PSA screening reduces deaths by 29% in large trial

Atlanta-A final analysis of data after 11 years of follow-up in the European Randomised Study of Screening for Prostate Cancer (ERSPC) shows a highly statistically significant benefit of PSA screening for reducing prostate cancer mortality.

The recently published results (N Engl J Med 2012; 15:366:981-90) confirm data from an interim analysis for 9 years of follow-up (N Engl J Med 2009; 360:1320-8). However, with longer follow-up, the benefits of screening increased, according to data presented at a late-breaking session at the AUA annual meeting in Atlanta.

In the final analysis, there were 462 prostate cancer deaths in the control arm, representing 8.6% of detected cancers and 0.52% of all control men. In the screening arm, there were 299 prostate cancer deaths, translating to 4.3% of detected cancers and 0.4% of the men.

The absolute risk reduction of prostate cancer mortality by screening was 1.07/1,000 men randomized, up 30% from the earlier analysis, and there were also reductions from the interim to the final analysis in both the number of men needed to invite to screen to avoid one man dying from prostate cancer (1,410 to 936) and the number needed to diagnose (48 to 33).

All of these numbers are likely to change with additional follow-up. Nevertheless, longer follow-up is needed to assess the effect of PSA-based screening, said Dr. Roobol.

'Screening needs to be individualized'

"It is obvious from the ERSPC data that some men will benefit from PSA-based screening, but some men will not and might be harmed due to unnecessary testing, overdiagnosis, and overtreatment of potentially indolent disease. It is also obvious that screening on the basis of the PSA test needs to be individualized. In other words, there is no indication for a population-based program," she commented.

"Currently, the situation for men who consider having a PSA test remains unchanged. The outcome needs to be balanced against the number of screens, biopsies, and potential risks of overdiagnosis. Men who decide to be PSA tested should make this decision on the basis of balanced information."

The ERSPC is a multicenter trial including eight centers and randomizing men ages 50 to 75 years to screening or a control arm. The final analyses excluded data from one center where there was short follow-up.

There were 182,160 men randomized, including 162,388 within a predefined core age group of 55 to 69 years; 72,891 men were in the screening group and 89,352 were randomized to control (145 patients died during the randomization process). PSA screening was conducted every 2 years at one center and otherwise every 4 years. In the final analysis, the average number of screening tests per patient was 2.3.

According to the protocol, lateralized sextant biopsy would be performed in men in the screening group whose PSA was ≥3.0 ng/mL. Biopsy had a positive predictive value of 24%.

In the final analysis, the incidence of prostate cancer in the screening and control arms was 9.6% and 6.0%, respectively, both up from the previous analysis (8.2% and 4.8%, respectively). The excess incidence of prostate cancer in the screening versus control arm was 35 per 1,000 men, nearly the same as previously reported, and due mainly to detection of low-risk cancers (T1-T2 with Gleason 6), which represented 60% of screening group prostate cancers and 42% of those in the control arm.

Incidence of cancers that were high risk (any stage with Gleason 8-10 or T4) or advanced at diagnosis was lower in the screening group than in the control arm (10% vs. 20.4%, respectively).

"Now, we are continuing to work on the basis of the unique ERSPC dataset. For the study group, the question is not if to screen, but how and who should be screened," Dr. Roobol said.