ABSTRACT NUMBER: ESOC2026A644 PREDICTION OF ENDOVASCULAR TREATMENT OUTCOMES USING PERFUSION PROFILES OF ELOQUENT BRAIN REGIONS: A MULTICENTER STUDY

Abstract Background and aims Pivotal endovascular thrombectomy (EVT) outcomes prediction for acute large vessel occlusion (LVO) based on clinical-core and core-penumbra mismatch was suboptimal. We hypothesised that the perfusion status of eloquent brain regions is more important than the global cerebral perfusion in post-EVT prognostication. We aimed to determine EVT outcomes in patients with critical hypoperfusion in eloquent brain regions. Methods In this multicentre retrospective study, we retrieved patients with acute middle cerebral artery or internal carotid artery occlusion who received computed tomographic perfusion prior to successful EVT from 6 hospitals in Hong Kong and mainland China from January 2020 to July 2025. We assessed the association between the eloquent regional infarct core volume and the primary outcome in a priori set of brain regions. LASSO regression was performed to select potential eloquent regions with dichotomized regional infarct involvement. LASSO-selected regions with priori set were subjected to eloquent models using multivariable logistic regression. All regression models were adjusted for global infarct core volume and other clinical demographics. Results 480 acute LVO patients (222[46.2%] poor recovery) were identified during the study period, while 54 more cases (26[48.1%] poor recovery) were identified for the external validation cohort. Multivariate logistic regression revealed that eloquent brain regions in priori set was significantly associated with poor functional recovery (Figure-1). The eloquent models showed a higher predictive ability compared with the control model in the main cohort and external validation cohort respectively (Figure-2). Conclusions The inclusion of eloquent brain regions may significantly improve the prognostic accuracy of successful EVT. Conflict of interest Nothing to disclose Figure 1 - belongs to Results Figure 2 - belongs to Results