ABSTRACT
BACKGROUND: Prostate cancers (PCs) with similar characteristics at the time of diagnosis can have very different disease outcomes. Conventional biomarkers of PC still lack precision in identifying individuals at high risk of PC recurrence. While many candidate biomarkers are proposed in the literature, few are in clinical practice as they lack rigorous validation. This study prospectively enrolled an independent phase III cohort to evaluate the clinical utility of zinc-alpha 2-glycoprotein (AZGP1) as a prognostic biomarker in localized PC. PATIENTS AND METHODS: In our multicentre, prospective phase III study, AZGP1 status in 347 radical prostatectomy specimens was assayed by immunohistochemistry in a NATA-accredited laboratory. The AZGP1 score was assessed in a multivariable model incorporating established prognostic factors. We also report extended outcomes from our previous phase II study. The primary endpoint was biochemical relapse-free survival (BRFS). Secondary endpoints were metastasis-free survival (MFS) and PC-specific survival (PCSS). RESULTS: In the phase II cohort, with a median follow-up of 15.8 years, low/absent AZGP1 expression was an independent predictor of poor BRFS (HR, 1.4; 95% CI, 1.1-1.9; P = 0.03), MFS (HR, 2.8; 95% CI, 1.2-6.6; P = 0.02) and PCSS (HR, 3.8; 95% CI, 1.5-9.5; P = 0.005). These results were validated in our prospective phase III cohort. Low/absent AZGP1 expression independently predicted for BRFS (HR, 1.9; 95% CI, 1.1-3.3; P = 0.02), with shorter MFS (HR, 2.0; 95% CI, 1.1-3.4; P = 0.02). AZGP1 improved the discriminatory value when incorporated into existing prognostic risk models. CONCLUSION: Our study provides prospective phase III validation that absent/low AZGP1 expression provides independent prognostic value in PC. This study provides robust evidence for the incorporation of this biomarker into clinical practice.
