High immunoexpression of Ki67, EZH2, and SMYD3 in diagnostic prostate biopsies independently predicts outcome in patients with prostate cancer.

Introduction: Overtreatment is a major concern in patients with prostate cancer (PCa). Prognostic biomarkers discriminating indolent from aggressive disease in prostate biopsy are urgently needed. We aimed to evaluate the prognostic value of Ki67, EZH2, LSD1, and SMYD3 immunoexpression in diagnostic biopsies from a cohort of PCa patients with long term follow-up.

Materials And Methods: A series of 189 consecutive prostate biopsies diagnosed with PCa (1997-2001) in a cancer center was included in the study, with follow-up last updated in November 2016. Biopsies were reviewed and graded according to 2016 WHO criteria. Immunohistochemistry was performed in the most representative block. Nuclear staining was assessed using digital image analysis. Study outcomes included disease-specific, disease-free, and progression-free survival. Statistical analysis was tabulated using SPSS version 22.0. Survival curves and hazard ratios (HRs) were estimated using Kaplan-Meyer and Cox-regression models, respectively. Statistical significance was set at P<0.05.

Results: The proportion of patients who completed the study was 177/189 (94%). In univariable analysis, high Ki67, EZH2, and SMYD3 immunoexpression associated with significantly worse disease-specific survival (HR = 1.86, 95% CI: 1.05-3.29; HR = 1.87, 95% CI: 1.10-3.27; HR = 2.68, 95% CI: 1.02-7.92). In multivariable analysis, the 3 biomarkers displayed significantly worse DSS adjusted for CAPRA score (HR = 1.78, 95% CI: 1.01-3.16; HR = 1.93, 95% CI: 1.12-3.32; HR = 2.71, 95% CI: 1.04-7.10). Among patients with low/intermediate risk CAPRA score, high Ki67 immunoexpression identified those more prone to experience disease recurrence (HR = 9.20, 95% CI: 1.27-66.44) and progression (HR = 2.97, 95% CI: 1.05-8.43).

Conclusions: High Ki67, EZH2, and SMYD3 immunoexpression, adjusted for standard clinicopathological parameters, independently predicts outcome in patients with PCa, at diagnosis. This might assist in discriminating indolent from aggressive PCa, improving treatment selection.