Results bolster diagnostic value of prostate cancer biomarker

November 1, 2008

The prostate cancer biomarker TMPRSS2-ERG is superior to DNA-based assays.

Boston-TMPRSS2-ERG has recently emerged as a promising biomarker for prostate cancer. At the AUA annual meeting in Orlando, FL, one group of researchers reported the findings from their multicenter study investigating the prevalence of TMPRSS2-ERG fusion prostate cancer in U.S. men undergoing prostate biopsy. Investigators in a second study evaluated the relative performance of DNA methylation, PCA3, and TMPRSS2-ERG gene fusion as prostate cancer biomarkers in expressed prostatic secretions.

The study of TMPRSS2-ERG fusion prostate cancer prevalence evaluated a single representative slide from 140 men undergoing prostate biopsy at five clinical centers in Massachusetts and Michigan; 134 of the 140 slides were assessable by fluorescence in situ hybridization (FISH) for TMPRSS2-ERG fusion.

Based on histologic evaluation performed by pathologists blinded to fusion status, prostate cancer was diagnosed in 100 men. The prevalence of TMPRSS2-ERG positive fusion status was 46% among the prostate cancer cores and 0% in the benign group, reported first author Juan-Miguel Mosquera, MD, of the department of pathology at Brigham and Women's Hospital, Boston.

"Ours is the first report of TMPRSS2-ERG gene fusion prostate cancer prevalence in North American men undergoing prostate needle biopsy, and the results are similar to the German prostatectomy series. Considering that previous studies show associations between positive gene fusion status and more advanced prostate cancer, perhaps the difference in the prevalence rates between the prostatectomy and TURP specimens reflects a higher incidence of indolent cancers in the TURP specimens."

The researchers also sought to identify clinical and histologic features associated with TMPRSS2-ERG fusion-positive fusion status. Clinical variables assessed included age, serum PSA, PSA density, prostate size, race, and Gleason score. Five histologic features were assessed: blue-tinged mucin, macronucleoli, intraductal tumor spread, cribriform growth pattern, and signet-cell features.

In univariate analysis, non-Caucasian patients, including African-Americans and Hispanics, were significantly less likely than Caucasians were to have TMPRSS2-ERG fusion-positive prostate cancer. In multivariate logistic regression analysis, lower serum PSA density and three morphologic features- blue-tinged mucin, macronucleoli, and cribriform growth pattern-were associated with positive TMPRSS2-ERG fusion status.

Small subsets of cases from each institution were exchanged to validate the FISH evaluation. The results showed complete agreement between the two centers on TMPRSS2-ERG fusion status.

Three predictive tests compared

The study comparing DNA methylation, PCA3, and TMPRSS2-ERG gene fusion as noninvasive predictors of biopsy outcomes analyzed samples from 74 men referred for transrectal ultrasound-guided prostate biopsy because of elevated PSA or abnormal digital rectal exam.

Based on receiver operating characteristic (ROC) analysis, each biomarker was evaluated for improved performance in predicting the biopsy results compared with the baseline of PSA and DRE. While each biomarker demonstrated diagnostic value, the combination of PSA, DRE, and TMPRSS2-ERG offered the best performance, reported co-author Steven S. Smith, PhD, professor of molecular science at the City of Hope's Beckman Research Institute, Duarte, CA.

"There are a number of noninvasive specimens (expressed prostatic secretion, post-massage urine, and ejaculate) that could contain predictive biomarkers," said co-investigator Laura Crocitto, MD, assistant professor of surgery, urology and urologic oncology at City of Hope.

"Available data suggest that EPS may represent an improvement in diagnostic reliability compared with post-massage urine," added Dr. Smith, "but we are currently investigating the equivalency of these specimens. In addition, we are trying to improve signal recovery in our DNA methylation assay, which has not performed as well in EPS as one might expect from surgical specimens.

"Compared to the RNA-based assays for TMPRSS2-ERG or PCA3, the DNA-based assay for methylation is also more time-consuming to perform, since it involves multiple biomarker sites."