Association between region-specific epicardial adipose tissue and CT-derived fractional flow reserve-defined myocardial ischemia: a retrospective coronary CTA study stratified by hypertension and type 2 diabetes mellitus

Background Epicardial adipose tissue (EAT) is a biologically active fat depot adjacent to the myocardium and coronary arteries and has been implicated in coronary atherosclerosis and myocardial ischemia. Although CT-derived fractional flow reserve (CT-FFR) provides integrated anatomic-functional assessment from coronary CTA, data on region-specific EAT phenotypes, particularly in patients with hypertension and type 2 diabetes mellitus (T2DM), remain limited. Purpose To investigate the association between region-specific EAT thickness and CT-FFR-defined myocardial ischemia and to assess its independent and incremental value in patients stratified by hypertension and T2DM. Methods In this retrospective study, consecutive individuals who underwent coronary CTA and CT-FFR analysis were included. Myocardial ischemia was defined as distal CT-FFR ≤ 0.80 in at least one major coronary artery. Region-specific EAT thicknesses were measured manually by two independent readers on standardized imaging planes. Multivariable logistic regression, ROC analysis, net reclassification improvement (NRI), integrated discrimination improvement (IDI), interaction analysis, and bootstrap internal validation were performed. Results A total of 576 participants were analyzed, including 148 controls, 131 patients with hypertension only, 145 with T2DM only, and 152 with both conditions. Compared with controls, ischemic subgroups had greater EAT thickness in multiple regions, with the largest difference observed for left atrioventricular groove EAT (LAVG-EAT) ( p < 0.001). In the fully adjusted model, LAVG-EAT remained independently associated with myocardial ischemia (OR 1.41, 95% CI 1.26–1.58; p < 0.001). LAVG-EAT yielded an AUC of 0.807, and its addition to the clinical plus coronary calcium burden model improved discrimination (AUC 0.886–0.912; DeLong p = 0.024), with a continuous NRI of 0.168 and an IDI of 0.023. The optimism-corrected AUC was 0.907. Conclusion Region-specific EAT, particularly LAVG-EAT, was independently associated with CT-FFR-defined myocardial ischemia and provided incremental discriminatory value beyond conventional clinical variables and coronary calcium burden. LAVG-EAT may serve as a practical adjunctive imaging marker for ischemia identification and risk stratification in patients with hypertension and/or T2DM.