This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Grandin, C. B. Articles by Cosnard, G. Search for Related Content PubMed PubMed Citation Articles by Grandin, C. B. Articles by Cosnard, G. Pubmed/NCBI databases Medline Plus Health Information MRI Scans Stroke Related Collections Acute Cerebral Infarction Computerized tomography and Magnetic Resonance Imaging

(Stroke. 2001;32:1147.)
© 2001 American Heart Association, Inc.

Original Contributions

Usefulness of Magnetic Resonance–Derived Quantitative Measurements of Cerebral Blood Flow and Volume in Prediction of Infarct Growth in Hyperacute Stroke

Cécile B. Grandin, MD; Thierry P. Duprez, MD; Anne M. Smith, PhD; Fréderic Mataigne, MD; André Peeters, MD; Catherine Oppenheim, MD Guy Cosnard, MD

From the Department of Medical Imaging (C.B.G., T.P.D., A.M.S., F.M., G.C.) and the Unit of Neurology (A.P.), Cliniques Universitaires St. Luc, Université Catholique de Louvain, Brussels, Belgium, and the Unit of Neuroradiology (C.O.), GH Pitié-Salpétrière, Paris, France.

Correspondence to Cécile B. Grandin, MD, Cliniques Universitaires St. Luc, Department of Medical Imaging (MRI Unit), 10 Hippocrate Ave, B-1200 Brussels, Belgium. E-mail grandin{at}rdgn.ucl.ac.be

	Abstract

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Background and Purpose—The identification of the tissue at risk for infarction remains challenging in stroke patients. In this study, we evaluated the value of quantitative cerebral blood flow (CBF) and cerebral blood volume (CBV) measurements in the prediction of infarct growth in hyperacute stroke.

Methods—Fluid-attenuated inversion recovery (FLAIR), diffusion-weighted (DW), and gradient-echo echo-planar perfusion-weighted (PW) sequences were obtained in 66 patients within 6 hours of stroke onset; ischemia was confirmed on follow-up FLAIR images. We delineated the following: (1) the initial infarct on DW images, (2) the area of hemodynamic disturbance on mean transit time (MTT) maps, and (3) the final infarct on follow-up FLAIR images. MTT, CBF, and CBV were calculated in the following areas: area of initial infarct (INF), area of infarct growth (IGR, final minus initial infarct), the hemodynamically disturbed area that remained viable (OLI, hemodynamic disturbance minus final infarct), and all contralateral mirror regions.

Results—Compared with mirror regions, the MTT in abnormal areas was always prolonged. The respective mean±SD CBF and CBV values were as follows: for INF, 28±16 mL/min per 100 g and 6.9±2.7%; for IGR, 36±20 mL/min per 100 g and 8.9±3.1%; for OLI, 50±17 mL/min per 100 g and 11.2±3%; and for mirror regions, 64±23 mL/min per 100 g and 8.7±2.5%. The CBV and CBF values were significantly different between all abnormal areas (except for the CBF between INF and IGR). In the area of DW/PW mismatch, a combined CBF or CBV threshold of 35 or 8.2, respectively, predicted evolution to infarction with a sensitivity of 81% and a specificity of 76%.

Conclusions—Quantitative measurements of CBF and CBV in hyperacute stroke may help to predict infarct growth and to select the subjects who will benefit from thrombolysis.

Key Words: cerebral blood flow • cerebral infarction • magnetic resonance imaging, diffusion-weighted • magnetic resonance imaging, perfusion-weighted

	Introduction

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Most pharmacological interventions in the acute phase of ischemic stroke are based on the concept of ischemic penumbra, a region of reversible ischemia that is still viable but will eventually evolve to infarction.^{1 2 3 4} Abnormalities observed on diffusion-weighted (DW) images allow an early identification of severely ischemic brain regions that typically evolve into infarction and are thought to represent the irreversible ischemic core.^{4 5 6 7} On the other hand, perfusion-weighted (PW) imaging provides information about the hemodynamic status of the cerebral tissue.^{6 7 8 9} Temporal parameters, such as the mean transit time (MTT) or the time to peak, are very sensitive in identifying areas of hemodynamically impaired perfusion, which are frequently larger than the DW lesions during the first hours of stroke evolution.^{4 8 9} A subsequent infarct enlargement has been described in the region of DW/PW mismatch, supporting the hypothesis that this area reflects the ischemic penumbra.^{4 6 10} However, the DW/PW mismatch region may constitute not only the area at risk for infarction but also oligemic tissue with blood flow above the critical viability threshold that is not at risk for infarction.^{7 8 9} Therefore, it is crucial to quantify the importance of the perfusion deficit within the DW/PW mismatch area to distinguish the real area at risk from the oligemic tissue that does not constitute a target for thrombolysis.⁹ Our hypothesis is that quantitative measures of cerebral blood flow (CBF) and cerebral blood volume (CBV) may be useful for this purpose.

In the present study, we used whole-brain magnetic resonance (MR) bolus tracking data for quantifying CBF, CBV, and MTT in 3 functional regions defined on the retrospective analysis of a population of "nontreated" hyperacute stroke patients. The ischemic core was defined as the initial DW lesion.^{7 9 11} The difference between the initial DW lesion and the final infarct volume demonstrated by a follow-up MR study corresponded to the region that was initially viable but had evolved to infarction and defined the area of infarct growth.¹¹ As proposed by Schlaug at al,⁷ the area of infarct growth may represent an operational definition of the penumbra and may be used to retrospectively identify the tissue at risk for infarction. The oligemic area was defined as the region that was hemodynamically disturbed but did not evolve to infarction and was represented by the regional difference between the hemodynamically disturbed area defined on MTT maps and the final infarct volume.¹¹

Our hypothesis was that absolute quantification of perfusion might help to distinguish these 3 functional regions and to predict infarct growth. The identification of the real area at risk for infarction could help target patients who would potentially benefit from thrombolytic or neuroprotective treatment and provide a tool to characterize stroke patients enrolled in therapeutic trials.

	Subjects and Methods

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Patients
Sixty-six patients were retrospectively included in the present study. They were selected from our hyperacute stroke patient database, which included all subjects presenting at our hospital with a sudden focal neurological deficit suggesting stroke and who had had an MR examination within 6 hours after symptom onset. In accordance with our institutional guidelines, no CT was performed before MR examination.

The inclusion criteria were as follows: (1) no hemorrhagic lesion, (2) acute cerebral infarct confirmed on follow-up MRI obtained 1 to 3 days after symptom onset, (3) identical MRI protocol including at least fluid-attenuated inversion recovery (FLAIR), DW, and PW sequences at the hyperacute phase and a follow-up FLAIR sequence, and (4) similar treatment based on arterial blood pressure control and administration of 300 mg/d acetylsalicylic acid. Among the 139 patients enrolled from January 1998 to March 2000, 73 were excluded because of (1) observation of a hemorrhage (n=10) or a tumor (n=1) on the initial MRI, (2) no acute ischemic lesion visualized on the follow-up MRI (n=20), (3) incomplete MRI protocol (n=18, including 4 patients who died before the follow-up MRI), different MRI protocol (n=4), archiving problems (n=3), motion artifacts (n=2), artifacts due to ferromagnetic material or technical problems (n=4), or poor bolus of the contrast agent for the perfusion study (n=2), or (4) thrombolytic therapy (n=9). Eight agitated patients included in the present study required sedation with an intravenous administration of midazolam (Dormicum).

The 66 selected patients (41 men and 25 women, mean age 69±13 years, range 26 to 91 years) presented with stroke in various arterial territories. Fifty-three lesions were located in the anterior circulation (38 middle cerebral arteries [MCAs], 7 anterior choroidal arteries or deep perforating arteries, 1 MCA+anterior choroidal artery, 4 anterior cerebral arteries, 1 MCA+anterior cerebral artery, and 2 watershed areas between MCAs and posterior cerebral arteries). Thirteen lesions were in areas perfused by the posterior circulation (3 posterior cerebral arteries, 3 cerebellar arteries, 1 posterior cerebral artery+cerebellar artery, 5 basilar arteries, and 1 basilar artery+cerebellar artery). The mean European stroke scale at admission was 63±25, with a range of 9 to 96 per 100.

MRI Protocol
At the hyperacute phase (delay between MRI and symptom onset: mean 3.2±1.3 hours, range 1 to 6 hours), the typical MRI protocol consisted of (1) sagittal gradient-echo T1-weighted images (scout view and hemorrhage detection), (2) axial fast T2-weighted and FLAIR sequences (detection of old and subacute lesions and gross examination of the main arteries), (3) 3D time-of-flight MR angiography of intracranial arteries, and (4) axial DW and bolus-tracking PW sequences (total imaging time 25 minutes). The follow-up MRI obtained 36±19 hours after the first examination consisted of a sagittal gradient-echo T1-weighted sequence followed by an axial fast FLAIR sequence. In the more recent studies, an axial DW sequence was added. All images were acquired on a 1.5-T GE Signa Echospeed scanner (GE Medical System). Only the FLAIR, DW, and PW sequences were used in the present study, and all these images were acquired in the anterior commissure–posterior commissure plane with 5-mm slice thickness, 0.5-mm slice gap, 24x24-cm² field of view, and 24 slices per volume, enabling whole-brain coverage. The acquisition parameters of the fast FLAIR sequence were as follows: repetition time 10 002 ms, echo time 148 ms, inversion time 2200 ms, and matrix 256x160. A T2-weighted reference volume (b=0 s/mm²) and DW images were acquired with a single-shot, echo-planar, spin-echo sequence (repetition time 4500 ms, echo time 95 ms, and matrix 96x64). The diffusion trace volume was calculated from 3 DW volumes with the diffusion gradients sequentially applied along each of the x, y, and z directions and with b=1000 s/mm², =32 ms, =39 ms, and G=22 mT/m. The PW images were acquired by using the dynamic first-pass bolus-tracking method and a single-shot, echo-planar, gradient-echo sequence (repetition time 2300 ms, echo time 30 ms, and matrix 96x64) starting either at the same time or a few seconds after the beginning of contrast material injection. A dose of 0.1 mmol/kg Gd-DTPA (Magnevist, Schering AG) was injected at a rate of 10 mL/s in a peripheral vein through an 18-gauge catheter by an MR-compatible power injector (Spectris, Medrad Inc), followed by a flush with 30 mL saline. The sequence duration was 46 seconds, and 20 brain volumes were acquired.

Data Processing
All images were processed on an independent workstation (Sun Ultra 1/200, Sun Microsystems) by using homemade programs written in C and interactive data language (Research System Inc).¹²

Parametric Perfusion Maps
The MR signal intensity was converted to relative gadolinium concentration to obtain concentration-versus-time curves on a voxel-by-voxel basis without smoothing. The curves were fitted to a variate function, and several parameters derived from these curves were calculated to create parametric maps. Only the apparent MTT (apMTT), defined as the first moment of the measured tissue curve, was used in the present study.

Regions of Interest
All images were spatially coregistered to the first volume of the PW sequence to superimpose the regions of interest (ROIs) delineated on each type of image.¹²

Two neuroradiologists, blinded as to other images and clinical symptoms, independently drew 3D ROIs by manual contouring of the trace DW and apMTT images. The contour of the abnormal bright area on DW images defined the initial infarct ROI, and the contour of the area showing a prolonged apMTT defined the hemodynamically disturbed ROI. For volumetric measurements, the DW images were chosen because of the better contrast between abnormal and normal brain tissue compared with apparent diffusion coefficient maps.^{4 7} To avoid potential misinterpretation due to a residual T2 shine-through effect, images obtained at b=0 were available for comparison. Similarly, apMTT maps were chosen to delineate perfusion abnormalities because, as opposed to CBV and CBF, this parameter is homogeneous throughout the whole healthy brain and is highly sensitive in detecting hemodynamic disturbance, providing an excellent contrast between normal and abnormal areas.^{7 8 12} The follow-up FLAIR images were used to define the final infarct ROI. A consensus was established between the 2 neuroradiologists, who drew an ROI delineating the abnormal bright area corresponding to the acute stroke, with the initial DW and FLAIR images being available for comparison. In the case of large infarct, the vasogenic edema artificially increased the final infarct volume, collapsing the ventricles and displacing the sulci. In these cases, the ROI was redrawn so that its contours corresponded to the same anatomic area without edema (Figure 1). The good delineation of the anatomic landmarks on FLAIR images facilitated this correction by comparing the initial and follow-up images and justified the use of the FLAIR instead of the DW sequence for determination of the final infarct. To test the validity of using FLAIR images for this purpose, ROIs were also delineated on follow-up DW images obtained in a subset of 10 patients. For all ROIs, the software generated a mirror ROI that was placed in normal contralateral regions. Finally, the 3 areas defined in our stroke model and their contralateral mirrors were obtained: (1) the ischemic core corresponding to the initial infarct contour (INF), (2) the area of infarct growth corresponding to the final minus the initial infarct contours (IGR), and (3) the oligemic area that was hemodynamically disturbed but remained viable, corresponding to the apMTT minus the final infarct contours (OLI).

View larger version (55K): [in this window] [in a new window]   Figure 1. Patient A was an 84-year-old woman scanned 3 hours after the onset of aphasia and right hemiparesis and 2 days later. The volume of hemodynamic disturbance (white ROI on apMTT map, 162 mL) was larger than the ischemic core (green ROI on initial diffusion images, 35 mL), but the infarct did not grow (according to final infarct drawn on follow-up FLAIR images, the yellow ROI corresponds to the initial contouring, and the red ROI was corrected for the edema by extrapolating the initial morphology of the sulci from the initial FLAIR images). The respective CBV (%) and CBF (mL/min per 100 g) values were 8.2 and 10 in the ischemic core, 11.4 and 63 in the DW/PW mismatch area, and 8.6 and 86 in mirror normal regions (blue ROIs). The flow in the mismatch area was far above the critical viability threshold, and this region was not at risk for infarction, discarding the indication of thrombolysis. Patient B was a 64-year-old man scanned 6 hours after the onset of left hemiplegia and 3 days later. The volume of hemodynamic disturbance (white ROI on apMTT map, 298 mL) was larger than the ischemic core (green ROI on initial diffusion images, 76 mL), and the infarct did grow. On follow-up FLAIR images, a hemorrhagic transformation occurred in the deep gray nuclei, collapsing the ventricle in combination with the edema. To correct for this effect, the yellow ROI contouring the final infarct was modified (red ROI). The respective CBV (%) and CBF (mL/min per 100 g) values were 4.4 and 9 in the ischemic core, 8.0 and 25 in the area of infarct growth (red minus green ROI), 10.3 and 40 in the fraction of DW/PW mismatch area that remained viable (white minus red ROI), and 6.7 and 64 in mirror normal regions (blue ROIs). The quantification of perfusion data allowed us to identify the area at risk that became infarcted, making this patient a candidate for thrombolysis if all the usual contraindications were not present.

Calculation of Quantitative CBF, CBV, and MTT Values
The arterial input function (AIF) was determined independently by the 2 neuroradiologists and was used for the quantification of CBF, CBV, and MTT in their own ROIs. The mean tissue curve of the voxels contained in each ROI was fitted and deconvolved by the AIF with use of a Fourier transform method.^{12 13} By application of the central volume principle, the CBV (%) was calculated as follows:

where C_m(t) is the measured tissue curve in the ROI, C_AIF(t) is the AIF curve, and is a constant taking into account the density of brain tissue and the difference in hematocrit between capillaries and large vessels. The CBF (mL/min per 100 g) was calculated as follows:

where C(t) is the deconvolved tissue curve and C_max is the maximum of this deconvolved curve. Finally, the MTT (seconds) was calculated as follows:

MTT=CBV/CBF

The measured values were reported only for the ROI >1 mL for the initial infarct and for the ROI >2 mL for the areas obtained by subtraction to minimize the errors related to imprecise coregistration and contouring.

Statistical Analysis
The volumes and the CBF, CBV, and MTT values reported in the results were averaged from the measurements obtained by the 2 observers. The continue variables characterizing a group were expressed as their mean±SD. Parametric tests were used according to the central limit theorem or because the data were normally distributed. The level of significance for 2-tailed tests was fixed at P=0.05, corrected for multiple comparisons when necessary. Unpaired Student t tests were used to compare volumes between groups. Pearson linear correlations were used (1) to evaluate the relationship between different types of volumes measured within a group and (2) to test the interobserver variability in measured volumes. The CBV, CBF, and MTT values were compared between areas by ANOVA with Student-Newman-Keuls post hoc tests. The ability of CBF or CBV values to discriminate IGR from OLI was evaluated by use of receiver operating characteristic (ROC) curve analysis.

	Results

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Characteristics of the Study Population
The volumes measured by the 2 neuroradiologists correlated well for INF (R²=0.987, P<0.001) and for the apMTT ROI (R²=0.868, P<0.001). The correlation between the final infarct volumes measured on follow-up FLAIR and DW images was excellent (R²=0.999, P<0.001). The mean final infarct volume was 34±55 mL, with 18 lesions <3 mL. Between the initial and follow-up MRI, the infarct volume remained stable in 43 of 66 patients (mean 14±23 mL). During the same period, an increase of the infarct volume was observed in 18 of 66 patients (mean volume increased from 64±62 to 85±79 mL). In 5 of 66 patients, the volume of infarcted tissue slightly decreased (mean volume decreased from 24±32 to 19±31 mL). The lesions that did not grow presented an initial infarct volume significantly smaller than the lesions that did grow (P<0.001), and a significant correlation was observed between the volume of INF and the volume of enlargement (R²=0.470, P<0.001). An area of hemodynamic disturbance (prolonged apMTT) was observed in 59 of 66 patients (mean volume 95±84 mL). This volume was larger than the volume of INF (DW/PW mismatch) in 47 patients. A DW/PW mismatch was observed in all patients who presented an extension of the ischemic core but also in 63% of patients who did not experience any growing of the initial infarct. This led to a determination coefficient of R²=0.058 (P=0.05) between the volume of DW/PW mismatch and the volume of infarct growth.

Quantification of Perfusion Parameters
It was not possible to quantify the perfusion in 1 patient because a reliable measure of the AIF could not be obtained (the AIF had a double peak, and the patient had atrial fibrillation and mitral regurgitation). The mean±SD values of MTT, CBV, and CBF calculated in INF (n=57), IGR (n=16), OLI (n=46), and the contralateral normal areas (n=119) are reported in Table 1, and individual CBV and CBF measurements are displayed in Figure 2. The MTT was significantly longer when abnormal areas were compared with normal areas, but it did not allow us to distinguish INF from IGR and IGR from OLI. The CBV was significantly different between all abnormal areas, with the lower value being observed in INF and the higher value being observed in OLI. The CBV was also significantly different when INF and OLI were compared with normal regions. However, the CBV measured in IGR and in normal areas was similar. The CBF was significantly different between the 4 areas of interest, except between INF and IGR. When only the abnormal areas were considered, a CBV threshold of 4.3% or a CBF threshold of 17 mL/min per 100 g always defined irreversibly infarcted areas, but such low values were observed in only 11 of 57 infarcted regions for CBV or 15 of 57 infarcted regions for CBF. Similarly, under a CBV threshold of 6.3% or a CBF threshold of 28 mL/min per 100 g, the areas were always nonviable on follow-up, but such low values were observed in only 29 of 73 nonviable regions (INF+IGR) for the CBV and 38 of 73 regions for the CBF.

View this table: [in this window] [in a new window]   Table 1. MTT, CBF, and CBV Measured in the 3 Functional Areas Defined in Our Stroke Model and in Normal Contralateral Mirror Regions

View larger version (22K): [in this window] [in a new window]   Figure 2. Individual measurements of CBF plotted against CBV for the 3 functional areas defined in our stroke model and in normal contralateral mirror regions. Lines represent the CBF and CBV thresholds that can be combined for the best prediction of infarct growth in the DW/PW mismatch region.

Figure 3 shows the ROC curves that evaluate the ability of MTT, CBV, and CBF values to predict infarct growth in the area of DW/PW mismatch. The sensitivity and specificity of some illustrative MTT, CBV, and CBF thresholds are presented in Table 2. The diagnostic performance of the MTT was significantly lower than that of the CBF. The CBF cutoff led to a slightly better (but not significantly different) accuracy compared with the CBV cutoff. Combining CBV and CBF thresholds slightly increased the accuracy of the test. The combined thresholds giving the highest accuracy (35 mL/min per 100 g for the CBF or 8.2% for the CBV) differentiated IGR from OLI with a sensitivity of 81% and a specificity of 76% and are illustrated in Figure 2.

View larger version (23K): [in this window] [in a new window]   Figure 3. ROC curves established to evaluate the ability of CBF, CBV, and MTT to predict infarct growth in the area of DW/PW mismatch. There was no significant difference between the diagnostic performance of CBF and CBV measurements, but the diagnostic performance of MTT was significantly less than that of CBF (P=0.014). The sensitivity and specificity for some illustrative cutoff values are reported in Table 2.

View this table: [in this window] [in a new window]   Table 2. Sensitivity and Specificity of Some Illustrative MTT, CBF, and CBV Thresholds for Predicting Infarct Growth in the Area of DW/PW Mismatch

	Discussion

Top Abstract Introduction Subjects and Methods Results Discussion References

 
In the present study, we have shown that quantitative CBV and CBF measurements were useful in the prediction of infarct growth in hyperacute stroke. Our data were derived from the retrospective analysis of patients presenting with hyperacute stroke of various sizes in various arterial territories because the indication of thrombolysis could be discussed in all subtypes of stroke.¹ In our study population, we observed a relatively small percentage (27%) of infarct growth that may be explained by the inclusion of many small infarcts.

The focus of the present study was the identification of the area at risk for infarction because it is the main target of hyperacute stroke therapy.² On the basis of a qualitative analysis of MR data, it has been proposed that the DW/PW mismatch represented the area at risk for infarction.^{4 6 8 10} In a nonselected stroke population, we confirmed that all patients with an infarct growth presented a mismatch, whereas the presence of a mismatch did not indicate that the infarct would eventually increase. Moreover, the enlargement volume was usually much smaller than the mismatch area, illustrating the insufficiency of this qualitative approach in identification of the area at risk for infarction. Like many other authors, we defined the DW/PW mismatch as the DW/apMTT mismatch because MTT maps clearly differentiate normal from hemodynamically impaired tissue in both gray and white matter.^{7 8 10 11 12} However, it is well known that qualitative apMTT maps substantially overestimate the region at risk because the zone of hemodynamic disturbance also encompasses the tissue supplied by collateral vessels, which is not at risk for infarction.¹⁴ The quantification of the severity of the perfusion deficit within the mismatch region is therefore essential for the identification of the actual tissue at risk.

The MTT was significantly prolonged in all abnormal areas, but it did not allow us to distinguish IGR from OLI. This confirms other reports of the literature illustrating that MTT is very sensitive in the differentiation of abnormal from normal regions but is not as good as CBF or CBV in characterizing the hemodynamically impaired areas and identifying the area at risk for infarction.^{7 8}

The CBF was significantly different between all functional regions defined in our stroke model, except between INF and IGR. It was postulated that the ischemic core could be reliably defined on DW images, and the quantification of CBF did indeed help us to distinguish IGR from OLI. Schematically, we could consider that the CBF in the area at risk fell in the approximate range of 20 to 42 mL/min per 100 g. Under 20 mL/min per 100 g, the tissue was probably irreversibly infarcted, and at >42 mL/min per 100 g, the tissue was probably not at risk for infarction (Figure 2).

The CBV was able to differentiate the 3 abnormal areas, but the mean CBV in IGR was similar to that of normal tissue. These results reflect the compensatory vasodilatation, leading to a high CBV in OLI, and the progressive failure of the autoregulation in IGR and INF, leading to a decrease of the CBV to "normal" values in IGR and to low values in INF.^{15 16}

The thresholds derived from the present study are higher than those reported from PET or xenon-CT studies.^{17 18 19} We have previously shown that our method for calculating CBV, CBF, and MTT was suitable in stroke patients and had minimal interuser variance.¹² The CBV and CBF values calculated in normal areas were in agreement with previous results reported in the literature, although they were slightly higher than those obtained with PET.¹² A systematic bias leading to the overestimation of CBV and CBF may be introduced by the underestimation of the AIF.¹² Theoretical models predict that the absence of correction for the delay and the dispersion of the AIF would lead only to a slight underestimation of slow flows in areas supplied by collaterals.²⁰ The Fourier transform method used in the present study is insensitive to the delay,¹³ and even if no attempt was made to correct for the dispersion, the error introduced by this factor alone is expected to be small and does not explain the observed differences between our thresholds and those reported from the PET literature. Besides methodological considerations, differences in the study population and in the definition of the areas of interest may also explain the observed differences. We are currently undertaking a study that will compare the regional CBF and CBV obtained with PET and MRI in the same subjects (normal volunteers and patients) and will also compare the MR-measured AIF with that obtained from blood sampling.

It is expected that instant CBF and/or CBV values are not perfect predictors of tissue outcome. Brain ischemia is a dynamic process, and the critical viability thresholds are dependent on the duration of the ischemia.^{17 18 21 22 23} The CBF measured in an area at risk may remain under critical threshold, leading to eventual infarction. Conversely, the CBF may partially recover, and the tissue may remain viable. A sudden drop of CBF may also occur in a region in which the flow was initially normal, leading to a second stroke. Indeed, this unpredictable event was observed in 2 of our 66 patients.

Like others,^{7 9 10 11} we made the assumption that the entire hyperintense area on DW images represented irreversibly infarcted tissue, and we focused our study on the prediction of tissue viability in the area surrounding the DW decrease. In humans, DW abnormalities typically evolve into infarction,^{4 6 11} but data from animals studies and patients treated with thrombolysis have shown that DW changes are potentially reversible if the apparent diffusion coefficient reduction is mild and/or if normal CBF is rapidly reestablished.^{22 23 24 25} As defined in the present study, the area of infarct growth may not represent the entire penumbra, defined as a region of ischemic but viable tissue that will eventually evolve into infarction.² Nevertheless, it represents an operative definition of the area at risk for infarction that can be used in clinical practice.⁷

Although the mean CBF and CBV tended to be smaller in INF than in IGR, the analysis of individual data did not allow a reliable identification of the ischemic core because many of the CBF and CBV values measured in INF were above the threshold of irreversible infarction (Figure 2). These results might be explained by (1) the early reperfusion that spontaneously occurred in some infarcts, (2) the fact that part of INF might still be viable, and (3) a possible bias introduced by the method used for measuring CBV and CBF that might overestimate slow flows.

Another factor that could limit the accuracy of CBF and/or CBV thresholds is the heterogeneity of these parameters in normal brain regions. Our data were measured in ROIs encompassing both gray and white matter, but the viability threshold might be different according to the location. With respective normal CBF (mL/min per 100 g) and CBV (%) values of 66 and 9.6 in cortical gray matter and 28 and 3.9 in white matter,¹² one can speculate that a viability threshold of 35 mL/min per 100 g for CBF and 8.2% for CBV may be valid for the gray matter but not for the white matter. We are currently examining the possibility of segmenting MR images to define specific thresholds according to the location or to compare the parametric maps to a normal template.

The calculation of quantitative CBF and CBV values could have an important impact on clinical management in triaging the patients who will benefit most from thrombolysis. In our population of 66 hyperacute stroke patients, if the decision to treat with thrombolysis had been taken on the basis of the presence of a DW/PW mismatch (all other clinical criteria being fulfilled), 47 patients would have been treated, but this procedure would have saved tissue at risk from infarction in only 17 patients and would have been useless in 30 patients, exposing them to unnecessary risks of hemorrhagic complications. If a combined CBV or CBF threshold of 8.2% or 35 mL/min per 100 g, respectively, was applied, only 19 of 65 patients would have been treated with thrombolysis (the quantification failed in 1 patient). The procedure would have been useless in only 6 patients, but 3 patients who would have benefited from the treatment would have been missed (2 of them extended their infarct into normally perfused areas). The benefit of thrombolysis in patients with small infarcts is questionable. The quantification of CBF could help to identify those few patients in whom thrombolytic therapy might be beneficial. Among the 19 patients with an initial infarct lesion <3 cm³ (including 7 infratentorial infarcts), 9 presented a DW/PW mismatch, and 2 lesions did grow >1 cm³: 1 in a normally perfused area and 1 in a penumbral area presenting a CBF inferior to 35 mL/min per 100 g. In an emergency setting, the limited step for using absolute quantification of perfusion parameters is the determination of the AIF, which takes 5 to 10 minutes. To overcome this limitation, we are currently working on a fully automatic method for calculating the AIF.

The present retrospective study has generated operational CBV and CBF thresholds for infarct growth prediction that need to be validated in a prospective population. The perfusion data will be smoothed to increase the signal-to-noise ratio and to allow absolute quantification of CBF and CBV on a voxel-by-voxel basis. In the area of MTT/DW mismatch, automatic thresholding will highlight the voxels with CBF and CBV values under the critical viability threshold. We speculate that the identification of areas with CBF of <35 to 38 mL/min per 100 g or CBV of 8.2% to 9.6% in the DW/PW region might be a criterion for selection of patients for thrombolysis, avoiding unnecessary, risky, and costly therapy in those with higher blood flow or volume.

Received August 28, 2000; revision received January 10, 2001; accepted February 2, 2001.

	References

Top Abstract Introduction Subjects and Methods Results Discussion References

 
1. The National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group. Tissue plasminogen activator for acute ischemic stroke. N Engl J Med. 1995;333:1581–1587.[Abstract/Free Full Text]

2. Fisher M. Characterizing the target of acute stroke therapy. Stroke. 1997;28:866–872.[Abstract/Free Full Text]

3. Jansen O, Schellinger P, Fiebach J, Hacke W, Sartor K. Early recanalisation in acute ischaemic stroke saves tissue at risk defined by MRI. Lancet. 1999;353:2036–2037.[Medline] [Order article via Infotrieve]

4. Barber PA, Darby DG, Desmind PM, Yang Q, Gerraty RP, Jolley D, Donnan GA, Tress BM, Davis SM. Prediction of stroke outcome with echoplanar perfusion- and diffusion-weighted MRI. Neurology. 1998;51:418–426.[Abstract/Free Full Text]

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

6. Warach S, Dashe JF, Edelman RR. Clinical outcome in ischemic stroke predicted by early diffusion-weighted and perfusion magnetic resonance imaging: a preliminary analysis. J Cereb Blood Flow Metab. 1996;16:53–59.[Medline] [Order article via Infotrieve]

7. Schlaug G, Benfield A, Baird AE, Siewert B, Lövblad KO, Parker RA, Edelman RR, Warach S. The ischemic penumbra operationally defined by diffusion and perfusion MRI. Neurology. 1999;53:1528–1537.[Abstract/Free Full Text]

8. Sorensen AG, Copen WA, Østergaard L, Buonanno FS, Gonzales RG, Rordorf G, Rosen BR, Schwamm LH, Weisskoff RM, Koroshetz WJ. Hyperacute stroke: simultaneous measurement of relative cerebral blood volume, relative cerebral blood flow, and mean tissue transit time. Radiology. 1999;210:519–527.[Abstract/Free Full Text]

9. Neumann-Haefelin T, Wittsack HJ, Wenserski F, Siebler M, Seitz RJ, Mödder U, Freund HJ. Diffusion- and perfusion-weighted MRI: the DWI/PWI mismatch region in acute stroke. Stroke. 1999;30:1591–1597.[Abstract/Free Full Text]

10. Barber PA, Davis SM, Darby DG, Desmond PM, Gerraty RP, Yang Q, Jolley D, Donnan GA, Tress BM. Absent middle cerebral artery flow predicts the presence and evolution of the ischemic penumbra. Neurology. 1999;52:1125–1132.[Abstract/Free Full Text]

11. Liu Y, Karonen JO, Vanninen RL, Østergaard L, Roivainen R, Nuutinen J, Perkiö, Könönen M, Hämäläinen A, Vanninen EJ, et al. Cerebral hemodynamics in human acute ischemic stroke: a study with diffusion- and perfusion-weighted magnetic resonance imaging and SPECT. J Cereb Blood Flow Metab. 2000;20:910–920.[Medline] [Order article via Infotrieve]

12. Smith AM, Grandin CB, Duprez T, Mataigne F, Cosnard G. Whole brain quantitative CBF, CBV and MTT measurements using MRI bolus tracking: implementation and application to data acquired from hyperacute stroke patients. J Magn Reson Imaging. 2000;12:400–410.[Medline] [Order article via Infotrieve]

13. Smith AM, Grandin CB, Duprez T, Mataigne F, Cosnard G. Whole brain quantitative CBF and CBV measurements using MRI bolus tracking: comparison of methodologies. Magn Reson Med. 2000;43:559–564.[Medline] [Order article via Infotrieve]

14. Neumann-Haefelin T, Wittsack HJ, Fink GR, Wenserski F, Li TQ, Seitz RJ, Siebler M, Mödder U, Freund HJ. Diffusion- and perfusion-weighted MRI: influence of severe carotid artery stenosis on the DWI/PWI mismatch in acute stroke. Stroke. 2000;31:1311–1317.[Abstract/Free Full Text]

15. Hatazawa J, Shimosegawa E, Toyoshima H, Ardekani BA, Suzuki A, Okudera T, Miura Y. Cerebral blood volume in acute brain infarction: a combined study with dynamic susceptibility contrast MRI and ^99mTc-HMPAO-SPECT. Stroke. 1999;30:800–806.[Abstract/Free Full Text]

16. Yonas H, Pindzola RR. Physiological determination of cerebrovascular reserves and its use in clinical management. Cerebrovasc Brain Metab Rev. 1994;6:325–340.[Medline] [Order article via Infotrieve]

17. Furlan M, Marchal G, Viader F, Derlon JM, Baron JC. Spontaneous neurological recovery after stroke and the fate of the ischemic penumbra. Ann Neurol. 1996;40:216–226.[Medline] [Order article via Infotrieve]

18. Heiss WD, Huber M, Fink GR, Herholz K, Pietrzyk U, Wagner R, Wienhard K. Progressive derangement of periinfarct viable tissue in ischemic stroke. J Cereb Blood Flow Metab. 1992;12:193–203.[Medline] [Order article via Infotrieve]

19. Yonas H, Sekhar L, Johnson DW, Gur D. Determination of irreversible ischemia by xenon-enhanced computed tomographic monitoring of cerebral blood flow in patients with symptomatic vasospasm. Neurosurgery. 1989;24:368–372.[Medline] [Order article via Infotrieve]

20. Østergaard L, Weisskoff RM, Chesler DA, Gyldensted C, Rosen BR. High resolution measurements of cerebral blood flow using intravascular tracer bolus passages, I: mathematical approach and statistical analysis. Magn Reson Med. 1996;36:715–725.[Medline] [Order article via Infotrieve]

21. Ueda T, Sakaki S, Yuh WTC, Nochide I, Ohta S. Outcome in acute stroke with successful intra-arterial thrombolysis and predictive value of initial single-photon emission-computed tomography. J Cereb Blood Flow Metab. 1999;19:99–108.[Medline] [Order article via Infotrieve]

22. Miyabe M, Mori S, van Zijl PC, Kirsch JR, Eleff SM, Koehler RC, Traystman RJ. Correlation of the average water diffusion constant with cerebral blood flow and ischemic damage after transient middle cerebral artery occlusion in cats. J Cereb Blood Flow Metab. 1996;16:881–891.[Medline] [Order article via Infotrieve]

23. Minematsu K, Li L, Sotak CH, Davis MA, Fisher M. Reversible focal ischemic injury demonstrated by diffusion-weighted magnetic resonance imaging in rats. Stroke. 1992;23:1304–1311.[Abstract/Free Full Text]

24. Kidwell CS, Saver JL, Mattiello J, Starkman S, Vinuela F, Duckwiler G, Gobin YP, Jahan R, Vespa P, Kalafut M, et al. Thrombolytic reversal of acute human cerebral ischemic injury shown by diffusion/perfusion magnetic resonance imaging. Ann Neurol. 2000;47:462–469.[Medline] [Order article via Infotrieve]

25. Schellinger PD, Jansen O, Fiebach JB, Heiland S, Steiner T, Schwab S, Pohlers O, Ryssel H, Sartor K, Hacke W. Monitoring intravenous recombinant tissue plasminogen activator thrombolysis for acute ischemic stroke with diffusion and perfusion MRI. Stroke. 2000;31:1318–1328.[Abstract/Free Full Text]

This article has been cited by other articles:

				R. E. Latchaw, M. J. Alberts, M. H. Lev, J. J. Connors, R. E. Harbaugh, R. T. Higashida, R. Hobson, C. S. Kidwell, W. J. Koroshetz, V. Mathews, et al. Recommendations for Imaging of Acute Ischemic Stroke: A Scientific Statement From the American Heart Association Stroke, November 1, 2009; 40(11): 3646 - 3678. [Full Text] [PDF]

				O. Zaro-Weber, W. Moeller-Hartmann, W.-D. Heiss, and J. Sobesky The Performance of MRI-Based Cerebral Blood Flow Measurements in Acute and Subacute Stroke Compared With 15O-Water Positron Emission Tomography: Identification of Penumbral Flow Stroke, July 1, 2009; 40(7): 2413 - 2421. [Abstract] [Full Text] [PDF]

				J. A. Alawneh, R. R. Moustafa, and J.-C. Baron Hemodynamic Factors and Perfusion Abnormalities in Early Neurological Deterioration Stroke, June 1, 2009; 40(6): e443 - e450. [Abstract] [Full Text] [PDF]

				H. Ay, E. M. Arsava, M. Vangel, B. Oner, M. Zhu, O. Wu, A. Singhal, W. J. Koroshetz, and A. G. Sorensen Interexaminer Difference in Infarct Volume Measurements on MRI: A Source of Variance in Stroke Research Stroke, April 1, 2008; 39(4): 1171 - 1176. [Abstract] [Full Text] [PDF]

				M. Takasawa, P. S. Jones, J. V. Guadagno, S. Christensen, T. D. Fryer, S. Harding, J. H. Gillard, G. B. Williams, F. I. Aigbirhio, E. A. Warburton, et al. How Reliable Is Perfusion MR in Acute Stroke?: Validation and Determination of the Penumbra Threshold Against Quantitative PET Stroke, March 1, 2008; 39(3): 870 - 877. [Abstract] [Full Text] [PDF]

				E. Bandera, M. Botteri, C. Minelli, A. Sutton, K. R. Abrams, and N. Latronico Cerebral Blood Flow Threshold of Ischemic Penumbra and Infarct Core in Acute Ischemic Stroke: A Systematic Review Stroke, May 1, 2006; 37(5): 1334 - 1339. [Abstract] [Full Text] [PDF]

				R.G. Gonzalez Imaging-guided acute ischemic stroke therapy: From "time is brain" to "physiology is brain". AJNR Am. J. Neuroradiol., April 1, 2006; 27(4): 728 - 735. [Abstract] [Full Text] [PDF]

				D. G. Na, E. Y. Kim, J. W. Ryoo, K. H. Lee, H. G. Roh, S. S. Kim, I. C. Song, and K.-H. Chang CT Sign of Brain Swelling without Concomitant Parenchymal Hypoattenuation: Comparison with Diffusion- and Perfusion-weighted MR Imaging Radiology, June 1, 2005; 235(3): 992 - 948. [Abstract] [Full Text] [PDF]

				D. G. Na, V. N. Thijs, G. W. Albers, M. E. Moseley, and M. P. Marks Diffusion-Weighted MR Imaging in Acute Ischemia: Value of Apparent Diffusion Coefficient and Signal Intensity Thresholds in Predicting Tissue at Risk and Final Infarct Size AJNR Am. J. Neuroradiol., September 1, 2004; 25(8): 1331 - 1336. [Abstract] [Full Text] [PDF]

				M. Yamada, S. Yoshimura, Y. Kaku, T. Iwama, H. Watarai, T. Andoh, and N. Sakai Prediction of Neurologic Deterioration in Patients with Lacunar Infarction in the Territory of the Lenticulostriate Artery Using Perfusion CT AJNR Am. J. Neuroradiol., March 1, 2004; 25(3): 402 - 408. [Abstract] [Full Text] [PDF]

				S. B. Coutts, J. E. Simon, A. I. Tomanek, P. A. Barber, J. Chan, M. E. Hudon, J. R. Mitchell, R. Frayne, M. Eliasziw, A. M. Buchan, et al. Reliability of Assessing Percentage of Diffusion-Perfusion Mismatch Stroke, July 1, 2003; 34(7): 1681 - 1683. [Abstract] [Full Text] [PDF]

				L. C. Shih, J. L. Saver, J. R. Alger, S. Starkman, M. C. Leary, F. Vinuela, G. Duckwiler, Y. P. Gobin, R. Jahan, J. P. Villablanca, et al. Perfusion-Weighted Magnetic Resonance Imaging Thresholds Identifying Core, Irreversibly Infarcted Tissue Stroke, June 1, 2003; 34(6): 1425 - 1430. [Abstract] [Full Text] [PDF]

				P. W. Schaefer, Y. Ozsunar, J. He, L. M. Hamberg, G. J. Hunter, A. G. Sorensen, W. J. Koroshetz, and R. G. Gonzalez Assessing Tissue Viability with MR Diffusion and Perfusion Imaging AJNR Am. J. Neuroradiol., March 1, 2003; 24(3): 436 - 443. [Abstract] [Full Text] [PDF]

				S.-M. Huang, N.-H. Chow, H.-L. Lee, and T.-J. Wu The Value of Color Flow Doppler Ultrasonography of the Superior Thyroid Artery in the Surgical Management of Graves Disease Arch Surg, February 1, 2003; 138(2): 146 - 151. [Abstract] [Full Text] [PDF]

				J. Fiehler, M. von Bezold, T. Kucinski, R. Knab, B. Eckert, O. Wittkugel, H. Zeumer, and J. Rother Cerebral Blood Flow Predicts Lesion Growth in Acute Stroke Patients Stroke, October 1, 2002; 33(10): 2421 - 2425. [Abstract] [Full Text] [PDF]

				T. Kucinski, O. Vaterlein, V. Glauche, J. Fiehler, E. Klotz, B. Eckert, C. Koch, J. Rother, and H. Zeumer Correlation of Apparent Diffusion Coefficient and Computed Tomography Density in Acute Ischemic Stroke Stroke, July 1, 2002; 33(7): 1786 - 1791. [Abstract] [Full Text] [PDF]

				F. Calamante, D.G. Gadian, and A. Connelly Quantification of Perfusion Using Bolus Tracking Magnetic Resonance Imaging in Stroke: Assumptions, Limitations, and Potential Implications for Clinical Use Stroke, April 1, 2002; 33(4): 1146 - 1151. [Abstract] [Full Text] [PDF]

				S. Warach Tissue Viability Thresholds in Acute Stroke: The 4-Factor Model Stroke, November 1, 2001; 32(11): 2460 - 2461. [Full Text] [PDF]

				C. Oppenheim, C. Grandin, Y. Samson, A. Smith, T. Duprez, C. Marsault, and G. Cosnard Is There an Apparent Diffusion Coefficient Threshold in Predicting Tissue Viability in Hyperacute Stroke? Stroke, November 1, 2001; 32(11): 2486 - 2491. [Abstract] [Full Text] [PDF]

				S. Schneweis, M. Grond, F. Staub, G. Brinker, M. Neveling, C. Dohmen, R. Graf, W.-D. Heiss, and A. Shuaib Predictive Value of Neurochemical Monitoring in Large Middle Cerebral Artery Infarction Editorial Comment Stroke, August 1, 2001; 32(8): 1863 - 1867. [Abstract] [Full Text] [PDF]

				C. B. Grandin, T. P. Duprez, A. M. Smith, C. Oppenheim, A. Peeters, A. R. Robert, and G. Cosnard Which MR-derived Perfusion Parameters are the Best Predictors of Infarct Growth in Hyperacute Stroke? Comparative Study between Relative and Quantitative Measurements Radiology, May 1, 2002; 223(2): 361 - 370. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Grandin, C. B. Articles by Cosnard, G. Search for Related Content PubMed PubMed Citation Articles by Grandin, C. B. Articles by Cosnard, G. Pubmed/NCBI databases Medline Plus Health Information MRI Scans Stroke Related Collections Acute Cerebral Infarction Computerized tomography and Magnetic Resonance Imaging