Abstract
Objectives: In acute ischemic stroke the pattern of a perfusion-imaging (PI) lesion larger than the diffusion-weighted imaging (DWI) lesion may be a marker of the ischemic penumbra. We hypothesized that acute middle cerebral artery (MCA) occlusion would predict the presence of presumed “penumbral” patterns (PI > DWI), ischemic core evolution, and stroke outcome.
Methods: Echoplanar PI, DWI, and magnetic resonance angiography (MRA) were performed in 26 patients with MCA territory stroke. Imaging and clinical studies (Canadian Neurological Scale, Barthel Index, and Rankin Scale) were performed within 24 hours of onset and repeated at days 4 and 90.
Results: MCA flow was absent in 9 of 26 patients. This was associated with larger acute PI and DWI lesions, greater PI/DWI mismatch, early DWI lesion expansion, larger final infarct size, worse clinical outcome (p < 0.01) and provided independent prognostic information (multiple linear regression analysis, p < 0.05). Acute penumbral patterns were present in 14 of 26 patients. Most of these patients (9 of 14) had no MCA flow, whereas all nonpenumbral patients (PI ≤ DWI lesion) had MCA flow (p < 0.001). Penumbral-pattern patients with absent MCA flow had greater DWI lesion expansion (p < 0.05) and worse clinical outcome (Rankin Scale score, p < 0.05).
Conclusions: Absent MCA flow on MRA predicts the presence of a presumed penumbral pattern on acute PI and DWI and worse stroke outcome. Combined MRA, PI, and DWI can identify individual patients at risk of ischemic core progression and the potential to respond to thrombolytic therapy beyond 3 hours.
Thrombolytic therapy with tissue plasminogen activator (t-PA) in ischemic stroke patients improves outcome if given within 3 hours of onset.1 The majority of stroke patients, however, present beyond this therapeutic window, and not all patients meeting the strict criteria for treatment obtain benefit. There is a need for a rapid and noninvasive imaging modality to both better select patients for thrombolytic therapy and investigate whether this treatment can be given later than 3 hours in selected patients.
The combination of echoplanar MRI (EPI), perfusion imaging (PI), and diffusion-weighted imaging (DWI) allows the identification of different patterns of acute PI and DWI lesions from which predictions as to likely lesion evolution and stroke outcome can be made.2-6 When DWI lesions occur within larger PI lesions, they typically expand into the surrounding hypoperfused tissue. Such a pattern may mark the presence of the ischemic penumbra,2 which is defined as functionally impaired but potentially salvageable ischemic brain tissue surrounding the irreversibly damaged ischemic core.7 Patients with such presumed “penumbral” patterns would therefore be expected to benefit most from reperfusion therapies.2,6,8
Magnetic resonance angiography (MRA) provides different but complementary information to PI and DWI. MRA depicts vascular anatomy and flow dynamics in the distal internal carotid artery (ICA) and the circle of Willis and its major branches, and is useful in the evaluation of large vessel occlusive disease.9 Studies examining the relationship between MRA and dynamic contrast-enhanced PI have found good qualitative and quantitative correlations between major vessel occlusion and PI lesions10,11 as well as increased relative cerebral blood flow with middle cerebral artery (MCA) recanalization.12 Therefore, although the effects of proximal arterial occlusion are determined in part by the presence and degree of collateral circulation,13 ongoing MCA occlusion indicates tissue at risk of hypoperfusion and therefore infarction.
A recent study by Rordorf et al.6 found that MCA territory stroke could be classified into clinically and pathophysiologically important subtypes based on MRA, PI, and DWI. Patients with MCA stem occlusions had greater areas of PI/DWI mismatch and a final infarct size larger than the initial DWI lesion. The authors suggested this group of patients may gain the most benefit from reperfusion therapy. However, MRA, PI, and DWI examinations were performed only once, and clinical information was not reported; therefore, the dynamic relationships between lesions and neurologic and functional state were not examined. Furthermore, final infarct size was measured from CT or MRI studies obtained only 3 to 7 days after stroke onset so that the presence of subacute edema may have lead to overestimation of final infarct size.
In this prospective serial study we examined the relationship between changes in MCA patency and the subacute evolution of PI and DWI lesions and outcome at 3 months in ischemic stroke patients. We hypothesized that the absence of MCA flow, through either ICA or MCA occlusion, would be associated with larger PI lesions, greater expansion of DWI lesions over time, and worse clinical and radiologic outcome. Better delineation of these relationships may have important implications for the selection of patients for thrombolytic therapy.
Methods.
Patients.
Twenty-six patients (18 men, 8 women; age 66.6 ± 13.1 years, range 37 to 90) with acute MCA territory cerebral infarction (19 left and 7 right) were prospectively recruited at The Royal Melbourne Hospital between September 1996 and June 1998. We studied MCA territory stroke because the MCA is the vessel most frequently involved in stroke syndromes13 and is readily visualized on MRA. Stroke onset was defined as the time the patient was last known to be without neurologic deficit. The first 16 patients are the subjects of a previous report in which PI and DWI lesion volumes were found to predict stroke outcome.2
Patients were excluded if they had cerebral hemorrhage, preexisting significant nonischemic neurologic deficits (including dementia or extrapyramidal disease), or a history of prior stroke that would hamper interpretation of clinical and radiologic data. Patients enrolled in stroke therapy trials were not excluded. The study was performed with the approval of our institution’s Ethics Committee, and written informed consent was obtained from the patient or next of kin.
Clinical assessment.
The Canadian Neurological Scale (CNS), a validated neurologic impairment score,14 was performed just before the acute and subacute imaging studies. Outcome clinical assessments were performed on the same day as the final MR study and consisted of a repeat CNS score, the Barthel Index (BI), and the Rankin Scale (RS).15 The BI is a validated functional disability score, and the RS is a validated handicap scale. All clinical assessments were performed by a neurologist or neurology resident trained in their administration and without knowledge of the MRI results.
Imaging.
All MRIs were obtained using a 1.5-T EPI-equipped whole body scanner (Signa Horizon SR 120, General Electric, Milwaukee, WI). Sequences were always performed in the same order, with an initial T1-weighted sagittal localizer, diffusion-weighted sequence, magnetic resonance spectroscopy, perfusion sequence, a proton-density and T2-weighted fast spin double echo sequence (repetition time [TR]/echo time [TE]/TE, 3,500/10/60 msec), EPI spin-echo sequence, phase contrast MRA, and finally a contrast-enhanced T1-weighted sequence. Similar slice positions were used to facilitate comparisons. Only the MRA, DWI, PI, and T2-weighted imaging are reported here, with a total “table time” of approximately 20 to 30 minutes for these sequences.
Perfusion imaging.
Perfusion images in the first 16 patients were obtained using an EPI spin-echo sequence as previously described.2 In the subsequent 10 patients, an EPI gradient-echo sequence with a TR/TE of 2,000/70 msec was used. A total of 10 slices were obtained, with a slice thickness of 6 mm with a 1-mm gap, a matrix of 256 × 128, and a field of view of 40 × 20 cm. Images were obtained at 40 time points per slice, with a total imaging time of 1 minute 21 seconds. A Gd-DTPA bolus (0.1 mmol/kg) was administered by a power injector (Spectris MR injector, MEDRAD, Indianola, PA) at a rate of 5 mL/sec via an 18-gauge catheter in the antecubital fossa. The signal intensity time curve obtained was processed on a voxel-by-voxel basis to determine a time-to-bolus peak map, which we have termed the “relative mean transit time” (rMTT) map. We used rMTT maps to assess PI abnormality because they give the most visually distinct perfusion deficit borders2,3 and result in PI lesions of greater volume than other hemodynamic parameter maps.
Diffusion imaging.
DWI was obtained using a multislice, single shot spin-echo EPI sequence. The rapid acquisition times made cardiac or respiratory gating and special head restraint unnecessary. Slice thickness was 6 mm with a 1-mm gap, the number of slices set to include the whole brain (average of 15), with a matrix size of 256 × 128 and field of view of 40 × 20 cm. The remainder of the protocol in the first 22 patients was as previously described but briefly, resulted in five b values of increasing magnitude from 0 to 1,200 sec/mm2 applied in three orthogonal directions.2 The protocol was changed in the most recent five patients, with TR/TE of 10,000/100 msec, and the diffusion gradient strength varied so that there were three b values of increasing magnitude from 0 to 1,000 sec/mm2.16 The diffusion sensitizing gradient was applied in six directions (xy, xz, yz, x, y, z); however, for the purposes of this study, analyses were performed from the average of the measurements taken in the x, y, and z orthogonal directions. This gave the trace of the diffusion tensor, which is reported to minimize the effects of diffusion anisotropy.17 Imaging time was 2 minutes 8 seconds.
Magnetic resonance angiography.
MRAs were obtained using a 2-D phase contrast sequence in the region of the circle of Willis with slab thickness of 10 mm (1-mm gap) and velocity encoding speeds of 70 cm/sec. Gradients were applied in all three orthogonal directions, TR/TE 25/7.5 msec, flip angle 30°, matrix of 256 × 128, field of view 20 × 20 cm, and number of excitations = 2. Imaging time was 1 minute 50 seconds.
Data analysis.
Post-processing of MR images was performed using customized software based on a commercial image analysis application AVS (Advanced Visualization Systems, Waltham, MA), using an Indigo 2 workstation (Silicon Graphics Inc., Mountain View, CA). Quantitative analysis methods of the DWI and T2-weighted imaging have been described previously.2 On the rMTT maps, there was a clear visual distinction between the hyperintense region of abnormal rMTT and surrounding tissue, and regions of interest (ROIs) were outlined using a manual pixel-wise method. Known anatomic markings such as ventricles and large sulci were taken into account. The area of the ROI was multiplied by the slice thickness plus the interslice gap and then summed. Analyses used the average of two measurements taken on separate occasions by two of the investigators trained in the technique and blinded to the clinical data. Intra- and interobserver variability was less than 15%, which compares favorably with other reports.5
The MRAs were presented separately to two neuroradiologists, blinded to clinical data and the results of the other MR sequences, in a random patient and time of scan order and examined for evidence of vessel occlusion. Interobserver agreement occurred in 90% (64 of 71) of all MRAs, with 100% agreement in cases of MCA stem occlusion or ICA occlusion with no MCA flow. In the seven studies in which there was disagreement, the MRAs were jointly reanalyzed, and a final decision was reached by consensus. However, because of this disagreement, patients were dichotomized and analyzed based on the presence or absence of MCA flow.
Statistical analysis.
Demographic and time-of-scan data are presented as mean values ± standard deviation. Dependent variables are compared using nonparametric techniques except where normality of data could be proven, in which case parametric equivalents are preferred and presented as mean difference with 95% confidence intervals (CIs). Corrections are made for paired data and unequal variance where required. The Pearson product moment correlation coefficient and multiple linear regression analysis were used to compare the strength of association between variables. Results were considered statistically significant at the 5% level.
Results.
Seventy-two MRA studies were performed in conjunction with PI, DWI, and T2-weighted imaging examinations. Acute MRA and DWI studies were performed in all 26 patients (12.1 ± 7.6 hours, with nine studies within 6 hours), with PI studies in 25 (no acute PI in one patient because of technical difficulties). All patients had subacute PI and DWI studies (4.5 ± 2.3 days), with MRA studies in 25 (MRA abandoned in one because of patient restlessness). Twenty-one patients had outcome MRA and T2-weighted imaging studies (90.1 ± 30.3 days). Five patients did not have outcome imaging studies: one was unable to tolerate the MRI study, two declined further MRI but not clinical studies, and two patients had died. Clinical and imaging results from the acute and subacute studies in these patients were included in the analysis. In addition, outcome clinical scores were available in all five of these patients and were also included in the analysis; the two patients who had died were assigned outcome CNS and BI scores of 0 and a modified RS score of 5, and the remaining three had outcome clinical studies performed. Two patients were enrolled in the ECASS II trial of t-PA18 (both received placebo) and two others in trials of putative neuroprotective agents (1 Tirilazad, 1 Cerestat).
Nine of the 26 patients (35%) had evidence of MCA stem occlusion (8 patients) or ICA occlusion with no significant collateral MCA flow (1 patient) and are referred to as the “no MCA flow” group (figure 1) . In the remaining 17 patients (65%), MCA flow was present although in one patient there was ICA occlusion with MCA flow supplied by collateral circulation. These patients are referred to as the “MCA flow” group (figures 2 and 3⇓). There were no significant differences in time to acute study between patients with and without MCA flow (table) .
Figure 1. Echoplanar studies in Patient 5 (one of the earlier patients in this series) at 4 hours (A–C) and 5 days (D–F) from stroke onset. (A) Acute magnetic resonance angiography (MRA) showing occlusion of the right middle cerebral artery (MCA) (arrow). (B) Acute perfusion-imaging (PI) deficit (relative mean transit time map) in right MCA territory shows as a hyperintense region. (C) Acute isotropic diffusion-weighted imaging (DWI) shows early infarct as subtle hyperintensity. Note that the acute PI deficit is larger than the DWI lesion. (D) Subacute MRA studies show recanalization of the right MCA stem. (E) Subacute PI shows resolution of the PI deficit consistent with reperfusion with (F) DWI lesion expansion between the acute and subacute studies. This patient died before outcome studies could be obtained.
Figure 2. Echoplanar studies in Patient 24 (one of the later patients in this series) at 9.5 hours (A–C) and 3 days (D–F) from stroke onset. (A) Normal acute magnetic resonance angiography (MRA). (B) Acute right middle cerebral artery branch territory perfusion-imaging (PI) deficit (relative mean transit time map). (C) Acute isotropic diffusion-weighted imaging (DWI) shows early infarct as subtle hyperintensity within region of hypoperfusion. Note that the acute PI deficit is larger than the DWI lesion. (D) Normal subacute MRA. (E) Subacute PI shows resolution of the PI deficit consistent with reperfusion with (F) DWI lesion expansion between the acute and subacute studies.
Figure 3. T2-weighted imaging in Patient 24 at 94 days shows final infarct size approximates subacute diffusion-weighted imaging lesion.
Comparison of patients with acute middle cerebral artery (MCA) blood flow and with MCA occlusion on MR angiography
In the MCA flow group there was evidence on acute DWI (subsequently confirmed on final imaging studies) of subcortical infarctions in six patients and cortical branch infarctions in nine. In addition, one patient had a lacunar infarct, and another patient had MCA/posterior cerebral artery watershed territory infarctions. Of the six patients with subcortical lesions on acute DWI, three had normal MRA and PI studies, one had a small perfusion lesion that corresponded to the DWI lesion, and two had cortical branch perfusion lesions distant to the subcortical DWI lesions. Of the nine patients with cortical branch lesions on acute DWI, two had normal MRA and PI studies, four had normal MRA studies and small cortical PI lesions, and three had first order MCA branch occlusions. Of these three, one had normal PI, one had a PI lesion that involved approximately two-thirds of the MCA territory, and one did not have acute PI owing to technical difficulties. Thus, 9 of 17 (53%) patients with MCA flow had PI lesions (42.7 ± 44.0 cm3, range 0.9 to 128.9).
All nine patients with absent MCA flow had acute PI lesion volumes that were significantly larger than the PI lesions in patients with MCA flow (see table). The absent MCA flow group also had larger acute and subacute DWI lesion volumes, larger final infarct sizes, and worse clinical outcome and included the only two patients who died during the study.
In an earlier study that included 16 of the patients reported here,2 we found that acute PI and DWI lesion volumes correlated with clinical outcome and final infarct size. In univariate analysis, the presence or absence of MCA flow correlated with clinical outcome (CNS, r = 0.55, p = 0.01; BI, r = 0.54, p = 0.003; and RS, r = 0.58, p = 0.003) and final infarct size (r = 0.77, p = 0.0001). In a multiple linear regression analysis comparing the predictive value of the absence of MCA flow, acute CNS score, and acute PI and DWI lesion volumes, the absence of MCA flow provided independent prognostic information for 3-month CNS (coef = −4.14, p = 0.04) and RS scores (coef = 2.27, p = 0.04).
The mean acute PI deficit in all patients studied was larger than the mean acute DWI lesion (mean volume difference = 44.1 cm3, CI 16.7 cm3 to 71.3, p < 0.005). However, this difference was only significant in patients with no MCA flow (mean PI/DWI mismatch = 99.9 cm3, CI 41.4 to 158.4 cm3, p < 0.005) and not in those patients with MCA flow (mean PI/DWI mismatch = 12.6 cm3, CI −4.2 to 29.4 cm3, p = 0.13). In patients with absent MCA flow the final infarct size was smaller than the acute PI lesion in eight of nine patients and larger by 13% in the remaining patient. In 14 of 17 patients with MCA flow, the final infarct size was smaller than the acute PI or DWI lesion. The final infarct size in the remaining three MCA flow patients was larger than the acute PI and DWI lesions, but in each of these cases the final infarct size was less than 1.5 cm3.
In patients with MCA flow, there was a correlation between acute PI and DWI lesion volumes (r = 0.73, p = 0.001). In contrast, there was no correlation between acute PI and DWI lesion volumes where MCA flow was absent (r = 0.29, p = 0.45), mainly as the result of increased variability in the DWI lesion volume.
The overall mean DWI lesion volume (in all 26 patients studied) expanded between the acute and subacute studies (mean volume difference = 13.5 cm3, CI 3.5 to 23.5 cm3, p = 0.01); however, this expansion was greater in the group with absent MCA flow (mean volume difference = 35.1 cm3, CI 9.8 to 60.4 cm3, p = 0.01). These changes were accompanied by an overall reduction in the mean PI deficit volume by the subacute studies (mean volume difference = 46.2 cm3, CI 17.8 to 74.0 cm3, p < 0.005), consistent with reperfusion. The PI lesion volume decrease, however, was greatest in the group with no MCA flow (mean volume difference = 74.5 cm3, CI 2.4 to 146.6 cm3, p < 0.05). Thus the absence of MCA flow predicted a dynamic state with greater expansion of the DWI lesions and greater contraction of the PI lesions between the acute and subacute studies (see figure 1).
Of the nine patients with no MCA flow, eight went on to have subacute MRA studies. Restoration of MCA flow with recanalization of the ICA or MCA by the subacute studies was seen in seven of these eight patients (88%) and was accompanied by an improvement in the mean CNS score (1.4 ± 1.9 units), larger than that seen in the rest of the patients (0.9 ± 1.4 units) but not reaching statistical significance (p = 0.26, Mann-Whitney U test) possibly because of the small patient numbers. The only patient without recanalization had no change in the CNS score between the acute and subacute studies and died before outcome studies were obtained. In addition, the contraction of the PI deficit volume between the acute and subacute studies was significantly larger in those in whom there had been restoration of MCA blood flow (mean volume difference = 97.2 cm3, CI 9.9 to 184.5 cm3, p < 0.05). Outcome MRA studies were performed in 21 patients, and MCA flow was present in all.
Based on the acute MRI studies, patients with a PI deficit larger than the DWI lesion volume were labeled the “presumed penumbral” group. The remaining patients in whom the PI deficit volume was either absent or of a size similar to or smaller than the DWI lesion volume were labeled the “nonpenumbral group.” Of the 25 patients who had acute PI and DWI studies, 14 (56%) were in the presumed penumbral group and 11 (44%) were in the nonpenumbral group. There was a significant evolution of the PI and DWI lesions between the acute and subacute studies in the presumed penumbral patients with expansion of the DWI lesion volume (mean DWI increase = 24.4 cm3, CI 7.1 to 41.7 cm3, p = 0.009) and contraction of the PI lesion volume (mean PI decrease = 79.4 cm3, CI 34.7 to 124.1 cm3, p = 0.002). In contrast, there was no significant evolution in the nonpenumbral group of patients (mean DWI increase = 0.9 cm3, CI −1.2 to 2.9 cm3, p = 0.38; and mean PI decrease = 4.0 cm3, CI −1.4 to 9.3 cm3, p = 0.13).
The presumed penumbral group of patients included all nine with absent acute MCA flow and five with MCA flow (see figures 2 and 3⇑). All nonpenumbral patients had acute MCA flow. These results differed significantly from those expected by chance (p < 0.05, Fisher exact probability test). Thus the absence of MCA flow predicted patients with an acute PI deficit that was larger than the DWI lesion and who may therefore have an ischemic penumbra. Comparing the nine patients in the presumed penumbral group with absent MCA flow with the five patients in this group with MCA flow, there was no difference in time to acute scan (mean difference = 0.5 hours, CI −9.1 to 10.0 hours, p = 0.92). However, patients with no MCA flow had greater subacute DWI lesion expansion (mean expansion difference = 29.19 cm3, CI 4.4 to 55.4 cm3, p = 0.03), worse clinical outcome (RS; 3.7 ± 1.7 versus 1.2 ± 1.8, p < 0.05 and CNS; 6.9 ± 4.7 versus 10.8 ± 1.1, p = 0.05, Mann-Whitney U test), and larger final infarct size (mean difference = 89.2 cm3, CI 27.3 to 151.2 cm3, p = 0.01). Thus within the presumed penumbral group, patients with absent acute MCA flow had greater evolution of the DWI lesion leading to a larger final infarct and worse clinical outcome.
Discussion.
The major findings of this study are that the absence of MCA flow on MRA correlates with larger acute PI and DWI lesion volumes, the presence of presumed penumbral patterns (PI > DWI lesion volume), greater early DWI lesion expansion, larger final infarct size, and worse clinical outcome. After adjustment for other acute variables, the absence of MCA flow was also an independent predictor of clinical outcome. All patients with nonpenumbral patterns (PI ≤ DWI lesion volume or no PI lesion) had MCA flow; however, the lack of proximal vessel occlusion did not necessarily exclude a penumbral pattern. To our knowledge this is the first study to examine the relationship among MRA, PI, and DWI, as well as clinical outcome measures and infarct size at 3 months.
The presence or absence of MCA occlusion correlated with PI lesion volume, with proximal vessel occlusion indicating a larger volume of tissue at risk of hypoperfusion. As a group, patients with MCA flow had smaller acute PI lesions compared with patients without MCA flow. All patients with absent MCA flow had PI deficits, and restoration of MCA flow between the acute and subacute studies was associated with a reduction in PI lesion volume, indicating reperfusion. Previous studies have also found good qualitative and quantitative correlations between the presence of vessel occlusions and lesions on MR PI6,10,11 or SPECT,19 but PI has been limited to single levels only10,11 and included patients imaged more than 24 hours after stroke onset.10
Just over one-half (53%) the patients with MCA flow had PI lesions. Thus the presence of MCA flow on MRA does not necessarily exclude a PI lesion. We hypothesize that emboli may have lodged in MCA branches leading to distal vascular occlusion at a level below the spatial resolution of these MRA sequences, although an alternative possibility is the “no reflow” phenomenon.20,21 An earlier study has found that MRA abnormalities are not reliably seen with infarcts less than 2 cm in diameter.10 The MRA sequences performed in this study were optimized to give as much information in as short a time as possible in unwell and frequently uncooperative patients, and this required a compromise between image quality and acquisition time. Therefore, although major vascular occlusion was always identified, small branch occlusions were less reliably shown.
Absent MCA flow was associated with larger DWI lesion volumes. DWI identifies regions of restricted diffusional movement of water in the ischemic brain22-24 that occur as a function of the degree, duration, and intrinsic vulnerability of the ischemic brain to cerebral hypoperfusion. Because proximal vascular occlusions are associated with larger PI lesions, it is not surprising that absent MCA flow is associated with larger DWI lesion volumes. This result is in contrast to a previous report in which no statistically significant difference was seen in the DWI lesion volumes between patients with and without MCA flow,6 despite the DWI lesions in each of these two groups being of similar volume to those found in the current study. Failure to reach the 5% level of statistical significance may be related to the smaller patient numbers in the earlier study (17 versus 26).
Just over one-half (56%) of all patients had acute PI lesions that were of larger volume than the DWI lesions, i.e., presumed penumbral patterns. In this study, the nine penumbral-pattern patients without MCA flow had greater subacute DWI lesion expansion, worse clinical outcome, and a larger final infarct size than the five penumbral patients with MCA flow. Therefore, although patients with a PI lesion larger than the DWI lesion are at risk of DWI expansion and worse stroke outcome, those without MCA flow are at even greater risk. Five patients in the presumed penumbral group had MCA flow; so although the absence of MCA flow provides prognostic information in addition to that gained from PI and DWI, MRA alone cannot predict the presence or absence of a presumed penumbral pattern.
The description of patients with an initial PI lesion of greater volume than the DWI lesion as penumbral is purely speculative. It is not yet established that this PI/DWI mismatch identifies the existence of an ischemic penumbra. As originally described, the penumbra referred to tissue in which cerebral blood flow was below the range at which electrical activity became impaired but above the threshold at which potassium gradients across cell membranes were disturbed (an indication of cell death).7 It has been suggested that DWI lesions identify the irreversibly damaged tissue of the ischemic core.2,6 Diffusion lesions typically evolve into infarction,4,8,22 and there are close correlations between DWI lesion volume and final infarct size.2,5,25-28 They may, however, be reversed with very early restoration of cerebral blood flow.24,29 We have never seen more than a 10% decrease in DWI lesion volumes measured in the first week after a stroke, but none of our patients have been treated with t-PA, and early thrombolytic treatment may lead to reversal of at least some portion of the DWI lesion. Nevertheless, DWI lesions may remain useful markers of the ischemic core before thrombolytic treatment. In addition, DWI lesions typically expand into surrounding hypoperfused tissue, mirroring the variability expected with the ischemic penumbra. No such evolution occurs if the DWI lesion is smaller than or the same size as the PI lesion.2-6 We therefore argue that a pattern of a PI lesion that is of larger volume than a DWI lesion may indirectly mark the presence of the ischemic penumbra and have labeled this pattern as such to emphasize the risk of infarct core (DWI lesion) expansion.
The relationship between MCA flow and DWI and PI lesion volume is complex. In patients with MCA flow there was a strong correlation between acute DWI and PI lesion volumes. In contrast, there was no correlation between DWI and PI lesion volumes in patients with absent MCA flow. This lack of correlation was mainly the result of greater DWI lesion volume variability in relation to PI lesion size. All patients with absent MCA flow acutely had presumed penumbral patterns, and DWI lesion expansion occurred between the acute and subacute studies in each of these cases. These findings are similar to the study by Rordorf et al.6 in which patients with MCA occlusion had larger infarcts on CTs or MRIs performed 3 to 7 days after stroke onset compared with acute DWI lesions. The greater variability in DWI lesion size with respect to PI lesions and greater subacute DWI lesion evolution in patients with no MCA flow supports the hypothesis that the acute pattern of a PI lesion larger than a DWI lesion may be a marker of the presence of the ischemic penumbra.
In only one of nine patients with absent MCA flow was the final infarct size (98.4 cm3) larger than the acute PI lesion (86.0 cm3), and in this case the difference was not great. This is in contrast to the report by Rordorf et al.6 in which 5 of 10 patients without MCA stem flow had a final infarct size larger than the acute PI lesion as measured from a cerebral blood volume (CBV) map. Possible reasons for this apparent difference with the present study include expansion of the ischemic zone beyond the initial perfusion deficit as a result of excitotoxic damage30,31 or overestimation of final infarct size, which was measured from conventional CT or MRI at 3 to 7 days when there is likely to be some contribution to infarct volume from ischemic edema. In addition, we have found that in the first 24 hours after stroke onset, acute rMTT maps give PI lesions of greater volume than other hemodynamic parameter maps so that acute CBV maps may underestimate the extent of hypoperfusion.
The presence or absence of MCA flow on MRA predicted stroke outcome. Patients with no MCA flow had worse clinical outcome and larger final infarct size at 3 months than those with MCA flow. Acute PI and DWI lesion volumes correlate with clinical outcome and final infarct size.2,28 However, the presence or absence of MRA flow provided independent prognostic information on clinical outcome. Patients with MCA flow had smaller PI lesion volumes, with almost half having no PI lesions at all. These patients had either reperfused or had proximal vessel obstruction that had cleared and presumably resulted in downstream embolization by the time of the acute imaging studies, or only ever had distal blood flow obstruction. Regardless of the mechanism, these patients had less tissue at risk of infarction compared with patients with proximal obstruction at the time of imaging.
Stroke patients with presumed penumbral patterns may have the most potential to benefit from reperfusion therapies, whereas nonpenumbral patients have little potential to benefit from such therapy2,6,8 unless there is a PI deficit and treatment can be given early enough to reverse the DWI lesion. The time to acute MR study in the penumbral group as a whole was 11.7 hours, with a range of 2.5 to 23 hours. Such prolonged penumbral persistence has previously been noted with functional MRI2,3 and PET32-34 studies. Thus, potentially salvageable ischemic tissue may be present in selected patients beyond 3 hours. The results from this study suggest that combined MRA, PI, and DWI may identify stroke patients at greatest risk of ischemic core expansion and therefore allow selection of individual patients with the potential to respond to thrombolytic therapy both within and beyond a 3-hour time window. These hypotheses now require investigation by prospective, randomized clinical trials.
Acknowledgments
Supported by the National Health and Medical Research Council, the National Stroke Foundation, and the Neurological Foundation of New Zealand, V.J. Chapman Research Fellowship (P.A.B.).
- Received September 1, 1998.
- Accepted December 19, 1998.
References
Letters: Rapid online correspondence
REQUIREMENTS
You must ensure that your Disclosures have been updated within the previous six months. Please go to our Submission Site to add or update your Disclosure information.
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
You May Also be Interested in
Alert Me
Recommended articles
Prediction of stroke outcome with echoplanar perfusion- and diffusion-weighted MRI
Evaluation of hyperacute infarct volume using ASPECTS and brain CT perfusion core volume
Combined 1H MR spectroscopy and diffusion-weighted MRI improves the prediction of stroke outcome
Ischemic diffusion lesion reversal is uncommon and rarely alters perfusion-diffusion mismatch