2003 2005 Magnetom Flash 2

8/14/2019 2003 2005 Magnetom Flash 2

1/134

smedical

Issue no. 2.2003

RSNA Edition MAGNETOMF L A S H

Summertime is

MAGNETOM World Meeting Time

8/14/2019 2003 2005 Magnetom Flash 2

2/134

Topic Page

EDITORIAL

Summertime is MAGNETOM World Meeting Time 4

TECHNOLOGY CORNER

iPAT Applications in Clinical Routine and Beyond:Imaging from Head to Toe 6

Motion Under Control with Prospective Acquisition Correction(PACE) 22

MRCP with 2D PACE 25

A Quantum Leap in MR Tomography: Tim [76x32] 32

Introducing MAGNETOM Avanto:The Revolution Begins Now 36

Imaging the Whole Body Viewing the Entire Person 40

MAGNETOM WORLD EVENTS MIAMI

2nd Annual MAGNETOM World SummitSeptember 17-19, 2003 44

SUPER TECHNOLOGISTS

Crues-Kressel Award 75

R2-HIC: A Practical Method for Measuring Liver Iron Levels 76

MRI SAFETY

Guidelines to Prevent Excessive Heatingand Burns 80

Institute for Magnetic Resonance Safety,Education and Research 81

Guidelines for the Managementof the Post-Operative Patient 84

The information presented in MAGNETOM Flash is for illustration only and is not intended to be relied upon by thereader for instruction as to the practice of medicine. Any health care practitioner reading this information is remindedthat they must use their own learning, training and expertise in dealing with their individual patients. This material

does not substitute for that duty and is not intended by Siemens Medical Solutions, Inc. to be used for any purpose inthat regard.

2

MAGNETOM FLASH

Content

8/14/2019 2003 2005 Magnetom Flash 2

3/134

3www.SiemensMedical.com/MAGNETOM-World

Topic Page

EVENTS

Annual CT/MR Users Seminar Review 86

Siemens Promotes a First-Ever Meeting BetweenMR and CT Users within Mercosur 88

MAGNETOM World MeetingCardiac Imaging Symposium for Asia 96

MAGNETOM Forum 2003 A Norwegian MAGNETOM World Users Meeting 103

MAGNETOM World Activities in India 104

Siemens MAGNETOM User Club (SMUC) Meetingin ngelholm, Sweden... 108

GASTROINTESTINAL IMAGING

MR Colonography as an InterdisciplinaryCooperative Project 110

ULTRA HIGH-FIELD

Cerebrospinal Fluid Flow Measurements Initial Results at 3.0 T 116

Intracranial 3D ToF MRA with Parallel AcquisitionTechniques at 1.5 T and 3.0 T 120

MAGNETOM SITE VISIT

High Workflow of Maestro Class Systemat a Brazilian Clinic 122

CARDIO VASCULAR

Bilateral Four Channel Phased Array Carotid Coilfrom Machnet 126

Self-Gated Cardiac CineVirtually Eliminates ECG Triggering 128

How to Improve your 3D ToF witha Few Drops of Gadolinium 131

8/14/2019 2003 2005 Magnetom Flash 2

4/134

Our Editorial Team decision hadbeen to include a report on one ortwo MAGNETOM World meetings ineach Flash magazine. So, as we wereplanning this latest issue of Flash,we thought that MAGNETOM WorldSummit and one other meetingwould be covered in detail. Then myold friend Gustavo Gonzalves Ribeirocalled from Brazil: Nejat, he saidexcitedly, you cant believe howenthusiastic our customers wereabout the MAGNETOM World at theirmeeting in Mercosur. They say theyhad never seen anything like thisbefore, for the quality of the scientificcontributions and the fun. He sentme the photos and a report of themeeting. He was right. I was evenworried that we would have difficultymatching that level of organizationwith our global MAGNETOM WorldSummit.

The CT/MR Users seminar in the USis a unique tradition and the envy of

all other medical imaging companiesand exotic New Orleans was thechosen location for this years mee-ting.

Cardiac MR is maybe the mostprominent aspect of Siemens MR. Weare so far ahead of the field thatthere is scarcely any competition.Proof of this are our CMR ambassadors MAGNETOM cardiac MR users andleaders in the scientific community intheir specific cardiac imaging areas.This year, in addition to our Miamimeeting of CMR Ambassadors, wealso held another MAGNETOM Worldmeeting for cardiac MR users in Asia.

The MAGNETOM World summit inMiami was an amazing event. Noteven Hurricane Lisa could disruptthe wonderful organization of angathering that brought togetherpeople from the four corners of theworld. I would like to thank againRaya Dubner from US organizationand team colleague Heike Schindlerfor the meticulous and imaginativeorganization of even the smallestdetails of this meeting.

Do not think that we always searchfor sun, sea and sand when we planour MAGNETOM World meetings.Stimulating meetings took place thissummer in the cool and refinedcountries of Sweden and Norway,where more than 200 people gatheredto exchange information and learnabout latest developments.

The MAGNETOM World meeting inMumbai, India, was also an opportu-nity to provide WIP sequences to ourcustomers and get valuable feedbackfor optimizing our new sequences.

4

MAGNETOM FLASH

Editorial

Summertime is

MAGNETOM World Meeting Time

8/14/2019 2003 2005 Magnetom Flash 2

5/134

Marion Hellinger, MTRAMR Marketing-Application Training,Erlangen

Milind Dhamankar, M.D.MR Marketing-Applications, Erlangen

Dagmar Thomsik-Schrpfer, Ph.D.MR Marketing-Products,Erlangen

Peter Kreisler, Ph.D.Collaborations &Applications, Erlangen

Charlie Collins, B.S.R.T.Market Manager (USA),Erlangen

Gary R. McNeal, MS(BME)

Advanced Application SpecialistCardiovascular MR ImagingSiemens Medical Solutions USA

Lisa Reid,US Installed Base Manager,Malvern, PA

Michael Wendt, Ph.D.US R&D Collaborations,Malvern, PA

Helmuth Schultze-Haakh,Ph.D.US R&D Collaborations,Malvern, PA

Judy Behrens,R.T. (MR) (CT)Adv. Clinical ApplicationsSpecialist

Raya DubnerDesign Editor,Malvern, PA

Achim Riedl

Technical Support,Erlangen

Tony Enright, Ph.D.Asia Pacific Collaboration,Australia

Enjoy this issue of Flash.

A. Nejat Bengi, M.D.Editor in Chief


This is what our MAGNETOM World isall about: communication. We havethe opportunity to get feedbackfrom a worldwide community whichcan only help to get our productsever closer to perfection.

On the theme of perfection, we havea new MAGNETOM family memberwhich is probably about to revolutio-nize the MR world: the MAGNETOMAvanto, which employs the state-of-the-art technology we have named

as Tim (Total Imaging Matrix). Timis the first seamless, whole bodysurface coil design that combines 76seamlessly integrated coil elementswith up to 32 RF Channels, openingthe door to the most advancedclinical applications available today.

With reports on 6 meetings, Tim,Avanto, iPAT and much more, Flash isan essential source of informationand a great read.

Editorial Team

We thank Harald Werner, Antje Hellwich, Lawrence Tallentireand Iman Staab for their editorial help.

8/14/2019 2003 2005 Magnetom Flash 2

6/134

Correspondence:

Dr. Olaf DietrichLudwig-Maximilians-University ofMunichDepartment of Clinical Radiology GrohadernMarchioninistrae 15D-81377 MunichGermany

Introduction

With the release of syngo MR 2002B,an important new feature has becomeavailable for routine examination:integrated Parallel AcquisitionTechniques (iPAT). The general ideabehind iPAT is to acquire image datasimultaneously by two or morereceiver coils with different spatialsensitivities. Initially, this technique

was motivated by the wish to acce-lerate image acquisition withoutreducing the spatial resolution of theimage. However, it turned out thatiPAT provides several other advan-tages in various MR applications.

Technical background

Acquisition of MR images works bysubsequently acquiring phase-encoded lines in k-space. These lines

of data are finally transformed intothe image slice (or slab, in the case of3D acquisitions) by a mathematicalprocess called Fourrier transformation.An important property of data ink-space is that the density or distanceof the lines in k-space correspondsinversely to the field of view (FOV) ofthe final image, whereas the datarange in k-space corresponds to thespatial resolution of the image.Therefore, reducing the line samplingdensity by a factor of 2 (by not

acquiring every other line) leads toan image with half the FOV in phase-encoding direction in comparisonwith the original image. In this case,the acquisition time is also reducedby a factor of 2 as is well known fromusing rectangular FOVs. iPAT methodsuse exactly this effect to accelerateimage acquisition, but withoutdecreasing the FOV due to a specialiPAT image reconstruction. Using thecomplementary data from the differentreceiver coils, the missing lines ink-space can be calculated duringimage reconstruction.

There are two groups of iPATalgorithms: algorithms that explicitlycalculate missing k-space lines beforeFourier transforming the data,and algorithms that first reconstructimages with reduced FOV for allreceiver coil elements and then mergethese different images into one withfull FOV. syngo MR 2002B providesboth types of algorithms. GeneRalizedAuto-calibrating Partially Parallel

Acquisition (GRAPPA) is an algorithmof the first type [1] and is based onanother well-known algorithm of thesame class called SiMultaneousAcquisition of Spatial Harmonics(SMASH) [2]. The best-known algo-rithm of the second type has beencalled SENSitivity-Encoded (SENSE)MRI [3] and a modified SENSE algo-rithm called mSENSE is availableunder syngo MR 2002B. It dependson the specific application, suchas the anatomical region and pulse

sequence, as to which of these twoiPAT algorithms will yield betterimage quality.

Both GRAPPA and mSENSE algorithmsrequire some additional informationabout the spatial coil sensitivities, i.e.which part of the FOV is covered byeach coil element. This informationcan be acquired as a separate extrascan with low resolution or, typicalfor the GRAPPA and mSENSE algo-rithm, by additionally acquiring some

of the missing data lines in the centerof k-space (so-called reference lines)integrated into the acquisition.

Generally, the signal-to-noise ratio(SNR) in iPAT images is decreasedcompared to acquisitions with thefull k-space data. This is the sameeffect as in conventional imagingwith rectangular FOV: acquiringfewer lines in k-space decreases theSNR of the image. Additionally, iPATimages suffer an SNR loss due to thespecial reconstruction scheme: thiseffect depends on the efficacy of the

6

MAGNETOM FLASH

iPAT Applications in Clinical Routine and

Beyond: Imaging from Head to ToeOlaf Dietrich, Ph.D.Stefan O. Schoenberg, M.D.

Department of Clinical Radiology Grohadern(Chairman: Maximilian F. Reiser,M.D.), Ludwig-Maximilians-University of Munich, Germany

8/14/2019 2003 2005 Magnetom Flash 2

7/134

geometry of coil distribution and isdescribed by the so-called geometryfactor g [3].

Advantages anddisadvantages of iPAT

The obvious advantage of iPAT is theacceleration of imaging due to thereduced number of phase-encoding

lines to be acquired. With an accele-ration factor (or PAT factor) of 2, i.e.acquisition of only every second linein k-space, the imaging time isreduced by close to 50 % dependingon the number of reference lines.This can be used to decrease theoverall examination time and thusimprove the patient throughput andexamination efficacy. Alternatively,the spatial image resolution can beimproved in an iPAT scan comparedto a conventional scan of the same

duration. Both shorter scan times andhigher resolution are especiallyimportant in breath-hold imaging:either breath-hold times can beshortened or the spatial resolutioncan be improved without prolongingthe breath-hold time.

Another important iPAT application isdynamic imaging, such as measure-ments of perfusion or cardiac function,because image acceleration allowsfor a higher temporal resolution.However, as mentioned above, the

resulting SNR will be decreasedcompared to non-iPAT acquisitions,so iPAT is especially useful forhigh-SNR applications like contrast-enhanced angiography.

Using iPAT to acquire more averagesin the same total scan time canimprove image quality, particularly inanatomical areas that are prone tomotion artifacts. iPAT imaging is lesssensitive to motion, because everysingle acquisition is shorter than in aconventional sequence. By averagingimage data, the iPAT-related SNR loss

is almost compensated and remainingmotion artifacts are further reduced.

Single-shot pulse sequences, likeecho-planar imaging (EPI) or half-Fourier acquired single-shot turbospin echo (HASTE), often suffer fromimage artifacts due to their long echotrains. EPI is especially sensitive tosusceptibility artifacts, whereasHASTE images often appear blurred

due to the T2-related signal decayduring the readout of the echo train.Both problems can be reduced byapplying iPAT to shorten the lengthof the echo train without loss ofspatial resolution. In contrast toother pulse sequences, single-shotmethods can even gain SNR due toiPAT because late echoes withrelatively low signal intensity that areacquired in conventional sequencesare not contained in the shortenediPAT echo train.

In conclusion, iPAT can be advan-tageous in very different applicationswith very different ways of usingiPAT. This is demonstrated in thefollowing sections with examplesranging from clinical routine imagingto advanced study protocols. Imagingin all presented applications is perfor-med on a 1.5 T MAGNETOM SonataMaestro Class system. Standardsequences are used in most cases;however, some applications requiresequences from special work in

progress (WIP) packages by SiemensMedical Solutions*.

Diffusion tensor imaging

Diffusion tensor imaging (DTI) is anadvanced MR imaging technique formeasuring the strength, anisotropy,and direction of water diffusionin tissue. The term water diffusionrefers to the property of all watermolecules to move stochastically dueto their thermal energy (Brownianmotion). The extent of this motion is

restricted by tissue properties,particularly by the cellular micro-structure and the spatial orientationof cells. Especially in fiber structureslike muscle tissue or the cerebralwhite matter, molecular motion isrestricted by cell membranes ormyelin sheaths, and molecules movepreferably parallel to the fiberdirection whereas diffusion orthogo-nal to the fiber direction is decreased.

Thus, the resulting water diffusion isanisotropic. Information about thediffusion strength (apparent diffusioncoefficient, ADC), diffusion anisotropy,and diffusion direction are containedin the so-called diffusion tensor, amathematical object (symmetric 3x3matrix) consisting of 6 independentnumbers.

The most common pulse sequencesto measure the diffusion tensor arediffusion-weighted EPI sequenceswith diffusion gradients applied in at

least six different directions [4, 5].Single-shot EPI sequences have theadvantage that imaging is fast (about100 ms/image) and thus very insensi-tive to motion. However, EPI sequencesare very prone to susceptibilityartifacts manifesting as distortions inthe frontal brain and the cranial base.This disadvantage can be overcomeby using iPAT sequences to shortenthe length of the EPI gradient echotrain. Hence, we use a spin echo EPIdiffusion sequence with GRAPPA

reconstruction, an acceleration factorof 2, and 24 reference lines for DTIexaminations. A dedicated iPAT headcoil consisting of 8 surface coilelements (Fig. 1) provides the requirednumber of receiver channels.

Acquisition with the 8-channelhead coil results in images with animproved SNR compared to thestandard quadrature head coil (Fig. 2).


TECHNOLOGY CORNER

* The information about this product ispreliminary. The product is under development

and is not commercially available in the U.S.,and its future availability cannot be ensured.

8/14/2019 2003 2005 Magnetom Flash 2

8/134

8

MAGNETOM FLASH

This can be explained by the smallerdiameter of the 8-channel head coil(24 cm vs. 26 cm) and the reducedcoil size in cranio-caudal direction.Images were acquired witha 128x128 matrix in 36 slices,230x230 mm2 FOV, phase-encodingin anterior-posterior direction, a slicethickness of 3.6 mm, 10 averages,and an iPAT factor of 2 (24 referencelines); the echo time was 71 ms and

the repeat time 6000 ms. DTI withiPAT displays less distortion artifactsthan DTI with conventional EPIsequences (Fig. 3). Evaluating thediffusion-weighted images, parametermaps with the mean ADC, the diffu-sion anisotropy, and the main directionof diffusion can be calculated (Fig. 4).Although an N/2 artifact in phase-encoding direction (anterior-posterior)is visible in some of the original EPIimages, this artifact seems not toinfluence the calculated parameter

maps. An additional advantage ofiPAT imaging is the reduced durationof the readout that allows for theacquisition of an increased numberof slices within the given repetitiontime (TR) compared to conventionalsequences, e.g., 38 slices withoutiPAT vs. 50 slices with iPAT given aTR of 6000 ms.

Larynx imaging

Magnetic resonance imaging of thelarynx is difficult due to tissue motioncaused by swallowing and respiration[6]. Generally, MR pulse sequenceswith long acquisition times are moresensitive to motion than fast imagingsequences or even single-shotsequences. Therefore, MRI of thelarynx at conventional speed oftenleads to images with excessivemotion artifacts.

Since moving organs like the larynxcan be more clearly visualized byreducing the image acquisition time,

Figure 1 8-channel phased-arrayhead coil for acquisition of iPAT data.8 surface coil elements are locatedcylindrically around the AP axis; theinner diameter of the coil is 24 cm.

Figure 2 Comparison of SNR with standard head coil (a) and 8-channel

head coil (b). Both images are acquired with identical sequence parameters(diffusion-weighted EPI sequence, b = 1000 s/mm2, no averaging) andwithout iPAT.

Figure 3 Comparison of spin echo EPI images in two slices with (a, c) and

without (b, d) iPAT (GRAPPA algorithm). Some obvious susceptibility artifactsare marked with arrows in (b) and (d).

8/14/2019 2003 2005 Magnetom Flash 2

9/134

iPAT sequences can reduce thesensitivity to motion artifacts byaccelerating image acquisition whilemaintaining the same image resolu-tion. Thus, we use T1-weightedand T2-weighted iPAT sequences forroutine larynx imaging, e.g. inpatients suffering from suspectedlaryngeal carcinoma. A pair of dedi-cated iPAT surface coil systems with2x6 coil elements is arranged around

the lower part of the head and neckof the patient. To compensate for theiPAT intrinsic SNR loss, we increasethe number of acquisitions to 5(T1-weighted gradient echo sequencewith TR/TE = 176 ms/4.8 ms, 24* iPATreference lines) and 3 (T2-weightedTSE sequence with TR/TE = 3970 ms/89 ms, echo train length 15, 45* iPATreference lines). Both sequenceshave a 512x384 matrix, a FOV of300x281.3 mm2 and a slice thicknessof 3.5 mm. The GRAPPA algorithm is

used for iPAT reconstruction.The acquired images show a betterdelineation of tumor extent and lessmotion artifacts than MRI usingnon-iPAT techniques, thus allowingaccurate diagnosis of laryngealcarcinoma (Fig. 5). In our experience,MRI with iPAT using a flexible 12-element phased-array coil is suitablefor reliable diagnosis when laryngealcarcinoma is suspected. Generally, inimaging moving tissue it appearspreferable to acquire more averages

with reduced imaging time usingiPAT and thus gain images withidentical resolution and comparableSNR but less motion artifacts thanin conventional non-iPAT imaging.


TECHNOLOGY CORNER

Figure 5 Axial MRI scan through the supraglottic larynx, all images showthe same scan position. The images demonstrate a large supraglottic tumorinfiltrating the preepiglottic space, the left paraglottic space and the leftaryepiglottic fold.(a) T1-weighted image showing muscle-isointense tumor.(b) T1-weighted image demonstrating contrast-enhancement of the tumor.(c) T2-weighted axial image showing slightly hyperintense tumor tissue.

(d) T1-weighted fatsat image demonstrating contrast enhancement of thetumor. Spinocellular carcinoma was found at surgery.

Figure 4 Examples of DTI evaluation from diffusion-weighted iPAT images: ADC map (a, d, g) with diffusioncoefficients from 0 (black) to 2.5x10 -3 mm2/s (white);fractional anisotropy (b, e, h) from 0 (black) to 1 (white);color-coded main diffusion direction (c, f, i),left-right: red, anterior-posterior: green, cranio-caudal:blue. Note especially the high anisotropy and left-rightdiffusion direction in the corpus callosum.

* In each case the actual number of linesmeasured is half the number of reference linesmentioned in the text. This is due to the fact

that every other line is already part of the iPATscan.

8/14/2019 2003 2005 Magnetom Flash 2

10/134

Lung imaging

MR screening of infiltrates

Radiological lung screening is typicallyperformed either by conventionalx-ray (CXR) examination or by high-resolution computed tomography(HR-CT) of the thorax and thereforeexposes patients to a considerableamount of radiation, particularlyafter repeated examinations. This

radiation dose could be reduced byusing MRI as screening modalityinstead of x-ray based methods.Unfortunately, MRI of the lung is stilla technical challenge because of thevery low proton density of the lungtissue and the strong variation ofsusceptibility leading to very shortT2* relaxation times. Both factorstogether are the reason for very lowMR signal intensities from lungparenchyma and hence for a lowSNR. A further difficulty in lung MRI is

tissue motion because of respirationand cardiac motion.

The introduction of iPAT opened newpossibilities to lung imaging withT2-weighted HASTE sequences. Themain disadvantage of conventionalHASTE sequences is the blurring ofimages caused by the long echo trainand the T2-related signal decayduring its readout; this effect severelylimited the actual maximum imageresolution [7]. By using iPAT, theecho train can be reduced to half of

its original length and thus blurringartifacts are reduced. Since lateechoes with low signal intensity arenot acquired, SNR can even improvecompared to non-iPAT sequences.Additionally, the image acquisition isaccelerated such that more slices canbe acquired during one breath-holdperiod.

To evaluate the use of iPAT HASTEsequences for lung screening, wecompared HR-CT and MR images inimmunosuppressed patients withsymptoms of pneumonia but normal

10

MAGNETOM FLASH

Figure 6 iPAT HASTE images of ahealthy volunteer reconstructed withGRAPPA (a) and mSENSE (b) algorithm.Note the reconstruction artifactssuperimposing the spine in (b).

Figure 7 Ground glass infiltrate inimmunosuppressed patient.Multi-detector HR-CT (a) and iPATHASTE MRI (b).

Figure 8 Small irregular infiltratesin immunosuppressed patient.Multi-detector HR-CT (a) and iPATHASTE MRI (b).

8/14/2019 2003 2005 Magnetom Flash 2

11/134

or unspecific CXR. After comparingiPAT images reconstructed with theGRAPPA and mSENSE algorithm(Fig. 6), we decided to use the GRAPPAalgorithm because of the occurrenceof reconstruction artifacts in theimage center of the mSENSE images.Coronal slices of the lung are acqui-red with a FOV of 400x400 mm2 anda resolution of 320x320 pixels; axialslices with a FOV of 400x320 mm2

and a 320x256 matrix. The slicethickness is 8 mm and the TE is 27 msin both sequences. To reduce theecho train length as far as possible, aSiemens WIP* sequence was used.Examples of the findings are shownin Figs. 7-10.

We found that lung MRI with iPATHASTE sequences is nearly as good asHR-CT for the detection of pulmonaryinfiltrates with only few false-negativeand false-positive cases such thatMRI can be recommended especially

as a follow-up tool after initial HR-CTdiagnosis.

MR angiography and perfusionimaging

Contrast-enhanced vascular lung MRIrequires a good spatial resolutionand especially in the case of perfu-sion imaging also a good temporalresolution. Both are limited by thebreath-hold duration for the patientrather than by SNR considerationsdue to the high-contrast situation incontrast-enhanced MRI. Experienceson MR perfusion imaging of the lungare still limited and various approa-ches like conventional FLASH andHASTE sequences or flow-sensitiveinversion recovery techniques areused [8-10]. However, using iPATtechniques, both temporal andspatial resolution can be significantly


TECHNOLOGY CORNER

Figure 9 Discrete atypical infiltratesin immunosuppressed patient.Multi-detector HR-CT (a) and iPATHASTE MRI (b).

Figure 10 Discrete atypical infiltratesin immunosuppressed patient.Multi-detector HR-CT (a) and iPATHASTE MRI (b).



8/14/2019 2003 2005 Magnetom Flash 2

12/134

increased compared with conventionalimaging.

We therefore added iPAT FLASHsequences to our protocol for 3Dcontrast-enhanced MR angiography(MRA) and MR perfusion imaging ofpatients with primary and secondarypulmonary arterial hypertension.Using GRAPPA with the optimized12-element iPAT coil, a temporal

resolution of 1.2 seconds per phase ispossible for dynamic perfusionimaging, acquiring 25 dynamicphases in 30 seconds; the imageresolution is 1.5x3.0x4.0 mm3 acqui-red with a 256x128 matrix in 24slices. High resolution angiogramscan be acquired with a 512 matrix(0.8x1.0x1.6 mm3 voxel size) in 20seconds breath-hold time. For bothdynamic perfusion and high-resolutionangiography, an iPAT accelerationfactor of 2 is used with 24 referencelines*.

Image examples of these sequencesare shown in Figs. 11 and 12. Usingthe parallel acquisition technique,excellent visualization of subsegmen-tal vessels is possible in the angio-graphic images. Time-resolvedperfusion imaging allows a reliabledetection of small segmental andsubsegmental perfusion defects.Using non-iPAT methods, visualizationof perfusion defects and intravascularthrombi is generally possible as well,

although with lower temporal andspatial resolution than using iPATmethods. In conclusion, we couldsubstantially improve the temporalresolution as well as the spatialresolution by using iPAT.

Functional cardiac imaging

Global and regional cardiacfunction

Cardiac magnetic resonance imaginghas been extensively used in theassessment of global and regional

myocardial function. There is nodoubt that MRI represents the currentstandard of reference. Althoughdataset acquisition can be performedin virtually any plane, the calculationof functional parameters is mostcommonly based on a stack of slicesin double oblique short-axis orien-tation. To allow for high spatial aswell as high temporal resolution, thecurrent sequence techniques acquire

a single-slice cine data set eachbreath-hold. Although innovations ofrecent years allowed for a speed upof techniques, a completion of astandardized functional study stilltakes about 10-15 minutes includingpatient recovery periods. Real-timeimaging techniques using steady-state free precession (SSFP) sequencessuch as TrueFISP, allow for a majorspeed up in data acquisition due tothe completion of a short-axis datasetwithin a single breath-hold [11, 12].

However, this comes along witha restriction in spatial and temporalresolution. And in terms of volumetricaccuracy, temporal resolution isfar more crucial than spatial resolutionas recently shown by Miller andco-workers [13]. The current recom-mendation for functional cardiacimaging requests a temporal resolutionof 50 ms or even better.

iPAT allows this criterion to be metwhen implemented in conjunction

with real-time TrueFISP. Comparedto previous studies performed byBarkhausen and Lee [11, 12], thetemporal resolution that can beachieved is in the order of 45-50 ms.And as most recently shown, thisimprovement in temporal resolutionnow leads to an accuracy of resultscomparable to that of segmentedTrueFISP [14] (Figs. 13-16); the iPATimages are acquired with an accele-ration factor of 2 and 12 referencelines. Accordingly, iPAT not onlyallows for dramatic time savings incardiac function analysis without

12

MAGNETOM FLASH

Figure 11 Dynamic pulmonaryperfusion imaging using iPAT toacquire a slab of 24 images each1.2 seconds. Perfusion defects areshown in the left upper lobe andright lower lobe.

Figure 12 Example of an iPAThigh-resolution pulmonary MRangiography of the same patient asin Fig. 11. A significant reduction ofarterial enhancement can bedemonstrated in the left upper lobeand right lower lobe due to centralthromboembolic occlusions.

8/14/2019 2003 2005 Magnetom Flash 2

13/134


TECHNOLOGY CORNER

Figure 13 Short axis view of a malepatient with impaired right ventricularfunction due to pericardial disease.

Images acquired with a segmentedTrueFISP technique showing diastole(a) and systole (b).

Figure 14 A single slice of themulti-slice iPAT real-time cine data setat exactly the same slice position as

in Fig. 13. Comparable time points in(a) diastole and (b) systole.

Figure 15 Patient after myocardialinfarction with ischemic dilatatingcardiomyopathy. Diastolic (a) andsystolic (b) images acquired withsegmented TrueFISP show almost nochange in ventricular shape (ejectionfraction < 25 %).

Figure 16 Comparison with Fig. 15at identical slice position.iPAT real-time TrueFISP shows thesame marked thinning of the anteriorwall without change within diastole(a) and systole (b).



8/14/2019 2003 2005 Magnetom Flash 2

14/134

losing accuracy of volumetric results,but even allows for a multiplanardata acquisition within a singlebreath-hold.

Based on experiences and comparisons,the GRAPPA reconstruction shows amore robust image quality in cardiacMRI than mSENSE [1]. Due to thehigher sensitivity of SENSE-relatedmethods to folding artifacts, these

techniques seem to be less usefulin cardiac imaging because of thenecessary larger FOV that leads eitherto an additional loss of spatial resolu-tion or to loss of acquisition timewhen more phase-encoding lines areacquired. Apart from the use of iPATwith real-time techniques, it alsoallows for a further improvement ofspatial or temporal resolution insegmented single-slice acquisitionscompared to standard techniques(Fig. 17). In general, when using cine

TrueFISP techniques, the loss in SNRdue to iPAT is almost negligible.

Myocardial perfusion imaging*

Myocardial perfusion imaging is apromising and rapidly increasing fieldin cardiac MRI. The rapid developmentof scanner hardware allows also foran improvement in sequence techno-logies which has been of a majorbenefit in techniques that require an

ultra-fast data acquisition such asmyocardial perfusion imaging. MRperfusion imaging has intrinsicbenefits compared to routinely usedtechniques of nuclear medicine suchas single photon emission computedtomography (SPECT) imaging or evenpositron emission tomography (PET)which represents the current goldstandard in clinical perfusion imaging.Apart from the higher spatial resolu-tion and lack of radiation exposure,myocardial MR perfusion imaginghas no attenuation problem relatedto anatomical limitations.

14

MAGNETOM FLASH

Figure 17 Combinationof segmented cine TrueFISPwith iPAT. In comparisonto standard techniques(a; pixel size 1.5x1.5 mm2),the use of iPAT allowsfor a marked increase ofspatial resolution(b; pixel size 1x1 mm2).

Figure 18 Images of MRperfusion data sets using asaturation recovery Turbo-FLASH technique. Compa-rison of a non-iPAT Turbo-FLASH technique (a) withan iPAT (GRAPPA) Turbo-FLASH technique (b). Thereis considerable more noisewithin the iPAT imagewhich hampers depictionof perfusion abnormalitiesbased on the low SNR.

Figure 19 Saturation recovery TrueFISPperfusion images combined with iPAT(GRAPPA) and a high in-plane resolution of2.1x2.1 mm2. In contrast to saturationrecovery TurboFLASH, the spatial resolutionis still high enough to follow signal dynamics.



8/14/2019 2003 2005 Magnetom Flash 2

15/134

With the use of magnetization-prepared TurboFLASH techniques,however, such as saturation recoveryTurboFLASH, the SNR has reached alimit (Fig. 18). Therefore a combina-tion with iPAT techniques seems tobe of less benefit, as a further loss inSNR is produced. Newly developedtechniques for myocardial perfusionimaging based on SSFP techniquesare currently under investigation

[15]. In comparison with TurboFLASHtechniques, these sequences (avail-able as Siemens WIP package) showa considerable higher intrinsic SNR,therefore allowing for a minimal tomoderate loss of SNR when combi-ned with iPAT (Fig. 19).

Liver imaging

MR liver imaging is most severelyrestricted by respiratory movement.

Therefore, image quality was con-siderably improved with the intro-duction of T2-weighted turbo spinecho (TSE) and single-shot sequences.With these techniques, breath-holdexaminations of the liver becamepossible, which most authors considersuperior to conventional spin echosequences [16-18]. Generally, amaximum breath-hold time of about20 seconds, tolerable even for patientsin bad health condition, is a limitingparameter for all sequences used for

liver imaging.

Respiratory-triggered T2-weightedsequences have been studied as analternative to the breath-hold strategywith contradictory results [16, 19, 20].An advantage of respiratory-triggeredsequences is the ability to performhigh-resolution examinations with5 mm slice thickness which has notbeen possible with breath-holdsequences due to the limited breath-hold time. An attempt to overcomethis limitation of breath-hold imagingby examinations with multiple breath-

holds such that the liver is examinedin several stacks of slices, may resultin parts of the liver being missed ifthe patient does not meet the sameposition of the diaphragm in allstacks [19].

The development of iPAT allows fora substantial reduction of acquisitiontime, and thus breath-hold sequenceswith improved spatial resolutioncan be used. 2D navigator-basedtechniques, as the Prospective Acqui-sition Correction (PACE) technique,known from cardiac imaging, canadapt the stacks of slices accordingto the respiratory position by registe-ring the diaphragm position, sothat the whole liver can be coveredeven if the patient does not hold hisbreath at the same position [21].

We compared four high-resolutionT2-weighted sequences with 5 mmslice thickness and a 320x240-256

matrix for routine liver imaging:a breath-hold TSE sequence with andwithout iPAT and PACE (echo trainlength: 27, TR = 2120 ms, TE = 87 ms,4 breath-hold cycles, iPAT factor 2,24 reference lines*), and a respira-tory-triggered TSE sequence with andwithout iPAT (echo train length: 25,min. TR = 2680 ms, TE = 117 ms,iPAT factor 2, 24 reference lines).All images were acquired with a 12-element surface coil system optimizedfor iPAT applications. A respiration

belt was used for triggering. The aimwas to demonstrate the feasibility ofiPAT and PACE for T2-weighted liverimaging and to evaluate imagequality of the different sequences.

Image examples of all sequences areshown in Figs. 20 and 21. In general,imaging with iPAT reduced theacquisition time by almost 50%without visible SNR loss. Comparing

breath-hold and respiratory-triggeredtechniques, the latter turned out tobe more robust in patients whereasno difference in image quality wasobserved in volunteers. An explana-tion for this result is that patientshave more difficulty with the breath-hold period of up to 20 seconds. Inconclusion, iPAT liver examinationswith respiratory triggering appearto be the most robust approach for

clinical routine examinations.

High-resolution renalMR angiography

Three dimensional gadolinium-enhanced magnetic resonanceangiography (3D-Gd-MRA) hasgained high popularity as a non-invasive imaging alternative forgrading of renal artery stenosis [22].High accuracies of over 90 % have

been reported by numerous resear-chers in the past five years [23].Nevertheless, the technique is stillnotoriously known for over-gradinghigh-grade renal artery stenoses andmissing low-grade lesions, therebylimiting its overall clinical acceptance[24]. A recent Dutch multi-centertrial presented less encouragingresults with overall accuracies of only85 % compared to DSA. In addition,no reliable data on grading of stenosesof the more distal main renal artery

or segmental arteries exists yet [25].

One major limiting factor is spatialresolution. For standard breath-holdacquisitions with bolus administrationof extracellular, non-intravasculargadolinium chelates, the maximumachievable spatial resolution repre-sents a compromise between scantime, anatomic coverage and SNR.Current imaging protocols usuallyobtain images with a maximum of1.5 mm3 isotropic resolution whichstill represents 5 to 7 fold less thanthat of digital subtraction angiography


TECHNOLOGY CORNER



8/14/2019 2003 2005 Magnetom Flash 2

16/134

16

MAGNETOM FLASH

Figure 20 Examples of T2-weighted liver images:

Breath-hold sequence without iPAT (a) and with iPAT (b),respiratory triggering without iPAT (c) and with iPAT (d).Note the markedly reduced artifacts in the breath-holdsequence with iPAT (b) of this subject, who hadproblems holding his breath. Respiratory triggering aswell compensated this problem.

Figure 21 Examples of T2-weighted liver images:

Breath-hold sequence without iPAT (a) and with iPAT (b),respiratory triggering without iPAT (c) and with iPAT (d).Respiratory triggering shows less motion artifacts andbetter delineation of the diffuse HCC due to better T2contrast.

Figure 22 Multiplanar reformatsof a high-resolution 3D-Gd-MRA. Inthe cross-sectional reformats of thevessel (lower image series) eventhe area of the stenotic lumen canbe clearly demonstrated due to theisotropic spatial resolution.

Figure 23 Coronal MIP image ofhigh-resolution MRA with iPAT (a)reveals excellent agreement toDSA (b). Both high-grade renal arterystenoses are seen including theresidual vessel lumen. Note theabsence of any major aliasingartifacts in the center of the image.

Figure 24 Comparison of propaga-ted aliasing artifacts in the samepatient using the mSENSE (a) andGRAPPA (b) algorithm. The field of

view was on purpose set to only32 cm to enforce aliasing of thearms. In the mSENSE images severeartifacts occur in the center of theimage (arrows) while these artifactsare virtually absent on the GRAPPAimages.

8/14/2019 2003 2005 Magnetom Flash 2

17/134

(DSA). In a renal artery with a diame-ter of 7-8 mm, an isotropic voxelsize of at least 1 mm3 is required foraccurate depiction of a 90 % reduc-tion in lumen diameter.

Parallel acquisition techniques allowfor improvement of spatial resolutionwithout prolonging data acquisitionand are well suited for images with ahigh SNR such as 3D-Gd-MRA. Based

on previous calculations, it is expectedthat voxel sizes of less than 1 mm3

are substantially limited by SNRconstraints [26]. Therefore, it wasour aim to increase spatial resolutionto maximum values within this range.The iPAT strategy was applied on an8-channel MAGNETOM Sonata MaestroClass System in combination with afast 3D FLASH sequence (TR = 3.79,TE = 1.3, Bandwidth = 350 Hz/pixel,flip angle = 25). Nearly isotropicdata sets with a spatial resolution of0.8x0.8x1 mm3 could be acquiredwithin 23 seconds [27]. For signalreception the 12-element array coilsystem was used. An accelerationfactor of 2 was used with 24 referencelines for auto-calibration of the coils.For data acquisition and reconstruc-tion, the GRAPPA and mSENSE algo-rithms were compared in terms ofartifacts. To improve the contrast-to-noise ratio, the one molar contrastagent gadobutrol (Gadovist, ScheringAG, Germany) was administered ata dose of 1.25 mmol/kg body weight

with an injection rate of 2 ml/s.

In the iPAT images, SNR decreasedby a factor of about 1.5 compared tothe data without iPAT. This decreasein SNR could be visually noticed inthe source images, however theintravascular signal was still accept-able. In the MIP images, the overalldecrease in SNR was hardly detected.

The high-resolution renal 3D-Gd-MRAdata sets were compared to selectivex-ray angiography in more than20 patients with renal artery stenosis

ranging from 20 % luminal narrowingto occlusion. Image analysis of theisotropic data sets consisted of multi-planar reformats along the vesselaxis to assess the degree of diameterreduction. In addition, reformatsperpendicular to the vessel axis wereperformed to assess the degree ofreduction of vessel area. Usingmultiplanar reformats the degree ofstenosis was correctly assessed in 18

of 20 patients. In 2 cases, the degreeof stenosis was overestimated.However, when reformats wereperformed in the isotropic data setsperpendicular to the vessel axis, allstenoses could be correctly identifiedcompared to x-ray angiography(Figs. 22 and 23).

One limitation is the propagationof aliasing artifacts into the center ofthe image. These artifacts could betheoretically avoided by extendingthe FOV in the left-right direction sothat no aliasing occurs at all. Inclinical practice, however, this wouldmean a substantial increase in scantime, in particular in large patients.In addition, not all patients are ableto put their arms over their heads.Therefore some degree of aliasinginto the margins of the FOV has tobe accepted. Using the GRAPPAalgorithm, artifacts propagating fromtissue outside the FOV into thecenter of the image were kept at aminimum. Only slight ring-like

artifacts occurred, which did notaffect the image interpretation. How-ever, when the mSENSE techniquewas alternatively used, these artifactswere more severe (Fig. 24).

In conclusion, high-resolution renal3D-Gd-MRA using iPAT allows forsubstantial improvement of spatialresolution, thereby increasing dia-gnostic accuracy compared to digitalsubtraction angiography. Usingthe GRAPPA based algorithm, artifactspropagating into the center ofthe FOV can be kept at a minimum.

Whole-body imaging

Because of the recent improvementsin hardware and software and thelack of ionizing radiation, magneticresonance imaging has becomea candidate for screening imaging[28]. We have developed an MRexamination which combines wellestablished components includingfunctional cardiac imaging together

with myocardial perfusion* imaging,imaging of the lung, brain, an overallview of liver, kidneys, spleen, andpancreas, as well as the arterialsystem.

The whole-body examination isperformed in two parts. In the firstpart, the patient is in a head firstposition; the spine array, two bodyarrays, and a head array are used asreceiver coils. In the second part, thepatient is in a feet first position; thespine array, the large FOV adapter,

one or two body arrays (dependingon the height of the patient), and theperipheral angio array are used asreceiver coils. iPAT with an accelera-tion factor of two is applied for mostscans of the examination includingreal-time TrueFISP imaging of theheart (Fig. 25), high-resolutionimaging of the lung (Fig. 26) as wellas dynamic cardiac perfusion* MRIwith TrueFISP. In addition, iPAT of3D-Gd-MRA in combination with thelarge FOV adapter is performed for

all studies allowing a total scan timeof only 62 seconds to cover the areafrom the thoracic aorta down tothe toes at a spatial resolution of lessthan 1.4x1.0x1.5 mm3 (Fig. 27).

By applying the GRAPPA algorithmwith its integrated auto-calibrationscan, it is possible to use flexiblecombinations of receiver coils with aflexible choice of iPAT directions and


TECHNOLOGY CORNER



8/14/2019 2003 2005 Magnetom Flash 2

18/134

18

MAGNETOM FLASH

Figure 25 Single breath-holdevaluation of global cardiac functionwith real-time iPAT TrueFISP.

Figure 26 Imaging of the brain,lungs, and abdomen as part of thewhole-body screening examination.

Figure 27 Example of gadolinium-enhanced MR angiography as part ofthe whole-body screening examina-tion. Note the excellent visualization

of vessel segments down to thepedal arch. No stenoses are present.

8/14/2019 2003 2005 Magnetom Flash 2

19/134

to move the patient table for differentiPAT acquisitions. The advantageof iPAT in this kind of exam is thecoverage of a large anatomic regionand gain of time.

In the last two months twenty indi-viduals, referred by their physicianwhile participating in a managerhealthcare program, underwent thewhole-body scan in our department.

All twenty individuals tolerated theMR examination well. Compared tothe conventional examination techni-ques like ultrasound and ECG, wehave established a more comprehen-sive exam within reasonable scantime. First results of pathologicfindings (scar in lung, aortic stenosis,renal artery stenosis) show goodcorrelation with the gold standardexaminations.

ConclusionThis overview on applications andongoing studies in different areas ofthe body supports the current trendto use parallel imaging in the majorityof clinical scan protocols. The generaladvantages of parallel imaging arenow well established. This includesthe possibility for higher spatial reso-lution for 3D-Gd-MRA with shorterbreath-holds, thereby potentiallyimproving the accuracy of this tech-

nique for grading of renal arterystenosis. The combination of time-resolved and high-resolution 3D-Gd-MRA improves the detection anddifferentiation of pulmonary hyper-tension. The use of shorter echotrains for single shot HASTE or echoplanar imaging results in less imagedistortion and less signal decay.Initial results show benefits for EPIdiffusion tensor imaging in the brainas well as detection of early infiltratesin the lung with HASTE imaging.Imaging with multiple averages inshorter acquisition times improves

visualization of tumors in areas withincreased motion such as the larynx.Higher temporal resolution improvesthe accuracy of cardiac real-timetechniques using SSFP sequences tomeasure global cardiac functionwithin a single breath-hold.

In addition to the general benefitsof parallel imaging, the iPAT methodsGRAPPA and mSENSE feature someunique advantages. Artifacts in thecenter of coronal images resultingfrom aliasing of tissue outsidethe FOV are substantially suppressedusing the GRAPPA algorithm. TheiPAT algorithms with auto-calibrationintegrated into the individual scanare less sensitive to patient motionthan other parallel imaging techniqueswith a single measurement of thecoil sensitivity profiles at the start ofthe examination. In addition, theIPA (Integrated Panoramic Array)allows a flexible combination of

multiple receiver coil systems. There-fore, large anatomic coverage withvarious receiver coils and a flexiblechoice of the iPAT directions ispossible. This is particularly helpfulfor whole-body imaging wheremultiple receiver coil systems arecombined to scan the entire bodywith parallel imaging techniques.Scan time for a complete cardio-vascular exam is substantially reducedwhile spatial and temporal resolutionof the individual scans are preserved.

In conclusion, iPAT can be used toimprove most clinical protocols forcomprehensive morphologic andfunctional imaging. Depending onthe specific application its mainadvantages are a decrease in imagingartifacts or an increase in speed,spatial, or temporal resolution.

Co-workers on parallelimaging

Roger Eibel (pulmonary imaging)

Wilhelm Flatz (imaging of the larynx)

Peter Herzog (pulmonary imaging)

Armin Huber (cardiac imaging)

Wolfgang Klinger (MR technician)

Harald Kramer (whole-body imagingand screening)

Konstantin Nikolaou (cardiac

imaging and pulmonary imaging)Carola Schmid (MR technician)

Frank Stadie (MR technician)

Robert Stahl (diffusion tensorimaging)

Anja Struwe (MR technician)

Bernd J. Wintersperger(cardiac imaging)

Christoph Zech (abdominal imaging)

AcknowledgementsWe would like to thank the MagneticResonance Development Departmentof Siemens Medical Systems andespecially Mathias Nittka, BertholdKiefer, and Rolf Sauter for theirtechnical support.


TECHNOLOGY CORNER

8/14/2019 2003 2005 Magnetom Flash 2

20/134

References:[ 1 ] Griswold MA, Jakob PM, Heidemann RM,Nittka M, Jellus V, Wang J, Kiefer B, Haase A.Generalized autocalibrating partially parallelacquisitions (GRAPPA).Magn Reson Med 2002; 47: 12021210

[ 2 ] Sodickson DK, Manning WJ. Simultaneousacquisition of spatial harmonics (SMASH):fast imaging with radiofrequency coil arrays.Magn Reson Med 1997; 38: 591603

[ 3 ] Pruessmann KP, Weiger M, Scheidegger MB,Boesiger P.SENSE: sensitivity encoding for fast MRI.Magn Reson Med 1999; 42: 952962

[ 4 ] Basser PJ, Pierpaoli C.A simplified method to measure the diffusiontensor from seven MR images.Magn Reson Med 1998; 39: 928934

[ 5 ] Le Bihan D, Mangin JF, Poupon C, Clark CA,Pappata S, Molko N, Chabriat H.Diffusion tensor imaging: concepts andapplications.J Magn Reson Imaging 2001; 13: 534546

[ 6 ] Keberl M, Kenn W, Hahn D.Current concepts in imaging of laryngeal andhypopharyngeal cancer.Eur Radiol 2002; 12: 16721683

[ 7 ] Biederer J, Busse I, Grimm J, Reuter M,

Muhle C, Freitag S, Heller M.Sensitivity of MRI in detecting alveolar Infiltra-tes: Experimental studies.Rofo Fortschr Geb Rontgenstr Neuen BildgebVerfahr 2002; 174: 10331039

[ 8 ]Amundsen T, Torheim G, Kvistad KA,Waage A, Bjermer L, Nordlid KK, Johnsen H,Asberg A, Haraldseth O.Perfusion abnormalities in pulmonary embolismstudied with perfusion MRI and ventilation-perfusion scintigraphy: an intra-modality andinter-modality agreement study.J Magn Reson Imaging 2002; 15: 386394

[ 9 ] Hatabu H, Tadamura E, Prasad PV, Chen Q,Buxton R, Edelman RR.Noninvasive pulmonary perfusion imaging bySTAR-HASTE sequence.Magn Reson Med 2000; 44: 808812

[ 10 ] Mai VM, Hagspiel KD, Christopher JM,Do HM, Altes T, Knight-Scott J, Stith AL, Maier T,Berr SS.Perfusion imaging of the human lung usingflow-sensitive alternating inversion recoverywith an extra radiofrequency pulse (FAIRER).Magn Reson Imaging 1999; 17: 355361

[ 11 ] Barkhausen J, Goyen M, Ruhm SG,Eggebrecht H, Debatin JF, Ladd ME.Assessment of ventricular function with singlebreath-hold real-time steady-state free preces-sion cine MR imaging.Am J Roentgenol 2002; 178: 731735

[ 12 ] Lee VS, Resnick D, Bundy JM,Simonetti OP, Lee P, Weinreb JC.Cardiac function: MR evaluation in one breathhold with real-time true fast imaging withsteady-state precession.Radiology 2002; 222: 835842

[ 13 ] Miller S, Simonetti OP, Carr J, Kramer U,Finn JP. MR imaging of the heart with cine truefast imaging with steady-state precession:influence of spatial and temporal resolutions onleft ventricular functional parameters.Radiology 2002; 223: 263269

[ 14 ] Wintersperger BJ, Nikolaou K, Dietrich O,Rieber, J, Nittka M, Reiser MF, Schoenberg SO.Single breath-hold real-time cine MR imaging:Improved temporal resolution using generali-zed autocalibrating partially parallel acquisition(GRAPPA) algorithm.Eur Radiol 2003; 13: 19311936

[ 15 ] Schreiber WG, Schmitt M, Kalden P,Mohrs OK, Kreitner KF, Thelen M.Dynamic contrast-enhanced myocardialperfusion imaging using saturation-preparedTrueFISP.J Magn Reson Imaging 2002; 16: 641652

[ 16 ] Katayama M, Masui T, Kobayashi S, Ito T,Takahashi M, Sakahara H, Nozaki A, Kabasawa H.Fat-suppressed T2-weighted MRI of the liver:comparison of respiratory-triggered fast spin-echo, breath-hold single-shot fast spin-echo,

and breath-hold fast-recovery fast spin-echosequences.J Magn Reson Imaging 2001; 14: 439449

[ 17 ] Gaa J, Hatabu H, Jenkins RL, Finn JP,Edelman RR.Liver masses: replacement of conventionalT2-weighted spin-echo MR imaging withbreath-hold MR imaging.Radiology 1996; 200: 459464

[ 18 ] Hori M, Murakami T, Kim T, KanematsuM, Tsuda K, Takahashi S, Takamura M, Hoshi H,Nakamura H.Single breath-hold T2-weighted MR imaging ofthe liver: value of single-shot fast spin-echo andmultishot spin-echo echoplanar imaging.AJR Am J Roentgenol 2000; 174: 14231431

[ 19 ]Augui J, Vignaux O, Argaud C, Coste J,Gouya H, Legmann P.Liver: T2-weighted MR imaging with breath-hold fast-recovery optimized fast spin-echocompared with breath-hold half-Fourier andnon-breath-hold respiratory-triggered fastspin-echo pulse sequences.Radiology 2002; 223: 853859

[ 20 ] Tang Y, Yamashita Y, Namimoto T, Abe Y,Takahashi M.Liver T2-weighted MR imaging: comparison offast and conventional half-Fourier single-shotturbo spin-echo, breath-hold turbo spin-echo,and respiratory-triggered turbo spin-echosequences.Radiology 1997; 203: 766772

[ 21 ] Liu YL, Riederer SJ, Rossman PJ,Grimm RC, Debbins JP, Ehman RL. A monitoring,feed-back, and triggering system for reproduciblebreath-hold MR imaging.Magn Reson Med 1993; 30: 507511

[ 22 ] Prince MR, Narasimham DL, Stanley JC,Chenevert TL, Williams DM, Marx MV, Cho KJ.Breath-hold gadolinium-enhanced MR angio-graphy of the abdominal aorta and its majorbranches.Radiology 1995; 197: 785792

[ 23 ] Dong Q, Schoenberg SO, Carlos RC,Neimatallah M, Cho KJ, Williams DM,Kazanjian SN, Prince MR.Diagnosis of renal vascular disease.Radiographics 1999; 19: 15351554

[ 24 ] Schoenberg SO, Bock M, Knopp MV,Essig M, Laub G, Hawighorst H, Zuna I,Kallinowski F, van Kaick G.Renal arteries: Optimization of 3D GadoliniumMR Angiography with bolus-timing independentfast multiphase acquisition in a singlebreath-hold. Radiology 1999; 201: 667676

[ 25 ] Schoenberg SO, Knopp MV, Londy F,Krishnan S, Zuna I, Lang N, Essig M, HawighorstH, Maki JH, Stafford-Johnson D, Kallinowski F,Chenevert TL, Prince MR.Morphologic and functional magnetic resonanceimaging of renal artery stenosis: a multireadertricenter study.

J Am Soc Nephrology 2002; 13: 158169.[ 26 ] Heid O.The outer limits of contrast enhanced MR angio(gradient performance, resolution, speed, etc).In: Proceedings of the tenth internationalworkshop on magnetic resonance angiography,Park City, Utah.The International MR Angio Club 1998: 75

[ 27 ] Schoenberg SO, Rieger J, Johannson LO,Dietrich O, Bock M, Prince MR, Reiser MF.Diagnosis of renal artery stenosis with magneticresonance angiography update 2003.Nephrol Dial Transplant 2003, 18: 12521256

[ 28 ] Goyen M, Herborn CU, Kroger K,Lauenstein TC, Debatin JF, Ruehm SG.Detection of atherosclerosis: systemic imaging

for systemic disease with whole-body three-dimensional MR angiography/initial experience.Radiology 2003; 227: 277282

20

MAGNETOM FLASH

8/14/2019 2003 2005 Magnetom Flash 2

21/134

What do

you see?

Proven Outcomes in Radiology.

It begins with you. By understanding what

you need most were able to develop solutions that are

most valuable to you. The advances weve made have

helped radiologists provide more informed diagnoses

in a shorter period of time. Dramatically improve

clinical workflow. Explore more non-invasive methods.

And identify diseases in earlier stages.

Our goal is clear. To help you achieve sustainable,

meaningful results. Results that come from integrating

medical technology, IT, management consulting and

services in a way that only Siemens can. See what we

see. Tangible solutions.

www.siemens.com/medical Results may vary. Data on file.

Siemens Medical Solutions that help

We see a way to offer the worlds fastest CT scanner with 0.37s rotation time

We see a way to do seamless whole-body imaging with MR in as little as12 minutes

We see a way to quadruple patient throughput in PET/CT

8/14/2019 2003 2005 Magnetom Flash 2

22/134

In a variety of MRI applications,

motion can adversely affect imagequality. Since patient motion cannotbe controlled sufficiently in all cases,Siemens has developed strategiesto maintain image quality despitemotion. Correction of motion effectsat the post-processing stage wouldbe one approach. However, there isnothing better than acquiring gooddata in the first place, and strategiesto account for motion during acquisi-tion are currently offered on SiemensMR scanners. These Inline techni-

ques for coping with motion arecollectively termed PACE (ProspectiveAcquisition CorrEction). Correspon-ding to the spatial dimensions ofthe dataset used for calculating theadjustment, these techniques aretermed 1D PACE, 2D PACE, and 3DPACE. The first two are used mainlyto deal with breathing motion, whilethe third one is applied for motionadjustment in neurological studies.

1D PACE and 2D PACEMethod

The fastest method of detectingmotion is 1D PACE (also known as anavigator technique). It typicallyrequires only 30 ms and is usedprimarily for minimizing the effectsof breathing motion in cardiacexams. For this purpose, a single lineof data from a pencil-shaped volumethat crosses the diaphragm is acquired.The volume is interactively placed(Fig. 1) in such a way that the positionof the diaphragm can be calculated

and used for motion correction inreal time. In 2D PACE, an image isacquired by means of a low-resolutiongradient echo sequence featuringa low flip angle; this ensures thatmagnetization is not saturated, sothat dark lines in the image areavoided. The user places a small boxacross the diaphragm on the 2D

image (Fig. 2). The change in signalintensity along the axis of the box isused to determine the position of thediaphragm. Since a 2D image providesmore information than a single line,this method is very robust. The timeneeded to acquire an image for 2DPACE is around 100 ms. The highlyreliable 2D PACE technique is uniqueto Siemens.

The advantages afforded by 1Dand 2D PACE can be used in a varietyof ways.

Application:Multiple breath-hold examinations

For patients who can hold their breathfor only a short time, the acquisitioncan be split up into multiple breath-holds. The information about thediaphragm position allows the opera-tor to monitor the breathing patternof the patient online. Furthermore,acquisition of slices during differentbreath-holds can be aligned in orderto compensate for imperfect repro-

ducibility of the breathhold position.In this way, gaps between slices oroverlaps are avoided.

Without PACE, the operator wouldhave to visually inspect the image-stacks and determine if there aregaps or overlaps between them atedious process that is highly opera-tor-dependent. During this time thepatient would have to remain in thescanner, since it might be necessaryto cover gaps with additional scans.Therefore, a lack of PACE capabilitieswould cause unnecessary and costlyprolongation of the total exam-time,which in turn would lead to decreasedpatient compliance and comfort.

22

MAGNETOM FLASH

Motion Under Control with Prospective

Acquisition Correction (PACE)Michael Szimtenings, Ph.D.Siemens Medical Solutions USA,Inc.

Figure 1 Positioningof the pencil-shapedvolume across thediaphragm for 1D PACEin an axial (a) andcoronal (b) plane. Thecross-section of thepencil-shaped volume isdefined by the intersec-tion of the two turquoiseboxes in the axial plane.The length of the pencil-

shaped volume isdepicted in the coronalplane (turquoise box).

Figure 2 Selection of the 2D area(turquoise box) used for detection ofthe diaphragm position when using2D PACE. Half the box should coverthe lungs, the other half the liver.

8/14/2019 2003 2005 Magnetom Flash 2

23/134

Application:Breathe freely with PACE

For some patients, even the shortestbreath-hold duration might be toodemanding; or, patients may beunable to follow breathing commandsdue to impairments in mental status.In such cases, PACE allows for imagingwhile the patient is breathing freely.During a short learning phase, thebreathing of the patient is analyzedand the central position of an accep-tance window is calculated auto-matically. Next, the gated acquisitionbegins: slices are acquired only whenthe diaphragm position falls withinthe acceptance window. Here, theslice positions of different scans canalso be aligned based on informationabout the position of the diaphragm.Without PACE, it would be extremelydifficult (or even impossible) toperform useful MR studies in patientswho cannot hold their breath.

1D PACE or 2D PACE

Whether 1D PACE or 2D PACE shouldbe used depends on the application.Cardiac exams benefit from thespeed of 1D PACE. In order to obtaincine-images with high frame rates,motion detection should be as fastas possible. Also, saturation in thepencil-shaped volume is not a problem,since it can be placed outside the

heart. Fig. 3 shows images of coronaryarteries acquired in this way. Forabdominal imaging, 2D PACE is thebest choice, since the scan timeextension is not significant. On multi-breathhold exams, for example,breath-hold times are extended onlyby a tenth of a second as a result ofusing 2D PACE.

3D PACE

Functional MR imaging (fMRI) isanother application where motion

detection, and instantaneous adjust-ment of the acquisition accordingto this information, are crucial.Here, complete multi-slice EPI data-sets of the head are acquired in rapidsuccession during presentation ofvarious stimuli. In order for thestatistical analysis to be successful,the datasets need to be alignedperfectly. For this purpose, each 3Ddataset is compared with the previousone and the translation as well asthe rotation of the head are calculated(and displayed) in real-time. Thesoftware is able to compensate forrotations and translations in all6 degrees of freedom. The techniquecan therefore account, in real time,for any so-called rigid-body motion.For acquisition of the next dataset,slice position and orientation are

adjusted according to the alteredposition of the head.

For 3D PACE, no additional dataacquisition is needed since thedetection of motion is done on theactual imaging data, which is typicallyreacquired every 2-4 seconds. Toaccount for potential motion effectseven within this short period of time,a further retrospective correction isapplied (in realtime) to the data. Theinterval between acquisitions canbe as low as 100 ms for the hardware

to be able to adjust to the movement.3D PACE is a feature unique toSiemens scanners. Its usefulness canbe seen in Fig. 4: without motioncorrection at all, or with retrospectivecorrection only, the fMRI activationmaps are much less meaningful(statistically significant differencesare lost in the motion-inducednoise). Without 3D PACE, fMRIstudies such as the one shown inFig. 5 would be much noisier andmay even turn out to be entirelyuseless due to motion artifacts.

Conclusion

The essential feature of SiemensPACE technology is the prospective

adjustment of an acquisitions scanparameters in order to minimizemotion artifacts. With the help of 1Dand 2D PACE, breathing motion canbe monitored and corrected, and thevariability of breath-hold positionsin multiple breath-hold exams can bevirtually eliminated. 1D PACE takesvery little extra time, making it idealfor cardiac MR exams. 2D PACEfeatures small flip-angles, leaving themagnetization in the volume ofinterest practically undisturbed. It isalso a very robust technique, makingfree-breathing abdominal MR imaging


TECHNOLOGY CORNER

Figure 3 Images of coronary arteries acquired using 1D PACE.In (a) the left coronary artery is visible brightly due to the bright-bloodcontrast inherent to the TrueFISP sequence. In (b) the right coronaryartery is displayed as a dark line, since the black-blood contrastpreparation of the TSE sequence makes the blood signal disappear.

8/14/2019 2003 2005 Magnetom Flash 2

24/134

a clinical reality. 3D PACE is capableof detecting, and correcting for,linear and rotational motion in6 degrees of freedom and in real-time a feature found only onSiemens MR scanners. The advanced

real-time feedback capabilities ofSiemens MR systems are fullyexploited in the three versions ofPACE to provide a more comfortableexam for patients and to producesharper, more meaningful diagnosticimages.

24

MAGNETOM FLASH

Figure 5 Activation of right andleft primary motor cortex as detectedby fMRI on a 1.5 T MAGNETOMSonata system. The paradigm wasalternating (30s/30s) left- and right-handed finger tapping. For thefunctional study, which took 4.0 minto acquire, 3D PACE real time motioncorrection was employed. An anatomi-cal dataset was pre-acquired in6.3 min. The use of 3D PACE improvesfMRI results by ensuring morerobust activation detection and bettersuppression of motion artifacts.

Figure 4 Activation maps of an fMRIstudy, during which the volunteerperforms nodding head motions of1.5 degrees in correlation with astimulus. Data were acquired withoutmotion correction at all (a), withretrospective motion correction only(b), and with 3D PACE (c). The virtualelimination of pixels falsely showingactivation is clearly seen in the3D PACE image. Only the real diffe-rences between regional activationsare shown in (c).

8/14/2019 2003 2005 Magnetom Flash 2

25/134

ment of normal pancreatic paren-chyma with atrophy, fibrosis andcalcification as well as ductal dilata-tion, strictures and calculi) even thedilated side branches of the pancreaticduct can be seen.

Pancreatic pseudocysts are encapsu-lated collections of pancreatic fluidcaused by acute/chronic pancreatitisand they are well shown. MRCPis far more sensitive in detection ofpseudocysts than ERCP.

A comprehensive MR imaging assess-ment of the pancreatobiliary tractnormally includes, in addition to thestrong-T2-weighted Magnetic Reso-nance Cholangio-Pancreatography(MRCP), a (normal) T2-weightedsequence (TSE), a T1-weighted FLASHwith fat saturation and a dynamicpre/post-gadolinium T1-weighted fatsuppressed FLASH sequence. Thelatter are not subjects dealt with in

this article.Strong-T2-weighted imaging display-ing selectively non-flowing fluid inthe biliary and pancreatic ducts canbe done with:

Thick slab and/or thin slicescoronal/axial T2-weighted single shotTSE/HASTEor with 3D TSE/HASTE techniques followedby a MIP postprocessing.

Superposition of fluid-filled bowelportions can easily be identified bysubsequently performing multiangleMIP projections or by administratingnegative contrast media. Neverthelesswhen using 3D techniques it is veryimportant also to review the originaldata to avoid the loss or change ofinformation due to postprocessingand to view also the conventionalaxial T1- and T2-weighted images.

Difficulties in visualizing thepancreatico-biliary system might beencountered due to either:

motion induced artifacts(respiratory-, peristaltic-, cardiac-motion) or

the long measurement time forobtaining high resolution.

Option 1:Breath-hold technique

Respiratory motion artifacts can be

largely eliminated to a great extentby using breath-hold techniquesin combination with T1-weightedFLASH sequences with spectral fatsaturation or in-/out-of-phase or withT2-weighted single shot TSE/HASTEsequences (thick slab or thin slices).

When running T2-weighted single-shot techniques due to the measure-ment time limited by the breath-holdduration, very high resolution is noteasy to reach and the SNR is diminis-hed in comparison with multishot-

segmented TSE sequences. Further-more, blurring due to T2 decay mayoccur when long echo trains aresetup.

Option 2: Free breathingtechnique (2D PACE MRCP)

Another technique which providesgood image quality with freebreathing is the 2D PACE MR Chol-angiography (Available with syngoMR 2004A). The advantages are:

higher spatial resolution better delineation of smallstructures, strictures and smallsecondary ducts as well as fillingdefects

reduction of motion effects inpatients who have difficulty or areunable to breath-hold.


TECHNOLOGY CORNER

This article addresses new MRCP

techniques which will be implementedwithin syngo MR 2004A* for highresolution imaging using free brea-thing navigator-triggered (2D PACE)3D TSE-Restore sequences (at 1.5 T-3 T)or 3D HASTE with inversion recovery(at 0.2 T).

Definition:

Magnetic Resonance (MR) Cholangio-Pancreatography (MRCP) is an imaging

technique that noninvasively depictsbiliary and pancreatic ducts. Thistechnique shows good correlationwith the Endoscopic RetrogradeCholangioPancreatography (ERCP).

(Strong)T2-weighted sequenceswith fluid being bright allow optimalvisualization of the anatomy andpathology of the biliary, and pancreaticducts. Unlike other techniques likeERCP, MRCP also allows visualizinganatomy beyond these obstructions.

Biliary and pancreatic ductal stonesare seen as filling defects, irregu-larities like strictures, dilatations,pancreatic cystic lesions, complexperipancratic fluid collections andislet cell tumors can be visualized.

Strong T2-weighted TSE/HASTEsequences are the best to show peri-pancreatic edema, fluid collections.

With pathologic changes e.g. chronicpancreatitis (an inflammatory processof the pancreas with irreversibleexocrine and endocrine dysfunction,characterized by permanent replace-

MRCP with 2D PACE

Wilhelm Horger

Application DevelopmentMREA-Clinical

Alto Stemmer

Application DevelopmentMREA-Clinical

* The information about this product ispreliminary. The product is under developmentand is not commercially available in the U.S.,and its future availability cannot be ensured.

The safety of imaging fetuses, infants has notbeen establised.

8/14/2019 2003 2005 Magnetom Flash 2

26/134

26

MAGNETOM FLASH

Chronic pancreatitis secondarybranches of the pancreatic ductare clearly visible(Courtesy Prim.Univ.Doz. Dr. Gerd

Reuther / Wien).

Subcapsular liver cyst and multiplesmall cysts(Courtesy Dr. Markus Henschel /Bremen).

Obstructive intrapancreatic diverticle(Courtesy Dr. Markus Henschel /Bremen).

Dilated biliary system.Chronic pancreatitis(Courtesy Prof. Janisch / Erlangen).

Chronic pancreatitis(Courtesy Prof. Janisch / Erlangen).

Chronic pancreatitis chain of beads appearance of thepancreatic duct

(Courtesy Prof. Janisch / Erlangen).

Dillated gall bladder and bile ducts2D PACE, free breathing(Courtesy Prof. Janisch / Erlangen).

Big gallstone in gall bladder(Courtesy Prof. Janisch / Erlangen).

Dilated biliary system. Chronicpancreatitis(Courtesy Prof. Janisch / Erlangen).

Some image examples obtained with the new technique:

8/14/2019 2003 2005 Magnetom Flash 2

27/134


TECHNOLOGY CORNER

Cystadenocarcinoma(Courtesy Prim.Univ.Doz. Dr. GerdReuther / Wien).

MAGNETOM Trio: volunteer withnormal appearance of biliary ducts.

MAGNETOM Concerto: volunteerwith normal appearance of biliaryand pancreatic ducts.

Figure 1 Physio-PACE parametercard.

How to perform 2D PACEMRCP?

Respiratory triggering reducesmotion artifacts by synchronizinganatomical data acquisition with therespiratory cycle. Navigator trigge-ring uses a navigator (i.e. MR signals)to monitor the respiratory motion.This distinguishes the new techniquefrom the respiratory triggering

technique available with the productsoftware, which uses a respiratorybelt to retrieve patients breathingpattern. The respiratory belt is notneeded for navigator triggering.

Navigator triggering is available withthe TSE PACE and with the HASTEPACE sequence.

Both sequences offer the possibilityto select a -90 RF pulse (the so calledrestore pulse) at the end of the echotrain. This pulse flips the transverse

magnetization back into longitudinaldirection, which shortens the spinrelaxation time.

To plan a navigator triggeredmeasurement, proceed as follows:

Select the Trigger option listedunder Respiratory control on thePhysio-PACE card.

Position the navigator on theedge of the diaphragm in the coronallocalizer. Navigator positioning is

the same for all respiratory controlmodes with navigator support and isdescribed in detail in the SiemensApplications Guide.

Plan the slices as normal. Set upthe imaging parameters (e.g. numbern of slices per concatenation, turbofactor) to ensure that the acquisitionduration TAcq is between one-thirdand one-half of the expected averagerespiratory-period. The tool tip of themeasurement time provides theacquisition duration. In an interleavedmulti-slice measurement, the acqui-

Exam set-up

sition duration is n-times the timeneeded to acquire a single echo train1.The respiratory period is defined asthe time interval from one maximuminspiration to next maximum inspira-tion. The respiratory period is roughly5 seconds in healthy adults but maybe substantially shorter for childrenor in the case of illness. The respiratory

period can be measured in a briefinitial measurement. Select the Scoutmode option in the Physio-PACEparameter card and start the acquisi-tion. After a complete respiratoryinterval (i.e. end inspiration has beendetected at least twice) the medianrespiratory period is displayed asRespiratory cycle in the upper leftcorner of the image shown in theOnline display. Remember to deselectthe Scout mode option prior to theactual measurement.

The predicted total scan time shownin the upper left corner of the cardstack is the minimum possible scantime. The actual total scan time willbe longer, depending on the actuallength of the respiratory period.The second line of the scan time tooltip provides the number of requiredrespiratory cycles in the form 5+X.

1 If a selective preparation pulse is used, theacquisition duration also includes the timeneeded to play out the preparation pulse and

the time between the inversion pulse and the90-excitation pulse (TI-time).

8/14/2019 2003 2005 Magnetom Flash 2

28/134

28

MAGNETOM FLASH

5 respiratory cycles for the learningphase of the trigger algorithm plusX respiratory cycles for the imagingphase. The actual total scan time willtherefore be close to:

Actual total scan time (5+X)

*Average respiratory period

Ask the patient to breathe regularlythroughout the measurement andstart the acquisition.

Principle of the navigatortriggered sequence

The navigator-triggered sequencecan be split into two parts. The firstpart is the initial learning phase,which is needed by the trigger algo-rithm to ascertain the patientsbreathing pattern. The second part isthe imaging phase during whichthe imaging data are acquired which

are needed to reconstruct the anato-mical images.

Imaging phase

At the beginning of the imagingphase, the navigator is repeated at aconstant interval Scout-TR to trackthe diaphragm position. As soon asthe series of detected diaphragmpositions fulfills the trigger condition,the sequence stops repeating the

navigator and executes the first blocof the anatomical imaging sequence(see Fig. 2). In the case of the inter-leaved TSE-sequence, a block acquiresn echo trains-one per slice of thecurrent concatenation. In the case ofthe single-shot HASTE-sequence theblock acquires one complete slice.400 ms after the anatomical dataacquisition period has finished thesequence plays out the navigatoragain, to find the next suited respira-tory phase. This cycle is repeateduntil all anatomical data have beenacquired.

If the standard setting is used thetrigger condition is:

i. The series of detected diaphragmpositions must be rising; i.e.the patient must not breathe in.

ii. The latest detected diaphragmposition must fall within apredefined acceptance window.

The respiratory curve shown in the

Online display during the imagingphase is incomplete (Fig. 3). Duringthe anatomical data acquisitionperiod the navigator is not played outand therefore the respiratory tracecan not be continued. As soon as thesystem detects the onset of expira-tion, the acceptance window isshown as a yellow box in the Onlinedisplay. The vertical edge width ofthe yellow box is equal to the valueof the parameter Acceptance win-dow on the Physio-PACE card. The

central position of the acceptancewindow (the so-called trigger level)is either determined by the systemduring the learning phase or can beadjusted manually

Learning phase

The initial learning phase requires 5respiratory cycles. The learning phaseis needed by the triggering routineto set the central position of the

acceptance window during theimaging phase2. During the learningphase the breathing pattern is shownin the Online display. Beginningwith the second complete respiratorycycle a red box visualizes the pro-posed anatomical data acquisitionperiod. The location of these boxes isbased on the parameter settingand the evaluation of the previousrespiratory cycles. The horizontaledge width of the red boxes is deter-mined by the aforementioned acqui-sition duration. The vertical positionand edge width of each box was set

that the box encloses the wholediaphragm trace during the proposedacquisition period. The parametersetting is fine for a certain respiratorycycle, if the data are acquired in therelaxed position near end expiration.If the horizontal edge width ofthe red boxes is comparable to, orgreater than, one respiratory period(horizontal distance from maximuminspiration to next maximum inspira-

tion), the measurement must bestopped and the acquisition durationmust be reduced. This is necessaryto avoid artifacts and a low triggerrate. In the case of the TSE-sequencea smaller turbo factor or a reducednumber of slices per concatenationshortens the acquisition duration.In the case of the single shot HASTEsequence the base/phase resolutionor the field of view in phase encodingdirection can be reduced. In eithercase, increasing the bandwidth per

pixel shortens the acquisition dura-tion.

2 Note that even if the trigger threshold is setmanually, a learning phase is needed since thecentral position of the acceptance window isnot the sole function of the trigger threshold.The central position of the acceptance windowalways depends on statistic quantities calcula-

ted from the series of diaphragm positionsmeasured during the learning phase.

8/14/2019 2003 2005 Magnetom Flash 2

29/134


Figure 2 Timing diagram for the imaging phase of the navigator triggeredsequence. The thin blue curve is the diaphragm position as a function of time.The upper gray boxes visualize the acceptance window. The acceptancewindow is interrupted while the patient breathes in, since the system nevertriggers during inspiration. On the lower left side the navigators are shown:these are repeated at a constant interval Scout-TR to track the diaphragmposition. As soon as the detected diaphragm position falls within the acceptancewindow, the sequence stops repeating the navigator and executes the firstblock of the anatomical imaging sequence. In the case of the interleaved TSEsequence, a block acquires n echo trains one per slice of the current concate-nation. In the case of the single-shot HASTE sequence, the block acquires one

complete slice. Acquisition duration is the time needed to execute the block.400 ms after the anatomical imaging block is finished, the navigators arerepeated again until the trigger condition is fulfilled within the next breathingcycle.

Figure 3 Respiratory curve of the Trigger option. The turquoise dotted windowon the left (which shows the navigator position) marks the learning phase ofthe trigger algorithm. During the learning phase red boxes visualise the propo-sed anatomical data acquisition periods. The location of these boxes is basedon the parameter setting and the evaluation of the previous respiratory cycles.The parameter setting is fine for a certain respiratory cycle, if the data isacquired in the relaxed position near the end of expiration. If the horizontaledge-width of the red boxes is comparable or greater than one respiratoryperiod (horizontal distance from maximum inspiration to next maximuminspiration) the measurement must be stopped and the acquisition durationmust be reduced. On the right half of the figure, the respiratory trace during theimaging phase is shown. As soon as the system detects rising signal (onset of

expiration), the acceptance window is shown as a yellow box. If the detecteddiaphragm displacement (green curve) falls into the acceptance window, thebasic anatomical imaging block is executed. During anatomical data acquisitionthe respiratory curve is not continued. The number of acquired scans in relationto the total number of scans to be acquired is shown in the upper left corner(here Scan 6/34). The trigger period is the median temporal displacementbetween two trigger events. If the system triggers once per respiratory cycle,the trigger period is equal to the respiratory period. The last image text line inthe upper left corner shows the trigger threshold.

TECHNOLOGY CORNER

8/14/2019 2003 2005 Magnetom Flash 2

30/134

Proven Outcomes in MR.

Penetrating. Scrutinizing. Head to toe. Front

to back. And side to side. Tim (Total imaging matrix)

takes it all in. And in the process, opens up countless

new possibilities. Tim brings together, for the first

time ever, 76 matrix coil elements and up to 32 RF

channels. All of which can be freely combined in any

way. The highest signal-to-noise ratio possible today,

while still enabling seamless, whole-body imaging with

a total FoV of 205 cm (6 9). Tim is not just another

round of enhancements. But a transforming technology

that does so much more. So you can, too. See for

yourself at www.Siemens.com/Tim.

Siemens Medical Solutions that help

Tim

sees all.www.siemens.com/medical

We see a way to do whole-body imaging with MR in as little as 12 minutes

We see a way to seamlessly scan up to 205 cm with local coil quality

8/14/2019 2003 2005 Magnetom Flash 2

31/134

Results may vary. Data on file.

We see a way to evaluate systemic diseases in one MR exam without any patient or coil repositioning

We see a way to do MR imaging with an increased signal-to-noise of up to100%

8/14/2019 2003 2005 Magnetom Flash 2

32/134

This fall, Siemens will be introducing

Tim, a revolutionary new addition tothe traditional MRI process. Tim, theunique Total imaging matrix techno-logy, brings MR tomography greaterperformance than it has ever hadbefore.

In the past, MR technology waslimited to array coils offering a maxi-mum of only eight receiving channels.Until Siemens introduced the uniqueIPA (Integrated Panoramic Array)concept in 1997, the possibility of

varying the combination of coils forscanning larger areas of the body hasbeen very limited or even nonexistent.Now, for the first time in the historyof MR, the user can individually selectthe specific exams for the desiredanatomy and no longer has to dealwith patient repositioning or coilreconfigurations. Tim features asmany as 76 (!) seamlessly integratedcoil elements, along with exactly32 independent receiver channels.These can be used flexibly in any

combination to create a whole-bodyimaging matrix, supporting a totalfield of view of 205 cm (6 9).Conventional metastases evaluationof the whole body, however, requireschanging coils and repositioning thepatient for each anatomical area ofinterest, e.g. head, thoracic, abdomi-nal, pelvis, etc With the introductionof Tim, workflow and patient comfortlevel are vastly improved by simpli-fying the MR process and shorteningexamination times. Both the matrixcoils and the patient need only bepositioned once for all desired

exams, as multiple channelsallow a unique and almost unlimitedscanning flexibility.

Scan times cut in half

Usually, even with the existing coilconcept, only about a 150-cm field ofview could be scanned with surfacecoils and a sufficient signal-to-noise

ratio (SNR). For whole-body coverage,the patient needs to be repositioned,i.e. the coil setup has to be reconfigu-red. The only alternative, up to now,has been to perform whole-body MRimaging without surface coils, i.e.just with the integrated body coil,which, however, significantly reducesimage quality. Thus, MRI technicianshave always had to choose in thepast between total body coverageand adequate image quality. Timerases that difficult decision by

allowing up to 100 percent moreSNR. The result: up to 205 cm (6 9)can now be scanned at maximumSNR requiring only 12 minutes.Thats a phenomenal time reductionof more than 50 percent!

Tim can help making accurate dia-gnoses, even for extremely compleximaging needs. Tims flexibilitymeans that if something unexpectedis spotted, the region of interest canbe expanded instantly, without anyrepositioning of the patient or any

coil reconfigurations.

Unlimited parallel imaging

Tims reach is virtually boundless.Today, so-called Parallel Imaging(Parallel Acquisition Technique, PAT)already allows faster acquisition withhigh image quality. This is a distinctadvantage, for instance, in MRIs ofmoving organs, because motionartefacts are significantly reduced oreven eliminated. Unfortunately, thedisadvantages of PAT include limited

fields of view, usually based on thespecific PAT coils required, and afinite number of receiver channels.Here, Tim shows its real added value.It removes all those negatives byallowing Parallel Imaging along thepatients entire body in a total field ofview of 205 cm. The very highest PATfactors are now available in all threedimensions: from head to feet,anterior to posterior, and left to right.

That means that the highest acquisi-tion speeds and image resolutionscan be achieved without the need forany specific PAT coils.

Intelligent assistance

Tim goes even further to intelligentlyassist with MRIs: For the first time,it makes Parallel Imaging easy byrecommending the maximum PATfactors for whichever application is

selected. And the Tim Assistant helpsfinding the selected coil elements,correct patient position, and appro-priate MR protocol, assuring theintegrated Parallel Acquisition Tech-nique (iPAT) configuration for eachparticular need. Siemens continuesto uncompromisingly pursue its goalof truly optimizing workflow: WithTims new Intelligent Coil Control,technicians can control all coils, bothfixed and flexible, and their corres-ponding elements. This is all aimed at

making Parallel Imaging easy andmore efficient, as well as integratingit into the clinical routine.

In addition, with iPAT, Siemens offerstwo forms of PAT: GRAPPA (Genera-lized Autocalibrating Partial ParallelAcquisition) and mSENSE (modifiedSensitivity Encoding), thus increasi

2003 2005 Magnetom Flash 2

Documents