Abstract
BACKGROUND
The Alberta Stroke Program Early Computed Tomography Score and core volume on preoperative imaging are key predictors of clinical outcomes following endovascular therapy in patients with a large ischemic core. Although the corticospinal tract is essential for motor function, its prognostic impact in patients with a large ischemic core remains unclear.
METHODS
This multicenter retrospective study analyzed preoperative imaging data from patients with Alberta Stroke Program Early Computed Tomography Score ≤5 who underwent endovascular therapy. The presence of lesions in the posterior corona radiata and lesions in the primary motor cortex was assessed. A good outcome was defined as a modified Rankin Scale score ≤3 at 90 days. The association between lesions in the posterior corona radiata, lesions in the primary motor cortex, and good outcome was analyzed using univariable and stepwise multivariable logistic regression, with variable selection based on Akaike information criterion corrected for small sample size.
RESULTS
Among 107 patients, 37 (34.6%) achieved a good outcome. In univariable analysis, neither lesions in the posterior corona radiata nor core volume was significantly associated with a good outcome. In stepwise multivariable logistic regression, modified Rankin Scale score before onset (odds ratio, 0.30 [95% CI, 0.10–0.73]), cardioembolism (odds ratio, 0.25 [95% CI, 0.08–0.76]), absence of lesions in the primary motor cortex involvement (odds ratio, 13.49 [95% CI, 3.75–63.45]), and shorter onset‐to‐reperfusion time (odds ratio, 0.996 [95% CI, 0.992–0.998]) were independent predictors. Alberta Stroke Program Early Computed Tomography Score and the absence of multiple artery occlusion were retained in the final model but were not statistically significant.
CONCLUSION
Absence of lesions in the primary motor cortex involvement was independently associated with good outcome after endovascular therapy in patients with large ischemic core, suggesting its potential utility as a complementary imaging marker in this population.
Keywords: ASPECTS, corticospinal tract, endovascular therapy, large ischemic core
Several randomized controlled trials have demonstrated the efficacy of endovascular therapy (EVT) for large vessel occlusion in patients with a large ischemic core (LIC), which is based on an Alberta Stroke Program Early Computed Tomography (CT) Score (ASPECTS) of ≤5. 1 , 2 , 3 , 4 , 5 However, the lower ASPECTS and upper core volume thresholds at which EVT remains effective remain unknown.
A secondary analysis of the RESCUE‐Japan LIMIT (Recovery by Endovascular Salvage for Cerebral Ultra‐Acute Embolism–Japan Large Ischemic Core Trial), which included patients with ASPECTS of 3–5, suggested that EVT was not beneficial for patients with ASPECTS ≤3. 6 Similarly, another study indicated that patients with a core volume ≥128 mL might not derive benefit from EVT. 7 In contrast, the LASTE (Large Stroke Therapy Evaluation) trial, 5 which included patients with ASPECTS of 0–5, demonstrated the efficacy of EVT even in patients with ASPECTS of 0–2 and a core volume >150 mL. Additionally, a subgroup analysis of SELECT2: A Randomized Controlled Trial to Optimize Patient's Selection for Endovascular Treatment in Acute Ischemic Stroke, which included patients with ASPECTS of 3–5, reported EVT efficacy even in patients with a core volume >150 mL. 2
These discrepancies likely reflect the limitations of severity assessment based on the ASPECTS and core volume. Although these metrics serve as simplified and objective measures of stroke severity, they do not sufficiently account for the complex functional organization of the human brain. Notably, most clinical trials on EVT for acute ischemic stroke evaluated patient functional outcomes using the modified Rankin Scale (mRS), which predominantly assesses patient motor function. 8 Therefore, the involvement of brain regions related to motor function would critically affect study outcomes. A previous study suggested that the degree of recanalization of cerebral vessels perfusing the primary motor cortex after EVT is a better predictor of prognosis than the Thrombolysis in Cerebral Infarction (TICI) classification, which is a purely volumetric reperfusion assessment tool. 9 However, the current eligibility criteria for EVT primarily depend on ASPECTS or infarct volume before treatment, which does not differentiate eloquent and noneloquent regions. This could lead to overestimation or underestimation of EVT benefits in terms of functional outcomes.
In clinical practice, particularly in patients with LIC where the salvageable penumbra is relatively limited, preoperative assessment of whether the infarcted area affects motor function‐critical regions may have a direct impact on functional prognosis and EVT decision‐making. However, the region most critical for such motor functional prognosis, independent of ASPECTS regions, has not been sufficiently investigated.
Given these considerations, this study aimed to evaluate whether preoperatively identified ischemic lesions in functionally critical motor‐related regions, specifically the posterior corona radiata and primary motor cortex, 10 , 11 , 12 , 13 could serve as important predictors of clinical outcomes following EVT in patients with LIC.
Nonstandard Abbreviations and Acronyms
- ASPECTS
Alberta Stroke Program Early Computed Tomography Score
- DWI
diffusion‐weighted imaging
- EVT
endovascular therapy
- LIC
large ischemic core
- LPCR
lesions in the posterior corona radiata
- LPMC
lesions in the primary motor cortex
- mRS
modified Rankin Scale
- mTICI
modified Thrombolysis in Cerebral Infarction
- NIHSS
National Institutes of Health Stroke Scale
CLINICAL PERSPECTIVE
What Is New?
This study identified that the absence of lesions in the primary motor cortex on baseline imaging is independently associated with favorable 90‐day functional outcomes in large ischemic core patients undergoing endovascular therapy, providing prognostic information not captured by Alberta Stroke Program Early Computed Tomography Score alone.
What Are the Clinical Implications?
Incorporating the evaluation of primary motor cortex involvement into pretreatment imaging. assessment may enhance prognostic accuracy and support individualized treatment planning in patients with large ischemic core stroke.
METHODS
The data that support the findings of this study are available from the corresponding author on reasonable request.
Study Population
This retrospective analysis included only Japanese patients with acute ischemic stroke who underwent EVT, using prospectively collected databases from 2 hospitals in Japan. Data collection began in 2016 and 2020 at each hospital, respectively, and included data up to March 2024. The inclusion criteria were as follows: (1) occlusion of the internal carotid artery or the M1 or M2 segment of the middle cerebral artery confirmed on CT or magnetic resonance angiography before EVT; (2) an ASPECTS of ≤5 on CT or diffusion‐weighted magnetic resonance imaging (DWI‐MRI); (3) an mRS score of ≤2 before onset; and (4) achieved successful reperfusion, defined as modified Thrombolysis in Cerebral Infarction (mTICI) ≥2b. Patients with bilateral occlusions were excluded from this study. EVT eligibility was determined by the attending physician at each facility. The study design and analysis were approved by the Ethics Committee of the Japanese Kumamoto Red Cross Hospital (approval number: 666). Informed consent was obtained through an opt‐out process at each facility. Information regarding the study was made publicly available, allowing patients to decline participation if desired. This study was conducted and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology guidelines. 14
Data Collection
The following patient clinical data were collected: age, sex, vascular risk factors, atrial fibrillation, premorbid mRS score, National Institutes of Health Stroke Scale (NIHSS) score on admission, stroke classification, diagnostic imaging findings, use of intravenous thrombolysis, and onset‐to‐reperfusion time. The ASPECTS on admission diagnostic imaging was determined using CT or DWI, as assessed by the attending physician. Occluded vessels were also identified by the attending physician based on CT or magnetic resonance angiography before EVT. The ischemic core was assessed only in cases where MRI was performed. The maximal visual extent of the DWI (b1000 image) positive lesion was drawn manually using careful windowing adjustments. To avoid a T2 shine‐through, the apparent diffusion coefficient image was used to determine the area. The ischemic core volume was calculated by multiplying the area by the slice thickness (5 mm) for each slice. Collateral status was evaluated using the American Society of Intervention and Therapeutic Neuroradiology/Society of Interventional Radiology grading system on angiography prior to EVT. 15 Recanalization results were evaluated by the attending neurointerventionalist using the mTICI grade. 16 Symptomatic and any intracranial hemorrhage were evaluated using a CT performed within 24 hours of the EVT procedure. Symptomatic intracranial hemorrhage was defined as a ≥4‐point worsening in the NIHSS Score. 17 Clinical outcomes were assessed using a mailed mRS questionnaire 90 days after EVT. An mRS score of 0–3 at 90 days was considered indicative of a good outcome. 1
Evaluation of Lesions in the Posterior Corona Radiata and Primary Motor Cortex
The presence of lesions in the posterior corona radiata (LPCR) and primary motor cortex (LPMC) was evaluated using pretreatment DWI or CT images, if an MRI was not performed. The evaluation was independently conducted by 3 specialists certified by the Japanese Society of Neurology and the Japanese Society for Neuroendovascular Therapy (S.N., K.K., and Y.N.). The evaluators were blinded to all clinical data. Prior to the formal image assessment, the 3 raters conducted a training session using approximately 10 cases that were not included in the study cohort, in order to standardize the evaluation criteria for LPCR and LPMC. S.N. (Evaluator 1) assessed all imaging data, whereas K.K. and Y.N. (Evaluators 2) evaluated imaging data from their respective institutions. LPCR and LPMC were considered positive only if they were confluent. Small spot‐like lesions were considered negative. To evaluate LPCR, horizontal lines (lines A and B) were drawn to divide the lateral ventricles into 3 equal sections (Figure 1A) using the anterior and posterior horns as reference points. Lesions intersecting line B were considered as LPCR. 12 The presence or absence of LPMC was assessed on axial slices above the level where the lateral ventricle disappears, based on lesion involvement of the primary motor cortex (Figure 1B). Interobserver reliability for identifying LPCR and LPMC was assessed using the κ‐coefficient to measure interobserver agreement between Evaluator 1 and Evaluator 2. Discrepancies were resolved through consensus discussion.
Figure 1.
Imaging‐based evaluation of LPCR and LPMC. A, Evaluation of LPCR. (a) LPCR‐negative case. (b, c) LPCR‐positive cases with lesions interesting line B. B, Evaluation of LPMC. The area enclosed by the dotted line is evaluated as LPMC‐positive. LPCR indicates lesions in the posterior corona radiata; and LPMC, lesions in the primary motor cortex.
Statistical Analysis
Statistical analyses were primarily performed using JMP 18.1.1 (SAS Institute, Inc, Cary, NC, USA), and statistical significance was set at P<0.05. The patient characteristics were compared between the good outcome group and the poor outcome group. Continuous variables are reported as mean±SD or median with interquartile range. Categorical variables are expressed as frequencies and percentages. Comparisons of continuous variables were performed using the Student's t‐test or the Wilcoxon rank‐sum test, depending on the data distribution. Categorical variables were examined using the χ 2 test or Fisher's exact test, as appropriate. To examine the correlation between ASPECTS and the manually measured core volume, we analyzed the median core volume for each ASPECTS. Multivariable logistic regression was performed to identify independent predictors of a good outcome. Stepwise multivariable logistic regression was performed using candidate variables selected based on the 2 criteria: (1) a variable with a P<0.05 in the univariable analysis and (2) clinically relevant as reported in previous literature. The following variables were included: age, mRS score before onset, NIHSS score, cardioembolism, ASPECTS on admission, LPCR, LPMC, multiple artery occlusion, intravenous thrombolysis, and onset‐to‐reperfusion time. To reduce overfitting and address potential multicollinearity, stepwise logistic regression was conducted using bidirectional selection based on the Akaike information criterion corrected for small sample size. To assess the discriminative performance of the final multivariable logistic regression model, the area under the receiver operating characteristic curve was calculated. All odds ratios (ORs), 95% CIs, and P values were obtained directly from JMP output based on the final logistic regression model.
As a sensitivity analysis, Firth's penalized likelihood logistic regression was applied to address quasi‐complete separation in the final model. This analysis was performed exclusively using the logistf package in R (version 4.3.2; R Foundation for Statistical Computing) to obtain reduced and stable estimates. Additionally, core volume was substituted for ASPECTS to test the robustness of the final model. In a separate model, age and NIHSS score were additionally included to account for their potential clinical importance. Patients without core volume data were excluded, and missing data were not imputed. To address potential variability in lesion classification, 2 additional models were constructed using high‐specificity and high‐sensitivity definitions based on interrater agreement. In the high‐specificity analysis, infarction in the LPCR or LPMC was recorded only when all raters agreed; in the high‐sensitivity analysis, infarction was considered present if identified by any rater. Finally, to explore whether the prognostic value of LPMC involvement differed across clinically relevant subgroups, we performed stratified multivariable logistic regression analyses by constructing separate models within each subgroup. These subgroup‐specific models were restricted to individuals belonging to each stratum (eg, only patients treated at Facility A, only those treated at Facility B, only male patients, or only female patients) and included the same covariates as the final model in the main analysis. Adjusted ORs with 95% CIs were calculated for each subgroup. To evaluate potential effect modification, omnibus P values for interaction were derived for each subgroup comparison.
All statistical values, including P values and CIs, were reported as generated by JMP without further adjustment. Slight discrepancies were observed between P values and 95% CI due to rounding differences or differences in statistical estimation methods.
RESULTS
A flow chart of the study is shown in Figure 2. A total of 107 patients were included in the study. Of these, 37 (34.6%) patients achieved a good outcome at 90 days (Table 1). The mean age was 75.4±14.8 years, and 64 (59.8%) patients were men. The median admission NIHSS score was 21 (interquartile range: 17–28). The median ASPECTS on admission was 4 (interquartile range: 3–5). Preoperative imaging to assess ASPECTS, LPCR, and LPMC was performed using DWI in 88 (82.2%) patients. The κ‐coefficient for interobserver agreement in identifying LPCR and LPMC was 0.75 (95% CI, 0.62–0.89) and 0.70 (95% CI, 0.56–0.83), respectively. A total of 77 (72.0%) patients had LPCR, and 67 (62.6%) had LPMC (Table 1). Among these, 84.4% of LPCR‐positive cases and 85.1% of LPMC‐positive cases were assessed using DWI, and 57 (85.1%) were assessed as positive on DWI, respectively. The median core volume was 79.7 (interquartile range: 51–123) mL, with 14 missing values. As the ASPECTS increased, a corresponding decrease in core volume was observed, following a generally consistent trend (Figure S1).
Figure 2.
Study flow chart. ASPECTS indicates Alberta Stroke Program Early Computed Tomography score; EVT, endovascular therapy; mRS, modified Rankin Scale.; and mTICI, modified Thrombolysis in Cerebral Infarction.
Table 1.
Characteristics of Patients
| Variables |
Total (n = 107) |
Good outcome (n = 37) |
Poor outcome (n = 70) |
P value |
|---|---|---|---|---|
| Age, y; mean±SD | 75.4 ± 14.8 | 69.6 ± 16.4 | 78.4 ± 13.0 | 0.003 |
| Male sex, n (%) | 64 (59.8) | 26 (70.3) | 38 (54.3) | 0.11 |
| Medical history | ||||
| Hypertension | 70 (65.4) | 22 (59.5) | 48 (68.6) | 0.35 |
| Diabetes | 24 (22.4) | 8 (21.6) | 16 (22.9) | 0.88 |
| Hyperlipidemia | 29 (27.1) | 14 (37.8) | 15 (21.4) | 0.07 |
| Atrial fibrillation | 64 (59.8) | 18 (48.7) | 46 (65.7) | 0.09 |
| Current smoker | 36 (33.6) | 14 (37.8) | 22 (31.4) | 0.50 |
| mRS score before onset, median (IQR) | 0 (0–1) | 0 (0–0) | 0 (0–1) | 0.04 |
| NIHSS score, median (IQR) | 21 (17–28) | 20 (13–27) | 23 (18–29) | 0.10 |
| Cardioembolism, n (%) | 74 (69.2) | 20 (54.1) | 54 (77.1) | 0.01 |
| Left artery occlusion, n (%) | 53 (49.5) | 18 (48.7) | 35 (50.0) | 0.89 |
| ASPECTS on admission, median (IQR) | 4 (3–5) | 4 (3–5) | 3.5 (2–5) | 0.02 |
| MRI‐ASPECTS, n (%) | 88 (82.2) | 29 (78.4) | 59 (84.3) | 0.45 |
| ASPECTS 3–5, n (%) | 86 (80.4) | 34 (91.9) | 52 (74.3) | 0.03 |
| LPCR, n (%) | 77 (72.0) | 24 (64.9) | 53 (75.7) | 0.23 |
| LPMC, n (%) | 67 (62.6) | 14 (37.8) | 53 (75.7) | 0.0001 |
| Core volume, mL; median (IQR) | 79.7 (51–123) | 79.3 (52–132) | 83.4 (45–104) | 0.46 |
| Occlusion site | 0.44 | |||
| ICA, n (%) | 49 (45.8) | 14 (37.8) | 35 (50.0) | |
| MCA M1 segment, n (%) | 47 (43.9) | 18 (48.7) | 29 (41.4) | |
| MCA M2 segment, n (%) | 11 (10.3) | 5 (13.5) | 6 (8.6) | |
| Multiple artery occlusion, n (%) | 8 (7.5) | 0 (0) | 8 (11.4) | 0.049 |
| Intravenous thrombolysis, n (%) | 22 (20.6) | 11 (29.7) | 11 (15.7) | 0.09 |
| OTR, min, median (IQR) | 324 (210–567) | 277 (194–363) | 358 (224–631) | 0.03 |
| ASTIN/SIR score, median (IQR) | 1 (0–2) | 2 (0–2) | 1 (0–2) | 0.41 |
| mTICI 2c/3, n (%) | 66 (61.7) | 25 (67.6) | 41 (58.6) | 0.36 |
| No. of passes to achieve mTICI ≥2b, median (IQR) | 1 (1–2) | 1 (1–2) | 1 (1–3) | 0.45 |
| Symptomatic ICH, n (%) | 6 (5.6) | 1 (2.7) | 5 (7.1) | 0.66 |
| Any ICH, n (%) | 50 (46.7) | 13 (35.1) | 37 (52.9) | 0.08 |
ASPECTS indicates Alberta Stroke Program Early Computed Tomography score; ASTIN/SIR, American Society of Intervention and Therapeutic Neuroradiology/Society of Interventional Radiology grades; ICA, internal carotid artery; ICH, intracranial hemorrhage; IQR, interquartile range; LCR, lesions in the posterior corona radiata; LPMC, lesion in the primary motor cortex; MCA, middle cerebral artery; MRI, magnetic resonance imaging; mRS, modified Rankin Scale; mTICI, modified thrombolysis in cerebral infarction; NIHSS, National Institutes of Health Stroke Scale; and OTR, onset‐to‐reperfusion time.
Univariable analysis revealed that the patients with a good outcome were significantly younger (69.6 versus 78.4 years old, P = 0.003) and had a lower prevalence of cardioembolic stroke (54.1% versus 77.1%, P = 0.02). The ASPECTS on admission imaging were higher in patients with a good outcome (median: 4 versus 3.5, P = 0.03). LPCR was more frequently observed in the poor outcome group than in the good outcome group; however, this difference was not statistically significant (64.9% versus 75.7%, P = 0.23). LPMC was observed significantly less frequently in patients with a good outcome than in those with a poor outcome (37.8% versus 75.7%, P = 0.0001). The prevalence of multiple artery occlusions was also lower, and the onset‐to‐reperfusion time was shorter in patients with a good outcome (0% versus 11.4%, P = 0.049, and 277 versus 358, P = 0.03, respectively).
Stepwise multivariable logistic regression identified 4 independent predictors of good outcome: mRS before onset (OR, 0.30 [95% CI, 0.10–0.73], P = 0.02), cardioembolism (OR, 0.25 [95% CI, 0.08–0.76], P = 0.02), absence of LPMC involvement (OR, 13.49 [95% CI, 3.75–63.45], P = 0.0002), and shorter onset‐to‐reperfusion time (OR, 0.996 [95% CI, 0.992–0.998], P = 0.004) as independent predictors of good outcome (Table 2). There was a nonsignificant trend toward higher ASPECTS being associated with a good outcome (OR, 1.59 [95% CI, 1.02–2.62], P = 0.0501). Absence of multiple artery occlusion was also retained in the final model; its coefficient estimate (β = 11.186) was unstable due to quasi‐complete separation, and statistical significance was not determined. Age, NIHSS score, intravenous thrombolysis, and LPCR involvement were not retained in the final model. The final multivariable logistic regression model demonstrated good discriminative ability, with the area under the receiver operating characteristic curve of 0.886.
Table 2.
Final Multivariable Logistic Regression Model for Predicting Good Outcome
| Variables | OR | 95% CI | P value |
|---|---|---|---|
| mRS score before onset | 0.30 | 0.10–0.73 | 0.02 |
| Cardioembolism | 0.25 | 0.08–0.76 | 0.02 |
| ASPECTS | 1.59 | 1.02–2.62 | 0.0501 |
| Absence of LPMC | 13.49 | 3.75–63.45 | 0.0002 |
| Absence of multiple artery occlusion | NA | NA | 0.9995 |
| Shorter OTR, per 1 min | 0.996 | 0.992–0.998 | 0.004 |
ASPECTS indicates Alberta Stroke Program Early Computed Tomography Score;LPMC, lesions in the primary motor cortex; mRS, modified Rankin Scale; NA, not applicable; OR, odds ratio; and OTR, onset‐to‐reperfusion time.
As a sensitivity analysis, Firth regression enabled estimation of the association between absence of multiple artery occlusion and good outcome, yielding an OR of 9.89 (95% CI, 0.85–1404.26, P = 0.07), although this association was not statistically significant (Table S1). Sensitivity analyses replacing ASPECTS with core volume showed consistent results with the main findings (Table S2). Additional models including age and NIHSS score confirmed the independent association of LPMC involvement with good outcome (Table S3). High‐specificity and high‐sensitivity definitions of LPCR and LPMC based on interrater agreement yielded similar trends, supporting the robustness of lesion classification (Table S4 and S5). Stratified multivariable logistic regression analyses demonstrated that the absence of LPMC involvement was significantly associated with a good outcome in several subgroups, including Facility A, age 71 to85 years, male, without intravenous thrombolysis, and those with right‐ and left‐sided occlusion (Table S6). However, effect estimates could not be calculated for certain strata (eg, Facility B, age >85 years, female sex, with intravenous thrombolysis) due to model nonconvergence or quasi‐complete separation, and these were noted as “NA.” The interaction P value for sex was 0.051, suggesting a trend toward interaction, although it did not reach statistical significance.
DISCUSSION
In this retrospective observational study of 107 patients with large ischemic core who underwent EVT, the absence of LPMC involvement identified primarily on preoperative DWI‐MRI was independently associated with a good outcome.
This supports the hypothesis that preservation of the primary motor cortex plays a critical role in functional recovery even in patients with substantial infarct burden. In contrast, LPCR involvement showed no significant association. Additionally, mRS before onset, cardioembolism, and shorter onset‐to‐reperfusion time were also independently associated with a good outcome. Although the absence of multiple artery occlusion was retained in the final stepwise model, its coefficient estimate was unstable due to quasi‐complete separation and did not reach statistical significance. ASPECTS was retained in the final model, but the association did not reach statistical significance. In a sensitivity analysis, substituting core volume for ASPECTS, core volume was not retained, and the results for other predictors remained consistent. These findings suggest that evaluating infarct topography—particularly involving functionally eloquent regions—may offer complementary prognostic information to traditional imaging markers such as ASPECTS and core infarct volume, which have been widely used in EVT selection criteria and validated in numerous clinical trials for patients with LIC.
A previous study demonstrated that the extent of reperfusion in the vessels supplying the primary motor cortex was independently and strongly associated with favorable outcomes, whereas the overall degree of partial reperfusion (mTICI 2b versus 2a) showed no significant correlation with outcomes. 9 These findings emphasize that reperfusion of functional tissue in the primary motor cortex is more closely associated with good clinical outcomes than volumetric assessments. Additionally, functional improvement at 90 days following EVT in LIC cases has been reported to be lower than in non‐LIC cases. 18 In large vessel occlusion with LIC, impairment of the primary motor cortex may further reduce the likelihood of symptom improvement.
In contrast, LPCR involvement was not significantly associated with a good outcome in patients with LIC. One study found that infarcts in the corticospinal tract region—including the internal capsule, caudate, lentiform nucleus, and central white matter, such as the corona radiata—were not significantly associated with poor clinical outcomes at 90 days in patients with embolic carotid‐T or M1 occlusions. 19 Another study comparing EVT‐treated proximal and distal M1 occlusions reported no significant difference in functional outcomes despite more frequent basal ganglia and corona radiata lesions in proximal occlusions. These findings suggest that white matter, including the corona radiata, may have greater ischemic tolerance than gray matter, 20 and that early reperfusion via EVT may have prevented infarct progression despite its appearance on imaging. Additionally, white matter plasticity may have minimized corticospinal tract damage. 21 Thus, LPCR was not considered to be associated with a good outcome even in cases with LIC.
ASPECTS has been widely validated as a prognostic marker and is routinely used as an inclusion criterion in EVT trials; however, its association with a good outcome in our final model was only marginally significant. This result may reflect partial collinearity with other variables or limited statistical power. Nonetheless, ASPECTS remains a valuable prognostic tool, as its utility has been supported by several studies evaluating the efficacy of EVT in patients with LIC. 6 , 22 Our findings further suggested that assessing infarct topography—particularly in eloquent areas such as the primary motor cortex—may offer additional prognostic value.
Core volume, which is frequently used in patient selection and prognostication, also did not remain in the final model in either the primary or sensitivity analyses. Although core volume provides important information regarding overall infarct burden, previous studies have reported that no consensus exists on a definitive upper core volume threshold for EVT. 1 , 2 , 3 , 4 , 5 In large vessel occlusion with LIC, no significant difference in clinical outcomes has been reported between mTICI 2b and mTICI 2c/3, suggesting that salvaging eloquent areas may be more critical than reducing final infarct core volume. 23 Additionally, a multicenter retrospective registry study including patients with LIC found that visually assessed ASPECTS was a more reliable predictor of functional outcomes at 3 months than automated CT perfusion‐derived ischemic core volume, supporting our findings. 24
Notably, the absence of multiple artery occlusion was also retained in the final model. However, due to quasi‐complete separation, its coefficient estimate was unstable, and the statistical inference was not reliable. This may reflect data sparsity or imbalance rather than a true absence of association. Larger prospective studies are warranted to better clarify its prognostic value.
The association remained robust in a sensitivity analysis that included age and NIHSS score in the final model, addressing concerns regarding potential confounding or model underfitting. Additionally, the association between LPMC involvement and good outcome remained consistent across all sensitivity analyses, including models using stricter or more liberal lesion definitions based on interrater agreement. In contrast, LPCR involvement showed no significant association in any model. Subgroup analyses using stratified multivariable models showed that the absence of LPMC involvement was significantly associated with a good outcome in several subgroups. However, no statistically significant effect modification was observed based on omnibus interaction P values. Notably, the interaction P value for sex was 0.051, which did not meet the conventional threshold for significance but indicated a possible sex‐related difference in prognostic relevance. Bonkhoff et al 25 reported that in women, lesions associated with poor outcomes were more widely distributed, particularly involving cortical regions and across the hemisphere, which may influence outcome prediction. Accordingly, lesions outside the primary motor cortex may also contribute substantially to outcome determination in women, potentially diminishing the relative influence of LPMC involvement. The inability to estimate effect sizes in certain subgroups likely reflects small sample sizes and sparse outcome events, which may have led to quasi‐complete separation and unstable estimates with wide CI. For example, estimates were not available for subgroups such as age 75–85 years or right‐sided occlusion, and some estimates showed wide CIs, likely due to limited sample size and variability in outcomes.
This study has several important limitations. First, although we included clinically relevant confounders such as age, NIHSS score, and intravenous thrombolysis in the stepwise logistic regression, these variables were not retained in the final model due to the statistical criteria. Therefore, their potential influence cannot be fully excluded. The relatively small sample size may have limited statistical power, particularly for patients with extremely large ischemic strokes (ASPECTS≤2), who represented only 19.6% of this study. Additionally, quasi‐complete separation was observed for the variable “multiple artery occlusion,” resulting in an unstable estimate despite its inclusion in the final model. This reflects a modeling limitation in small or imbalanced data sets, and the potential influence of this variable on outcomes should be interpreted with caution. Second, the majority cohort (82.2%) underwent MRI‐based preoperative imaging. Although MRI allows for greater sensitivity in detecting early ischemic changes, particularly in white matter, CT remains the standard in many regions worldwide. Additionally, core volume was assessed using DWI‐positivity, which may lead to overestimation, as up to 25% of DWI‐positive lesions can be reversible with prompt recanalization. 26 Therefore, the generalizability of our findings may be limited in settings where only CT was used. 27 Third, recanalization of vessels supplying the primary motor cortex was not evaluated. However, given that this study focused on the corticospinal tract lesions on preoperative imaging, and that >90% of cases with mTICI ≥2b reportedly achieve at least partial recanalization in these vessels, the impact on our findings is likely minimal. 9 Fourth, this retrospective study was subject to potential selection and measurement bias. EVT eligibility was determined by treating physicians, which may have led to underrepresentation of patients with poor baseline profiles, such as extensive LPMC or LPCR involvement. Additionally, ASPECTS was assessed by the treating physicians rather than through blinded or centralized adjudication. Given its moderate interrater variability, especially in patients with large ischemic cores, potential misestimation of ASPECTS may have influenced its observed prognostic association in this cohort. Finally, all patients enrolled in this study were Japanese, and racial differences were not examined. Therefore, the findings may not be fully generalizable.
Further prospective studies with larger cohorts, standardized imaging review, and complete outcome data are necessary to validate our findings and refine EVT decision‐making for patients with LIC.
CONCLUSION
In the preoperative imaging evaluation of patients with LIC, the absence of LPMC involvement was independently associated with a good outcome. These findings suggest that incorporating assessment of primary motor cortex involvement may provide complementary information to existing criteria such as ASPECTS and infarct volume. Although LPMC may serve as an important independent prognostic factor for the efficacy of EVT in patients with LIC, further prospective validation is needed before it can be applied to treatment decision‐making in clinical settings.
Sources of Funding
The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not‐for‐ profit sectors.
Disclosure
Dr Nagao reports lecture fee from Medtronic, Kaneka, Daiichi Sankyo, Asahi‐Intec, Terumo. Corporation, Japan Lifeline Co, and Johnson and Johnson. Dr Shindo reports lecturer fees from Medtronic, Kaneka, Stryker, Daiichi Sankyo, Asahi‐Intec, Ezai, Bayer, Abbot medical, Medicos Hirata, and Johnson and Johnson. Dr Ueda reports research grants and lecturer fees from Alnylam and Pfizer. Dr Nakajima reports lecture fees from Eisai, Amgen, Daiichi Sankyo, Kowa, Medtronic, Otsuka Pharmaceutical Co, Otsuka Pharmaceutical Factory, Japan Blood Products Organization, Stryker, Takeda, Biogen, Byer, and Sumitomo Pharma. Dr Namitome, Dr Kawamoto, Dr Kuroki, Dr Hayashi, Dr K. Terasaki, and Dr S. Terasaki have no conflict of interest to declare.
Ethics Statement
This study was approved by the Ethics Committee of Japanese Kumamoto Red Cross Hospital (Approval Number: 666). Written informed consent from each patient was waived because we used clinical information obtained in routine clinical practice.
Patient Consent for Publication
Not applicable.
Supporting information
Supplemental Figure S1: The correlation between ASPECTS and core volume.
Supplemental Table S1: Sensitivity analysis using Firth logistic regression to address quasi‐complete separation issues in the final models.
Supplemental Table S2: Stepwise multivariate logistic regression including core volume in place of ASPECTS.
Supplemental Table S3: Multivariable logistic regression model including age and NIHSS score in addition to the variables retained in the final model.
Supplemental Table S4: Multivariable logistic regression using a high‐specificity definition of LPCR and LPMC.
Supplemental Table S5: Multivariable logistic regression using a high‐sensitivity definition of LPCR and LPMC.
Supplemental Table S6: Subgroup analyses of the association between absence of LPMC involvement and good outcome.
Acknowledgments
The authors have nothing to report.
REFERENCES
- 1. Yoshimura, S , Sakai, N , Yamagami, H , Uchida, K , Beppu, M , Toyoda, K , Matsumaru, Y , Matsumoto, Y , Kimura, K , Takeuchi, M , et al. Endovascular therapy for acute stroke with a large ischemic region. N Engl J Med. 2022;386:1303‐1313. [DOI] [PubMed] [Google Scholar]
- 2. Sarraj, A , Hassan, AE , Abraham, MG , Ortega‐Gutierrez, S , Kasner, SE , Hussain, MS , Chen, M , Blackburn, S , Sitton, CW , Churilov, L , et al. Trial of endovascular thrombectomy for lage ischemic strokes. N Engl J Med. 2023;388:1259‐1271. [DOI] [PubMed] [Google Scholar]
- 3. Huo, X , Ma, G , Tong, Xu , Zhang, X , Pan, Y , Nguyen, TN , Yuan, G , Han, H , Chen, W , Wei, M , et al. Trial of endovascular therapy for acute ischemic stroke with large infarct. N Engl J Med. 2023;388:1272‐1283. [DOI] [PubMed] [Google Scholar]
- 4. Bendszus, M , Fiehler, J , Subtil, F , Bonekamp, S , Aamodt, AH , Fuentes, B , Gizewski, ER , Hill, MD , Krajina, A , Pierot, L , et al. Endovascular thrombectomy for acute ischemic stroke with established large infarct: Multicentre, open‐label, randomized trial. Lancet. 2023;402:1753‐1763. [DOI] [PubMed] [Google Scholar]
- 5. Costalat, V , Jovin, TG , Albucher, JF , Cognard, C , Henon, H , Nouri, N , Gory, B , Richard, S , Marnat, G , Sibon, I , et al. Trial of thrombectomy for stroke with a large infarct of unrestricted size. N Engl J Med. 2024;390:1677‐1689. [DOI] [PubMed] [Google Scholar]
- 6. Uchida, K , Shindo, S , Yoshimura, S , Toyoda, K , Sakai, N , Yamagami, H , Matsumaru, Y , Matsumoto, Y , Kimura, K , Ishikura, R , et al. Association between alberta stroke program early computed tomography score and efficacy and safety outcomes with endovascular therapy in patients with stroke from large‐vessel occlusion. a secondary analysis of the recovery by endovascular salvage for cerebral ultra‐acute embolism‐japan large ischemic core trial (RESCUE‐Japan LIMIT). JAMA Neurol. 2022;79:1260‐1266. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Inoue, M , Yoshimoto, T , Yamagami, H , Toyoda, K , Sakai, N , Matsumaru, Y , Matsumoto, Y , Kimura, K , Ishikura, R , Uchida, K , et al. Yoshimura, S , Expanding the treatable imaging profile in patients with large ischemic stroke: subanalysis from a randomized clinical trial. Stroke. 2024;55:1730‐1738. [DOI] [PubMed] [Google Scholar]
- 8. Banks, JL , Marotta, CA . Outcomes validity and reliability of the modified Rankin scale: implications for stroke clinical trials: a literature review and synthesis. Stroke. 2007;38:1091‐1096. [DOI] [PubMed] [Google Scholar]
- 9. Raychev, R , Saber, H , Saver, JL , Hinman, JD , Brown, S , Vinuela, F , Duckwiler, G , Jahan, R , Tateshima, S , Szeder, V , et al. Impact of eloquent motor cortex‐tissue reperfusion beyond the traditional thrombolysis in cerebral infarction (TICI) scoring after thrombectomy. J Neurointerv Surg. 2021;13:990‐994. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Ross, ED . Localization of the pyramidal tract in the internal capsule using whole‐brain dissection. Neurology. 1980;30:59‐64. [Google Scholar]
- 11. Yagishita, A , Nakano, I , Oda, M , Hirano, A , Location of the corticospinal tract in the internal capsule at MR imaging. Radiology. 1994;191:455‐460. [DOI] [PubMed] [Google Scholar]
- 12. Konishi, J , Yamada, K , Kizu, O , Ito, H , Sugimura, K , Yoshikawa, K , Nakagawa, M , Nishimura, T , MR tractography for evaluation of functional recovery from Lenticulostriate infarcts. Neurology. 2005;64:108‐113. [DOI] [PubMed] [Google Scholar]
- 13. Gao, Y , Zhang, Ke , Liu, H , Zong, Ce , Yang, H , Yao, Y , Xu, Y , Lesion indexes predict early neurologic deterioration in lenticulostriate single small subcortical infarction. AJNR Am J Neuroradiol. 2024;45:568‐573. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. von Elm, E , Altman, DG , Egger, M , Pocock, SJ , Gøtzsche, PC , Vandenbroucke, JP , The strengthening the reporting of observational studies in epidemiology (STROBE) statement: Guidelines for reporting observational studies. Int J Surg. 2014;12:1495‐1499. [DOI] [PubMed] [Google Scholar]
- 15. Higashida, RT , Furlan, AJ , Trial design and reporting standards for intra‐arterial cerebral thrombolysis for acute ischemic stroke. Stroke. 2003;34:109‐137. [Google Scholar]
- 16. Almekhlafi, MA , Mishra, S , Desai, JA , Nambiar, V , Volny, O , Goel, A , Eesa, M , Demchuk, AM , Menon, BK , Goyal, M , Not all “Successful” angiographic reperfusion patients are an equal validation of a modified TICI scoring system. Interv Neuroradiol. 2014;20:21‐27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Wahlgren, N , Ahmed, N , Dávalos, A , Ford, GA , Grond, M , Hacke, W , Hennerici, MG , Kaste, M , Kuelkens, S , Larrue, V , et al. Thrombolysis with alteplase for acute ischemic stroke in the Sage Implementation of Thrombolysis in Stroke‐Monitoring Study (SITS‐MOST): An observational study. Lancet. 2007;369:275‐282. [DOI] [PubMed] [Google Scholar]
- 18. Li, J , Duan, J , Zhang, L , Chen, J , Duan, Y , Yang, B , Low (0‐5) Alberta Stroke Program Early Computed Tomography Score on admission predictive of worse functional outcome after mechanical thrombectomy for anterior circulation large vessel occlusion. Eur J Med Res. 2023;28:266. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Ko, HC . The clinical outcomes of mechanical thrombectomy for proximal M1 occlusion involving lenticulostriate perforators: Is it worse than distal M1 occlusion? Retrospective observational study. Medicine. 2021;100:e27036. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Koga, M , Reutens, DC , Wright, P , Phan, T , Markus, R , Pedreira, B , Fitt, G , Lim, I , Donnan, GA , The existence and evolution of diffusion‐perfusion mismatched tissue in white and gray matter after acute stroke. Stroke. 2005;36:2132‐2137. [DOI] [PubMed] [Google Scholar]
- 21. Sampaio‐Baptista, C , Johansen‐Berg, H . White matter plasticity in the adult brain. Neuron. 2017;96:1239‐1251. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Sarraj, A , Hassan, AE , Abraham, MG , Ortega‐Gutierrez, S , Kasner, SE , Hussain, MS , Chen, M , Churilov, L , Johns, H , Sitton, CW , et al. Endovascular thrombectomy for large ischemic stroke across ischemic injury and penumbra profiles. JAMA. 2024;331:750‐763. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Namitome, S , Uchida, K , Shindo, S , Yoshimura, S , Sakai, N , Yamagami, H , Toyoda, K , Matsumaru, Y , Matsumoto, Y , Kimura, K , et al. Number of passes of endovascular therapy for stroke with a large ischemic core: secondary analysis of RESCUE‐Japan LIMIT. Stroke. 2023;54:1985‐1992. [DOI] [PubMed] [Google Scholar]
- 24. Dittrich, TD , Nguyen, A , Sporns, PB , Toebak, AM , Kriemler, LF , Rudin, S , Zietz, A , Wagner, B , Barinka, F , Hänsel, M , et al. Large ischemic core defined by visually assessed ASPECTS predicts functional outcomes comparably accurate to automated CT perfusion in the 6‐24 h window. Eur Stroke J. 2024;10:552‐559. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Bonkhoff, AK , Schirmer, MD , Bretzner, M , Hong, S , Regenhardt, RW , Brudfors, M , Donahue, KL , Nardin, MJ , Dalca, AV , Giese, A‐K , et al. Outcome after acute ischemic stroke is linked to sex‐specific lesion pattern. Nat Commun. 2021;12:3289. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Nagaraja, N , Forder, JR , Warach, S , Merino, JG , Reversible diffusion‐weighted imaging lesions in acute ischemic stroke: A systematic review. Neurology. 2020;93:571‐87. [Google Scholar]
- 27. Barber, PA , Demchuk, AM , Zhang, J , Buchan, AM Validity and reliability of a quantitative computed tomography score in predicting outcome of hyperacute stroke before thrombolytic therapy. ASPECTS Study Group. Alberta Stroke Program Early CT Score. Lancet. 2000;355:1670‐1674. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental Figure S1: The correlation between ASPECTS and core volume.
Supplemental Table S1: Sensitivity analysis using Firth logistic regression to address quasi‐complete separation issues in the final models.
Supplemental Table S2: Stepwise multivariate logistic regression including core volume in place of ASPECTS.
Supplemental Table S3: Multivariable logistic regression model including age and NIHSS score in addition to the variables retained in the final model.
Supplemental Table S4: Multivariable logistic regression using a high‐specificity definition of LPCR and LPMC.
Supplemental Table S5: Multivariable logistic regression using a high‐sensitivity definition of LPCR and LPMC.
Supplemental Table S6: Subgroup analyses of the association between absence of LPMC involvement and good outcome.