Skip to main content
. 2023 Oct 5;13:16743. doi: 10.1038/s41598-023-42331-7

Figure 2.

Figure 2

Biomarkers improve prediction of prevalent AF. (a) ANGPT2, BMP10, and FGF23 provided added value to clinical variables (age, sex, BMI) as denoted by the increase of AUC from 0.655 (95% CI 0.620, 0.690) to 0.785 (95% CI 0.756, 0.815) for the apparent model; [AUC 0.743 (95% CI 0.712, 0.775) after optimism adjustment using bootstrapping]. (b) Calibration curve of the apparent model demonstrating near perfect calibration as expected for model development. ANGPT2 angiopoietin 2, AUC area under the ROC curve, BMP10 bone morphogenetic protein 10, CI confidence intervals, FGF23 fibroblast growth factor 23.