Presentation Authors: Zachary T. Kornberg*, Matthew R. Cooperberg, Janet E. Cowan, June M. Chan, Jeffry P. Simko, Imelda Tenggara, Peter R. Carroll, San Francisco, CA
Introduction: The OncotypeDXÂ® Genomic Prostate Score (GPS) test is an RNA expression assay that can be performed on needle-core biopsies from men with prostate cancer (PCa). GPS has previously been validated as a predictor of adverse pathology in men with low-risk prostate cancer who undergo primary radical prostatectomy (RP). We sought to determine whether GPS was associated with increased risk of adverse pathology for men enrolled on active surveillance (AS) who later underwent RP.
Methods: Of 1,662 men enrolled on AS at the University of California San Francisco (UCSF) who consented for prospective data collection, we evaluated 215 men on AS with Gleason score (GS) 3+3 and low volume (â‰¤33% positive cores) GS 3+4 PCa who underwent GPS testing at diagnostic or confirmatory biopsy (ie. within 24 months). Patients had stage T1 or T2 disease, PSA < 20, and clinical Cancer of the Prostate Risk Assessment (CAPRA) score < 6. The primary outcome was adverse pathology at delayed RP, defined as GS â‰¥ 4+3, and/or stageâ‰¥ pT3a or pN1. We performed Cox proportional hazards regression with inverse probability censored weights (IPCW) to evaluate association between GPS and adverse pathology, adjusting for age at diagnosis, clinical CAPRA at diagnosis, PSA density at time of GPS, whether the patient had an additional GPS test, and whether the biopsy was performed at UCSF.
Results: 83 percent of men (N=179) were low risk, and 17 percent of men (N=36) were intermediate risk by CAPRA scoring. Median GPS was 26.4 (interquartile range [IQR]: 18.8, 34.6). Median time from diagnosis to RP was 23 months (IQR: 15, 40). 121 men had adverse pathology on delayed RP at a median time of 26 months (IQR 16, 43) to prostatectomy. In a multivariate proportional hazards model with IPCW, GPS was associated with increased risk of adverse pathology at delayed RP (Hazard Ratio [HR] per 5 units: 1.11, 95 Confidence Interval [CI]: 1.02, 1.21, p=0.01). Age at diagnosis was the only other covariate associated with adverse pathology.
Conclusions: In patients who undergo radical prostatectomy after a relatively short period on active surveillance, a higher Genomic Prostate Score is associated with increased risk for adverse pathology.
Source of Funding: Department of Defense Transformative Impact Award W81XWH-13-2-0074