Safety and clinical validity of tumor genomic testing for guiding active surveillance selection in men with NCCN intermediate-risk prostate cancer.

Abstract

336 Background: Clinicopathologic features are considered the gold standard for predicting prostate cancer (PCa) disease severity and guiding decisions on the use of active surveillance (AS) or definitive treatment (DT). However, tumor genomic testing has been shown to provide valuable information as an addition to clinical risk stratification measures, allowing for more accurate identification of AS candidates. This study evaluated the safety of genomic testing for guiding AS selection in patients with National Comprehensive Cancer Network (NCCN) intermediate-risk PCa. Methods: In this retrospective observational cohort study, patients with NCCN intermediate-risk PCa underwent Prolaris testing at diagnosis from 9/2015–12/2018. The Prolaris test combines a molecular cell cycle progression score with clinicopathologic features (USCF Cancer of the Prostate Risk Assessment [CAPRA] score, inclusive of Gleason score), generating a Combined Clinical Risk (CCR) score to identify patients as candidates for AS or DT. Patients with a CCR-based 10-year disease-specific mortality (DSM) risk ≤3.2% were recommended to AS. Clinical follow up data, including initial treatment selection and time to metastasis, were retrospectively collected from the treating clinics. Kaplan-Meier estimates were used to evaluate the safety of AS in patients who initially selected AS. Cox proportional hazards models were used to evaluate CCR as a predictor of metastasis in all patients. Results: In total, 3204 patients were included in this analysis, with 1468 (46%) recommended to AS and 1736 (54%) recommended to DT. Among the 973 (30%) patients who initially selected AS, 613 (63%) were recommended to AS and 360 (37%) were recommended to DT. In patients who were recommended to AS and initially selected AS, the observed risk of metastasis within 5 years of diagnosis was 0.37% (95% confidence interval [CI], 0.09%–1.47%). Patients who were recommended to AS had a significantly lower rate of metastasis than patients who were recommended to DT after accounting for initial treatment decision (AS vs DT) and CAPRA score (hazard ratio 4.20 [95% CI, 1.41–18.04], P-value = 0.03). Conclusions: For patients with NCCN intermediate-risk PCa who were identified by genomic testing as candidates for AS, AS is safe based on the very low observed metastasis risk. The CCR score is a strong predictor of metastasis beyond clinicopathologic factors.