Fig. 2

Receiver operating characteristic (ROC) curves of the prediction model for the statistically significant prognostic factors. ROC curves of SVM classifier for pathological differentiation: (a1) training set (AUC, 0.871; std., 0.037; sensitivity, 80.6%; specificity, 89.2%); (a2) test set (AUC, 0.862; 95% CI, 0.750–0.967; sensitivity, 83.3%; specificity, 85.0%). ROC curves of MLP classifier for T stage: (b1) training set (AUC, 0.824; std., 0.087; sensitivity, 80.4%; specificity, 90.0%); (b2) test set (AUC, 0.809; 95% CI, 0.690–0.905; sensitivity, 76.2%; specificity, 74.1%). ROC curves of RF classifier for N stage: (c1) training set (AUC, 0.794; std., 0.100; sensitivity, 100.0%; specificity, 95.4%); (c2) test set (AUC, 0.746; 95% CI, 0.622–0.872; sensitivity, 79.3%; specificity, 72.2%)