To the Editor.—The article by Dr Catalona and colleagues1 claims that the free PSA assay can reduce the number of biopsies performed in men with normal DRE findings and total PSA levels between 4.0 and 10.0 ng/mL, while missing only a minimal number of patients with prostate cancer. However, a critical examination of their data shows that the test has significant weaknesses and in fact is almost worthless.
Quantitative evaluation of the data reveals how many patients would be missed by a test that is 95% sensitive. The article by Catalona et al shows that the initial screen (normal DRE findings, total PSA of 4.0-10.0 ng/mL) identifies a population of men with a 25% prevalence of prostate carcinoma. Therefore, a hypothetical group of 400 such men would include 100 men with prostate carcinoma and 300 disease-free men. The free PSA test, used at a cutoff of 25% as recommended by the article, has a sensitivity of 95% and a specificity of 20%. At 95% sensitivity, the test would correctly identify 95 of the 100 patients with cancer, missing 5 of them. At 20% specificity, only 60 of the 300 unaffected patients would be classified as not needing biopsies (Table 1).