Figure 1
From: Optimal cut-point definition in biomarkers: the case of censored failure time outcome
Cut-point finding methods background. Panel A. ROC-based objective functions. J(cJ) is the thick line segment. CZ(cCZ) is the area of the dotted rectangle and ER(cER) is the thin line segment. Panel B. Behaviour of the population objective functions J(c) (thick line), CZ(c) (dotted line) and ER(c) (thin line). CZ(c) is multiplied by 1.10. In the two panels, the biomarker X is generated according to N(2.56,1) or N(0,1), depending on whether Z ≤ τ or Z > τ, respectively, leading to J = 0.8, CZ = 0.81, ER = 0.14.