Prostate specific antigen (PSA) has become essential to the follow-up of radical treatment for T1-T2 tumours. Various assays are available, but require a correlation coefficient to homogenize their results. PSA is probably the most reliable marker for the follow-up of radical prostatectomy (RP), as this operation should make PSA undetectable after 3 weeks. Highly sensitive tests, with a limit of detection of 0.1 ng/ml, allow the earlier laboratory detection of tumour escape (20 to 45%). Anastomotic biopsies are positive in 35 to 50% of cases. Seminal vesicle invasion and positive resection margins are more frequently associated with recurrence. The doubling time and rate of progression of PSA after RP can be used to distinguish local recurrence from metastasis. Urinary PSA is not useful in the follow-up of RP, as it is secreted by the periurethral glands. The use of the PSA after radical radiotherapy is less clearly established, as this treatment is not designed to eliminate all prostatic tissue or render PSA undetectable. Therapeutic efficacy is situated between 1 and 1.5 ng/ml according to the tests and is achieved in approximately 40% of cases after 4 years. A PSA level greater than 3 ng/ml at 3 months is indicative of a poor prognosis. Prospects for the future include the use of highly sensitive assays and reverse transcriptase polymerase chain reaction (RT-PCR) to detect circulating prostatic cells. The use of PSA has led to a re-evaluation of the efficacy of radical treatments and could influence the indications for adjuvant treatments.