Nonprofit advocacy group Friends of Cancer Research held a meeting last week on the future of oncology and other diagnostic tests, reporting progress in some of its assay harmonization programs but highlighting struggles faced by other proposed programs in an evolving and uncertain regulatory landscape.
The nonprofit hosted speakers from the US Food and Drug Administration, industry, and academia. Jeff Shuren, director of the US Food and Drug Administration’s Center for Devices and Radiological Health, updated attendees on what he called the “elephant in the room,” the agency’s proposed rule for the oversight of laboratory-developed tests (LDTs), which it issued last October.
During the 60-day comment period on the rule, the FDA received more than 6,500 comments, which it is now reviewing as it moves to finalize the legislation.
An exploration by GenomeWeb last month examined patterns in the comments, with roughly 2,900 in favor of the rule and about 3,700 against, many of which came in the form of copies of the same form letter.
Although Shuren said he could not say much about the agency’s progress, he reiterated the FDA’s rationale for enforcing its oversight of LDTs, based on issues it has seen across disease areas with tests’ analytical and clinical validity.
As an example, he referred to New York state’s laboratory test assessment program, which he said has seen more than half of the tests that come through its doors over the last 30 years fail to pass its internal bar on a first try.
One area of particular concern has been in oncology, where Shuren said that data shows that whether or not a patient gets the right cancer treatment “can depend more on the lab you get tested in than on your tumor biology.”
As an interim step as the larger proposed rule moves toward finalization, the FDA also launched a voluntary pilot program last year, which is intended to reduce the risks of LDTs used to select patients for oncology therapies.
Under the pilot the agency asked pharmaceutical manufacturers to volunteer to provide performance information for tests used to enroll patients in clinical trials supporting drug approval.
The FDA would then use that data to try to come up with minimum performance specifications that could guide the development of similar tests. As of last week, Shuren said, no companies have been enrolled in the pilot, although some have approached and are in discussions with the agency.
Challenges to participation, he said, include the fact that when companies know what test is being used — which they don’t always — the labs involved have not been willing to share the data.
For its part, FOCR has been trying to address the issue of oncology test quality through harmonization projects that seek to compare existing tests across unified datasets, determine areas of discordance, and seek solutions so that different tests can be realigned to yield equivalent results.
Concurrent with the meeting, FOCR said it was launching the newest of these, a research project focused on improving biomarker assessment across computational pathology platforms, which will focus initially on HER2 biomarker testing. Prior efforts have tackled tumor mutational burden and homologous recombination deficiency, or HRD.
Hillary Andrews, FOCR’s director of regulatory and research partnerships, shared new data from the HRD harmonization effort during the meeting last week.
HRD is what FOCR considers a “complex biomarker” because tests developed for it can assess a variety of different measurable indicators to create an HRD score, including alterations in specific genes or their downstream effects.
“The complexity leads to different definitions of what constitutes HRD,” Andrews said. “There are different assays that have different cut points or thresholds. And that leads to inconsistency and how HRD is measured and interpreted by patients and providers … that may impact treatment decisions and ultimately patient outcomes.”
Investigators presented some initial in silico data from the project in November 2022 and have now followed that up with clinical samples from 90 ovarian cancer patients.
According to Andrews, all patients in the new analysis were treatment-naïve and had samples collected between 2012 and 2022. Seventeen participating assay developers independently sequenced the samples and then measured and reported HRD, as well as information about the content of their respective assays, which varied significantly.
A key aspect of the analysis, Andrews said, is that the field lacks a gold standard for HRD. This meant the team had to assess test concordance based on positive percent agreement and negative percent agreement amongst pairs of assays.
Overall, investigators calculated that the median and interquartile range was 83 percent for positive agreement and 80 percent for negative agreement, a result that Andrews called “moderate agreement.”
Exploring the impact of potential interfering factors, the team found that assay agreement was better, in general, when patients had a BRCA mutation. The inter-assay agreement for HRD-negative calls was also improved among samples with CCNE1 amplification.
Things like race, tumor purity, and DNA quality did not affect test agreement, nor did tests’ status as research-use-only versus clinically approved.
In a follow-on panel discussion, MD Anderson pathologist Alex Lazar said that in his opinion, it’s actually “remarkable” that the group saw the degree of concordance that it did “because all of these tests are constructed very, very differently from each other.”
“I think it’s important to remember that no assay for HRT that I’m aware of perfectly predicts who’s going to benefit, or who’s going to be sensitive to these therapies,” added Lisa McShane, an associate director at the National Cancer Institute. “With that in mind, we have to realize that even if two assays have a fair amount of disagreement, maybe they’re picking up different parts of the population that are still going to [benefit] from PARP inhibitors or [be] sensitive to platinum agents.”
“Often people feel like, well, my assay is better because I can identify more people who are HRD[-positive], therefore, more people have access to the drug,” McShane said. “That is not the goal. The goal is to ask, ‘Can we identify more accurately those who are going to benefit from these agents?'”
Rebecca Arend, an associate professor at the University of Alabama, Birmingham, said that the overall takeaway is that assay developers have more work to do to understand what these discordances mean. “All the trials we’ve done usually only use one assay … so I urge all these diagnostic companies … if we can incorporate multiple assays in these prospective trials where we do have hundreds or thousands of ovarian cancer patients, that’s what we need.”
As HRD tests, other companion diagnostics, oncology assays, and molecular diagnostics in general face up to the anticipated finalization of the FDA’s enforcement rule, a bright spot FOCR panelists discussed was the agency’s announcement last week of its plans to reclassify some high-risk in vitro diagnostics as moderate risk.
If those diagnostics are reclassified as moderate risk, or Class II, test makers could then obtain clearance through the 510(k) premarket notification pathway rather than the more burdensome premarket approval pathway.
Brittany Schuck from the FDA’s Center for Devices and Radiological Health said that most oncology companion diagnostics, as well as some infectious disease tests, would likely be eligible for this downgrading.
Carly McWilliams, head of North America regulatory policy at Roche Diagnostics, cheered the move, saying that “putting that many more tests in Class II, it would just completely shift resources. It would fundamentally make that program so much more sustainable.”
“I think just that kind of creative thinking by the agency just reinforces how these new regulatory frameworks could be so impactful to innovation in the field,” she added.
Regarding the stalled oncology diagnostics pilot, Anthony Sireci, senior VP of clinical biomarker and diagnostics development at Loxo Oncology, said that if the FDA can overcome its recruitment challenges, the benefits to the precision oncology field could be great.
The program basically codifies what Loxo has tried to do internally, he added. “Many patients are diagnosed before we launch our drug trials so why on earth would we ask them to be retested if in fact they’ve been tested on high-quality tests,” Sireci said. As such, Loxo uses a “mix of validations” before allowing enrollment, “but essentially we think that this pilot allows us to codify that and allows other sponsors to do that.”
“A bad-quality test isn’t great, but no test is also not great, so I think it does a nice job of striking that balance and we hope to … help shape what we’re asking from these labs,” Sireci said.