Figure 4 Nomograms for predicting 1-, 3-, and 5-year survival of patients by training cohort. (A) Nomogram of cancer-specific survival. (B) Nomogram of overall survival.Nomogram validationThe nomogram was assessed using the area under the curve (AUC) and concordance index (C-index). In the training cohort, as shown in (Figure 5), the C-index of CSS and OS prediction nomograms were 0.609 (95%CI, 0.598 to 0.619) and 0.697 (95%CI, 0.686 to 0.708) , and the AUC in predicting 1-, 3- and 5-year CSS were 0.746, 0.741 and 0.731, the AUC in predicting 1-, 3- and 5-year OS was 0.636, 0.642 and 0.633. Respectively, while in the validation cohort the C-index of CSS and OS prediction nomograms were 0.682 (95%CI, 0.665 to 0.698) and 0.605 (95%CI, 0.589 to 0.621), and in predicting 1-, 3-, and 5-year CSS were 0.732, 0.720 and 0.710, the AUC in predicting 1-, 3-, and 5-year OS was 0.609, 0.633 and 0.631(Figure 6). Calibration curves of survival rates at 1, 3, and 5 years in the training cohort(Figure 7) and validation cohort (Figure 8) showed satisfactory internal modelfitting capabilities. Excellent discrimination and calibration demonstrated the accuracy of the prognostic model.