Prognostic capabilities and clinical utility of cell cycle progression testing, prostate imaging reporting and data system, version 2, and clinicopathologic data in management of localized prostate cancer

• Published in: Urologic oncology 2021-06, Vol.39 (6), p.366.e19-366.e28
• Main Authors: Morris, David S., Woods, J. Scott, Edwards, Byard, Lenz, Lauren, Logan, Jennifer, Flake, Darl D., Mabey, Brent, Bishoff, Jay T., Cohen, Todd, Stone, Steven
• Format: Article
• Language: English
• Subjects: CCP score; Gene expression classifier; Multiparametric magnetic resonance imaging; PI-RADS; Prognostic; Prostate cancer
• ISSN: 1078-1439; 1873-2496
• DOI: 10.1016/j.urolonc.2020.11.016

•PI-RADS and cell cycle risk scores are only modestly correlated.•The cell cycle progression score predicted tumor grade better than PI-RADS.•Molecular information is essential for accurate prostate cancer risk assessment. To compare the prognostic capabilities and clinical utility of the cell cycle progression (CCP) gene expression classifier test, multiparametric magnetic resonance imaging (mpMRI) with Prostate Imaging Reporting and Data System (PI-RADS) scoring, and clinicopathologic data in select prostate cancer (PCa) medical management scenarios. Retrospective, observational analysis of patients (N = 222) ascertained sequentially from a single urology practice from January 2015 to June 2018. Men were included if they had localized PCa, a CCP score, and an mpMRI PI-RADS v2 score. Cohort 1 (n = 156): men with newly diagnosed PCa, with or without a previous negative biopsy. Cohort 2 (n = 66): men who initiated active surveillance (AS) without CCP testing, but who received the test during AS. CCP was combined with the UCSF Cancer of the Prostate Risk Assessment (CAPRA) score to produce a clinical cell-cycle risk (CCR) score, which was reported in the context of a validated AS threshold. Spearman's rank correlation test was used to evaluate correlations between variables. Generalized linear models were used to predict binary Gleason score category and medical management selection (AS or curative therapy). Likelihood-ratio tests were used to determine predictor significance in both univariable and multivariable models. In the combined cohorts, modest but significant correlations were observed between PI-RADS score and CCP (rs = 0.22, P = 8.1 × 10−4), CAPRA (rs= 0.36, P = 4.8 × 10−8), or CCR (rs = 0.37, P = 2.0 × 10−8), suggesting that much of the prognostic information captured by these measures is independent. When accounting for CAPRA and PI-RADS score, CCP was a significant predictor of higher-grade tumor after radical prostatectomy, with the resected tumor approximately 4 times more likely to harbor Gleason ≥4+3 per 1-unit increase in CCP in Cohort 1 (Odds Ratio [OR], 4.10 [95% confidence interval [CI], 1.46, 14.12], P = 0.006) and in the combined cohorts (OR, 3.72 [95% CI, 1.39, 11.88], P = 0.008). On multivariable analysis, PI-RADS score was not a significant predictor of post-radical prostatectomy Gleason score. Both CCP and CCR were significant and independent predictors of AS versus curative therapy in Cohort 1 on multivariable analysis, with each