NEW YORK – Foundation Medicine researchers are investigating how the company's estimation of tumor fraction, something it provides on patient reports as a quality metric, might have more specific clinical uses.
In a study of more than 23,000 samples sequenced as part of the company's routine clinical testing, investigators found that cases with elevated tumor fraction are extremely likely to have contained enough circulating tumor DNA that a negative result can be relied on without the need to reflex to confirmatory tissue sequencing.
Reporting the results in JCO Precision Oncology last month, the authors wrote that this could allow clinicians and labs to prioritize tissue reflex testing for low-ctDNA samples, potentially sparing patients unnecessary repeat biopsies and labs unnecessary tests.
FoundationOne liquid and other analogous competitors in the market have become more widely used in recent years, especially after the US Food and Drug Administration began to approve companion diagnostic versions of assays like Foundation's.
Nevertheless, a remaining challenge for blood-based cancer sequencing is that the scarcity of tumor DNA in the blood relative to a tissue biopsy sample means negative test results can't necessarily be trusted by clinicians — something stressed in the FDA's labeling documentation for approved CDx assays, as well as by payors who have policies for reimbursing these tests.
Companies providing these tests have advocated for stepwise or even simultaneous testing protocols where both liquid and tissue tests are ordered upfront.
Under these schemes, patients benefit from the relative ease and rapidity of liquid biopsy without fear that an actionable finding might be missed. But the healthcare system also bears the weight of potentially unnecessary tests, Hanna Tukachinsky, senior translational scientist for clinical development and emerging biomarkers at Foundation Medicine, said in an interview.
Thus far, there hasn't been a proven way to determine if a liquid biopsy negative result is a true negative or just a product of not enough mutated DNA entering the blood. But Foundation has begun to collect evidence that its method for estimating tumor fraction (TF) in liquid biopsy samples — an aneuploidy-based calculation — could potentially serve that purpose.
In the company's study, investigators identified a subset of the larger 23,000-patient clinical cohort who had appropriately timed matched tissue sequencing — 613 lung cancers, 292 breast cancers, 279 colorectal cancers, and 105 pancreatic cancers.
The results showed that the FoundationOne liquid's sensitivity in detecting driver alterations ranged from about 58 percent to 86 percent overall, when compared to matched tissue results. However, in cases with TF of at least 10 percent, sensitivity was consistently at or near 100 percent.
Given those results, the researchers decided to test whether elevated TF could be used to rule out the presence of an actionable alteration, focusing on National Comprehensive Cancer Network-recommended biomarkers for non-small cell lung cancer (NSCLC) and colorectal cancer (CRC).
Among 613 NSCLC cases, the overall negative predictive value (NPV) of liquid biopsy was only about 65 percent. But for the subset of 129 patients with at least 10 percent TF predicted, NPV rose to 97 percent. For 279 CRC cases, overall NPV was also 65 percent, rising to 100 percent in the subset of 120 pairs with elevated TF.
Tukachinsky said that Foundation doesn't really have a stake in whether patients get liquid biopsy or tissue testing or both.
"As a portfolio company we do both, so we definitely don't mind if tissue is reflexed to," she said. "What we do want to emphasize is that in some cases it does depend on the biomarker you're looking for. You're interested in a BRCA deletion? Guess what, you almost always have to reflex to tissue."
"But if you're a physician and you're sending in a liquid biopsy and it comes back negative, but you do have that tumor fraction elevated flag on your report and you're only interested in, let's say PIK3CA hotspot mutations, what we want to say is that there's reduced value in going back to the tissue. … You're welcome to do it. But it's about which patients do you prioritize," she said.
Right now, this type of decision-making would be left to individual oncologists. Foundation Medicine isn't noting on patient reports that a TF of 10 or above means they don't necessarily have to reflex a negative result.
"You have to publish evidence before you can convince regulatory bodies of clinical utility [so] right now it's a line on the report that I do think that a lot of physicians are not sure about how to use effectively. We're mainly advocating for it to be used as a quality metric," Tukachinsky said.
But there could be a future where some more directed use of TF becomes possible. "Looking forward we certainly want to make the report even more understandable," she said. "There's definitely limitations on what we can say right on the report, but we … counsel physicians who have the time and energy to call in. And eventually in the future, this is part of building that evidence to show how tumor fraction is useful."
Aside from the question of tumor fraction and the potential to better triage tissue reflex testing, Tukachinsky added that the overall data from the full 23,000-patient cohort were encouraging in and of themselves, showing a promising rate of actionable findings.
"There's a perception that you send liquid biopsies and all you get is [clonal hematopoiesis] mutations or nothing back. And I don't think that's true. [We are seeing] more than half coming back with something relevant. So that's encouraging," she said.