Skip to main content
. 2018 Oct 16;8:438. doi: 10.3389/fonc.2018.00438

Figure 3.

Figure 3

Calibration plot of observed vs. predicted probabilityof PCa (A) and CSPCa (B) after leave-one-out cross validation, demonstrating excellent concordance. Decision curve analyses demonstrating net benefit between the threshold probabilities of ~15 and ~80% for the model predicting PCa (C) and between the threshold probabilities of ~10 and ~60% for the model predicting CSPCA (D).