Figure 3
From: Prospects for developing an accurate diagnostic biomarker panel for low prevalence cancers
Characteristics of biomarker panels with 99% specificity. The modeled sensitivity of biomarker panels is given for panels consisting of between 10 and 100 individual biomarkers. The three panels represent different assumptions about the individual biomarkers: Chart A indicates panel characteristics assuming all biomarkers in the panel yield 45% sensitivity at the 95% specificity threshold (e.g. TIMP-1), Chart B indicates panel characteristics assuming all biomarkers in the panel yield 32% sensitivity at the 95% specificity threshold (the average of the 9 biomarkers examined), and chart C indicates panel characteristics assuming all biomarkers in the panel yield 19% sensitivity at the 95% specificity threshold (e.g. haptoglobin). Different assumptions about the mean correlation ratios between individual biomarkers constituting the panels are indicated by the color legend. Numbers above the data points indicate the minimum number of biomarkers in the panels required to be above the 95% specificity threshold in order to make a positive diagnosis of PDAC.