Abstract Number: Esoc2026a1583 Hypoperfusion Topography as a Predictor of Clinical Improvement After Mechanical Thrombectomy in Anterior Circulation Stroke

Abstract Background and aims Mechanical thrombectomy (MT) is an effective stroke treatment, preventing damage to hypoperfused regions suffering from reversible ischemia. Recent studies have shown that lesion topography is crucial to determine stroke impairment and recovery. In this study, we tested whether topography of rescued hypoperfusion predicts clinical improvement after MT and compared the prediction of topographical measures with demographic and volumetric variables. Methods We retrospectively enrolled patients with large-vessel occlusion stroke treated with MT at the Stroke Unit and Neurology Clinic of Padova between January 2018 and June 2022. Clinical improvement was quantified as the difference in NIHSS between admission and 7 days. Individual maps of rescued tissue were obtained by subtracting 1-week lesion masks from baseline hypoperfused regions and normalized to a standard template. Using the Lesion Quantification Toolkit, we quantified disconnection of key white matter tracts (cortico-spinal, superior longitudinal fasciculus, arcuate fasciculus) and large-scale brain network involvement. Clinical improvement was used as the dependent variable in a regression model with bootstrapping. We compared demographic, volumetric, and topographic models based on R2adjusted values and tested the same predictors using hierarchical linear regression models. Results We included a total of n=56 patients (72±13 years old). Topographical model showed the highest predictive accuracy (R2 = 0.34), outperforming demographic (R2 = 0.07) and volumetric models (R2 = 0.15). The combined model integrating all predictors achieved the best performance (R2 = 0.42), significantly improving predictive accuracy (F-test-p-value=0.048). Conclusions These results demonstrate that hypoperfusion topography is a key determinant of clinical improvement after successful MT. Conflict of interest All authors have nothing to disclose. Figure 1 - belongs to Background and aims Figure 2 - belongs to Methods Figure 3 - belongs to Results