Fig. 6: Receiver operating characteristic curves for all-cause mortality and cardiovascular mortality using models.
From: Predicting mortality from AI cardiac volumes mass and coronary calcium on chest computed tomography
Receiver operating characteristic curves for all-cause mortality and cardiovascular mortality for clinical (age, sex, medical history - gray), deep learning (DL) derived imaging variables (coronary artery calcium, cardiac volumes, shape index, eccentricity index, and left ventricular mass index – red), and a combined model incorporating all variables (DL derived and clinical variables - blue). Variables in the combined model were integrated using logistic regression analyses. The center of the bar represents area under the receiver operating characteristic curve (AUC), with 95% confidence intervals (CI) shown with whisker plots. AUC is compared using DeLong’s test. Source data is available.
