Prostate Cancer Gene Test Could Boost Active Surveillance
Published Online: Thursday, July 25, 2013
Matthew Cooperberg, MD, MPH
A validation study showed that combining results of the Genomic Prostate Score (GPS) with standard risk-stratification criteria more than doubled the number of men who could qualify for active surveillance as opposed to immediate aggressive therapy. The standard criteria alone suggested 5% to10% of men were suitable for surveillance, and that increased to about 25% with the addition of GPS results.
The GPS demonstrated precision for assigning patients to higher-risk categories, as well as lower-risk categories.
“For every 20-point increase in the (100-point) GPS, the likelihood of adverse pathology doubled,” said Matthew Cooperberg, MD, MPH, a urologic oncologist at the University of California, San Francisco. “Almost half of the patients included in this study had at least a 5% shift in risk, higher or lower, on the basis of GPS results.”
The GPS comprises a 17-gene panel that incorporates information from biologic pathways involved in stromal response, cellular organization, androgen signaling, and proliferation, as well as five reference genes. Studies have shown that the assay can accurately assess gene expression in prostate cancer specimens as small as 1 mm.
Cooperberg presented results from a study involving 395 patients with low- and intermediate-risk prostate cancer, all of whom had been treated by radical prostatectomy. The GPS was used to evaluate specimens obtained by needle biopsy. The study’s primary outcome was the correlation of GPS findings with final pathology, as determined by the prostatectomy specimen.
Investigators evaluated the assay’s ability to predict the presence of high-grade disease and pathologic stage T3. The results showed that every 20-unit increase in the GPS increased the odds ratio (OR) for high-grade disease by 2.48 and the likelihood of pT3 disease by 2.20 (P<.001 for both evaluations). Cooperberg said a 20-unit change was analogous to comparing the top and bottom quartiles of patients.
A prespecified analysis showed that the GPS results derived from prostate biopsies predicted pathology on radical prostatectomy independent of Gleason score.
In three different multivariable models that incorporated all known information about a tumor specimen, the GPS retained its ability to predict the presence of high-grade and/or high-stage disease (OR 1.85-2.13, P=.003 to P<.001).
Investigators also examined the incremental precision of the GPS when added to two standard risk-stratification systems: the Cancer of the Prostate Risk Assessment (CAPRA) and the National Comprehensive Cancer Network (NCCN) risk-assessment system.
Combining the GPS with CAPRA resulted in at least a 5% shift in risk in 49% of patients. The change was toward more favorable status in 26% of cases and toward less favorable status in 23%.
By themselves, the CAPRA and NCCN risk-stratification schemes predicted that 5% to10% of men had at least an 80% likelihood of organ-confined disease, making them candidates for active surveillance. When the GPS results were added to either risk-stratification system, the proportion of men qualifying as candidates for active surveillance increased to 24% to 26%.
“The GPS adds independent predictive information beyond all standard clinical and pathological data,” Cooperberg said. “The GPS assesses underlying biology from very small biopsy tumor volumes, helping address tumor heterogeneity and biopsy undersampling to more accurately predict overall disease aggressiveness.
“Incorporation of the GPS enables more accurate identification of a larger population of patients who can more confidently choose active surveillance as an initial management strategy.”
Cooperberg M, Simko J, Falzarano S, et al. Development and validation of the biopsy-based genomic prostate score as a predictor of high grade or extracapsular prostate cancer to improve patient selection for active surveillance. Presented at: the Annual Meeting of the American Urological Association; May 4-8, 2013; San Diego, CA. Abstract 2131.
Most Popular Right Now