INTRODUCTION
Endovascular thrombectomy (EVT) is the current standard of care for acute ischemic stroke (AIS) caused by large vessel occlusion (LVO), offering higher reperfusion rates and improved clinical outcomes [1]. The modified Thrombolysis in Cerebral Infarction (mTICI) grading system is widely used to assess reperfusion success, with grades 2b and 3 commonly considered indicative of successful recanalization in randomized EVT trials [2]. Advances in thrombectomy devices have significantly improved recanalization rates, distinguishing recent successful trials from earlier unsuccessful efforts [3,4]. However, 44.4–54.5% of AIS patients with LVO remain disabled despite achieving successful recanalization [5-10].
Futile recanalization is primarily attributed to a large ischemic core prior to EVT and poor collateral circulation, both of which are well-established predictors of poor outcomes [11-13]. While previous studies have predominantly emphasized ischemic core volume, the prognostic relevance of infarct location has received comparatively less attention. Basal ganglia (BG) infarction, resulting from occlusion of the lenticulostriate arteries—non-collateralized end arteries originating from the M1 segment of the middle cerebral artery (MCA)—presents a major challenge in AIS patients with M1 occlusion. Despite timely recanalization, BG infarction may progress rapidly, and is particularly vulnerable to reperfusion injury, often leading to hemorrhagic transformation (HT) [14-16]. Moreover, due to the BG’s anatomical proximity to key motor pathways in the internal capsule and corona radiata, infarction in this region frequently results in profound neurological deficits, significantly impairing functional recovery. However, the prognostic impact of BG infarction in stroke patients undergoing EVT remains controversial, particularly with respect to successful recanalization [17-19].
This study aimed to evaluate the association between BG infarction and clinical outcomes in AIS patients with LVO who achieved successful recanalization through EVT. By focusing on infarct location, this study seeks to improve understanding of factors contributing to futile recanalization.
MATERIALS AND METHODS
Population
This study included consecutive patients treated with EVT between March 2016 and January 2019 at Gyeonsang National University Hospital. There was no upper age limit for inclusion. Eligible patients had AIS due to LVO in the internal carotid artery (ICA) or M1 segment of MCA, with evidence of diffusion-weighted imaging (DWI)/fluid attenuated inversion imaging (FLAIR) mismatch and any neurological deficit, defined as a National Institute of Health Stroke Scale (NIHSS) score of at least 1. Only patients who achieved successful recanalization, defined as a mTICI grade of 2b or 3, were included. Patients with missing 3-month modified Rankin Scale (mRS) data or without pre-procedural DWI scan were excluded. Alberta Stroke Program Early Computed Tomography Score (ASPECTS) based on DWI was visually estimated [20]. The stroke etiology was classified as large artery atherosclerosis, cardio-embolism, other determined causes, or undetermined according to the Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria. All relevant data including demographics, risk factors, prior medications, baseline NIHSS scores, and procedural details were collected from our institution’s stroke database. HT was defined as parenchymal hematoma type 1 or 2 based on European Cooperative Acute Stroke Study (ECASS) classification [21]. Good clinical outcome was defined as a mRS score of 0–2 at 3 months, and poor outcome as a mRS score of 3–6. Futile recanalization was defined as a poor outcome despite successful recanalization (mTICI 2b or 3) following EVT.
Brain Imaging Protocol and Diffusion-Weighted Imaging Assessment
All patients initially underwent a non-contrast-enhanced brain computed tomography (CT) scan to exclude hemorrhagic stroke. Intravenous thrombolysis with tPA (0.9 mg/kg) was administered within 4.5 hours of last known well, according to standard eligible criteria. Magnetic resonance imaging (MRI) was subsequently performed, including DWI, FLAIR, susceptibility-weighted imaging, and contrast-enhanced magnetic resonance angiography. A fast MRI protocol with a total acquisition time of approximately 8 minutes was implemented; the actual time required varied depending on patient transportation and cooperation. MRI was performed using a 3T scanner (Ingenia; Philips Healthcare) equipped with a 20-channel neurovascular head coil. DWI sequences were acquired as single-shot echo planar images in the axial plane with the following parameters: TR (repetition time), 2,800 ms; TE (echo time), 78 ms; b, 1,000 s/mm2; field of view, 22 cm; matrix size, 148×145; and 20 slices with 5-mm slice thickness and 2-mm interslice gap. Apparent diffusion coefficient (ADC) maps were calculated using two b-values (b=0 and b=1,000 s/mm2). BG infarction was defined as any signal change in the BG regions, including caudate nucleus, putamen, and globus pallidus, on DWI or ADC maps (Fig. 1). Two reviewers assessed the presence of BG infarction, and inter-rater reliability was evaluated using kappa statistics. Any discrepancies in assessment were resolved by consensus.
Thrombectomy Procedure
EVT procedures were performed by 3 experienced interventionists using a femoral artery approach under local anesthesia. Occlusion sites were evaluated with digital subtraction angiography after the placement of a guiding catheter proximal to the occlusion in the cervical segment of the ICA or common carotid artery. Device selection, including guiding catheters, microcatheters, and thrombectomy devices was at the discretion of the interventionist based on arterial tortuosity and vascular anatomy. For stent retrieval, a microcatheter was navigated over a microwire beyond the clot, and the stent was deployed across the clot and retrieved with simultaneous aspiration using a 50-mL syringe through the guiding catheter. For direct aspiration, the aspiration catheter was advanced to the proximal end of the clot, and aspiration was applied using a syringe or pump. If recanalization was unsuccessful with a device, alternative modalities were employed at the interventionist’s judgement. Adjuvant intra-arterial tirofiban (up to 2 mg) was administered when necessary to prevent re-occlusion in cases of residual stenosis. Balloon angioplasty and/or intracranial or extracranial stenting were performed, as appropriate, in patients with severe intracranial or extracranial steno-occlusive disease.
Statistical Analysis
Categorical variables were compared using the chi-squared test, and continuous variables were analyzed using the Student’s t-test. Demographic, clinical, and procedural variables were compared between groups stratified by clinical outcome (good vs. poor) and by the presence or absence of BG infarction. Inter-rater reliability for the presence of BG infarction on DWI was assessed using Cohen’s kappa (k) statistics. Variables with a P-value<0.05 in the univariate analysis were included in multivariate binary logistic regression to identify predictors of futile recanalization. Statistical significance was defined as P<0.05. Multicollinearity analysis was conducted post hoc due to the potential collinearity between the presence of BG infarction and other key predictors that showed statistical significance in the multivariate model—namely, NIHSS score, DWI-ASPECTS, and HT. To evaluate potential collinearity, variance inflation factors (VIFs) and Pearson correlation coefficients were calculated. A VIF >5 or a Pearson correlation coefficient >0.7 was considered indicative of multicollinearity. All statistical analyses were performed using IBM SPSS Statistics version 25 (IBM Co.).
RESULTS
Of the 448 patients who underwent EVT for AIS, 287 had AIS with LVO of ICA and MCA M1. Successful recanalization (mTICI 2b or 3) was achieved in 230 patients (80.1%). Patients with mTICI grades 0, 1 (n=29), or 2a (n=28) were excluded, along with those lacking 3-month mRS (n=6) and preprocedural DWI data (n=2). Ultimately, 222 patients (107 males, 48.2%; mean age, 71.7±11.4 years) met the inclusion criteria and were included in the analysis (Fig. 2).
Among these 222 patients, 160 patients (72.1%) had BG infarction on pre-intervention DWI. Inter-rater agreement for BG infarction status was excellent (n=222, k=0.865, P<0.001). Compared to the no BG infarction group, patients with BG infarction had a higher frequency of embolic stroke, higher NIHSS scores, lower ASPECTS, less frequent use of tirofiban, and more HT (Table 1).
At the 3-month follow-up, 111 patients (50%) achieved a good outcome (mRS 0–2), while the remaining 111 (50%) had a poor outcome (mRS 3–6). Univariate analysis identified age, NIHSS score, DWI-ASPECTS, use of tirofiban, thrombectomy technique, the presence of BG infarction, atrial fibrillation, and HT as factors associated with clinical outcomes at 3 months (P<0.05; Table 2). Multivariate logistic regression analysis revealed 5 independent predictors of futile recanalization: older age (odds ratio [OR], 1.10; 95% confidence interval [CI], 1.06–1.15; P<0.001), higher NIHSS scores (OR, 1.20; 95% CI, 1.10–1.31; P<0.001), lower DWI-ASPECTS (OR, 0.79; 95% CI, 0.66–0.94; P=0.009), occurrence of HT (OR, 2.97; 95% CI, 1.11–7.95; P=0.031) and the presence of BG infarction (OR, 4.14; 95% CI, 1.70–10.06; P=0.002) (Table 3). Multicollinearity analysis showed no significant collinearity among predictors; All VIFs were below the commonly accepted threshold of 5 and Pearson correlation coefficients also did not exceed 0.7 (Supplementary Table 1).
DISCUSSION
Despite successful angiographic recanalization, the lack of clinical benefits remains a major challenge in the treatment of AIS in the endovascular era. This study identified 5 independent predictors of futile recanalization in patients with AIS and LVO who underwent successful EVT: older age, higher NIHSS scores, lower DWI-ASPECTS, development of HT and the presence of BG infarction.
The DWI-ASPECTS, an MRI-based adaptation of the original ASPECTS, is widely recognized for its correlation with infarct volume and superior inter-rater reliability compared to CT-based version [22-24]. The presence of BG infarction may confound the DWI-ASPECTS, as involvement of the BG can contribute to the score. Additionally, BG infarction is often associated with higher baseline NIHSS scores due to the involvement of motor pathways within the internal capsule and adjacent structures. Despite these associations, our analysis demonstrated that BG infarction remained an independent predictor of poor functional outcome after adjusting for both DWI-ASPECTS and NIHSS scores. While correlation analysis showed significant associations among BG infarction, NIHSS, and DWI-ASPECTS, multivariate logistic regression confirmed that each variable independently predicted poor outcomes, with no evidence of multicollinearity. These findings suggest that BG infarction contributes to clinical deterioration through mechanisms beyond its impact on initial stroke severity or infarct volume, highlighting the unique prognostic significance of infarct location.
One such mechanism may be an increased risk of HT. Prior studies have demonstrated that BG infarction is associated with higher rates of HT, likely due to occlusion of the M1 segment of MCA, which supplies the lenticulostriate arteries—terminal, non-collateralized perforators [15-18]. Recanalization via EVT may exacerbate reperfusion injury, leading to blood-brain barrier disruption and subsequently HT [15,16,25]. However, the prognostic significance of HT remains controversial. While Loh et al. [18] and Horie et al. [19] reported no significant association between HT and poor outcomes in patients with successful recanalization, our study identified HT as an independent predictor of unfavorable outcomes, even among patients who achieved successful recanalization. This discrepancy may stem from differences in sample size as our cohort included a larger number of patients with BG infarction and successful recanalization (n=160) compared to the previous studies. Another contributing factor is the heterogeneity in HT definitions. We defined HT as parenchymal hematoma type 1 or 2, whereas prior studies included all forms of HT, including hemorrhagic infarction. This heterogeneity in classification may account for differences in reported associations with outcomes.
Importantly, BG infarction remained an independent predictor of poor outcome irrespective of HT occurrence, suggesting the presence of additional deleterious mechanisms. One plausible explanation is that the damage of BG may impede post-stroke motor recovery, particularly gait function. Previous studies have linked BG infarction with poor motor recovery and impaired ambulation [26,27]. Lesions involving the caudate nucleus, putamen, and globus pallidus may disrupt neural reorganization and functional compensation, thereby limiting the recovery of voluntary motor control. This recovery delay may persist beyond the acute phase and continue to affect long-term outcomes, even independently of initial stroke severity as reflected by baseline NIHSS scores.
Consistent with prior literature, older age and higher baseline NIHSS scores were independently associated with poor outcomes despite successful recanalization [5-8,28]. Nonetheless, EVT remains beneficial compared to standard medical treatment, even among elderly patients or those with severe strokes [29]. These findings highlight the need for individualized risk-benefit assessments when selecting candidates for EVT in these high-risk populations.
This study has several limitations. First, it was based on single-center retrospective registry data, which are subject to selection bias and potential confounding. Collateral status, ischemic core volume, and mismatch profile assessed by advanced imaging techniques were not systematically evaluated. Moreover, strict ASPECTS thresholds were not applied for EVT candidate selection. Instead, EVT was performed in patients with DWI/FLAIR mismatch who were deemed likely to benefit clinically, resulting in relatively broad inclusion criteria. Recent randomized controlled trials, however, have demonstrated the efficacy of EVT even in patients with an ASPECTS ≤5 [30,31]. Notably, in our study, 50% of patients achieved a 3-month mRS score of 0–2, and the rate of successful recanalization was 80.1%, both comparable to those reported in previous EVT trials for LVO [32,33]. Furthermore, only 3.5% of patients (n=6 with missing 3-month mRS data and n=2 without preprocedural MRI) were excluded from the total population (n=230), thereby minimizing the potential for selection bias. Second, the use of MRI for EVT candidate selection, rather than CT, may have led to treatment delays. However, our fast MRI protocol (acquisition time <8 minutes) helped reduced such delays, with a mean door-to-puncture time was 110.8 minutes (median, 95 minutes), suggesting minimal impact on treatment timelines. Third, BG infarctions were not stratified by extent. Any high signal in the BG on DWI was classified as BG infarction, regardless of lesion size. However, a previous study using such stratification found no significant differences in outcomes between subgroups, except for hospital stay duration in the post hoc analysis [17]. Finally, several variables identified as predictors of futile recanalization in prior studies, such as pre-stroke disability, diabetes, C-reactive protein, and longer onset-to-admission time, were not confirmed in our study [8,10,28]. Those studies included larger sample sizes with high-quality data or were based on meta-analyses, whereas our study was limited by the inherent constraints of single-center retrospective data, affecting generalizability.
CONCLUSION
Recanalization significantly improves outcomes in AIS patients with LVO. However, approximately half of these patients continue to experience poor outcomes despite successful recanalization. In this study, older age, higher baseline NIHSS scores, lower DWI-ASPECTS, HT, and BG infarction were identified as factors associated with futile recanalization. Notably, BG infarction was independently associated with poor outcomes, regardless of infarction volume, initial stroke severity, or the presence of HT. These findings underscore the prognostic importance of infarct location, particularly within the BG. Tailored EVT decision-making strategies for patients with BG infarction should be explored.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Full text via PMC
Download Citation
Supplement
Print
