Abstract 114: The Predictive Ability of the CTA Spot Sign for Hematoma Expansion is Dependent on Time Since ICH Onset: A Systematic Review and Patient-level Meta-analysis.
Background: Hematoma expansion (HE) occurs in up to 40% of patients with intracerebral hemorrhage (ICH), and predicts poor clinical outcome. Contrast extravasation following CT-angiography (CTA), termed “spot sign”, identifies patients at highest risk of HE. However, the prevalence and predictive values of the spot sign varies across studies, possibly due to differences in onset-to-CTA time. We therefore performed a patient-level meta-analysis to define the relationship between onset-to-CTA time and the prevalence & predictive value of spot sign, and the size of HE. Methods: We searched the Cochrane Central Register of Controlled Trials, the Cochrane Library Database of Systematic Reviews, MEDLINE and EMBASE for studies of CTA spot sign prevalence and HE. We pooled data on the prevalence and predictive values for significant HE (defined as either 6mL or 33% growth of ICH) for patients with ICH stratified by onset-to-CTA time: <3hours, 3-6 hours, >6hours. We used chi-square analysis to assess the spot sign in each time strata, and two-way ANOVA to compare across time strata. Results: We identified ICH spot sign databases derived from 7 countries and 14 centers (n=705). Prevalence of spot sign decreased with increasing onset-to-CTA time (Table; p<0.001). The subset with follow-up scans used for HE analysis (n=582) revealed spot sign sensitivity and PPV were highest in the earliest time strata, whereas specificity and NPV were highest in the latest time strata (Table). Spot positive patients had greatest absolute HE in the earlier CTA time strata (median spot positive growth 6.8mL, 5.6mL, 5.2mL for 6hr respectively; p<0.001; means in Table). Conclusion: Prevalence, predictive values and magnitude of effect of the spot sign are dependent on onset-to-CTA timing; these results are relevant to both ICH trial design and acute management.