Intracerebral Hemorrhage Expansion in Late Presenters

BACKGROUND AND OBJECTIVES: The benefit of therapies targeting hematoma expansion (HE) in acute intracerebral hemorrhage (ICH) is highly time dependent and greatest within the first 2 hours after onset. We aimed to develop imaging-based "target expansion criteria" to identify late presenters (2-12 hours from onset/last seen well) who remain at high risk of HE. To achieve this goal, a practical risk score (target expansion score) was developed. METHODS: We conducted a multicenter, international, retrospective cohort study. To mirror a clinical trial population, we excluded patients with brainstem ICH, anticoagulation, and baseline ICH volume >100 mL. The primary outcome was HE (>33%/6 mL). The following imaging predictors of HE were analyzed: high ultra-early hematoma growth (uHG) (baseline volume/onset to imaging time >10 mL/h), hypodensities on noncontrast CT (NCCT), and spot sign on CT angiography (CTA). The study population was randomly divided into development and validation cohorts, and a weighted score was derived from estimates from multivariable logistic regression. RESULTS: Overall, 492 patients were included: 345 (70.1%) in the development and 147 (29.9%) in the validation cohort. A 4-point score was created, based on independent predictors of HE (high uHG: 1 point, NCCT hypodensities: 1 point, and CTA spot sign: 2 points). The score showed good discrimination in the development and validation cohorts (area under the curve [AUC] 0.77, 95% CI 0.71-0.83, and AUC 0.79, 95% CI 0.71-0.87, respectively). We observed a nonlinear increase in the frequency of HE and mean absolute hematoma growth with higher scores: 9.8% and 0.9 mL (score 0), 17.1% and 3.9 mL (score 1), 48.3% and 11.5 mL (score 2), 70.8% and 13.0 mL (score 3), and 87.5% and 26.9 mL (score 4). These findings were confirmed in the validation cohort. A dichotomized score greater than or equal to 2 predicted HE with 0.61 sensitivity and 0.85 specificity in the development cohort and 0.64 sensitivity with 0.84 specificity in the validation cohort. DISCUSSION: Imaging-based target expansion criteria, defined as a target expansion score ≥2, identified late-presenting patients with ICH at high risk of HE. These findings might inform future trials targeting HE and increase the pool of eligible patients.