The Impact of Pre-Treatment PSA on Risk Stratification in Men with Gleason 6 Prostate Cancer: Implications for Active Surveillance
Sina Monfared, BS1, Aaron Fleishman, MPH2, Ruslan Korets, MD2, Peter Chang, MD, MPH2, Andrew Wagner, MD2, Glenn Bubley, MD2, Irving Kaplan, MD2, Aria Olumi, MD2, Boris Gershman, MD2.
1Boston University School of Medicine, Boston, MA, 2Beth Israel Deaconess Medical Center, Boston, MA.
BACKGROUND: The optimal treatment for men with Gleason 6 prostate cancer and a PSA above 10 ng/mL is uncertain. Although active surveillance is the preferred management strategy for men with low-risk disease, there are limited data to support the safety of active surveillance in men with favorable-intermediate risk disease due only to a discordant PSA above 10 ng/mL. We therefore evaluated the impact of pre-treatment PSA on risk-stratification in men with Gleason 6 prostate cancer.
METHODS: We identified men aged 18-75 with cT1-2 cN0 cM0, pre-treatment PSA <20 ng/mL, Gleason 6 prostate cancer diagnosed from 2010-2016 in the National Cancer Database (NCDB) who underwent radical prostatectomy. The primary outcomes were Gleason score upgrading to 7-10 and adverse pathologic features at prostatectomy (i.e., pT3a, pT3b, or pN1). The associations of patient and disease features with each outcome were evaluated using bivariate analyses as well as univariable and multivariable logistic regression. To evaluate for non-linear relationships between PSA and each outcome, we examined predicted marginal event rates standardized for baseline characteristics with PSA modeled using restricted cubic splines with four knots.
RESULTS: A total of 79,036 patients were included in the cohort. Mean age at diagnosis was 59.5 (SD 6.8) years, and median pre-treatment PSA was 5.1 (IQR 4.1-6.8) ng/mL. At prostatectomy, Gleason score upgrading was identified in 49.7% of patients, including 41.2% with Gleason 3+4, 5.3% with Gleason 4+3, and 1.5% with Gleason 8-10. Adverse pathologic features were identified in 10.5% of patients, including 8.9% with pT3a disease, 1.4% with pT3b disease, and 0.2% with pN1 disease. In unadjusted analyses, patients with pre-treatment PSA ≥10 ng/mL had higher rates of Gleason core upgrading (58.8% vs 47.9%, p<0.001) and adverse pathologic features (19.7% vs 10.0%, p<0.001) compared to patients with a PSA <10 ng/mL. In multivariable analyses that adjusted for patient and disease features, PSA ≥10 ng/mL was associated with statistically significantly increased risks of Gleason score upgrading (OR 1.47, 95% CI 1.39-1.55) and adverse pathologic features (OR 2.14, 95% CI 2.00-2.29). When modeled as a non-linear continuous covariate, PSA was associated with increased adjusted rates of Gleason score upgrading (Figure 1) and adverse pathologic features (Figure 2) without a clear dichotomization at a threshold of 10 ng/mL.
CONCLUSIONS: In this surgical cohort, higher pre-treatment PSA was independently associated with increased risks of Gleason score upgrading and adverse pathological features at prostatectomy in men with Gleason 6 prostate cancer. Flexible modeling of the relationship between PSA and each outcome did not support dichotomization at a threshold of 10 ng/mL. These results can be used to improve patient risk-stratification for active surveillance. 
Back to 2020 Abstracts