Background
Over-expression of prostate-specific membrane antigen (PSMA) in tumor tissue and serum has been linked to increased risk of biochemical recurrence in surgically treated prostate cancer patients, but no studies have assessed its association with disease-specific mortality.
Methods
We examined whether high PSMA protein expression in prostate tumor tissue was associated with lethal disease, and with tumor biomarkers of progression, among participants of two US-based cohorts (n=902, diagnosed 1983–2004). We used Cox proportional hazards regression to calculate multivariable hazard ratios (HR) and 95% confidence intervals (CI) of lethal prostate cancer, defined as disease-specific death or development of distant metastases (n=95). Partial Spearman rank correlation coefficients were used to correlate PSMA with tumor biomarkers.
Results
During an average 13 years of follow-up, higher PSMA expression at prostatectomy was significantly associated with lethal prostate cancer (age-adjusted HRQuartile(Q)4vs.Q1=2.42; p-trend<0.01). This association was attenuated and non-significant (multivariable-adjusted HRQ4vs.Q1=1.01; p-trend=0.52) after further adjusting for Gleason score and PSA at diagnosis. High PSMA expression was significantly (p<0.05) correlated with higher Gleason score and PSA at diagnosis, increased tumor angiogenesis, lower vitamin D receptor and androgen receptor expression, and absence of ERG expression.
Conclusions
High tumor PSMA expression was not an independent predictor of lethal prostate cancer in the current study. PSMA expression likely captures, in part, malignant features of Gleason grade and tumor angiogenesis.
Impact
PSMA is not a strong candidate biomarker for predicting prostate cancer-specific mortality in surgically treated patients.