The use of PSA in the diagnosis of prostate cancer is controversial. This is due to false-positive results caused by benign prostatic hyperplasia (BPH). Different forms of circulating PSA have recently been described. Initial studies indicate that the fraction of free PSA is lower in prostate cancer than in BPH, therefore suggesting that its measurement could be of some diagnostic value. We have assessed the serum value of the percentage of free/total PSA in the differential diagnosis of BPH and prostate cancer. The levels of PSA and free PSA (DELFIA) were measured in 145 BPH patients and 56 prostate cancer patients. Free PSA is a small fraction of PSA, and significantly lower levels are being found in prostate cancer. In those patients with a PSA level between 2 and 25 micrograms/l, the determination of the percentage of free PSA increased the diagnostic efficiency of PSA, while reducing the number of negative biopsies. We conclude that free PSA may be a useful marker for the diagnosis of prostate cancer.