iwMDS 2025 | Evolving strategies in clinical trial design for MDS: defining response criteria

Maximilian Stahl

All right, we’re here in sunny Lisbon, Portugal at the iwMDS meeting for 2025. My name is Max Stahl, I’m from Yale University and I’m here with two colleagues and friends.

Rami Komrokji

I’m Rami Komrokji, from Moffitt Cancer Center in Tampa, Florida.

Coleman Lindsley

And I’m Coleman Lindsley from Dana-Farber Cancer Institute in Boston, Massachusetts.

Maximilian Stahl

Wonderful. So I think today, Rami and Coleman, we discussed clinical trial design a lot in the morning, and we had a particular focus on how should we measure response or benefit of clinical trials. And Rami talked about lower-risk trials, and Coleman and I talked about higher-risk trials. So maybe I’ll just start with Rami. A couple of reflections of what you see, how we currently assess response in lower-risk MDS trials, and maybe what we should do in the future. What should we do different? What could we improve?

Rami Komrokji

Absolutely. So, yeah, as we discussed earlier today, obviously, the gold standard in lower risk MDS had been rendering the patient’s transfusion independent. So the rate of transfusion independence had been the gold standard for drug approval. And I think we should be moving beyond that. Like we are starting already in clinical trial looking at time to transfusion dependency as an endpoint. but I think as we understand more the natural history of the lower risk MDS and the causes of morbidity and mortality in those patients we should be looking at other things. We showed some data on like the rate of cardiovascular events for example in those patients and that if we look at the mortality and lower risk it’s almost one-third are gonna go to higher risk one-third cardiac disease and the rest is variety so I think incorporation of like rate of progression to higher risk MDS, AML, incorporation of cardiovascular events, if we’re going to start looking at like an event-free survival endpoints, it’s going to be very relevant. And I think the interesting part for this, it’s not just really now improving the cytopenia. I think we are realizing that in MDS, it’s probably more than that, that the inflammatory milieu, the underlying CHIP mutations maybe are contributing beyond just the simple cytopenia and its related complications.

Maximilian Stahl

Yeah, I think that’s wonderful. I think it’s also incredibly complex. You know, it’s certainly easier to look at transfusion independence. How would you envision, I guess, incorporating it in some of the trials? You mentioned some anti-inflammatory therapies. What are kind of the key things you would look at in order to get to that important endpoints that I think you mentioned, how would you measure that?

Rami Komrokji

Yeah, I think obviously first, historically, we’ve done a lot of the trials with anti-inflammatory drugs so late in the game, where like some of the original trials we were conducting, we were requiring patients to have HMA failure, which I think at that point, really, the disease is so clonal and then evolved to a point where it’s, you know, I think we should move those agents earlier. And the endpoints are going to be, still some of them are going to be the traditional improvement in cytopenias. But I think, you know, as we heard today, a lot of, you know, such a common theme about looking at cardiovascular events, we try to look at things like rate of those events, hospitalization for those events. You know, I don’t know if we have a very good biomarker for those diseases to incorporate in trials, but it’s really like, you know, it may not be the classical, you know, standard improvement in cytopenia. And maybe even the selection for those patients should not be requiring patients that have profound cytopenia. Maybe taking patients that have mild cytopenia and looking at different aspects of the disease.

Maximilian Stahl

Right, yeah. And I think we’re also lucky to have a very enthusiastic cardiologist with us today, which I think shows that we can benefit from learning from what they look at. And so let me switch gears and ask Coleman here. We’re very glad that Coleman could join us. I think he’s always very thought-provoking, but also thoughtful about what he tries to provoke. And he talked a little bit about molecular clearance, how should we incorporate genetic testing into clinical trials. Maybe Coleman, you can start with just summarizing a couple of the key takeaways that you would want people from our field, the clinicians, but also from the company perspective or from the regulators to take away how should we do trials to be able to look at all those things?

Coleman Lindsley

I think it’s a very challenging question that reflects where we are as a field right now. So the last 10 or 15 years have been transformative for our understanding of how disease genetics interacts with disease biology. And I think we’re at an inflection point, to be honest, in terms of sorting out where and how we apply that knowledge to clinical trial interpretation, clinical trial design, and the regulatory evaluation process. And so I think in terms of the takeaways from our discussion this morning, I think it’s time to look back at a series of large Phase III trials that have recently been completed in MDS and dig deep into their molecular characteristics, both at baseline, what the patient starts with, and then how those genetics change over time in order to better understand which patients are likely to respond, are responding, or likely to progress or are progressing. And so what we’re proposing is a strategy that balances more qualitative assessment of the disease. And by that, I mean key defining genetic changes that dictate the biology and how the phenotype or the clinical manifestations of the disease. That qualitative part with more quantitative parts, the abundance of a particular clone or subclone, and how tracking that over time may reflect the disease status. Some people talk about, and I think we all collectively think about MRD or measurable residual disease as something that we could track for this. I think it’s becoming apparent that we don’t have a shared definition of MRD. And that really poses a challenge. And maybe it’s time to take a step back, reevaluate that term in this more composite or hybrid way of the quantitative and qualitative. In terms of your last part of your question, the collaborative approach that we need at this point. So I think we’re out of the raw discovery phase and into the translation phase for clinical genetics. And that’s going to require academic partnership with industry and regulatory agencies to come up with consensus strategies for how to do this together.

Maximilian Stahl

Yeah, that’s wonderful. I think you have been the kind of master of learning from trials that have, so quotation mark failed and kind of really getting as much data from them, mainly in AML. It’s glad that you’re going full steam into MDS. But maybe even before a trial fails and we kind of do a dissection on autopsy of that trial, what would you wish would be kind of like incorporated right from the beginning? And this is, of course, directed at the people who design the trials, but also the companies that then conduct those trials. What would you think, from your perspective, going backwards in a prospective way, what would be most helpful, I think, for the field to learn those type of things?

Coleman Lindsley

Yeah, I think we’ve learned without a doubt that MDS and AML are heterogeneous diseases. And that heterogeneity impacts response and resistance to treatment. I think we can say that’s true. Yeah. Right. And some drugs in these negative trials or positive trials, we’ve been able to look retrospectively and say this group of patients has zero response and this group of patients has the bulk of the response for this drug. And I think it’s important in order to best capture a positive signal in a trial and a negative signal is to incorporate this analysis earlier on. And this may in earlier phases of the investigation of a drug in the Phase I and Phase II setting may help a company target that drug to the right populations. It may help them broaden it if it’s agnostic to the genetics.

Maximilian Stahl

If appropriate.

Coleman Lindsley

If appropriate, right. And then the design of Phase III trials may be such that we as a community can collectively target that therapy to the right group of patients and know how it fits in the landscape of treatment. So I very much favor the incorporation of correlative genetics into all of these clinical trials. And even not to restrict inclusion, but rather to understand the outcomes. And I think you acknowledge an important state in our field right now, which is a number of negative trials when we define it at the whole population level. But that doesn’t necessarily tell us that it’s negative in every population.

Maximilian Stahl

Right, right. And maybe just one last question, maybe directed at both of you really. It’s like, what I struggle a lot with is that we look at certain predictors, genetic or otherwise, but it’s complex because, you know, often we look at it for IPSSM, we looked at it in untreated patients, right? But then there’s a treatment context, different types of therapy, intensive chemotherapy, less intensive chemotherapy, target therapy, immunotherapy, and transplant. You know, how do you guys both as kind of a translational scientist, but also as a clinical scientist and clinical trialist to kind of untangle this complexity of context that you review certain responses or even molecular patterns? Maybe I start with Rami.

Rami Komrokji

Yeah, no, I think it’s very well taken point. And there is no doubt, you know, all of those, even prognostic models, it’s really in the context of therapy. And there are several examples from other diseases where things that used to be considered a high-risk feature are not anymore. And the closest example to us now is really the AML risk criteria that now we know what applies to intensive chemotherapy does not apply to the HMA anti-BCL2, for example. And I think as things evolve in MDS, we are probably going to go that way. And I totally agree. I think now it’s like a very heterogeneous group of patients we are collecting together that eventually when we get active therapies like for example DEL5Q we know that now that there are actually different predictive factors of response prognostic factors than the traditional like even using the IPSSM or IPSSR would not apply to that. I have the feeling it’s the same in the MDSSF3B1. In general, those patients do very well. But if you start really digging there, you’ll find a small group that their response to treatment and outcomes are going to be different. And that speaks to the heterogeneity and what, you know, we were talking about in terms of the genotype-phenotype association. And I think simplistically, we just lumped so many diseases under one, and we will just keep moving in that direction. But I think the point is very well taken so.

Maximilian Stahl

That’s your last word for Coleman yeah.

Coleman Lindsley

Sure. I think what you’re highlighting is the importance of context for interpreting disease biology. And so I think it compels us to look under every treatment paradigm. We need to evaluate what a specific biological group means and so what might be meaningful in intensive chemo versus venetoclax based versus immunotherapy might be different but our ability to answer that collectively as a community depends on collective buy-in to that idea that we’re going to analyze everyone comprehensively and uniformly so that we can develop these large enough data sets to understand these small subgroups which likely are quite meaningful. But it really takes numbers and collaboration and shared commitment to that.

Maximilian Stahl

Perfect. Wonderful.

Rami Komrokji

I have a question to pose for you. So you addressed the importance of complete response before transplant and should we aim for that. So maybe you can tell us a bit about that. But my question to you, even beyond transplant, Like, since we have all those trials now and we invested in a Phase III trials, should CR or what would be the primary endpoint for at least the early readout studies to say, this is promising enough that we should go into a Phase III?

Maximilian Stahl

Yeah, yeah, that’s a great question. I mean, I think to the first part of your question, I guess the purpose of my study was to show that we should be careful in demanding certain responses without clearly showing that it actually benefits a patient. In particular, that we don’t take response as necessarily a reflection of what we do with treatment, but that we account for what the baseline biology it is. And we always know this, that patients who have a better baseline biology respond better. And if there’s no added benefit of responding that is not already accounted for by baseline biology, I think it’s very dangerous to keep treating a patient in order to get that response while maybe delaying an important treatment with transplant. That was, I think, was the first part. The second part I think that you brought up is kind of a million dollar question, what should be our endpoint, you know. And I think it’s a little bit treatment context dependent, but in terms of complete remission, I think it has been shown that complete remission is a helpful marker, particularly in patients who are not transplanted. You know, if you have complete recovery of your blood counts, you do better and you don’t require transfusions and your risk of progression to AML is less. But I think what is often forgotten is the time aspect. So I think a durable CR is meaningful for patients. A CR that lasts for a cycle is not meaningful. But currently, we kind of treat those the same. So I think that would be one of the many things you could say about that.

Rami Komrokji

Absolutely. Absolutely.

Maximilian Stahl

All right. Wonderful. This was fantastic. And I really enjoyed sharing the session with both of you.

Rami Komrokji

Always. Thank you.

Coleman Lindsley

Thank you.

This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.