Fig. 4

Receiver operating characteristic (ROC) curve comparison. A ROC curves of EAT density and volume in predicting the presence of metabolic syndrome. The area under the curve (AUC) in EAT density was 0.731 (95% CI 0.639–0.823), the best cutoff point was − 76.0HU with the sensitivity of 66.2% and the specificity of 74.5%, the predictive positive value (PPV) was 71.6%. The AUC in EAT volume was 0.694 (95% CI 0.597–0.791), the best cutoff point was 143.5cm3 with the sensitivity of 66.2% and the specificity of 72.5%, the PPV was 64.7%. B ROC curves of EAT density and volume in predicting the more severe metabolic syndrome. The AUC in EAT density was 0.735 (95% CI 0.649–0.822), the best cutoff point was − 72.1HU with the sensitivity of 87.2% and the specificity of 54.5%, the PPV was 73.7%. The AUC in EAT volume was 0.662 (95% CI 0.569–0.755), the best cutoff point was 134.3cm3 with the sensitivity of 69.2% and the specificity of 61.8%, the PPV was 62.4%. All p < 0.01. CI = confidence interval, EAT epicardial adipose tissue