Fig. 5

Construction and validation of a new model and nomogram for breast cancer diagnosis. A1 ROC curves of the new model for predicting breast cancer in the training set, validation set, training and validation sets. A2 ROC curves of the new model, CA153, CEA, CEA and CA153 for predicting breast cancer in training and validation sets. B1–3 Calibration curves of the new model for the training set, validation set, training and validation sets. C Violin plots that show distributions of predicted values for the new model, CA153, CEA, CEA and CA153. D1 The nomogram and D2 its example of the model. The overall probability is calculated by taking the sum of the risk points. For each parameter, its risk point can be determined by drawing a vertical line straight up from the parameter's value to the “Points” axis. In order to determine the probability of breast cancer, a vertical line is drawn intersecting the “Total points” with the “Pr()” line