Fig. 4: Assessment of the comprehensive evaluation and predictive prognostic ability of the models.

The radar plot visualizes the overall performance of the predictive models on the training set (a), external vali-set1 (b), and external vali-set2 (c). Decision curve analysis for the predictive models in the training set (d) and the two validation sets (d–f). g–j Calibration curves of the predictive models in the three datasets. Kaplan‒Meier curves of progression-free survival in the outcome set based on the median of the radiomics model score (k) and clinical–radiomics score (l).