Skip to main content
Fig. 3 | Diagnostic and Prognostic Research

Fig. 3

From: A novel method for interrogating receiver operating characteristic curves for assessing prognostic tests

Fig. 3

Illustrations of positive predictive value thresholds. a Impact of prevalence: equi-PPV lines for fixed positive predictive value (PPV = 0.50) and three different prevalence values; i.e. 0.05, 0.1 and 0.2. b Defining minimum predictive performance: division of the “ROC space” into four quadrants as defined by a PPV cut-off (PPV ≥ 0.50) and a sensitivity cut-off (S n  ≥ 0.50) corresponding minimum test requirements for a test to deliver clinically relevant prognostic performance (hypothetical). Only tests whose ROC curves cross both cut-off lines and have points in the upper-left quadrant (green) do outperform the minimum test requirements

Back to article page