Data from The <i>TMPRSS2:ERG</i> Rearrangement, ERG Expression, and Prostate Cancer Outcomes: A Cohort Study and Meta-analysis

Background: Whether the genomic rearrangement transmembrane protease, serine 2 (TMPRSS2):v-ets erythroblastosis virus E26 oncogene homolog (ERG) has prognostic value in prostate cancer is unclear.

Methods: Among men with prostate cancer in the prospective Physicians' Health and Health Professionals Follow-Up Studies, we identified rearrangement status by immunohistochemical assessment of ERG protein expression. We used Cox models to examine associations of ERG overexpression with biochemical recurrence and lethal disease (distant metastases or cancer-specific mortality). In a meta-analysis including 47 additional studies, we used random-effects models to estimate associations between rearrangement status and outcomes.

Results: The cohort consisted of 1,180 men treated with radical prostatectomy between 1983 and 2005. During a median follow-up of 12.6 years, 266 men experienced recurrence and 85 men developed lethal disease. We found no significant association between ERG overexpression and biochemical recurrence [hazard ratio (HR), 0.99; 95% confidence interval (CI), 0.78–1.26] or lethal disease (HR, 0.93; 95% CI, 0.61–1.43). The meta-analysis of prostatectomy series included 5,074 men followed for biochemical recurrence (1,623 events), and 2,049 men followed for lethal disease (131 events). TMPRSS2:ERG was associated with stage at diagnosis [risk ratio (RR)_{≥T3 vs. T2}, 1.23; 95% CI, 1.16–1.30) but not with biochemical recurrence (RR, 1.00; 95% CI, 0.86–1.17) or lethal disease (RR, 0.99; 95% CI, 0.47–2.09).

Conclusions: These results suggest that TMPRSS2:ERG, or ERG overexpression, is associated with tumor stage but does not strongly predict recurrence or mortality among men treated with radical prostatectomy.

Impact: This is the largest prospective cohort study to examine associations of ERG overexpression and lethal prostate cancer among men treated with radical prostatectomy. Cancer Epidemiol Biomarkers Prev; 21(9); 1497–509. ©2012 AACR.