Background: Most biomarkers in prostate cancer have only been evaluated in surgical cohorts. The value of these biomarkers in a different therapy context remains unclear. Our objective was to test a panel of surgical biomarkers for prognostic value in men treated by external beam radiotherapy (EBRT) and primary androgen deprivation therapy (PADT).
Methods: The Fluidigm® PCR array was used for multi-transcript profiling of laser microdissected tumours from archival formalin-fixed diagnostic biopsies of patients treated by EBRT or PADT. Cases were matched for disease characteristics and had known 5 year biochemical relapse outcomes (n = 60). Results were validated by immunohistochemistry in a custom needle biopsy tissue microarray. Six biomarkers previously tested only in surgical cohorts were analysed (PTEN, E-Cadherin, EGFR, EZH2, PSMA, MSMB). Transcript and protein expression was correlated with clinical outcome analysed using Kruskal Wallis, Fisher's test and Cox proportional hazard model.
Results: Altered expression of E-Cadherin (p = 0.008) was associated with early relapse after EBRT. In PADT treated men however only altered MSMB transcript was prognostic for early relapse (p = 0.001). The remaining biomarkers however did not demonstrate prognostic ability in either cohort. In a separate tissue array we validated altered E-Cadherin protein as a predictor of early relapse after EBRT (n = 47) (HR 0.34, CI p = 0.02) but not in PADT treated men (n = 63).
Conclusion: We demonstrate proof of principle of multiple transcript profiling in archival diagnostic biopsies of non-surgically treated men for biomarker discovery. We identify a role for E-Cadherin as a novel biomarker of early relapse following EBRT.