Association of SPINK1 expression and TMPRSS2:ERG fusion with prognosis in endocrine-treated prostate cancer.

PURPOSE The aim of the study was to examine whether TMPRSS2:ERG fusion or SPINK1 protein expression is associated with hormone responsiveness of prostate cancer and can thus be used as a biomarker. EXPERIMENTAL DESIGN Diagnostic needle biopsies from prostate cancer patients primarily treated by endocrine therapy were evaluated for TMPRSS2:ERG fusion with fluorescence in situ hybridization and SPINK1 protein expression with immunohistochemistry. RESULTS The frequency of TMPRSS2:ERG fusion in 178 biopsies of hormonally treated patients was 34%. Of the fusion-positive cases, 71% showed deletion between the two genes, and 23% showed gain of the fusion. The fusion was associated with high Ki-67 staining (P=0.001), age at diagnosis (P=0.024), and tumor area (P=0.006), but not with Gleason score, T stage, M stage, prostate-specific antigen (PSA), or progression-free survival. Strong positive SPINK1 expression was found in 11% (21 of 186) of the biopsies. SPINK1-positive cases had significantly shorter progression-free survival compared with SPINK1-negative cases (P=0.001). The expression was not associated with any other clinicopathologic variables studied. In a multivariate analysis, SPINK1 expression showed independent prognostic value, with a relative risk of 2.3 (95% confidence interval, 1.1-4.6). SPINK1 expression and the fusion were not associated with each other. CONCLUSIONS There was no association between TMPRSS2:ERG fusion and prognosis, suggesting that TMPRSS2:ERG rearrangement does not implicate hormone dependence of the cancer. SPINK1 expression, found in approximately 10% of prostate cancers, was associated with aggressive form of the disease and could serve as a biomarker in endocrine-treated prostate cancer.