Figure 2
From: Translation of proteomic biomarkers into FDA approved cancer diagnostics: issues and challenges
Hypothetical ROC curves. (A) A hypothetical non-parametric ROC plot. Each open square corresponds to the sensitivity and (1 minus specificity) values obtained for a particular decision threshold. The dashed diagonal line corresponds to the random chance line. The hashed region corresponds to the PAUC for the range of specificities between 68% and 100%. (B) Two hypothetical ROC curves with identical AUCs but different performances over the range of thresholds.