Risk-Based Selection for Active Surveillance: Results of the Movember Foundation's Global Action Plan Prostate Cancer Active Surveillance (GAP3) Initiative
Sophie M Bruinsma,Daan Nieboer,Monique J Roobol,Chris H Bangma,Jan F M Verbeek,Vincent Gnanapragasam,Mieke Van Hemelrijck,Mark Frydenberg,Lui-Shiong Lee,Riccardo Valdagni,Christopher Logothetis,Ewout W Steyerberg,the Movember Foundation's Global Action Plan Prostate Cancer Active Surveillance (GAP3) Consortium,Bruce Trock,Behfar Ehdaie,Peter Carroll,Christopher Filson,Jeri Kim,Todd Morgan,Laurence Klotz,Tom Pickles,Eric Hyndman,Caroline Moore,Prokar Dasgupta,Chris Bangma,Monique Roobol,Arnauld Villers,Antti Rannikko,Antoinette Perry,Jonas Hugosson,Jose Rubio-Briones,Anders Bjartell,Lukas Hefermehl,Lee Lui Shiong,Yoshiyuki Kakehi,Mikio Sugimoto,Byung Ha Chung,Theo van der Kwast,Henk Obbink,Wim van der Linden,Tim Hulsen,Cees de Jonge,Mike Kattan,Ji Xinge,Kenneth Muir,Artitaya Lophatananon,Michael Fahey,Ewout Steyerberg,Liying Zhang,Wei Guo,Nicole Benfante,Janet Cowan,Dattatraya Patil,Emily Tolosa,Tae-Kyung Kim,Alexandre Mamedov,Vincent LaPointe,Trafford Crump,Vasilis Stavrinides,Jenna Kimberly-Duffell,Aida Santaolalla,Jonathan Olivier,Tiziana Rancati,Helén Ahlgren,Juanma Mascarós,Annica Löfgren,Kurt Lehmann,Catherine Han Lin,Hiromi Hirama,Kwang Suk Lee,Guido Jenster,Anssi Auvinen,Masoom Haider,Kees van Bochove,Ballentine Carter,Sam Gledhill,Mark Buzza,Michelle Kouspou,Sophie Bruinsma,Jozien Helleman
DOI: https://doi.org/10.1097/JU.0000000000001700
Abstract:Purpose: We sought to identify and validate known predictors of disease reclassification at 1 or 4 years to support risk-based selection of patients suitable for active surveillance. Materials and methods: An individual participant data meta-analysis using data from 25 established cohorts within the Movember Foundations GAP3 Consortium. In total 5,530 men were included. Disease reclassification was defined as any increase in Gleason grade group at biopsy at 1 and 4 years. Associations were estimated using random effect logistic regression models. The discriminative ability of combinations of predictors was assessed in an internal-external validation procedure using the AUC curve. Results: Among the 5,570 men evaluated at 1 year, we found 815 reclassifications to higher Gleason grade group at biopsy (pooled reclassification rate 13%, range 0% to 31%). Important predictors were age, prostate specific antigen, prostate volume, T-stage and number of biopsy cores with prostate cancer. Among the 1,515 men evaluated at 4 years, we found 205 reclassifications (pooled reclassification rates 14%, range 3% to 40%), with similar predictors. The average areas under the receiver operating characteristic curve at internal-external validation were 0.68 and 0.61 for 1-year and 4-year reclassification, respectively. Conclusions: Disease reclassification occurs typically in 13% to 14% of biopsies at 1 and 4 years after the start of active surveillance with substantial between-study heterogeneity. Current guidelines might be extended by considering prostate volume to improve individualized selection for active surveillance. Additional predictors are needed to improve patient selection for active surveillance.