a clear trendP value = .076
Although the specified significance threshold was not reached, there was a clear trend toward an improved predictive discrimination of the hybridized ML model with respect to the expert-based model in the test data (Concordance index, P value = .076; AUROC, P value = .731), while a significant difference was patent between the hybridized ML model and the calcium score-based model (Concordance index, P value = .014; AUROC, P value = .025).