Fig. 1

Diagnostic model for LC based on CEA, CYFRA21-1 and NSE. A and D ROC and corresponding AUCs for diagnostic prediction by cd-score, CEA, CYFRA21-1 and NSE in training data set (A) and test data set (D). B and E Calibration curves which show the relationship between the predicted probabilities base on the diagnostic model (cd-score) and actual values of the training data set (B) and test data set (E). C and F Decision curve analysis (DCA) of diagnosis based on diagnostic model (cd-score) and each individual TM in training data set (C) and test data set (F). G Nomogram for predicting LC diagnosis based on cd-score. LC, lung cancer. CEA, carcinoembryonic antigen (CEA). CYFRA21-1, cytokeratin-19 fragment. NSE, and neuron-specific enolase