PSA screening cuts prostate cancer deaths by 32%, ERSPC data indicate

Milan, Italy-New evidence from the second-largest section of the European Randomized Study of Screening for Prostate Cancer (ERSPC) suggests that prostate cancer screening is beneficial, particularly in the core age group of 55 to 69 years.

Prostate cancer screening remains controversial, as groups such as the U.S. Preventive Services Task Force contend that the harms of PSA-based screening outweigh the benefits.

“Overall, too many men are being screened, too often, and confronted with over-diagnosis and surgical over-treatment. However, our current data suggests that the benefits of screening may be becoming more obvious,” said first author Monique J. Roobol, PhD, associate professor of urology at Erasmus University Medical Center, Rotterdam, the Netherlands, who worked with the ERSPC Study Group and presented the findings at the European Association of Urology annual congress in Milan, Italy.

“These new data show that systematic PSA-based screening reduces prostate cancer-specific mortality by 32% in the age range of 55 to 69 years. Starting screening in men over 70 years is, however, ineffective. The roughly twofold higher incidence of prostate cancer found in the screening arm of our investigation underlines the importance of identifying those men who would benefit from screening.”

The Rotterdam section of the ERSPC is the second largest ERSPC study. Investigators invited 88,283 men ages 54 to 74 years to enter the trial from 1993 to 1999. Active screening is still ongoing. The 42,376 men (34,833 men age 55-69 years, core age group) who gave their informed consent were then randomized into either the screening (21,210 men) or control arm (21,166 men).

ERSPC investigators screened participants every 4 years and used a PSA cutoff of ≥3.0 ng/mL (the cutoff was ≥4.0 ng/mL in 1993-‘97) and/or abnormal digital rectal exam/transrectal ultrasound as an indication for sextant biopsies. They obtained data on prostate cancer incidence and mortality through yearly linkages with the databases of the Cancer Registry and National Statistics and analyzed the data with respect to prostate cancer-specific mortality using Poisson regression.

After a median follow-up of 12.8 years, investigators screened 19,765 (94.2%) men at least once and detected 2,674 cases of prostate cancer, of which 561 (21%) were interval prostate cancers (those that emerge between two screenings). The positive predictive value (PPV) of PSA 3.0 ng/mL or higher was about 20% and stable over the four screening rounds. The overall cancer detection rate in the 21,210 men who were screened was 10.1%, which rose to 12.7% when including interval prostate cancer and prostate cancer diagnosed after age 74.

The PPV for PSA ≥10.0 ng/mL declined after the first screening, while the PPV in the PSA range of 3.0 to 9.9 ng/mL remained fairly constant. The rate of prostate cancer detected in the control arm was 6.8% (1,430 men), resulting in an excess incidence of 59 cases of prostate cancer per 1,000 men randomized (61 prostate cancers per 1,000 in age group 55-69 years).

Majority of cancers T1c

The majority of screen-detected cases were T1c cancers. The rate of T1c disease increased from 41.5% in the first screening to 71% in the fourth screening round. More advanced and high-grade prostate cancers were found among interval prostate cancers and prostate cancers detected in the control arm, Dr. Roobol noted.

The relative risk (RR) reduction for mortality in men aged 55 to 74 years was significant, at 20% (p=.042). In the core age group of 55 to 69, the RR reduction was 31.6% (p=.004). An adjustment for non-compliance in the screening arm resulted in an RR reduction of 33% in men ages 55 to 69 years. The RR reduction in the screened age group of 70 to 74 years was –14.1% (p=.504), indicating no benefit from screening.

“This is a very important new insight. We hope this study helps move us toward improving the harm/benefit ratio of screening for prostate cancer. Multivariate risk approaches, including long-term predictions on the basis of relevant clinical data and comorbidity, are currently the subject of many studies and seem the way forward,” Dr. Roobol said.UT