Development and validation of an artificial neural network prognostic model after gastrectomy for gastric carcinoma: An international multicenter cohort study.

摘要: Background Recently, artificial neural network (ANN) methods have also been adopted to deal with the complex multidimensional nonlinear relationship between clinicopathologic variables and survival for patients gastric cancer. Using a multinational cohort, this study aimed develop validate an ANN-based prediction model Methods Patients cancer who underwent gastrectomy in Chinese center, Japanese recorded Surveillance, Epidemiology, End Results database, respectively, were included study. Multilayer perceptron was used model. Time-dependent receiver operating characteristic (ROC) curves, area under curves (AUCs), decision curve analysis (DCA) compare ANN previous models. An nine input nodes, hidden two output nodes constructed. These three cohort's data showed that AUC of 0.795, 0.836, 0.850 5-year prediction, respectively. In calibration analysis, ANN-predicted had high consistency actual survival. Comparison DCA time-dependent ROC models good stable capability compared all cohorts. Conclusions The has significantly better discriminative allows individualized prediction. This versatility Eastern Western clinical application value.