This Article Abstract 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 Marchal, G. Articles by Baron, J. C. Search for Related Content PubMed PubMed Citation Articles by Marchal, G. Articles by Baron, J. C.

(Stroke. 1996;27:599-606.)
© 1996 American Heart Association, Inc.

Articles

Prolonged Persistence of Substantial Volumes of Potentially Viable Brain Tissue After Stroke

A Correlative PET-CT Study With Voxel-Based Data Analysis

Gilles Marchal, MD; Vincent Beaudouin, BSc; Patrice Rioux, MD, PhD; Vincent de la Sayette, MD; François Le Doze, MD; Fausto Viader, MD; Jean Michel Derlon, MD Jean Claude Baron, MD

From Cyceron (G.M., V.B., J.M.D., J.C.B.), INSERM U320 (G.M., P.R., F.V., J.C.B.), CEA DSV/DRM (V.B.), and Services de Neurologie, CHU Cote de Nacre (V. de la S., F.L.D., F.V.), University of Caen, France.

Correspondence to Dr J.C. Baron, INSERM U320, Centre Cyceron, Bd Becquerel, BP, 5229, 14074 Caen, France. E-mail inserm-u320@cyceron.fr.

	Abstract

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Background and Purpose The existence in humans of brain tissue at risk for infarction but potentially viable (eg, the penumbra) remains unproven. One retrospective operational definition of such tissue includes its final infarction despite a relatively preserved or even normal cerebral metabolic rate of oxygen (CMRO₂) in the early hours after stroke onset. Although previous positron emission tomography (PET) studies identified tissue whose CMRO₂ declined from the acute to the subacute stage, in principle compatible with deteriorating penumbra, they all lacked a coregistered CT scan mapping of final infarct and an objective three-dimensional PET data analysis, while many patients were studied in the subacute (up to 48 hours) phase. We have evaluated whether tissue with CMRO₂ ranging above a threshold for presumably irreversible damage in the first 18 hours of middle cerebral artery territory stroke, but below it in the chronic stage, could be retrospectively identified within the final infarct volume.

Methods Our data bank comprises 30 consecutive patients with first-ever middle cerebral artery territory stroke prospectively studied with PET within the first 18 hours after clinical onset; the ¹⁵O equilibrium method was used to measure cerebral blood flow and CMRO₂. All survivors with the following criteria were eligible for the present study: (1) technically adequate chronic-stage PET performed in the same stereotaxic conditions, (2) coregistered CT scan also performed in the chronic stage, and (3) an infarct of sufficient dimension (>16 mm diameter) on late CT. Corresponding CT scan cuts and PET slices were exactly realigned, and the outlines of CT hypodensities were superimposed on the corresponding CMRO₂ matrix. Infarcted voxels with CMRO₂ values less than or greater than 1.40 mL/100 mL per minute (ie, the generally accepted threshold for irreversible damage) were automatically identified and projected on matrices of all other PET parameters and for both PET studies.

Results Eight patients (mean age, 78 years) were eligible for the present study. The acute-stage PET study was performed 7 to 17 hours after stroke onset and the chronic-stage PET 13 to 41 days later. Within the final infarct, mean CMRO₂ fell significantly from the acute- to the chronic-stage PET study (P<.001). Eventually infarcted voxels with acute-stage CMRO₂ values above the threshold were found in each of these eight patients; they were most often situated near the infarct borders and constituted 10% to 52% (mean, 32%) of the final infarct volume. The acute-stage CMRO₂ in these voxels ranged up to 4.13 mL/100 mL per minute but fell below 1.40 mL/100 mL per minute in 93% of them at the chronic-stage PET. In 7 of 8 patients the acute-stage mean cerebral blood flow ranged from 10 to 22 mL/100 mL per minute, and the mean oxygen extraction fraction was markedly increased (>0.70) in these voxels, consistent with a penumbral state.

Conclusions In a strictly homogeneous sample of prospectively studied patients, we have identified, up to 17 hours after stroke onset, substantial volumes of tissue with CMRO₂ well above the assumed threshold for viability that nevertheless spontaneously evolved toward necrosis. This tissue exhibited penumbral ranges of both cerebral blood flow and oxygen extraction fraction and thus could represent the part of penumbra that might be saved with appropriate therapy.

Key Words: cerebral blood flow • cerebral infarction • cerebral ischemia • positron emission tomography

	Introduction

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Animal studies have clearly shown that the brain is more resistant to ischemia than previously thought. Thus, after cerebrovascular occlusion, the transition from reversible ischemia to irreversible damage is a process that spans several hours.^{1 2 3 4} To demonstrate brain tissue at risk for infarction but still potentially recoverable (ie, the ischemic penumbra^{5 6} ) in humans is crucial, since it would constitute a well-defined target for early therapy. In a three-dimensional manner PET assesses both cerebral perfusion and CMRO₂, an index of the local integrated synaptic activity. One straightforward operational definition of at-risk tissue includes an unfavorable final outcome (ie, infarction) despite a relatively preserved or perhaps even normal CMRO₂ in the early hours after stroke. Although three previous studies^{7 8 9} reported deterioration of CMRO₂ from the acute to the subacute stage of stroke and interpreted this as transition from ischemia to necrosis, none directly documented the presence of still metabolically active tissue within the final CT-infarcted area in the acute stage. More precisely, since none of these early studies included late CT scans coregistered with PET, the topographical relationships between acute-stage physiology and final infarction have not been assessed. Therefore, the existence of an initially at-risk but eventually infarcted tissue has remained undocumented. With the exception of the study by Wise et al,⁷ most reported patients were investigated approximately 24 hours or later after stroke onset, a time when only little, if any, such tissue would be expected to still exist, while all studies except that of Heiss et al⁹ used first-generation PET systems with poor spatial resolution. Finally, in no previous study was the PET image analysis both objective (ie, with an automatic identification of the at-risk tissue) and comprehensive with respect to either three-dimensional infarct topography or time course of PET parameters in the at-risk tissue.

The present study was performed according to a prospective design that included (1) the investigation of patients in both the acute (within 18 hours after stroke onset) and the subacute stages; (2) a high-resolution PET system; (3) late CT scans coregistered with PET; and (4) an original, comprehensive, and objective PET data analysis. The specific aims of this study were (1) to identify, within the volume of the ultimately infarcted brain, that portion of tissue, if any, that in the acute stage was still consuming oxygen to amounts presumably associated with viability, but whose CMRO₂ had deteriorated below the threshold for irreversibility 3 weeks later and (2) to characterize the hemodynamic status of such tissue, if any, in the acute stage in relation to the concept of the ischemic penumbra.

	Subjects and Methods

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Patient Selection and Study Design
Of 30 consecutive patients (mean age, 73±10 years) with first-ever MCA territory ischemia prospectively studied with PET within 18 hours of stroke onset, 8 (4 women, 4 men; mean age, 78.4±8.0 years; range, 67 to 90 years) fulfilled all of the following criteria for eligibility: (1) survival until the late CT scan; (2) acceptance of follow-up PET study; (3) MCA territory infarct larger than 16 mm in its greatest diameter, as measured on high-resolution chronic-stage (14 to 63 days after onset) CT scanning, to avoid as much as possible the effects of partial voluming on measured PET variables; and (4) technically adequate PET studies at both scanning sessions (see below). A priori or post hoc causes of exclusion were neurological deficit lasting less than 24 hours, coma, organ failure, classic lacunar syndrome, cerebral hemorrhage or other lesions on admission CT scan, or secondary neurological event within the period of observation. The neurological deficit was quantified at admission and 60 days later with the use of the MCA stroke scale of Orgogozo and Dartigues,¹⁰ which is a validated scale for MCA territory infarction. Clinical deterioration or recovery was quantified with the indices of Martinez-Vila et al¹¹ [(day 60-day 0)/day 0 for deterioration, when the score at day 60 is below the score at day 0; (day 60-day 0)/(100-day 0) for recovery, when the score at day 60 is above the score at day 0]. In addition to standard medical workup, cervical and, if possible, transcranial Doppler ultrasound studies were systematically performed to help determine the mechanism of stroke. Informed consent was given by each patient or their relatives. This PET procedure has been approved by the ethics committee of Caen.

PET Study
Each patient underwent two PET studies: the first one began 7 to 17 hours after onset of clinical signs (PET_l), and the second one was performed 13 to 41 days later (PET₂). We used a high-resolution seven-slice PET camera (LETI TTV03; intrinsic resolution for ¹⁵O, 7x7x9 mm; full width at half maximum; coordinates x, y, z) and the classic ¹⁵O equilibrium method¹² with cerebral blood volume correction,¹³ following a procedure described in detail elsewhere.¹⁴ The scans were performed at rest in dimmed light. Blood pressure, heart rate, and electrocardiogram were continuously monitored. The head was positioned according to the glabella-inion line (Fox method¹⁵ ) and was repositioned in exactly the same coordinates for the second study (if technical difficulties of repositioning arose, the subject was retrospectively excluded). The seven PET slices were made parallel to the glabella-inion line, and their centers were -4, +8, +20, +32, +44, +56, and +68 mm relative to the glabella-inion line. A transmission scan with germanium 68 was performed before each study. Before the multiple steps that lead from the raw images to parametric maps were applied, the raw images were checked for adequate alignment and, whenever necessary, the computerized realignment method of Woods et al¹⁶ was applied. From the blood and brain radioactivities, regional CBF, cerebral blood volume, OEF, and CMRO₂ were calculated pixel by pixel according to classic equations and the procedure described by Marchal et al.¹⁴

Late CT scans
Late CT scans were performed 14 to 63 days after onset (mean±SD, 44±18 days) with the use of a CGR CE-12000 model scanner (Compagnie Générale de Radiologie; resolution, 1.5x1.5 mm; coordinates x, y). Seven cuts parallel to the glabella-inion line (as determined on a scout view and with reference to the bony landmarks determined on the same subject's lateral skull x-ray film obtained at PET scanning), centered at the same levels as the PET slices and with a thickness of 2 mm, were performed in each patient.

PET Data Analysis
We used a comprehensive procedure to analyze the PET data. This original method is based on a voxel-by-voxel analysis¹⁷ and the combination of SV and functional ROIs. To this end, the corresponding CT and PET matrices (from both PET studies) were first realigned according to the CT-defined outer brain outline by means of an interactive software running on SiliconGraphics/Indigo WorkStations. Then, to reduce both the volume of data and the statistical noise in each resulting voxel, the 1x1x9-mm PET matrix for each plane and parameter was transformed by averaging into an 8x8x9-mm matrix.

SV ROI
Without knowledge of the PET data, the infarcted area was outlined on the late CT scan as the area of hypodensity present on the relevant cuts, and these contours were projected onto the corresponding PET 8x8x9-mm voxel matrices (Fig 1). The SV ROI was defined as the cumulative set of 8x8x9-mm PET voxels contained within the infarct outlines across all relevant planes. (Regarding those voxels overlapping the infarct outline, we considered as belonging to the SV "infarct" ROI all 8x8x9-mm voxels with more than half of their 1x1x9-mm original voxels included within the infarct outlines [see Fig 1].) This procedure thus allowed us to examine the acute-stage CMRO₂ values for each individual 8x8x9-mm infarcted voxel and address the issue of at-risk tissue as defined above.

View larger version (40K): [in this window] [in a new window]   Figure 1. Illustrative example shows the method used for PET data analysis, which combines structural and functional ROIs. On the left, the infarct area is delineated on late CT scan; after PET-CT image realignment and PET matrix transformation (see "Subjects and Methods"), voxels with CMRO2 less and greater than 1.40 mL/100 mL per minute (in blue and yellow, respectively) are projected within the ultimate infarct for both PET studies. In this illustration only one brain cut is shown.

CMRO₂-Based Functional Voxel Classification (PET₁)
To assess the 8x8x9-mm voxels belonging to the SV ROI according to their viability status (ie, still viable or already irreversibly compromised), they were automatically classified into two subsets with respect to their acute-stage CMRO₂ value being either below or above 1.40 mL/100 mL per minute, ie, the widely accepted CMRO₂ critical threshold for irreversible damage.^{18 19 20} These earlier studies have shown that above this threshold the tissue is still viable but might deteriorate, while below this threshold it would already be irreversibly damaged. In the present study we used this threshold in an operational sense to single out that part of tissue that will spontaneously become infarcted but was still presumably viable in the acute stage. An example of this procedure is shown in Fig 1 (see also Fig 3). The voxels, so identified according to their CMRO₂ subset and their x, y, z coordinates, were then projected onto the matrices of all other PET parameters and for both PET studies. Because they would represent the initially at-risk but eventually infarcted tissue, the 8x8x9-mm voxels within the SV ROI with acute-stage CMRO₂ values greater than 1.40 mL/100 mL per minute will be referred to below as VOIs.

View larger version (18K): [in this window] [in a new window]   Figure 3. For each of the 8 eligible patients, the individual CMRO2 values of each eventually infarcted voxel with CMRO2 greater and less than 1.40 mL/100 mL per minute (on the left and right, respectively) in the acute stage (PET1) and the corresponding CMRO2 values of these voxels in the chronic stage (PET2) are represented. The horizontal lines at CMRO2=1.40 mL/100 mL per minute represent the putative irreversibility threshold (see "Subjects and Methods"). For the vast majority of voxels above the threshold, the CMRO2 fell below 1.40 mL/100 mL per minute at PET2, indicating a metabolic deterioration. Voxels with CMRO2 below the threshold at PET1 remained in the very low range at PET2, which suggests that they represented early irreversible damage.

Mirror ROI
To compare the findings in the finally infarcted tissue with the contralateral homologous territory, the SV (ie, infarct) ROIs were copied by symmetry onto the nonaffected hemisphere for each relevant PET plane by means of a dedicated software. The same voxel-based analysis described above was then applied to these "mirror" ROIs.

Determination of the Time Course of CMRO₂
Since the above procedure was simultaneously validated for both sets of CMRO₂ data (PET₁ and PET₂) and each voxel is identified according to its x, y, and z coordinates, it was possible to follow the changes in CMRO₂ values from PET₁ to PET₂ within each voxel belonging to both the SV ROI and the mirror ROI.

Volume of Infarct
An index of the infarcted tissue volume was operationally estimated by summing all 8x8x9-mm voxels assigned as "infarcted" across all the relevant CT scan cuts.

Statistical Procedure
We analyzed the data using the nonparametric Wilcoxon test and the Spearman rank test for small samples. Student's paired t test and the Pearson linear regression analysis were applied for the larger samples of selected voxels; an ANOVA with repeated measures was performed on the whole sample of voxels.

	Results

Top Abstract Introduction Subjects and Methods Results Discussion References

 
General Clinical and PET Data
The demographic, clinical, ultrasound, and CT data of the 8 eligible patients are summarized in Table 1. One patient (patient 3) dramatically improved, 2 others (patients 1 and 2) made an intermediate recovery, 4 (patients 4, 5, 7, and 8) were essentially stable, and 1 (patient 6) deteriorated.

View this table: [in this window] [in a new window]   Table 1. Clinical, Ultrasound, CT, and PET Characteristics

Evolution of Mean CMRO₂ from PET₁ to PET₂ Within the Infarct and the Mirror Region
The global statistical analysis concerning all voxels within the infarct volume (ie, whatever their initial CMRO₂ value) across the whole patient sample revealed a highly significant decline in CMRO₂ from PET₁ to PET₂ (from 1.17±0.73 to 0.59±0.42 mL/100 mL per minute [mean±SD]; P<.001 by ANOVA with repeated measures). In the mirror region, the CMRO₂ also significantly decreased from PET₁ to PET₂ (from 2.36 to 1.81 mL/100 mL per minute; P<.001 by ANOVA), but the percent decline was significantly smaller than in the infarct (P=.001, paired t test). Across the sample, the CMRO₂ was significantly higher in the mirror than in the infarct ROIs at both PET studies (P<.02, Wilcoxon).

Assessment of the CMRO₂ Threshold
To confirm the CMRO₂ threshold below which tissue is probably irreversibly damaged, we sought to determine the upper CMRO₂ 95% confidence limit for the infarcted voxels, as measured at PET₂, across all 8 patients. This confidence limit was defined from the entire population of infarcted voxels from all patients. To do this, histograms of each patient's CMRO₂ values were first constructed. This revealed frequent nongaussian distributions that reflected the different infarct volumes, ie, the larger the infarct, the more frequent the near-zero values. To control for this effect, which presumably represents partial voluming, each subject's histogram was normalized by the individual infarct volume. These normalized histograms were then summed across all subjects and rescaled at the median CMRO₂ value for each bin. This allowed us to obtain a weighted mean CMRO₂, which allowed us in turn to calculate the upper 95% confidence limit (as mean±2 SD). The weighted mean±SD of PET₂ CMRO₂ within the infarct area was 0.81±0.28 mL/100 mL per minute, yielding a 95% upper limit of 1.37 mL/100 mL per minute, which corresponds well to the operational threshold we extracted from the literature (ie, 1.40 mL/100 g per minute).

Ultimately Infarcted Tissue and Mirror Region
In each of the 8 patients studied, VOIs (ie, with CMRO₂ >1.40 mL/100 mL per minute at PET₁) were present within the area of final infarct, as shown in Fig 2. Most VOIs were located near the borders of the infarct. Individually, VOIs represented a substantial part of the final necrotic tissue (range, 10% to 52% for a mean of 32%; Table 2). There was no significant relationship between these percentages and the infarct volumes. In the mirror region, VOIs represented on average 76% of the voxel sample (range, 52% to 96%), which is a highly significantly larger fraction than in the infarct (P<.001, ²).

View larger version (61K): [in this window] [in a new window]   Figure 2. For each of the 8 patients eligible for this study, the relevant CT scan cuts showing, superimposed in yellow, the VOIs (ie, with CMRO2 >1.40 mL/100 mL per minute at the acute-stage PET study) within the final infarct area (outlined in white) are displayed. The VOIs were most often situated near the infarct borders, and their volume ranged from 10% to 52% of the infarct volume.

View this table: [in this window] [in a new window]   Table 2. Volume Indices of Final Infarct and VOIs1

Evolution of CMRO₂ From PET_l to PET₂ in VOIs
For the vast majority of VOIs in the final infarct, the CMRO₂ fell dramatically below 1.40 mL/100 mL per minute at PET₂. Across all 8 patients, this was the case for 277 of 297 VOIs (93%), and this decline was individually significant in each patient (P<.05 to .001 by Wilcoxon or paired t test for the entire VOI sample; Fig 3 and Table 3). Such a dramatic deterioration was true even for VOIs with CMRO₂ values as high as 4.13 mL/100 mL per minute at PET₁ (Fig 3). In contrast, in mirror regions only 19% of the VOIs exhibited such a decrease in CMRO₂ (P<.001 by ², compared with the final infarct).

View this table: [in this window] [in a new window]   Table 3. Mean (±SD) CMRO2, CBF, and OEF Values in VOIs at PET1 (Acute Stage) and PET2 (Chronic Stage) for the Eight Eligible Patients

CBF and OEF Values in VOIs
The mean CBF and OEF values for VOIs within the infarct area at PET₁ and PET₂ are shown in Table 3. At PET₁ the mean CBF ranged from 10 to 27 mL/100 mL per minute. Mean PET₁ OEF values were high (>0.70) in all patients but one (patient 5; note that this is the patient with both the highest CBF and lowest CMRO₂ of the whole sample).

From PET₁ to PET₂ there occurred a significant decline in mean CBF in 3 patients, a significant increase in 3 more, and no significant change in the remaining 2. The mean OEF significantly fell in all patients except again patient 5.

In the mirror region, mean CBF across the patient sample was within the normal range and significantly higher than in the infarct at both PET studies (mean±SD, 27.6±6.3 and 26.4±5.8 mL/100 mL per minute for the mirror, and 16.3±5.5 and 17.8±5.0 mL/100 mL per minute for the infarct at PET₁ and PET₂, respectively; P<.02 for both comparisons, Wilcoxon). The OEF was also in the normal range and significantly different in the mirror and infarct regions at both PET studies, but this time in opposite ways (mean±SD, 0.564±0.143 and 0.463±0.091 for the mirror, and 0.753±0.152 and 0.309±0.104 for the infarct at PET₁ and PET₂, respectively; P<.02 for both comparisons, Wilcoxon). Furthermore, there was no significant change in either CBF or OEF from PET₁ to PET₂ in the mirror regions.

Subset of Voxels With PET₁ CMRO₂ Below 1.40 mL/100 mL per Minute in the Final Infarct
In the vast majority (437 of 451, or 97%) of voxels with initial CMRO₂ below 1.40 mL/100 mL per minute in the infarct region, the PET₂ CMRO₂ either remained stable (ie, below the threshold) or exhibited a further decline (Fig 3). The corresponding OEF, which was generally low at PET₁, remained so or showed a further decline at PET₂ (data not shown). Individually, the decreases in CMRO₂ and OEF from PET₁ to PET₂ were statistically significant in 4 and 7 patients, respectively. The CBF significantly increased from PET₁ to PET₂ in 7 of 8 patients (data not shown).

	Discussion

Top Abstract Introduction Subjects and Methods Results Discussion References

 
In each of our 8 eligible patients, a substantial fraction of the finally infarcted tissue exhibited acutely a CMRO₂ greater than 1.40 mL/100 mL per minute and thus was potentially viable; the CMRO₂ in this tissue fell below this threshold at PET₂. One strong support to the view that this tissue was indeed potentially viable is the fact that, in 7 of our patients, it exhibited both a high OEF and a CBF in the penumbral range in the acute stage. To rule out potential methodological artifacts, we assessed the contralateral symmetrical mirror region as a reference. Although the mean CMRO₂ did exhibit a significant decrease from PET₁ to PET₂ in the mirror, this decrease was significantly relatively smaller, and the mirror CMRO₂ was significantly higher at both studies than that for the infarct. Furthermore, the fraction of voxels with CMRO₂ greater than 1.40 mL/100 mL per minute was significantly larger in the mirror region at PET₁, while the proportion of these voxels that deteriorated below this threshold at PET₂ was dramatically smaller for the mirror region. Finally, the mirror region did not exhibit the same pattern of CBF and OEF changes at PET_l and the consequent evolution from PET₁ to PET₂, as in the infarct area, in a highly significant way. These findings, which strongly contrast these two homologous regions, thus rule out methodological artifacts and emphasize our interpretation regarding the at-risk tissue. With respect to the metabolic decline from PET₁ to PET₂ in the mirror, we believe that this reflects neither a global effect nor a methodological problem, since in the present sample the ipsilateral cerebellar metabolic rate of oxygen did not change significantly from PET₁ to PET₂ (data not shown), as previously reported in a larger sample.²¹ The most likely explanation for this metabolic decline in the mirror region relates to the phenomenon of contralateral hemisphere "diaschisis,"²² which appears to develop several days after stroke (see Andrews²³ for review), presumably reflecting transcallosal fiber degeneration.^{24 25 26} Thus, the mirror region of an infarct may not be the most appropriate reference but was elected here because it has the same anatomic architecture as the affected region. A final issue concerns the existence of voxels with CMRO₂ below the chosen threshold in the mirror regions, which could appear surprising at first consideration. It presumably reflects the expected partial volume effects with neighboring ventricles and enlarged sulci as well as lower metabolism in white matter (see below for further discussion); as expected, however, these voxels were significantly less numerous in the mirror than the infarct regions at PET₁ (24% versus 68%; P<.001 by ²).

Our findings are consistent with earlier PET studies but are documented here with considerably improved methodology. Wise et al⁷ first documented a deterioration of CMRO₂ from the acute to the chronic stage in tissue exhibiting high OEF acutely and interpreted this as the transition from ischemia toward infarction; however, they used a low-resolution PET camera and reported only cases with extensive stroke; furthermore, the topographical relationships between tissue exhibiting this time course of CMRO₂ and the final infarct were not assessed. Hakim et al,⁸ using a voxel-by-voxel functional threshold procedure, reported a metabolic deterioration in the tissue with CBF between 12 and 18 mL/100 g per minute (designated "penumbral"). They used a low-resolution device and restricted their analysis to the brain surface despite the subcortical extension of the lesions, while only very few patients were studied in the acute (<24 hours) stage. Heiss et al,⁹ in their three-dimensional approach, reported a progressive metabolic derangement in the tissue bordering the initially most severely hypoperfused/hypometabolic tissue (mean stroke to PET interval, 23 hours). However, their procedure to define the "infarct core" was based on a visual inspection of PET images rather than an objective and preestablished threshold (indeed, in only 10 of their 16 cases were the CBF or CMRO₂ values in these regions at or below the previously reported thresholds for viable tissue). Furthermore, since there was no PET-CT coregistration, the "infarct border zone" was defined by concentric rims of fixed width, which therefore may have inadvertently included eventually noninfarcted tissue and excluded truly compromised tissue. Thus, neither the volume nor the exact topography of the deteriorating tissue was assessed in that work. In contrast, our study concerned only patients studied within 17 hours of stroke onset, applied a comprehensive and objective three-dimensional analysis procedure performed on coregistered PET-CT data sets, and was based on the combination of structural (ie, the irregular hypodense areas delineated on late CT scans) and functional (ie, voxels with values above a critical threshold) ROIs. In addition, rigorous eligibility criteria were established, which entailed the exclusion of patients with small infarcts or with PET_l-PET₂ or PET-CT misregistration.

In this study we used an operational CMRO₂ threshold of 1.40 mL/100 mL per minute to distinguish potentially viable from already irreversibly damaged voxels. Three previous PET investigations attempted to determine whether there exists a threshold of CMRO₂ below which the affected tissue would consistently be irreversibly damaged.^{18 19 20} All three studies achieved this by comparing the CMRO₂ ranges in brain regions intact or ultimately infarcted on delayed CT scans. They concurred in showing that CMRO₂ in established infarcts was consistently below a threshold ranging from 1.3 to 1.7 mL/100 g per minute, and all areas with acute-stage CMRO₂ below 1.5 mL/100 g per minute (as early as 2 to 6 hours after onset²⁰ ) consistently turned into infarction. Conversely, ischemic areas with CMRO₂ values above the threshold could either evolve toward necrosis or remain morphologically intact, ie, they were potentially viable. Thus, based on these investigations and following earlier studies of the same kind,^{8 9} we selected an operational value of 1.40 mL/100 mL per minute. We document here with an independent method that the CMRO₂ in more than 97.5% of the infarcted voxels indeed ranged below this threshold at the chronic-stage PET study. Furthermore, and in agreement with Heiss et al,⁹ we show here that virtually all voxels with CMRO₂ below 1.40 mL/100 mL per minute in the acute stage remained below this threshold in the chronic stage, consistent with the concept of irreversible damage. If the concept of a CMRO₂ threshold below which tissue cannot escape infarction seems well accepted, it remains that its exact value is unknown. Thus, it may vary with the type of tissue (eg, gray versus white matter, striatum versus cortex) and the duration of ischemia (with its value possibly increasing with time), although there is no firm evidence for either at the present time. In addition, voxels with CMRO₂ below this threshold may be found in nonischemic white matter, which normally has a twofold to fourfold lower CMRO₂ than gray matter,²⁷ especially near the ventricles (as a result of partial volume effects) or with high-resolution PET, which presumably explains our present findings for the mirror region. Regardless, the use of a lower CMRO₂ threshold in our study would have resulted in even larger fractions of potentially viable tissue, thereby strengthening rather than weakening our conclusions. As a matter of fact, our main finding is independent of the threshold concept, since voxels with CMRO₂ far above 1.40 mL/100 mL per minute did exhibit marked metabolic deterioration and evolve toward infarction (Fig 3).

One strength of our voxel-by-voxel method was to allow the analysis of CBF and OEF in voxels selected on the basis of both their CMRO₂ and their final outcome. In all our patients but one (patient 5), the mean CBF in the above-threshold CMRO₂ voxels ranged from 10 to 22 mL/100 mL per minute (Table 3), well within the classic penumbral range.⁶ Since the mean OEF was consistently above 0.70 and up to 1.00 in each of these 7 patients, this pattern of changes would be consistent with a still-evolving penumbra.²⁸ The relative preservation of oxygen metabolism in the at-risk tissue hours after stroke onset may be explained by oxygen supply being reduced but sufficient for maintaining mitochondrial respiration, as shown by the fact that CBF was only moderately reduced (in the penumbral range) and OEF was markedly increased.²⁸ Based on rodent studies of focal cerebral ischemia,²⁹ we assume that this represents energy-efficient respiration. Sequential baboon PET studies have documented a progressive expansion of profound hypometabolism from the initial core over at least 24 hours after permanent MCA occlusion,^{30 31} indicating that the CMRO₂ may be temporarily preserved in ultimately infarcted tissue. The mechanisms of such progressive metabolic deterioration remain undefined but may represent gradual neuronal attrition as a result of recurrent depolarization waves and glutamate release from the ischemic core, leading to depletion of energy stores and subsequent failure of transmembrane Na⁺-K⁺ homeostasis and membrane damage.³²

In patient 5 the tissue with above-threshold CMRO₂ displayed a higher range of CBF and no increase of OEF, suggesting that it was not penumbral but partially reperfused; the CMRO₂ lay close to the threshold (Table 3), and thus the PET study may have actually captured the final transition toward irreversibility. In their seminal studies, Powers et al¹⁹ also noted some apparent overlap between irreversibly damaged and still viable tissue around the threshold value. However, because this patient also had the smallest percentage of VOIs relative to infarct volume and only few VOIs were detected (Table 2), we cannot exclude a methodologically related error.

We are the first to document, within the ultimate infarct, the existence of a substantial volume of tissue with penumbral characteristics up to 17 hours after stroke onset. Although this tissue spontaneously evolved toward necrosis, it could represent the at-risk tissue that can be saved with appropriate therapy. To save these areas would represent a substantial benefit in terms of neurological function.³³ Although this study does not directly prove this tissue's viability (that is, we do not show actual survival with anti-ischemic therapy), we do have preliminary evidence in baboons that early (6 hours) MCA recanalization reverses the otherwise ineluctable spread of profound hypometabolism.³⁴ Were this also to be the case in humans, then the present results would speak in favor of extending the therapeutic window to 17 hours in appropriately selected patients.³⁵

	Selected Abbreviations and Acronyms

 

CBF = cerebral blood flow CMRO2 = cerebral metabolic rate of oxygen MCA = middle cerebral artery OEF = oxygen extraction fraction PET = positron emission tomography ROI = region of interest SV = structural-volumic VOI = voxel of interest

	Acknowledgments

 
This study was supported by grant 82/90 from CNAM-INSERM to Drs Marchal and Baron. We are grateful to the physicians of the Emergency Department and to the cyclotron and computer science technicians for their help in this work.

Received July 25, 1995; revision received November 6, 1995; accepted November 21, 1995.

	References

Top Abstract Introduction Subjects and Methods Results Discussion References

 

1. Jones TH, Morawetz RB, Crowell RM, Marcoux FW, FitzGibbon SJ, DeGirolami U, Ojemann RG. Thresholds of focal cerebral ischemia in awake monkeys. J Neurosurg. 1981;54:773-782. [Medline] [Order article via Infotrieve]
2. DeGirolami U, Crowell RM, Marcoux FW. Selective necrosis and total necrosis in focal cerebral ischemia: neuropathologic observations on experimental middle cerebral artery occlusion in the Macaque monkey. J Neuropathol Exp Neurol. 1984;43:57-71. [Medline] [Order article via Infotrieve]
3. Weinstein PR, Anderson GG, Telles DA. Neurological deficit and cerebral infarction after temporary middle cerebral artery occlusion in unanesthetized cats. Stroke. 1986;17:318-324. [Abstract/Free Full Text]
4. Kaplan B, Brint S, Tanabe J, Jacewicz M, Wang XJ, Pulsinelli W. Temporal thresholds for neocortical infarction in rats subjected to reversible focal cerebral ischemia. Stroke. 1991;22:1032-1039. [Abstract/Free Full Text]
5. Astrup J, Siesjö BK, Symon L. Thresholds in cerebral ischemia: the ischemic penumbra. Stroke. 1981;12:723-725. [Free Full Text]
6. Lassen NA, Fieschi C, Lenzi GL. Ischemic penumbra and neuronal death: comments on the therapeutic window in acute stroke with particular reference to thrombolytic therapy. Cerebrovasc Dis. 1991;1(suppl 1):32-35.
7. Wise RJS, Bernardi S, Frackowiak RSJ, Legg NJ, Jones T. Serial observations on the pathophysiology of acute stroke. Brain. 1983;106:197-222. [Abstract/Free Full Text]
8. Hakim AM, Evans AC, Berger L, Kuwabara H, Worsley K, Marchal G, Biel C, Pokrupa R, Diksic M, Meyer E, Gjedde A, Marrett S. The effect of nimodipine on the evolution of human cerebral infarction studied by PET. J Cereb Blood Flow Metab.. 1989;9:523-534. [Medline] [Order article via Infotrieve]
9. 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]
10. Orgogozo JM, Dartigues JF. Methodology of clinical trials in acute cerebral ischemia: survival, functional and neurological outcome measures. Cerebrovasc Dis. 1991;1(suppl 1):100-116.
11. Martinez-Vila E, Guillén F, Villanueva JA, Matias-Guiu J, Bigorra J, Gil P, Carbonell A, Martinez-Lage JM. Placebo-controlled trial of nimodipine in the treatment of acute ischemic cerebral infarction. Stroke. 1990;21:1023-1028. [Abstract/Free Full Text]
12. Frackowiak RSJ, Lenzi GL, Jones T, Heather JD. Quantitative measurement of regional cerebral blood flow and oxygen metabolism in man using ¹⁵O and positron emission tomography: theory, procedure, and normal values. J Comput Assist Tomogr. 1980;4:727-736. [Medline] [Order article via Infotrieve]
13. Pantano P, Baron JC, Crouzel C, Collard P, Sirou P, Samson Y. The ¹⁵O continuous-inhalation method: correction for intravascular signal using C¹⁵O. Eur J Nucl Med. 1985;10:387-391. [Medline] [Order article via Infotrieve]
14. Marchal G, Rioux P, Petit-Taboué MC, Sette G, Travère JM, Le Poec C, Courtheoux P, Derlon JM, Baron JC. Regional cerebral oxygen consumption, blood flow, and blood volume in healthy human aging. Arch Neurol. 1992;49:1013-1020. [Abstract]
15. Fox PT, Perlmutter JS, Raichle ME. A stereotactic method of anatomical localization for positron emission tomography. J Comput Assist Tomogr. 1985;9:141-153. [Medline] [Order article via Infotrieve]
16. Woods RP, Cherry SR, Mazziotta JC. Rapid automated algorithm for aligning and reslicing PET images. J Comput Assist Tomogr. 1992;16:620-633. [Medline] [Order article via Infotrieve]
17. Marchal G, Beaudouin V, Serrati C, Rioux P, Viader F, Baron JC. New automated analysis of metabolic PET images: application to acute ischemic stroke. Cerebrovasc Dis. 1992;2:235. Abstract.
18. Baron JC, Rougemont D, Bousser MG, Lebrun-Grandié P, Iba-Zizen MT, Chiras J. Local CBF, oxygen extraction fraction (OEF), and CMRO₂: prognostic value in recent supratentorial infarction in humans. J Cereb Blood Flow Metab. 1983;3(suppl 1):1-2. Abstract.
19. Powers WJ, Grubb RL, Darriet D, Raichle ME. Cerebral blood flow and cerebral metabolic rate of oxygen requirements for cerebral function and viability in humans. J Cereb Blood Flow Metab. 1985;5:600-608. [Medline] [Order article via Infotrieve]
20. Ackerman RH, Lev MH, Mackay BC, Katz PM, Babikian VL, Alpert NM, Correia JA, Panagos PD, Senda M. PET studies in acute stroke: findings and relevance to therapy. J Cereb Blood Flow Metab. 1989;9(suppl 1):S359. Abstract.
21. Serrati C, Marchal G, Rioux P, Viader F, Petit-Taboué MC, Lochon P, Luet D, Derlon JM, Baron JC. Contralateral cerebellar hypometabolism: a predictor for stroke outcome? J Neurol Neurosurg Psychiatry. 1994;57:174-179. [Abstract]
22. Lenzi GL, Frackowiak RSJ, Jones T. Cerebral oxygen metabolism and blood flow in human cerebral ischemic infarction. J Cereb Blood Flow Metab. 1982;2:321-335. [Medline] [Order article via Infotrieve]
23. Andrews RJ. Transhemispheric diaschisis: a review and comment. Stroke. 1991;22:943-949. [Abstract/Free Full Text]
24. Kataoka K, Hayakawa T, Yamada K, Mushiroi T, Kurada R, Mogami H. Neuronal network disturbance after focal ischemia in rats. Stroke. 1989;20:1226-1235. [Abstract/Free Full Text]
25. Yamauchi H, Pagani M, Fukuyama H, Ouchi Y, Nagahama Y, Matsuzaki S, Kimura J, Yonekura Y, Konishi J. Progression of atrophy of the corpus callosum with deterioration of cerebral cortical oxygen metabolism after carotid artery occlusion: a follow up study with MRI and PET. J Neurol Neurosurg Psychiatry. 1995;59:420-426. [Abstract]
26. Iglesias S, Marchal G, Rioux P, Beaudouin V, de la Sayette V, Le Doze F, Derlon JM, Viader F, Baron JC. A PET study of the role of the contralateral cerebral hemisphere in early neurological recovery after acute MCA ischemic stroke. J Cereb Blood Flow Metab. 1995;15(suppl 1):S184. Abstract.
27. Lebrun-Grandié P, Baron JC, Soussaline F, Loc'h C, Sastre J, Bousser MG. Coupling between regional cerebral blood flow and oxygen consumption in the normal human brain: a study with positron tomography and oxygen 15. Arch Neurol. 1983;40:230-236. [Abstract]
28. Baron JC. Pathophysiology of acute cerebral ischemia: PET studies in humans. Cerebrovasc Dis. 1991;1(suppl 1):22-31.
29. Nowicki JP, Gotti B, Poignet H. Temporal changes in mitochondrial activity after irreversible focal cerebral ischaemia in the mouse. J Cereb Blood Flow Metab. 1991;11(suppl 2):S513. Abstract.
30. Pappata S, Fiorelli M, Rommel T, Hartmann A, Dettmers C, Yamaguchi T, Chabriat H, Poline JB, Crouzel C, Di Giamberardino L, Baron JC. PET study of changes in local brain hemodynamics and oxygen metabolism after unilateral middle cerebral artery occlusion in baboons. J Cereb Blood Flow Metab. 1993;13:416-424. [Medline] [Order article via Infotrieve]
31. Touzani O, Young AR, Derlon JM, Beaudouin V, Marchal G, Rioux P, Mezenge F, Baron JC, Mackenzie ET. Sequential studies of severely hypometabolic tissue volumes after permanent middle cerebral artery occlusion: a positron emission tomographic investigation in anaesthetized baboons. Stroke. 1995;26:2112-2119. [Abstract/Free Full Text]
32. Pulsinelli W. Pathophysiology of acute ischaemic stroke. The Lancet. 1992;339:533-536. [Medline] [Order article via Infotrieve]
33. Brott T, Marler JR, Olinger CP, Adams HP, Tomsick T, Barsan WG, Biller J, Eberle R, Hertzberg V, Walker M. Measurements of acute cerebral infarction: lesion size by computed tomography. Stroke. 1989;20:871-875. [Abstract/Free Full Text]
34. Touzani O, Young AR, Derlon JM, Baron JC, MacKenzie ET. Evolution of severely hypometabolic tissue: temporary versus permanent middle cerebral artery occlusion (MCAO) in the baboon. J Cereb Blood Flow Metab. 1995;15(suppl 1):S327. Abstract.
35. Baron JC, von Kummer R, del Zoppo GJ. Treatment of acute ischemic stroke: challenging the concept of a rigid and universal time window. Stroke. 1995;26:2219-2221.[Free Full Text]

This article has been cited by other articles:

				A. Bruno, T. A. Kent, B. M. Coull, R. R. Shankar, C. Saha, K. J. Becker, B. M. Kissela, and L. S. Williams Treatment of Hyperglycemia In Ischemic Stroke (THIS): A Randomized Pilot Trial Stroke, February 1, 2008; 39(2): 384 - 389. [Abstract] [Full Text] [PDF]

				A. H. Tayal, R. Gupta, H. Yonas, T. Jovin, K. Uchino, M. Hammer, L. Wechsler, and J. M. Gebel Quantitative Perihematomal Blood Flow in Spontaneous Intracerebral Hemorrhage Predicts In-Hospital Functional Outcome Stroke, February 1, 2007; 38(2): 319 - 324. [Abstract] [Full Text] [PDF]

				D. Kim, R. Jahan, S. Starkman, A. Abolian, C.S. Kidwell, F. Vinuela, G.R. Duckwiler, B. Ovbiagele, P.M. Vespa, S. Selco, et al. Endovascular Mechanical Clot Retrieval in a Broad Ischemic Stroke Cohort AJNR Am. J. Neuroradiol., November 1, 2006; 27(10): 2048 - 2052. [Abstract] [Full Text] [PDF]

				L. A. Steiner and P. J. D. Andrews Monitoring the injured brain: ICP and CBF Br. J. Anaesth., July 1, 2006; 97(1): 26 - 38. [Abstract] [Full Text] [PDF]

				J. V. Guadagno, P. S. Jones, T. D. Fryer, O. Barret, F. I. Aigbirhio, T. A. Carpenter, C. J. Price, J. H. Gillard, E. A. Warburton, and J.-C. Baron Local Relationships Between Restricted Water Diffusion and Oxygen Consumption in the Ischemic Human Brain Stroke, July 1, 2006; 37(7): 1741 - 1748. [Abstract] [Full Text] [PDF]

				S. Rojas, A. Martin, C. Justicia, C. Falcon, N. Bargallo, A. Chamorro, and A. M. Planas Modest MRI Signal Intensity Changes Precede Delayed Cortical Necrosis After Transient Focal Ischemia in the Rat Stroke, June 1, 2006; 37(6): 1525 - 1532. [Abstract] [Full Text] [PDF]

				A. S. Cunningham, R. Salvador, J. P. Coles, D. A. Chatfield, P. G. Bradley, A. J. Johnston, L. A. Steiner, T. D. Fryer, F. I. Aigbirhio, P. Smielewski, et al. Physiological thresholds for irreversible tissue damage in contusional regions following traumatic brain injury Brain, August 1, 2005; 128(8): 1931 - 1942. [Abstract] [Full Text] [PDF]

				T. G. Phan, G. A. Donnan, P. M. Wright, and D. C. Reutens A Digital Map of Middle Cerebral Artery Infarcts Associated With Middle Cerebral Artery Trunk and Branch Occlusion Stroke, May 1, 2005; 36(5): 986 - 991. [Abstract] [Full Text] [PDF]

				W.-D. Heiss, J. Sobesky, U. v. Smekal, L. W. Kracht, F.-G. Lehnhardt, A. Thiel, A. H. Jacobs, and K. Lackner Probability of Cortical Infarction Predicted by Flumazenil Binding and Diffusion-Weighted Imaging Signal Intensity: A Comparative Positron Emission Tomography/Magnetic Resonance Imaging Study in Early Ischemic Stroke Stroke, August 1, 2004; 35(8): 1892 - 1898. [Abstract] [Full Text] [PDF]

				K. Saita, M. Chen, N. J. Spratt, M. J. Porritt, G. T. Liberatore, S. J. Read, C. R. Levi, G. A. Donnan, U. Ackermann, H. J. Tochon-Danguy, et al. Imaging the Ischemic Penumbra with 18F-Fluoromisonidazole in a Rat Model of Ischemic Stroke Stroke, April 1, 2004; 35(4): 975 - 980. [Abstract] [Full Text] [PDF]

				H S Markus Cerebral perfusion and stroke J. Neurol. Neurosurg. Psychiatry, March 1, 2004; 75(3): 353 - 361. [Abstract] [Full Text] [PDF]

				R. Markus, D.C. Reutens, S. Kazui, S. Read, P. Wright, B.R. Chambers, J.I. Sachinidis, H.J. Tochon-Danguy, and G.A. Donnan Topography and Temporal Evolution of Hypoxic Viable Tissue Identified by 18F-Fluoromisonidazole Positron Emission Tomography in Humans After Ischemic Stroke Stroke, November 1, 2003; 34(11): 2646 - 2652. [Abstract] [Full Text] [PDF]

				C. S. Kidwell, J. R. Alger, and J. L. Saver Beyond Mismatch: Evolving Paradigms in Imaging the Ischemic Penumbra With Multimodal Magnetic Resonance Imaging Stroke, November 1, 2003; 34(11): 2729 - 2735. [Abstract] [Full Text] [PDF]

				T. G. Jovin, H. Yonas, J. M. Gebel, E. Kanal, Y. F. Chang, S. Z. Grahovac, S. Goldstein, and L. R. Wechsler The Cortical Ischemic Core and Not the Consistently Present Penumbra Is a Determinant of Clinical Outcome in Acute Middle Cerebral Artery Occlusion Stroke, October 1, 2003; 34(10): 2426 - 2433. [Abstract] [Full Text] [PDF]

				M. Fisher Recommendations for Advancing Development of Acute Stroke Therapies: Stroke Therapy Academic Industry Roundtable 3 Stroke, June 1, 2003; 34(6): 1539 - 1546. [Abstract] [Full Text] [PDF]

				H.-C. Song, H.-S. Bom, K. H. Cho, B. C. Kim, J.-J. Seo, C.-G. Kim, D. J. Yang, and E. E. Kim Prognostication of Recovery in Patients With Acute Ischemic Stroke Through the Use of Brain SPECT With Technetium-99m--Labeled Metronidazole Stroke, April 1, 2003; 34(4): 982 - 986. [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]

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