Let's Not Throw the Baby out With the Bathwater in Prostate Cancer Screening

Abstract

In 2012, the United States Preventive Services Task Force recommended against prostate-specific antigen (PSA) –based screening for prostate cancer (grade D recommendation) primarily on the basis of the results of two large randomized clinical trials: the US Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial and the European Randomized Study of Screening for Prostate Cancer (ERSPC). In the article accompanying this editorial, Shoag et al present their analysis of questionnaire data sent to a subset of participants in the prostate component of the PLCO study. They again illustrate the high contamination rate in the control arm of the study and call into question the internal validity of PLCO and, ultimately, whether the negative study finding can be used to inform the discussion around the effectiveness of PSA-based screening. The United States Preventive Services Task Force is currently in the process of updating their recommendation concerning PSA-based screening. Given the data presented by Shoag et al, the panel will have to seriously reconsider the validity of the PLCO trial in their deliberations. However, even if they were to completely discount PLCO, it is not entirely clear that the magnitude of benefit currently seen in the ERSPC trial will be enough to support a change in the recommendation. The latest report from ERSPC indicates that, with 13-year follow-up, 27 cases of prostate cancer need to be detected (NND) to prevent one prostate cancer–related death (NND). Although modeling studies extrapolating follow-up in ERSPC out to 25 years or more indicate that NND ultimately may be less than 10, these simulation studies are dependent on myriad assumptions, which may or may not prove to be true. To this end, the task force may choose to only consider the actual randomized clinical trial evidence collected from ERSPC to date, which currently indicates a small to moderate benefit at best. The documented (albeit small) benefit of PSA screening identified in ERSPC must then be weighed against the potential harms of screening. For example, an elevated PSA test often leads to a prostate biopsy. In a recent study of 1,147 men undergoing prostate biopsy in the Prostate Testing for Cancer and Treatment study, 31.8% reported grade 2 adverse events (defined as symptoms causing moderate/major problems or requiring contact with the health care system) and 1.4% required hospital admission within 30 days. Although no deaths were noted in the cohort, these adverse events undoubtedly resulted in discomfort and anxiety, with at least a temporary negative impact on overall health. Furthermore, in patients in whom prostate cancer is diagnosed and treated, both surgery and radiation are associated with a considerable risk of sexual, urinary, or bowel problems. This may be acceptable in patients with higher-risk disease, who are likely to experience a survival benefit from aggressive treatment. Given the widely recognized risk of overdiagnosis associated with population-based annual PSA screening, patients diagnosed with low-risk clinically indolent disease who elect aggressive therapy are unlikely to garner any survival advantage and are, therefore, at considerable risk of overtreatment. Recent data demonstrating increased use of active surveillance in American men diagnosed with low-risk disease is reassuring, however, as we attempt to optimize the balance of predicted benefit associated with prostate cancer treatment and the predicted morbidity of therapy. Increasing penetration of active surveillance will likely mitigate some of the risk of overtreatment, but, given ongoing challenges associated with accurately classifying indolent and clinically relevant disease, it certainly does not eliminate it. If one weighs the benefits of population-based screening seen in ERSPC (completely dismissing PLCO) against the potential harms noted in these and other studies, it is not completely unreasonable to conclude that the benefits of population-wide annual PSA screening do not outweigh the harms. Does this mean that we should abandon PSA-based screening altogether? Given the benefit seen in ERSPC, the answer is clearly no. Rather, we should focus on identifying better screening strategies that optimize the risk-benefit ratio by personalizing the approach to the target population’s underlying risk of disease. Consider that the ERSPC consisted of eight unique sites that used eight similar but distinct protocols for prostate cancer screening that varied by age of initiation, screening interval, and threshold for diagnostic biopsy. Of these eight sites, only two (Sweden and the Netherlands) noted a statistically significant difference in prostate cancer–specific survival, although the magnitude of observed differences and the size of these cohorts were sufficient to drive the results of the parent study. The observed heterogeneity in outcomes between ERSPC sites is likely the result of variation in the intensity and delivery of the intervention (PSA screening) superimposed on differences in underlying risk of disease between