Fig. 4

The performance of the scoring system to predict the probability of cardiac injury in AML patients. ROC curves and AUCs to evaluate the prediction accuracy in training set (A), testing set (B) and the external validation set (C). Calibration curves to assess the agreement of actual probabilities and predicted probabilities for prediction accuracy in the training set (D), testing set (E) and the external validation set (F). Cardiac injury predicting effects of the risk score in nomograms. Our nomogram divided participants into two subgroups according to the optimal threshold of ROC. The proportion of cardiac injury in each subgroup is in (G), P < 0.001. The OS was also assessed in different risk subgroups with P < 0.001 (H). ROC, receiver operating characteristic; AUC, area under the ROC curve. OS, overall survival