EAU 2018: Development of PSA Screening Nomograms for 15-Year Prediction of Prostate Cancer Diagnosis, Mortality, and All-Cause Mortality

Copenhagen, Denmark (UroToday.com)  Dr. Carlsson presented their group’s work developing PSA screening nomograms for 15-year prediction of prostate cancer diagnosis, mortality and all-cause mortality. Recommendations for PSA screening encourage shared decision making between patients and providers, rather than population-based screening. Nomograms provide individualized risk predictions potentially enhancing informed decisions to undergo screening. The objective of this study was to develop PSA screening nomograms using demographic and follow-up data from two large randomized screening trials: the Göteborg Screening Trial [1] and Prostate, Lung, Colon, and Ovarian (PLCO) Cancer Screening Trial [2].

For this study, patients with screening prior to trial entry were excluded, leaving 59,951 total subjects, which comprised 19,899 from Goteborg and 40,052 from PLCO. Demographic information including age, family history of prostate cancer diagnosis, ethnicity, Charlson Comorbidity Index, marital status, education, and control vs. screening arm were used as predictive variables. If screened, baseline PSA was used as an additional continuous variable. Models for all-cause mortality were first developed, then models predicting prostate cancer diagnosis and prostate cancer mortality were developed using competing risk analysis. Predictive accuracy was assessed using the concordance index (c-index) and calibration plots.

The authors found that on multivariable analysis, all variables were significant (p<0.05) for predicting at least one outcome. Nomograms predicting prostate cancer diagnosis, prostate cancer mortality, and all-cause mortality demonstrated reasonable discrimination: c-index 0.60, 0.65, 0.70, respectively. When a single baseline PSA was included, discriminative accuracy improved for prostate cancer diagnosis and prostate cancer mortality, but remained the same for all-cause mortality: c-index 0.78, 0.74, 0.69, respectively. Estimates were well-calibrated for all endpoints.

Dr. Carlsson and colleagues concluded that using demographic information allowed development of PSA screening nomograms, with good discrimination and calibration in predicting all three outcomes. A single baseline PSA improved nomogram discrimination substantially for prostate cancer diagnosis and prostate cancer mortality. The authors plan on studying the implementation of these nomograms into clinical practice to aid decisions for previously unscreened patients, or whether to continue PSA screening for patients with a known prior PSA.


Presented by: S Carlsson, MD Cleveland Clinic Foundation, Cleveland, OH, USA
Co-Authors: Brooks M, Zajichek A, Chagin K, Hugosson J, Kattan M, Stephenson A

Written by: Zachary Klaassen, MD, Urologic Oncology Fellow, University of Toronto, Princess Margaret Cancer Centre, twitter: @zklaassen_md at the 2018 European Association of Urology Meeting EAU18, 16-20 March, 2018 Copenhagen, Denmark

References:

1. Hugosson J, Carlsson S, Aus G, et al. Mortality results from the Goteborg randomized population-based prostate-cancer screening trial. Lancet Oncol 2010;11(8):725-732.
2. Andriole GL, Crawford ED, Grubb RL 3rd, et al. Mortality results from a randomized prostate-cancer screening trial. N Engl J Med 2009;360:1310-1319.