In this study our aim was to investigate the efficacy of free to total PSA ratio in discrimination of benign prostate hyperplasia and prostate cancer. A total of 194 patients, 52 to 82 years old (mean 66.06 +/- 0.47 years) with PSA levels between 4 to 20 ng/mL were included into this study. Each patient underwent sextant prostate biopsy under transrectal ultrasound guidance. The patients were divided into two groups as PSA 4-10 and 10-20 ng/mL. Patients with benign and malign results were compared with respect to age, total PSA level, free PSA level and free/total (f/t) PSA ratio. Biopsies revealed prostate cancer in 16 of 130 patients (12.3%) with serum PSA 4-10 ng/mL and in 10 of 64 patients (15.6%) with serum PSA 10-20 ng/ml. In both PSA groups free PSA and f/t PSA levels were statistically significant, where total PSA levels were not. In patients with 4-20 ng/mL total PSA levels and a cut off level of < 0.18 for f/t PSA, the sensitivity, specificity and positive predictive value for prostate cancer were 88.5%, 53.6% and 20.4% respectively. Higher levels of PSA suggest prostate cancer, but still additional parameters are needed for patients with PSA 4-20 ng/mL, such as free PSA and f/t PSA. Although a cut off level of < 0.18 for f/t PSA seems to be the most accurate one to discriminate benign and malign diseases further studies on larger groups of patients are needed.