Abstract / Summary
Background: Conventional markers of risk such as coronary artery calcium (CAC) scores are lower and less predictive in Black adults, despite a high prevalence of Type 2 diabetes mellitus (T2DM) and Metabolic Syndrome We evaluated whether incorporation of epicardial adipose tissue (EAT) features capture prognostic signal beyond CAC and social deprivation index, in predicting major adverse cardiovascular events. Methods: Adults with T2DM undergoing non-contrast CT calcium scoring from the CLARIFY registry (NCT04075162), including 1024 non-Hispanic Black participants and 3072 propensity-matched non-Hispanic White participants were included. EAT was automatically segmented and radiomic features were extracted and selected using survival-based machine learning. Cox proportional hazards models were constructed using CAC, Social Deprivation Index (SDI), and EAT features alone and in combination. The primary endpoint was 3-point major adverse cardiovascular events (MACE-3P), defined as cardiovascular death, non-fatal myocardial infarction, or non-fatal stroke. Model discrimination, reclassification, and calibration were assessed. Results: During a median follow-up of approximately 5 years, 206 (20%) Black participants and 422 (14%) matched White participants experienced MACE-3P In Black participants, the SDI+CAC+EAT model demonstrated the highest discrimination (C-index 0.62) and significantly improved reclassification compared with CAC alone (NRI) of 0.29 (95% CI 0.15–0.56; p < 0.0007). In contrast, the SDI+CAC+EAT model showed no improvement in discrimination when compared to CAC alone (C-index 0.67 v. 0.67) in the White cohort. Conclusion: In adults with T2DM, EAT may serve as a complimentary imaging marker when combined with CAC scores to provide incremental prognostic information or refine risk characterization.