Chest-CT-based radiomics feature of epicardial adipose tissue for screening coronary atherosclerosis

This study aims to investigate the diagnostic value of chest-CT epicardial adipose tissue (EAT) radiomics feature in coronary atherosclerotic stenosis.

Clinical data from 215 individuals who underwent coronary angiography and chest-CT scan from January to July 2022 at our institution were retrospectively analyzed. Based on the coronary angiography results, the total population, men, and women were divided into the CAD group and non-CAD group. radiomics feature of EAT at the level of the bifurcation of the left-main coronary artery on the transverse level of chest CT were measured. The features contain both first-order feature and shape-order feature.The differences between groups were analyzed using the t test or Chi-square test. The diagnostic efficacy of each parameter in diagnosing atherosclerotic stenosis of coronary arteries was assessed by plotting the receiver operating characteristic (ROC) curve.

First-order features: Mean, IntDen, Median, and RawIntDen; shape-order features: Area, Perim, Round, and BSA index; and clinical index: HbA1c showed statistical significance between the CAD group and the non-CAD group. The ROC curve analysis demonstrated high diagnostic efficacy, with the best for diagnostic efficacy being Median for the first-order feature parameter (AUC, 0.753; 95% confidence interval [CI], 0.689-0.817; t = 4.785, p < 0.001), Round for the shape-order feature (AUC, 0.775; 95% CI, 0.714-0.836; t = 7.842, p < 0.001), and HbA1c for the clinical index (AUC, 0.797; 95% CI, 0.783-0.856; t = 6.406, p < 0.001). After dividing the participants into male and female subgroups, the best diagnostic efficacy was observed with the BSA index for men (AUC, 0.743; 95% CI, 0.656-0.829; t = 5.128, p < 0.001) and Round for women (AUC, 0.871; 95% CI, 0.793-0.949; t = 7.247, p < 0.001).

Median, Round in radiomics feature of EAT on chest CT may play a role in the assessment of coronary atherosclerotic stenosis.