Fig. 2

The least absolute shrinkage and selection operator (LASSO) Cox regression algorithm were used to select optimal radiomic features. The selected tuning parameter (Lambda) in LASSO model through tenfold cross-validation were with minimum criteria in predicting (a) PFS and (b) OS. LASSO coefficient distributions of the radiomic features in predicting (c) PFS and (d) OS. The dotted vertical lines indicate the optimal values using the minimum criteria