MDS updates from iwMDS-iwMPN 2024 | Response assessment & regulatory considerations in MDS

Rami Komorokji:

Hi, I am Rami Komorokji. I’m vice chair of Malignant Hematology at Moffitt Cancer Center in Tampa, Florida. We just had a very interesting session about regulatory endpoints and response assessment in MDS. I had the pleasure of learning from my colleagues on the stage, and I’ll have my colleagues introduce themselves starting with Katharina.

Katharina Gotze:

Hi, Katharina Gotze from the Technical University of Munich in Germany.

Andrew Brunner:

Hi, I am Andrew Brunner from Massachusetts General Hospital in Boston, Massachusetts.

Lisa Pleyer:

And I’m Lisa Pleyer from the Paracelsus Medical University Hospital in Salzburg, Austria.

Rami Komorokji:

It’s always a pleasure to have you on board. So maybe we’ll start just a little bit, a brief summary of what everybody presented. I thought the session went very well and really highlighting many of the challenges we have. So again, I’ll probably start with Katharina.

Katharina Gotze:

Okay. So I was asked to speak about how to approach MDS-IB2, so with above 10% blasts or MDS/AML in clinical trials and for regulatory agencies. And I think we’re faced with the conundrum that we have these two classifications, WHO and ICC, which maybe are two sides of the same coin, and we don’t have a whole lot of options for these patients at the moment in the MDS world. So only azacitidine, or allotransplant, or in the US also decitabine for these patients. And the MDS/AML category does open up some AML treatment regimens for these patients, which could be beneficial for the patients. Even though I do think this new category is also a political issue a little bit, but at least for targeted agents, I think it does allow us to put patients on these trials, which could benefit patients that we think are good enough for intensive treatment.

Rami Komorokji:

Absolutely, and I totally agree. I think we were just also talking about the line between MDS, AML, and I think in my opinions, as I said, I think most of the AML we encounter in older patients is really coming from MDS, whether it’s diagnosed or not, it’s hard to tell. They don’t have to have the antecedent history, but if you look into the ontogeny of the disease, I think most of the older AML patients are. And I think to your point, those are sometimes arbitrary cutoffs so that in trials we started restricting ourselves even too much with those arbitrary cutoffs. So-

Katharina Gotze:

I think things might change if we get the ability to have more targeted agents, like we’re having menin inhibitors. So if we see KMT2A rearrangements in a secondary AML or a patient with 10% blasts, then we might go for that targeted treatment when it’s available. So that might change the landscape a bit, but it’s a big issue right now and I think hopefully it will go away if we’re able to harmonize the two classifications.

Rami Komorokji:

Absolutely. So Andrew, can you tell us a little bit about your?

Andrew Brunner:

Yeah, absolutely. Yeah, so I was here to discuss a little bit about the need for standardization about how we report studies. And I think that it is great how, as a field, we have really understood a lot more about the biology of disease. Our understanding of heterogeneity of the disease and what to expect from a given treatment has really advanced, but it makes comparing one study to another and trying to understand what to expect out of a given therapeutic, really challenging. And we don’t have a great way to interpret how one result might compare to another. And even when we get results from a final manuscript, it often is hard to know, well, is this giving us a full picture of how this relates to a given population? I think that there’s a few areas where we really need to be cautious about as a group.

One, is about what to expect from a patient population, and that has to do with the heterogeneity in low-risk patients around their cytopenias. So who actually needs a transfusion? Who has other cytopenias that might need intervention? How bad are those at baseline? And do we need to adjust the expected response based on the baseline attributes of a given patient? And in high-risk disease, we know patients who have bad actors, bad mutations, bad disease features, but I don’t know that it always reflects how we choose therapies. And that gets a little bit into this idea of MDS versus AML. If you have something which is sensitive to a cytotoxic agent or treatment intensification that might be falling more into this AML-type therapy. If you have something where really you need to reprogram the bone marrow, it doesn’t work, it’s molecularly damaged, that probably requires a different approach. And so I think we just need to be cognizant of who we enroll on trials and what to expect from a given patient.

A point that I think we may not have realized, although it makes sense is that if you have really bad disease, you have p53 mutations, complex cytogenetics, you have a really messy disease but don’t have increased blasts, I can put you on a study. There’s no way you can ever achieve a CR by our definitions. And so how do we identify clinical benefit for those kinds of patients, and how do we move the needle when a lot of our disease is going to be molecularly defined? And so I think that just being aware of it as we develop new therapies, looking at novel targets, looking at groups, defined molecularly, is the right direction to go, but we have to have the right expectations about what to get out the given therapy.

Rami Komorokji:

And to your point, and I think you’ve published on this before, restoring effective hematopoiesis, how do we measure that? What is really a robust, effective hematopoiesis for those patients? I really was interested in the slides you showed is the number of CRs you need to see in an earlier phase to translate that, because I think what we’ve seen in the most recent studies, one after one, the Phase II looks very promising, CR rates. We take to Phase III, it drops down. So I think that was enlightening for me, that you really have to have very high CR rates in a Phase I, Phase II before you say I’m going to commit to that. Or even… I think that’s now the direction if we’re doing randomized clinical trials for CR in a Phase II randomized, is what’s the magnitude or the delta in the CR that will translate to survival advantage? Because I was looking at the slides you showed and it’s like even a 10% CR difference didn’t make it. So-

Andrew Brunner:

Yeah, no, the bar is pretty high. Until you have a sizable number of patients, the certainty around who’s benefiting can be really challenging. That’s not to say that CR is the be-all, end-all. CR is a high bar and most patients will not achieve a CR. But you’re absolutely right, if we’re using an early phase study, the number of people to know that you’re actually really improving upon a historic, the majority of patients really have to be hitting a true CR. I think we as a group still need to identify less than CR responses and other metrics that are meaningful, but we also have to validate those in azacitidine alone or with any given therapy. And so many people are still benefiting, and to your point about what blood count matters, any blood count that keeps you away from my clinic is probably good. And so there’s a lot of ways to put that together, but as a field, we do probably need to validate benchmarks so that when we are trying to be a historical control, like azacitidine, we’re not fooling ourselves early into thinking that something is a slam-dunk.

Rami Komorokji:

All right. No, absolutely agree. And finally, I think Lisa, your talk was the most exciting. Dr. Bennett, the father of hemepath and bone marrows was in the audience. I expected him to stand up and say, “No, you need the bone marrow on everybody.” And he did not, which means you succeeded in convincing the audience a little bit. So can you tell us a little bit?

Lisa Pleyer:

Yeah. Well, I was asked to talk on the very controversial topic of whether we still need bone marrow in patients with MDS and/or AML for diagnosing, the risk stratifying and also monitoring treatment response. So to wrap it up, I basically started off with showing what happens in the real world. And so bone marrow evaluations are only performed in 50% of patients to diagnose in the US, and also in only 50% of patients who receive non-intensive therapies. And therefore it’s really important to be able to use other parameters other than bone marrow blasts for all of these things. And I just basically threw out bullet points from all of the different topics in the areas that, in my opinion, are going to be important and should gain more importance. And that includes NGS analysis, flow cytometry, and perhaps also other CBC parameters. And yes, so there are a lot of groups working on these things and I think that the future will be an exciting in this regard.

Katharina Gotze:

I thought it was super interesting that you showed with the red cell distribution width, because that’s one of the biggest parameters in the clonal hematopoiesis risk score as well. And I’ve actually started documenting this on all of my MDS patients because if you do start looking at it, you do see that, oh, the higher risk ones are the ones where you’re not sure if it’s going to be MDS and the ones that have higher-

Lisa Pleyer:

There’s so much data that we’re assessing but we’re not using-

Rami Komorokji:

Right.

Lisa Pleyer:

… and implementing yet. And I think AI is going to help us there decide what is most relevant for our patients.

Rami Komorokji:

And I think one of the most practical things is, similar work you did and presented last year, is that the peripheral blood is reasonable or the concordance is very high with the bone marrow in terms of assessing mutations, because that’s a very common scenario we see. We get asked, patients already got their bone marrow, but in community practices in the USA, they don’t do molecular profiling all the time at the diagnosis. And then your work shows that it’s okay to get those on the peripheral blood and the concordance is very high. I think that has a very high practicality in terms of patient management and even if we are getting to a point where we will be assessing responses molecularly, that you could probably do serial peripheral blood, you don’t have to repeat a bone marrow every time. So that’s really very practical and one of the largest data sets to address that question. So thank you for that.

Lisa Pleyer:

It’s also going to be really interesting in the future to see for longitudinal assessments. What happens with the VAF, which mutations disappear under treatment, which potentially appear in perhaps still responding patients before they grow out or before they lose response? So I think there is still so much that we can, and need, to and have to learn from. And, yeah, it will be interesting.

Rami Komorokji:

The other thing I was thinking of is maybe the next step, futuristic will be digital imaging of peripheral blood and does that, if you start having that correlate with better than just the eye of the hematopathologist, that will see a morphology that we don’t see now in the peripheral blood that will replace the bone marrow. So…

Lisa Pleyer:

I think that’s a very interesting concept, but it’s the same with digital imaging of FISH analyses. So we’re currently doing that, but we’re still manually checking it. The machine, it aids you in doing the analysis, but it still doesn’t replace the humans. I think we’ll probably be able to get there, but I don’t think we’re quite there yet.

Rami Komorokji:

Absolutely, absolutely. Yeah, and on the final part, on my part, obviously we had some heated discussion at the end about… So I was trying to address the lower-risk MDS and what’s meaningful or what are the endpoints that we use to measure modifying the disease. And my point was, modifying the disease does not have to be always overall survival, because there are so many other things in medicine, diseases that our endpoint is not improving over the overall survival all the time. If the disease burden is symptomatic, that affects patient’s daily activities and quality of life, it’s meaningful to modify that or improve that. So I think the discussion was obviously we have the traditional endpoint, blood cell transfusion independency, improving quality of life that we struggle in every trial to objectively demonstrate that. And maybe creative endpoints, like time to transfusion independency, transfusion free survival. And then we saw a bit debate about patients that are symptomatic anemia, do you pull the trigger on treatment before they become transfusion independent? Where is the trigger to start the treatment? So I’m interested in anybody’s thoughts on that.

Katharina Gotze:

I think one of the problems with the quality of life is the time point where we’re assessing it, because the patients come into their blood counts or their therapy and that’s when they’re getting the paperwork to see what their quality of life is, and if they just had a transfusion before or not, that will impact how the quality of life is assessed. So I think that’s a really difficult issue and we’ve seen that in the MEDALIST trial where it didn’t really pan out even though as physicians treating the patients, we really felt like we were making a difference in the quality of life. So that was a disconnect for sure.

Rami Komorokji:

Absolutely.

Andrew Brunner:

Yeah, it has been good to see so much interest in trying to improve upon how we assess the patient-specific outcomes, patient-reported outcomes. It is challenging too because I think people, humans are resilient, right? You get used to your hemoglobin. I see somebody and I have to go back five, 10 years to understand how much have you actually reduced in your life because of your anemia. And people adjust, and so it can be really hard when you’re trying to change something that’s really a global problem and everybody has a different threshold to some degree of where intervention is needed. And I think that we heard that in the discussion a little bit, is this idea of, all right, when are things bad enough that we really need to change course? And that’s going to change for every person.

I think that there are some novel technologies I think that are really interesting. I would love to have ring data looking at VO2 max, and how much are you up and about, and number of steps per day, and having areas under the curve before and after an intervention. And can we be smarter about using some of the technologies that we have to try to identify those? Because it is tough when you take a survey, catching somebody, how your day’s going influences a survey. You need a lot of surveys to wipe out that noise. Everybody started from a different baseline. And so I think I’m optimistic that some of the advances and the increased interest in this space can make it a little bit more granular, a little bit more refined, and then maybe we can start to see some of these differences where it’s more obvious, oh, you do need to really start, look at what you’ve stopped doing in your life.

Rami Komorokji:

Right, and it’s very challenging from regulatory point to get approvals based on that. I think that the general principle is you either demonstrate that you live longer or live better. That was what we were hearing from the FDA yesterday, and sometimes either of them is hard to prove via trials. So absolutely, so…

Lisa Pleyer:

Especially regarding the lower-risk MDS, what you were saying, because so much is going to happen afterwards.

Rami Komorokji:

Right.

Lisa Pleyer:

So overall survival I think is going to be very… different by what happens. Do they have access to trial, do they have access to this drug, to that drug, do they respond to the following treatment? So I think time to next treatment, again would be a good endpoint.

Rami Komorokji:

Absolutely.

Katharina Gotze:

Kind of like with the breast cancer or the myeloma patients.

Rami Komorokji:

Right. I think that would be a meaningful, I actually didn’t think of putting it up there, but time to next treatment will be very meaningful. Andrew and I talk all the time about even this question of sequence of therapies, you have three options now with X response, duration. How do you sequence them? How do you select the patients that have the better chance of response to A, rather than B? What is the chance of B working after A? It really gets complicated a little bit. It’s not an easy decision. So thank you again very much. It was a pleasure to be with you on this session and look forward for future meetings. So thank you.

All: thank you.