A0584
Title: A mixture model to estimate prostate cancer progression risk and indolent fraction in active surveillance
Authors: Chi Hyun Lee - Yonsei University (Korea, South) [presenting]
Yibai Zhao - Fred Hutch Cancer Center (United States)
Ruth Etzioni - Fred Hutchinson Cancer Research Center (United States)
Abstract: Active surveillance has become a widely accepted management strategy to mitigate overtreatment among low-risk prostate cancer patients. Prostate cancer is monitored through biopsies at scheduled visits to detect cancer upgrade to high-risk, leading to interval-censored outcomes when estimating the time to progression. In addition to interval censoring, patient heterogeneity presents further challenges: Some individuals may have undetected high-grade cancer at baseline due to imperfect diagnostics, while others may carry indolent cancers that never progress. All individuals are also subject to misclassification during follow-up due to biopsy imperfection. A mixture model is assumed for prevalent cases, progressive, and indolent cancers, where the proportional hazards model incorporates time-independent or time-varying covariates. A semiparametric likelihood-based approach is proposed to handle interval-censored observations while accounting for biopsy misclassification. Its performance is assessed via simulation and is applied to the Canary Prostate Active Surveillance Study to evaluate risk factors for progression and estimate the indolent fraction under varying biopsy sensitivity.
