Fig. 3.

Selection of radiomics features using LASSO logistic regression and the predictive accuracy of the radiomics signature. a Selection of the tuning parameter (λ) in the LASSO model via tenfold cross-validation based on minimum criteria. b The coefficients have been plotted vs. log(λ). c The final retained features with nonzero coefficients. d Radiomics score distribution in the training and validation cohorts; the optimum cutoff value was −0.41