Prediction of Early Neurological Deterioration UsingDiffusion- and Perfusion-Weighted Imaging in Hyperacute

Middle Cerebral Artery Ischemic StrokeJuan F. Arenillas, MD; Álex Rovira, MD; Carlos A. Molina, MD; Elisenda Grivé, MD;

Joan Montaner, MD; José Álvarez-Sabín, MD, PhD

Background and Purpose—Early neurological deterioration (END) occurs in approximately one third of all ischemicstroke patients and is associated with a poor outcome. Our study sought to assess the value of ultra-early MRI in theprediction of END in stroke patients.

Methods—Between August 1999 and November 2001, 38 stroke patients with a proven middle cerebral artery (MCA) orintracranial internal carotid artery (ICA) occlusion on MR angiography underwent perfusion-weighted imaging (PWI)and diffusion-weighted imaging (DWI) within 6 hours after onset, and 30 fulfilled all inclusion criteria. Control DWIand MR angiography were performed between days 3 and 5. Cranial CT was performed to rule out hemorrhagictransformation. Vascular risk factors, temperature, blood pressure, glycemia, and blood count were assessed onadmission. National Institutes of Health Stroke Scale (NIHSS) scores were obtained at baseline and at 6, 12, 24, and 48hours. At the same time points, transcranial Doppler (TCD) examinations were conducted to assess arterialrecanalization. END was defined as an increase in the NIHSS score �4. A logistic regression model was applied todetect independent predictors of END. The Kruskal-Wallis test was used to evaluate the relationship between infarctgrowth and duration of vessel occlusion.

Results—Initial MR angiography showed an occlusion of intracranial ICA in 7 patients (23.3%), of proximal MCA in 14(46.6%), and of distal MCA in the remaining 9 (30%). A PWI-DWI mismatch �20% was observed in 28 patients(93.3%). END occurred in 7 patients (23.3%). Baseline NIHSS score (P�0.05), proximal site of occlusion (P�0.002),initial DWI (P�0.002) and PWI (P�0.003) volumes, and reduced PWI-DWI mismatch (P�0.038) were associated withEND in the univariate analysis. Only hyperacute DWI volume remained as a predictor of END when a logisticregression model was applied (odds ratio, 11.5; 95% CI, 2.31 to 57.10; P�0.0028). A receiver operator characteristiccurve identified a cutoff point of DWI �89 cm3 (sensitivity, 85.7%; specificity, 95.7%) to predict END. A gradedresponse was seen in DWI lesion expansion in relation to duration of arterial occlusion (P�0.017).

Conclusions—Ultra-early DWI is a powerful predictor of END after MCA or intracranial ICA occlusion. (Stroke. 2002;33:2197-2205.)

Key Words: brain edema � magnetic resonance imaging � stroke, acute � stroke outcome� ultrasonography, Doppler, transcranial

Early neurological deterioration (END) occurs during theacute phase in approximately one third of all ischemic

stroke patients and is associated with a significantly greatermorbidity and mortality.1–4 Previous studies demonstratedthat the extension of early focal hypodensity and mass effecton initial CT scan represent the strongest predictors of ENDamong clinical, neuroradiological, and biochemical vari-ables.1,2 Thus, END has been suggested to be determinedmainly by cerebral edema provoked by the ischemic injurythat follows an arterial occlusion.2 Moreover, clinical wors-ening during the first 48 hours after onset of cerebral ischemiamay be influenced by various and heterogeneous factors,

See Editorial Comment, page 2204

including the underlying vascular lesion, the capacity ofcollateral circulation, and biochemical processes such asexcitotoxic neurotransmission and neuroinflammation.5–8 Inthis setting, experimental and human stroke studies provideevidence that early rather than later application of neuropro-tective and aggressive therapies that may modify the afore-mentioned mechanisms underlying END could lead to abetter outcome.9–11

MRI has proven to be extremely sensitive to cerebralischemia. Several studies have shown good correlations

Received March 13, 2002; final revision received May 6, 2002; accepted May 13, 2002.From the Cerebrovascular Unit (J.F.A., C.A.M., J.M., J.A-S.) and Magnetic Resonance Unit (A.R., E.G.), Vall d’Hebron Hospital, Barcelona, Spain.Correspondence to Juan F. Arenillas, MD, Cerebrovascular Unit, Department of Neurology, Hospital Vall d’Hebron, Barcelona, Spain. E-mail

juanfarenillas@terra.es© 2002 American Heart Association, Inc.

Stroke is available at http://www.strokeaha.org DOI: 10.1161/01.STR.0000027861.75884.DF

2197 by guest on August 18, 2015http://stroke.ahajournals.org/Downloaded from

between MRI findings, stroke severity, and long-term clinicaloutcome in acute12–14 and hyperacute15–17 ischemic strokepatients. Furthermore, diffusion-weighted imaging (DWI) isable to detect restricted interstitial water diffusion related tocellular energetic failure, membrane dysfunction, and subse-quent cytotoxic edema within ischemic parenchyma,18 butearlier within the hyperacute phase of ischemic stroke19 andwith a greater sensitivity20 than CT scan. Moreover, DWIlesion volume obtained within 14 hours after onset of middlecerebral artery (MCA) occlusion was recently found to be anindependent predictor of malignant MCA infarction.21 How-ever, evidence of the usefulness of ultra-early (�6 hours)DWI and perfusion-weighted imaging (PWI) findings topredict early clinical evolution in MCA stroke is lacking. Inaddition, the impact of the time course of arterial recanaliza-tion on early clinical and radiological outcomes has not beensystematically assessed in previous MRI studies with MCAstroke patients.

The purpose of this study was to investigate the value ofDWI and PWI performed within the first 6 hours after onsetof documented MCA occlusion to predict END, which wouldhelp to define the role of MRI in the selection of patients whowould obtain benefit from early neuroprotective or aggressivetherapeutic strategies. We also sought to determine therelationship between time to artery recanalization as detectedby transcranial Doppler ultrasonography (TCD), DWI lesionenlargement, and the occurrence of END.

Subjects and MethodsInclusion criteria were as follows: patients presenting with symptomsattributable to MCA territory ischemia; performance of DWI, PWI,and MR angiography (MRA) within the first 6 hours from symptomonset; DWI signal hyperintensity within MCA territory; MRAdemonstration of occlusion affecting intracranial internal carotidartery (ICA) or MCA; and informed consent obtained from thepatients or their relatives. Exclusion criteria comprised the follow-ing: initial DWI defect outside MCA territory; absence of arterialocclusion on admission MRA; previous disability related to stroke orknown neurological illness; END associated with parenchymalhematoma on control cranial CT scan; END as result of newischemic events caused by occlusion of initially nonaffected vesselsdemonstrable by means of control MRA or TCD; and END afterreocclusion of the previously recanalized and initially occludedvessel demonstrated by TCD.

This study was approved by the local ethics committee.

Patient SelectionBetween August 1999 and November 2001, of a total of 610hyperacute stroke patients attended at our Cerebrovascular Unit, 38underwent DWI, PWI, and MRA within the first 6 hours fromsymptom onset. Eight patients had to be excluded. In 2 of them, DWIdisplayed posterior cerebral artery territory infarctions. MRAshowed no arterial occlusion in 2 patients who had anterior choroidalartery infarctions on DWI. Proximal MCA stenosis, but not occlu-sion, was observed on initial MRA in 1 patient, who was alsoexcluded. One patient suffered a severely disabling myopathy thatmade it very difficult to interpret National Institutes of Health StrokeScale (NIHSS) scores. Two patients who presented END showedparenchymal hematomas on control CT scan. Finally, 30 patientsfulfilled all criteria to enter this study.

Clinical AssessmentBaseline examinations included a medical history, physical exami-nation, routine blood biochemistry and blood count, ECG, and chest

x-ray. Stroke onset was defined as the last time the patient wasknown to be without any neurological deficit. Neurological exami-nations were performed on admission and at 6, 12, 24, and 48 hoursafter stroke onset. Stroke severity was assessed with the NIHSS. Allclinical scores were obtained by a video-trained stroke neurologist ora senior neurology resident, certified to apply the NIHSS.22 ENDwas defined as an increase in the NIHSS score by �4 points duringthe first 48 hours after symptom onset.23

Age, sex, cigarette smoking (defined as present if the patient hadsmoked at least an average of 10 cigarettes per day during the past5 years), and medical history of hypertension, diabetes, hypercho-lesterolemia, diagnosed coronary heart disease, and intermittentclaudication were recorded. Systolic and diastolic blood pressurevalues, body temperature, and glucose level were determined onadmission. Leukocyte count and fibrinogen levels were obtainedfrom baseline blood samples and recorded for further analysis.

Clinical outcome at day 90 was evaluated by means of themodified Rankin Scale (mRS). A score �2 was considered indica-tive of poor outcome.

MRI ProtocolMRI was performed with a 1.5-T whole body imager system with24-mT/m gradient strength, 300-ms rise time, and an echo-planar–capable receiver equipped with a gradient overdrive (MagnetomVision Plus, Siemens Medical Systems). The images obtainedincluded the following: (1) axial diffusion-weighted echo-planarspin-echo sequence (4000/100/2 [repetition time {TR}/echo time{TE}/acquisitions]); (2) axial perfusion-weighted echo-planargradient-echo sequence (2000/60/40 [TR/TE/acquisitions]); and (3)MRA (30/5.4/15 [TR/TE/flip angle]).

Diffusion-weighted (DW) images were obtained with a single-shotspin-echo echo-planar pulse sequence with diffusion gradient bvalues of 0, 500, and 1000 s/mm2 along all 3 orthogonal axes over 15axial sections, 5-mm-thick sections, interslice gap of 1.5 mm,240-mm field of view, and 96�128 matrix. The acquisition time forthe DW images was 56 seconds. To minimize the effects of diffusionanisotropy, the DW data were automatically processed to yieldstandard isotropic DW images.

Perfusion-weighted (PW) images were acquired by using thedynamic first pass of a 0.1-mmol/kg bolus of gadolinium-basedcontrast material (Magnevist, Schering AG) for selected 13- to15-section positions measured 40 times sequentially (acquisitiontime, 2 seconds for each measure). The bolus of 15 mL of contrastmaterial was injected in the antecubital vein by using an MR-compatible power injector (Spectris, Medrad Inc) and an injectionspeed of 5 mL/s for 3 seconds, starting 5 seconds after initiating thesequence, followed by a flush with 15 mL saline. The PW sequencegenerated a time-to-peak (TTP) map for each section position thatwas immediately available for interpretation at the console with allthe other images. Perfusion images were obtained with the use of5-mm-thick sections, interslice gap of 1.5 mm, 240-mm field ofview, and 128�128 matrix.

Tissue abnormality was considered in areas of high signal inten-sity on both DW images (reflecting decreased water motion) andTTP maps (reflecting delayed bolus arrival).

Volume measurements of the extension of tissue abnormality onDW images and on TTP maps were performed by a manual tracingtechnique by a neuroradiologist (A.R.), who was blind to clinical andTCD data. The perimeter of the area of abnormal high signalintensity was traced on each DW image and TTP map. All measuredareas were multiplied by the slice distance to obtain the total lesionvolumes for both DW images and TTP maps.

For MRA, we used a 3-dimensional time-of-flight sequence, with1.5-mm-thick sections, 200-mm field of view, and 200�512 matrix,with a total acquisition time of 156 seconds.

In most patients, a follow-up MRI examination was performedbetween days 3 and 5 after stroke onset. This examination includedDW images, MRA, and an additional transverse T2-weighted fastspin-echo (3000/85/2 [TR/TE/excitations]) or fast fluid-attenuatedinversion recovery (FLAIR) (9000/110/2200/2 [TR/TE/inversiontime/excitations]) sequence.

2198 Stroke September 2002

by guest on August 18, 2015http://stroke.ahajournals.org/Downloaded from

Computed TomographyFifteen patients (50%) underwent cranial CT scan on admission. Theremaining 15 were included in the study before local approval ofthrombolytic therapy. Cranial CT scan was performed systematicallyat the time END occurred to rule out hemorrhagic transformation.Presence and type of hemorrhagic transformation were definedaccording to previously published criteria.24,25 We excluded patientsin whom a parenchymal hematoma was considered the cause ofEND.

TCD ProtocolSerial TCD examinations were conducted on admission and at 6, 12,24, and 48 hours from stroke onset. Baseline and follow-up studieswere performed by the same stroke neurologist. TCD recordingswere performed with the use of a Multi-Dop X/TCD (DWLElektronische Systeme GmbH) device, with a hand-held transducerin a range-gated, pulsed-wave mode at a frequency of 2 MHz. Astandard method of insonation without compression testing, aspreviously described, was used.26,27 MCAs, anterior cerebral arteries,and posterior cerebral arteries were insonated through the temporalwindow at a depth between 45 and 65 mm. Mandatory initial MRAocclusion corresponded to TCD findings in all cases. Proximal MCAocclusion was defined as the absence of flow or the presence ofminimal flow signal throughout the MCA, accompanied by flowdiversion in the ipsilateral anterior cerebral artery and posteriorcerebral artery. Distal MCA occlusion was defined as a diffusedampening of the mean blood flow velocity in the affected MCA�21% compared with the contralateral MCA.28 Recanalization onfollow-up TCD recordings was diagnosed when a dampened ornormal waveform appeared in a previously demonstrated proximalMCA occlusion or when a previously dampened waveform camewithin the normal range in a previously occluded distal MCA.28 Nochange in the abnormal waveforms indicated that no recanalizationhad occurred.

Therapeutic ConsiderationsFive patients were included in clinical trials of potentially neuropro-tective drugs shortly after MRI had been performed. After localapproval of thrombolytic therapy (July 2000), 7 patients of thefinally 30 included received intravenous recombinant tissue plasmin-ogen activator (rtPA). Thrombolysis was performed according toEuropean Cooperative Acute Stroke Study (ECASS) II criteriawithin the 6-hour time window.29 The indication for rtPA was basedsolely on the clinical status and CT findings, so that MRI in no casesupposed a delay in the beginning of the treatment. Anticoagulanttherapy was started in the absence of hemorrhagic transformation onthe control CT scan performed at 48 hours after stroke onset, whenindicated. Until then, all patients received subcutaneous low-molecular-weight heparin as prophylaxis for deep venousthrombosis.

Blood pressure, temperature, and glucose levels were managedfollowing the European Stroke Initiative recommendations.30

Statistical AnalysisStatistical analyses were made by use of the SPSS statisticalpackage, version 9.0. We used �2 tests to compare rates or propor-tions of discrete variables and Mann-Whitney U tests to assessstatistical differences between continuous variables. The Spearmantest was used to study the correlation between baseline NIHSS scoreand initial DWI and PWI lesion volumes. Baseline variables werecompared to detect potential predictors of END, and those showinga P�0.05 were included in a logistic regression model. To calculatethe sensitivity and specificity of baseline DWI or PWI volumes topredict END, a receiver operator characteristic (ROC) curve wasconfigured, and cutoff values with the highest sensitivity andspecificity were included in the final logistic regression analysis.Results were expressed as adjusted odds ratios and corresponding95% CIs. Follow-up variables were compared to detect factorsassociated with END. Kruskal-Wallis tests were performed toanalyze the distribution of initial DWI and PWI lesion volumes

according to site of arterial occlusion and to study the relationshipbetween duration of vessel occlusion and DWI defect volumeexpansion from baseline to control examination at days 3 to 5.

A value of P�0.05 was considered statistically significant.

ResultsBaseline VariablesSeventeen women and 13 men were finally included. Demo-graphic characteristics of the study population and vascularrisk factors are shown in Table 1. Mean age was 70.6�12.4years (range, 41 to 90 years). Forty percent of patients werehypertensive.

Mean time from stroke onset to baseline MRI was223.2�95.6 minutes (range, 95 to 359 minutes). Initial MRAshowed an intracranial ICA occlusion in 7 patients (23.3%),a proximal MCA occlusion in 14 (46.6%), and an occludeddistal MCA in the remaining 9 (30%). Median NIHSS scoreon admission was 16.5 (range, 5 to 22). Median baseline PWIand DWI lesion volumes were 226 cm3 (range, 3 to 366 cm3)and 15 cm3 (range, 1 to 366 cm3), respectively. Twenty-eightpatients (93.3%) had a PWI-DWI mismatch �20% on base-line MRI. In the remaining 2 cases (1 patient with anintracranial ICA occlusion and another 1 with a distal MCAocclusion), no mismatch was observed.

A significant positive correlation was observed betweeninitial NIHSS score and DWI volume (Spearman’s r�0.674,P�0.001) but not between baseline NIHSS score and PWIlesion volume (r�0.257, P�0.196).

Figure 1 shows the significant differences in initial DWIand PWI volumes according to location of arterial occlusion.Patients with more proximal arterial occlusions had signifi-cantly greater PWI lesion volumes (P�0.003, Kruskal-Wallistest), DWI volumes (P�0.49), and a more reduced PWI-DWImismatch (P�0.022) on initial MRI.

Potential Predictors of ENDEND occurred in 7 patients (23.3%). In 5 of them (71.4%),NIHSS score increase was mainly due to a decrease of thelevel of consciousness initiated between 36 and 48 hours afterstroke onset. All of these patients died within the next 5 daysbecause of massive cerebral edema, raised intracranial pres-sure, and transtentorial herniation. The remaining 2 patientshad a worsening in limb strength, survived, and scored 5 onthe mRS at day 90. Presence of END was significantlyassociated with a poor long-term outcome (P�0.002). Table2 summarizes MRI, clinical, and TCD data from these 7patients. None of them had received rtPA therapy. Of theremaining 23 patients, only 1 died of a nonneurologicalcause, at day 21 after stroke onset.

TABLE 1. Characteristics of the Study Population

Age, y (SD) 70.6 (�12.42)

Sex (F), No. (%) 17 (56.7)

Medical history, No. (%)

Hypertension 12 (40)

Diabetes 5 (16.7)

Hypercholesterolemia 7 (23.3)

Coronary disease 4 (13.3)

Arenillas et al Ultra-Early DWI and Deterioration in MCA Stroke 2199

by guest on August 18, 2015http://stroke.ahajournals.org/Downloaded from

Table 3 shows baseline variables associated with theposterior occurrence of END. Baseline NIHSS score(P�0.049), proximal occlusion (P�0.002), initial DWI(P�0.002) and PWI (P�0.003) volumes, and reduced PWI-DWI mismatch (P�0.038) were significantly associated withEND in the univariate analysis. A trend toward significancewas observed for higher glucose levels (P�0.074). Further-more, there were no significant differences in time to MRIbetween the patient groups. Only hyperacute DWI lesionvolume remained as an independent predictor of END whena logistic regression model was applied (odds ratio, 11.5;95% CI, 2.31 to 57.1; P�0.0028). A ROC curve provided acutoff point of DWI �89 cm3 (sensitivity, 85.7%; specificity,95.7%) that better predicted END.

Follow-Up Variables Associated With ENDControl MRI and TCD follow-up data were compared todetect variables associated with END. A second MRI couldnot be performed in 4 of the 7 patients that suffered ENDbecause of their deteriorated clinical state. Median DWI

volume at day 3 to 5 was 27.5 cm3 (range, 1 to 451 cm3). Anenlargement of initial DWI lesion volume was observed in 22of 26 patients (84.6%). Initial and final DWI volumescoincided in 2 patients (7.7%). A partial reversal of initialDWI defect was noted in 2 patients (7.7%) in whom veryearly recanalization occurred. Greater DWI lesion volumesobserved on control MRI (P�0.018) and more pronouncedDWI lesion enlargement between baseline and control exam-inations (P�0.01) were the radiological variables associatedwith END. Control cranial CT scan performed after END hadbeen detected, in those patients who could not undergo DWI,showed large carotid territory hypodensity and signs ofspace-occupying brain edema such as midline shift. Figure 2shows baseline and control MRI of patient 6 of Table 2.Moreover, TCD-determined duration of arterial occlusionwas significantly associated with END (P�0.001). Of the 7END patients, 5 showed no recanalization of intracranialICA, and in the remaining 2, proximal MCA recanalizedbetween 12 and 24 hours. No patient with arterial occlusionduration of �12 hours suffered END in our series. Further-more, in none of the patients with baseline DWI defectvolumes �89 cm3, the cutoff value that better predicted END,was arterial recanalization detected within the first 12 hours.

In addition, a graded response was observed in DWIvolume expansion in relation to duration of arterial occlusion,as shown in Figure 3 (P�0.017, Kruskal-Wallis test).

DiscussionThe present study demonstrates that DWI lesion volumeobtained within the first 6 hours after intracranial ICA orMCA occlusion is a powerful predictor of further neurolog-ical deterioration and provides scientific evidence that ENDis mainly determined by the severity and extent of earlyischemic injury.

Previous studies have identified extension of early CT scanhypodensity as an independent predictor of END in acutestroke patients.1,2 Furthermore, CT scan may detect cytotoxicpostischemic cerebral edema that causes a decrease in x-rayattenuation, which would be visible earlier after more severeischemic damage.31 Moreover, large hyperacute CT scanhypodensity32–34 and diffusely attenuated corticomedullarycontrast35 have been shown to be highly specific, but onlymoderately sensitive, to predict fatal ischemic brain edema

Figure 1. Differences in baseline DWI and PWI lesion volumesaccording to location of arterial occlusion. Patients with moreproximal occlusions had significantly greater PWI (P�0.003 byKruskal-Wallis test) and DWI (P�0.049) volumes at baseline.Asterisks indicate outlier patients. Int indicates intracranial; prox,proximal; and dist, distal.

TABLE 2. Summary of Data From Patients Who Experienced END

PatientNo.

Age,y/Sex Occluded Artery

Time toMRI, min

BaselineNIHSS

PWI,cm3

BaselineDWI, cm3

PW-DWMismatch, %

Recanalization,h*

Control DWI,cm3

3-MonthOutcome

1 63/M L proximal MCA 359 21 305 126 58.7 12–24 250 mRS�5

2 75/M L proximal MCA 244 13 182 4 97.8 12–24 50 mRS�5

3 56/F R intracranial ICA 135 19 325 211 35.1 No � � � Death

4 67/M R intracranial ICA 289 18 320 108 66 No � � � Death

5 71/M R intracranial ICA 145 18 366 366 0 No � � � Death

6 80/F R intracranial ICA 120 22 349 244 24 No 451 Death

7 57/M R intracranial ICA 230 22 350 228 30.7 No � � � Death

L indicates left; R, right. PWI and DWI values are lesion volumes. All reported deaths occurred within the first week after stroke onset. None ofthese patients received rtPA. Patients 1 and 7 were included in clinical trials of neuroprotective agents.

*Time interval in which recanalization was detected.

2200 Stroke September 2002

by guest on August 18, 2015http://stroke.ahajournals.org/Downloaded from

and ominous stroke outcome. In this context, comparativestudies have shown that PWI and DWI are much moresensitive and reliable than CT in the detection of earlyischemic underlying pathology.20,36 In our study, ultra-earlyDWI lesion volume emerged as the most robust predictor ofEND among a series of clinical and neuroradiological vari-ables. These findings are in agreement with recent studiesshowing that DWI lesion volume within the first few hoursafter MCA or intracranial ICA occlusion strongly predicts

clinical deterioration due to massive postischemic edema.21,37

All of our patients were imaged within the first 6 hours afterstroke onset, the most attractive time window in which earlydecisive therapeutic decisions may have to be taken. Incontrast to previous MRI studies, presence and location ofarterial occlusion and time course of recanalization weresystematically assessed.

In our series, END appeared basically in the context of severepostischemic cerebral edema, which may lead to neurologicalworsening through mass effect and raised intracranial pressure,with conversion of areas of oligemia to critically hypoperfusedtissue and compromise of regional blood flow in previouslyclinically silent neighbor territories.38 Although the impossibilityof performing a follow-up MRI in 4 patients limited our analysisof other factors related to END, DWI lesion expansion wassignificantly associated with clinical deterioration. However,despite the marked increase of irreversible lesion volume ob-served during the first days in most patients (84.6%), only 23.3%of patients experienced END, which is in agreement withprevious studies.16 Thus, recruitment of ischemic penumbrawithin the infarct core may not always correspond to clinicalprogression in stroke patients. In addition, our analysis of thevariables associated with END is limited by the fact that 5patients participated in clinical trials of neuroprotective agentsand 7 were treated with rtPA, because of the potential impact ontissue viability of these substances.

TABLE 3. Univariate Analysis of Potential Predictors of END

END

PYes (n�7) No (n�23)

Age, mean�SD, y 67 �9 71.7�13.2 0.169

Sex (F), No. (%) 2 (29) 15 (65) 0.190

Hypertensive, No. (%) 4 (57) 8 (35) 0.392

Diabetes, No. (%) 2 (29) 3 (13) 0.565

Hypercholesterolemia, No. (%) 2 (29) 5 (22) 0.708

Coronary disease, No. (%) 2 (29) 2 (9) 0.225

Time to MRI, mean�SD, min 217.42�89.03 225�100.53 0.856

Left hemisphere, No. (%) 2 (29) 13 (57) 0.390

rtPA, No. (%) 0 (0) 7 (30) 0.518

Neuroprotective trials, No. (%) 2 (29) 3 (13) 0.6

Baseline NIHSS score, mean�SD 19�3.1 14.9�4.9 0.049*

Temperature, mean�SD, °C 36.7�0.56 36.5�0.4 0.618

Systolic blood pressure, mean�SD, mm Hg 159.2�39.8 150.0�27.7 0.645

Diastolic blood pressure,mean�SD, mm Hg

91.1�21.5 87.7�24.7 0.604

Glucose levels, mean�SD, mg/dL 223.5�89.9 168.5�95.4 0.074

Leukocytes, mean�SD, mm3 10257�4077 8219�2234 0.243

Fibrinogen, mean�SD, �mol/L 4.82�1.8 3.77�1.13 0.135

Intracranial ICA occlusion, No. (%) 5 (71) 2 (9) 0.002*

DWI volume, mean�SD, cm3 183.8�116.1 21.1�25.5 0.002*

PWI volume, mean�SD, cm3 312.8�67.9 168.1�89.0 0.003*

Percent PW-DW mismatch, �SD 42.7�32.8 79.05�25.8 0.038*

Results of univariate analysis of baseline variables potentially associated with the posterioroccurrence of END are shown.

*Variables included in the final logistic regression model (see Results).

Figure 2. a, Initial DWI obtained 2 hours after onset of left hemi-plegia in an 80-year-old woman (patient 6 in Table 2). b,Follow-up T2-weighted image performed at day 3, showingenlargement of the infarcted area and severe mass effect. Thepatient died at day 5.

Arenillas et al Ultra-Early DWI and Deterioration in MCA Stroke 2201

by guest on August 18, 2015http://stroke.ahajournals.org/Downloaded from

Hyperacute MRI findings may provide information aboutthe dynamic evolution of ischemic brain tissue. Ultra-earlyDWI volume better discriminated the patients that laterdeteriorated among the patients with greater baseline PWIlesions. Moreover, larger hyperacute DWI lesions may reflecta more severe early ischemic insult with a more rapidtransition from ischemia to infarction, which may be condi-tioned by a proximal arterial occlusion and a poor collateralcirculation.21,32 In this setting, early DWI volume might beconsidered a surrogate marker of the intensity of the molec-ular response triggered by cerebral ischemia, which includesall the biochemical mechanisms that have been previouslyrelated to END. These comprise the release of neuroexcita-tory amino acids, which contribute to cytotoxic edemaformation,5,6 and neuroinflammation, which leads to furthercellular destruction.3,7,8 Our study supports previous observa-tions showing that both extent and severity of the initialischemic insult determine the risk of developing massivebrain edema that leads to early neurological worsening.21,32,39

The altered hemodynamic conditions, set by the initialarterial occlusion, and the hypoperfused tissue viabilitythresholds may vary over time depending on the evolution ofthe affected vessel and on the state of collateral circulation.40

Therefore, early MRI findings may reflect only the worstpossible clinical and morphological course if the arterialocclusion persists.17 In this context, duration of occlusion wasassociated with the occurrence of END in our series. Previousstudies have demonstrated that even delayed recanalization isassociated with smaller infarct size and better clinical out-come.17,41 In accord with these findings, a graded responsewas observed in DWI volume enlargement in relation toduration of arterial occlusion. Interestingly, partial reversal ofinitial DWI volume was seen in 2 patients with ultra-earlyrecanalization, as reported previously.42 Nevertheless, wefailed to demonstrate whether earlier recanalization mayabort the fatal evolution predicted by larger hyperacute DWIlesions. In agreement with previous studies,17 we observedthat patients with more proximal occlusions had larger PWI

and DWI lesions at baseline and longer duration of arterialocclusions.

In the present study the extension and severity of the initialischemic injury were estimated by DWI lesion volumequantitative measurement, which can be obtained rapidlywith current processing techniques. Moreover, apparent dif-fusion coefficient maps better discriminate irreversibly dam-aged tissue within the heterogeneous DWI hypersignal,43 andtheir use might have improved the predictive power of ourmodel. In addition, our system uses TTP maps to obtain PWI.The impact of other techniques of dynamic susceptibilitycontrast imaging, such as the more sensitive mean to transittime maps determined by deconvolution,44 on the predictivevalue of PWI deserves further study.

In conclusion, patients showing hyperacute DWI largerlesions after intracranial ICA or MCA occlusion are at ahigher risk of END, a process with a severe outcome. Wesuggest that ultra-early MRI evaluation may identify, withvery high sensitivity and specificity, the subset of ischemicstroke patients who would obtain benefit from early admin-istration of neuroprotective therapies, including aggressivestrategies such as hypothermia and decompressivecraniectomy.

References1. Dávalos A, Toni D, Iweins F, Lesaffre E, Bastianello S, Castillo J, for the

ECASS Group. Neurological deterioration in acute ischemic stroke:potential predictors and associated factors in the European CooperativeAcute Stroke Study (ECASS) I. Stroke. 1999;30:2631–2636.

2. Toni D, Fiorelli M, Gentile M, Bastianello S, Sacchetti ML, Argentino C,Pozzili C, Fieschi C. Progressing neurological deficit secondary to acuteischemic stroke: a study of predictability, pathogenesis and prognosis.Arch Neurol. 1995;52:670–675.

3. Castillo J, Leira R. Predictors of deteriorating cerebral infarct: role ofinflammatory mechanisms: would its early treatment be useful? Cere-brovasc Dis. 2001;11(suppl 1):40–48.

4. Tei H, Uchiyama S, Ohara K, Kobayashi M, Uchiyama Y, Fukuzawa M.Deteriorating ischemic stroke in 4 clinical categories classified by theOxfordshire Community Stroke Project. Stroke. 2000;31:2049–2054.

5. Castillo J, Dávalos A, Noya M. Progression of ischemic stroke andexcitotoxic aminoacids. Lancet. 1997;349:79–83.

Figure 3. Relationship between diffusionlesion expansion and time of recanaliza-tion. Graded response was observed inDWI enlargement in relation to durationof arterial occlusion (P�0.017 byKruskal-Wallis test). Circles indicate out-lier patients.

2202 Stroke September 2002

by guest on August 18, 2015http://stroke.ahajournals.org/Downloaded from

6. Dávalos A, Castillo J, Serena J, Noya M. Duration of glutamate releaseafter acute ischemic stroke. Stroke. 1997;28:708–710.

7. Vila N, Castillo J, Dávalos A, Chamorro A. Proinflammatory cytokinesand early neurological worsening in ischemic stroke. Stroke. 2000;31:2325–2329.

8. Montaner J, Alvarez-Sabín J, Molina C, Anglés A, Abilleira S, ArenillasJ, González MA, Monasterio J. Matrix metalloproteinase expression afterhuman cardioembolic stroke: temporal profile and relation to neurologicalimpairment. Stroke. 2001;32:1759–1766.

9. Doerfler A, Forsting M, Reith W, Staff C, Heiland S, Schabitz WR, vonKummer R, Hacke W, Sartor K. Decompressive craniectomy in a ratmodel of “malignant” cerebral hemispheric stroke: experimental supportfor an aggressive therapeutic approach. J Neurosurg. 1996;85:853–859.

10. Schwab S, Steiner T, Aschoff A, Schwarz S, Steiner HH, Jansen O,Hacke W. Early hemicraniectomy in patients with complete middlecerebral artery infarction. Stroke. 1998;29:1888–1893.

11. Schwab S, Schwartz S, Spranger M, Keller E, Bertram M, Hacke W.Moderate hypothermia in the treatment of patients with severe middlecerebral artery infarction. Stroke. 1998;29:2461–2466.

12. Lövblad K-O, Baird A, Schlaug G, Benfield A, Siewert B, Voetsch B,Connor A, Burzynski C, Edelman R, Warach S. Ischemic lesion volumesin acute stroke by diffusion-weighted magnetic resonance imaging cor-relate with clinical outcome. Ann Neurol. 1997;42:164–170.

13. Barber PA, Darby DG, Desmond PM, Yang Q, Gerraty RP, Jolley D,Donnan GA, Tress BM, Davis SM. Prediction of stroke outcome withechoplanar perfusion and diffusion-weighted MRI. Neurology. 1998;51:418–426.

14. Baird A, Lövblad K-O, Dashe J, Connor A, Burzynski C, Schlaug G,Straroselskaya I, Edelman R, Warach S. Clinical correlations of diffusionand perfusion lesion volumes in acute ischemic stroke. Cerebrovasc Dis.2000;10:441–448.

15. Tong DC, Yenari MA, Albers GW, O’Brien M, Marks MP, Moseley ME.Correlation of perfusion- and diffusion weighted MRI with NIHSS scorein acute (�6.5 hour) ischemic stroke. Neurology. 1998;50:864–870.

16. Beaulieau C, de Crespigny A, Tong DC, Moseley ME, Albers GW, MarksMP. Longitudinal magnetic resonance imaging study of perfusion anddiffusion in stroke: evolution of lesion volume and correlation withclinical outcome. Ann Neurol. 1999;46:568–578.

17. Schellinger PD, Fiebach JB, Jansen O, Ringleb PA, Mohr A, Steiner T,Heiland S, Schwab S, Pohlers O, Ryssel H, Orakcioglu B, Sartor K,Hacke W. Stroke magnetic resonance imaging within 6 hours after onsetof hyperacute cerebral ischemia. Ann Neurol. 2001;49:460–469.

18. van Everdingen KJ, van der Grond J, Kappelle LJ, Ramos LMP, MaliWPTM. Diffusion-weighted magnetic resonance imaging in acute stroke.Stroke. 1998;29:1783–1790.

19. Yoneda Y, Tokui K, Hanihara T, Kitagaki H, Tabuchi M, Mori E.Diffusion-weighted magnetic resonance imaging: detection of ischemicinjury 39 minutes after onset in a stroke patient. Ann Neurol. 1999;45:794–797.

20. Fiebach J, Jansen O, Schellinger P, Knauth M, Hartmann M, Heiland S,Ryssel H, Pohlers O, Hacke W, Sartor K. Comparison of CT withdiffusion-weighted MRI in patients with hyperacute stroke. Neuroradi-ology. 2001;43:628–632.

21. Oppenheim C, Samson Y, Manaï R, Lalam T, Vandamme X, Crozier S,Srour A, Cornu P, Dormont D, Rancurel G, Marsault C. Prediction ofmalignant middle cerebral artery infarction by diffusion-weightedimaging. Stroke. 2000;31:2175–2181.

22. Lyden P, Brott T, Tilley B, Welch KMA, Mascha EJ, Levine S, HaleyEC, Grotta J, Marler J, and the NINDS t-PA Stroke Study Group.Improved reliability of the NIH Stroke Scale using video training. Stroke.1994;25:2220–2226.

23. The National Institutes of Neurological Disorders and Stroke rt-PAStroke Study Group. Tissue plasminogen activator for acute ischemicstroke. N Engl J Med. 1995;333:1581–1587.

24. Pessin M, del Zoppo G, Estol C. Thrombolytic agents in the treatment ofstroke. Clin Neuropharmacol. 1990;13:271–289.

25. Hacke W, Kaste M, Fieschi C, Toni D, Lesaffre E, von Kummer R,Boysen G, Bluhmki E, Hoxter G, Mahagne M, Hennerici M. Intravenousthrombolysis with recombinant tissue plasminogen activator for acute

hemispheric stroke: the European Cooperative Acute Stroke Study(ECASS). JAMA. 1995;274:1017–1025.

26. Lindegaard K-F, Bakke SJ, Aaslid R, Nornes H. Doppler diagnosis ofintracranial occlusive disorders. J Neurol Neurosurg Psychiatry. 1986;49:510–518.

27. Aaslid R, Markwalder T-M, Nornes H. Noninvasive transcranial Dopplerultrasound recording of flow velocity in basal cerebral arteries. J Neu-rosurg. 1982;57:769–774.

28. Zanette EM, Roberti C, Mancici G, Pozzilli C, Bragoni M, Toni D.Spontaneous middle cerebral artery reperfusion in ischemic stroke: afollow-up study with transcranial Doppler. Stroke. 1995;26:430–433.

29. Hacke W, Kaste M, Fieschi C, von Kummer R, Davalos A, Meier D,Larrue V, Bluhmki E, Davis S, Donnan G, Scheneider D, Diez-Tejedor E,Trouillas P. Randomised double-blind placebo-controlled trial ofthrombolytic therapy with intravenous alteplase in acute ischemic stroke(ECASS II). Lancet. 1998;352:1245–1251.

30. European Stroke Council, European Neurological Society, and EuropeanFederation of Neurological Societies. European Stroke Initiative recom-mendations for stroke management. Cerebrovasc Dis. 2000;10:335–351.

31. Berrouschot J, Barthel H, von Kummer R, Knapp WH, Hesse S, SchneiderD. 99mTechnetium-ethyl-cysteinate-dimer single-photon emission CT canpredict fatal ischemic brain edema. Stroke. 1998;29:2556–2562.

32. von Kummer R, Meyding-Lamade U, Forsting M, Rosin L, Rieke K,Hacke W, Sartor K. Sensitivity and prognostic value of early CT inocclusion of the middle cerebral artery trunk. AJNR Am J Neuroradiol.1994;15:9–15.

33. Krieger DW, Demchuk AM, Kasner SE, Jauss M, Hantson L. Earlyclinical and radiological predictors of fatal brain swelling in ischemicstroke. Stroke. 1999;30:287–292.

34. Kasner SE, Demchuk AM, Berrouschot J, Schmutzhard E, Harms L,Verro P, Chalela JA, Abbur R, McGrade H, Christou I, Krieger DW.Predictors of fatal brain edema in massive hemispheric ischemic stroke.Stroke. 2001;32:2117–2123.

35. Haring HP, Dilitz E, Pallua A, Hessenberger G, Kampfl A, Pfausler B,Schmutzhard E. Attenuated corticomedullary contrast: an early cerebralcomputed tomography sign indicating malignant middle cerebral arteryinfarction: a case control study. Stroke. 1999;30:1076–1082.

36. Jaillard A, Hommel M, Baird AE, Linfante I, Llinas RH, Caplan LR,Edelmann RR, Warach S. Significance of early CT signs in acute stroke:a CT scan-diffusion MRI study. Cerebrovasc Dis. 2002;13:47–56.

37. Thijs VN, Lansberg MG, Beaulieu C, Marks MP, Moseley ME, AlbersGW. Is early ischemic lesion volume on diffusion-weighted imaging anindependent predictor of stroke outcome? A multivariate analysis. Stroke.2000;31:2597–2602.

38. Hacke W, Schwab S, Horn M, Spranger M, De Georgia M, von KummerR. Malignant middle cerebral artery territory infarction: clinical courseand prognostic signs. Arch Neurol. 1996;53:309–315.

39. Firlik AD, Yonas H, Kauffmann AM, Weschler LR, Jungreis CA, FukuiMB, Williams RL. Relationship between cerebral blood flow and thedevelopment of swelling and life-threatening herniation in acute ischemicstroke. J Neurosurg. 1998;89:243–249.

40. Warach S. Tissue viability thresholds in acute stroke: the 4-factor model.Stroke. 2001;32:2460–2461.

41. Molina CA, Montaner J, Abilleira S, Arenillas JF, Ribó M, Huertas R,Romero F, Álvarez-Sabín J. Time course of tissue plasminogen activa-tor-induced recanalization in acute cardioembolic stroke: a case-controlstudy. Stroke. 2001;32:2821–2827.

42. Kidwell CS, Saver JL, Mattielo J, Starkman S, Vinuela F, Duckwiler G,Gobin YP, Jahan R, Vespa P, Kalafut M, Alger JR. Thrombolytic reversalof acute human cerebral ischemic injury shown by diffusion/perfusionmagnetic resonance imaging. Ann Neurol. 2000;47:462–469.

43. Oppenheim C, Grandin C, Samson Y, Smith A, Duprez T, Marsault C,Cosnard G. Is there an apparent diffusion coefficient threshold in pre-dicting tissue viability in hyperacute stroke? Stroke. 2001;32:2486–2491.

44. Yamada K, Wu O, Gonzalez RG, Bakker D, Ostergaard L, Copen WA,Weisskoff RM, Rosen BR, Yagi K, Nishimura T, Sorensen AG. Magneticresonance perfusion-weighted imaging of acute cerebral infarction: effectof the calculation methods and underlying vasculopathy. Stroke. 2002;33:87–94.

Arenillas et al Ultra-Early DWI and Deterioration in MCA Stroke 2203

by guest on August 18, 2015http://stroke.ahajournals.org/Downloaded from

Editorial Comment

Diffusion-Weighted MRI: Back to the Future

The accompanying article demonstrates the capacity ofdiffusion-weighted imaging (DWI) to be a powerful predictorof early neurological deterioration. The authors have ele-gantly performed a study of 38 patients with middle cerebralartery stroke in which they found an increase in the NationalInstitutes of Health Stroke Scale (NIHSS) score �4 inpatients to be associated with acute DWI signs that canpredict this deterioration. Imaging was performed within 6hours of onset, which represents the time window for studiesof thrombolysis. While the number of patients studied is stillsmall, this report raises an important issue, namely, thecapacity of DWI to serve not only as a marker of stroke butas a report of its severity. This article clearly answers thisquestion positively: indeed, all patients who presented earlyneurological deterioration had greater DWI lesion volumeson the acute scan. This is of extreme importance since the aimof imaging should be not only to determine the presence ofischemia but to demonstrate evolution, be it positive ornegative. Indeed, CT has been demonstrated to be a predictorof bad outcome: not only could it demonstrate absent hem-orrhage, but it could show signs heralding malignant ischemictransformation.1 This is why CT has remained the method ofchoice for evaluation of patients in studies of thrombolysis.

With the introduction of modern echo-planar fast MRtechnology, the ability of DWI to image acute changesassociated with ischemia was demonstrated. Together withfurther refinements in MR angiography, perfusion imaging,and spectroscopy, it is able to obtain in vivo data that werepreviously obtainable only in animal models of ischemia.2

However, from a clinical perspective, DWI has not been asthoroughly studied as CT. Indeed, DWI has entered ourcurrent armamentarium for imaging of acute stroke with fewvalidation studies being available. After the initial enthusiasmabout a method often thought to demonstrate acute strokeunequivocally, more questions were raised than answeredabout its implementation. Initial studies showed it to besensitive and to correlate with neurological status,2–4 butmore data were needed for DWI to gain wide acceptance. Inan era of evidence-based medicine, when solid data wereneeded to strengthen initial resistance to thrombolysis, italmost looked as though the bright future of DWI was behindit.5

However, with studies such as that reported in the accom-panying article, DWI is now showing its capacity to fulfill theearly promises made in the laboratory setting: DWI demon-strates its capacity not only as a diagnostic tool but also as aresearch tool. The authors again report that DWI correlatesstrongly with initial NIHSS score, further strengthening theclinical validity of previous studies.

DWI not only provides us with important clinical informa-tion by acutely showing lesion extension but is also able todeliver important physiopathological data through measures

of the apparent diffusion coefficient (ADC). Indeed, despitecontradictory findings concerning the exact nature of thechanges underlying the acute decrease in ADC, many recentreports tend to point to its validity. Reports have shownreversal of lesions to be possible after intervention, when bothDWI volumes and ADC lesions are measured.6–11

Most importantly, Oppenheim et al12 found a slight markeddecrease in the ADC in areas that would present infarctgrowth. In another study they found that patients with a lowerADC measured in the lesion would develop hemorrhagictransformation.13 Oppenheim et al found that lesion volumeand ADC value measurements together could predict malig-nant infarction better than DWI lesion alone. The authors ofthe accompanying article are encouraged to further investi-gate this phenomenon, which would undoubtedly strengthentheir assertions.

In addition to well-established MRI methods such asT2-weighted imaging and MR angiography, newer tech-niques such as diffusion tensor imaging are opening the wayfor new approaches to clinical ischemia management andresearch. Indeed, changes in water anisotropy might occur atan early point, heralding structural changes taking place in thepenumbra.14,15

Additionally, it is becoming clear that the changes seen onany kind of neuroimaging study should not be interpretedalone but are to be seen as part of a whole. This was recentlypointed out in an important study by Baird et al,16 who foundthat factors such as DWI/perfusion-weighted imagingchanges, NIHSS score, and time from onset of symptomsshould be taken into account; this was more recently summa-rized as a 4-factor model. This stresses not only the multi-modality of methods but the multidisciplinary managementthat is becoming increasingly evident in the treatment ofacute stroke.17 This is important to take into account when thedata presented in the accompanying article are considered.

Therefore, we conclude that, as stated by the authors, DWIseems to be usable as an early means to select patients formore aggressive therapy. Thus, DWI has again found itsplace both in the present as a clinical tool and in the future asa source of clinical experimentation.

Karl-Olof Lövblad, MD, Guest EditorNeuroradiology Unit

Department of RadiologyUniversity Hospital of Geneva

Geneva, Switzerland

References1. von Kummer R, Allen KL, Holle R, Bozzao L, Bastianello S, Manelfe C,

Bluhmki E, Ringleb P, Meier DH, Hacke W. Acute stroke: usefulness ofearly CT findings before thrombolytic therapy. Radiology. 1997;205:327–333.

2204 Stroke September 2002

by guest on August 18, 2015http://stroke.ahajournals.org/Downloaded from

2. Warach S, Gaa J, Siewert B, Wielopolski P, Edelman R. Acute humanstroke studied by whole brain echo planar diffusion-weighted magneticresonance imaging. Ann Neurol. 1995;37:231–241.

3. Lovblad KO, Laubach HJ, Baird AE, Curtin F, Schlaug G, Edelman RR,Warach S. Clinical experience with diffusion-weighted MR in patientswith acute stroke. AJNR Am J Neuroradiol. 1998;19:1061–1066.

4. Lövblad K-O, Baird A, Schlaug G, Benfield A, Siewert B, Voetsch B,Connor A, Burzynski C, Edelman R, Warach S. Ischemic lesion volumesin acute stroke by diffusion-weighted magnetic resonance imaging cor-relate with clinical outcome. Ann Neurol. 1997;42:164–170.

5. Keir SL, Wardlaw JM. Systematic review of diffusion and perfusionimaging in acute ischemic stroke. Stroke. 2000; 2000;31:2723–2731.

6. Marks MP, Tong DC, Beaulieu C, Albers GW, de Crespigny A, MoseleyME. Evaluation of early reperfusion and IV tPA therapy using diffusion-and perfusion-weighted MRI. Neurology. 1999;52:1792–1798.

7. Kidwell CS, Saver JL, Mattiello J Starkman S, Vinuela F, Duckwiler G,Gobin YP, Jahan R, Vespa P, Kalafut M, Alger JR. Thrombolytic reversalof acute human cerebral ischemic injury shown by diffusion/perfusionmagnetic resonance imaging. Ann Neurol. 2000;47:462–469.

8. Lansberg MG, Tong DC, Norbash AM, Yenari MA, Moseley ME. Intra-arterial rtPA treatment of stroke assessed by diffusion- and perfusion-weighted MRI. Stroke. 1999;30:678–680.

9. Schellinger PD, Jansen O, Fiebach JB, Heiland S, Steiner T, Schwab S,Pohlers O, Ryssel H, Sartor K, Hacke W. Monitoring intravenous recom-binant plasminogen activator thrombolysis for acute ischemic stroke withdiffusion and perfusion MRI. Stroke. 2000;31:1318–1328.

10. Jansen O, Schellinger P, Fiebach J, Hacke W, Sartor K. Early recanal-isation in acute ischaemic stroke saves tissue at risk defined by MRI.Lancet. 1999;353:2036–2037.

11. Taleb M, Lovblad KO, El-Koussy M, Guzman R, Bassetti C, Arnold M,Oswald H, Remonda L, Schroth G. Reperfusion demonstrated byapparent diffusion coefficient mapping after local intra-arterialthrombolysis for ischaemic stroke. Neuroradiology. 2001;43:591–594.

12. Oppenheim C, Grandin C, Samson Y, Smith A, Duprez T, Marsault C,Cosnard G. Is there an apparent diffusion coefficient threshold in pre-dicting tissue viability in hyperacute stroke? Stroke. 2001;32:2486–2491.

13. Oppenheim C, Samson Y, Dormont D, Crozier S, Manai R, Rancurel G,Fredy D, Marsault C. DWI prediction of symptomatic hemorrhagic trans-formation in acute MCA infarct. J Neuroradiol. 2002;29:6–13.

14. Basser PJ, Pierpaoli C. Microstructural and physiological features oftissues elucidated by quantitative-diffusion-tensor MRI. J Magn ResonImaging. 1996;B111:209–219.

15. Sorensen AG, Wu O, Copen WA, Davis TL, Gonzalez RG, KoroshetzWJ, Reese TG, Rosen BR, Wedeen VJ, Weisskoff RM. Human acutecerebral ischemia: detection of changes in water diffusion anisotropy byusing MR imaging. Radiology. 1999;212:785–792.

16. Baird AE, Dambrosia J, Janket S, Eichbaum Q, Chaves C, Silver B,Barber PA, Parsons M, Darby D, Davis S, Caplan LR, Edelman RE,Warach S. A three-item scale for the early prediction of stroke recovery.Lancet. 2001;357:2095–2099.

17. Warach S. Tissue viability thresholds in acute stroke: the 4-factor model.Stroke. 2001;32:2460–2461.

Arenillas et al Ultra-Early DWI and Deterioration in MCA Stroke 2205

by guest on August 18, 2015http://stroke.ahajournals.org/Downloaded from

Álvarez-SabínJuan F. Arenillas, Álex Rovira, Carlos A. Molina, Elisenda Grivé, Joan Montaner and José

Imaging in Hyperacute Middle Cerebral Artery Ischemic StrokePrediction of Early Neurological Deterioration Using Diffusion- and Perfusion-Weighted

Print ISSN: 0039-2499. Online ISSN: 1524-4628 Copyright © 2002 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Stroke doi: 10.1161/01.STR.0000027861.75884.DF

2002;33:2197-2205Stroke. 

http://stroke.ahajournals.org/content/33/9/2197World Wide Web at:

The online version of this article, along with updated information and services, is located on the

  http://stroke.ahajournals.org//subscriptions/

is online at: Stroke Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer process is available in the

Request Permissions in the middle column of the Web page under Services. Further information about thisOnce the online version of the published article for which permission is being requested is located, click

can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.Strokein Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:

by guest on August 18, 2015http://stroke.ahajournals.org/Downloaded from