PD08-04 VALIDATION OF A RISK CALCULATOR PREDICTING BIOPSY OUTCOME IN PROSTATE CANCER TREATED WITH ACTIVE SURVEILLANCE

Abstract

You have accessJournal of UrologyProstate Cancer: Localized: Active Surveillance III1 Apr 2016PD08-04 VALIDATION OF A RISK CALCULATOR PREDICTING BIOPSY OUTCOME IN PROSTATE CANCER TREATED WITH ACTIVE SURVEILLANCE Daan Nieboer, Monique J. Roobol, Ewout W. Steyerberg, and Chris H. Bangma Daan NieboerDaan Nieboer More articles by this author , Monique J. RoobolMonique J. Roobol More articles by this author , Ewout W. SteyerbergEwout W. Steyerberg More articles by this author , and Chris H. BangmaChris H. Bangma More articles by this author View All Author Informationhttps://doi.org/10.1016/j.juro.2016.02.2820AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookTwitterLinked InEmail INTRODUCTION AND OBJECTIVES Prostate cancer patients on active surveillance (AS) are followed using repeat prostate biopsies (Bx), a burdensome intervention not without risks for patients. Most Bx, indicated by clinical indicators like PSA change and/or pre-defined time since diagnosis, do not show disease progression or risk reclassification. Identifying patients at low risk of progression or reclassification may reduce the number of these unnecessary Bx. Our objective was to validate the Canary-EDRN active surveillance biopsy risk calculator, which estimates the probability of upgrading or reclassification on biopsy (gleason grade≥7 or more than 34% of biopsy cores positive for cancer) using the patients age, months since last biopsy, last observed PSA level, percentage of cores positive for cancer on last biopsy and the number of prior negative biopsies. METHODS We used data from the Movember Global Action Plan (GAP-3) project, a global consortium of AS studies comprising of more than 10,000 PC patients. Patients were included from the prostate cancer research international: active surveillance (PRIAS), an American cohort (from Johns Hopkins), and a Canadian cohort (from Sunnybrook). These cohorts contained 5,129 follow-up biopsies in 3,412 PC patients. Calibration of the risk calculator was assessed graphically and the discriminative ability was quantified using the area under the receiver operating characteristic curve (AUC). Clinical usefulness was assessed using decision curves analysis (DCA). RESULTS The AUC at external validation was 0.64 in PRIAS, 0.73 in the American cohort and 0.66 in the Canadian cohort. The probability of upgrading or risk reclassification was somewhat over-estimated in PRIAS and the Canadian cohort and somewhat under-estimated in the American cohort. DCA showed that the model was clinical useful in each cohort for risk thresholds between 10% and 30%. CONCLUSIONS The risk calculator provided moderate discrimination in 3 cohorts of the GAP-3 consortium, but showed some miscalibration. Before applying the risk calculator in clinical practice recalibration may be required. © 2016FiguresReferencesRelatedDetails Volume 195Issue 4SApril 2016Page: e229-e230 Advertisement Copyright & Permissions© 2016MetricsAuthor Information Daan Nieboer More articles by this author Monique J. Roobol More articles by this author Ewout W. Steyerberg More articles by this author Chris H. Bangma More articles by this author Expand All Advertisement Advertisement PDF downloadLoading ...