Figure 1.

Receiver operating characteristic plots by randomly selected training (50%) and validation (50%) neural network models on patients with complete variables (N = 211). A semi-quantitative graphic presentation of the covariates relevance is presented for training and validation models. Full names of coded variables are reported in Additional File 2, Table S1. Keep = 1 means that covariate may stay in the model. Note that Gini's coefficients are practically identical for training and validation neural network models, respectively 0.742 and 0.741 (ROC AUC: 0.871 and 0.870, respectively). Therefore, training and validation neural network models have a very high, yet similar, accuracy and define a set of 5 predictive covariates useful to index long-term mortality in patients operated for Type A ascending aorta dissection.