 1, 3, 6 months and yearly.Results: Technical success was achieved in all but one cases, who died aftersurgery due to major stroke. Aneurysms were excluded and proximal entry oc-cluded in all cases. Early complications were: 1 acute limb ischemia, 1 minorstroke and 1 major stroke with death after open surgery. No complications wereobserved in the follow-up.Conclusion: Combined repair of aortic arch aneurysms or dissections by se-quential TST and endovascular stent-graft placement is a safe and effective treat-ment option. This technique avoids complications of descending aorta opensurgery and enlarges endovascular indications.

B-744 11:33

Endovascular repair of inflammatory aortic aneurysms: Long-term resultsS. Puchner, R.A. Bucek, C. Loewe, T. Hoelzenbein, J. Lammer; Vienna/AT([email protected])

Purpose: To report the long term follow-up results of endovascular aneurysmrepair (EVAR) of inflammatory aortic aneurysms (IAA).Methods and Materials: In a retrospective cohort study we identified 8 patientssuffering from IAA based on findings of the pre-interventional computed tomog-raphy angiography (CTA) who underwent EVAR. Primary and follow-up imageswere reviewed by two observers for changes of the aneurysm sac diameter, ofthe periaortic fibrosis (PAF), of potential renal impairment and the procedure re-lated and long term complications.Results: EVAR was successfully completed in all patients. 5 patients had a min-imum follow-up period of 36 months (median 65 months). The maximum diame-ter of their aneurysm sac decreased from median 57 mm (47-95) to median 45 mm(36-60), resulting in a median relative regression of 21.1% (4.8-63.1). PAF re-gressed in all of these patients, prior hydronephrosis was present in two patients,remaining unchanged in one of them and disappearing in the other one. In twopatients a type 3 endoleak was detected and overstented during the follow-upperiod.Conclusion: EVAR of IAA excludes the aneurysm and seems to reduce the sizeof the aneurysm sac and the extent of PAF with an acceptable peri-interventionaland long term morbidity.

B-745 11:42

Fenestration as therapeutic method in treatment of abdominal aorticbranches malperfusion in course of chronic descending aorta dissectionM. Januszewicz, K. Milczarek, J. Szmidt, Z. Galazka, M. Gola, O. Rowinski;Warsaw/PL ([email protected])

Purpose: To evaluate results of endovascular treatment with different techniquesof fenestration in hypoperfusion of abdominal aorta branches caused by descend-ing aorta dissection.Methods and Materials: Between 2002-2005.11 patients (2 women, 9 men, meanage 61.3 y) with chronic descending aorta dissection (2 after type A dissectiontreated with Bental technique, 9 with type B dissection treated previously withthoracic stent-grafts) were treated because of symptomatic visceral and iliacmalperfusion. Renal ischemia was a cause of 5 interventions, intestinal ischemiawas an indication to treatment in 4 patients and lower limb ischemia in 2 patients.Patients were treated with balloon fenestrations (3), scissors-technique fenestra-tions (3) and with stent implantation (5 including 3 renal stents, 1 coeliac trunk,and 1 iliac stent). 16-row computed tomography was a diagnostic and controlmethod in all cases. The method of treatment was chosen depending on Williamsscale of malperfusion.Results: All 11 patients were successfully treated for ischemia with no earlycomplications. In 2 patients reintervention (accessory fenestration) was neces-sary in a 3-6 month period.

TT TTTu

esd

au

esd

au

esd

au

esd

au

esd

ayy yyy

Scientific Sessions

B 311C D E FA G

Conclusion: Endovascular fenestration is a safe and effective treatment in vis-ceral malperfusion complicating chronic aortic dissection.

B-746 11:51

Endovascular stent grafting for complicated type B aortic dissectionG.J. Munneke, M. Thompson, A.-M. Belli, R. Morgan; London/UK([email protected])

Purpose: Open surgery is associated with considerable morbidity and mortalityin patients with complicated type B aortic dissection. We report our 5 year expe-rience for the use of thoracic aortic stent grafts in these patients.Methods and Materials: Thoracic stent grafts were deployed in 35 patients (28men) with acute complicated type B aortic dissections (rupture, end organ is-chaemia or continued pain) and chronic type B dissections which had undergoneaneurysmal dilatation. The patients were followed up with computed tomographyat 1.3 and 6 months post stenting and then yearly.Results: Initial technical success was 97%. One patient required a further proce-dure on the initial admission to seal other entry sites. Three patients died in thepost operative period giving a 30 day mortality of 8.5%. In two patients (6%)proximal (type A) extension of the dissection occurred subsequent to stenting.No patient developed post operative paraplegia. Follow-up averaged 14 months(range 1-57 months). In that time the false lumen failed to thrombosis or wasreperfused in six patients (17%). There were two late deaths which were unrelat-ed to the procedure (6%).Conclusion: The use of stent grafts in selected patients with type B dissectionprovides an attractive alternative treatment option. The procedure can be accom-plished with acceptable initial morbidity and mortality when compared to surgery.However long term follow-up is essential to monitor the durability of these devic-es.

10:30 - 12:00 Room K

Genitourinary

SS 1807Kidneys, including renal transplantsModerators:T. El Diasty; Mansoura/E.G.M. Memarsadeghi; Vienna/AT

B-747 10:30

Comparison of 2D-SR-turboFLASH with and without parallel imaging with2D-SR-trueFISP for semiquantitative assessment of first pass renalperfusion at 1.5 TH.J. Michaely1, H. Kramer1, N. Oesingmann2, K.-P. Lodemann3, M.F. Reiser1,S.O. Schönberg1; 1Munich/DE, 2Erlangen/DE, 3Konstanz/DE([email protected])

Purpose: To evaluate the optimal technique for first pass renal perfusion imagingat 1.5 T.Methods and Materials: 15 healthy male volunteers (25-38 years) underwent aprospective intra-individual comparison study of renal perfusion measurementon a 1.5 T MR-scanner (Avanto, Siemens). Each volunteer underwent three ex-ams on three different days with three different sequences after administration of7 ml Gd-BOPTA (MultiHance®, Bracco-Altana Pharma): Saturation-recovery (SR)turboFLASH without parallel imaging (TFl), SR-trueFISP (TFi) and SR-turbo-FLASH with parallel imaging (TFl-PI). The temporal resolution of the sequenceswas 4 slices/s (TFl, TFi) and 6 slices/s (TFl-PI). Field of view (450x373 mm²) andmatrix (192x134) were kept constant for all sequences. Evaluation comprised thesemiquantitative assessment of first pass perfusion parameters mean transit time(MTT), time to peak signal intensity (TTP), maximal upslope (MUS) and maximalsignal intensity (MSI). In addition the baseline and peak signal-to-noise ratio (SNR)were assessed as well as the image quality which was rated by two radiologistsin consensus.Results: The MSI and the MUS were significantly lower with TFi (272 A.U./42 A.U./s) than with TFl (400 A.U./64 A.U./s) or TFl-PI (440 A.U./70 A.U./s). In contrastthe MTT (14.8s) and TTP (11.1s) were independent of the technique used. Theimage quality was rated best for TFl. TFi revealed the highest baseline SNR butthe relative SNR increase after contrast was lowest compared to the TFL andTFL-PI.Conclusion: In conclusion, it seems that at 1.5 T, TFl without parallel imagingrepresents the best technique for first pass renal perfusion imaging.

B-748 10:39

Diffusion weighted imaging in the evaluation of renal damage: A preliminaryreportS.F. Carbone, E. Gaggiotti, M.A. Mazzei, F. Mazzei, V. Ricci, L. Volterrani;Siena/IT ([email protected])

Purpose: Magnetic resonance diffusion weighted imaging (MR-DWI) is useful toassess proton motion by computation of an apparent diffusion coefficient (ADC).This property could be used to assess renal damage, with special regard to uni-lateral dysfunction. The aim of this study is to estimate the correlation betweenADC and stage of chronic renal failure.Methods and Materials: Twenty patients (mean age 51.6 years, range 38-66years) underwent MR examination on a 1.5 T system; in 10 cases there was ahistory of hypertension, in 3 diabetes, in 3 obstructive uropathy, Takayasu's dis-ease and nephrovascular hypertension respectively. Four subjects without knownrenal pathology were used as a normal reference. Glomerular filtration rate (GFR)assessed by Cockroft-Gault's equation was used as a functional marker. Imagingprotocol consisted of the following sequences: Fast-field echo T1-weighted, TSET2-weighted and Single-Shot DWI echo-planar sequence (b-value 600 sec/mm2),with application of SENSitivity encoding factor of 2, by using a four channel phased-array body coil. In five patients SS-DWI sequence was repeated after iv adminis-tration of 1 mg of furosemide. The ADC was correlated with the GFR by Pearson'stest.Results: ADC was 2.37 ± 0.25*10-3 mm2/sec in subjects with normal GFR and2.05 ± 0.33*10-3 mm2/sec (p < 0.05) in subjects with altered GFR; a significantdifference was found between stage III and IV (p < 0.01). A good correlation wasfound between GFR and ADC (r = 0.76; p < 0.01), with no significant modifica-tion after furosemide administration.Conclusion: DWI is a feasible MR technique to assess intrarenal pathology. Oth-er studies need to evaluate this finding on a more extensive population.

B-749 10:48

Arrival time parametric imaging: A new ultrasound technique forquantifying perfusion of kidney graftsT. Fischer, A. Thomas, J. Rudolph, B. Hamm, S. Filimonow; Berlin/DE([email protected])

Purpose: Evaluation of a new standardized ultrasound (US)technique for diag-nosis of acute rejection of kidney grafts.Methods and Materials: 22 kidney recipients underwent US examination follow-ing administration of 1.6 ml US contrast medium (USCM, SonoVue™) 6 daysafter transplantation. The examinations were performed on an Aplio US system(Toshiba™). The difference in time from the first increase in signal intensity be-tween the renal artery and renal cortex was determined. Subsequently the tem-poral course of contrast enhancement in the area of the renal artery following thefirst peak was recorded over 10 sec and depicted in color using a Windows-based software. The resulting color-time images were evaluated by three readerswho rated the images on an analogue scale from 1 (normal) to 5 (abnormal).Results: Twelve of the 22 patients had an uneventful clinical course. US demon-strated rapid inflow of the USCM into the renal cortex. The calculated time differ-ence was 1.0 ± 0.4 s. The score assigned to the parametric images was 1.7 ± 0.8.Eight of the 22 patients received a biopsy and showed histologically proven re-jection. The time difference in this group was twice as high as in the normal group(2.2 ± 0.7 s, p < 0.05). The scores were in the abnormal range (3.7 ± 1.6, p < 0.05).Two patients with perirenal hematoma also showed high scores, without rejec-tion.Conclusion: Acute rejection and perirenal hematoma are associated with a de-layed signal increase in the renal cortex. This information can be provided with asingle image with standardized color display of the temporal course of USCMinflow.

B-751 10:57

Renal perfusion abnormalities: Evaluation with contrast enhanced US usinglow-acoustic power contrast specific modesM. Bertolotto1, A. Martegani2, L. Aiani2, A. Mandic1, C. Borghi2, L. Calderan1,M.A. Cova1; 1Trieste/IT, 2Como/IT ([email protected])

Purpose: To investigate the contribution of contrast enhanced US using low-acoustic-power contrast-specific modes after sulphur hexafluoride-filled micro-bubble contrast agents in evaluation of patients with renal lesions presenting asperfusion abnormalities.Methods and Materials: Contrast enhanced US (CEUS) scans of 49 consecu-tive patients with renal perfusion abnormalities were reviewed. There were 12

Scientific Sessions

312 C D E FBA G

renal infarcts, 3 acute cortical necrosis (ACN), 6 renal abscesses, 15 acutepyelonephritis, 6 atheroembolic renal disease, 5 not perfused areas after bluntabdominal trauma, 2 segmental renal artery stenosis. Cysts, renal tumors, renalcortical scars and scarred kidneys were excluded.Results: An excellent depiction of normal renal vasculature was obtained evenin elderly, obese and nephropathic patients. Renal infarcts presented as wedge-shaped segmental perfusion defects. In ACN only small cortical areas were per-fused, while the hilar renal vessels enhanced normally. Abscesses presentedwith rounded perfusion defects. Pyelonephritis presented with wedge-shapedareas of decreased enhancement (4/15), slightly hypoperfused areas (7/15), orbarely visible perfusion defects (4/15). In patients with atheroembolic renal dis-ease bilateral hypoperfused areas can be identified during the early arterial phase.Segmental renal artery stenosis presented as segmental areas with delayedenhancement.Conclusion: CEUS allows excellent depiction of renal lesions presenting as per-fusion abnormalities.

B-752 11:06

Day case renal biopsies - desirable or dangerous?A.J. Sebastian, S. Nair, R. Panigrahi, A.M.B. Bowker; York/UK([email protected])

Purpose: To determine the rate and timing of complications after percutaneousultrasound guided renal biopsy (PRB) and to assess if this biopsy can be per-formed as a day case procedure.Methods and Materials: 141 consecutive renal biopsies performed during the 4year study period between January 2001 and December 2004 were retrospec-tively assessed to identify post biopsy complications and the adequacy of thebiopsy samples. The patients were all monitored for a minimum duration of 24hours after the procedure. A median of two 18G samples (range 1-5) were takenfrom each patient. Diagnostic samples were obtained in 138/141 (98%) patients.Results: One patient (0.7%) developed post procedure renal haemorrhage with-in 4 hours after the biopsy requiring selective renal arterial coil embolization andblood transfusion. There were no other major complications and there was noprocedure related mortality. Literature suggests that PRB performed using 14Gneedles had a major complication rate of 6.4%, with 33% of these complicationsbeing detected more than 8 hours after the procedure. We found no additionalbenefit in monitoring patients for 24 hours rather than for a period of 8 hours.Performing PRB as day case procedures will also be cost effective as admittingeach patient for overnight observation can cost more than £300.Conclusion: Day case PRB is a safe, cost effective procedure that will continueto provide renal biopsy samples of diagnostic quality. We aim to start performingPRB as day case procedures and re-audit our results.

B-753 11:15

Arterial spin labeling for kidney transplant monitoringC. de Bazelaire1, P. Martirosian2, O. Mathieu1, E. Pillebout1, P. Bourrier1,E. de Kerviler1, J. Frija1; 1Paris/FR, 2Tübingen/DE([email protected])

Purpose: Kidney transplants are monitored by MR imaging using a morphologi-cal approach based on the analysis of arteries, nephrogram, urinary cavities andthe surrounding structures. This study aimed to assess whether cortex perfusiondetermined by Arterial Spin Labeling (ASL) could help to evaluate the transplantkidney.Methods and Materials: Twenty patients with kidney transplants were includedin the study. The median delay between transplantation and ASL was 19 days.Reasons for MR imaging evaluation were anuria, unsatisfactory improvement inserum creatine or rise of serum creatine. MR imaging was performed on a Sie-mens-Symphony 1.5 Tesla scanner using the body coil for transmission and asurface coil array for reception. For ASL, a single slice FAIR method with a True-FISP imaging acquisition was used. Perfusion was measured on a coronal viewof the kidney in free breathing conditions. Imaging protocol included an MRAsequence (FLASH 3D) with contrast injection (Dotarem, Guerbet, France) and amorphological sequence (VIBE).Results: Perfusion was fully evaluated by ASL in all patients. The pelvic positionof kidney transplants did not cause reduced motion artefact. Renal cortex bloodflow was found to correlate with serum creatine (R=0.80). No correlation betweenperfusion and transplant artery stenosis was found. Perfusion maps were unre-lated to enhancement of the renal parenchyma.Conclusion: Our results suggest that physiological mechanisms could be high-lighted by ASL: 1) renal cortex perfusion and glomerular function are related, 2)renalcortex perfusion could be maintained for a long time despite significant re-nal artery stenosis and 3) renal enhancement does not reflect perfusion.

B-754 11:24

Diffusion-weighted MR imaging of renal allografts: A comparative study tohealthy volunteersH.C. Thoeny1, D. Zumstein1, S. Simon-Zoula1, U. Eisenberger1, F. De Keyzer2,C. Boesch1, F. Frey1, P. Vock1, P. Vermathen1; 1Berne/CH, 2Leuven/BE([email protected])

Purpose: To prospectively evaluate the feasibility and reproducibility of diffusion-weighted magnetic resonance imaging (DW-MRI) for the assessment of kidneyfunction in patients with transplanted kidneys and to correlate the results withage and sex matched native kidneys.Methods and Materials: Fifteen patients with transplanted kidneys and stablekidney function and 15 age and sex matched healthy volunteers underwent coro-nal DW-MRI with respiratory triggering on a 1.5 T MR system using a large set of10 b-values (b=0-900sec/mm2). The following diffusion parameters were calcu-lated: Apparent diffusion coefficient (ADCtot) from all b-values, a diffusion coeffi-cient (D) and the perfusion fraction F

p. Cortex and medulla were analyzed

separately. Serum creatinine concentration was obtained from each subject. Sevenof the 15 patients with transplanted kidneys underwent a repeat study using theidentical protocol to determine reproducibility.Results: In transplanted kidneys almost identical values were obtained in thecortex and medulla for ADCtot, D, and for Fp. In contrast, in native kidneys ADCtot,D and Fp were significantly higher in cortex than in medulla (p < 0.0001, p < 0.002and p < 0.01, respectively). The reproducibility study showed a very low CVw (co-efficient of variation within subjects) in cortex and medulla especially for ADCtot

and D of less than 3.2%. Diffusion and perfusion indices correlated with serumcreatinine concentrations.Conclusion: DW-MRI in patients with transplanted kidneys is feasible with highlyreproducible results and is able to detect differences in diffusion-based corti-comedullary differentiation as compared to native kidneys. It might therefore beable to noninvasively assess acute and chronic graft rejection after kidney trans-plantation.

B-755 11:33

Implication of dilated renal pericapsular vein in renal cell carcinoma on CTor MR imagingJ. Lee, S. Kim, S. Lee; Pusan/KR ([email protected])

Purpose: To evaluate the factors affecting dilatation of renal pericapsular veinsin renal cell carcinoma (RCC), and the relationship between dilated renal peri-capsular veins (DRPV) and hematogenous metastases to other organs.Methods and Materials: CT and MR images of 127 patients with pathologicallyproven RCC were retrospectively reviewed. CT or MR imaging was obtained withexcretory phase only, or both corticomedullary and excretory phase after admin-istration of contrast material. The parameters analysed on imaging were: (1) thediameter and number of DRPV; (2) the tumor size and degree of contour bulging;(3) the presence of renal vein thrombosis; and (4) the degree of tumor enhance-ment. In addition the degree of DRPV was compared to the nuclear grade andhistological subtypes of RCC by using the Spearman rank correlation coefficient.Results: DRPV were seen in 90 (70.9%) patients. Hematogenous metastases ofRCC to other organs were found in 15 (12%) patients. Metastases were found inthe lung (n = 9), liver (n = 3), brain (n = 1), vertebra (n = 1) and pleura (n = 1). Ofthe 15 patients with metastases, 14 patients showed a variable degree of DRPV.DRPV were identified only in patients with an exophytic mass larger than 2 cm indiameter and patients with renal vein thrombosis. Renal vein thrombosis waspresent in 6 patients. There was a weak correlation between the degree of DRPVand mean tumor size and degree of contour bulging (P <.05).Conclusion: Tumor size and exophytic growth of renal cell carcinoma were relat-ed to the development of DRPV. Hematogenous metastasis of RCC to other or-gans was high in patients with DRPV.

B-756 11:42

High resolution ASL perfusion imaging of the kidneys at 3 TeslaA. Boss, P. Martirosian, C.D. Claussen, H.-P. Schlemmer, F. Schick;Tübingen/DE ([email protected])

Purpose: Arterial spin labeling (ASL) provides a non-invasive modality for quan-titative perfusion measurements. The feasibility of high resolution ASL imaging ofthe kidneys was tested and proven at 3 Tesla using a flow-sensitive alternatinginversion recovery (FAIR) true fast imaging in steady precession (TrueFISP) tech-nique.Methods and Materials: An ASL sequence with FAIR spin-preparation and True-FISP signal detection strategy was adapted for high-resolution perfusion imag-

TT TTTu

esd

au

esd

au

esd

au

esd

au

esd

ayy yyy

Scientific Sessions

B 313C D E FA G

ing of the kidneys in a 3 Tesla whole-body scanner. Quantitative perfusion mapswere acquired in six healthy volunteers and two patients. To avoid banding arte-facts, which are prominent at 3 Tesla, a TrueFISP frequency scout was imple-mented. Perfusion maps with an in-plane resolution of 1.5 mm were recorded intransverse and coronal planes. For fast mapping of the whole kidney, an in-planeresolution of 2 mm was applied.Results: High-resolution perfusion images with excellent image quality could beobtained during a measuring time of 10 minutes in all subjects. The whole kidneycould be mapped with diagnostic image quality in less than 10 minutes. For allslices, a frequency offset of less than 100 Hz provided the kidneys free of True-FISP artifacts. Perfusion values of the renal cortex ranged from 250 up to 400 ml/100 g/min (mean cortical perfusion right kidney 316 ± 43, left 336 ± 40).Conclusion: Perfusion images of the kidney could be obtained with excellentimage quality by means of a 3 Tesla MR setting within a clinically applicablemeasuring time offering an alternative to conventional perfusion imaging withpotentially nephrotoxic contrast media.

10:30 - 12:00 Room L/M

Neuro

SS 1811Fetal brain, epilepsy and neuraxonal injuryModerators:L. Valanne; Helsinki/FIJ. Zizka; Hradec Králové/CZ

B-757 10:30

Ultra fast pre-natal MR imaging of the normal fetal posterior fossa: Acomparison with ultrasound imaging at the same gestational ageJ.R. Eames, P.D. Griffiths, E. Whitby; Sheffield/UK([email protected])

Purpose: There is sparse data on the MR appearance of the normal develop-ment of the structures of the fetal posterior fossa, and how prenatal MR imagingdata compares to US investigations at the same gestational age (GA). In thisstudy we present preliminary MR imaging data and compare that to published,GA matched ultrasound (US) data.Methods and Materials: Image sets originally obtained by using (T2 weighted)single-shot fast spin-echo, (EXPRESS) of 91 normal fetuses between GA's of 40and 20 weeks were retrospectively analysed using proprietary measuring soft-ware. Measurements were taken of the trans-cerebellar diameter (TCD) the crain-iocaudal (CCD), anteroposterior (AP), midline sagittal surface area (CSA)dimensions of the cerebellar vermis, the AP diameter of the fourth ventricle (AP4),anterior and posterior axial cerebrospinal fluid (CSF), and the midline sagittalsurface area of an area defined as the 'bony posterior fossa' (PF).Results: We present data from 63 normal fetuses. We have shown that there issignificant correlation between MR imaging and US measurements of fetal PFstructures and that there is a constant discrepancy between vermis CCD and APdiameters as measured by MR imaging.Conclusion: An in vivo MR template of normal fetal cerebellar (posterior fossa)development is essential for conclusive and accurate interpretation of the foetalbrain, to this end our data underlines previous work by Griffiths, Whitby and othergroups corroborating previously published MR depictions of the developing cere-bellum and posterior fossa and is consistent with published US data of similarage matched measurements.

B-758 10:39

Comparison of BOLD signal changes in the fetal brain at 3 T with local pO2measurementsU. Wedegärtner, N. Beindorff, A.N. Priest, H. Kooijman, H. Schröder, G. Adam;Hamburg/DE ([email protected])

Purpose: To compare BOLD signal changes (SI) with local pO2 measurementsin the fetal brain during hypoxia.Methods and Materials: A T2*-weighted single-shot gradient-echo sequence(TE 30 ms) was used to measure BOLD SI changes in the fetal sheep brainduring hypoxia. 9 pregnant ewes were examined in a 3 T scanner (Intera, PhilipsMedical Systems). Measurements were performed during a control period withnormal oxygen saturation, then during hypoxia and recovery induced by reduc-tion and subsequent increase of oxygen in the ventilated gas mixture. Regions ofinterest were placed in the fetal cerebrum to measure SI. The changes in therelaxation rate (Delta_R2*) were estimated for each plateau by ln[(SI (t)/SIcontrol)/

TE]. Fetuses were chronically instrumented with a probe in the left hemisphere tocontinuously measure the local pO2. The difference of the control and hypoxic/recovery pO2

values was calculated (Delta_pO2). Linear regression analysis was

used and a correlation coefficient (r) was calculated to indicate the correlationbetween Delta_R2* and Delta_pO2.Results: BOLD measurements were obtained during control (n = 15), hypoxia(n = 22) and recovery (n = 12). Mean values of relative SI% (SI (t)/SIcontrol) andpO2 were reduced from 100% and 14 mmHg pO2 [95% confidence interval 10-17] to 81% SI% [76-86] and 3 mmHg [2-4] and increased to 99% [93-105] and15 mmHg [11-20] during recovery. Delta_R2* and Delta_pO2 were correlated sig-nificantly (r = 0.64; p = 0.00005).Conclusion: BOLD SI changes reflect changes of local brain pO2. R2* measure-ments in the brain correlate with local brain pO2 values. Therefore R2* providesinformation about fetal brain tissue oxygenation.

B-760 10:48

Comparative analysis of MR spectroscopy and diffusion tensor imaging indiagnosing temporal lobe epilepsy on a 3.0 T MR scannerJ.-J. Lu, F. Feng, Z.-Y. Jin, H.-H. Fu; Beijing/CN ([email protected])

Purpose: It has been respectively shown that MR spectroscopy (MRS) and diffu-sion tensor imaging (DTI) can depict abnormalities in temporal lobe epilepsy (TLE).Our purpose was to use both techniques to assess TLE and compare their dis-criminative ability.Methods and Materials: Single-voxel proton MRS, DTI and routine MR imagingwere performed in 42 patients with TLE and in 29 age-matched controls using a3.0 T MR scanner. The voxel of MRS was placed in the hippocampal region bilat-erally in each subject. The N-acetylaspartate (NAA)/creatine and choline/creat-ine values were calculated automatically. DTI was performed with a b factor of1,000s/mm2 along 25 directions. Fractional anisotropy (FA) values were obtainedfrom heads of hippocampus and selected regions in temporal lobe white matterthrough Functool. The metabolic parameters and FA values were compared be-tween the patient and control groups. Receiver operating characteristic (ROC)analysis was utilized to compare the discriminative efficiency of the two tech-niques.Results: Patients with TLE had significantly lower NAA/Cr (1.09 vs 1.27, P < 0.01)in the ipsilateral hippocampal region and FA values in the head of ipsilateral hip-pocampus (0.138 vs 0.149, P < 0.05) and ipsilateral temporal lobe white matter(0.487 vs 0.521, P < 0.05). The Area under Curve was highest in NAA/Cr value(0.883), followed by FA values in temporal lobe white matter (0.716) and FA val-ues in hippocampal head region (0.623).Conclusion: The NAA/Cr values of MRS in the hippocampal region have compa-rably higher discriminative power than FA values of DTI in diagnosing TLE.

B-761 10:57

Metabolite abnormalities in potential pain processing brain regions offibromyalgia patients using MR spectroscopyM. Petrou, B.R. Foerster, X. Fan, D.J. Clauw, P.C. Sundgren; Ann Arbor, MI/US([email protected])

Purpose: To determine if there are significant differences in concentrations andratios of metabolites in different gray and white matter areas in the brain betweenfibromyalgia (FM) patients and normal controls using MR spectroscopy.Methods and Materials: 26 FM patients and 22 age-matched healthy controlsubjects underwent conventional pre and post-contrast enhanced MR imagingas well as 2D-CSI spectroscopy (TE/TR=144/1500 ms). Large volumes of inter-est were placed at the level of basal ganglia and also in the periventricular whitematter at the level of the centrum semiovale. Within the larger voxel, eighteen1 cm x 1 cm x 1 cm smaller voxels were placed in gray and white matter loca-tions that have been implicated in pain processing. Post-processing of the spec-tra was done with vendor-provided software (Functool, 2000 GE). NAA/Cr, Cho/Cr and Cho/NAA ratios were calculated for each voxel. Student's t-test was usedto determine significant differences in the metabolite ratios between the twogroups.Results: Mean Cho/Cr ratios were significantly higher in FM patients comparedto normal controls in the right prefrontal subcortical white matter (p = 0.02) andthe left parietal white matter (p = 0.04). Non-significant similar trends were seenin the left thalamus (p = 0.07) and the left internal capsule (p = 0.09).Conclusion: MR spectroscopy may be of use in evaluating patients with fibromy-algia. Our data suggests that there are regional metabolite differences betweenFM patients and healthy controls in areas that have been implicated in painprocessing. Larger studies are required to further validate this potential tool andcorrelate MR spectroscopy with clinical symptoms and response to therapies.

Scientific Sessions

314 C D E FBA G

B-762 11:06

Feasibility and initial comparison of ultrahigh field (≤≤≤≤≤ 7T) time-of-flight andphase contrast cerebral MRAJ.T. Heverhagen1, J. Duraj2, P. Schmalbrock1, M.R. Thompson2, D. Chakeres1,M.V. Knopp1; 1Columbus, OH/US, 2Cleveland, OH/US ([email protected])

Purpose: To evaluate TOF and PC MR angiography of the cerebral arteries in awhole body 7T MR scanner and to compare the results with clinical MR systemsat 1.5 T and 3 T.Methods and Materials: PC and TOF MRAs were acquired in a 7T MR scanner(Achieva, Philips) and in clinical 3 T (Achieva, Philips) and 1.5 T (Signa, GE)using birdcage head coils. 3D PC (TR: 8-15 ms; TE: 3-4 ms; FA: 12-20°; FoV:240 mm; Matrix: 512x384; averages:1) and 3D TOF (TR: 15-25 ms; TE: 3-4 ms;FA: 18-20°; FoV: 240 mm; Matrix: 512x384; averages:1) sequences with a slicethickness of 0.75 mm were acquired in the axial plane covering the vascular anat-omy of the circle of Willis. PC and TOF images were acquired in similar acquisi-tion times of about 5 minutes. The depiction of distal branches of the middle,posterior and anterior cerebral arteries was assessed.Results: Both methods demonstrated the same small arterial branches of thecircle of Willis. Besides the first order branches, e. g. frontopolar a., callosomar-ginal a. and pericallosal a., also second and third order branches, paracentral a.and medial frontal branches, were seen. The PCMRAs exhibited a better back-ground subtraction compared to the TOF sequence. In comparison to clinical1.5 T/3 T images, unenhanced MRA of the brain at 7T provides superior vesselcontrast and conspicuity. Ultrahigh field MRAs demonstrate higher order branch-es with smaller calibers than conventional MRAs.Conclusion: This study demonstrates the feasibility of both TOF and PCMRA ofthe brain in a whole body 7T MR scanner with improved arterial contrast andvessel conspicuity compared to lower clinical field strengths.

B-763 11:15

Optimized image processing in susceptibility weighted MR-imaging (SWI)A. Deistung1, A. Rauscher1, J. Sedlacik1, S. Witoszynskyj2, M. Barth3,J.R. Reichenbach1; 1Jena/DE, 2Vienna/AT, 3Nijmegen/NL([email protected])

Purpose: SWI visualizes small cerebral veins and susceptibility differences bycombining magnitude and phase information. Prior to using phase images in SWI,phase-wraps have to be eliminated. This study investigated the effects of differ-ent phase-unwrapping algorithms and phase-masks (weighting functions). Addi-tionally, a technique for visualizing the whole SWI data over maximum intensityprojection (MIP) was tested.Methods and Materials: High-resolution 3D scans were acquired with a veloci-ty-compensated gradient-echo sequence (TR/TE/FA = 67 ms/40-50 ms/25°) ona 1.5 T scanner (Vision, Siemens, Germany). Different filter sizes for homodynefiltering (2D and 3D) were analyzed in 6 healthy volunteers. The results werecompared with venograms created using a region-growing phase-unwrappingalgorithm in image-space. Corrected phase information was weighted with differ-ent functions and multiplied with the magnitude. For the MIP data were maskedand inverted.Results: The conspicuity of vessels depended critically on the quality of phase-unwrapping. Good results were obtained with filter sizes of 25% for 2D and 50%of each k-space dimension for 3D. With homodyne filtering there exists a trade-off between CNR and phase-wrap suppression in regions with field inhomoge-neities. For transverse slices the optimum weighting function maps negativephases between zero and one and sets positive phases to one. The MIP visualiz-es venous vessels well, whereas tissue is suppressed.Conclusion: Optimized homodyne filters provide a robust method for SWI dataprocessing, whereas in regions with field inhomogeneities phase-unwrapping byregion-growing yields superior results. Performing a MIP of SWI data allows vis-ualization of the vessel network similar to other angiographic methods.

B-764 11:24

A technique for assessing changes of connectivity of motor pathways inpatients with post-traumatic head injury and mostly recovered motorfunctions with 3 Tesla fMRI and DTIG. Luccichenti, A. Cherubini, P. Péran, C. Barba, R. Formisano, U. Sabatini;Rome/IT ([email protected])

Purpose: To develop a method for assessing functional plasticity of white matterin patients with post-traumatic head injury.Methods and Materials: Six right-handed patients with post traumatic head inju-ry involving motor pathways with hemiparesis mostly recovered after six months

from trauma, underwent the following MR protocol using a 3 T scanner: 1) fMRIright and left hand finger tapping using a standard block acquisition design; 2)DTI with the following parameters: 52 images with 1x1x2 mm3 voxel size in axialplane, NEX=16, b=0 and b=1000 s/mm2 in 12 directions; 3) 3D FLASH (1x1x1 mm3

voxel) and PD-T2 3 mm-thick images. The same protocol was performed in sixage-matched right-handed healthy volunteers. Diffusion and fractional anisotro-py maps were generated from DTI acquisition. Conventional fiber tracking recon-structions of pyramidal motor pathways were generated. Probabilistic connectivitymaps of pyramidal motor pathways were generated using cortical activation ob-tained from finger tapping fMRI as seed point.Results: In all patients, conventional fiber tracking algorithms showed structuralintegrity of pyramidal motor pathways, that were in all cases close but not in-volved by post-traumatic lesions. Finger tapping of the hand contralateral to theside of the lesion showed bilateral activation, while connectivity maps generatedfrom these active cortical areas showed increased probability of connection withpyramidal pathway contralateral to the side of the lesion (omolateral to fingertapping) through the corpus callosum.Conclusion: Probabilistic quantification of connectivity of brain areas may po-tentially allow the assessment of white matter plasticity, representing a powerfultool for understanding mechanisms of recovery and response to therapy.

B-765 11:33

Traumatic diffuse axonal injury (DAI): Diffusion-weighted MR imagingfindingsY. Li, X. Feng, W. Tang; Shanghai/CN ([email protected])

Purpose: Diffuse axonal injury (DAI) accounts for a significant portion of primaryintra-axial lesions in cases of traumatic brain injury. In this study we want to usediffusion-weighted MR imaging to characterize DAI in the setting of acute andsubacute traumatic brain injury.Methods and Materials: Nineteen patients (13 men and 6 women) ranging inage from 26 to 63 years were examined with conventional MR imaging (includingfast T2-weighted, FLAIR) as well as diffusion-weighted MR imaging at 1 to 18days after closed head trauma. According to the Glasgow Coma Scale (GCS), 14patients suffered from severe head injury (GCS < 8) and 5 patients had moderatehead injury (GCS 9-12). Lesions were characterized as DAI on the basis of theirlocation and their appearance on conventional MR images. Trace apparent diffu-sion coefficient (ADC) maps were computed off-line with the diffusion-weightedand base-line images. Areas of increased signal were identified on the diffusion-weighted images, and regions of interests were used to obtain trace ADC values.Results: In the nineteen patients studied, lesions were located in the cerebralhemispheric gray-white matter interface and subcortical white matter (14 cases),body and splenium of corpus callosum (3 cases), basal ganglia (1 csaes), dorso-lateral aspect of brainstem (1 case) and isotropic diffusion-weighted imagesshowed areas of increased signal with correspondingly decreased ADC. In onecase, decreased ADC was seen 18 days after the initial event.Conclusion: Decreased ADC can be demonstrated in patients with DAI in theacute setting and may persist into the subacute period, beyond that described forcytotoxic edema in ischemia.

B-766 11:42 !Magnetic resonance diffusion tensor imaging in diffuse axonal injury with a3-Tesla systemS. Wang, J. Dai, J. Ma, S. Li, L. Ai, H. Cheng; Beijing/CN([email protected])

Purpose: To use diffusion tensor imaging (DTI) to semiquantitatively analyze theseverity of diffuse axonal injury (DAI) in cases of traumatic brain injury (TBI).Methods and Materials: Twelve closed head trauma patients were prospectivelystudied with conventional MR imaging and DTI with a Siemens 3-Tesla MR sys-tem within fourteen days of injury. The fractional anisotropy (FA) values weremeasured at multiple regions of interest (ROI) by two radiologists in consensus.Results were compared with those of thirteen healthy control subjects. The FAvalues were correlated with the Glasgow Coma Scale (GCS) score in the traumagroup.Results: In comparison with the control group, the mean FA values of patientswere significantly reduced within the posterior limb of the internal capsule(P < 0.001) and corpus callosum (P < 0.05), no significant reduction within thethalami (P > 0.05). The mean FA value of the posterior limb of internal capsulewas significantly correlated with GCS score in trauma group (r = 0.823, P < 0.001).No correlation was found between the FA value of thalami and GCS scores in thetrauma group (P > 0.5).Conclusion: The findings indicated that white matter disruption or damage played

TT TTTu

esd

au

esd

au

esd

au

esd

au

esd

ayy yyy

Scientific Sessions

B 315C D E FA G

an important role in pathogenesis of closed traumatic brain injury. The measure-ment of FA value is a simple and objective procedure for semiquantitative analy-sis of brain injury severity.

10:30 - 12:00 Room N/O

Breast

SS 1802Biopsy and localisation techniquesModerators:W. Buchberger; Innsbruck/ATP. Viehweg; Dresden/DE

B-767 10:30

Stainless-steel versus titanium needles for MR imaging-guided large corebreast biopsy: Comparison for yield of breast tissueD. Flöry, M. Duenkelmeyer, M. Weber, S. Jaromi, L. Ponhold, T.H. Helbich;Vienna/AT ([email protected])

Purpose: Both stainless-steel and titanium needles are commercially availablefor MR imaging guided large-core breast biopsy (LCBB). The purpose of thisstudy was to compare these needles for their yield of breast tissue.Methods and Materials: Six 14G LCBB needles consistent of stainless steeland three consistent of titanium alloys were included. Each needle was used toobtain 50 specimens from fresh turkey breasts. Turkey breasts were immobilizedusing a stereotactic device (Noras, Germany) and positioned on the table of a1.5 T MR imager 60 cm from the gantry (Vision, Siemens, Germany). Specimenswere measured for length and weight.Results: LCBB could be successfully performed with each of the needles insidethe MR imaging suite. Average sample weight was 12.7 mg among the stainlesssteel needles and 10.4 mg for the needles consistent of titanium. Samples ob-tained with stainless steel needles showed an average length of 7.72 mm, com-pared to 7.23 mm for the titanium needles. Overall, stainless steel needlesdemonstrated significantly larger tissue harvest than titanium needles (p < 0.05).The three best performing needles were stainless steel needles (Somatex Biop-sy Handy, average sample weight 16.3 mg, followed by the True Core with 15.2 mg,and the Bard Magnum with 13.3 mg). The best titanium needles ranked fourth(Invivo Semi Automatic Biopsy Gun MR imaging, 12.3 mg) and fifth (Invivo Dou-ble Shoot Biopsy Gun MR imaging, 11.6 mg).Conclusion: Yield of breast tissue was significantly higher for stainless steelneedles compared to titanium alloys. Thus the use of the cheaper stainless steelneedles seems advantageous for MR imaging guided LCBB.

B-768 10:39

Initial experiences with skin marking for non-palpable breast cancer usingmultidetector-CTN. Shiraki, M. Kitase, M. Hara, Y. Shibamoto; Nagoya/JP

Purpose: To evaluate usefulness and problem of the skin marking using multide-tector (MD)-CT for non-palpable breast cancer.Methods and Materials: From January 2004 and August 2005, 12 patients un-derwent skin marking using MD-CT. Their age was 37-82 years (median, 52 years).Six had ductal carcinoma in situ, and 6 had invasive ductal carcinoma. 60 and180 seconds after administration of 370 mgI/ml non-ionic contrast media (CM),CT scanning was started. Radiologists identified abnormal density spots on 3 mmslice transverse images and marked foci in a felt-tip pen. Surgeons enucleatedtumors with a 1 cm margin from marking. We considered the margin as positivewhen foci existed within 5 mm from margin. We evaluated a rate of positive mar-gin, distance between foci and margin, and block number and axis of positivefoci.Results: Of the 12 cases, 5 (42%) had a positive margin. Mean + SD of distancebetween margin and foci were 19 + 13 mm. The number of positive block waswithin two in all margin-positive cases. Of 9 foci in margin-positive cases, mean +SD of distance between margin and foci was 4.4 + 1.1 mm. Six of 9 positive-margin foci (67%) existed on an axis from nipple to focus center.Conclusion: Since more than half of positive foci existed on the nipple-focusaxis and mean distance between margin and foci was 4.4 mm, improvement ofthe results would be expected if transverse images are examined in more detailand the surgical margin is slightly elongated.

B-769 10:48

US-guided core biopsy of suspicious microcalcifications using digitalmammography-guided skin markingW. Moon, N. Cho; Seoul/KR ([email protected])

Purpose: To evaluate whether digital mammography-guided skin marking canimprove the visualization and targeting of suspicious microcalcifications for US-guided biopsy.Methods and Materials: During a 10-month period, 68 consecutive women with68 mammographically suspicious microcalcifications (mean 2.4 ± 1.6 cm) with-out associated abnormalities underwent US with a 10-13-MHz transducer after afull-field digital mammography-guided skin marking. A fenestrated grid was usedfor a skin marking on suspicious calcific area at true lateral and craniocaudalviews. US detected lesions underwent US-guided 11-gauge vacuum-assistedbiopsies while lesions not seen on US underwent mammography-guided needlelocalization and surgical excision.Results: Of 68 microcalcifications, 66 (97%) lesions were identified as microcal-cifications alone or masses or ducts with microcalcifications and underwent US-guided 11-gauge vacuum-assisted biopsies. Two calcific clusters invisible on USunderwent mammography-guided needle localization and proved to be benign atsurgery. US-guided biopsy revealed 32 (49%) benign, 3 (5%) atypical ductal hy-perplasia (ADH), 22 (33%) ductal carcinoma in situ (DCIS) and 9 (14%) infiltrat-ing ductal carcinomas. US-guided biopsy retrieved calcifications in 61/66 (92%)lesions at specimen mammography. Failure to retrieve calcifications was morelikely in lesions not associated with mass on US (19%[5/26] vs 0%[0/40],p = 0.007), lesions with benign histologic result (14%[5/35] vs 0%[0/31], p = 0.055)or lesions 10 mm or smaller (19%[3/16] vs 4%[2/50], p = 0.087). Surgery revealedno carcinoma in 3 lesions that yielded ADH at US-guided biopsy and 5 (23%)infiltrating carcinomas in 22 lesions that yielded DCIS.Conclusion: Digital mammography-guided skin marking improved the visualiza-tion and targeting of suspicious microcalcifications for US-guided biopsy.

B-770 10:57

Radiologic-pathologic correlation of specimens from vacuum assistedbiopsies of patients with DCIS: Localization of calcificationsF. Diekmann, S.D. Diekmann, T. Fischer, E. Hein, B. Hamm, U. Bick; Berlin/DE([email protected])

Purpose: To determine whether the diagnosis of DCIS can be made using onlythe biopsy specimens with radiologically visible microcalcifications.Methods and Materials: One hundred patients with mammographically suspectedmicrocalcifications underwent 11G vacuum-assisted biopsy (Ethicon Endosur-gery) with removal of 12 tissue specimens per patient (resulting in a total of 1200specimens). All specimens were examined for the presence of microcalcifica-tions with digital mammography at 1.8x magnification at 22 kVp; 11 mAs; Mo/Mo(Senographe 2000D, GE Healthcare). In addition, all specimens of all patientswere evaluated histologically for microcalcifications and the presence of DCIS. Ina retrospective analysis, the 264 specimens from the 22 patients with histologi-cally demonstrated DCIS were analyzed for correlations between the radiologicdemonstration of microcalcifications in the specimens and the diagnosis of DCIS.Results: A DCIS was diagnosed in 83 of the 264 specimens from the 22 DCISpatients. Fifty-six of the 83 specimens (67.5%) contained radiologically detectedmicrocalcifications; no microcalcifications were detected radiologically in the re-maining 27 specimens. Conversely, of the 181 specimens without histologic dem-onstration of DCIS, 47 (26%) contained radiologically visible microcalcifications,134 did not. In 2 patients with radiological demonstration of microcalcifications,histology did not reveal DCIS in any of the specimens containing microcalcifica-tions.Conclusion: The results presented here suggest that full radiological workup ofall specimens is warranted since the diagnosis of DCIS was established fromspecimens not containing any radiological microcalcifications in some patientsand only 56 of the 83 DCIS specimens contained radiologically detected micro-calcifications.

B-771 11:06

Potential clip displacement following 11-gauge vacuum-assistedstereotactic biopsy: A comparison of a gelatin pledget/metallic clip versus aconventional localization clipS. Jaromi, C. Reiner, D. Flöry, G. Pfarl, M. Weber, T.H. Helbich; Vienna/AT([email protected])

Purpose: To assess a potential clip displacement of two biopsy marker clipsfollowing stereotactic 11-gauge vacuum assisted breast biopsy (VABB).

Scientific Sessions

316 C D E FBA G

Methods and Materials: In a prospective randomised study 100 conventionallocalization clips (Micromark™ II, Ethicon, USA) and 100 gelatin pledget/MetallicMarker clips (Gel Mark™ ultra, SenoRX Inc, USA) were deployed using a stere-otactic system (Mammotome™, Fisher Imaging, USA) following 11-g VABB. Pre-and post-interventional two-view mammograms were analyzed by two radiolo-gists, independently. The position of the markers were assessed. In case of adislocation, defined as a clip migration > 5 mm from the target, the distance anddirection of dislocation was recorded. The dislocation was scored on a three-point scale [1=no dislocation (≤ 5 mm); 2=minor dislocation (6-20 mm); 3=majordislocation (> 20 mm)].Results: Post-interventional mammograms revealed clip displacement in 43/200(21.5%) cases. Clip displacement was seen significantly more often in calcifica-tions (39/43; 90.7%) than in masses (4/43; 9.3%; p < 0.05). Displacement wasseen significantly more often in conventional localizing clips (30/100; 30%) thanin gelatin pledget/metallic clips (13/100; 13%; p0.05). Major dislocation was seensignificantly more often using conventional clips only (7% vs. 0%; p < 0.05).Conclusion: Following 11 g VABB clip displacement more than 20 mm was seensignificantly more often using conventional localization clips in comparison to thegelatine pledget/metallic clips. Conventional localization clips should no longerbe used following stereotactic 11 g VABB.

B-772 11:15

Galactography-guided stereotactic 11-gauge vacuum-assisted directionalbiopsy of intraductal lesions: No alternative to open surgeryC. Reiner, T.H. Helbich, D. Floery, S. Jaromi, C.C. Riedl, M. Fuchsjaeger;Vienna/AT ([email protected])

Purpose: The purpose of our study was to evaluate the accuracy of galactogra-phy-guided vacuum-assisted biopsy (VAB) and its feasibility as potential thera-peutic procedure for intraductal lesions of the breast.Methods and Materials: 16 patients (mean age 58.2 years) with nipple dischargeand galactographically verified intraductal lesions, were included in this prospec-tive trial. Galactography-guided stereotactic 11-gauge VAB was performed in 16lesions followed by open biopsy. VAB and surgical histology were correlated.Results: After VAB, histopathologic findings revealed ductal cancer in situ in five(31.3%) cases, invasive ductal cancer in three (18.7%), atypical in two (12.5%),and a benign result in six (37.5%) lesions. In ten cases (62.5%) histopathologyafter open biopsy confirmed the previous established diagnosis. Five (31.3%)carcinomas were missed or underestimated at VAB, showing one DCIS, two atyp-ical papillomas and two benign results at VAB. One (6.2%) atypical ductal hyper-plasia was underestimated as papilloma at VAB. At surgery tumor remnants werefound in all cases. In all patients nipple discharge ceased after open surgery(follow-up at least 18 months).Conclusion: Our results suggest that galactography-guided VAB is a feasiblemethod to evaluate intraductal lesions presenting with nipple discharge. As un-derestimation or false negative results occurred in six of 16 patients (37.5%) andcomplete removal cannot be assured due to the nature of lesions, 11-gauge VABcannot be recommended as alternative to open biopsy.

B-773 11:24

Stereotactic large-core needle breast biopsy (SLCNBB): Analysis of painand discomfort related to the biopsy procedureJ.M. Hemmer, J.C. Kelder, J.P.M. van Heesewijk; Nieuwegein/NL([email protected])

Purpose: This prospective study was designed to evaluate the significance ofvariables (duration of the procedure, type of breast tissue, number of biopsies,depth of the biopsies and the operator performing the procedure) and their effecton women's perception of pain and discomfort during SLCNBB.Methods and Materials: Between June 2003 and October 2005, 150 patientswere included with a non-palpable suspicious mammographic lesions, in the anal-ysis.14 Gauge biopsies were taken using a prone stereotactic table (Lorad stereogu-ide). Following their biopsy procedure the patients were asked to complete aquestionnaire.Results: 150 biopsy procedures were analyzed. There was no discomfort in lyingon the prone table. There is no relation between type of breast lesion and pain,underlying pathology and pain and performing operator and pain. The type ofbreast tissue is correlated with the pain experienced from biopsy (p = 0.0001).The depth of the biopsy correlates with pain from biopsy (p = 0.0028). There is acorrelation between the amount of biopsies and pain in the neck (p = 0.0188)and shoulder (p = 0.0366). The duration of the procedure is correlated with painexperienced in neck (p = 0.0116) but not with pain experienced from biopsy.

Conclusion: SLCNBB is not experienced as painful. No correlation was foundbetween the performing operator and pain. There is a correlation between dura-tion of the procedure and pain in neck. Pain from biopsy is correlated with depthand type of breast tissue.

B-774 11:33

Computer-navigated MR-guided vacuum-assisted breast biopsy: Afeasibility studyU. Mastalier, H. Fauster, A. Ruppert-Kohlmayr, M. Uggowitzer; Leoben/AT([email protected])

Purpose: To evaluate a recently available computer-navigated targeting systemfor MR-guided biopsy and localization of non-palpable breast lesions which arenot detectable with mammography or sonography.Methods and Materials: 14 patients underwent vacuum-assisted breast biopsy(VABB) because of BIRADS- IV and -V lesions, exclusively detected on prior MRexams. 14 lesions with a diameter between 6 and 36 mm underwent biopsy usinga computer-navigated target system including a dedicated workstation (Dy-naCAD®). Sequences for intervention consisted of a T1W FS- markersequenceand CE-dynamic sequences using a dedicated breast coil. 20-24 specimen weretaken of each lesion and underwent histological evaluation. Immediately afterbiopsy a MR-compatible clip with bioabsorbable attachment was placed via acoaxial needle into the biopsy site. Histologically proven malignant lesions un-derwent surgery guided by the clip within 2 weeks.Results: VABB could be performed in all lesions well tolerated by all patients(100%). 8/14 lesions (57%) were malignant (6 invasive carcinomas, 2 DCIS) whichcould be surgically confirmed. 6/14 lesions were benign (43%). Placement of aclip into the biopsy hole was possible in all cases. Surgery confirmed the pres-ence of the clip within or adjacent to the tumor in all surgically proven malignantlesions. No clip displacement occurred. Mean duration of the procedure was50 minutes.Conclusion: Computer-navigated MR-guided breast biopsy with DynaCAD® isa robust and highly accurate and feasible method for histological evaluation andlocalization of breast lesions which are neither detectable with mammographynor with sonography.

B-775 11:42

Vacuum-assisted biopsy systems for MR imaging-guided breast biopsy:Comparison for yield of breast tissueD. Flöry, M. Duenkelmeyer, S. Jaromi, M. Weber, C. Reiner, T.H. Helbich;Vienna/AT ([email protected])

Purpose: Vacuum assisted breast biopsy (VABB) under MR imaging guidance isincreasingly used for the histological verification of MR imaging-detected breastlesions that lack a correlation at mammography and ultrasound. The aim of thisstudy was to compare three commercially available VABB devices for their yieldof breast tissue.Methods and Materials: Three VABB devices [Mammotome (Ethicon USA), Va-cora (Bard, USA), and ATEC (Suros Surgical System, USA)] were included inthis study. Each device was used to obtain 50 specimens from fresh turkey breasts.Turkey breasts were immobilized using a stereotactic device (Noras, Germany)and positioned on the table of a 1.5 T magnetic resonance imager (Vision, Sie-mens, Germany) 60 cm from the gantry. Obtained specimens were measured forweight, length and fragmentation.Results: Biopsy could be successfully performed with each of the devices insidethe MR imaging suite. The ATEC device (mean sample length 25.5 mm, meansample weight 188.6) performed best, followed by the Vacora (mean sample length18 mm, mean sample weight 155.1) and the Mammotome (mean sample length19.8 mm, mean sample weight 77.6). Differences in mean sample weight werestatistically significant (p < 0.05). 15 samples obtained with the ATEC device yield-ed a total of 3004.8 mg of breast tissue. To retrieve the same amount of breasttissue, 21 passes were necessary with the Vacora, and 39 passes with the Mam-motome.Conclusion: Biopsies inside the MR imaging suite could be successfully per-formed with each of the devices. The ATEC device performed significantly betterthan the Vacora and the Mammotome.

TT TTTu

esd

au

esd

au

esd

au

esd

au

esd

ayy yyy

Scientific Sessions

B 317C D E FA G

B-776 11:51

Breast abscess treatment with ultrasound-guided percutaneous aspirationG. Jahr, P. Skaane, O. Westerheim; Oslo/NO ([email protected])

Purpose: To study the effect of needle aspiration of breast abscesses as alterna-tive to percutaneous pig-tail drainage or surgical incision.Methods and Materials: Ultrasonography (US) was performed on 63 patientswith suspected breast abscesses. Patients with phlegmonous mastitis were treatedwith antibiotics only. US-guided needle aspiration was performed with a 14-18 Gneedle, if an abscess was diagnosed. The patients were scheduled for short-term follow-up and the procedure was repeated in case of relapse. Patients withprogression or no therapeutic response were referred for surgical treatment.Results: 9 of the 63 patients were excluded because of complex etiological fac-tors. US showed phlegmonous mastitis in 4 patients. Of the remaining 50 pa-tients having an abscess formation, 43 were successfully treated with needleaspiration and antibiotics (median abscess diameter 26 mm, range 7-105 mm).24 abscesses had a diameter of more than 25 mm, of which only five had toundergo surgical drainage. Mean number of aspirations for each patient was 1.6(range 1-5). Median number of days from first aspiration until treatment was con-sidered successfully finished, was 13 days (range 4 to 70 days). Seven patientshad to be referred for surgical incision. Treatment success was not influenced byabscess size or duration of symptoms before therapy.Conclusion: US-guided abscess aspiration in combination with antibiotics as aminimally invasive procedure is an alternative to percutaneous pig-tail drainageand surgical incision, even in patients with larger abscess formations.

10:30 - 12:00 Room P

Vascular

SS 1815Veins and lymphaticsModerators:A.A. Azarine; Paris/FRM. Stajgis; Poznan/PL

B-777 10:30 !Recurrence after varicose vein surgery evaluated with color duplex imagingD. Tsetis, T. Kostas, C. Ioannou, S. Yarmenitis, M. Veligrantakis, K. Pagonidis,A. Katsamouris, N. Gourtsoyiannis; Iraklion/GR ([email protected])

Purpose: To evaluate the mechanisms responsible for recurrence after varicosevein surgery with color duplex imaging (CDI).Methods and Materials: One hundred and thirteen limbs (93 patients, female:69, mean age: 48 years), were preoperatively examined using CDI in order todetect the presence and the extent of venous reflux and to design the appropri-ate surgical procedure. All limbs were followed-up clinically and ultrasonograph-ically one month and 5-years after surgery.Results: Forty-two sources of reflux were identified among 28 (25%) of the 113operated limbs that had presented with recurrence during the follow-up period.The one-month postoperative examination revealed the preservation of 1) an in-competent greater saphenous vein (GSV) distal remnant in 6 limbs that weresubjected to a limited stripping (6/42, 14.3%) and 2) a saphenous stump withincompetent tributaries in 2 limbs (2/42, 4.8%).The 5 year follow-up examination showed: 1) the presence of refluxing serpen-tine tributaries arising from the ligated SFJ as a result of neovascularization in 13limbs (31%), 2) the expansion of reflux along the preoperatively competent GSVremnant in 4 limbs with limited stripping, due to chronic venous disease progres-sion (CVDP) (9.5%), 3) the development of reflux in the preoperatively compe-tent superficial veins (in non-operated areas) due to CVDP in 17 limbs (40.8%).Conclusion: The one-month postoperative ultrasonographic examination vali-dated the sufficiency and the appropriateness of the surgery performed, and the5-year examination, including both the operated and non-operated areas of thelimb, revealed unavoidable causes of recurrences such as neovascularizationand CVDP.

B-778 10:39

Diagnostic accuracy and immediate discharge rate of a DVT pathwayI. Britton1, P. Vlachou2, A. Morley-Davies1; 1Stoke-upon-Trent/UK, 2Boston/UK([email protected])

Purpose: Implementation of a combined clinical, haematological, venometer andDoppler DVT diagnostic pathway, was assessed for diagnostic accuracy and im-mediate discharge rate.Methods and Materials: Over a five-month period, 200 patients with suspectedDVT, had a Wells pre-test clinical probability score calculated. Exclusionary test-ing was performed for "low" and "moderate risk" groups with d-dimer and venom-eter. Positive results were clarified by Doppler. Patients were tracked through thepathway, and for subsequent events over a three-month period.Results: 57 had DVT (28.5%). 3 of 41 (7%) "low risk" patients, 18 of 86 (21%)"moderate risk" patients, and 36 of 73 (49%) of the "high-risk" group had DVT. 31(15.5%) patients were discharged following exclusionary testing. Negative pre-dictive value of venometer and d-dimer were 94% and 96% respectively, but PPVwas poor at 38% and 25% respectively. One patient re-presented following anegative Doppler with femoral vein thrombus. A second patient, with a negatived-dimer and venometer, had popliteal clot on USS, representing a 1% (2/200)error rate. Only 33, of potentially, 112 patients underwent venometry (defaultingto Doppler). Extrapolating the PPV of the venometer and d-dimer results in amaximum immediate discharge rate of 25%.Conclusion: The Wells clinical score is reproducible and reliable locally. Manag-ing patients with a combined pathway was 99% accurate. A 15.5% immediatedischarge rate was realised despite limited implementation of the pathway. Theresults emphasise a 7 day Doppler USS service as the most effective triage tool.

B-779 10:48

Lower limb deep venous thrombosis imagingR. Subramaniam1, T. Chou2, R. Heath2; 1Garran/AU, 2Hamilton/NZ([email protected])

Purpose: To establish the accuracy of 'Simplify' D dimer assay and Hamiltonscore for diagnosis of deep venous thrombosis (DVT).Methods: 453 patients who presented to emergency department with suspectedfirst episode of lower limb DVT were prospectively recruited from 2001 to 2003. Apreviously established pre test probability score (Hamilton Score), 'Simplify' Ddimer and a complete, single lower limb compression ultrasound examinationwere performed in all patients. All patients with a negative ultrasound examina-tion were followed up for 3 months.Results: Of the 227 patients with a negative 'Simplify' D dimer assay, 214 pa-tients had negative ultrasound examinations and 13 patients had isolated calfDVT. Among the 226 patients with a positive D dimer assay, 74 patients had DVTand 152 patients had negative ultrasound examinations for DVT. The sensitivity,specificity, positive and negative predictive values were 85.1% (95% CI 75.8 -91.8%), 58.5% (95% CI 53.4 - 63.5%), 32.7% (95% CI 26.6 - 38.9%) and 94.3%(95% CI 90.4 - 96.9%), respectively. 165 patients had an unlikely Hamilton scoreand a negative D dimer assay. The negative predictive value of 'Simplify' D dimerin an unlikely Hamilton score population was 98.8% (95% CI 95.7-99.8%).Conclusion: A negative Simplify D dimer and unlikely probability Hamilton scoreexclude DVT and lower limb ultrasound examinations are unnecessary in thesepatients.25% of cost related to DVT ultrasound imaging can be saved by applying thisdiagnostic strategy.

B-780 10:57

Evaluation of the deep venous system after orthopedic arthroplasty withmultidetector row CT: A prospective study in comparison with DopplersonographyJ. Kim, Y. Chun, S. Byun, W.-H. Kim; Incheon/KR ([email protected])

Purpose: Early detection and proper management of deep vein thrombosis (DVT)after major orthopedic surgery is very important to the prognosis. The aim of thisprospective study was to evaluate the ability of indirect multidetector row CTvenography (CTV) in detecting DVT after hip or knee arthroplasty in comparisonwith Doppler sonography.Methods and Materials: Sixty-two patients had undergone hip (n = 28), knee(n = 32) or both (n = 2) arthroplasty. For the detection of DVT, CTV and Dopplersonography was performed 8 - 40 days after the operation. CTV (collimation 16 x1.5 mm, table feed 24 mm) from diaphragm to feet was done. Doppler sonogra-phy was performed within 30 days (average 2 days) following CTV. In the detec-tion of DVT, the ability of CTV was evaluated in comparison with Dopplersonography.

Scientific Sessions

318 C D E FBA G

Results: In 28 (45%) of 62 patients, thrombosis was found in popliteal or calfveins on Doppler sonography. Compared with sonography, CTV has a sensitivityof 89%, a specificity of 97%, a positive and negative predictive value of 96% and92%, and an accuracy of 94%, respectively.Conclusion: Doppler sonography is time consuming and operator-dependent inthe detection of DVT. In this study, the diagnostic ability of CTV was comparableto that of Doppler sonography and beam-hardening artifact due to arthroplastydid not become a serious problem. Therefore, CTV is a promising alternativemodality in the evaluation of post-arthroplasty patients.

B-781 11:06

Chronic lower extremity deep venous thrombosis: Findings of indirect CTvenographyE.-A. Park, W. Lee, H. Jae, J. Chung, J. Park; Seoul/KR([email protected])

Purpose: To describe computed tomographic (CT) findings of chronic deep ve-nous thrombosis (DVT) in the lower extremity.Methods and Materials: Forty-seven consecutive patients (18 men, 29 women;mean age, 50.8 years; range, 13-76 years) confirmed as having acute DVT onCT underwent follow-up CT at least 2 months later. Mean duration of follow-upCT was 10. 6 months (range 2-42 months). Localization and extent of primaryDVT on initial CT, intraluminal residual thrombi, decreased caliber, luminal oblit-eration, intravenous fibrotic bands, ipsilateral muscular enlargement, and ipsilat-eral subcutaneous edema on follow-up CT were assessed retrospectively by tworadiologists in consensus. Frequencies of residual thrombi, decreased caliber,luminal obliteration, and fibrotic bands were calculated on each affected venoussegment; iliac, femoral, popliteal, and calf vein.Results: Nine patients (19.1%) were completely normal on follow-up CT. Thirty-eight patients (80.9%) showed abnormal findings on follow-up CT; intraluminalresidual thrombi (n = 7, 18.4%), fibrotic bands (n = 23, 60.5%), decreased calib-er (n = 16, 42.1%), luminal obliteration (n = 16, 42.1%), ipsilateral muscular en-largement (n = 28, 73.7%), and ipsilateral subcutaneous edema (n = 7, 18.4%).Most commonly affected site was iliac vein (n = 29/37, 78.4%), followed by fem-oral vein (n = 29/45, 64.4%), popliteal vein (n = 22/43, 51.2%), and calf vein(n = 15/34, 44.1%), in order of frequency.Conclusion: Fibrotic bands, decreased caliber, luminal obliteration, and ipsilat-eral muscular enlargement are common CT findings of chronic DVT in the lowerextremity.

B-782 11:15

Indirect upper extremity (UE) multidetector-row computed tomographic(MDCT) venography: Initial technical experienceJ.C. Hellinger, D. Fleischmann, S. Cheung, G.D. Rubin; Stanford, CA/US([email protected])

Purpose: To assess the UE central and peripheral venous enhancement pat-terns when performing indirect UE-MDCTV.Methods and Materials: 17 UE 16-channel MDCT venograms were performedfor either vascular mapping (N=6), occlusive disease (N=5), failing hemodialysisaccess (N=3), trauma (N=2), or vasculitis (N=1). Slice thickness (1.25 mm), re-construction interval (0.8), and table speed (55 mm/sec) were identical. Imageswere acquired 60 seconds following the arrival of ionic contrast medium (concen-tration 350 mgI/ml; volume 90-140 ml; rate 3-5 cc/sec) to the aortic arch. For eachUE, enhancement values were recorded in 8 regions (SVC, brachiocephalic, sub-clavian, axillary, brachial, upper cephalic, median cubital, and forearm deep veins)resulting in a total of 136 analyzed venous segments. All studies were reviewedby two radiologists for qualitative enhancement (0-4 scale).Results: 96% (N=130) of the vascular segments were well depicted and of diag-nostic image quality (mean 3.25). Segments deemed to have poorenhancement (N=6) had attenuation values of < 90 HU (mean 1.2). The meanvenous attenuation value across all segments was 137 HU ± 23 SD (92-182, 95%CI). Central veins (N=51, mean 147 HU) yielded 13% greaterenhancement (p = 0.005) than peripheral veins (N=85, mean 130 HU). Within thecentral venous segments, 100% had values above 95 HU (range 103-195 HU)and 90%(N=46) achieved values ≥ 115 HU. Peripheral veins revealed no signifi-cant difference between venous attenuation in the upper arm (mean 131 HU) orelbow/forearm (mean 128 HU). 93% of peripheral venous segments reached val-ues above 95 HU (range 95-214 HU) while 59% (N=48) achieved values ≥ 115 HU.Conclusion: Indirect MDCT venography can be used to image the UE centraland peripheral veins. A delay of 60 seconds after aortic arch enhancement isnecessary to achieve robust UE venous enhancement.

B-783 11:24

Impact of gadolinium MR venography on the management of patients withhaemodialysis-related thoracic venous diseaseP.S. Goh, S.C. Wang, C. Au; Singapore/SG ([email protected])

Purpose: To assess the impact of gadolinium MR venography on the manage-ment of patients with haemodialysis-related thoracic venous disease.Methods and Materials: Retrospective study of 44 consecutive patients withsuspected haemodialysis related thoracic venous disease presenting with heador arm swelling, failure of AV fistula or dialysis line placement, or history of mul-tiple central lines was performed over 20 consecutive months. MR imaging find-ings, results of subsequent interventional procedures, and medical records werereviewed. Scans were performed on a GE or Siemens 1.5 T MR system, with fastSPGR or FLASH sequences, an injection system for triggered gadolinium deliv-ery, and multiphasic image acquisition. Post-processing with image subtraction,and both maximum intensity and 3D volumetric projections from different anglesof view was used.Results: 21 female and 23 male patients with end stage renal disease, between17 to 80 years old were evaluated. Multiple stenotic and occlusive lesions weredemonstrated, most commonly involving the right internal jugular vein (23). 18patients had subsequent interventional procedures - 9 had venography confirm-ing mediastinal venous stenosis with 5 successful venoplasty and stenting, andthe other 9 had dialysis catheter venous access successfully via the non affectedintenal jugular vein (7) or subclavian vein (2). MR findings were also used in 23patients to plan AV fistula creation or revision, and in 9 to refer for peritonealdialysis.Conclusion: Gadolinium MR venography is a useful non-invasive investigationwhich has significant impact for optimising the management of patients with tho-racic venous disease.

B-784 11:33

MR imaging in Paget-Schroetter syndromeM. Reiter, D. Tscholakoff, W. Kopsa, M. Haumer, S. Sabeti, E. Minar,R.A. Bucek; Vienna/AT ([email protected])

Purpose: To evaluate anatomic structures of patients with idiopathic subclavianvein thrombosis (SVT) compared to healthy controls and to evaluate a potentialcorrelation between clinical provocative tests and the results of MR imaging (MRI).Methods and Materials: Both thoracic outlets of 29 consecutive patients withidiopathic SVT and 29 healthy controls were investigated by MR imaging with thearms alongside the body, patients were additionally imaged after a postural maneu-ver. We performed analyses of the costoclavicular and the interscalene spacesand evaluated the thoracic outlets for the presence of aberrant muscles.Results: Asymmetric interscalene spaces were not related to the presence ofSVT. We detected an aberrant muscle, the musculus subclavius posticus, in fourshoulders of patients, three of them symptomatic. In the remaining patients andin all controls no anatomical variant was detected. The minimum costoclaviculardistance (MCD) of patients was significantly smaller than in controls (P < 0.001).The thickness of the subclavius muscle did not differ between patients and con-trols. The fat surrounding the subclavian vessels had a significant influence onthe MCD, because patients with less fat exhibited a smaller MCD. The results ofclinical provocative tests were not significantly related to any measurement orany anatomical structure determined by MR imaging and exhibited a sensitivityof about 60% for detection of arm sides with SVT.Conclusion: A smaller MCD was the only anatomical finding in patients withspontaneous SVT. Clinical evaluation including provocative tests had a low sen-sitivity and did not correlate with anatomical conditions evaluated by MR imag-ing.

B-785 11:42

Lymphoscintigraphy in lower extremity lymphoedema: Relevance of visualgrading using summed lymphatic scoreA. Singh, W.E. Svensson, D.J. Towey, L. Rahman, J. Murrell, K.S. Nijran,A. Al-Nahhas; London/UK ([email protected])

Purpose: Lymphoscintigraphy is used in the evaluation of lower extremity lym-phoedema. We evaluated the role of visual grading using a Summed LymphaticScore (SLS) and measure of tracer clearance from injection sites in the interpre-tation of lymphoscintigrams.Methods and Materials: 99mTc-albumin nanocolloid lymphoscintigraphy was per-formed in 13 patients (7 males, 6 females; age range 18-74) for 26 lower extrem-ities (21 oedematous and 5 non-oedematous). Clinical records were reviewed forall patients. Images were acquired at 30 and 120 minutes using a Siemens E.

TT TTTu

esd

au

esd

au

esd

au

esd

au

esd

ayy yyy

Scientific Sessions

B 319C D E FA G

CAM-180 dual-detector gamma camera. Qualitative assessment using visual grad-ing (SLS) was performed. Quantitative parameters assessed were percentageilio-inguinal node tracer uptake at 120 min, and tracer clearance from injectionsites. Statistical analysis was performed to assess the data.Results: Lymphoedema was established in 14 oedematous extremities (primary= 9, secondary = 5) and excluded in 7 oedematous extremities for other reasons(Prader-Willi syndrome = 2, venous insufficiency = 2, surgery = 1, lipoedema =2). Lymphatic function of 5 non-oedematous extremities was also analysed. Ear-ly lymphatic filling was observed in 7, and dermal flow in 4 extremities. For SLS,sensitivity = 71%, specificity = 40%, PPV = 83.3%, NPV = 25%; and for tracerclearance from injection site, sensitivity = 55.5%, specificity = 80%, PPV = 90.9%,NPV = 33.3% were calculated.Conclusion: Our findings suggest that visual grading of lymphoscintigrams maybe standardised using summed lymphatic score and combined with a measureof tracer clearance from injection sites for establishing the diagnosis of lowerextremity lymphoedema.

B-786 11:51

High-resolution magnetic resonance lymphangiography in patients withprimary and secondary lymphedemaC. Lohrmann1, E. Foeldi2, O. Speck1, M. Langer1; 1Freiburg/DE,2Hinterzarten/DE ([email protected])

Purpose: To evaluate the feasibility of high-resolution magnetic resonance lym-phangiography with intracutaneous injection of gadodiamide in patients with pri-mary and secondary lymphedema.Methods and Materials: 15 patients with lymphedema of the lower extremitiesunderwent high-resolution magnetic resonance lymphangiography. Gadodiamidewas injected into the dorsal aspect of both feet. First, the extent and distributionof the lymphedema was evaluated using a heavily T2-weighted 3D-TSE sequence.For high-resolution magnetic resonance lymphangiography a 3D spoiled gradi-ent-echo sequence (VIBE) was used.Results: No complications were observed during or after the examination. In allpatients the lymphedema demonstrated an epifascial distribution. In 14 patients,lymphatic vessels extending from the injection site were reliably detected 15 min-utes after injection. After 5 minutes, concomitant venous enhancement was de-tected in all patients. In 13 patients, the lymphatic vessels in the upper leg couldbe seen. In 14 patients, the inguinal lymph nodes with external iliac lymphaticpathways were reliably depicted at 35 minutes. The external iliac lymph nodeswere observed merely in 6 patients. Collateral vessels with dermal back-flowwere seen in 12 patients. A lymphocutaneous fistula and a lymphocele with theafferent lymphatic vessel were detected in one patient.Conclusion: High-resolution magnetic resonance lymphangiography is feasiblein non-invasively visualizing the lymphatic vessels and accompanying complica-tions in patients with primary and secondary lymphedema. The method is notaimed at the depiction of lymph node morphology, but provides complementaryinformation about the lymphatic vessels when lymph nodes are examined withsuper-paramagnetic iron oxide particles.

Scientific Sessions

320 C D E FBA G

10.1007/s10406-006-0175-4.pdf - Springer LINK

Documents