Objective: To determine whether prostate-specific antigen (PSA)-based screening alters the proportion of organ-confined prostate cancers detected.
Design: A prospective, nonrandomized, serial PSA-based screening trial (follow-up from 6 to 37 months), and a concurrent comparison group.
Setting: General community outpatient screening program based at a university center.
Patients: The study group consisted of 10,251 men aged 50 years and older (mean and median age, 63 years; mean and median age of patients who underwent biopsies, 66 years) who presented to a prostate cancer screening program and consented to phlebotomy. The comparison group consisted of 266 concurrently studied private patients in the same age range (mean and median age, 68 years) who were referred for prostatic ultrasonography and biopsy because of an abnormal digital rectal examination (DRE).
Main outcome measure: Proportion detected with clinically or pathologically advanced prostate cancer.
Results: The men were divided into three groups: the comparison group, the initial PSA-based screening group, and the serial PSA-based screening group. The proportions of prostate cancers detected that were clinically or pathologically advanced were as follows: comparison group, 57% (27/47); initial PSA-based screening group, 37% (91/244); and serial PSA-based screening group, 29% (37/129). Screened patients had a lower proportion of advanced cancers than the comparison group (chi 2 [2] = 12.3; P = .002); this advantage was observed principally in patients younger than 70 years. Surgical staging revealed that the cancer was microscopically focal and well differentiated (possibly latent cancer) in 2.5% (1/40) of the nonscreened group, 2.9% (7/244) of the initially screened group, and 7.8% (10/129) of the serially screened group (generalized Fisher's Exact Test, P = .08).
Conclusion: Screening based on PSA identifies some men with prostate cancer who have a significantly increased proportion of organ-confined tumors compared with those detected through evaluation for an abnormal DRE alone.