Figure 4. Model discrimination and performance in the training set. (A) Receiver operating characteristic curves for nomogram-based prognostic prediction. (B) Calibration plot examining estimation accuracy. (C) DCA assessing clinical utility.