CHICAGO (GenomeWeb) – Studies presented at this week's annual meeting of the American Society of Clinical Oncology contained new evidence supporting the use of Foundation Medicine's targeted sequencing test FoundationOne to extrapolate a cancer patient's genome-wide tumor mutation burden in order to potentially guide the use of immunotherapeutic drugs.
Foundation Medicine has said it plans to develop a CLIA-certified version of this tumor mutation burden analysis, which it will provide to physicians as part of FoundationOne and FoundationOne Heme reports by the third quarter of 2016.
In the meantime, the company and its collaborators are building evidence for the clinical utility of this added analysis, with numerous studies presented at ASCO this week, most supporting the plausibility of using FoundationOne to infer genome-wide mutation loads but some also linking the test to patient outcomes.
During a clinical science symposium that focused on biomarkers for response to anti-PD-1 drugs, researchers from Memorial Sloan-Kettering Cancer Center, led by Jonathan Rosenberg, presented a retrospective study of participants in the Imvigor210 trial of Roche's Tecentriq (atezolizumab) in metastatic urothelial cancer.
The investigators tested a number of assays — Ventana's SP142 PD-L1 immunohistochemistry assay, which was approved last month as a complementary diagnostic to atezolizumab; RNA sequencing to assess gene expression; and Foundation Medicine's 315-gene FoundationOne test to extrapolate genome-wide mutational burden.
Overall, Rosenberg and his colleagues concluded that PD-L1 expression, tumor mutation burden, and TCGA subtype can act as independent predictors of atezolizumab sensitivity.
In the same symposium, another group of researchers from Vanderbilt University, Foundation Medicine, Adaptive Biotechnologies, and the University of Texas, Southwestern shared data from a study in which they used FoundationOne as a proxy for overall mutational load in melanoma samples.
The results demonstrated a strong correlation between mutational burden, as measured by the FoundationOne panel, and the exome-wide frequency of mutations identified in 300 melanoma samples from TCGA.
Beyond these oral presentations, the conference also saw a handful of poster presentations by Foundation Medicine and academic colleagues.
In one, Foundation Medicine bioinformatics director Garrett Frampton and co-authors highlighted results from a study in which they used FoundationOne sequencing data from over 60,000 clinical cases across more then 400 cancer types to determine patterns of tumor mutation burden (TMB) across a wide spectrum of cancers.
The team identified numerous cancer types — for example certain types of stomach and uterine cancers — in which they saw many high-TMB cases, but for which cancer immunotherapy is not currently approved. This finding, the group wrote, supports other similar results and suggests a need for broader pan-cancer access to clinical trials for these therapies.
The group also reported on their validation of the accuracy of the FoundationOne-based TMB calculation by comparing it to whole-exome tumor mutation levels. For this, they compared TCGA whole-exome variant calls for 7,000 subjects to variant calls representing only the portion of the genome covered by FoundationOne.
This analysis, as well as a small second validation study in 16 matched samples that received clinical FoundationOne testing and whole-exome sequencing, both showed a high correlation between the FoundationOne TMB assessment and exome-wide TMB.
In a subset of about 35,000 samples, the researchers also compared FoundationOne-based TMB to microsatellite instability (MSI) — another known predictor of immunotherapy response. While most samples with high MSI also showed a high TMB, there were many microsatellite-stable samples that were nonetheless highly mutated. Overall, the authors wrote, only 15 percent of the high TMB cases also had high MSI, implying that TMB could help identify additional patients who might benefit from immunotherapy and would be missed by MSI testing alone.
In another poster, researchers led by Thomas George from the University of Florida shared additional data from a study comparing tumor mutation burden as assessed by FoundationOne with microsatellite instability status in more than 2,000 subjects with colorectal cancer.
While all of the patients in this analysis that had high microsatellite instability also had high TMB based on FoundationOne, there were also more than 400 subjects who had microsatellite stable cancers, but a high TMB.
Though the study did not look at immunotherapy responses in this MS-stable, TMB-high group, the authors were able to show that this group and the MSI-high patients had gene signatures in common. According to George and his co-authors, the results suggest FoundationOne-based TMB might be able to pick out more immunotherapy responders than MSI status alone, although that would have to be borne out in therapeutic outcomes data.
In yet another study, this one in lung cancer patients, researchers led by David Spigel of the Sarah Cannon Research Institute evaluated FoundationOne-based TMB status, PD-L1 expression, and MSI status.
Their results showed that while only a fraction of lung cancers analyzed tested positive for high MSI or for PD-L1 expression, a significant number had high TMB, implying that if TMB truly is a marker of response to immunotherapy in these patients, it appears to be a more sensitive tool to identify possibly responsive patients.
According to Foundation Medicine Chief Medical Officer Vincent Miller, who spoke with GenomeWeb before the ASCO conference, the company also plans to incorporate a microsatellite stability status readout into its test reports.
"We have completed a robust validation with PCR and IHC, which shows near perfect concordance between our calls on MSI status and those by the 'conventional modalities'," Miller said.
Physicians more familiar with MSI as a marker of immunotherapy response than TMB may be better served by having this particular marker pulled out on its own, Miler explained. At the same time, data presented at the conference highlighted that many patients that do not have high MSI still can have high TMB, suggesting they should be reported independently.
In addition to the two studies presented at the conference, in melanoma and urothelial cancer, a few other pieces of early data at the meeting also connected FoundationOne-based TMB status with immunotherapy response.
In one, researchers led by Alessandro Santin of Yale University used FoundationOne to assess targetable mutations, TMB, and MSI status in samples from patients with endometrial adenocarcinoma.
The results demonstrated that TMB might identify additional patients likely to respond to immunotherapy that MSI testing does not. In addition, the Yale group's poster mentioned briefly that both patients with high MSI and patients with stable microsatellites but high TMB ,including two cases initially excluded by standard testing, who started on immune checkpoint inhibitors have shown early clinical benefit.
As excitement around the new class of cancer immunotherapies continues to grow, evidence is also accumulating that there will be a need for multiple biomarkers or approaches for predicting immunotherapy response and that no single measure — whether PD-L1 expression, mutation burden, or T-cell status — should be used on its own to exclude patients from immunotherapy.
This was a point of agreement amongst speakers at ASCO, and was also recently highlighted in Science by researchers from the Netherlands Cancer Institute and the University of California, Los Angeles, who proposed a multi-pronged approach they call a "cancer immunogram" — which includes mutational burden, PD-L1, and a number of other immunological factors — for assessing an individual cancer patient's likelihood of responding to cancer immunotherapies.
"While there are likely multiple determinants of response to these classes of agents, there is emerging data that TMB is one of the most powerful we have, at least as a single measure or metric," Miller told GenomeWeb.
"What we've seen and shown to date … is that we can, as a natural byproduct of the [FoundationOne] test — not really changing anything about the assay, get an estimate of tumor mutational burden which correlates well with overall mutational burden as assessed by whole-exome sequencing," Miller said.
The utility of a FoundationOne-based TMB test will likely vary from cancer type to cancer type, he added. For example, in melanoma, where responsiveness to checkpoint inhibitors is high, it would be improper to use TMB to exclude patients from therapy, but it might help identify patients who are likely to have a more durable versus a more transient response, guiding the latter toward combination regimens.
The company is planning to make this type of analysis clinically available this year. In the meantime, Miller said that physicians who have appropriately consented their patients can order a TMB analysis as a research use-only service.