MDS updates from iwMDS-iwMPN 2024 | Updates in LR-MDS treatment: imetelstat and luspatercept, standardizing PROs, and using AI to improve risk stratification

Hetty Carraway:

Welcome to today’s session and recap on the International Working Group for MDS. What a fantastic day we have had and just exciting data in both lower-risk and high-risk MDS. And I’m here with several of the excellent presenters from today. And I want to begin with a question around some of the new and novel therapies in lower-risk MDS. So I’ll begin with you Dr. Platzbecker. Can you update and share with us some of the key take home points from your presentation today on telomerase inhibition?

Uwe Platzbecker:

I think the major take home message, I hope at least, for today, was that the community, MDS community, is pretty happy about the recent approval of imetelstat in lower-risk MDS patients, which was actually based on the large Phase III trial and I tried to summarize the background for the Phase III study and also the outcome and the recent discussion on the potential benefit of telomerase inhibition in lower-risk MDS patients. As you know, the label is restricted to a specific population of lower-risk MDS patients, namely EPO/ESA refractory patients with a certain transfusion threshold, which is four units of red blood cells every eight weeks. And this was the inclusion criteria within the pivotal IMerge trial, which met the primary endpoint with a significantly higher rate of transfusion independence compared to the placebo arm.

And I think the key messages of the clinical trial were, first, that in the majority of the responders after eight weeks, also this translated into a long-term response, so more than one year, the rate was actually 18%. Number two, there was disease-modifying activity as expressed by a decline of the allelic burden of the mutations, also of the ring sideroblasts, also some side genetic responses. So although this did not translate into an overall survival benefit with the analysis we have so far, it looks like the drug, in contrast to luspatercept, may alter at least the biology of the disease, with outcome I think needs to be read out with further follow-up. And last but not least, there is some considerable toxicity with the agent, which I’m not actually, I’m not afraid of because we give R-CHOP and all these things all the time to lymphoma and other patients. So I think we can do a very good surveillance on neutropenia, thrombocytopenia, but I think for someone who’s inexperienced with the drug, this needs to be educated before and it can be easily managed sometimes with dose reductions or dose delay.

So I think that’s the take home message. We need to know a little bit more in future studies also that the disease-modifying activity needs to be explored a little bit better and I think it’s time potentially for studies with combining this agent with other agents in lower-risk MDS. So I think good news, 2024 two approvals, luspatercept as a first-line therapy, imetelstat as a maybe all comers second-line therapy.

Hetty Carraway:

Just really exciting data and so wonderful to have such a quick recap. I mean it’s a lot to cover in a short period of time and really delighted to hear that feedback from you just with regards to the success story from this agent and also the success for our patients that are transfusion dependent. And again, just highlighting how important it is and what the true benefit is to patients in terms of transfusion independence and the large increase that patients are seeing when they’re responding to this drug. And I think if there are concerns around the drug with regard to toxicity, are there any special things that people should do or pay attention to if they’re using this drug for patients?

Uwe Platzbecker:

I think the prerequisite of the trial and also the inclusion criteria, I also highlighted this today, is that apart from a severe transfusion dependency, those patients had actually normal blood counts. So normal neutrophils and platelet counts. So I think don’t treat a pancytopenic high transfusion burden, lower-risk MDS patient with imetelstat. I would definitely be cautious at this stage. So I think these patients should be highly selected. I’m not saying that those patients are not around. I think we have a lot of those patients eligible for imetelstat, but especially during the first cycle, the neutrophil count can go down significantly. So sometimes GCSF is needed, sometimes platelet transfusions, but lower-risk MDS is not metastatic breast cancer. So it’s more of a long run disease. So be patient, dose reduce, dose delay, give GCSF, talk to the patient, be calm. You know, that’s, I think, the very important message and don’t give up if the drug actually causes a significant hematotoxicity during the first cycle.

Hetty Carraway:

Right. Just because some of those toxicities, once patients get through them, they then can lead to benefit. So really important messages there and exciting to think about the future with regard to combinations with other therapies. And also I remember you saying that maybe figuring out the best dose maybe with those combinations is also in the future for us to work on. So hand in hand with that and the exciting FDA approval for that, it’s easy to then turn to some of the things that you presented Dr. Zeidan around luspa and interested in your recap and take-home points from your presentation today.

Amer Zeidan:

Yeah, so as Uwe was saying, it’s a very exciting era for lower-risk MDS patients. We have two drugs now that have been approved in the last four years in the US and I think many of us have worked with luspatercept. There are other second-in-class agents coming through this. So it’s really I think the start of a very new era for management of lower-risk MDS.

So luspatercept is given subcutaneously every three weeks. It’s generally an easy drug to give. The most common kind of side effect that we worry about is some fatigue, but it tends to be generally not leading to discontinuation, generally improves after two to three cycles. And I think of it more closer to growth factors rather than imetelstat what Dr. Platzbecker was talking about in terms of issues related to the cytopenia. So it’s easier to give and it’s certainly effective.

Initially the data was suggesting that the efficacy is limited to the ring sideroblasts positive patients, but now we know that it works in both patients with ring sideroblasts and without ring sideroblasts. So the main update we discussed in this meeting was some longer-term follow-up data from the COMMANDS trial that led to the frontline approval of the drug. And I think what we have shared is that the efficacy is still there. I think the cumulative period of transfusion independence is up to two and a half to three years. Around 65 to 70% of patients do respond to the drug, slightly lower in ring sideroblasts-negative patients.

But I think in my mind the most exciting kind of developments are going to relate to how do we select patients? How do we combine some of those agents? How do we become more ambitious in terms of defining some of the endpoints?

So far we have focused on transfusion independence as the main endpoint, but I do think hopefully with time we can even aim for things like overall survival improvement as we improve, not only improve the anemia, but really I think we should try to resolve the anemia completely now that we have effective and safe drugs. So I think there is more to come and I think it’s becoming more clear that some of the disease burden, in my opinion, for MDS patients is not fully appreciated by the wider community. I think anemia tends to be somewhat minimized. And I think fighting some of those perceptions, even by the regulators, I think are very important because this is a disease that leads to death of most patients. Most patients with MDS, lower-risk MDS will not die with MDS, they will die from MDS and I think even if they don’t progress to leukemia. So I think trying to modulate that with some of those newer agents is going to be very important.

Hetty Carraway:

Well, I really appreciate your comments around that. It’s really amazing to be witnessing the approval of these two fantastic agents, really changing the landscape for the armamentarium for therapy for our patients and coming together at this very meeting and thinking about how can we improve the future for our patients and also advocate for important changes in terms of how we measure improvements for them and their disease.

We’ve been very fortunate to have a number of people at the meeting help us think through what should be some of these endpoints and also how to measure patient-reported outcomes. So Dr. Efficace, share with us your take home points from your presentation today.

Fabio Efficace:

Well, considering the evolving treatment landscape that we have seen in MDS, it is now so critical to understand what is the patient perspective about the treatment. And when we say to understand the patient perspective, we need to do this in a methodological exam fashion. So we need to address a number of methodological aspects and we also need to make some efforts. And this group have made already important steps to standardize the process and to be able to compare patient-reported outcomes across different trials because this is now a major challenge.

And while in this presentation we also touched upon the use of PRO assessment in routine practice, which is a different setting than measuring PRO in a clinical trial setting, and in clinical practice there might be a number of benefits by assessing in a systematic way the patient perspective in terms of symptoms for example. Of course we might not use lengthy questionnaires in clinical practice because we need to be pragmatic, so we want to select a core set of specific items. But this type of feedback that the patient can provide us can really improve a number of outcomes in terms of treatment adherence, in terms of clinical response. Of course, providing that all this assessment has been done by thoughtfully considering a number of logistical and administrative aspects that are to be addressed before implementing this. So providing that all this framework is appropriately set up, I mean this type of use of PRO in clinical practice might really provide a number of advantages.

Hetty Carraway:

Yeah, it’s such a challenging population because of the anemia and the fatigue that comes with this disorder. And some of the things that were pointed out is that the control group is often being supported with transfusions and then we have to measure fatigue levels when we think fatigue is driven by that very anemia that we’re trying to fix with the drug that’s being investigated. So we’re hoping that you’ll fix this in our PRO and the optimal PRO platform for our patients with MDS as well as the working group, and put our heads together to think about how to best attack this for our patients. So it’s exciting to think about together.

And then finally, Dr. Mosquera, you presented some work today about using artificial intelligence and how to improve our assessment of patients with myelodysplastic syndrome. And it sounds like your background also in myelofibrosis has really led to moving this field forward. So share with us the things you presented and your key take-home points.

Adrian Mosquera Orgueira:

So basically from our perspective in Spain and as we have a broad relationship with Latino countries, for example, we are very sensitive about the unavailability of many of these countries to have genetic testing available for myeloid disorders. And we have came across a big percentage of… 80% of people in the world don’t have accessible MDS testing, at least in a timely fashion, for clinical decision making. So we thought it was a good idea, both in the myelofibrosis and the MDS settings in Spain who have great registry data from thousands of patients with long follow-up, to try to further exploit the information that was collected in those databases even though this information did not have MDS testing available, at least at the very beginning of the studies.

And what we did in both cases was to exploit the full granularity of the data, exploit that granularity with machine learning to risk stratify both myelofibrosis and MDS, obviously separately, and we came across two models with stunning accuracy, which we then actually compared with other models based on genomic and clinical parameters like for example, the MIPSS70 score in myelofibrosis or the molecular IPSS score in MDS. And surprisingly we found that the accuracies of both models were very similar for overall survival prediction. And then we have built on top of this a new version of our risk stratification models that incorporate NGS testing. So what we’ve discovered, and this is very interesting, is that the demographic and clinical parameters in myelofibrosis and demographic clinical lab-based and cytogenetic parameters in MDS are very robust for overall survival prediction. And that mutations add a little bit more prognostication for leukemic transformation prediction.

So we now have models for risk stratification based on available traditional information that can be applicable anywhere in the world, virtually. And now you also have these new tools that you can refine your prognostic estimation for a patient with a genetic annotation. And finally, we’ve just released some data about our project within the EBMT to predict bad outcomes before an allogenic stem cell transplantation in myelofibrosis patients. So our conclusion is that our approach, taking these thousands of patients from the EBMT database and applying machine learning to predict responses based on clinical, disease-related and transplant-related variables, we could actually identify a higher proportion of high-risk patients who have rates of mortality, which are over 50% at two years and non relapse-related mortality at 35% two years. And these patients have similar outcomes of other high-risk definitions, for example, that of the CIBMTR, but the number of patients, it’s more than double. So we believe this can be the starting point to consider which patient to avoid doing an allogeneic stem cell transplantation and in which patient trying to use other medical approaches that are now available.

And just to end up, we are also immersed in a revolution by generative AI, so one of the things that we are now exploring a lot is the generation of decision support systems based on large language models, such as OpenAI or YAMA, for supporting clinical decision making. And we are now exploring this in different hematologic disorders. The results are really stunning, to be honest. The responses that these systems generate when the information provided by the experts is accurate, is really interesting, and it will probably lead to a change in the way we communicate and update medical information.

Hetty Carraway:

Sounds amazing, just the work that you’re doing, thinking about how to assess patients and then even how to expect or predict responses to therapies, and then how to manage the toxicities from those therapies. So pretty amazing work and really fantastic to be hearing these sorts of pieces of information for our MDS community. Thank you so much for the opportunity to be a part of the International Working Group here focused on the management of patients with MDS. Really amazing to be witnessing the use and approval of these two fantastic agents, imetelstat and luspatercept, and to be moving the field forward for our patients with MDS and how to best identify prognosis, response to therapy, toxicities, and patient-reported outcomes. So thank you very much and really appreciate you being here and a central part of today’s work.