publication date: Jul. 7, 2017
NCI-MATCH keeps enrolling—Targeted mutations are proving to be less common than estimated
By Paul Goldberg
The NCI-MATCH trial has met its goal of screening 6,000 patients, but it will not stop there.
The landmark precision medicine trial, which is essentially a collection of single-arm phase II studies, recently cleared four laboratories to identify patients who are getting tested as part of their care. If actionable mutations are found, these patients could become eligible for NCI-MATCH.
MATCH stands for Molecular Analysis for Therapy Choice.
NCI-MATCH is being expanded because of an important discovery: In the patients tested so far in NCI-MATCH every tumor gene abnormality has turned out to be less common than expected in this study population. The mutation prevalence rates found in the trial range from zero to 3.47 percent.
This means that tens of thousands of patients will need to be screened to fill rare tumor cohorts, said Keith Flaherty, the ECOG-ACRIN Cancer Research Group study chair and director of Developmental Therapeutics at the Massachusetts General Hospital Cancer Center, and professor of medicine at Harvard Medical School.
There are 35 patients in each rare variant subprotocol.
How long can NCI-MATCH keep going?
“We’ll certainly stop when we get the 35 patients in all of our subprotocols,” Flaherty said to The Cancer Letter. “Once we complete all the subprotocols that have been launched, then that will be the formal end of the study.”
A conversation with Flaherty appears here.
In the trial’s original iteration, NCI paid for biopsies and tumor sequencing for patients, and if actionable mutations were found, these patients were assigned to receive investigational targeted treatments based on genomic characteristics of their tumors. Now, NCI-MATCH will accrue patients for treatment based on sequencing they get as part of their care.
Patients referred by laboratories will enroll in NCI-MATCH for a screening assessment. If they meet all of the subprotocol inclusion and exclusion criteria, they may be enrolled on the treatment arm, where they would receive drugs at no charge.
The trial has 30 treatment arms, and eight of them—the ones with more common mutations—have reached their 35-patient enrollment goals. One arm closed completely because the company discontinued development of the drug.
Of the eight arms, four continue to enroll patients because their accrual goals were increased to 70 patients to ensure that those with matching gene abnormalities would have a treatment opportunity. This leaves 25 arms that are currently accepting patients.
The latest information on all treatment arms and the number of patients accrued within them is posted here.
To help the trial adjust to the higher volume of patient cases coming in from the external laboratories, the new study entry process is ramping up via a demonstration project, whereby the labs are initially referring only those patients who match to 19 of the arms that address gene variants with the lowest prevalence rates found so far in the trial (1.5 percent or less).
The potential for the designated labs to report on the remaining open treatment arms will be considered, along with six new subprotocols in development, once the demonstration project is complete, ECOG-ACRIN officials said.
In the old days, trials had starting rules, stopping rules, accrual targets. There were also fewer arms. In most cases, two seemed to suffice. There were protocol amendments, sure, but few trials if any were this open-ended.
NCI-MATCH is something completely different, said Barbara Conley, associate director of the Cancer Diagnosis Program in the NCI Division of Cancer Treatment and Diagnosis. Conley is the NCI study chair for the trial.
“It’s a different mindset. This is a different kind of trial,” Conley said to The Cancer Letter. “You know in the background are these other precision medicine trials, some of which seem to show that it’s better to do it that way, others of which have been kind of a bust. And so, when we designed this trial, we designed it carefully to have levels of evidence for every—what we call actionable mutation of interest and every treatment. There’s no treatment in MATCH that hasn’t shown a signal of activity somewhere in some tumor in a patient.”
A conversation with Conley appears here.
Altogether, over 1,100 treatment sites are involved in MATCH, having received the required NCI central IRB approval to do so. The trial has required multi-center involvement from the start because of its size and scope.
In the past, MD Anderson Cancer Center received the biopsies for initial processing, and the isolated nucleic acids were sequenced at one of four laboratories running a single NGS assay: Yale University, Mass General, Leidos/Frederick and MD Anderson Cancer Center received the tumor samples, and then MD Anderson.
All laboratories used the same variant calling and results were fed into the software platform MATCHBox for assigning patients to treatment. The trial’s panel of experts reviewed every patient case before the results were sent back to the treating site.
As the trial changed, NCI cleared two commercial vendors—Foundation Medicine Inc. and Caris Life Sciences—to identify patients through their normal testing ordered for clinical reasons. Collaboration agreements with two academic labs—MD Anderson Cancer Center and Memorial Sloan Kettering Cancer Center—allow testing of their own patients using their institutional assays.
This type of study entry process has never been done before in an NCI-sponsored trial.
Additional testing providers may be qualified in the future, but for now, the demonstration project must be completed first, ECOG-ACRIN officials said. The demonstration project aims to build an understanding of implementation standards that will be required for broader use.
ECOG-ACRIN, the clinical research group that runs the trial, which is also known as EAY131, had no trouble accruing patients. The 6,000-patient screening target was met nearly two years ahead of schedule, with an uncharacteristically large number of patients coming from community sites, many of them in the upper Midwest.
Initially, NCI and ECOG-ACRIN hoped to screen 25 percent of patients with rare or uncommon cancers, defined as those other than breast, colorectal, non-small cell lung or prostate.
“We are surprised that over 60 percent of patients came into the trial with less common types of cancer, which far surpassed our goal,” said Peter O’Dwyer, co-chair of ECOG-ACRIN and director of the Developmental Therapeutics Program at the Abramson Cancer Center and professor of medicine at the University of Pennsylvania.
Though it’s too early to see publications, NCI’s James Doroshow said at a recent advisory committee meeting that a successful result has been obtained in at least one of the trial’s treatment arms.
“I was told over the weekend that I am allowed to tell you that there actually are objective responses on this trial. I can’t tell you what the responses are, or the arms, because since I’m not on the steering committee, I don’t know which arms were successful so far,” Doroshow, NCI deputy director for clinical and translational research and director of the Division of Cancer Treatment and Diagnosis, said at a joint meeting of the National Cancer Advisory Board and the Board of Scientific Advisors June 20. “But there are objective responses, and they’ll start to get reported out over the next 12 to 18 months.”
NCI-MATCH is a signal-finding trial whose findings could enable moving drugs to larger, more definitive trials.
Doroshow’s comments appear here.
In May, FDA, for the first time in history, approved an agent based purely on a mutation, without regard for tissue or site involved. The “site-agnostic” approval was granted to Merck’s pembrolizumab (Keytruda) for adult and pediatric patients with unresectable or metastatic, microsatellite instability-high or mismatch repair deficient solid tumors that have progressed following prior treatment and who have no satisfactory alternative treatment options or with MSI-H or dMMR colorectal cancer that has progressed following treatment with a fluoropyrimidine, oxaliplatin, and irinotecan.
ECOG-ACRIN and NCI fully vetted the four labs based on three criteria:
A mechanism that can review tens of thousands of patient cases is necessary to identify small subsets with tumor gene abnormalities.
These labs are already testing a high volume of patients.
The comprehensive and highly-validated genomic profiles are particularly sensitive to the tumor gene abnormalities being studied in the trial.
The labs are able to provide assay results in a format that can be uploaded into MATCHbox, the trial’s informatics system that generates treatment assignment information for the trial’s panel of experts to review.
“I think NCI-MATCH is paradigmatic of how we do trials and how we make drugs available,” said said Vincent Miller, chief medical officer at Foundation Medicine. “Molecular profiling for inclusion into clinical trials not only improves patient access to needed therapies but it also potentially speeds development for drug makers by optimizing trial recruitment and enrollment based on biomarker testing.”
It would be impractical to screen all patients for specific mutations. However, checking for possible genomically-guided treatment options can be made an element of standard care.
“The fact that MATCH is open at 1,100 sites makes it something that’s not just opened at Memorial and MD Anderson and Dana-Farber,” Miller said. “This is for you, meaning a community-based medical oncologist. It brings these guys into the fold where cancer genomics are central for practice—they are not peripheral or tangential—and also making clinical trials much more real, rather than something done at big cancer centers, because 85 percent of cancer care is rendered in the community. If genomic profiling becomes standard, then 85 percent of these patients who are needed for trials should be identified in the community setting.”
Though no one keeps track of the number of patients whose tumors are sequenced in the course of their care, Flaherty estimates that about 100,000 such tests are performed annually.
This is likely close to what is actually happening. During the first quarter of 2017, Foundation Medicine, a public company, reported that it had performed 13,933 tests. This represented a 55 percent year-over-year growth. The company said it expects to deliver approximately 53,000 to 56,000 clinical tests in 2017.
“It is paramount that laboratories across the country are providing accurate and reliable results because they have direct impact on the care provided to cancer patients,” said John Marshall, chief medical officer of Caris Life Sciences, associate director for clinical care for the Georgetown Lombardi Comprehensive Cancer Center and the chief of the Division of Hematology-Oncology at MedStar Georgetown University Hospital.
Caris, a privately held company, said it expects to perform 30,000 tests that would be relevant to NCI-MATCH.
Prevalence of mutations in NCI-MATCH was recently assessed in a paper from the American Association for Cancer Research Project GENIE, a data-sharing consortium focused on generating an evidence base for precision cancer medicine by integrating clinical-grade cancer genomic data with clinical outcome data for tens of thousands of cancer patients treated at multiple institutions worldwide.
To demonstrate that the data in GENIE has clinical utility, the study, ran an estimate of clinical actionability across multiple cancer types (>30%) and prediction of accrual rates to the MATCH trial. The results accurately reflect recently reported actual match rates.