Purpose: We investigate the validity of prostate specific antigen (PSA) as a screening test for prostate cancer.
Materials and methods: A registry of serum samples drawn from 1968 to 1976 from 21,387 men was linked to the Finnish Cancer Registry. During followup from 1968 to 1991, 104 prostate cancers were identified. A matched case control design with incidence density sampling and nested in the serum sample bank was applied, and PSA was assessed.
Results: The estimated sensitivity of the test was 44% and specificity 94% at a cutoff of 4.0 microg./l. in the total material. The sensitivity had improved to 86% in patients diagnosed in 5 years after the sample drawing. The test had a better sensitivity (93%) and specificity (96%) in men younger than 65 years at the time of the sample drawing compared to those older. The sensitivity further improved to 100% with a cutoff of 2.5 microg./l.
Conclusions: PSA is a valid screening test for prostate cancer, which compares favorably with mammography for breast cancer. However, until an effect on mortality has been shown, routine screening cannot be recommended.