The aim of the present study was to evaluate the clinical value of the pretreatment serum testosterone (T) level as a potential predictor of prostate cancer risk in screening for prostate cancer.The subjects were 420 patients suspected of having prostate cancer who underwent prostate biopsy, and whose pretreatment T levels were recorded. We checked for association between the presence of prostate cancer and the following clinical factors: pretreatment serum T level, age, pretreatment prostate-specific antigen (PSA) level, digital rectal examination findings, ratio of free to total PSA, prostate volume, and PSA density (PSAD).Overall, there was no significant difference in mean pretreatment T level between patients diagnosed with cancer (3.9±2.4ng/ml) and patients diagnosed with benign prostate disease (BPD; 3.7±1.7ng/ml); diagnosis was based on prostate biopsy. However, among patients with PSA <10ng/ml, the pretreatment T level was significantly higher in patients diagnosed with prostate cancer (4.2±2.6ng/ml) than in patients diagnosed with BPD (3.6±1.4ng/ml) (p=0.007); a similar trend was observed among patients with PSAD <0.15ng/ml/cc. Multivariate analysis indicated that pretreatment T level was an independent significant predictor of positive prostate biopsy (p=0.020). Additionally, the serum T level was significantly lower in patients with a Gleason score ≥7 (3.7±2.1ng/ml) versus a score <7 (4.2±1.7ng/ml) (p=0.030). Also, serum T levels were significantly higher in well-differentiated prostate cancer (4.8±2.1ng/ml) versus moderately differentiated (3.8±1.3ng/ml) or poorly differentiated (3.7±1.4ng/ml) (p<0.01).Among relatively low-risk patients, serum T level was an independent significant predictor of positive prostate biopsy, suggesting that the efficiency of prostate cancer screening can be improved by including measurement of serum T level.