Abstract Body: Introduction: Intracerebral hemorrhage (ICH) volume and expansion are important predictors of clinical outcome. Recent results show volumetric thresholds for prediction of poor outcome differ based on the anatomic location (lobar vs deep) of the ICH. In this study we add further atlas-based anatomic detail including lobar and subcortical locations to determine location-specific volume and expansion thresholds for the prediction of poor clinical outcome, hypothesizing that these cut-offs will vary by location.
Methods: We analyzed 286 subjects from the MISTIE3 trial that had not undergone surgery prior to imaging analysis. Admission and follow-up CT scans at 24 hours ± 6 hours were analyzed. ICH locations (Basal Ganglia, Thalamus, Frontal, Occipital, Parietal, Temporal) were determined using adjudication and atlas-based methods from day 30 scans. ICH expansion was defined as a 33% or 6ml volume increase. Poor outcome was defined as modified Rankin Scale 4-6. We performed univariate and multivariate analysis using ICH location, age, ICH volumes, time from ictus to scan to identify variables associated with expansion and outcome (P < 0.05). Cutoffs at each location for volume and expansion, and their sensitivity and specificity for predicting outcome were determined using receiver operator characteristic curves.
Results: Baseline CT volumes averaged 42.68ml with 41% showing expansion. Multivariate results indicated that the volume in temporal, parietal, occipital and basal ganglia locations and volume expansion in the temporal, parietal, basal ganglia were significant predictors of poor outcome. Volume thresholds for the prediction of poor outcome at the different lobar locations were frontal (46.4ml), temporal (53.7ml), parietal (48.9ml), occipital (57.6ml), basal ganglia (45.0ml). Expansion thresholds for prediction of poor outcome were frontal (13.7ml), temporal (1.20ml), parietal (2.52ml), occipital (6.1ml), basal ganglia (6.20ml). Temporal, parietal and basal ganglia volume and expansion thresholds were significant.
Conclusion: Our results add specific lobar locations and their threshold values. Our volumetric expansion associations with outcome suggest that small expansions in the temporal and parietal lobes may be more clinically significant than in other locations. These results indicate that targeting of therapeutic interventions to reduce ICH expansion may have different effects based on specific lobar and deep locations.