An analysis of multiple biomarkers to better predict prostate cancer metastasis and death after radical prostatectomy

Abstract

<jats:p> 54 </jats:p><jats:p> Background: Identification of potentially lethal disease at the time of diagnosis with localized prostate cancer (PCa) remains a significant clinical issue despite a plethora of candidate biomarkers. This study evaluates a range of biomarkers previously associated with biochemical relapse (BR) in localized PCa to determine whether a combined expression model can improve detection of clinically significant cases. Methods: The Australian PCa Research Centre NSW has completed 23 studies of molecular biomarkers associated with BR in a well-described localized PCa cohort (n=324, median followup 16 years). 12 studies were excluded due to missing data. Each biomarker was analyzed as a marker for metastatic-free survival (MFS) and prostate cancer specific survival (PCSS) and then used to develop a prognostic model for clinical outcomes incorporating clinico-pathological factors. This model is currently being validated in an independent cohort. Results: The PCa cohort experienced 39 metastatic relapses (12%) and 23 PCa deaths (7%). Of 12 biomarkers (AR, AZPG1, C0S, Cyclin D1a, Cyclin D1b, E6AP, H3K18Ac, H3K4me2, Ki67, p53, PML, SGTA) assessed, only AZGP1 and Ki67 were associated with MFS (HR 2.9, 95% CI, 1.4-5.6; P=0.002, and HR 1.2, 95% CI, 1.0-1.4; P=0.03, respectively) and PCSS (HR 4.2, 95% CI, 1.7-10.5; P=0.002; and HR 1.2, 95% CI, 1.0-1.5; P=0.04, respectively). The combined panel of AZGP1 and Ki67 was an independent predictor of MFS (HR 1.9, 95% CI, 1.1-3.2; P=0.01), and PCSS (HR 3.3, 95% CI, 1.5-7.3; P=0.002) when modeled with known clinicopathological variables. The panel was more robust in predicting MFS and PCSS compared to the individual biomarkers alone and superior to other prognostic models (See table). Data from the validation cohort will be available for the meeting. Conclusions: Our novel signature of AZPG1 and Ki67 improves existing prognostication tools in predicting PCa metastasis and death. [Table: see text] </jats:p>