Construction and Evaluation of a Predictive Nomogram

The independent clinicopathological covariates, identified by multivariate Cox regression analysis, were taken into consideration to assemble a nomogram, providing clinicians with a clinically relevant quantitative approach for predicting individualized survival probability in BC patients (Harrell et al., 1996; Zhang and Kattan, 2017). Additionally, the calibration plots were graphically depicted to evaluate the consistency between the predicted probability of derived nomogram and actual situation (Balachandran et al., 2015). And Harrell’s concordance index (c-index) was utilized to estimate the discrimination ability of the nomogram. Moreover, decision curve analysis (DCA) was performed to explore the clinical usefulness of the nomogram, further providing straightforward information concerning whether clinical decision-making based on a predictive model will do more good than harm, in contrast to abstract statistical concepts (Fitzgerald et al., 2015; Vickers and Elkin, 2016; Zhang et al., 2018d).