Development of an International Prostate Cancer Risk Tool Integrating Data from Multiple Heterogeneous Cohorts

• Activity Number: 296 - SPEED: Biometrics - Methods and Application, Part 1
• Type: Contributed
• Date/Time: Tuesday, July 30, 2019 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Donna Ankerst* and Johanna Tolksdorf
• Companies: Technical University of Munich and Technical University of Munich
• Keywords: risk model; prediction; random-effects models; logistic regression; heterogeneous data; multiple cohorts

Abstract

The Prostate Biopsy Collaborative Group (PBCG) collects prostate cancer risk factors and biopsy outcomes from multiple international centers with the objective to improve biopsy decision-making. We recently developed a global risk prediction model for outcomes on prostate biopsy, and posted it online for worldwide use. Construction of the logistic regression model required integrating data from 8492 biopsies from ten diverse centers across North America and Europe. It involved decisions on choice of model, for example, whether to include random effects, as well as the detection of outlying cohorts. We outline here the data visualization and cohort-permutation internal validation strategy used to develop the PBCG model, the utility of which makes it applicable for future multi-cohort modeling applications.