NEW YORK – Inivata's research efforts are bearing fruit, laying groundwork for expansion of its liquid biopsy technology from genotyping for lung cancer targeted therapy into immunotherapy prediction and potentially response monitoring for multiple cancer therapies.
A study published last month by Inivata researchers and company collaborators demonstrated that its lung cancer liquid biopsy technology not only detects driver mutations to guide molecularly targeted therapies, but could also help predict immunotherapy response in two ways: via algorithmic assessment of mutation status before treatment, and by measuring response signals during the first few weeks of a patient's therapy.
Inivata is only in the first stages of getting its lung cancer genotyping assay, InVision, into the hands of oncologists. That test received a positive local coverage determination from Medicare contractor Palmetto GBA this spring. Inivata CEO Clive Morris said in an interview that the firm is still transitioning from a "pre-commercial" or pilot phase. Morris noted there is "a lot that has to happen" — such as building billing systems, creating a portal for delivery of results, and developing supply chains — before the firm can get the test into the hands of physicians over the next year.
But with the publication last month in the journal Lung Cancer, along with related conference presentations over the course of this year, Inivata is also making clear that it is already exploring expanded utility and validating new technology offshoots.
In the study, investigators from Inivata and several research institutes and hospitals in Toulouse, France, used the InVision panel to test baseline, and in some cases one-month post-treatment plasma samples from a group of 97 patients treated with an immune checkpoint inhibitor in the second line.
Of the 97 individuals, 87 had samples that met criteria for the Inivata test: 39 known responders who had at least six months of progression-free survival on treatment and 47 non-responders who showed progressive disease at their first evaluation.
According to the authors, the Inivata test picked up alterations in ctDNA in 67 of the 86 baseline samples. Certain molecular features in the cohort were associated with poor response to immunotherapy, including the presence of a PTEN or STK11 mutation. Others, like mutations in KRAS and TP53, predicted better outcomes.
The team used these findings to develop what they called an "immune score" that classified patients into two groups. Patients with a low immune score had median progression-free survival of just two months compared with patients with a high immune score who averaged 14 months PFS.
According to Morris, although there are a lot of steps the company would have to take to bring this new method for immunotherapy prediction into the clinic, the fact that it works within the company's current test design could reduce some hurdles.
"This is different than PD-L1 or [tumor mutational burden] ... We can do this using the same blood sample [as our existing test] but we analyze mutation pattern in a different way," Morris said. "Instead of looking at whether a samples is ALK positive or EGFR positive, or BRAF positive … we integrate all the DNA info from the sample [and use] a five-step algorithm or cascade to put people in a group where they are either good candidates for PD1 inhibition or they are not."
"We can run it at the same time as our existing assay. We have to run different analytics but that can be added fairly easily," he added, since the front end of sample collection and sequencing would stay the same.
That said, this is not an FDA-approved indication," Morris said. "It's potentially a new companion diagnostic approach. We will have to think about how we work that through regulatory frameworks," potentialy in partnership with biopharmaceutical partners.
According to the study authors, limitations of the study include its retrospective nature and a lack of a non-treatment arm to help confirm that the algorithm is truly predictive regarding immunotherapy rather than generally prognostic. Inivata will have to address these lingering questions in future work.
Importantly, the Toulouse study did not stop with the "immune score" assessment. Investigators also tested patients' blood after one month of treatment, measuring early changes in circulating tumor DNA.
Although a number of biomarkers are being investigated with great enthusiasm for immunotherapy, the field has not had as much clear-cut success with upfront predictors as in certain areas of targeted therapy, where particular gene mutations like ALK or EGFR can offer highly accurate, binary classification of whether a patient will respond to a particular drug.
In immunotherapy, biomarkers like PD-L1 expression, tumor mutational burden — and even microsatellite instability, which became officially recognized as a companion diagnostic marker for Merck's immunotherapy drug pembrolizumab in 2017 — predict response much more imperfectly.
Because oncologists don't want to deny patients a drug that has a chance of benefiting them, research continues to try to optimize the existing predictive markers, and to identify new ones in other areas like the microbiome, the immune-cell repertoire, or in immune cell gene expression that can make the choice a more reliable one.
But at the same time, other studies are also looking into whether treatment with these therapies might be better personalized not by a pre-treatment marker at all, but by putting patients on the therapy and looking for early changes in things like circulating DNA or tumor cells, which forewarn what will later turn out to be either a good response or a poor response.
Inivata and its collaborators also investigated this in their study, in 65 specimens from patients who had an on-treatment sample taken. According to the authors, early changes in circulating tumor DNA allele fraction were strongly correlated with clinical outcomes. Individuals whose ctDNA fraction decreased any amount had 14-month PFS, while those whose mutant alleles increased survived just 2 months without disease progression.
"We started off with a one-off testing model, and obviously you have to start somewhere … but [this is] how we see this type of testing evolving over time," Morris said.
In some ways, the integration of both an upfront predictor and early monitoring could be especially powerful, he added, and is one reason why Inivata is interested in investigating its technology for both.
"Ultimately what you want to do is the best for the patients at any one time," he said. Even the best targeted therapy biomarkers are still not perfect —a canonical example like EGFR "might be 60 to 70 predictive," Morris said, but there is still a group that doesn't respond.
The company hopes its early experience in the recent study may presage that something like its immune score method can help bump up immunotherapy prediction closer to that 60 or 70 percent prediction.
Then, with early on-treatment monitoring, liquid biopsy could potentially add even more benefit.
"If we can get up to 70 percent … that's a great benefit to patients," he said. And then, with early on-treatment monitoring, "we can see within a few weeks if [a patient] is one of that 70 percent or not."
Of course, Morris added, many questions would have to be addressed to translate this from promising early research to the clinic. But he said that is where Inivata is hoping to move, and where the liquid biopsy field as a whole seems to be traveling.
"This is definitely being done today [in the research environment], but we have to produce datasets to say not just that this is a good thing to do ... scientifically — that it's valid and robust and reliable— but [also] that it's good money to pay for a test to do this."
Inivata and competitors exploring monitoring may also find themselves able to demonstrate utility for monitoring in other ways. In that vein, Morris also highlighted researcher that is accumulating for his company's technology in tracking things like emerging resistance mutations for patients being treated with various genomically targeted medicines.
In another study published last week by the company in Clinical Cancer Research, researchers were able to identify novel resistance mutations and patterns to the ALK inhibitor lorlatinib using longitudinal liquid biopsy testing. Although it's not yet established that being able to track these changes would improve patient outcomes, the possibility of a new avenue for personalizing treatment is obviously of great interest to oncologists and patients, Morris said.
Inivata has been active in collecting evidence around this type of resistance monitoring for other drugs, as well. The firm shared data in poster presentations this year at the European Society for Medical Oncology and the American Society of Clinical Oncology annual meetings from similar studies in patients treated with BRAF inhibitors and anti-EGFR therapy.
Competing firms like Guardant Health have also begun reporting increasing data in this vein.
According to Morris, although the type of technology that may work best for these types of advanced cancer indications is different than what may be needed in earlier cancer contexts, Inivata is also working on ways to address that area.
Liquid biopsy monitoring for patients after an early-stage cancer is surgically removed, for example, is also receiving increasing attention, with Natera leading the commercialization push with its Signatera assays, which involve the creation of individualized assays for each patient based on a profile of their tumor tissue exome.
Morris said that Inivata hasn't published anything on its work along these lines, but hinted that the company believes it has methods in hand that may work in similar ways to Natera's, but with added advantages. The company hopes to share data on this at conferences this year.