The Myelodysplastic Syndromes Channel is supported with funding from Geron (Silver).
VJHemOnc is an independent medical education platform. Supporters, including channel supporters, have no influence over the production of content. The levels of sponsorship listed are reflective of the amount of funding given to support the channel.
FDA perspectives on drug approval and clinical trial endpoints in MDS
Nina Kim, U.S. Food and Drug Administration (FDA), Silver Spring, MD, discusses regulatory considerations for drug approval in myelodysplastic syndromes (MDS), focusing on clinical trial design and the selection of meaningful endpoints. She highlights the role of overall survival, transfusion independence, and durable treatment responses in supporting regulatory approval, referencing the Phase III IMerge trial (NCT02598661) and emphasizing the need for further validation of emerging response criteria. This presentation took place at the 4th International Workshop on Myelodysplastic Syndromes & Myeloproliferative Neoplasms (iwMDS/iwMPN) 2026, held in Washington, DC.
These works are owned by Magdalen Medical Publishing (MMP) and are protected by copyright laws and treaties around the world. All rights are reserved.
Transcript
Good morning, everyone. thank you to the organizers for the invitation. So my name is Nina Kim, and I’m a clinical reviewer in the FDA’s Division of Hematologic Malignancies One. And I’m here today to share an FDA clinical perspective on trial design and endpoints for MDS. So I have no disclosures, and these views are my own and do not reflect official FDA policy.
So I want to begin with why MDS presents such distinctive challenges for clinical development...
Good morning, everyone. thank you to the organizers for the invitation. So my name is Nina Kim, and I’m a clinical reviewer in the FDA’s Division of Hematologic Malignancies One. And I’m here today to share an FDA clinical perspective on trial design and endpoints for MDS. So I have no disclosures, and these views are my own and do not reflect official FDA policy.
So I want to begin with why MDS presents such distinctive challenges for clinical development. So lower and higher risk MDS are in many respects fundamentally different diseases. So as this group knows well, a patient with lower risk MDS may live for over a decade with anemia as a dominant clinical presentation, while a patient with higher risk disease may progress to AML within months. And these two populations often require different endpoints, different comparators, and different follow-up strategies.
On top of this, evolving classification systems and response definitions complicate population definition across trials and comparison of efficacy to historical controls. And so, broadly speaking, randomized designs are strongly preferred for confirmatory trials and MDS. Single-arm trials are increasingly difficult to interpret because the benchmarks for historical comparisons have shifted over time.
And we’ve seen repeatedly that drugs with high response rates in single-arm trials often fail to demonstrate benefit in large randomized trials. Trials should also enroll populations representative of real-world MDS patients, including older adults, patients with renal or hepatic impairment, and patients from diverse racial and ethnic backgrounds. Blinding is recommended to prevent bias.
And we specifically encourage sponsors to engage with us early, particularly when considering adaptive designs or novel endpoints. which brings us specifically to the topic of endpoints. So the FDA ICH E9 guidance for clinical trials states that there should be sufficient evidence that the primary variable or primary endpoint can provide a valid and reliable measure of some clinically relevant and important treatment benefit.
So I think there are two key components here that are worth emphasizing. So the endpoint must be valid and reliably measured, and it must capture a clinically meaningful treatment benefit. So these are the fundamental criteria to keep in mind when evaluating the adequacy of any clinical trial endpoint, not just in MDS.
So, what does this actually mean? So, I think there are three terms that are worth clarifying because we at the FDA refer to these a lot when discussing endpoints, and I think they often cause a lot of confusion in practice. So the first is direct clinical benefit. A direct clinical benefit endpoint directly measures how a patient feels, functions, or survives. And overall survival is the clearest example of this. A validated surrogate endpoint, on the other hand, predicts direct clinical benefit, and as the term suggests, has been fully clinically validated to be a surrogate for a long-term clinical outcome like overall survival. In this case, the reliability has to be so high that you could essentially fully replace the direct clinical benefit outcome with this endpoint. Now it’s worth noting that true validated surrogates are exceptionally rare in oncology, but one example would be CR rate at 30 months for first-line follicular lymphomatherapies, in which there’s clear trial level data based on multiple randomized trials to support surrogacy. And then there are surrogate or intermediate clinical endpoints reasonably likely to predict clinical benefit, which are endpoints that measure a therapeutic effect earlier than morbidity or mortality. In this case, there could be some data to suggest that the endpoint predicts better survival or function, but the data is perhaps not robust enough to rise to the level of true surrogacy.
So an example of this is CR rate for frontline AML therapies. Now the first two here may be used to support regular approval, whereas the last may be used to support accelerated approval. Now unfortunately in MDS, no endpoints have been formally validated as surrogates for overall survival. that said, we have historically accepted durable CR plus PR rate, transfusion independence, and overall survival for regular approval in MDS on the basis that these endpoints represent direct clinical benefit. However, there are important limitations to each of these, even as direct clinical benefit measures when it comes to MDS. So overall survival is the gold standard, but it takes a long time and a large sample size to demonstrate benefit. For higher risk disease, this is not typically an issue. And an example of the use of overall survival was the azacitidine approval in 2004. However, it’s often not a realistic primary endpoint for lower-risk disease where survival can span a decade or longer.
And so transfusion independence is an important endpoint, particularly for lower risk MDS, and has been used to support regular approvals for lenalidomide, luspatercept, and most recently imetelstat. However, despite transfusion dependence being a poor prognostic marker, it’s not clear that reversing it independently improves overall survival. So, in two recent trials, overall survival was similar between active and placebo arms, despite higher transfusion independence rates with treatment, suggesting that transfusion independence could be prognostic rather than predictive of a treatment’s impact on overall survival. Now, approvals have still been granted on the basis that achieving transfusion independence is inherently clinically meaningful, but this does emphasize the point that overall survival should always be evaluated at minimum as a secondary endpoint to ensure that there is no safety signal for overall survival detriment.
There’s also been some debate about what duration of transfusion independence is clinically meaningful. So we’ve historically accepted eight weeks, but recent trials have demonstrated high placebo response rates at this threshold as shown here. And so in recent years, we have been recommending that sponsors assess at least 16 weeks of transfusion independence to demonstrate clinical meaningfulness as the placebo effect diminishes with longer durations, as seen in the IMerge trial on the right.
Finally, CR plus PR has historically supported approvals for HMAs in higher risk MDS since both by definition require restoration of trilineage hematopoiesis, which has inherent clinical benefit. However, the relevance of this endpoint for lower risk MDS is perhaps less clear, and several trials have shown that a promising response rate in early studies does not necessarily translate to an overall survival benefit in phase three. And in some cases, there’s actually been a trend towards survival detriment. With that said, we recognize that the IWG 2023 criteria represent a sincere effort to improve upon the prior framework by eliminating responses like marrow CR and stable disease, for which the evidence did not support clinical meaningfulness as shown here, where marrow CR was associated with an overall survival similar to or worse than stable disease, and clearly worse than a true CR. And so we have been supportive of sponsors incorporating IWG 2023 criteria prospectively in clinical trials.
That said, I would like to emphasize that some of the endpoints proposed in IWG 2023 from a regulatory standpoint remain provisional, as acknowledged in the paper that proposed them. And specifically the less than CR endpoints such as CRL, CRH, and ORR, which includes these types of less than CR responses. And since these endpoints do not require full restoration of trilineage hematopoiesis, I think their clinical meaningfulness is more ambiguous perhaps than a true CR or PR. And so further prospective validation is needed.
Now, what about hematologic improvement? So this is a question that we encounter a lot from sponsors, and we acknowledge that some HI responses, specifically HIE in patients with low baseline transfusion burden and major HIE response in those with high baseline transfusion burden, actually do require transfusion independence by definition, which is clinically meaningful. However, transfusion reduction or isolated improvements in hemoglobin platelets or neutrophils alone, which are also considered HI responses by this definition, are not necessarily considered direct clinical benefit. And so further validation is needed to establish the meaningfulness of these specific thresholds and their predictive value for direct clinical benefit.
That said, we recognize that there have been efforts to validate the newer response criteria, and these are important first steps. for regulatory purposes, prospective validation using raw data from multiple positive and negative randomized trials remains necessary. And my statistical colleague, Dr. Avadani, will go into more detail about the kind of various levels of evidence needed in the presentation to follow.
So, in summary, durable CR plus PR, transfusion independence, and overall survival remain the endpoints that primarily support marketing applications in MDS at this time. We recommend assessing at least 16-week transfusion independence to demonstrate clinical meaningfulness with concurrent evaluation of overall survival for safety. Less than CR endpoints and hematologic improvement require further validation, and we encourage prospective incorporation of the IWG criteria in clinical trials to build that evidence base.
Finally, I just wanted to note that we published a draft guidance on MDS drug development last year describing our current thinking on trial design and endpoints for disease modifying therapies in MDS. A final version incorporating community feedback will be issued in the near future. And so I encourage everybody to take a look at that. Thank you for your attention.
Read more...
