Infarct-volume prognostic value depends on outcome ascertainment and validation design in public stroke MRI datasets.

PURPOSE: To determine how participant-level outcome missingness and center-level endpoint absence influence estimates of the incremental prognostic value of infarct volume in public stroke magnetic resonance imaging datasets. METHODS: We analyzed two public cohorts separately. In the Stroke Outcome Optimization Project (SOOP), 948 of 1106 S records had a valid acute lesion mask and core predictors; 620 had discharge modified Rankin Scale (mRS). Models containing age, sex, and admission National Institutes of Health Stroke Scale, with or without log lesion volume, underwent repeated nested cross-validation and inverse-probability weighting. In ISLES'24, models developed in center 1 (n = 87) were evaluated in center 2 (n = 39) for 3-month mRS and compared with pooled random cross-validation. The primary measure was the ranked probability score (RPS) difference. RESULTS: In SOOP, observed outcomes were associated with larger lesions than missing outcomes (median, 15.08 vs 3.68 mL; standardized mean difference for log volume, 0.486). Adding volume improved RPS by 0.0609 (95 % confidence interval [CI], 0.0314-0.0893); the weighted improvement was 0.0538 (95 % CI, 0.0267-0.0799). In ISLES'24, discharge mRS was absent from center 2. For 3-month mRS, the center-held-out RPS improvement was 0.0488 (95 % CI, -0.1228 to 0.2473), compared with 0.1351 in pooled random validation. Both models overpredicted disability in center 2. CONCLUSIONS: Outcome ascertainment and validation design changed the apparent prognostic contribution of infarct volume. Endpoint-by-center auditing should precede interpretation of multicenter imaging performance.