Fig. 2

a Calibration of the VENUSS score and VENUSS group in predicting recurrence. The grey line represents the performance of an ideal prognostic model, while the purple and blue solid lines represent the performance of the VENUSS score and group, respectively. The graphs indicate good calibration. b Smoothed decision curves of VENUSS and other risk definitions predicting PRCC recurrence. Assuming that patients with PRCC would be treated differently (i.e. would be included in adjuvant trials), the net benefit of VENUSS is plotted against threshold probabilities and compared to UISS, TNM and Leibovich groups as well as the strategy of putting all or none into an adjuvant trial. Compared with UISS, TNM and the 2018 Leibovich prognostic group, VENUSS showed an improved net benefit between 10 and 40% threshold probability