Table 2.

Performance comparison between univariate model and XGBoost model

Model	Accuracy(%)	Sensitivity(%)	Specificity(%)	AUC (95% CI)
XGBoost	74.09	79.57	71.12	0.82 (0.79–0.82)
f/tPSA	71.82	75.46	67.07	0.75 (0.72–0.76)*
tPSA	64.70	68.33	59.40	0.68 (0.65–0.70)*
fPSA	57.04	60.36	50.00	0.61 (0.58–0.63)*

* The AUC of XGBoost was significantly compared with that of f/tPSA, tPSA and fPSA models (each P<0.001)

f/tPSA free-to-total PSA ratio, tPSA total prostate-specific antigen, fPSA free prostate-specific antigen