Despite FDA Clearance for NantHealth Test, Benefits of TMB Biomarker Unclear

This story has been updated to clarify details of Foundation Medicine's FoundationOne CDx. 

CHICAGO – Although NantHealth won US Food and Drug Administration 510(k) clearance last week for its Omics Core whole-exome test for determining tumor mutational burden, the company cannot market it as a test that can predict whether cancer patients will respond to a specific immunotherapy. 

The Omics Core technology is a whole-exome tumor-normal in vitro diagnostic that measures overall tumor mutational burden in cancer tissue and reports somatic mutations in 468 cancer-relevant genes. "The test is intended to provide information on somatic mutations (point mutations and small insertions and deletions) and tumor mutational burden (TMB) for use by qualified healthcare professionals in accordance with professional guidelines, and is not conclusive or prescriptive for labeled use of any specific therapeutic product," NantHealth said in a statement.

While this is the first whole-exome test that is FDA cleared specifically for assessing tumor mutational burden, and there is growing interest in exploring the use of TMB to identify best responders to checkpoint inhibitors, the lack of a predictive indication tied to a specific drug may be a limiting factor for the adoption of the test. 

In an email, NantHealth CMO Sandeep "Bobby" Reddy said that the Culver City, California-based company will market Omics Core based on the "utility of the test for assessment of multiple targets which correlate with multiple drugs and TMB for immunotherapy." Reddy said that this is no different than how, for example, Foundation Medicine is handling TMB. The company garnered FDA approval for its FoundationOne CDx in 2017, and lists TMB in test reports, but it's not one of the FDA's greenlit companion diagnostic indications associated with a specific drug. 

Reddy said that NantHealth is partnering with several pharmaceutical companies in pursuit of a companion diagnostic indication that would allow the firm to claim that Omics Core could predict response to specific drugs.

For example, he said, NantHealth is involved in the Quantum Immuno-Oncology Lifelong Trial (QUILT) 2.023 study investigating first-line pembrolizumab (Merck's Keytruda) in combination with another compound, called N803, for treatment of non-small cell lung cancer. QUILT is an initiative of NantHealth Founder and CEO Patrick Soon-Shiong's Cancer Moonshot 2020, designed to test novel combinations of vaccines, cell-based immunotherapy, metronomic chemotherapy, low-dose radiotherapy, and immunomodulators in patients who have undergone next-generation whole-genome, transcriptome, and quantitative proteomic analysis.

Even though NantHealth's test currently lacks a predictive indication, the company believes that its whole-exome test does have advantages over other tests on the market that report TMB. Reddy said that because Omics Core analyzes the entire exome, it "is better able to identify neoantigens than a panel test."

NantHealth predicated its FDA submission on Memorial Sloan Kettering Cancer Center's MSK-IMPACT next-generation sequencing assay. However, Omics Core differs from the MSK-IMPACT 468-gene panel in that it sequences the entire human exome and offers information about tumor mutational burden.

NantHealth said that the Omics Core bundle reports on the same 468 cancer-relevant genes as the MSK test down to 2 percent allele frequency to help guide physician decisions for cancer therapies. Reporting of tumor mutational burden is based on both the total number of somatic nonsynonymous exonic variants in the exome and on estimating mutation rate by counting all somatic, synonymous, and nonsynonymous variants detected in gene coding regions, then dividing by the approximate size of the exome, according to the company.

To highlight another advantage of its test, NantHealth also noted that the FoundationOne CDx assay only analyzes tumor tissue. "FoundationOne CDx identifies both somatic (acquired) and germline (inherited) mutations, but does not distinguish between the two on the test report," a representative for Foundation said over e-mail. 

"Omics Core is tumor-normal and thus adds a layer of increased accuracy to reliably detect both true somatic variants and germline variants," Reddy said. "The full exome allows for actual measurement of the TMB, rather than a calculated estimate from a panel test, and so the real difference for patients is precision, which is really the point of doing a precision medicine test."

Greater insight on analytical validity

In the absence of any predictive indications, however, the clearance offers more insight into the performance of NantHealth's test than anything else, said Jeff Allen, CEO of Friends of Cancer Research. "I'm gathering from the clearance that they were able to show a high degree of analytical validity for their test," he said. "It certainly gives them a status of validity for use in clinical trials."

However, NantHealth would have to demonstrate that its TMB test can identify best responders to specific immunotherapies before Omics Core could become a clinically valid companion diagnostic test, Allen said. The focus on whole-exome sequencing might give NantHealth an advantage in this regard, he added.

According to FDA's decision memorandum for the test, NantHealth reports results for point mutations and small indels in protein-coding exons in 468 genes. Omics Core does not report mutations in 203 exons because of low sequence coverage, poor mapping quality, or high GC content, the agency said in the document. 

Omics Core reports results under one of two categories: "Cancer Mutations with Evidence of Clinical Significance" or "Cancer Mutations with Potential Clinical Significance." Whether a detected variant falls in one of these categories depends on the level of supporting clinical evidence for that variant. The assay measures TMB as mutations per megabase and includes this value under "Cancer Mutations with Potential Clinical Significance," according to the decision memo.

FDA assessed the accuracy of Omics Core by comparing the assay's results to values produced by a reference lab. This testing involved 401 FFPE tumor samples, representing 2,634 single-nucleotide variants, 125 small insertions, and 313 small deletions.

Omics Core successfully detected mutations in all 401 samples, though there were three false positives across two genes and eight false negatives across eight genes, the regulatory agency said.

NantHealth submitted data across multiple sequencing runs, on the limit of detection and confirmation scores, and compared its test's ability to analyze more than 400 samples from different tumor types against an external sequencing laboratory, according to Reddy. The company also calculated positive percent agreement and positive predictive value for multiple variants for both the "clinically significant" 468 genes and the entire exome, he said.

"The validation for individual point mutations and small insertions and deletions was focused on the 468 genes reported in MSK-IMPACT due to their relevance in oncology, however we performed comparisons of our entire exome sequencing against the external lab for the purposes of validating our TMB measure," Reddy explained. He did not comment on any performance-related differences between Omics Core and MSK-IMPACT.

The comparison study that NantHealth submitted to the FDA was from an external lab using different sequencing machines, capture kits, and variant-calling pipelines. Reddy said that Omics Core is able to identify all variant classes, including fusions and novel viral integration sites.  

NantHealth's WES test reports the actual TMB as the number of nonsynonymous variants detected in the tumor sample. Reddy said that this assay follows the framework of a 2015 Science paper, in which researchers from Memorial-Sloan Kettering and other institutions found a median of 200 nonsynonymous mutations per sample in a retrospective analysis of pembrolizumab responsiveness for non-small cell lung cancer, which caused NantHealth to set a cutoff of 200 for its own test.

NantHealth has not yet announced a price for Omics Core as a standalone product, though Reddy said that GPS Cancer, a solid tumor whole-genome and transcriptome sequencing assay, carries a list price of $10,500. GPS Cancer has a turnaround time of 21 days.

'Reimbursement conundrum'

The company has said for the better part of a year that it would not be able to reach profitability without solving the "reimbursement conundrum" associated with these assays. In a conference call following the release of NantHealth's third quarter financial results this month, CFO Bob Petrou said that FDA clearance for the tests would open the door to much-needed payor reimbursement.

The Centers for Medicare and Medicaid Services has granted national coverage for NGS tests ordered for advanced cancer patients who have an indication addressed by an FDA-approved CDx claim. While NantHealth cannot market Omics Core for specific indications, Reddy said that GPS Cancer does have some private reimbursement contracts.  

"We are confident that FDA approval will help with further contracting," Reddy said. He said that the company is pursuing the CDx path as well as other unspecified means of securing payor coverage for its tests.

NantHealth has sharply curtailed the fulfillment of molecular tests in the last year or so to help stem the sea of red ink. In the third quarter, the company performed fewer than 50 of its GPS-branded MDx tests, including GPS Cancer and the Liquid GPS blood-based test. As recently as Q4 2018, NantHealth had fulfilled more than 1,000 of the tests.

For the three months ended Sept. 30, NantHealth posted a net loss of $16.4 million, or $.15 per diluted share, compared to a net loss of $97.5 million, or $0.89 per share, in the same period a year earlier. The company said this month that it became cash flow positive for the first time during the 2019 third quarter.

Reddy said that a favorable reimbursement determination for its molecular tests would be enough to get NantHealth into the black for the first time. He said that the company's "software enterprises" were profitable.

Sequencing and molecular analysis accounted for a little more than 1 percent of NantHealth's overall Q3 sales of $22.4 million.

Predictive value unclear

While the FDA clearance is a major step for NantHealth, the diagnostic value of TMB as a predictive biomarker remains unclear.

Two studies presented at the International Association for the Study of Lung Cancer meeting in September already shed doubt in the minds of some oncologists regarding the predictive value of TMB.

One study, from Fondazione IRCCS Istituto Nazionale dei Tumori in Milan, Italy, said that TMB was "not significantly associated" with the efficacy of pembrolizumab as first-line therapy for metastatic non-squamous, non-small cell lung cancer.

The other, led by the Abramson Cancer Center of the University of Pennsylvania, also found that TMB, as determined by WES and tumor-normal comparison, was not significantly associated with objective response rate, progression-free survival, or overall survival for pembrolizumab plus chemotherapy or chemotherapy alone in advanced, non-squamous NSCLC patients.

"While perhaps this may not be an actionable biomarker at this point in time in lung cancer, there's still a lot of active research and [there are] other areas, other types of cancers that probably need to be further elucidated," FOCR's Allen said.

Luis Diaz, head of solid tumor oncology at MSK, added that in view of recent data showing the lack of an association between survival outcomes and TMB, "there will need to be a prospective study to determine the role to TMB with immunotherapy."

NantHealth's Reddy maintained that testing is still sometimes justified. "There is general consensus that TMB is a useful biomarker for immunotherapy response in many cancers and across different immunotherapeutics, and so clinically it is a useful adjunctive tool to help assess the likelihood of benefit with this class of agents," he said.

Reddy said that there will be a learning curve as oncologists adapt to actual measurement of TMB from the current standard of calculated TMB. "We do expect a transition period as the field adopts the more precise standard," he said.

Jason Rosenbaum, a clinical pathologist at Penn's Perelman School of Medicine, however, questioned whether a whole-exome test would even produce a more accurate measure of TMB.

"The adoption of any test offering TMB is hindered by the fact that TMB, even from whole-exome sequencing, is a calculation rather than a traditional biomarker or analyte. The assumptions underlying a TMB calculation can make a significant difference to the output and we don't know what the correct assumptions are yet," he said in an email.

While conceding that WES leads to a more accurate mutation count, Rosenbaum said that "it is almost as certain, though, that an accurate accounting of mutations will be less clinically meaningful than a weighted count of the most relevant mutations."

According to Rosenbaum, "The field awaits the proper studies demonstrating what TMB should even be." He declined to comment on any specific assays, including NantHealth Omics Core and MSK-IMPACT. "No one has an adequate solution yet," he said.

Allen remains cautiously optimistic about the usefulness of tumor mutational burden.

"You want to retain the ability to be able to develop a comprehensive test for a complex biomarker," he said. "I think time will tell and the clinical studies will continue to evolve to better define what the exact role is that TMB can and should play." 

Allen added that his Washington, D.C.-based nonprofit has identified more than 100 clinical trials that incorporated measurement of TMB in some form in various cancers.

"This shows that there's a lot of interest within the field," he said, while also recognizing that clinical data about TMB so far has been mixed. "Additional research is needed to better characterize where TMB may be a predictive biomarker or where it isn't," he said. "Ultimately, that will guide the utility of TMB as a potential marker or a diagnostic tool."