Fig. 6.

Performance of the developed models in the training and test cohorts using five-fold cross-validation. For each iteration, 80% of patients were used as the training cohort and the remaining patients as the test cohort.