NEW YORK – A little more than a year after launching clinical cancer liquid biopsy testing, diagnostics firm BillionToOne hopes to soon gain Medicare coverage for one assay, called Northstar Select, and to publish the first data for Northstar Response, its methylation-based treatment response monitoring test.
The company, which was initially focused on noninvasive prenatal testing, began its foray into oncology in 2021, having long viewed cancer testing as a natural application for its technology, which uses quantitative counting templates, or QCTs, synthetic DNA molecules that are spiked into a cell-free DNA sample prior to PCR amplification.
The QCTs share certain sequences with a gene or genes of interest and thus amplify at the same rate. Counting the number of QCT molecules in the resulting sequencing data can help determine the number of DNA molecules that are derived from a particular gene in the sample.
"In cell-free DNA testing in particular, you have to amplify the sample a million- to a billion-fold before you can sequence it," said BillionToOne CEO Oguzhan Atay, "and that amplification and sequencing process is very error prone, and there's so much effort that goes into making it just a little better."
"People design new polymerases that have slightly lower error rates. There are sequencing company wars about who has the lower error rate. And then there is a huge selection of companies that try to differentiate themselves over slight differences in bioinformatics. They are trying to squeeze a little bit more signal from that sequencing data," he added.
BillionToOne has tried to make a run around these issues by correcting for errors before any of these processes.
"It's garbage in, garbage out," Atay said. "Once you have those errors baked into your sequencing reads, it's very difficult to get a much higher signal."
After performing clinical validation studies, BillionToOne launched its two tests early last year through its CLIA-certified lab. Both tests are intended for late-stage cancer patients — the first to inform molecularly targeted therapy decision-making and the second to monitor circulating tumor DNA changes in patients on treatment as a readout of their response.
The company is hoping to soon publish its first study describing the validation of the second test, Northstar Response, and offered a preview of the data in a preprint posted last year.
Atay said BillionToOne doesn't like to get too into competitive dynamics, but he said the study touches on the advantage of the company's QCT methodology for monitoring late-stage patients because it allows for a broad, tumor-naïve assay.
"We firmly believe that in that setting, you have to be tumor naïve and looking at a broad cancer signal … because these patients are on very strong therapies and their tumors are evolving," he said. A tumor-informed approach that takes a snapshot of a cancer at a particular time just doesn't make sense, he added. "It misses that tumor evolution and can result in problematic results."
In their preprint, company investigators described the development of Northstar Response, which targets more than 500 genome locations, which the firm derived from TCGA data, that are hypermethylated in cancer compared to controls.
In a variety of contrived samples, the assay was able to accurately read out different tumor fractions, with its score closely matching the stepwise increase in tumor fraction from sample to sample.
The assay made accurate calls for step increases as small as 0.5 percent tumor fraction in the contrived samples. Overall, the investigators calculated a limit of detection between 0.25 percent and 0.5 percent tumor fraction for the test, and further experiments confirmed that it could accurately detect changes in tumor fraction across 12 cancer types.
In a small set of clinical samples, the company was able to detect initial decreases in tumor fraction in patients after they began treatment and subsequent increases later on. The test's methylation scores also showed high concordance with clinical outcomes. All six patients who had an increase in methylation detected progressed shortly before or after their sample was collected. In three of the six, the increase was measured before clinical signs appeared.
Atay said that although these initial results were strong, the version of the paper going through peer review features updated data using a newer version of the test that achieved a five times lower limit of detection.
Although the oncology field is still mulling the potential utility of this type of real-time response monitoring, the hope is that compared to imaging, it could accelerate the pace with which physicians can detect progression and move their patients to another therapy. Future interventional clinical utility studies will be needed to bear that out.
Atay didn't share a precise number of customers, but said that adoption of both cancer tests has been much faster than anticipated, given that the firm's primary validation data has yet to be published.
Researchers at the University of Florida in Gainesville are some of the early adopters, and they shared early observational data at a meeting of the American Society of Clinical Oncology this January, concluding that the Northstar Response assay could consistently detect ctDNA across a variety of advanced GI cancers.
The company has a significantly smaller sales team in oncology than in its legacy business of prenatal testing. "But even with that, we are seeing a lot of key opinion leaders, as well as oncologists, starting to use the tests," Atay said. He guessed that this has to do with the tests' improved sensitivity compared to commercial alternatives.
Data on the company's therapy selection test is currently under review by the Centers for Medicare and Medicaid Services' MolDx program, but Atay said that BillionToOne performed a head-to-head study in which clinicians sent it samples from the same blood draw that they also supplied to their choice of commercial company for testing.
In that study, investigators concluded that Northstar Select could detect about 50 percent more actionable mutations for therapy selection compared to other liquid biopsy tests.
"Our limit of detection is two to five times lower than [that of] other liquid biopsies, and it turns out that there are quite a few mutations that are in that range that are being missed by existing liquid biopsies," he claimed, without naming those other tests.
That advantage may be driving initial adoption of the assay, he hopes, and then therapy response monitoring would become an added bonus. He estimated that about 95 percent of customers have adopted both tests.
Northstar Select is eligible for Medicare coverage under an existing umbrella LCD, as long as it passes its MolDx technology assessment. Since commercial payors have begun to follow that policy, Atay said, BillionToOne is already being reimbursed for some of its therapy selection testing.
On the therapy response testing side, he said the company anticipates it will take a bit longer to gain reimbursement because there aren't policies in place already that cover all the indications in which the test can be used.
Looking to the future, Atay said that minimal residual disease testing would be a logical next step, since the base technology of the response monitoring assay would port to that application in a "relatively straightforward manner."
He added that while that could be a future area of growth for the firm, BillionToOne believes that response monitoring in later-stage patients is "actually a very big unmet medical need right now" because most existing assays are tumor-informed and unable to adjust to tumor evolution.