New research led by Friends of Cancer Research demonstrates that decreases in circulating tumor DNA after initiation of treatment are associated with improved overall survival in patients with advanced non-small cell lung cancer treated with immunotherapy or chemotherapy.
The result is an early but promising signal that ctDNA—as a measurement of minimal residual disease—may be an effective intermediate endpoint for accelerated approval in NSCLC.
Until now, MRD has been used as an intermediate endpoint for approval in multiple myeloma and acute lymphoblastic leukemia, and ctDNA would constitute a jump to solid tumors (The Cancer Letter, April 26, 2024; March 29, 2018).
The study is one component of the larger, Friends-led “ctDNA for Monitoring Treatment Response,” or ctMoniTR, project. ctMoniTR is an overarching research effort that involves pharmaceutical companies, diagnostic labs, government health officials, patient advocates, and academic researchers.
Advancements in cancer treatments have led to a paradigm where new trials are taking longer and longer to reach the point of completion.
Jeff Allen
Friends has taken up the managerial mantle necessary for endpoint development and has begun to collect data from clinical trials that incorporate ctDNA as an intermediate endpoint to assess how advanced disease responds to treatment. The group is also collaborating with statisticians at Cancer Research And Biostatistics, a non-profit research organization, to assess across multiple clinical trials whether ctDNA change reflects response to treatment.
The project is an attempt to synchronize and coordinate the sprawling efforts and interests of the many groups involved in developing data to support the use of a novel biomarker—from the outset. The ctMoniTR project has already begun aggregating data across multiple trials early in the endpoint development process. Indeed, in the most recent study, Friends researchers aggregated data from four clinical trials
Friends hopes that, with this approach, generation of evidence will proceed “faster than if any single organization tried to do so alone,” according to the Friends website.
The need for intermediate endpoints is becoming increasingly urgent in oncology drug development, Jeff Allen, president and CEO of Friends of Cancer Research, said to The Cancer Letter.
“Advancements in cancer treatments have led to a paradigm where new trials are taking longer and longer to reach the point of completion,” Allen said. “So, early indicators of a beneficial therapy are becoming all the more important in order to keep progress moving forward and to make sure that there is timely patient access to new beneficial therapy.”
However, novel endpoints come with increased risk of exposing patients to ineffective or harmful therapies. According to Tatiana Prowell, a medical officer in the FDA Office of Oncologic Diseases and associate professor of oncology in the Division of Women’s Malignancies at the Johns Hopkins Kimmel Comprehensive Cancer Center, novel endpoints present the following challenges:
- They may not be associated with long-term clinical benefit
- They may compromise assessment of conventional endpoints
- They may discover new serious adverse events in postmarket setting
- They may pose operational challenges & create delays in development
- They may increase postmarketing requirements
- They may result in unfavorable global regulatory or payer decisions
- They may result in loss of public confidence & reputational risks
Prowell spoke at a Sept. 11 workshop on endpoint validation hosted by FDA and the American Association for Cancer Research.
Finding the appropriate risk-benefit balance is an ongoing challenge, Richard Pazdur, director of FDA’s Oncology Center of Excellence and the acting director of the Office of Oncologic Diseases in the Center for Drug Evaluation and Research, said at the workshop.
Said Pazdur:
The evaluation of endpoints is really a journey, and we’ve been having this journey ever since I came to the FDA in 1999. I remember when I was applying for the job at that time, there was actually competition for government jobs, believe it or not. I had to travel to Rockville [MD] from Houston, Texas and gave a lecture on endpoints in oncology drug development.
And what was said basically then to me is ‘Rick, what the FDA expects for drug development and oncology is not one randomized trial, [but] two randomized trials that show an overall survival benefit.’
Well, things have obviously changed. And I think they’ve changed for the better in terms of flexibility. And I think we’re going to be talking throughout this morning about the balance that one needs about the rigor of endpoint development as well as the expediency of getting patients necessary drugs.
And always there is an inherent tension in the amount of information that we need versus the needs that patients have.
In recent years, two endpoints have been accepted by FDA for use as an intermediate endpoints to support accelerated approval—minimal residual disease in multiple myeloma and pathological complete response in the neoadjuvant treatment of high-risk, early-stage breast cancer.
In April 2024, the FDA Oncologic Drugs Advisory Committee unanimously voted in favor of the use of minimal residual disease as an early endpoint in trials seeking accelerated approval in multiple myeloma.
The ODAC meeting chronicled the “herculean effort,” over a decade in the making, driven by a Sylvester Cancer Center researcher, Carl Ola Landgren, who—skepticism of many colleagues notwithstanding—believed in the potential of MRD to make effective drugs available more quickly to multiple myeloma patients with no other options (The Cancer Letter, April 26, 2024).
Landgren started the EVIDENCE meta-analysis in 2009 as an interagency initiative between himself, as an intramural NCI employee, FDA, and researchers at the National Heart, Lung and Blood Institute.
“I was convinced that this is really the way to go,” Landgren said in a conversation with The Cancer Letter at the time.
Said Landgren:
It actually came out of my observation as a physician-scientist, when I saw what happened to our patients that we were treating in the Clinical Center, in NIH Building 10 in Bethesda. We started treating patients with better and better drugs. We had access to very powerful drugs back in 2009 already.
We started seeing that the majority of our patients achieved an overall response. The overall response in multiple myeloma requires only a 50% reduction of the disease, and I started seeing about 90% of our patients were going in that direction—we were close to 95%, and eventually like 98% or so.
And then I was looking to see how many of the patients were going into remission, the complete response, and that number started going up as well. We were getting close to 90, 95%.
So, then it crossed my mind that if every individual with whatever condition you follow as a physician, if every person you test has a positive test, that’s the day you can stop doing the test, because you already know the answer.
Also, I was thinking if all our patients are eventually going to be in remission, either we stop testing them because we already know, or we start thinking about better tests. So, I got really interested in developing minimal residual disease tests.
While the story of MRD as an endpoint for accelerated approval in multiple myeloma is now recognized asa success story, the research initiative was largely disjointed, piecemeal, and could have benefited from early synchronization efforts.
When the teams involved in the development of the MRD endpoint brought the application to FDA, there were inconsistencies in the data collection timing and biomarker cutoffs of the clinical trials that were being submitted as part of the meta-analyses supporting MRD’s use as an early endpoint in MM.
The ctMoniTR project is an attempt to prevent these issues of inefficiency from the get-go.
“MRD was a good stepping stone,” Allen said. “It was actually one of our motivations for setting up this study when we did, because that’s been a long story in the making—to have that number of clinical trials.
“And what you saw through that process is that it wasn’t a super coordinated effort. It was, ‘How do you take all of this information around the recent experience around minimal residual disease, and what do you make of it, and what are the analyses that you can then conduct across numerous clinical trials?’”
Friends of Cancer Research saw that ctDNA was at risk of falling into the same administrative messiness and stepped in.
Said Allen:
When we started this project, there were some reasonable studies that were starting to point to the fact that reducing, at a molecular level, the presence of cancer to a sufficient degree following treatment that it would correspond with with long-term benefit. But at the time, they were kind of smaller exploratory phase II studies that weren’t intended or powered to necessarily fully characterize the change in ctDNA and its association with long-term benefits.
When we put this together, what we wanted to aggregate, because it took so long in the MRD space, was, could we have a more concerted effort by bringing together multiple different trials in order to sort of work through this process a little bit more?
The current study, published Sept. 21 in the Journal for ImmunoTherapy of Cancer, was an analysis of pooled data from four randomized clinical trials involving 918 patients treated with anti-PD(L)1 therapy (with or without chemotherapy) or chemotherapy alone. Researchers assessed outcomes associated with ctDNA molecular response vs. non-molecular response using three MR thresholds (≥50% reduction, ≥90%, and 100% clearance) at two post-treatment timepoints: an early window (up to seven weeks) and a later window (seven to 13 weeks).
When we started this project, there were some reasonable studies that were starting to point to the fact that reducing, at a molecular level, the presence of cancer to a sufficient degree following treatment that it would correspond with with long-term benefit.
Jeff Allen
“Our findings suggest that across all three molecular response thresholds, changes in ctDNA are associated with OS in patients with advanced non-small cell lung cancer treated with either anti-PD(L)1 or chemotherapy. Additional research is planned as part of the ctMoniTR project to evaluate molecular response thresholds across different tumor types and treatment modalities,” Nevine Zariffa, founder of the NMD Group and Friends’ consultant for the ctMoniTR Project, said in a statement.
Trial-level meta-analyses and prospective studies will also be essential next steps to advancing and validating ctDNA as an early endpoint for regulatory decision making, according to Friends.
“There are things that I think still are regularly debated around the aspects of the performance of the test used to detect ctDNA—even the appropriate timing for collecting ctDNA. I think those will require some prospective analysis in order to set up what the parameters should be for evaluating ctDNA into the future. And that’s what we’re planning to hopefully try and instigate into the future here.
Friends is primarily attempting to lay out the “trajectory of evidence” necessary to get ctDNA over the “finish line” of early endpoint development.
“Again, we are in a different position because we are trying to encourage others that have these trials of getting-up-and-running to embed this into [their] trials,” Allen said. “And to do that, we’re hoping to lay the trajectory of evidence that will be needed in order to get across the finish line.
“Obviously, I’m biased, but I think we’ve done a good job at providing the foundational evidence and setting up some of the potential methodology to do so. And as we sort of wrap up this component of the project, we’re hoping to really sort of adapt that and encourage, you know, where can we start moving quickly on the prospective side? To confirm some of the things that we’ve seen from this initial analysis?”
While the ctMoniTR project is a step in the right direction in terms of planning ahead for coordinated endpoint development, challenges remain.
“Given our vantage point as the third party convener of all of this data, we were sort of left with the data that we had access to,” Allen said. “So, we are very fortunate for all of the partners that donated data to this project—about a dozen different companies that enabled access to their trials. It was all retrospective data, so we sort of had to make the best with what we had.”
Indeed, there is still opportunity for improvement.
FDA is in a unique position to help coordinate and harmonize early efforts, Mark Stewart, vice president of science policy at Friends of Cancer Research, said to The Cancer Letter.
Said Stewart:
One opportunity is thinking about some type of dedicated resourcing that could assist FDA in optimal partners in this process. FDA, obviously as companies are exploring these various endpoints and biomarkers, they’re in the position to see all of this playing out on their end as companies come to them with their study protocols. I’m not saying they need to be the ones that solve the issues around standardization and harmonization, but they could certainly be the organization that kind of triggers or helps prioritize endpoints that they’re seeing come across that would benefit from harmonization or standardization.
Because that is the rate-limiting step.
You’ll have all these in independent studies read out, and on their own, they provide insights and excitement, but they don’t do much given what FDA’s expectations are to actually help validate the endpoint. And so, you have to kind of leverage all those different studies then say, “Well, here’s where we need to standardize.”
Had all that happened, say four years prior, when those protocols were being developed, we’d be where we want to be much quicker.
So, how can we make that front end more efficient? And what triggers can be in place to help facilitate those discussions to happen then rather than when all the studies read out? And then you gotta wait for a whole nother round of prospective trials to be initiated to do anything with the data.
Though purists swear by the Prentice Criteria to validate surrogacy, FDA officials have said that it is generally impractical to use these criteria in real-world situations, and post-accelerated-approval randomized trials would weed out agents that don’t improve outcomes.
The Prentice Criteria state that (1) the surrogate must be a correlate of the true clinical endpoint, and (2) the treatment effect on the surrogate should capture the full effect of treatment on the clinical endpoint.
The evaluation of endpoints is really a journey, and we’ve been having this journey ever since I came to the FDA in 1999.
Richard Pazdur
Additionally, thanks to recent regulation surrounding confirmatory trials, the field can weather more risk on the front end, Allen said. The Food and Drug Omnibus Reform Act (FDORA), which was enacted on Dec. 29, 2022, provided FDA with legal authorities for the accelerated approval pathway. Under these authorities, the FDA can refuse approval if confirmatory clinical trials are not considered underway prior to approval (The Cancer Letter, March 29, 2024).
“We have seen accelerated approval be the subject of policy debate for many years,” Allen said. “That is not something new. It’s not something that’s likely going to change. There will be mixed opinions as to even the impact that it’s had overall, including in oncology. But there was a substantial effort by Congress a couple of years ago to further strengthen that backside of things, the confirmatory side of things.
“And so, with that in place, and the recognition that in most cases at this point, accelerated approvals are being required to have their confirmatory studies underway sufficiently enrolled as a safeguard to make sure that the evidence necessary to confirm will be generated, I do think it’s maybe a time we’re thinking about now that we strengthen that sort of safety net on the backside of the program. Will that enable additional flexibility to the front end of the program in thinking about the expediency in which a new endpoint might be able to be leveraged?”
