Introduction: The spot sign is the strongest radiological marker of hematoma expansion (HE) in acute intracerebral hemorrhage (ICH). Yet clinical trials failed to show benefits from hemostatic agents in patients selected based on its presence. Hypothesis: Multi-phase CTA (mCTA) might improve HE stratification risk of spot sign. We aim to describe spot sign evolution patterns on mCTA and assess their association and predictive performance for HE. Methods: Acute ICH patients investigated with non-contrast CT and mCTA were included from Foothills Medical Center, Calgary (development cohort) and Vall d’Hebron Hospital, Barcelona (validation cohort). Spot sign mCTA patterns were: no spot sign (pattern 0), no expansion (pattern I), circular expansion (pattern II), and linear/chaotic expansion (pattern III) ( Figure 1-2 ). Logistic regression adjusted for baseline covariates assessed associations with HE (≥6mL or 33% volume increase), severe HE (≥12.5mL or 66%), and absolute/relative growth. Predictive performance (c-statistics) was compared between spot sign patterns and simple spot sign presence. Results: In the development cohort (n=217), median onset-to-imaging time was 225 (IQR=109-392) minutes, 67 (30.9%) patients were spot sign positive and 51 (23.5%) experienced HE: pattern 0=8.7%; pattern I=38.8%, pattern II=50%, and pattern III=71.1% (adjusted OR 3.23 2.26-4.62 per each pattern increase) ( Table 1 ). The predictive performance of spot sign pattern was not superior to simple spot sign presence (c-statistics 0.810 vs. 0.772; p=0.059). In the validation cohort (n=274), median onset-to-imaging time was 120 (IQR=78-241) minutes, 76 (27.7%) patients were spot sign positive and 67 (24.5%) experienced HE: pattern 0=14.5%, pattern I=50%, pattern II=23.5%, and pattern III=63.4% (adjusted OR 1.92 95%CI=1.49-2.49 per each pattern increase). The predictive performance of spot sign pattern was not superior to simple spot sign presence (c-statistics 0.707 vs. 0.712; p=0.727). Exploratory analyses showed a stronger association and better prediction when patterns I and II were combined, as well as in patients presenting after two hours from symptom onset. Conclusion: We describe a pathophysiology-based classification of spot sign evolution patterns, showing improved HE risk stratification. Pattern III (linear/chaotic expansion) had markedly higher HE risk than other patterns, suggesting that the type—not just the presence—of spot sign expansion drives HE risk.
Peres et al. (Thu,) studied this question.