Evaluation of the ischemic penumbra focusing on the venous drainage: The role of susceptibility weighted imaging (SWI) in pediatric ischemic cerebral stroke

Page 1

ARTICLE IN PRESS+ModelNEURAD-407; No. of Pages 9

Journal of Neuroradiology (2013) xxx, xxx—xxx

Available online at

www.sciencedirect.com

ORIGINAL ARTICLE

Evaluation of the ischemic penumbra focusing on thevenous drainage: The role of susceptibility weightedimaging (SWI) in pediatric ischemic cerebral stroke

Avner Meoded, Andrea Poretti, Jane E. Benson, Aylin Tekes,Thierry A.G.M. Huisman ∗

Division of Pediatric Radiology, Russell H. Morgan Department of Radiology and Radiological Sciences, Johns Hopkins Hospital,Sheikh Zayed Tower, Room 4174, 1800, Orleans Street, Baltimore, MD 21287-0842, USA

KEYWORDSIschemic stroke;Children;Magnetic resonanceimaging;Susceptibilityweighted imaging;Ischemic penumbra

SummaryBackground and purpose: Susceptibility weighted imaging (SWI) allows the study of the intracra-nial venous vasculature based on the paramagnetic susceptibility effects of deoxygenated blood.Prominent hypointense draining veins have been revealed in ischemic brain tissue by SWI. Thegoal of our study was to evaluate whether a match or mismatch between territorial changes inthe venous drainage of ischemic brain tissue, as identified by SWI and diffusion restriction, canshow a ‘venous ischemic penumbra’.Materials and methods: Eight children with a confirmed diagnosis of acute pediatric arterialischemic stroke (PAIS) were included in this preliminary study. All had undergone an acutestandard magnetic resonance imaging (MRI) study with diffusion-weighted imaging (DWI) andSWI sequences. SWI scans were semi-quantitatively evaluated for signal intensity and caliber ofboth the intramedullary and sulcal veins. In addition, SWI abnormalities were compared withDWI images for match/mismatch of signal alterations, and the acute MRI data were comparedwith follow-up scans.Results: A total of 17 vascular territories showed infarction. SWI hypointensity in sulcal andintramedullary veins was found in 77% and 94% of the infarcted territories, respectively, whilethe caliber of the sulcal and intramedullary veins was increased in 64% and 88% of the infarctedareas, respectively. SWI/DWI match was observed in 88% of the vascular territories, whereas mis-match was noted in two; follow-up neuroimaging showed infarct progression into the mismatchareas.

Please cite this article in press as: Meoded A, et al. Evaluadrainage: The role of susceptibility weighted imaging (SWI) in

http://dx.doi.org/10.1016/j.neurad.2013.04.002

Conclusion: Our study showed that, in children, high-quality SWI studies focused on venousdrainage can provide important non-invasive data on critically perfused brain tissue at risk ofinfarct progression. SWI is therneuroimaging techniques for acu© 2013 Elsevier Masson SAS. All

∗ Corresponding author. Tel.: +1 410 955 6454; fax: +1 410 502 3633.E-mail address: [email protected] (T.A.G.M. Huisman).

0150-9861/$ – see front matter © 2013 Elsevier Masson SAS. All rights rehttp://dx.doi.org/10.1016/j.neurad.2013.04.002

efore a valuable MR tool that can be added to the battery of

tion of the ischemic penumbra focusing on the venouspediatric ischemic cerebral stroke. J Neuroradiol (2013),

te PAIS.rights reserved.

served.

Page 2

IN+ModelN

2

I

P1stecb

sprog[mwmatm

hratioetldris

sciDgdndvmpibTcamtpm

•

•

•

M

S

I

•

•

•

e

aJpcf

r

S

Anottbpso2(tTnss

D

D(iatt

ARTICLEEURAD-407; No. of Pages 9

ntroduction

ediatric arterial ischemic stroke (PAIS) affects 2 to3/100,000 children per year [1] and is associated withevere long-term neurological sequelae [2—8]. Althoughherapeutic guidelines for acute PAIS are still lacking [9],mpirical thrombolytic therapy is increasingly being used inhildren [2,10—14]. However, the diagnosis of PAIS must firste confirmed and mimickers ruled out.

Conventional and advanced magnetic resonance (MR)equences such as diffusion-weighted imaging (DWI) anderfusion-weighted imaging (PWI) can confirm PAIS andule out mimickers [15]. In addition, combined analysisf DWI/PWI data can identify tissue at risk of pro-ression to infarction—–the so-called ‘ischemic penumbra’16]. The ischemic core is thought to be represented byatching areas of hypoperfusion and restricted diffusion,hile the penumbra shows critical hypoperfusion withoutatching diffusion restriction (DWI/PWI mismatch). Early

nd reliable identification of potentially salvageable brainissue is essential and could serve as a guide to treat-ent.Dynamic susceptibility contrast-enhanced (DSC) PWI is,

owever, rarely performed in children. DSC PWI requiresapid intravenous bolus injection of paramagnetic contrastgents and, therefore, a large venous catheter. In addi-ion, DSC PWI relies on the determination of an arterialnput function (AIF), chosen from among the main arteriesf the circle of Willis; as a result, there is no unique AIF forvery pixel [17]. Arterial spin labeling (ASL) is an alternativeool for assessing cerebral blood flow, but its accessibility isimited and its validation in PAIS has not yet been clearlyemonstrated [18]. Also, PWI focuses primarily on the arte-ial side of ischemic injury, whereas important hemodynamicnformation could also be collected by studying the venouside of ischemia.

SWI may serve as a widely available, easy-to-performequence that offers important information on venous vas-ulature. Its feasibility has already been demonstratedn pediatric and neonatal neurological disorders [19—24].ifferences in the susceptibility characteristics of oxy-enated and deoxygenated hemoglobin lead to a phaseifference between regions with deoxygenated blood andormally oxygenated surrounding tissue. Depending on theegree of blood oxygenation, SWI can display veins witharying degrees of signal loss. Greater conspicuity of theedullary draining veins has been shown in areas of impairederfusion, a result thought to be linked to an increasen deoxyhemoglobin concentration due to an imbalanceetween oxygen supply and demand in that area [25—27].his suggests that SWI may be able to indirectly detectritical brain perfusion by focusing on the venous ‘output’spects of brain perfusion. Correlating DWI and SWI imagesay also identify brain tissue at risk of infarction and poten-

ially identify a difference in the quality of the ischemicenumbra compared with more arterial-weighted DWI/PWIismatches.The aims of the present retrospective study were to:

Please cite this article in press as: Meoded A, et al. Evaluadrainage: The role of susceptibility weighted imaging (SWI) inhttp://dx.doi.org/10.1016/j.neurad.2013.04.002

semi-quantitatively evaluate venous drainage on SWIimages;

saTa

PRESSA. Meoded et al.

identify brain tissue at risk of infarction by comparing DWIand SWI images;

evaluate the accuracy of penumbra identification onfollow-up MRI scans in PAIS.

aterials and methods

ubjects

nclusion criteria for this study were:

confirmed diagnosis of acute PAIS (compatible clinicalfindings and matching MRI findings);

availability of at least one standard departmental pro-tocol MRI study, including T1- and T2-weighted MRI andDWI/SWI;

MRI study performed within 72 h of the onset of symptoms.

Patients with hemorrhagic or metabolic stroke werexcluded from the study.

Eligible patients were found through a search of our pedi-tric neuroradiology database for the time period betweenanuary 2009 and March 2011. Clinical histories of theatients were reviewed for stroke etiology, symptoms andlinical findings on admission, and the timing of acute andollow-up neuroimaging.

The study protocol had the approval of the institutionalesearch ethics board.

usceptibility weighted imaging

ll MRI studies were acquired on a 1.5 T clinical MRI scan-er (Magnetom Avanto, Siemens, Erlangen, Germany) usingur standard departmental protocol. For SWI sequences,he following parameters were routinely used: TR (repeti-ion time) = 48 ms; TE (echo time) = 40 ms; flip angle = 15◦;andwidth = 80 kHz; slice thickness = 1.2 mm with 128 sliceser slab; FOV (field of view) = 146 × 180 mm; and matrixize = 256 × 512. Acquisition time was 4.57 min with the usef integrated parallel acquisition techniques (iPAT) factor. During post-processing, minimum intensity projectionminIP) images were reconstructed, with an effective minIPhickness of 8 mm for neonates and 16 mm for older patients.he smaller effective minIP thickness of 8 mm was used ineonates to limit partial volume effects due to small brainize and any subsequent anatomical misregistration of ves-els that might masquerade as pathology.

iffusion-weighted images

WI maps were derived from diffusion tensor imagingDTI) scans. A single-shot spin-echo echoplanar imag-ng (EPI) axial DTI sequence with diffusion gradientslong 20 orthogonal directions was performed. An effec-ive high b value of 1000 s/mm2 was used for each ofhe 20 diffusion-encoding directions. An additional mea-

tion of the ischemic penumbra focusing on the venous pediatric ischemic cerebral stroke. J Neuroradiol (2013),

urement without diffusion weighting (b = 0 s/mm2) waslso performed. Acquisition parameters were TR = 7100 ms,E = 84 ms, slice thickness = 2.5 mm, FOV = 240 × 240 mmnd matrix size = 192 × 192. Parallel-imaging iPAT = 2 with

Page 3

ARTICLE IN PRESS+ModelNEURAD-407; No. of Pages 9

The role of susceptibility weighted imaging in pediatric ischemic stroke 3

Table 1 Demographic and imaging data for eight children with arterial ischemic stroke.

Patient Age/gender Time of first imaging(from symptom onset)

Infarcted vascular territory Time of follow-up imaging(from first imaging)

1 7 years/M < 6 h Left MCA (proximal + distal), right ACA 3 days/CT< 72 h Right distal MCA

2 1 month/F < 24 h Right distal MCA n.a.3 2 days/F < 24 h Right distal MCA, right PCA, left ACA 7 days/MR4 15 years/F < 24 h Right distal MCA 3 months/MR5 2 days/M < 24 h Left MCA (proximal + distal) n.a.6 4 months/F < 72 h Left distal MCA n.a.7 16 years/F < 24 h Right distal MCA 8 months/MR8 2 days/F < 24 h Right MCA (proximal + distal), left

distal MCA, right PCAn.a.

cere

fiaaitf

•

••

s

R

Ecp1fttatwsora

(

M: male; F: female; MCA: middle cerebral artery; ACA: anteriorartery; MR: magnetic resonance; n.a.: not available.

generalized autocalibrating partially parallel acquisition(GRAPPA) reconstruction was used. Acquisition time was5.21 min.

Image analysis

Analysis of all images was performed by two experiencedpediatric neuroradiologists (A.M. and T.H.) with consensus.For each cerebral hemisphere, four vascular territories weredefined, corresponding to areas perfused by the anteriorcerebral artery (ACA), proximal and distal middle cerebralartery (MCA), and posterior cerebral artery (PCA). Eachpatient’s ischemic lesion and respective arterial territorywere studied using conventional T1- and T2-weighted MRIand minIP images from SWI. Also evaluated were SWI sig-nal intensity and caliber (prominence) of the intramedullaryand sulcal veins in both hemispheres. Signal intensity ofthe intramedullary and sulcal veins was graded as similar(0), lower (1) or markedly lower (2) compared with ves-sels in matching non-infarcted regions in the contralateralhemisphere. Caliber of the intramedullary and sulcal veinswas graded as similar (0), mildly wider (up to twofold; 1)or markedly wider (more than twofold; 2) compared withvessels in matching non-infarcted regions in the contralat-eral hemisphere. In cases of bilateral strokes in the sameregion, comparisons were made in the anterior or posteriorcirculation territory according to primary lesion distribution.

To identify brain tissue at risk of infarction, minIPimages were fused/overlaid onto DWI images to obtain asingle DWI/SWI image, using an offline picture archivingand communication system (PACS) workstation. These fusedDWI/SWI images were used to evaluate the match or mis-match of DWI (restricted diffusion) and SWI (hypointensesignals in the involved intramedullary or sulcal veins), whichwas categorized as:

• area of restricted diffusion is smaller than the area with

Please cite this article in press as: Meoded A, et al. Evaluadrainage: The role of susceptibility weighted imaging (SWI) in

http://dx.doi.org/10.1016/j.neurad.2013.04.002

hypointense venous signals;• both areas are the same size;• area of restricted diffusion is larger than the area with

hypointense venous signals.

r(st

bral artery; CT: computed tomography; PCA: posterior cerebral

To test their accuracy in identifying critical brain per-usion, the acute SWI scans were compared with follow-upmages. Also evaluated was the match or mismatch betweenreas with hypointense venous signals on acute SWI imagesnd areas of gliosis or encephalomalacia on follow-upmages. The latter findings were assumed to depict infarctedissue. The match or mismatch between acute SWI andollow-up images was categorized as:

area with hypointense venous signals on acute SWI issmaller than the area with chronic ischemic changes onfollow-up images;

both areas are the same size; area with hypointense venous signals on acute SWI is

larger than the area with chronic ischemic changes onfollow-up images.

Also, additional SWI findings such as hemorrhagic conver-ion within the infarct area were assessed.

esults

ight children (six girls, two boys) met our study inclusionriteria (Table 1). Median age of the patients at acute strokeresentation was 0.2 years (mean 4.8 years, range 2 days to6 years). Clinical presentation included focal seizures inour patients, hemiparesis in two and paresthesia in one. Inhe remaining child with tetralogy of Fallot, PAIS was asymp-omatic and accidental puncture of the right internal carotidrtery (ICA) instigated an imaging work-up that led to detec-ion of a ‘clinically silent’ infarction. The etiology of PAISas cardioembolic in three patients due to partial occlu-

ion of the right ICA, and because of accidental puncturef the right ICA (as reported above) in one patient. Etiologyemained unclear in four children. No child had receivedcute thrombolytic therapy.

Time to MRI after symptom onset was 6 h in one patient12.5%), 6—24 h in six other patients (75%) and 24—72 h in the

tion of the ischemic penumbra focusing on the venouspediatric ischemic cerebral stroke. J Neuroradiol (2013),

emaining patient (12.5%; Table 1, Figs. 1—4). One patient#1) was scheduled for MRI because of a suspected right MCAtroke (onset of symptoms about 48 h before MRI), but lesshan 6 h before the neuroimaging took place, he developed

Page 4

ARTICLE IN PRESS+ModelNEURAD-407; No. of Pages 9

4 A. Meoded et al.

Figure 1 In patient #1 (7-year-old boy), axial (A) diffusion-weighted imaging (DWI), (B) apparent diffusion coefficient (ADC) mapsand (C) minimum intensity projection (minIP) susceptibility weighted imaging (SWI) show areas of restricted diffusion in the rightinsular cortex representing acute ischemia. Several small areas with restricted diffusion can also be seen in the cortex of theright frontal and left temporal lobes, and in the left basal ganglia. Corresponding signal characteristics on SWI include: (1) widehypointense intramedullary veins (short white arrow) in the left hemisphere matching the diffusion abnormality; (2) an area ofprominent hypointense sulcal veins (long white arrows) along the left hemisphere and right frontal lobe that is greater than thearea of DWI abnormality; and (3) complete attenuation of the SWI signal from intramedullary/sulcal veins in the occipital regionsb essiod

aoptti

swT9

saiIoco

ilaterally. D. Follow-up axial CT shows significant infarct progristribution matching the initially SWI hypointense sulcal veins.

new stroke involving the left MCA and right ACA. A totalf 17 vascular territories were involved on SWI; in fouratients (50%), one vascular territory was involved, whilewo patients (25%) had four affected territories each, andhe two remaining patients had two and three territoriesnfarcted, respectively (12.5%).

Results of the semi-quantitative analysis of signal inten-

Please cite this article in press as: Meoded A, et al. Evaluadrainage: The role of susceptibility weighted imaging (SWI) inhttp://dx.doi.org/10.1016/j.neurad.2013.04.002

ity and caliber of the intramedullary and sulcal veinsithin infarcted areas on minIP images are summarized in

able 2. Hypointense signals on SWI images were found in4% (16/17) of the intramedullary and 76% (13/17) of the

ivwI

n especially in the left hemisphere and right frontal lobe in a

ulcal veins draining the infarcted vascular territories. Inddition, increased caliber was found in 88% (15/17) of thentramedullary and 71% (12/17) of the sulcal draining veins.n the acute neuroimaging studies, a match between areasf restricted diffusion on DWI and apparent diffusion coeffi-ient (ADC) maps and areas with hypointense venous signalsn SWI images of intramedullary and sulcal veins was found

tion of the ischemic penumbra focusing on the venous pediatric ischemic cerebral stroke. J Neuroradiol (2013),

n 15 of the 17 (88%) infarcted vascular territories. In twoascular territories (12%), the area of restricted diffusionas smaller than the area with hypointense venous signals.

nterestingly, these territories in patient #1 had the shortest

Page 5

ARTICLE IN+ModelNEURAD-407; No. of Pages 9

The role of susceptibility weighted imaging in pediatric ischemic

Figure 2 Fused axial DWI/SWI images for patient #1 showhypointense intramedullary veins on SWI matching the area ofrestricted diffusion on DWI in the left MCA territory. The dilated

aatic

D

MfaaplwBbwvohupPEep

M

SWI hypointense sulcal veins are draining an area larger than theregion of restricted diffusion in the left MCA territory.

clinical prodromal phase (< 6 h). Hyperintense signals in the

Please cite this article in press as: Meoded A, et al. Evaluadrainage: The role of susceptibility weighted imaging (SWI) in

http://dx.doi.org/10.1016/j.neurad.2013.04.002

ischemic cortex were also seen on SWI in four involvedvascular territories (24%), two of which had hemorrhagiccomponents (12%).

•

Table 2 Evaluation of imaging findings from 17 vascular territori

SWI findings Veins Grade Distal MCA Pr

Right(n = 6)

Left(n = 4)

R(n

Signalintensity

Sulcal 0 2 1 0

1 2 2 1

2 2 1 0

Intramedullary 0 0 0 0

1 4 3 0

2 2 1 1

Caliberevaluation

Sulcal 0 2 2 0

1 2 1 1

2 2 1 0

Intramedullary 0 0 1 0

1 4 2 0

2 2 1 1

DWI/SWImismatch

DWI = SWI 6 3 1

DWI > SWI 0 0 0

DWI < SWI 0 1 0

SWI: susceptibility weighted imaging; MCA: middle cerebral artery; ACdiffusion-weighted imaging.

PRESS stroke 5

Follow-up studies were available for four children (50%)nd included eight infarcted areas. These were performedn average of 85 days after acute neuroimaging (medianime 48.5 days, range 3 days to 8 months). In all regions, thenitial venous signal hypointensity on SWI images matchedhronic ischemic changes.

iscussion

ultimodal MRI is the neuroimaging investigation of choiceor early evaluation of PAIS. While T2-weighted and fluid-ttenuated inverted recovery (FLAIR) images may showcutely infarcted tissue within 12—24 h of injury, DWI hasroven helpful by identifying ischemic lesions even ear-ier than that—–specifically, within hours in neonates andithin minutes in older children after the initial insult.y combining DWI and PWI, the core of the infarction cane differentiated from the ischemic penumbra, the regionith critical residual perfusion in which brain tissue is stilliable but at risk if infarction progresses. Matching areasf restricted diffusion, as revealed by DWI, with areas ofypoperfused brain, as shown by PWI, correlate with thesually irreversibly injured ischemic core. The ischemicenumbra is characterized by areas of hypoperfusion onWI, but no matching areas of restricted diffusion on DWI.arly identification of the ischemic penumbra could guidearly treatment options to limit or even prevent infarctrogression.

SWI has also been shown to be a useful complementaryR sequence in patients with ischemic stroke by:

tion of the ischemic penumbra focusing on the venouspediatric ischemic cerebral stroke. J Neuroradiol (2013),

detecting hemorrhagic components within the region ofinfarction to help distinguish ischemic from hemorrhagicstroke [28];

es in children with acute ischemic stroke.

oximal MCA ACA PCA Total (%)

ight = 1)

Left(n = 2)

Right(n = 1)

Left(n = 1)

Right(n = 2)

Left(n = 0)

1 0 0 0 0 240 0 0 0 0 291 1 1 2 0 470 1 0 0 0 61 0 0 1 0 531 0 1 1 0 41

2 0 0 0 0 350 0 0 0 0 240 1 1 2 0 410 1 0 0 0 121 0 0 1 0 471 0 1 1 0 41

2 0 1 2 0 880 0 0 0 0 00 1 0 0 0 12

A: anterior cerebral artery; PCA: posterior cerebral artery; DWI:

Page 6

ARTICLE IN PRESS+ModelNEURAD-407; No. of Pages 9

6 A. Meoded et al.

Figure 3 In patient #3 (2-day-old baby girl), (A) axial ADC maps and minIP SWI show extensive bilateral acute ischemia involvingthe ACA and PCA territories in the right cerebral hemisphere, and the MCA territory in the left cerebral hemisphere. SWI revealedprominent hypointense veins throughout the right hemisphere matching the restricted diffusion area; prominent venous structuresc

•

••

•

pcwcoaonip

grT[ircfitte[fo

ItvbivnIltgr[pe

wioAsnioT[ocippc

an also be seen in the left MCA territory.

demonstrating hypointense signals in the veins draininghypoperfused areas and directing the use of PWI [29];

detecting acute occlusive arterial thromboemboli [30]; quantifying microbleeds that predict the probability of

potential hemorrhagic transformation prior to throm-bolytic treatment [31];

detecting early hemorrhagic complications after intra-arterial thrombolysis [30].

In contrast to DSC PWI, SWI also has the potential torovide valuable information on the ‘venous side’ of criti-al brain perfusion and to reveal information that otherwiseould remain undetected by the more ‘arterial-weighted’ontrast-enhanced PWI approach. Also, most of the datan the role of SWI in stroke have been collected from thedult stroke population, whereas the present report focusesn the role of SWI in the evaluation of cerebral hemody-amics and tissue viability in pediatric patients with acuteschemic stroke. Indeed, SWI may be especially helpful in theediatric population as no contrast injection is necessary.

In stroke, focal brain ischemia results in impaired oxy-en metabolism due to inadequate (‘misery’) perfusion inelation to the high oxygen demands of cerebral tissue.his results in an increased oxygen extraction fraction (OEF)32,33] in an attempt to compensate for the reductionn arterial cerebral blood flow that may, at least tempo-arily, preserve the rate of regional oxygen metabolism. Thisompensatory mechanism increases the deoxyhemoglobinraction at the same time that oxyhemoglobin is decreasingn the veins that drain areas of impaired perfusion. Givenhe paramagnetic properties of deoxyhemoglobin, its rela-ive increase in the veins draining ischemic tissue results in

Please cite this article in press as: Meoded A, et al. Evaluadrainage: The role of susceptibility weighted imaging (SWI) inhttp://dx.doi.org/10.1016/j.neurad.2013.04.002

nhanced signal loss or hypointensity of those veins on SWI34—40]. In our present study, hypointense SWI signals wereound in 94% (16/17) of the intramedullary and 76% (13/17)f the sulcal veins draining infarcted vascular territories.

p

ad

n addition, increased caliber was noted in 88% (15/17) ofhe intramedullary and 71% (12/17) of the sulcal drainingeins. This prominence of the draining vessels may haveeen due to the more pronounced paramagnetic ‘bloom-ng’ effect of deoxygenated blood and/or slowing of corticalenous outflow. However, a few vascular territories showedormal SWI signal intensity and/or caliber in draining veins.n fact, normal SWI signal intensity was present in all vascu-ar territories where stroke onset was more than 24 h prioro neuroimaging. Normalization of the venous blood oxy-enation level-dependent (BOLD) signal has previously beeneported in adult stroke patients 12 h after stroke onset17]. Cell swelling and settling of edema were proposed asossible explanations of a decreased BOLD signal in strokevolution.

Only one of our patients had two vascular territoriesith multiple dilated SWI hypointense sulcal veins drain-

ng a brain area that was significantly larger than the areaf restricted diffusion seen on DWI (DWI/SWI mismatch).ccording to our hypothesis, such a mismatch could repre-ent brain tissue at risk of infarction. This patient receivedo reperfusion therapy, and follow-up neuroimaging showednfarct progression that exactly matched the initial areaf the SWI abnormality (dilated hypointense sulcal veins).hese findings as well as similar ones from previous studies25—27,41] support the role of SWI in predicting the extentf impaired perfusion in acute PAIS without the need forontrast media. In particular, one study of adults with acuteschemic stroke compared the ability of SWI versus PWI toredict stroke evolution [42]. The authors found that SWI canrovide perfusion information comparable to PWI, and con-luded that SWI/DWI mismatch may indicate the ischemic

tion of the ischemic penumbra focusing on the venous pediatric ischemic cerebral stroke. J Neuroradiol (2013),

enumbra and predict stroke evolution.The DWI/PWI mismatch concept was introduced to offer

n estimate of the ischemic penumbra. The perfusioneficit, measured as either a prolonged mean transit time

Page 7

ARTICLE IN PRESS+ModelNEURAD-407; No. of Pages 9

The role of susceptibility weighted imaging in pediatric ischemic stroke 7

Figure 4 In patient #7 (16-year-old girl), axial (A) ADC maps and (B) contrast minIP SWI show acute infarction in the right MCAterritory. SWI shows hypointensity of the intrasulcal venous structures in the corresponding infarcted area (B, arrows). C. Fused

nse inalac

[ic

wisemfmmpfia

axial DWI/SWI shows a match between areas with SWI hypointeD. Follow-up axial T2-weighted image shows gliosis/encephalom

or delay in time to peak, appears to be greater than the DWIdeficit in 70—80% of patients imaged within the first 6 h ofstroke onset [43—48]. The use of DWI/PWI mismatch helpsto select those patients who might benefit from early reper-fusion therapy. When cerebral perfusion was evaluated byfocusing on the venous drainage of ischemic or critically per-fused brain tissue, a DWI/SWI match was observed in 88% ofthe vascular territories. The DWI/SWI match was not surpris-ing as neuroimaging was performed more than 6 h after theonset of symptoms in nearly all of our patients. It could beargued that, at the time of neuroimaging, the brain tissue atrisk was already infarcted. Unfortunately, delayed diagnosiswith subsequent neurological sequelae and impaired qual-ity of life are well recognized in PAIS. Non-specific clinicalsigns, the high prevalence of stroke mimics and particularlythe low index of suspicion for a diagnosis of PAIS are themost common causes of delayed diagnosis [49].

Please cite this article in press as: Meoded A, et al. Evaluadrainage: The role of susceptibility weighted imaging (SWI) in

http://dx.doi.org/10.1016/j.neurad.2013.04.002

Two vascular territories showed hemorrhagic componentswithin the ischemic area. SWI had higher sensitivity forhemorrhage compared with gradient echo sequences andcomputed tomography (CT) in acute and subacute stroke

C

Ef

tramedullary and sulcal veins and restricted diffusion on DWI.ia in the area corresponding to the acute SWI signal alteration.

50]. The detection of hemorrhagic components within annfarcted area is important and may represent a contraindi-ation for thrombolysis.

Our present study had several limitations. Although thisas the largest reported cohort of children with acute

schemic stroke studied by SWI, the number of patients wastill small. Also, the study was retrospective; images werevaluated using a semi-quantitative approach (approximateeasurements may be difficult to reproduce), and the dif-

erence in slice thickness between DWI and minIP SWI imagesay have caused partial misregistration. In addition, for theajority of our patients, the first neuroimaging study waserformed at least 6 h after the onset of symptoms, andollow-up neuroimaging was only available for four children,ncluded CT and was performed at variable times after thecute study.

tion of the ischemic penumbra focusing on the venouspediatric ischemic cerebral stroke. J Neuroradiol (2013),

onclusion

mergent thrombolytic therapy is the treatment of choiceor acute stroke, but the risks mandate restricting its use

Page 8

IN+ModelN

8

tivsinawatlad

D

Tc

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

ARTICLEEURAD-407; No. of Pages 9

o only those patients most likely to benefit. A non-invasivemaging tool that can immediately identify and quantify sal-ageable brain tissue early on is very much needed. Ourtudy has shown that, in children, high-quality SWI stud-es that focus on venous drainage can provide importanton-invasive data concerning critically perfused brain tissuet risk of infarct progression, especially and/or exclusivelyith thrombolytic delay. This suggests that SWI is a valu-ble MR tool to be added to the battery of neuroimagingechniques for acute PAIS. Future prospective studies with aarger number of children, quantitative analysis of SWI data,nd comparisons of SWI and PWI findings are now needed toefinitively evaluate the role of SWI in PAIS.

isclosure of interest

he authors declare that they have no conflicts of interestoncerning this article.

eferences

[1] Amlie-Lefond C, Sebire G, Fullerton HJ. Recent develop-ments in childhood arterial ischaemic stroke. Lancet Neurol2008;7:425—35.

[2] Amlie-Lefond C, deVeber G, Chan AK, et al. Use of alteplasein childhood arterial ischaemic stroke: a multicentre, observa-tional, cohort study. Lancet Neurol 2009;8:530—6.

[3] deVeber GA, MacGregor D, Curtis R, Mayank S. Neurologic out-come in survivors of childhood arterial ischemic stroke andsinovenous thrombosis. J Child Neurol 2000;15:316—24.

[4] Fullerton HJ, Chetkovich DM, Wu YW, Smith WS, Johnston SC.Deaths from stroke in US children, 1979 to 1998. Neurology2002;59:34—9.

[5] Gabis LV, Yangala R, Lenn NJ. Time lag to diagnosis of stroke inchildren. Pediatrics 2002;110:924—8.

[6] Ganesan V, Hogan A, Shack N, Gordon A, Isaacs E, Kirkham FJ.Outcome after ischaemic stroke in childhood. Dev Med ChildNeurol 2000;42:455—61.

[7] Lynch JK, Hirtz DG, DeVeber G, Nelson KB. Report ofthe National Institute of Neurological Disorders and Strokeworkshop on perinatal and childhood stroke. Pediatrics2002;109:116—23.

[8] Rafay MF, Pontigon AM, Chiang J, et al. Delay to diag-nosis in acute pediatric arterial ischemic stroke. Stroke2009;40:58—64.

[9] Jordan LC. Stroke: thrombolysis in childhood arterial ischemicstroke. Nat Rev Neurol 2009;5:473—4.

10] Carlson MD, Leber S, Deveikis J, Silverstein FS. Successful useof rt-PA in pediatric stroke. Neurology 2001;57:157—8.

11] Benedict SL, Ni OK, Schloesser P, White KS, Bale Jr JF. Intra-arterial thrombolysis in a 2-year-old with cardioembolic stroke.J Child Neurol 2007;22:225—7.

12] Janjua N, Nasar A, Lynch JK, Qureshi AI. Thrombolysis forischemic stroke in children: data from the nationwide inpatientsample. Stroke 2007;38:1850—4.

13] Jordan LC. Assessment and treatment of stroke in children.Curr Treat Options Neurol 2008;10:399—409.

14] Amlie-Lefond C, Gill JC. Pharmacology in childhood arterialischemic stroke. Semin Pediatr Neurol 2010;17:237—44.

15] Shellhaas RA, Smith SE, O’Tool E, Licht DJ, Ichord RN. Mimicsof childhood stroke: characteristics of a prospective cohort.

Please cite this article in press as: Meoded A, et al. Evaluadrainage: The role of susceptibility weighted imaging (SWI) inhttp://dx.doi.org/10.1016/j.neurad.2013.04.002

Pediatrics 2006;118:704—9.16] Gadian DG, Calamante F, Kirkham FJ, et al. Diffusion and per-

fusion magnetic resonance imaging in childhood stroke. J ChildNeurol 2000;15:279—83.

[

PRESSA. Meoded et al.

17] Viallon M, Altrichter S, Pereira VM, et al. Combined use ofpulsed arterial spin-labeling and susceptibility-weighted imag-ing in stroke at 3T. Eur Neurol 2010;64:286—96.

18] Huisman TA, Sorensen AG. Perfusion-weighted magnetic res-onance imaging of the brain: techniques and application inchildren. Eur Radiol 2004;14:59—72.

19] Thomas B, Somasundaram S, Thamburaj K, et al. Clinical appli-cations of susceptibility weighted MR imaging of the brain - apictorial review. Neuroradiology 2008;50:105—16.

20] Tong KA, Ashwal S, Obenaus A, Nickerson JP, Kido D, HaackeEM. Susceptibility-weighted MR imaging: a review of clinicalapplications in children. AJNR Am J Neuroradiol 2008;29:9—17.

21] Haacke EM, Mittal S, Wu Z, Neelavalli J, Cheng YC.Susceptibility-weighted imaging: technical aspects and clinicalapplications, part 1. AJNR Am J Neuroradiol 2009;30:19—30.

22] Hodel J, Rodallec M, Gerber S, Blanc R, Maraval A, Caron S,et al. Susceptibility weighted magnetic resonance sequencesSWAN, SWI and VenoBOLD. Technical aspects and clinical appli-cations. J Neuroradiol 2012;39:71—86.

23] Verschuuren S, Poretti A, Buerki S, Lequin MH, Huisman TA.Susceptibility weighted imaging of the pediatric brain. AJR AmJ Roentgenol 2011;198:W440—9.

24] Meoded A, Poretti A, Northington FJ, Tekes A, Intrapiromkul J,Huisman TA. Susceptibility weighted imaging of the neonatalbrain. Clin Radiol 2012;67:793—801.

25] Hermier M, Nighoghossian N. Contribution of susceptibility-weighted imaging to acute stroke assessment. Stroke2004;35:1989—94.

26] Kesavadas C, Santhosh K, Thomas B, et al. Signal changes in cor-tical laminar necrosis-evidence from susceptibility-weightedmagnetic resonance imaging. Neuroradiology 2009;51:293—8.

27] Santhosh K, Kesavadas C, Thomas B, Gupta AK, ThamburajK, Kapilamoorthy TR. Susceptibility weighted imaging: a newtool in magnetic resonance imaging of stroke. Clin Radiol2009;64:74—83.

28] Wycliffe ND, Choe J, Holshouser B, Oyoyo UE, Haacke EM,Kido DK. Reliability in detection of hemorrhage in acutestroke by a new three-dimensional gradient recalled echosusceptibility-weighted imaging technique compared to com-puted tomography: a retrospective study. J Magn Reson Imaging2004;20:372—7.

29] Kesavadas C, Thomas B, Pendharakar H, Sylaja PN. Suscep-tibility weighted imaging: does it give information similarto perfusion weighted imaging in acute stroke? J Neurol2011;258:932—4.

30] Mittal S, Wu Z, Neelavalli J, Haacke EM. Susceptibility-weighted imaging: technical aspects and clinical applications,part 2. AJNR Am J Neuroradiol 2009;30:232—52.

31] Akter M, Hirai T, Hiai Y, et al. Detection of hemorrhagichypointense foci in the brain on susceptibility-weighted imag-ing clinical and phantom studies. Acad Radiol 2007;14:1011—9.

32] Heiss WD, Graf R, Wienhard K, et al. Dynamic penumbrademonstrated by sequential multitracer PET after middle cere-bral artery occlusion in cats. J Cereb Blood Flow Metab1994;14:892—902.

33] Sobesky J, Zaro Weber O, Lehnhardt FG, et al. Does the mis-match match the penumbra? Magnetic resonance imaging andpositron emission tomography in early ischemic stroke. Stroke2005;36:980—5.

34] Ogawa S, Lee TM, Kay AR, Tank DW. Brain magnetic resonanceimaging with contrast dependent on blood oxygenation. ProcNatl Acad Sci U S A 1990;87:9868—72.

35] Kavec M, Grohn OH, Kettunen MI, Silvennoinen MJ, PenttonenM, Kauppinen RA. Use of spin echo T(2) BOLD in assessment of

tion of the ischemic penumbra focusing on the venous pediatric ischemic cerebral stroke. J Neuroradiol (2013),

cerebral misery perfusion at 1.5 T. MAGMA 2001;12:32—9.36] Lee JM, Vo KD, An H, et al. Magnetic resonance cerebral

metabolic rate of oxygen utilization in hyperacute strokepatients. Ann Neurol 2003;53:227—32.

Page 9

IN+Model

emic

[

[

[

[

[

[

ARTICLENEURAD-407; No. of Pages 9

The role of susceptibility weighted imaging in pediatric isch

[37] Rauscher A, Sedlacik J, Barth M, Haacke EM, Reichen-bach JR. Nonnvasive assessment of vascular architecture andfunction during modulated blood oxygenation using suscepti-bility weighted magnetic resonance imaging. Magn Reson Med2005;54:87—95.

[38] Geisler BS, Brandhoff F, Fiehler J, et al. Blood-oxygen-level-dependent MRI allows metabolic description of tissue at risk inacute stroke patients. Stroke 2006;37:1778—84.

[39] Fushimi Y, Miki Y, Mori N, et al. Signal changes in the brain onsusceptibility-weighted imaging under reduced cerebral bloodflow: a preliminary study. J Neuroimaging 2010;20:255—9.

[40] Haacke EM, Tang J, Neelavalli J, Cheng YC. Susceptibilitymapping as a means to visualize veins and quantify oxygensaturation. J Magn Reson Imaging 2010;32:663—76.

[41] Chalian M, Tekes A, Meoded A, Poretti A, Huisman TA.Susceptibility-weighted imaging (SWI): a potential non-invasiveimaging tool for characterizing ischemic brain injury? J Neuro-radiol 2011;38:187—90.

[42] Kao HW, Tsai FY, Hasso AN. Predicting stroke evolution: compar-ison of susceptibility-weighted MR imaging with MR perfusion.

Please cite this article in press as: Meoded A, et al. Evaluadrainage: The role of susceptibility weighted imaging (SWI) in

http://dx.doi.org/10.1016/j.neurad.2013.04.002

Eur Radiol 2012;22:1397—403.[43] Ogata J, Yutani C, Imakita M, et al. Hemorrhagic infarct of the

brain without a reopening of the occluded arteries in cardioem-bolic stroke. Stroke 1989;20:876—83.

[

PRESS stroke 9

44] Sorensen AG, Buonanno FS, Gonzalez RG, et al. Hyperacutestroke: evaluation with combined multisection diffusion-weighted and hemodynamically weighted echo-planar MRimaging. Radiology 1996;199:391—401.

45] Patel MR, Edelman RR, Warach S. Detection of hyperacuteprimary intraparenchymal hemorrhage by magnetic resonanceimaging. Stroke 1996;27:2321—4.

46] Baird AE, Benfield A, Schlaug G, et al. Enlargementof human cerebral ischemic lesion volumes measured bydiffusion-weighted magnetic resonance imaging. Ann Neurol1997;41:581—9.

47] Barber PA, Darby DG, Desmond PM, et al. Prediction of strokeoutcome with echoplanar perfusion- and diffusion-weightedMRI. Neurology 1998;51:418—26.

48] van Everdingen KJ, van der Grond J, Kappelle LJ, Ramos LM,Mali WP. Diffusion-weighted magnetic resonance imaging inacute stroke. Stroke 1998;29:1783—90.

49] McGlennan C, Ganesan V. Delays in investigation and manage-ment of acute arterial ischaemic stroke in children. Dev MedChild Neurol 2008;50:537—40.

tion of the ischemic penumbra focusing on the venouspediatric ischemic cerebral stroke. J Neuroradiol (2013),

50] Goos JD, van der Flier WM, Knol DL, et al. Clinicalrelevance of improved microbleed detection by susceptibility-weighted magnetic resonance imaging. Stroke 2011;42:1894—900.