Fig. 3

Construction and validation of a prognostic model for pancreatic cancer. A A combination of machine learning predictive models calculates the C-index for each model on the training set and the validation set. B Forest map of univariate cox regression analysis showed hazard ratios of the seven selected genes. C Coefficients of genes in the prognostic model. D Kaplan–Meier curves of OS in the train, test sets and TCGA, ICGC cohorts based on the model showed longer survival time in low-risk groups. E Univariate and multivariate Cox regression analyses were performed to assess the prognostic value of the risk score in conjunction with additional clinical features. F A nomogram combining risk scores and clinical information was constructed in the TCGA dataset