Fig. 5

Developed the prediction nomogram based on selected radiomics and clinical features predicting 3-year PFS in training cohort (A). The probability of each predictor could be converted into scores according to the first scale “Points” at the top of the nomogram. After adding up the corresponding prediction probability at the bottom of the nomogram was the 3-year PFS. Calibration curves of nomogram in training (B), internal validation cohort (C) and external validation cohort (D), respectively. The X-axis represented the predicted probability estimated by nomogram, whereas the Y-axis represented the actual observed rates. The gray dashed line represented a perfect prediction by an ideal model, and the pink solid line represented the apparent prediction of nomogram. Calibration curves showed the actual probability corresponded closely to the prediction of nomogram. OSM: original, shape, Maximum2DDiameterRow; ONB: original, ngtdm, Busyness; WLGG: wavelet-LLL, glszm, GrayLevelNonUniformity; LGC: logarithm, glcm, Contrast