Longitudinal screening for prostate cancer with prostate-specific antigen.

OBJECTIVE: To determine for the first 4 years of serial prostate-specific antigen (PSA)-based screening trends in compliance, prevalence of abnormal screening test results, cancer detection rates, and stage and grade of cancers detected.
DESIGN: A community-based study of serial screening for prostate cancer with serum PSA measurements.
SETTING: University medical center.
SUBJECTS: A total of 10 248 male volunteers at least 50 years old who were screened at 6-month intervals for a minimum of 48 months. MAIN OUTCOME MEASURES: The proportion of men who returned for serial screening, the proportion with elevated PSA levels, the proportion with newly elevated PSA levels, prostate cancer detection rates, and the distribution of tumor stages and grades at diagnosis.
RESULTS: At 48 months, 79% of volunteers returned for screening. During this interval there was a decrease in the proportion of volunteers with serum PSA levels higher than 4.0 ng/mL (from 10% to 6%-7%), in cancer detection rates (from 3% to <1%), and in the proportion with clinically advanced cancer (from 6% to 2%). In men who underwent surgery, the proportion with high-grade cancer decreased (from 11% to 6%), and the proportion with pathologically advanced cancer was proportionately reduced but not significantly reduced (from 33% to 27%).
CONCLUSIONS: With serial PSA-based screening, the proportion of men with abnormal test results decreased, and the prostate cancer detection rate decreased to near the reported population-based incidence rate. There was also a shift to detection of cancers at an earlier clinical stage and detection of lower-grade cancers. If PSA screening is ultimately shown to be beneficial, then appropriate cost-benefit analyses will be required to determine how the shifts in cancer detection rate and cancer stage will affect the economics of serial PSA-based screening.