Fig. 6

a ROC curves of HRMRI findings and radiomic signature to detect the presence of RC PNI in training (left) and validation cohorts (right). b Calibration of the radiomic nomogram in training (left) and validation cohorts (right). The solid line represents the reference line with the observed value. The dotted line represents the performance of the hybrid nomogram, while the solid line is corrected for any bias in the hybrid nomogram. c DCA was performed to assess the clinical usefulness of the combined model in predicting PNI. The net benefit is measured by the y-axis. The red curves represent the combined model with Rad-score. The green lines represent the model without Rad-score. The blue curves represent all patients with PNI. The black lines represent patients without PNI. It indicated that using nomogram to predict PNI gains more benefit when the threshold probability from 0.05–0.5 and 0.6–0.75 for a doctor or a patient than “treat all” or “treat none”