NEW YORK (GenomeWeb) – A study presented this week at the American Society for Clinical Oncology annual meeting has answered some questions that clinicians may have regarding a new area of precision medicine — the identification of men with advanced prostate cancer who are unlikely to respond to widely used hormonal therapies.
This year saw the first commercial launch of an AR-V7 test for metastatic castration-resistant prostate cancer — developed by Epic Sciences and marketed by Genomic Health — as well as moves by Qiagen to create a research-use kit version of a separate assay developed by researchers at Johns Hopkins University.
The new data at ASCO, from a study called PROPHECY, provides an independent validation that the presence of AR-V7 splice variants in circulating tumor cells is predictive of poor response to the anti-androgen drugs abiraterone and enzalutamide, regardless of which test is used. But the data so far leave other questions unanswered.
Based on the results, the two tests both appeared to be strong negative predictors of response to hormonal therapy. Men who were AR-V7 positive by either assay had a significantly poorer outcome on these drugs than those who were AR-V7 negative.
But the tests do not appear to distinguish the same populations — they were only 82 percent concordant. And the study results don't yet paint a clear picture of why both tests might be analytically valid while diverging from each other in this way — something that could give pause to clinicians considering using one or the other.
The PROPHECY trial, led by Duke University oncologist Andrew Armstrong, was designed to provide an independent multi-center validation of AR-V7 as a biomarker in metastatic castration-resistant prostate cancer, as well as to explore the concordance of different available assays, and to provide a venue for broad genomic analyses of circulating tumor cells in this population that could lead to better tests down the line.
One significant downside to AR-V7, regardless of which test is used, is that although the biomarker is very specific, it only detects a subset of the total population of men who don't respond to hormonal therapy.
This held true in the PROPHECY trial. Using the Hopkins test, only three patients showed a response to hormonal therapy despite AR-V7 detection. With the Epic test, no patients that tested positive had a confirmed PSA decline.
However, Armstrong said during his conference presentation, "the lack of AR-V7 does not guarantee response or benefit."
Discussing the results further, Anthony Joshua of Canada's Princess Margaret Cancer Center reiterated that despite high specificity, PROPHECY confirmed that both of the tests in question pick out only a subset of the men who are unlikely to benefit from these drugs.
"Its Important to remember that a third of the men in this study with AR-V7 negative disease do not have a response at all," he said.
Armstrong did not present on PROPHECY's more investigative aspects during the conference, but other evidence has grown in the meantime that there may be other ways to identify hormonal therapy non-responders that could help fill the gaps that AR-V7 leaves open.
In a study last year in Annals of Oncology, for example, a team led by researchers at the Institute of Cancer Research, London showed that AR copy number changes in the circulating cell-free DNA of men with advanced prostate cancer were significantly associated with outcome in men treated with abiraterone and enzalutamide.
As part of that study, investigators analyzed samples from a cohort that was also tested with the AdnaTest method being advanced by Johns Hopkins. And in those men, AR-V7 was not predictive of therapy response, while AR copy number in cfDNA was.
Researchers working with Epic Sciences have also said they are investigating other capabilities of the company's platform to further stratify patients, including measuring the overall genomic and phenotypic heterogeneity of CTCs as a predictor of response to anti-androgen drugs.
In previous research he and his team found that patients whose CTC population showed a high degree of heterogeneity — either in their phenotypes or their genetic clonality — did not respond as well to hormone therapy as those with less heterogeneity.
Interestingly, the PROPHECY data indicated that CTC heterogeneity had a significant overlap with AR-V7 positivity. According to Armstrong, 63 percent of AR-V7-positive men had highly heterogeneous CTCs compared to only about 14 percent of AR-V7-negative men.
This suggests that the two factors may not be independent of each other, which would hamper an effort by Epic Sciences to use heterogeneity as way to boost its detection of non-responders. But that would need to be investigated and tested specifically to fine out one way or another.
As Epic has moved forward with its test, which received a favorable draft local coverage determination from Medicare contractor Palmetto GBA — it has presented its own data in an attempt to help distinguish itself from the Hopkins methodology, which Qiagen licensed last year and launched for research use only this April.
One factor that clearly distinguishes the two tests is that Epic's assay looks specifically for expression of AR-V7 in a circulating tumor cell's nucleus, which helps explain why in previous studies the test showed a much lower rate of AR-V7 detection compared to an RNA-based approach applied to lysed samples.
The PROPHECY data bear this out. As Epic has highlighted in other studies, its nucleus-specific method appears to avoid false positives. However, while the Hopkins method did result in a few false positives, it also detected a much larger number of true positives: about twice the number as the Epic assay.
The question of balance — minimizing false positives versus missing true positives — is one that clinicians will have to grapple with when considering using this type of testing in their practice, and in the future, whether to use one assay or another. Now that Qiagen has launched its AR-V7 RUO kit, CLIA labs could potentially use it to develop their own RNA-based AR-V7 tests in the same vein as the Hopkins assay.
In all, the trial results don't offer much of a clear takeaway in terms of test superiority. Both assays were strong predictors of both shorter progression-free survival and shorter overall survival. The Hopkins test demonstrated a hazard ratio of 2.4 for shorter PFS, and 3.9 for shorter OS. For the Epic test, the HR was 2.5 for PFS and 4.5 for OS.
When the researchers controlled for other variables, both tests maintained an independent predictive value in a multivariate model, although the Hopkins assay appeared to fare a little better than the Epic method.
According to Joshua, the study results may not be enough to persuade the clinic of a predictive value for AR-V7 apart from the role it plays, amongst various clinical factors, in marking prognosis.
"Further prospective trials may be needed to define predictive ability in order for those tests to guide clinical care," he said.
Clinicians, he argued, don't only need studies like PROSPECT, which show that these tests can predict which men are unlikely to respond to hormonal drugs, but also studies that define more precisely how such a tool can and should be used to guide care.
"The first question in my practice isn't should I order an AR-V7 assay, its rather, 'When should I recommend chemo for a man with prostate cancer.'"
The question for oncologists is not as simple as just whether to use abiraterone or enzalutamide versus chemotherapy. They need to know which patients who have progressed on one of these drugs can try another versus which need to be switched to chemo.
To truly answer that question, Joshua argued, clinicians should push for trials that employ these assays in a structure where researchers can gauge whether the tests actually improve outcomes by informing one or the other of these treatment pathways.
With clinical utility questions like this still outstanding, questions clearly remain in terms of how the commercial landscape for these assays will evolve.