Skip to main content
NCBI home page
Neurology logoLink to Neurology
. 2023 Feb 7;100(6):e627–e638. doi: 10.1212/WNL.0000000000201498

Association of Perfusion Lesion Variables With Functional Outcome in Patients With Mild Stroke and Large Vessel Occlusion Managed Medically

Peng Wang 1,*, Wenhuo Chen 1,*, Chushuang Chen 1,*, Andrew Bivard 1, Geng Yu 1, Mark W Parsons 1,†, Longting Lin 1,†,✉; on behalf of INSPIRE1
PMCID: PMC9946183  PMID: 36307224

Abstract

Background and Objectives

The best management of patients with mild stroke and large vessel occlusion (LVO) remains unclear. This study aimed to identify perfusion imaging predictors of poor functional outcome in such patients.

Methods

This cohort study retrospectively selected patients enrolled in the International Stroke Perfusion Imaging Registry between August 2011 and April 2022. The registry enrolled patients with acute ischemic stroke and with baseline CT perfusion scanned within 24 hours of stroke onset. This study identified patients with mild symptoms, defined by an NIH Stroke Scale score of ≤5. Patients with LVO of anterior circulation were selected. This study further selected patients who received medical management and excluded patients who received endovascular treatment. The primary outcome was poor functional outcome defined as a modified Rankin Scale of 3–6 at 3 months. Perfusion lesion was defined by delay time > 3 seconds on CTP. Regression analyses were used to identify clinical and imaging variables that predicted poor functional outcome.

Results

A total of 139 patients with mild stroke were included, of whom 27 (19%) had poor functional outcome. Patients with poor outcome, compared with those with good outcome, had much larger perfusion lesion volume (median 80 mL vs 41 mL, p < 0.001). Perfusion lesion was a significant predictor of poor outcome in either univariable regression (crude OR = 1.02, 95% CI = [1.01–1.03]) or multivariable regression model (adjusted OR = 1.01, 95% CI = [1.01–1.02]), adjusting for occlusion site, good collaterals, baseline stroke severity, age, IV thrombolysis (IVT), and onset to scan time. A perfusion lesion of 65 mL was the optimal cutpoint to identify poor functional outcome (sensitivity = 59%, specificity = 77%). Patients with perfusion lesion ≥65 mL, compared with patients with perfusion lesion <65 mL, showed a much higher rate of poor functional outcome (38% vs 11%, p < 0.001). Of the 139 patients in this study, 95 received IVT. Patients treated with or without IVT did not influence their outcomes (crude OR = 0.74, 95% CI = [0.31–1.78]).

Discussion

A perfusion lesion of ≥65 mL predicted poor functional outcome in mild stroke patients with LVO.

In patients with ischemic stroke and large vessel occlusion (LVO), endovascular thrombectomy (EVT) has become a standard of care up to 24 hours after onset. However, most randomized trials of EVT have excluded patients with mild symptoms,1,2 although up to 10% of patients with LVO present with mild symptoms,3 defined as an NIH Stroke Scale (NIHSS) score of ≤5. The dearth of level 1 evidence means that there are no clear guidelines on the best management of such patients. A number of observational and cohort studies have been conducted but have reached conflicting conclusions, with some suggesting a benefit from EVT4-6 whereas others reported no benefit of EVT over best medical management.7,8 Several studies9-11 even found that EVT increased the risk of hemorrhagic complications without improving clinical outcomes. A recent meta-analysis12 concluded that there was no benefit of EVT over best medical management overall. One possible explanation for the different findings is the lack of patient selection. CT perfusion (CTP) utilization is increasing in patient selection for EVT in the extended time window.1,2 However, previous studies that examined the treatment effect of EVT on patients with mild LVO stroke did not include CTP in their analyses.4-11

EVT shall be targeted at patients with mild LVO stroke who would deteriorate and result in poor outcome with best medical management. Patients with mild stroke symptoms often have a favorable natural history with best medical management, but approximately 20% of such patients with LVO experience early neurologic deterioration13 and 15% of patients with mild LVO stroke have poor functional outcomes.14 Previous attempts to identify clinical and radiologic predictors of poor outcome or early neurologic deterioration have had limited success.13 Previous studies suggest that the proximal occlusion site, an angiographic imaging marker, may help to predict poor prognosis of patients with mild LVO stroke15 and help to select such patients for endovascular treatment.16,17 CTP can provide more detail regarding the volume and severity of hypoperfusion, which may provide stronger imaging predictors of prognosis in mild stroke. This study aimed to identify predictors from CTP imaging markers for poor functional outcome of patients with mild stroke.

Methods

This was a retrospective cohort study of patients in the International Stroke Perfusion Imaging Registry (INSPIRE). Data in this study were collected between August 2011 and April 2022 and from 22 sites (9 Australian, 12 Chinese, and 1 Canadian). The follow-up period of INSPIRE participants was 3 months.

Standard Protocol Approvals, Registrations, and Patient Consents

Institutional ethical approvals were obtained, and written informed consent was obtained for each patient for their data to be used as part of the INSPIRE. The INSPIRE and all associated analyses are conducted in accordance with the Declaration of Helsinki.

Inclusion and Exclusion Criteria

Inclusion criteria were as follows: (1) Patients diagnosed with ischemic stroke and with baseline CTP scanned within 24 hours of stroke onset; (2) patients with mild stroke defined as a baseline NIHSS score of ≤5; and (3) patients with LVO of anterior circulation detected on baseline CT angiography (CTA), which included internal carotid artery, M1 segment of the middle cerebral artery, and proximal M2 segment of the middle cerebral artery. Patients with occlusion on the anterior cerebral artery and distal M2 segment of the middle cerebral artery were excluded; patients with occlusion located at posterior circulation were excluded; and patients with isolated cervical intracranial artery were also excluded. Patients who received endovascular treatment were excluded because the endovascular treatment might be proposed as a rescue treatment for clinical deterioration. Patients with incomplete 3-month follow-up information were further excluded in this study.

Clinical and Imaging Parameters

Imaging data were processed centrally by the INSPIRE imaging analysis team (L.L., C.C., A.B., and M.P.), blinded to clinical data. Baseline CT imaging included brain noncontrast CT (NCCT), CTP, and CTA, obtained with various CT scanners (64, 128, 256, or 320 detectors, with Toshiba [Tokyo, Japan], Siemens [Munich, Germany], or GE [Cleveland, OH] scanners). The CTP raw data were centrally processed by commercial software MIStar (Apollo Medical Imaging Technology, Melbourne, VIC, Australia). CTP parameters were generated by applying the mathematical algorithm of singular value decomposition with delay and dispersion correction.18 A threshold of >3 seconds in delay time (DT) was used to delineate the perfusion lesion, within which tissue with cerebral blood flow (CBF) ≤30% was identified as infarct core.19-21 Penumbra volume was defined as the DT > 3 s volume subtracting the CBF < 30% volume. Severe hypoperfusion lesion volume was defined as DT > 6 seconds. Collateral flow and occlusion sites were assessed on CTA. Collateral flow was assessed using the Miteff scale22 and dichotomized into good collateral flow if the occluded vessel was reconstituted by retrograde flow up to the distal end of the occlusion or poor collateral flow otherwise. The Alberta Stroke Program Early CT Score (ASPECTS) was estimated on NCCT.

The following clinical data were extracted from INSPIRE: baseline vascular risk factors, IV thrombolysis (IVT), time from onset to scan and time from onset to IVT treatment, prior disease history including prior history of stroke, drug taken history, stroke etiology of large artery atherosclerosis classified by the Trial of Org 10172 in Acute Stroke Treatment,23 blood glucose level, baseline NIHSS, follow-up NIHSS assessed at 24–72 hours after stroke, and 3-month modified Rankin Scale (mRS). The use of IVT was at the discretion of treating clinicians in accordance with local protocols.

Outcome Definitions

The primary outcome was poor functional outcome defined as a mRS of 3–6 vs good outcome of 0–2 at 3 months after stroke. The secondary outcome was early neurologic deterioration, defined as an increase of NIHSS ≥4 points from baseline to 24–72 hours without evidence of a parenchymal hemorrhage (PH). The safety outcome included PH and symptomatic intracerebral hemorrhage (sICH). PH was assessed on follow-up CT or MRI. sICH was defined according to ECASS II criteria24: any intracerebral hemorrhage with early neurologic deterioration of NIHSS increase ≥4 from baseline to 24-hour assessment.

Statistical Analysis

All statistical analyses were performed using STATA 13.0 (Stata Corp, College Station, TX), with CI set at 95% and an alpha level set at 0.05.

Patients with incomplete primary outcome data were excluded from the analysis. Continuous data were summarized as median and interquartile range; between-group differences were tested by the Wilcoxon rank-sum test. Categorical variables were described as proportions and compared with the Pearson χ2 test.

Univariable logistic regressions were used to identify CTP predictors and other potential predictors of poor functional outcome, which were then included in a multivariable logistic regression model. The multivariable regression model also included age, baseline NIHSS, IVT, and onset to scan time for their clinical significance. Goodness of fit was assessed using receiver operating characteristic. Cutpoint thresholds for continuous predictors were analyzed using ROC curves, with cutpoints selected to maximize the Youden index and resulting sensitivity and specificity described. The mRS at 3 months was also assessed as an ordinal outcome using ordinal logistic regression.

Sensitivity analyses were then conducted as follows: (1) redefinition of the primary outcome as poor outcome of 2–6 vs good outcome of 0–1; (2) redefinition of CTP predictors by a different algorithm, Tmax generated by standard singular value decomposition without delay or dispersion correction, and a threshold of Tmax > 6 seconds was used to delineate the perfusion lesion. Further sensitivity analyses were conducted on the following 2 subgroups: (1) patients with onset to scan time ≤6 hours and (2) patients with recanalization information assessed by modified Thrombolysis in Cerebral Infarction.25,26 The subgroup analysis plan and results are detailed in eAppendix 1 (http://links.lww.com/WNL/C457).

Data Availability

Anonymized data not published within this article will be made available by request from qualified investigators whose proposal of data use has been approved by an independent review committee.

Results

A total of 586 patients with mild stroke were identified in INSPIRE, of whom 263 demonstrated LVO located at the internal carotid artery, M1, or proximal M2 segment of the middle cerebral artery. Of the 263 patients, 112 patients were excluded for receiving endovascular treatment, and 151 patients were identified with best medical management. Of the 151 patients, 12 patients were excluded because of incomplete 3-month outcome data for this study. Therefore, a total of 139 patients were included in the final analyses of this study.

Compared with the 139 patients included in this study, the 12 patients missing 3-month outcome had comparable CTP imaging profiles, with a median perfusion lesion of 43 mL vs 50 mL (p = 0.767), penumbra volume of 35 mL vs 38 mL (p = 0.858), core volume of 3.0 mL vs 3.0 mL (p = 0.591), and severe hypoperfusion volume of 6.0 mL vs 6.0 mL (p = 0.958).

Patient Characteristics

Within the 139 total patient cohort, 19% (27/139) resulted in poor outcome at 3 months. Baseline patient characteristics stratified by poor vs good outcome are listed in Table 1. Patients with poor outcome had larger perfusion lesion volumes (median 80 mL vs 41 mL, p < 0.001), larger penumbra volumes (median 62 mL vs 34 mL, p < 0.001), and larger severe hypoperfusion lesion volumes (median 19 mL vs 5.0 mL, p = 0.006). Patients with poor outcome also had less good collateral flow (46% vs 72%, p = 0.011) and more proximal occlusion (ICA of 33% vs 11%, M1 of 33% vs 41%, and M2 of 33% vs 48%, p = 0.013) compared with patients with good outcome.

Table 1.

Patient Characteristics of Mild LVO Stroke

graphic file with name WNL-2022-201360t1.jpg

Open in a new tab

Regarding clinical factors, patients with poor outcome had slightly higher stroke severity scores at baseline (median NIHSS of 5 vs 3, p = 0.003). There was no difference between the 2 groups in age (median of 70 vs 66, p = 0.164), IVT rate (63% vs 70%, p = 0.503), onset to IVT treatment time (median 3.3 vs 2.8 hours, p = 0.257), or onset to CTP scan time (median 2.6 vs 2.3 hours, p = 0.439). Vascular risk factors, stroke etiology, and blood glucose showed no statistically significant difference between groups, as were other baseline imaging markers including core volume and ASPECTS (Table 1).

Predictors of Poor Outcome in Univariable Regression

Univariable regression confirmed the following CTP imaging markers as predictors of 3-month poor outcome: perfusion lesion volume (crude OR = 1.02 [1.01–1.03], p < 0.001), penumbra volume (crude OR = 1.02 [1.01–1.03], p = 0.001), and severe hypoperfusion lesion volume (crude OR = 1.03 [1.01–1.05], p = 0.016). With the increase of perfusion lesion volume, the odds of having a poor outcome increased significantly (Figure 1, A and B, AUC = 0.71 [0.60–0.83]). A similar relationship was observed between poor outcome and penumbra volume (Figure 1, C and D, AUC = 0.71 [0.60–0.82]). Severe hypoperfusion volume had slightly lower predictive power of poor outcome (AUC = 0.68 [0.55–0.78]).

Figure 1. CTP and 3-Month Poor Outcome of Patients With Mild LVO Stroke.

Figure 1

Open in a new tab

(A) Distribution of poor outcome across quartiles of perfusion lesion volume; (B) receiver operating characteristic curve of poor outcome predicted by perfusion lesion volume; the area under the curve is 0.71, with 95% CI of 0.60–0.83; the optimal cutpoint is 65 mL, with a sensitivity of 59% and a specificity of 77%; (C) distribution of poor outcome across quartiles of penumbra volume; and (D) receiver operating characteristic curve of poor outcome predicted by penumbra volume; the area under the curve is 0.71, with 95% CI of 0.60–0.83; the optimal cutpoint is 48 mL, with a sensitivity of 67% and a specificity of 70%. CTP = CT perfusion; LVO = large vessel occlusion.

In addition to CTP markers, univariable regression models showed a significant predictive power of poor outcome by the following variables: occlusion site (crude OR = 0.26 [0.09–0.80] for M1 vs ICA, crude OR = 0.22 [0.07–0.68] for M2 vs ICA), good collaterals (crude OR = 0.49 [0.28–0.88]), and baseline NIHSS (crude OR = 1.84 [1.25–2.73]). Distal occlusion site (M2 vs M1 vs ICA) and good collaterals correlated with perfusion lesion volume (correlation coefficient = 0.36 and 0.25, respectively, p = 0.003 and p <0.001), but their predictive powers of poor outcome were slightly lower than perfusion lesion volume (AUC of 0.63 [0.51–0.74] for occlusion site, 0.63 [0.52–0.74] for collaterals, and 0.71 [0.60–0.83] for perfusion lesion volume). Baseline NIHSS and perfusion lesion volume had a similar predictive power of poor outcome (AUC of 0.72 vs 0.71), but the 2 variables showed no significant correlation (correlation coefficient = 0.06, p = 0.511).

Predictors of Poor Outcome in Multivariable Regression Models

The predictive power of perfusion lesion volume, penumbra volume, and severe hypoperfusion volume was then tested in 3 multivariable regression models, respectively. The 3 CTP markers were tested in separate models because of the consideration of multicollinearity of these variables. In addition to CTP markers, the multivariable regression models included the following patient characteristics: occlusion site, good collaterals, baseline NIHSS, age, IVT, and onset to scan time. Although age (crude OR = 1.03 [0.99–1.06]), IVT (crude OR = 0.74 [0.31–1.78]), and onset to scan time (crude OR = 1.05 [0.91–1.21]) showed no significative predictive power in poor outcome in univariable regression, they were included in the multivariable regression models for their clinical significance.

In the multivariable regression model, perfusion lesion volume remained a significant predictor (adjusted OR = 1.01 [1.01–1.02], p = 0.039, model 1, Table 2). Increase of perfusion lesion volume by each 1 mL resulted in increasing the odds of poor outcome by 1% after adjusting for occlusion site, good collaterals, acute NIHSS, age, IVT, and onset to scan time. Penumbra volume had similar performance in predicting poor outcome in the multivariable regression model (adjusted OR = 1.01 [1.01–1.03], p = 0.043, Model 2, Table 2). However, severe hypoperfusion lesion volume showed no significance in predicting poor outcome in the multivariable regression model (adjusted OR = 1.01 [0.99–1.04], p = 0.190).

Table 2.

Multivariable Regression Models of Predicting 3-Month Poor Outcome of Patients With Mild LVO Stroke

graphic file with name WNL-2022-201360t2.jpg

Open in a new tab

In the multivariable regression models (model 1 and model 2, Table 2), in addition to perfusion lesion or penumbra volume, age and baseline NIHSS also showed a significant predictive power in poor outcome. Occlusion site and good collaterals lost its significance in predicting poor outcome in the multivariable regression models (model 1 and model 2, Table 2). IVT and onset to scan time remained nonsignificant in predicting poor outcome in the multivariable regression models (model 1 and model 2, Table 2).

Large Perfusion Lesion ≥65 mL Predicting Poor Outcome

Analysis of the ROC curve predicting poor outcome yielded an AUC of 0.71 (95% CI [0.60–0.83], Figure 1B) and a threshold of 65 mL as the optimum cutpoint for perfusion lesion to predict poor outcome. This cutpoint demonstrated a sensitivity of 59% and a specificity of 77%. Within the patient cohort, 42 patients had perfusion lesions ≥65 mL, 38% (16/42) of whom resulted in poor outcome compared with 11% of patients (11/97) who had perfusion lesion <65 mL. In univariable logistic regression, perfusion lesion ≥65 mL was associated with a greater-than 4.8-fold increase in the odds of poor outcome compared with perfusion lesion <65 mL (crude OR = 4.8 [1.99–11.64], p < 0.001). Patients with perfusion lesion ≥65 mL, compared with those with perfusion <65 mL, also shifted the ordinal functional outcome toward a higher disability scale (Figure 2A).

Figure 2. Ordinal Distribution of the 3-Month mRS for Patients With Mild LVO Stroke.

Figure 2

Open in a new tab

(A) Perfusion lesion volume <65 mL vs =65 mL and (B) penumbra volume <48 mL vs =48 mL. LVO = large vessel occlusion; mRS = modified Rankin Scale.

All patients in this study had hemorrhagic transformation assessed on follow-up scan. Of the 139 patients included in this study, only 3 patients (2%) had PH, and 2 patients (1%) had sICH. The rate of PH and sICH showed no difference in patients with baseline perfusion lesion ≥65 mL, compared with patients with perfusion lesion <65 mL (Table 3).

Table 3.

Patient Outcome of Mild LVO Stroke: Perfusion Lesion <65 vs ≥65 mL

graphic file with name WNL-2022-201360t3.jpg

Open in a new tab

Large Perfusion Lesion ≥65 mL Predicting Early Neurologic Deterioration

A total of 129 (of 139) patients within the cohort had 24-hour NIHSS recorded. Of the 129 patients, 15 patients (12%) experienced early neurologic deterioration. For the 15 patients with early neurologic deterioration, 53% (8/15) resulted in 3-month poor outcome, whereas only 14% (16/114) ended with poor outcome for patients without early neurologic deterioration. Early neurologic deterioration was closely correlated with poor outcome (p < 0.001).

Analysis of the ROC curve predicting early neurologic deterioration yielded an AUC of 0.78 (95% CI = [0.65–0.91], Figure 3A). It resulted an optimal cutpoint of 65 mL in perfusion lesion as well, with a sensitivity of 80% and a specificity of 75%. Of the 129 patients, 40 patients had baseline perfusion lesions ≥65 mL, 33% (13/40) of whom experienced early neurologic deterioration, whereas 97 patients had perfusion lesion volumes <65 mL, of whom only 3% (3/97) had early neurologic deterioration (p < 0.001, Table 3).

Figure 3. Receiver Operating Characteristic Curve of Early Neurologic Deterioration Predicted by Perfusion Lesion Volume (A) and Penumbra Volume (B).

Figure 3

Open in a new tab

Large Penumbra ≥48 mL Predicted Poor Outcome and Early Neurologic Deterioration

The performance of penumbra was similar to that of perfusion lesion in terms of predicting poor outcome (Figure 1, C and D, and Figure 2B) or predicting neurologic deterioration (Figure 3B). AUC analysis showed that the optimal threshold for penumbra was 48 mL in predicting poor outcome, with a sensitivity of 67% and a specificity of 70%. When stratifying patients with the penumbra threshold, 87 of 139 patients had penumbra volume <48 mL, and 52 of 139 patients had penumbra volume ≥48. The poor outcome rate was significantly higher in patients with penumbra ≥48 mL vs penumbra <48 mL (35% vs 10%, p < 0.001, Table 3). Regarding early neurologic deterioration, the rate was 24% for penumbra ≥48 mL and was only 4% for penumbra <48 mL. Penumbra ≥48 mL, compared with penumbra <48 mL, increased the odds of poor outcome by 4.6 folds (crude OR = 4.6 [1.87–11.23], p = 0.001). The large and small penumbra groups showed no difference in PH (2% vs 2%) or sICH (2% vs 1%, Table 3).

Sensitivity Analyses

When redefining poor outcome by mRS 2–6, 31% (43/139) of patients had poor outcome. Perfusion lesion still predicted the poor outcome significantly in either the univariable regression (crude OR = 1.01 [1.01–1.02], p = 0.003) or the multivariable regression model (adjusted OR = 1.01 [1.01–1.02], p = 0.031), adjusting for occlusion site, good collaterals, acute NIHSS, age, IVT, and onset to scan time. Perfusion lesion showed slightly lower predictive power in predicting the poor outcome defined by mRS 2–6 compared with that defined by mRS 3–6 (AUC of 0.63 [0.52–0.73] vs 0.71 [0.60–0.83]). When dichotomizing patients with perfusion lesion of 65 mL, the mRS 2–6 poor outcome was significantly higher for patients with large perfusion lesion of ≥65 mL than patients with perfusion lesion <65 mL (48% vs 24%, p < 0.005).

When redefining perfusion lesion with Tmax > 6 s, 4 of the 139 patients failed to generate Tmax perfusion lesion. For the remaining 135 patients, the perfusion lesion volume showed no significant difference to that measured by DT > 3 s (median 48 vs 50 mL, p = 0.214). Perfusion lesion measured by Tmax > 6 s had high concordance to that measured by DT > 3 s (concordance correlation coefficient = 0.87 [0.83–0.91], Figure 4A). Perfusion lesion measured by Tmax > 6 s and DT > 3 s showed a similar predictive power in poor outcome (mRS 3–6, Figure 4B). When dichotomizing patients with Tmax perfusion lesion of 65 mL, patients with large perfusion lesion of ≥65 mL resulted in a significantly higher poor outcome of mRS 3–6 than patients with perfusion lesion <65 mL (32% vs 14%, p = 0.011).

Figure 4. CTP Lesion Measured by DT > 3 s vs Tmax > 6 s.

Figure 4

Open in a new tab

(A) Concordance of perfusion lesion measured by DT > 3 s vs Tmax > 6 s on CTP and (B) receiver operating characteristic curve of DT > 3 s vs Tmax > 6 s in predicting 3-month poor outcome. CTP = CT perfusion; DT = delay time.

Discussion

This retrospective multisite cohort study indicates that total perfusion lesion volume and penumbra volume at baseline are strong predictors of poor prognosis of patients with mild LVO stroke. A perfusion lesion of ≥65 mL was associated with a greater-than 4.8-fold increase in poor functional outcome. A penumbra volume of ≥48 mL was associated with a greater-than 4.6-fold increase in poor functional outcome.

The relatively poor outcome of patients with large perfusion lesion might be explained by early neurologic deterioration. According to this study, patients with mild LVO stroke often have very small core volume (median volume of 6.0 mL). This is supported by previous studies (median core volume of 11 mL).27 If patients with mild LVO stroke presented with large perfusion lesion, they most likely would have large penumbra and small core. This is the group of patients who could potentially benefit from endovascular treatment that salvages the penumbra region with immediate restoration of blood flow. The effective blood restoration of endovascular treatment could also potentially reduce cerebral edema for patients with large perfusion lesion.28 The edema around the large perfusion lesion can be space occupying, leading to a gradual increase in intracranial pressure and further reductions in cerebral perfusion pressure. This could cause benign oligemic tissue to be recruited into the penumbra and eventually into the infarcted core.29,30 The demise of oligemia, without prompt blood restoration, might explain the early neurologic deterioration of patients with large perfusion lesion. In summary, findings of this study indicate that early neurologic deterioration is highly correlated with the 3-month poor outcome of patients with mild stroke. Therefore, detecting patients who might experience early neurologic deterioration is the key for acute mild stroke management. This study helps to screen patients who are likely to experience early neurologic deterioration with the CTP brain imaging markers.

The main clinical implication of this study is to select patients with mild LVO stroke for endovascular treatment. EVT with prompt restoration of blood flow could potentially prevent early neurologic deterioration of patients with mild LVO stroke and improve their functional outcome. However, previous mild LVO stroke studies showed limited efficacy of EVT in improving patients' outcome compared with best medical management.7,8 A recent meta-analysis of mild LVO stroke studies12 reported that there was no benefit of EVT over best medical management overall. One possible explanation is that no CTP was used in patient selection in the previous studies. This study indicates that if patients with mild LVO stroke had perfusion lesion <65 or penumbra lesion <48 mL, they would have very limited benefit from the EVT because 89% of patients had good outcome without EVT, meaning that it would be nearly impossible to show a benefit of EVT on top of that. For patients with mild LVO stroke, the targeted group for EVT treatment might be those with perfusion lesion ≥65 mL or penumbra lesion ≥48 mL. There could perhaps be more potential to improve on the good outcome rate in this group, as the good outcome rate was 62%. A well-designed randomized controlled trial using perfusion imaging selection is now needed.

This study identifies a biological marker that can predict patients with mild LVO stroke at higher risk of poor functional outcome and early neurologic deterioration. Previous studies attempting to find predictive markers of patients' outcome of mild LVO stroke have had little or no success. Only in recent 2 studies,27,31 CTP variables were identified as promising predictors. Findings of this study are supported by the previous study,27 reporting that patients with early neurologic deterioration had significant bigger perfusion lesion (93.5 vs 52.1 mL). Furthermore, findings of this study are also in line with the other recent study31 reporting a significant predictive power of perfusion variables for early neurologic deterioration. Compared with the previous 2 studies,27,31 this study confirms the predictive power of perfusion lesion on long-term outcome of patients with mild LVO stroke. In addition, this study has the advantage of larger sample size and adjusting for other baseline imaging variables, including collateral flow, ASPECTS, and occlusion sites. Another novelty of this study is using a well-validated definition of at-risk tissue (DT > 3 s),18-21 with additional delay and dispersion correction, meaning that we are less likely to have overestimated perfusion lesions. However, we do acknowledge the widely use of Tmax > 6 s in defining perfusion lesion as in the previous 2 mild stroke studies.27,31 Thus, we further validate our findings with the Tmax > 6 s definition in this study. This helps to generalize findings of this study across CTP algorithms and software.

This study indicates that CTP markers are stronger outcome predictors than angiographic markers of occlusion site or collateral grade. Proximal occlusion site has been reported to be a predictor of early neurologic deterioration of patients with mild stroke,15 and patients with mild stroke due to proximal occlusion site have been reported to benefit from endovascular treatment in recent studies.16,17 Collateral status has been considered as a major determinator of stroke outcome,32 and it has been reported to correlate with the prognosis of patients with mild stroke.14 In this study, we have confirmed the predictive power of occlusion sites and collateral grade in univariable regression models. However, when adjusting for perfusion lesion volume or penumbra lesion volume, occlusion sites and collateral grade lost its significance. One possible explanation is that occlusion site and collaterals affect patient outcome through perfusion lesion. Patients with either proximal occlusion or poor collaterals would result in larger perfusion lesion and thus lead to poor outcome of patients with mild LVO stroke. Thus, findings of this study support the utilization of CTP in mild stroke management.

This study has some limitations. First, this study used data from the INSPIRE that only recruits patients with acute CTP. This explains the relatively low patient recruitment rate (about 2.5 mild strokes a year per center). In some INSPIRE sites, patients were scanned with CTP when they were considered for reperfusion treatment. Such selection bias explains the high IVT rate in this cohort. Because of the selection bias, findings of this study are more applicable to patients with mild LVO stroke who do not have clinical contradictions for reperfusion treatment. Second, this is not a randomized controlled study. Perfusion lesion and penumbra size were adjusted for occlusion site, collateral status, acute stroke severity, age, IVT rate, and onset to scan time when assessing their predictive power of outcome in this study. However, residual confounding factors probably persist. This study did not assess the confounding effect of onset to needle time, because of the lack of such information in patients who did not receive IVT. For patients who did receive IVT, there was a trend toward delayed treatment in patients with poor outcome compared with good outcome. The trend was not significant, possibly because of a small number of patients with poor outcome. Further studies are required to investigate the role of IVT time in patients with mild stroke. Third, the CTP lesion thresholds derived from this study need to be tested in mild stroke studies that compare efficacy and safety of direct EVT vs best medical management. This study showed very low rate of bleeding rate (2%–3%) for patients with large perfusion lesion. However, the bleeding rate might be increased with EVT, and this might comprise the benefit of EVT treatment for patients with perfusion lesion ≥65 mL or penumbra ≥48 mL. Fourth, this study had recanalization status assessed at 24–72 hours. This explains why this study had such high recanalization rate with best medical management, and why the high recanalization rate did not predict good outcome. According to a previous meta-analysis,33 the spontaneous recanalization rate or recanalization from IVT can reach to >50% after 24 hours. Late spontaneous recanalization is the common natural outcome of vessel occlusions. However, only early recanalization while penumbral tissues persist to be salvaged has a substantial impact on outcome. This further emphasizes the importance of EVT with prompt restoration of blood flow for patients with mild stroke. Fifth, this study used mRS 0–2 to define good outcome, because of the consideration that this cohort had LVO and large baseline perfusion lesion. Because the mRS of 0–1 has been widely used in defining the outcome of mild stroke, findings of this study were further validated in a sensitivity analysis of good outcome of mRS 0–1 vs poor outcome of 2–6. Sixth, findings of this study are limited to patients with anterior circulation stroke. The CTP lesion thresholds derived from this study might be different for patients with posterior circulation stroke. In addition, the CTP lesion thresholds derived from this study had relatively low sensitivity in predicting poor outcome. Further research on the optimal perfusion or penumbral lesion is needed. Finally, this study had follow-up NIHSS assessed at 24–72 hours after stroke onset, rather than within the 24 hours of stroke onset. This might affect the efficacy of capturing early neurologic deterioration in this study.

In conclusion, total perfusion lesion volume ≥65 mL or penumbra volume ≥48 mL is associated with poor functional outcome in patients with mild stroke symptoms due to LVO. The benefits and risks of EVT in patients with mild LVO stroke should be tested in a randomized clinical trial, and such trials should involve the use of perfusion imaging for patient selection (larger lesions alone), or, at the very least, randomization should be stratified by baseline perfusion lesion.

Glossary

ASPECTS

Alberta Stroke Program Early CT Score

CBF

cerebral blood flow

CTP

CT perfusion

DT

delay time

EVT

endovascular thrombectomy

INSPIRE

International Stroke Perfusion Imaging Registry

IVT

IV thrombolysis

LVO

large vessel occlusion

NCCT

noncontrast CT

NIHSS

NIH Stroke Scale

PH

parenchymal hemorrhage

sICH

symptomatic intracerebral hemorrhage

Appendix 1. Authors

Appendix 1.

Open in a new tab

Appendix 2. Coinvestigators

Appendix 2.

Open in a new tab

Footnotes

CME Course: NPub.org/cmelist

Study Funding

The authors report no targeted funding.

Disclosure

Dr. P. Wang, Dr. W. Chen, Dr. C. Chen, Dr. A. Bivard, and Dr. Y. Geng report no disclosures. Dr. M.W. Parson reports grant from the National Health and Medical Research Council of Australia; Dr. M.W. Parson reports research partnership with Siemens, Canon, and Apollo Medical Imaging, outside the submitted work. Dr. L. Lin reports grant from BioMedTech Horizons program, the Medical Research Future Fund, Australia. Go to Neurology.org/N for full disclosures.

References

  • 1.Albers GW, Marks MP, Kemp S, et al. Thrombectomy for stroke at 6 to 16 hours with selection by perfusion imaging. N Engl J Med. 2018;378(8):708-718. doi: 10.1056/NEJMoa1713973 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Nogueira RG, Jadhav AP, Haussen DC, et al. Thrombectomy 6 to 24 hours after stroke with a mismatch between deficit and infarct. N Engl J Med. 2018;378(1):11-21. doi: 10.1056/NEJMoa1706442 [DOI] [PubMed] [Google Scholar]
  • 3.Heldner MR, Zubler C, Mattle HP, et al. National Institutes of Health Stroke Scale score and vessel occlusion in 2152 patients with acute ischemic stroke. Stroke. 2013;44(4):1153-1157. doi: 10.1161/STROKEAHA.111.000604 [DOI] [PubMed] [Google Scholar]
  • 4.Haussen DC, Bouslama M, Grossberg JA, et al. Too good to intervene? Thrombectomy for large vessel occlusion strokes with minimal symptoms: an intention-to-treat analysis. J Neurointerv Surg. 2017;9(10):917-921. doi: 10.1136/neurintsurg-2016-012633 [DOI] [PubMed] [Google Scholar]
  • 5.Haussen DC, Lima FO, Bouslama M, et al. Thrombectomy versus medical management for large vessel occlusion strokes with minimal symptoms: an analysis from STOPStroke and GESTOR cohorts. J NeuroInterventional Surg. 2018;10(4):325-329. doi: 10.1136/neurintsurg-2017-013243 [DOI] [PubMed] [Google Scholar]
  • 6.Nagel S, Bouslama M, Krause LU, et al. Mechanical thrombectomy in patients with milder strokes and large vessel occlusions. Stroke. 2018;49(10):2391-2397. doi: 10.1161/strokeaha.118.021106 [DOI] [PubMed] [Google Scholar]
  • 7.Dargazanli C, Arquizan C, Gory B, et al. Mechanical thrombectomy for minor and mild stroke patients harboring large vessel occlusion in the anterior circulation: a multicenter cohort study. Stroke. 2017;48(12):3274-3281. doi: 10.1161/strokeaha.117.018113 [DOI] [PubMed] [Google Scholar]
  • 8.Manno C, Disanto G, Bianco G, et al. Outcome of endovascular therapy in stroke with large vessel occlusion and mild symptoms. Neurology 2019;93(17):e1618–e1626. doi: 10.1212/wnl.0000000000008362 [DOI] [PubMed] [Google Scholar]
  • 9.Urra X, San Román L, Gil F, et al. Medical and endovascular treatment of patients with large vessel occlusion presenting with mild symptoms: an observational multicenter study. Cerebrovasc Dis. 2014;38(6):418-424. doi: 10.1159/000369121 [DOI] [PubMed] [Google Scholar]
  • 10.Sarraj A, Hassan A, Savitz SI, et al. Endovascular thrombectomy for mild strokes: how low should we go?. Stroke 2018;49(10):2398-2405. doi: 10.1161/STROKEAHA.118.022114 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Wolman DN, Marcellus DG, Lansberg MG, et al. Endovascular versus medical therapy for large-vessel anterior occlusive stroke presenting with mild symptoms. Int J Stroke. 2020;15(3):324-331. doi: 10.1177/1747493019873510 [DOI] [PubMed] [Google Scholar]
  • 12.Goyal N, Tsivgoulis G, Malhotra K, et al. Medical management vs mechanical thrombectomy for mild strokes: an international multicenter study and systematic review and meta-analysis. JAMA Neurol. 2020;77(1):16-24. doi: 10.1001/jamaneurol.2019.3112 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Saleem Y, Nogueira RG, Rodrigues GM, et al. Acute neurological deterioration in large vessel occlusions and mild symptoms managed medically. Stroke. 2020;51(5):1428-1434. doi: 10.1161/strokeaha.119.027011 [DOI] [PubMed] [Google Scholar]
  • 14.Kimmel ER, Al Kasab S, Harvey JB, et al. Absence of collaterals is associated with larger infarct volume and worse outcome in patients with large vessel occlusion and mild symptoms. J Stroke Cerebrovasc Dis. 2019;28(7):1987-1992. doi: 10.1016/j.jstrokecerebrovasdis.2019.03.032 [DOI] [PubMed] [Google Scholar]
  • 15.Seners P, Ben Hassen W, Lapergue B, et al. Prediction of early neurological deterioration in individuals with minor stroke and large vessel occlusion intended for intravenous thrombolysis alone. JAMA Neurol. 2021;78(3):321-328. doi: 10.1001/jamaneurol.2020.4557 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Seners P, Perrin C, Lapergue B, et al. Bridging therapy or IV thrombolysis in minor stroke with large vessel occlusion. Ann Neurol. 2020;88(1):160-169. doi: 10.1002/ana.25756 [DOI] [PubMed] [Google Scholar]
  • 17.Lin CH, Saver JL, Ovbiagele B, Tang SC, Lee M, Liebeskind DS. Effects of endovascular therapy for mild stroke due to proximal or M2 occlusions: meta-analysis. J Neurointerv Surg. Published Online March 15, 2022. doi: 10.1136/neurintsurg-2022-018662 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Lin L, Bivard A, Kleinig T, et al. Correction for delay and dispersion results in more accurate cerebral blood flow ischemic core measurement in acute stroke. Stroke. 2018;49(4):924-930. doi: 10.1161/STROKEAHA.117.019562 [DOI] [PubMed] [Google Scholar]
  • 19.Lin L, Bivard A, Krishnamurthy V, Levi CR, Parsons MW. Whole-brain CT perfusion to quantify acute ischemic penumbra and core. Radiology. 2016;279(3):876-887. doi: 10.1148/radiol.2015150319 [DOI] [PubMed] [Google Scholar]
  • 20.Chen C, Bivard A, Lin L, Levi CR, Spratt NJ, Parsons MW. Thresholds for infarction vary between gray matter and white matter in acute ischemic stroke: a CT perfusion study. J Cereb Blood Flow Metab. 2019;39(3):536-546. doi: 10.1177/0271678x17744453 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Bivard A, Levi C, Krishnamurthy V, et al. Perfusion computed tomography to assist decision making for stroke thrombolysis. Brain. 2015;138(Pt 7):1919-1931. doi: 10.1093/brain/awv071 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Miteff F, Levi CR, Bateman GA, Spratt N, McElduff P, Parsons MW. The independent predictive utility of computed tomography angiographic collateral status in acute ischaemic stroke. Brain 2009;132(Pt 8):2231-2238. doi: 10.1093/brain/awp155 [DOI] [PubMed] [Google Scholar]
  • 23.Adams HP Jr., Bendixen BH, Kappelle LJ, et al. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke. 1993;24(1):35-41. doi: 10.1161/01.str.24.1.35 [DOI] [PubMed] [Google Scholar]
  • 24.Hacke W, Kaste M, Fieschi C, et al. Randomised double-blind placebo-controlled trial of thrombolytic therapy with intravenous alteplase in acute ischaemic stroke (ECASS II). Second European-Australasian Acute Stroke Study Investigators. Lancet (London, England). 1998;352(9136):1245-1251. doi: 10.1016/s0140-6736(98)08020-9 [DOI] [PubMed] [Google Scholar]
  • 25.Dargazanli C, Fahed R, Blanc R, et al. Modified thrombolysis in cerebral infarction 2C/thrombolysis in cerebral infarction 3 reperfusion should Be the aim of mechanical thrombectomy: insights from the ASTER trial (contact aspiration versus stent retriever for successful revascularization). Stroke. 2018;49(5):1189-1196. doi: 10.1161/strokeaha.118.020700 [DOI] [PubMed] [Google Scholar]
  • 26.Goyal M, Fargen KM, Turk AS, et al. 2C or not 2C: defining an improved revascularization grading scale and the need for standardization of angiography outcomes in stroke trials. J Neurointerv Surg. 2014;6(2):83-86. doi: 10.1136/neurintsurg-2013-01066527 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Lee VH, Thakur G, Nimjee SM, et al. Early neurologic decline in acute ischemic stroke patients receiving thrombolysis with large vessel occlusion and mild deficits. J Neurointerv Surg. 2020;12(11):1085-1087. doi: 10.1136/neurintsurg-2020-015871 [DOI] [PubMed] [Google Scholar]
  • 28.Thorén M, Dixit A, Escudero-Martínez I, et al. Effect of recanalization on cerebral edema in ischemic stroke treated with thrombolysis and/or endovascular therapy. Stroke. 2020;51(1):216-223. doi: 10.1161/strokeaha.119.026692 [DOI] [PubMed] [Google Scholar]
  • 29.Alawneh JA, Moustafa RR, Baron JC. Hemodynamic factors and perfusion abnormalities in early neurological deterioration. Stroke. 2009;40(6):e443-e450. doi: 10.1161/strokeaha.108.532465 [DOI] [PubMed] [Google Scholar]
  • 30.Tisserand M, Seners P, Turc G, et al. Mechanisms of unexplained neurological deterioration after intravenous thrombolysis. Stroke. 2014;45(12):3527-3534. doi: 10.1161/strokeaha.114.006745 [DOI] [PubMed] [Google Scholar]
  • 31.Gwak DS, Kwon JA, Shim DH, Kim YW, Hwang YH. Perfusion and diffusion variables predict early neurological deterioration in minor stroke and large vessel occlusion. J stroke. 2021;23(1):61-68. doi: 10.5853/jos.2020.01466 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Liebeskind DS. Collateral circulation. Stroke. 2003;34(9):2279-2284. doi: 10.1161/01.Str.0000086465.41263.06 [DOI] [PubMed] [Google Scholar]
  • 33.Rha JH, Saver JL. The impact of recanalization on ischemic stroke outcome: a meta-analysis. Stroke. 2007;38(3):967-973. doi: 10.1161/01.Str.0000258112.14918.24 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

Anonymized data not published within this article will be made available by request from qualified investigators whose proposal of data use has been approved by an independent review committee.

Articles from Neurology are provided here courtesy of American Academy of Neurology

RESOURCES