Comparing Prognostic Utility of a Single-marker Immunohistochemistry Approach with Commercial Gene Expression Profiling Following Radical Prostatectomy

Despite the availability of numerous genomic predictors of prostate cancer (PCa) outcome, few comparative studies have been performed. To compare the prognostic utility of previously validated immunohistochemical (IHC) markers with an expression-based cell-cycle progression (CCP) score. We identified 424 men with localized PCa treated with radical prostatectomy (RP). IHC analysis was performed using a tissue microarray to examine the expression status of PTEN, Ki-67, and ERG compared with previously calculated CCP scores derived from 31 genes normalized to 15 housekeeper genes. Associations of IHC status and CCP scores, adjusted for clinical and pathologic characteristics were performed using Cox regression and competing risks regression to examine risk of biochemical recurrence (BCR), and metastasis or PCa-specific mortality (PCSM). We compared models using concordance index (c-index) testing. RP. Median age at treatment was 59 yr, and patients were followed for a median of 114 mo after RP. By 10 yr after RP, 27% experienced BCR and 4% developed metastasis or PCSM. In a multivariable model adjusted for Cancer of the Prostate Risk Assessment score (CAPRA-S), CCP was associated with risks of recurrence (hazard ratio [HR] 1.51, 95% confidence interval [CI] 1.08–2.11) and metastasis/PCSM (HR 2.15, 95% CI 1.36–3.39). PTEN loss was not associated with recurrence but was associated with metastasis/PCSM (HR 5.26, 95% CI 2.57–10.7), adjusted for CAPRA-S. The c-index for models consisting of PTEN status and CAPRA-S was similar (0.80) for risk of metastasis/PCSM when compared with CCP and CAPRA-S (0.81). Integration of Ki-67 and ERG status did not improve the c-index relative to CAPRA-S and PTEN alone. PTEN status offered comparable discrimination of the risk of metastasis or death from PCa relative to a commercial RNA amplification-based CCP assay. Efforts are warranted to reduce the cost of PCa prognostic tools in order to expand access. We compared a commercial genomic signature and the expression status of single genes to predict outcomes in men with prostate cancer who were treated with surgical removal. When accounting for clinical information about the patient's cancer, the status of the PTEN gene alone matched a multigene panel to predict which patient's cancer would metastasize or lead to death from the disease.