NEW YORK (GenomeWeb) – A Team from MD Anderson working to test and validate Biocartis' Idylla System for analysis of circulating tumor DNA, has shown that the platform can detect mutations with high sensitivity compared to tissue samples, and with identical results to other PCR-based liquid biopsy technologies like ddPCR or BEAMing.
The study, which focused on identifying BRAF mutations in late-stage cancer patients and was published today in Molecular Cancer Therapeutics, also found, somewhat counterintuitively, that higher frequencies of BRAF V600E mutations in the blood were associated with poorer outcomes overall and less responsiveness to BRAF inhibitors.
Filip Janku, the study's lead author and an assistant professor in MD Anderson's department of investigational cancer therapeutics, told GenomeWeb this week that other studies have also demonstrated this unexpected link between a lack of ctDNA mutations and better targeted therapy response.
"We are not the only group noticing this," Janku said. "We published a paper a year ago looking at KRAS, BRAF, EGFR, and PIK3CA, and had similar findings. … This is also in line with data published a few months ago from a study of patients treated in multiple trials of BRAF and MEK inhibitors."
He added, "One hypothesis may be that this reflects something about the biology of the tumor, [wherein] more shedding of DNA from cells undergoing apoptosis or necrosis reflects more aggressiveness in the tumor itself."
In the study, Janku and colleagues analyzed plasma samples collected from 160 patients with a range of advanced cancers who had known BRAFV600 mutation status determined by testing of paraffin-embedded tumor tissue samples.
The most common tumor types represented in the study were colorectal cancer and melanoma, but there were also patients with multiple other tumor types including non-small cell lung cancer and thyroid cancer.
The investigators then analyzed plasma samples from each subject — a subset of patients had multiple blood samples taken from different time points, but some had only one — for BRAF V600 mutations using the Idylla system.
Overall, the team found that the test had 88 percent concordance with tissue-based BRAF status when they compared only blood samples collected at baseline.
In several patients, however, while a baseline blood sample was BRAF V600E-negative, samples taken later in the disease course did begin to show the mutation. When Janku and his colleagues included these samples from later time points, the concordance between liquid and tissue biopsy rose to 90 percent.
According to the research team, blood-based BRAF testing on the Idylla platform was 73 percent sensitive and 98 percent specific, with a positive-predictive value of 96 percent and a negative-predictive value of 85 percent.
In cases where tumor tissue showed a BRAF mutation but blood samples did not, the group also took a second step and retested samples using two other PCR-based methods that are being advanced in the liquid biopsy space: Bio-Rad's Droplet Digital PCR and Sysmex Inostics' BEAMing.
The three platforms — Idylla, ddPCR, and BEAMing — yielded identical results for every tissue-positive, blood-negative sample.
According to Janku, this solidifies growing evidence that where liquid biopsy and tissue biopsy fail to correspond, at least in the context of PCR-based testing, it does not appear to be as a result of a failure of technology, but more likely a reflection of tumor biology.
Though small, Janku and his colleagues also included an analysis of the relationship between blood-borne V600E mutations and patient outcomes, which he said is mainly hypothesis generating at this point and will need to be followed up with prospective studies before the full clinical relevance can be determined.
As might be expected, the researchers found that the amount of BRAFV600 mutant cell-free DNA as detected by the Idylla system was predictive of overall survival, but the way it was predicted was somewhat counterintuitive: Those with less mutant DNA in circulation actually survived longer and responded better to targeted therapy than those with more circulating mutant DNA.
Patients with 2 percent or less mutant DNA had an average overall survival of 10.7 months, compared with 4.4 months in those who had more than 2 percent BRAFV600 mutations.
Moreover, in those whose baseline blood samples were negative for BRAFV600 mutations, the time to treatment failure with a BRAF or MEK inhibitor was 13.1 months on average, compared to those whose baseline samples were positive for the mutation, in whom treatment failure occurred within an average of three months.
A possible explanation, Janku said, could be that tumors that shed more mutated DNA into the blood are doing so because they are more aggressive, fast growing cancers, and thus, these patients have poorer outcomes. Further research will be required to determine the reason or mechanism behind the MD Anderson team's findings.
Janku and colleagues are working with the Idylla instrument in a research-use-only setting, and the platform is not currently available for clinical use in the US, although under current regulation, many US labs have employed research-use instruments in their CLIA laboratories for testing of clinical samples.
Biocartis received a €1.4 million ($1.5 million) grant from the Flemish Agency for Innovation by Science and Technology last fall to support the development of liquid biopsy versions of its solid tumor BRAF and KRAS diagnostics for melanoma and colorectal cancer.
In January, the company announced that it had signed an agreement with Merck to develop and commercialize a liquid biopsy test including an extended panel of RAS mutations as well as BRAF V600 mutation analysis for metastatic colorectal cancer on the Idylla platform.
The companies said they planned to make the panel available for research use only in the second half of this year and submit it for CE marking shortly thereafter.