Early detection of prostate cancer: the nature of cancers detected with current diagnostic tests.

Overview

The goal of an early detection program is not to detect all prostate cancers, but only to detect those that are potentially morbid or lethal (clinically important cancers). The prevalence of such cancers in the population of 50-year-old men can be estimated to be about 6%, using both epidemiologic and pathologic data. Screening trials using PSA and/or TRUS instead of or in addition to DRE have approached this rate of detection. As Fig 1 illustrates, screening appears to detect almost all of the clinically important cancers in the study population. The fear that these tests will detect a high proportion of latent, or clinically unimportant, cancers appears to be unfounded. Our studies and those of others support the concept that 85% to 90% of cancers detectable with current tests are clinically important. The detection of nonpalpable cancers by PSA or TRUS reduces the proportion of cancers detected at an advanced stage without increasing significantly the detection of latent or clinically unimportant cancers. Despite this analysis, we still lack definitive data to show that screening and treating early stage cancers will decrease the mortality rate of this disease. Such a conclusion will require a long-term, randomized screening trial or a comparison of mortality rates among large populations of men differing primarily in the extent of screening for prostate cancer.