Fig. 1

Model diagram schema. The process of the model diagram is shown in a different section. a presents the selection of the datasets from the Project Data Sphere’s Data Sharing Platform that contained all features for building the predictive models. b shows the data frames made by features in training datasets and then made a model to predict the BOR. c shows the method used to predict the BOR. In this method, two training datasets were considered, and test data was used to segment the patients. The novel method applied seven main CV techniques to explore the optimal CV techniques. d is the standard method that built the predictive models and used two training datasets. e identifies that in the standard method the best cutoff was collected from the optimal CV techniques applied to segment the patients. f shows, in the standard method, a probability threshold of 0.5 converted the score of patients into binary group membership. Finally, g shows the comparison of the performances among predictive models using the CPH and AFT models, in addition to, presenting the Kaplan-Meier Curve for each method