To compare the accuracy of serum total prostate-specific antigen (tPSA), free PSA (fPSA) and age in predicting prostate volume (PV), aiming to identify the simple and accurate biomarkers for predicting PV.

Age, PV, and PSA values of patients diagnosed with lower urinary tract symptoms/benign prostatic obstruction (LUTS/BPO) admitted to our hospital from January 2005 to December 2014 were collected. The data was processed using SPSS 13.0 statistical software. Pearson's linear correlation was utilized to describe the correlation between age, serum tPSA, serum fPSA and PV. Chi-square test and the receiver characteristic curve (ROC) were used to analyze and compare the accuracy of serum tPSA and fPSA in predicting PV.

A total of 6308 male patients were enrolled. Pearson's linear correlation analysis demonstrated that the correlation coefficients of age-PV, tPSA-PV and fPSA-PV were 0.197, 0.434, and 0.446, respectively (all P<0.05). When tPSA was 0-4 μg/L, the AUC-ROC of tPSA and fPSA in predicting PV in the (30-50), (50-70) and >70 ml groups were (0.617, 0.732, 0.761) and (0.625, 0.738, 0.767), respectively. When tPSA was 0-4 μg/L, the optimal cut-off values of tPSA and fPSA in predicting PV in the (30-50), (50-70), and PV> 70 ml groups were (1.3 μg/L, 1.6 μg/L, 2.0 μg/L)and (0.3 μg/L, 0.4 μg/L, 0.5 μg/L), respectively.

The serum fPSA possesses the highest positive correlation with PV in the male LUTS/BPO patients in China. Serum tPSA and fPSA can be considered as independent biomarkers to predict the PV in this population, which can be used as simple and convenient predictors for PV in clinical practice. fPSA yields higher prediction accuracy compared with tPSA.