AHA Scientific Sessions 2026 Abstract · Nov 2, 2026
Sex Differences in the Association Between AI-Quantified CCTA Plaque Composition and Incident Myocardial Infarction in Nonobstructive Coronary Artery Disease
By Lin Z, Guo W, Yin J et al.
Background: The prognostic impact of sex in coronary artery disease (CAD) has been widely explored, but little is known about how sex influences the relationship between quantitative coronary plaque composition and the risk of incident myocardial infarction (MI) in patients with nonobstructive CAD. Research question: To determine whether sex modifies the association between AI-quantified coronary plaque composition and incident MI in patients with nonobstructive CAD undergoing CCTA. Methods: We developed a deep learning pipeline for fully automated CCTA parameter extraction and conducted a retrospective analysis of the multicentre cohort, including 6161 patients with nonobstructive CAD undergoing CCTA (2506 women, 3655 men; median follow-up 8.0 years). AI-derived plaque metrics included non-calcified plaque volume (NCPV), calcified plaque volume (CPV), and total plaque volume (TPV). The primary outcome was defined as myocardial infarction (MI). Fine-Gray subdistribution hazard regression models were fitted for each plaque feature (TPV, NCPV, CPV) separately, adjusting for cardiovascular risk factors. The competing risk was all-cause death. The interaction analyses explored the potential relationship between plaque and sex. Results: At baseline, women were older than men (median 63 vs 61 years; P<0.001) and had a lower prevalence of diabetes (23.9% vs 28.7%; P<0.001). Men exhibited greater plaque burden across components (all P<0.001). In sex-stratified Fine-Gray analyses, higher plaque volumes were more strongly associated with incident MI in women than in men. For NCPV, adjusted subdistribution hazard ratios (sHRs) per unit increase were 1.104 (95% CI, 1.066-1.142) in women and 1.074 (95% CI, 1.057-1.092) in men. Corresponding sHRs for TPV were 1.070 (95% CI, 1.044-1.097) and 1.052 (95% CI, 1.038-1.066), respectively, and for CPV were 1.056 (95% CI, 1.027-1.085) and 1.023 (95% CI, 1.006-1.041), respectively. Sex significantly modified the associations between plaque composition and incident MI in both crude and multivariable-adjusted models (all interaction P<0.05). Conclusion: Among patients with nonobstructive CAD, sex significantly modifies the association between AI-quantified coronary plaque composition and incident MI. Sex-specific interpretation of quantitative CCTA plaque metrics may improve risk stratification for future MI in nonobstructive CAD.