Results of a randomized, population‐based study of biennial screening using serum prostate‐specific antigen measurement to detect prostate carcinoma

Abstract

BACKGROUND The purpose of the current study was to evaluate the effectiveness of a prostate carcinoma screening program in which serum prostate‐specific antigen (PSA) levels were measured. METHODS From a group of 20,000 men born between January 1, 1930, and December 31, 1944, 10,000 men were randomized into a screening group and 10,000 were randomized into a control group. Patients in the screening group were invited to undergo initial PSA testing between 1995 and 1996 and then were invited to receive testing every second year thereafter for 8 years (for a total of 4 PSA tests). Men with PSA levels ≥ 3 ng/mL (or ≥ 2.54 ng/mL, in the third and fourth screening rounds) were invited to undergo clinical investigation, which included sextant biopsy of the prostate. By linking to the regional cancer registry, the authors were able to obtain the true and expected incidence rates for the screening and control groups. RESULTS The screening participation rate was high (73%). A total of 884 malignancies have been detected to date, with 640 having been detected in the screening group. There was an early and marked shift toward more favorable disease stage and grade for malignancies detected on repeat screening. In the fourth screening round, only 2 of 82 detected malignancies were classified as advanced disease. Of the 227 screen‐detected tumors on which surgery was performed, only 20 (8.8%) had small volume (< 0.2 cm³). Forty‐three interval malignancies were detected, but only five were accompanied by symptoms. At 8 years, the cumulative disease incidence rate among screening participants was 7.3%, compared with 2.4% in the control arm. The incidence rate observed in the screening population corresponds to the cumulative incidence rate observed in the Swedish male population at age 72 years. CONCLUSIONS Biennial PSA screening was very successful in diagnosing prostate carcinoma at an early stage, when curative treatment typically is effective. In addition, the results regarding interval malignancies were favorable. Thus, decreased mortality should be observed on long‐term follow‐up. The lead time associated with screening appears to fall within the range described in earlier studies involving frozen sera (i.e., 5–9 years). Cancer 2004;100:1397–405. © 2004 American Cancer Society.