Additional value of the ratio of serum total testosterone to total prostate-specific antigen in the diagnosis of prostate cancer in a Chinese population

We investigated whether serum testosterone and testosterone/prostate-specific antigen ratio (T/PSA) might be prostate cancer (PCa) biomarkers. We retrospectively reviewed 92 patients with benign prostatic hyperplasia (BPH) and 164 patients with PCa treated at Zhongshan Hospital, China (April 2012 to November 2013). The BPH and PCa groups had similar serum total testosterone (median, 15.8 versus 16.3nmol/L).Compared with the BPH group, the PCa group had higher PSA (16.8 versus 5.1ng/ml) and lower free/total PSA (9.5% versus 19.3%) and T/PSA (1.37 versus 4.69) (all p<.001).Patients with PCa and PSA 20ng/ml had higher testosterone (17.5 versus 12.9ng/ml; p=.002) and T/PSA (2.24 versus 0.29; p<.001) than those with PSA >20ng/ml. Patients with PCa and Gleason score 7 had higher testosterone (18.3 versus 14.1ng/ml; p=.023) and T/PSA (1.93 versus 0.72; p<.001) than those with Gleason score >7. In patients with PSA 20ng/ml, T/PSA was higher in those with BPH than in those with PCa (4.69 versus 2.24; p<.001). ROC curve analysis yielded an AUC of 0.712; for the optimal cut-off of 4.43, specificity and sensitivity were 52% and 97% respectively. T/PSA may improve the accuracy of PCa diagnosis in patients with a PSA level 20ng/ml.