Fig. 5
(a) Decision curve analysis (DCA) for the predictive model in the training set. The net benefit curve for the training set shows the relationship between the high-risk threshold and the net benefit. The curve for the training set (blue) is compared with the “All” and “None” strategies, with the former showing higher net benefits at various threshold levels. (b) Decision curve analysis (DCA) for the predictive model in the validation set. The net benefit curve for the testing set (orange) is similar to the training set, but with slight variations. This demonstrates the model’s performance and its predictive ability in an independent testing set. (c) Clinical impact curve for the predictive model in the training set. The plot shows the number of individuals classified as high risk (blue) at each threshold and the number of those who had the event (orange dashed line). As the high-risk threshold increases, fewer individuals are categorized as high risk, while the number of true events (dashed line) gradually decreases. (d) Clinical impact curve for the predictive model in the validation set.
