Figure 2.

The LASSO algorithm and 10-fold cross-validation were applied to extract the optimal subset of radiomic features. ROC curves for the radiomic model in predicting RFS. (A) The AUC reached the peak corresponding to the optimal number of radiomic features when the ln (λ) value increased to 0.0925. Optimal features were determined by the AUC value. (B) LASSO coefficient profiles of the 94 radiomic features. The vertical line was drawn at the value determined by 10-fold cross-validation, where the optimal λ generated six non-zero coefficients. (C) ROC curve of the training cohort. (D) ROC curve of the validation cohort.