Abstracts that were granted an exception in accordance with ASCO's Conflict of Interest Policy are designated with a caret symbol (^).
Background: Using clinical parameters, prostate cancers can be risk-stratified to aid treatment decisions. However, novel biomarkers are needed to predict tumor behavior more accurately. We aimed to validate the ability of a previously described 46-gene expression panel consisting of 31 cell-cycle progression (CCP) genes and 15 “housekeeping” genes to predict recurrence in a contemporary cohort of men undergoing radical prostatectomy (RP). Methods: Patients undergoing RP at UCSF since 1994 with ≥5 years of follow-up were included. RNA was extracted from archival RP tissue, and CCP score was calculated as average expression of the 31 CCP genes. Recurrence was defined as 2 PSAs ≥0.2 ng/ml or any salvage treatment at least 6 months after surgery. CCP score association with recurrence was assessed in univariate survival analysis and adjusted for PSA at diagnosis, pathologic Gleason score, margin status, extracapsular extension, seminal vesicle invasion, lymph node invasion. These variables were considered as separate covariates and as measured by the validated CAPRA-S score. Results: 413 men were included, 82 recurred. Most had low- to intermediate-risk pathology (mean CAPRA-S 2.1 ± 1.8). CCP scores ranged from -1.62 to 2.16; the mean was -0.28 ± 0.63. CCP score was associated with recurrence, HR 2.1 per CCP score unit (95% CI 1.6-2.9), p<0.0001. With adjustment for pathologic variables, patient age, and year of surgery, HR was 2.0 per CCP score unit (1.4-2.9), p<0.0001; and with adjustment for CAPRA-S score, the HR was 1.7 per CCP score unit (1.3-2.3), p<0.001. By likelihood ratio testing, adding the CCP score to the CAPRA-S improved accuracy significantly (LR χ2 11.4, p<0.0001). Based on Kaplan-Meier analysis, CCP score adds useful discrimination to both the overall cohort and for patients with clinically low-risk disease (CAPRA-S score 0-2). Conclusions: In a contemporary post-RP cohort, the CCP score was validated to have significant prognostic accuracy even when controlling for clinical and pathologic data. The score has potential to substantially improve accuracy of risk-stratification, and lead to better decision-making with respect to timing and intensity of treatment.