Compartment-specific computed tomography hemorrhage volumetry and long-term outcomes after aneurysmal subarachnoid hemorrhage

BACKGROUND: Quantitative intracranial blood volume on admission computed tomography (CT) is a promising imaging biomarker after aneurysmal subarachnoid hemorrhage (aSAH), but its long-term prognostic relevance remains incompletely characterized. We assessed the association between total and compartment-specific hemorrhage volumes and long-term functional outcomes. METHODS: We performed a retrospective cohort study of consecutive patients with aSAH admitted to a tertiary care center. Total and compartment-specific hemorrhage volumes were quantified on admission CT. An unfavorable outcome was defined as Glasgow Outcome Scale score ≤ 3 at 6 months, 1 year, and 3 years. For each endpoint, candidate predictors were restricted to variables available at admission; variable selection used least absolute shrinkage and selection operator-penalized logistic regression with bootstrap stability selection, and selected predictors were refitted with Firth’s bias-reduced logistic regression. Discrimination and calibration were internally validated using bootstrap optimism correction. RESULTS: A total of 170 patients were analyzed. Unfavorable outcomes occurred in 40.0%, 39.4%, and 42.9% at 6 months, 1 year, and 3 years, respectively. Greater total hemorrhage volume was independently associated with unfavorable outcome at 6 months (odds ratio [OR]: 1.21/10 mL; 95% confidence interval [CI]: 1.02–1.43), and greater intraventricular hemorrhage volume at 3 years (OR: 1.71/10 mL; 95% CI: 1.06–2.74); total and intraventricular volumes were also retained at 1 year but did not reach independent significance after adjustment. Apparent AUCs were 0.829, 0.833, and 0.840, with optimism-corrected AUCs of 0.804, 0.801, and 0.803, respectively. CONCLUSIONS: Admission CT hemorrhage volumetry was independently associated with unfavorable outcomes at discrete follow-up assessments after aSAH and provided information complementary to clinical severity and comorbidity burden. These endpoint-specific associations should not be interpreted as evidence of a longitudinal or causal shift in prognostic relevance.