PSA is the most useful tumor marker for the diagnosis and treatment of prostate cancer. Its clinical use, however, still lacks the necessary sensitivity and specificity to be considered as ideal. In fact PSA is not specific for adenocarcinoma of the prostate: an elevated serum level of the marker does not necessarily mean malignant growth and normal levels too often hide an occult and potentially lethal cancer. With the discovery of different molecular PSA forms in the serum, an improved discrimination between benign prostate hyperplasia (BPH) and prostate cancer appears possible. This may be particularly useful in cases with equivocal PSA values and unpalpable prostate neoplasm to reduce the number of unnecessary prostate biopsies and increase the number of biopsies in cases without palpable or ultrasonic visible anomalies. The clinical use of PSA age-referenced levels is discussed. Their use is invaluable in screening programs where the routine adoption of age-specific values can help to pick-up younger patients with potentially curable prostate cancer and older patients with BPH where additional tests would be unnecessary. The role of PSA velocity (PASAV) s also discussed. An elevation rate of PSA of 0.75 ng/ml over 18 months on 3 serial samples appears to be the best cut-off to distinguish BPH from prostate cancer. However, the use of PSAV seems to be less useful in patients with elevated PSA levels and negative biopsy results. Free to total PSA ratio is probably the best parameter to reduce the number of unnecessary biopsies in men with a serum total PSA of 4 to 10 ng/ml. The advantages and limitations for different levels of cut-off are shown. A flow chart illustrating the role of various PSA "derivatives" in screening and subsequent evaluation of men over 50 years of age is also presented.