Firas Abdollah

MD

Henry Ford Hospital

Senior Fellow

PLCO (Learn more about this study)

PLCO-226

Aug 31, 2016

A RISK-STRATIFIED APPROACH TO PROSTATE CANCER SCREENING IN THE PROSTATE, LUNG, COLORECTAL AND OVARIAN TRIAL PARTICIPANTS

Screening for prostate cancer (PCa) has been a matter of debate over the past decade. Although the initiation of widespread prostate specific antigen (PSA) testing in the early 1990s has resulted in a dramatic drop in incidence of metastatic disease, its systematic use in men ≥50 could also favor the detection of indolent tumors, leading to some extent of overdiagnosis and overtreatment. Interestingly, the contemporary role of PSA screening has been examined in two landmark randomized controlled trials.
On one hand, the Prostate, Lung, Colorectal and Ovarian (PLCO) screening trial showed no impact of annual PSA testing and digital rectal examination on cancer-specific survival in US men. On the other hand, PCa screening using variable combinations of PSA testing and digital rectal examination, was associated with a cancer-specific survival benefit in the European Randomized Study of Screening for Prostate Cancer (ERSPC). However, 1410 men would need to be screened and 48 additional cases of prostate cancer would need to be treated to prevent one death from prostate cancer, which may have limited clinical interest.
Both trials have several strengths and limitations that were weighted one against each other, before the United States Preventive Services Task Force (USPSTF) released a first recommendation against PCa screening in men ≥ 75 in 2008; this extended to all men in 2012 through a second statement. The long-term consequences of such health-care policy decisions remain currently unknown but a significant rise in incidence of advanced and metastatic disease is logically expected within the next 10 years.
Although the clinical interest of a population-based screening for PCa is questionable according to the results from PLCO and ERSPC trials, we hypothesized that a more individualized approach could result in a significant cancer-specific survival benefit. To achieve our aim, we intend to use the PLCO dataset to develop a risk calculator including all available baseline characteristics to identify the men who are more likely to benefit from PCa screening in the US. Specifically, we will assess the impact of all baseline clinical characteristics on the screening effect by testing an interaction between the predicted risk of cancer-specific mortality and the receipt of PSA testing +/- digital rectal examination. The predicted risk of cancer-specific mortality will be calculated for each patient in both groups by using a multivariable Cox model including all available clinical covariates. This model will be developed in the non-screening group to establish a baseline predicted risk of cancer-specific mortality. Interaction between the predicted risk of cancer-specific mortality and the receipt of screening will be further graphed using a locally-weighted curves method, which could help to determine the threshold risk above which men may benefit from PCa screening.

To develop a risk calculator including all available baseline PLCO characteristics to identify the best candidates for individualized PCa screening in the US

Next, follow up studies have shown that the control arm of the PLCO did have a significant rate of ‘contamination’: a large proportion of men in the unscreened arm also underwent some form of PSA screening, artificially diluting the differences between the two arms. We wish to ascertain the true magnitude of the survival benefit (if any) associated with PSA screening by comparing the population that underwent PSA screening vs. no screening, treating men who underwent opportunistic screening (in the control arm) as the screened population.

Further, recent large scale RCTs (such as the ProTecT and the PIVOT studies) have also questioned the role of definitive treatment in the setting of localized/early stage prostate cancer. While these studies level 1 evidence with high internal validity, in order to be clinically relevant, their findings should be externally valid as well. As such, we wish to compare the demographics, tumor staging, treatment and outcomes for men diagnosed with prostate cancer in PLCO, to those meeting the inclusion criteria of these studies.

Thomas Seisen
Patrick Karabon
Deepansh Dalela
Akshay Sood
Mani Menon