Pretreatment Perfusion-Collateral Impairment Identifies Outcome Heterogeneity After Thrombectomy Reperfusion

PURPOSE: Outcomes remain heterogeneous after angiographically successful endovascular thrombectomy (EVT), and pretreatment perfusion-collateral status may contribute to this variability. We evaluated whether the Perfusion Collateral Impairment Score (PCIS), a composite of hypoperfusion intensity ratio, cerebral blood volume index, and prolonged venous transit, was associated with 90-day functional outcome across mTICI reperfusion groups after EVT. METHODS: This retrospective secondary analysis included patients from 2 hospitals with ICA, M1, or M2 occlusion who underwent EVT and had available pretreatment CT perfusion, PCIS, final mTICI grade, and valid 90-day modified Rankin Scale (mRS). PCIS was dichotomized as low burden (0–1) versus high burden (2–3), and successful reperfusion was defined as mTICI 2b–3. The primary outcome was 90-day mRS 0–2. Multivariable logistic regression among patients with mTICI 2b–3 adjusted for age, continuous admission NIHSS, ASPECTS, intravenous thrombolysis, admission glucose, occlusion site, and mTICI 2b versus 2c–3. Prespecified NIHSS-threshold, ordinal PCIS, component, and model-performance analyses were also performed. RESULTS: Among 194 patients, 172 achieved mTICI 2b–3 reperfusion. Favorable outcome occurred in 63/108 patients (58.3%) with PCIS 0–1 versus 27/64 (42.2%) with PCIS 2–3. In the primary model using continuous NIHSS, high PCIS burden was associated with lower odds of favorable outcome, although the confidence interval crossed unity (adjusted OR, 0.56; 95% CI, 0.25–1.25; p=.16). In the prespecified NIHSS >5 model, high PCIS burden remained independently associated with lower odds of favorable outcome (adjusted OR, 0.44; 95% CI, 0.20–0.94; p=.04). Each 1-point increase in ordinal PCIS was associated with lower odds of favorable outcome (adjusted OR, 0.67; 95% CI, 0.45–0.99; p=.05), with improved model fit (likelihood-ratio p=.04). PVT was the only independently associated component (adjusted OR, 0.44; 95% CI, 0.20–0.95; p=.04). CONCLUSIONS: Pretreatment PCIS identified clinically meaningful outcome heterogeneity after EVT. High PCIS burden remained independently associated with unfavorable outcome in the prespecified NIHSS >5 model, while ordinal PCIS demonstrated a graded relationship with outcome and improved model fit. These findings support PCIS as a complementary pretreatment risk-stratification marker after successful reperfusion.