With work ongoing in the oncology community to develop the evidence needed to support circulating tumor DNA as a regulatory efficacy endpoint, one US Food and Drug Administration official is making a plea for standardization.
Developing the evidence needed to support use of a novel endpoint is a “heavy lift” requiring many stakeholders and pooling data across numerous trials for meta-analyses purposes, the Oncology Center of Excellence’s Nicole Gormley said at a Friends of Cancer Research meeting.
“Standardization is absolutely key to try to make sure we collect the same information, and then that … improves the strength of our analyses later,” said Gormley, OCE’s associate director for oncology endpoint development and director of the Division of Hematologic Malignancies II. “If we could sort of have some degree of standardization in terms of the times that we’re assessing this, the thresholds, etc., it really could help then to advance the field altogether.”
FOCR is leading a multi-stakeholder effort to develop the evidence needed to use ctDNA as an early endpoint for regulatory decision-making. (Also see “Cancer Drug Endpoints: ctDNA Promising But Presents Many Challenges, Experts Say” – Pink Sheet, 1 Dec, 2021.)
The group’s 11 July meeting included a presentation of initial data from an evaluation of baseline ctDNA measurements across cancer types, stages and assays, as well as updates on the ctMoniTR project, which is using retrospective trial data to assess whether changes in ctDNA reflect response to treatment.
Wanted: More Data
ctDNA is tumor-derived fragmented DNA shed into a patient’s bloodstream that is not associated with cells. The quantity of ctDNA can vary by individual and depends on various factors, including the tumor type, location, stage, tumor burden and response to therapy.
In a May 2022 draft guidance, the FDA said ctDNA can be used in early-stage solid tumor clinical trials for patient selection, patient enrichment and response measurement. However, further data are needed to support the biomarker’s use as an endpoint reasonably likely to predict long-term outcome, the agency said. (Also see “ctDNA Can Be Used For Patient Selection and Enrichment, But Not As Early Efficacy Endpoint, US FDA Says” – Pink Sheet, 4 May, 2022.)
The guidance discussed how meta-analyses of data from randomized trials potentially could support use of ctDNA as an early endpoint for drug approval in an early-stage cancer setting.
FDA officials at the FOCR meeting said that while there is growing evidence of ctDNA’s potential as a prognostic biomarker, it needs to be prospectively incorporated more widely into clinical trials.
“We are at a stage where we need to collect more information. And how do we collect the information? Well, we collect it in the clinical trials and during development,” said Center for Drug Evaluation and Research Director Patrizia Cavazzoni.
“It’s very important that novel endpoints start being collected in clinical trials that utilize accepted endpoints so that then we can collect the data to see whether there are associations between accepted clinical endpoints such as PFS [and] overall survival when it comes to cancer,” she said.
“We need to have sufficient data. We need to do meta-analyses of the data from various clinical trials,” Cavazzoni said. “And it’s very important that developers come and talk to [OCE Director] Rick Pazdur and his team because it’s very important to have the right guideposts when these endpoints are included in clinical trials.”
OCE’s Gormley said using novel endpoints in clinical trials has to be a data-driven decision.
“We need to know if we’re going to incorporate this into a clinical trial what percent change is clinically meaningful? What difference between treatment arms would predict that this therapy is an effective therapy,” Gormley said.
“If there’s so much variability in the times when it was assessed or so much variability in the assays, it sort of limits your ability to draw meaningful conclusions from that data.” – FDA’s Nicole Gormley
“We need to know other aspects in terms of how to best define this,” she added. “What are the appropriate thresholds? In all of those decisions, when we’re thinking about how to use this as an endpoint for a subsequent trial, we need to be data-driven and data-informed based on our prior knowledge.”
“Those are some of the outstanding questions that … exist right now in terms of how to get ctDNA to a point where we can use it” as an endpoint, Gormley said.
Developing standardized approaches to measuring ctDNA in prospectively designed studies will be critical to addressing some of these questions, she said.
“When you think about trying to go back and pool this information or data from a retrospective meta-analysis or in some other manner, if there’s so much variability in the times when it was assessed or so much variability in the assays, it sort of limits your ability to draw meaningful conclusions from that data,” she said. “To the extent that there can be more standardization across the field, that really does, I think, have a potential to sort of [raise] all boats.”
Vladimir Jankovic, senior director of precision medicine at Regeneron Pharmaceuticals, Inc., suggested companies consider a step-wise approach to incorporating ctDNA assessments into studies, starting by using the existing data to prospectively define endpoints and include them as exploratory analyses. Including ctDNA endpoints as part of the statistical hierarchy with hypotheses testing may not be possible at this time because the effect size is not yet known, he said.
Data from these studies, which are specific to the assay, time of measurement and clinical setting, could then inform the statistical powering of future studies with ctDNA as an endpoint, Jankovic said.
“Realistically, this is where we need to start,” he said. “It doesn’t mean that we’re not going to do this in parallel, as well. Maybe there are some settings where we can be informed already right now. But I think those are at least two different levels of evidence.”
“Even if there’s not comfort in having a secondary endpoint with appropriate alpha allocation, still making sure that that’s assessed in trials across the board, such that then you actually have more data that you can evaluate and answer questions for how to subsequently use this,” is important, Gormley said.
Long-Term Follow-Up Remains Essential
Gormley was asked about lessons learned from the development of evidence to support other novel endpoints in specific cancer settings. She said identifying what threshold of change is actually clinically meaningful is important, suggesting the issue again hinges on standardization.
The FDA’s ctDNA draft guidance states that assessments can vary among laboratories and technologies used to detect the biomarker, which can result in discrepant results.
“Further standardization of assays will allow for better use of ctDNA in a regulatory setting and will allow for analyses across studies to validate the use of ctDNA,” the guidance states.
A determination of “key thresholds for the analytical performance that if you at least attain this threshold, that would correspond or correlate with long-term clinical benefit,” is necessary, Gormley said, noting the agency is taking a similar approach with the recently announced oncology drugs and in vitro diagnostics pilot project. (Also see “US FDA’s Cancer Drug/Diagnostic Pilot Program Aims To Close The Gap Between Regulation And Clinical Practice” – Pink Sheet, 13 Jul, 2023.)
However, Gormley urged caution because deeper responses are not always better. And if the response comes “at the expense of safety, you may still lose that correlation with long-term outcomes.”
“In many of the clinical drug trial designs that we talk about where we use these endpoints, we still continue to follow those patients for later long-term outcome,” she said.
“We use early endpoints, and we think that they’re really, really helpful to expedite drug development,” Gormley added. “But we still continue to follow those patients for the long-term outcomes to make sure that the clinical benefit that we expect from those early endpoints is actually there.”
‘Socialization’ Of Novel Endpoints
Cavazzoni also spoke about the need to build community support for novel endpoints that have been deemed sufficient for regulatory decision-making.
“It’s also very important that there be some socialization of a novel endpoint even after there is sufficient data and sufficient analysis have taken place, so that there is a … common understanding across the treatment and also the payer community in … acceptance of that endpoint,” she said.
“It’s also very important that there be some socialization of a novel endpoint even after there is sufficient data and sufficient analysis have taken place, so that there is a … common understanding across the treatment and also the payer community in … acceptance of that endpoint.” – FDA’s Patrizia Cavazzoni
Cavazzoni’s comments may reflect lessons learned from the FDA’s June 2021 accelerated approval of Biogen, Inc. and Eisai Co., Ltd.’s Alzheimer’s disease drug Aduhelm (aducanumab-avwa) on the basis of amyloid plaque reduction as a surrogate endpoint reasonably likely to predict clinical benefit.
The Aduhelm approval took many by surprise because then-Office of Neuroscience Director Billy Dunn expressly told a November 2020 advisory committee meeting that the agency was not viewing amyloid as a surrogate for aducanumab’s efficacy.
The Centers for Medicare and Medicaid Services limited reimbursement for aducanumab, and other anti-amyloid antibodies granted accelerated approval, only to use in clinical trials.
However, with the subsequent accelerated approval, and recent regular approval, of Eisai and Biogen’s Leqembi (lecanemab-irmb), amyloid plaque reduction seems to have become more accepted as a regulatory endpoint, although the FDA is not yet ready to declare it as a validated surrogate. (Also see “After Leqembi Approval, US FDA In No Rush To Declare Amyloid A Validated Surrogate Endpoint” – Pink Sheet, 8 Jul, 2023.)