4D-CTA Collateral Status Predicts Functional Outcome and Parenchymal Hematoma in Anterior Circulation Large Vessel Occlusion Stroke

Background Functional recovery after endovascular thrombectomy (EVT) for anterior-circulation large-vessel occlusion (LVO) stroke remains heterogeneous. Four-dimensional computed tomography angiography (4D-CTA) enables dynamic collateral assessment, but its utility for EVT triage remains uncertain. Methods We retrospectively analyzed 196 anterior-circulation LVO patients who underwent 4D-CTA. Collaterals were graded on a 0–4 scale and categorized as poor (0–1), intermediate (2), or good (3–4); for parenchymal hematoma (PH), collaterals were dichotomized as poor (0–1) versus non-poor (2–4). Primary outcomes were 90-day modified Rankin Scale (mRS) shift and PH occurrence. Multivariable ordinal logistic regression assessed mRS shift, and Firth logistic regression assessed PH. EVT subgroup models additionally adjusted for reperfusion status. Results Among 196 patients (59.2% EVT), median mRS (interquartile range [IQR]) improved across collateral tiers: 3 (1–4), 2 (1–3), and 1 (0–1) for poor, intermediate, and good collaterals, respectively ( P  < 0.001). Favorable outcome (mRS 0–2) increased from 46.2% to 96.2%. After adjustment, good collaterals predicted a favorable mRS shift (cOR 0.24, 95% confidence interval [CI] 0.08–0.66; P  = 0.005). PH occurred in 7.7%, with rates of 23.1%, 7.8%, and 0% across collateral tiers. Non-poor collaterals were protective against PH (odds ratio [OR] 0.18, 95% CI 0.04–0.80; P  = 0.024), including in the EVT subgroup (OR 0.13, 95% CI 0.03–0.70; P  = 0.018). No collateral-by-EVT interaction was observed ( P for interaction = 0.523). Conclusions 4D-CTA collateral status independently predicts 90-day functional outcomes and PH risk. Poor collaterals alone should not preclude EVT in otherwise eligible patients.