Educational content on VJHemOnc is intended for healthcare professionals only. By visiting this website and accessing this information you confirm that you are a healthcare professional.

The Myeloproliferative Neoplasms Channel is supported with funding from Takeda (Gold) and Kartos Therapeutics, Inc. (Bronze).

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.

Share this video  

Molecular endpoints in MPNs: challenges, evidence, and the AmbiMPN study

Jennifer O’Sullivan, MB, BCh, BAO, Guy’s and St Thomas’ NHS Foundation Trust, London, UK, presents on molecular endpoints in myeloproliferative neoplasms (MPNs). Dr O’Sullivan highlights growing evidence linking molecular response with outcomes, particularly in polycythemia vera, and outlines key challenges including assay standardization, response thresholds, durability, and additional mutations. She introduces AmbiMPN, a UK ambispective observational study of molecular response and clinical outcome, while also noting that trial-level validation is still needed before molecular response can serve as a surrogate endpoint. 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

Hey everybody, that was really interesting. I wish we would have more time for that, but talking about endpoints, I’m going to invite Jennifer O’Sullivan to talk about molecular endpoints. Thank you.

I don’t think I’ve ever… I first want to say thank you so much for inviting me to speak. I work at Guy’s with Claire very closely, and it’s really amazing to be here...

Hey everybody, that was really interesting. I wish we would have more time for that, but talking about endpoints, I’m going to invite Jennifer O’Sullivan to talk about molecular endpoints. Thank you.

I don’t think I’ve ever… I first want to say thank you so much for inviting me to speak. I work at Guy’s with Claire very closely, and it’s really amazing to be here. I don’t think I’ve ever had such an extensive background to a talk in my whole career.

Yeah, I think most people have sort of covered at length in many of the talks already what I’m going to cover. So I will sort of skip through sections that aren’t relevant and, you know, give you food for thought, and maybe what we’re trying to do in the UK to address some of what has just been discussed. OK, these are my disclosures.

So I guess, why are molecular endpoints such a hot topic, as you can just see, contentiously discussed at every meeting we go to and discussed at length? Here is just some of the data that’s been shown on the CALR antibodies. So there’s sort of a raft of mutant-targeting agents coming on stream in early-phase trials. This is one that’s looking really encouraging, and we’re seeing early deep molecular responses. So quite different to, say, interferon, where you see, sort of four or five years out, you do see molecular responses, and there are correlations with outcomes with interferon. These are early deep molecular responses. So, you know, so many questions: when should it be measured? Presumably it should be durable? And does this actually correlate with outcomes? And can we start actually looking at molecular response as a primary or co-primary endpoint, or how would that look? And I’m going to come on to maybe some thoughts I might have. I’m happy to discuss further.

I thought this was quite helpful for thinking about deep molecular responses and where we are with, say, the current endpoints that are more clinically based, with symptoms and spleen. This is an illustrative example to highlight a specific point, of two patients. These are example patients. Both of these patients are on a Phase III clinical trial of combination therapy. And you can see on one side of the plot the symptom score of the two patients, patient A and B, and they reduce slightly, but not significantly. But if you look at the allele burden of the patients, they diverge. So one patient’s allele burden increases slightly. I wouldn’t say necessarily progression, as I’ve stated, but, you know, 50–54% is unchanged, whereas one patient does achieve a molecular response.

And as Alison commented on, most of the MF trials have a 24-week primary outcome, so they’re very short. And we know these diseases develop over, you know, we know their JAK2 mutations are present at birth, and they take 20, 30, 40 years to develop. So is 24 weeks sort of… Are you really going to be able to tell you’re modifying the disease in a way that’s durable at that point? But in any case, you know, this discord… One example is navitoclax in TRANSFORM-1, which missed the key TSS secondary endpoint. But in the study, there was a link and a suggestion of VAF reduction associated with longer PFS, so a 20% reduction. So you can see that there’s a discord in the field, and sort of an issue that we need to address.

And can a molecular endpoint be a valid surrogate? Lots has been said about this, but these are some examples. And David sort of commented, and it has been talked about, myeloma. But I think this is to highlight for us in the field a route to approval, how it was done in myeloma. And I guess, listening in the last few hours, that took a very, very long time. Thousands of patients. I mean, this study, this couple of… [unclear] from 8,000 patients pooled from trials. I think we need to do better than… necessarily what we have in MF. I don’t know. I’ll come to it. But 8,000 patients is sort of eye-watering, right? And the biomarker for myeloma was MRD negativity, so a different marker. I’ll come on to thoughts about whether we should or shouldn’t start thinking about thresholds. But the measure is MRD negativity.

Evidence provided, you know, really came from randomized controlled trials. And you can see in August there was approval of iberdomide. And this was a co-primary endpoint, but it was initially based on MRD negativity, and the other primary endpoint, PFS, has yet to read out. So whether or not for MPN, could you apply molecular response as a co-primary with, as Amylou quite nicely said, maybe symptoms, maybe hematological response or survival? Don’t know. They’re still currently secondary endpoints.

I’m going to briefly go through this. I think everyone’s really familiar with all of the data that we have on outcomes with molecular responses. I’m not going to go into the MAJIC-PV trial, but I will highlight that when we looked in the stem cells at four or five years, so patients who had excellent molecular responses, who virtually cleared their allele burden in the peripheral blood at four or five years out, we looked at the CD34-positive cells, and they didn’t have any of the mutations, so it’s quite nice to see in PV. This has been replicated, correlating molecular response to outcomes, by other groups. So I think there is, you know, increasing, pretty convincing data in the space of polycythemia vera.

I think molecular response in myelofibrosis, there are some encouraging signals, but I think 20% VAF reduction needs a lot more validation before we can start just using it. Earlier, if I was to say, we haven’t really talked about the source of the allele burden. So, you know, varying trials may have used whole blood or granulocytes. Are they directly comparable? How they were sequenced, is that also directly comparable? So I think there are quite a lot of questions in terms of standardization. You know, CML, the exemplar of using a surrogate for survival, has very clear standardization of the assay. We’re so far from that still in MPN. Just as sort of a comment, looking at 20% is just a little bit subtle. And is that durable? I think none of the trial data at six months, 12 months… we don’t know that yet. And I don’t know if they are banking the material. Hopefully they are.

I think a really clear context… I mean, it is logical: you reduce the percentage of the mutant clone, you’ll do better. But this is quite a nice, obvious example. Whilst it’s single-center and published in the New England Journal in the transplant context, patients who cleared the mutation at day 30 post-allo had better outcomes, irrespective of JAK2 or CALR, and it outperformed donor chimerism. So I think it’s quite compelling, even though it’s in the context of transplant.

What about CALR versus JAK2? So for CALR molecular responses, there’s less data with standard therapies, and these are sort of busy plots with varying horizontal lines. So there are some responses. CMRs are very rarely reported, but there are some responses, certainly on interferon. We looked at this, this is an observational retrospective analysis, in a number of patients in varying centers, and there were molecular responses in patients on interferon and ruxolitinib, which has been reported and known, but we observed that with anagrelide, the VAF went up in the majority of patients. We extended this beyond CALR and looked in JAK2, to see, oh, is it specific to CALR that with anagrelide the VAF goes up? But it was across the board. And this was also shown in SURPASS-ET quite recently.

Whether allele burden increasing… does it matter? On the opposite end of molecular response being associated with good outcomes, does an increasing allele burden predict progression? This is the same analysis that we did. It was observational, so to try to mitigate bias, we did a landmark analysis and showed that the allele burden increasing over time was associated with progression. So it may be useful as a biomarker, but it really needs a lot of proper validation within external data sets and prospectively. But it’s certainly what you might think would happen.

And we haven’t talked at all, I don’t think, yet in the meeting about beyond the driver mutant clone. I’ve been, you know, thinking about how a lot of patients carry mutations in other clones. So I think MPN is a little bit more complex than maybe thinking about, say, myeloma, or a lot of CML patients. And we know they confer prognostic significance, particularly in MF. So thinking about splicing factor mutations, RAS pathway, TP53, and also some mechanisms of resistance to ruxolitinib. But are they relevant? And if we’re trying to think about a molecular response, should they be included as part of that, whether that be negativity or looking at the VAF of those as well? There’s not a huge amount of data, actually, correlating them robustly. There’s a very small analysis on the side, looking at TET2 mutations in PROUD-PV and CONTINUATION-PV, and those allele burdens reduced in those patients. But then in the DALIAH trial, emergent clones, I mean, these are clearly sort of CHIP clones, but these increased on the study. Does that matter? Do patients do more poorly? And, you know, DNMT3A in this analysis didn’t confer an adverse outcome. So I think the data are inconsistent on whether or not these are bystander CHIP, or I think it really probably does depend on the mutation. But there’s a lot of question marks, for me, around that. And serial, wider NGS beyond just the driver mutation is certainly needed.

So a lot of what I’m going to say now has already been said about what the challenges are in thinking about these. For me, standardizing the assay and harmonizing that across the board. So if something’s sequenced on a different platform, from a different source, from different trials, can we really carefully cross-compare, when it’s already such a challenge trying to work out if we already have sufficient data? And yeah, do we have enough evidence to start using it? Thinking about how the myeloma field did this, and maybe a specific consortium directed at endpoints, which I think has already been mentioned today, as is being looked at in AML. But that’s industry and academic jointly; I think there’s no way around that. Listening to the data-sharing agreements and the duration is sort of, yeah, slightly heartbreaking. Then trial designs, and how we can actually capture these diseases that have 20–30 years of progression, when we need earlier outcomes. And then what are the definitions that we need to think about for these? Partial molecular response, complete molecular response: are they appropriate? Are they clinically relevant?

Okay, so what do we need? I think Tanya talked about this, with clearer, better expertise, but we have encouraging signals associating molecular response with outcomes, and we really need to link that together at the trial level. And can we do this with the data we have already? I guess this is provocative, asking the question; I don’t know the answer. I think maybe in PV we might, actually. Pooling all of the trials would make sense. If the DNA is there, should they be resequenced? Possibly. Is the source the same? Does that matter? And is there sufficient follow-up? You can see here, these are only a small example of the studies, but for some of them, the follow-up isn’t necessarily terribly long.

So this brings me to an effort in the UK that we’re trying to address this with. This came about more from actually the patient-facing side of things. Patients in clinic, I don’t know about in the US and in other countries, but in the UK they ask us, “Can you check my allele burden?” all the time. They’re always asking about it and want to know, can you change the drug, add a drug, start a drug, do all sorts of things with the allele burden? They think we have enough information. So actually, that was the motivation for this study, and also coming to meetings and hearing about it being discussed at length.

So we have a CRUK-funded study that we’re going to start this year. It’s not simple, but there’s one overarching question: does a molecular response in an MPN driver correlate with better event-free survival? And we’re going to look at that in almost 1,300 patients. I’m leading this with Alesia Khan, who’s a hematologist working in Leeds, where the molecular testing will be centralized. We’re looking across PV, ET and myelofibrosis, particularly trying to enrich for myelofibrosis. I think there’s quite a reasonable amount of data in PV. At the moment we have funding for UK sites, but ideally, if we could open internationally, that would be fantastic. I don’t think we have enough patient numbers.

But patients essentially will enroll onto the study. They need to have a DNA sample stored. The term ambispective is being used, so we’re looking retrospectively and collecting DNA from diagnosis or start of treatment, and then we’re going to prospectively collect DNA annually, along with really detailed clinical information on these patients, so that we can try to capture at least more than, say, five years of follow-up by having the historical data. With this, we’re hoping to maybe better determine some of the molecular response thresholds, at least those that correlate: 50% may be correct, it may not be; it may depend on the disease subtype; maybe 20% in MF is correct. By centralizing the assay, we might remove some of the biases and concerns about different assays and different centers doing the test. And I’ve already spoken to the longitudinal follow-up by using this design.

I guess what this clearly isn’t, with this observational study, is randomized controlled data, but I think we will try to prioritize including patients who are already on academic and commercial clinical trials. And just to give you an illustrative example of why the term ambispective, and why we’re taking this approach: this is an example of a patient with PV on interferon. They enroll onto the study and they have an allele burden of 11%. And you can see they’re enrolled here for a number of years, and the VAF goes down slightly. But when they started interferon a number of years prior to that, their starting VAF was 42%, so you would miss that they had already achieved a molecular response at the time of the study. So it’s critical to establish, for any patient entering the study, whether they have or have not achieved a molecular response, or where they’re at in that.

So I’ll leave you with this slide, my own thoughts on a roadmap to molecular endpoints in MPN. I think we do have increasing evidence, and quite a lot of evidence, that molecular response is associated with better outcomes. I think there’s quite a lot to be defined in terms of what the thresholds are, when, and the durability. Thinking beyond JAK2 and CALR, I think it’s very important with mutant-directed therapies to know if it’s in the same clone or a different clone. If someone has a TP53 mutation in an independent clone and you remove, say, the MPN clone, what will happen to that patient? Standardizing the assays, hopefully with initiatives that are going on, and maybe with [study name unclear] we might provide more data supporting this. And then trial-level validation is certainly needed, whether by pooling the controlled trials we have already, which I completely agree with, and whether we have the data within those to do that and harmonize it. So with that, thank you very much. Happy to take questions.

AI-generated transcript. Please refer to the video as the authoritative source.

Read more...

Choose your cookie settings

VJHemOnc uses necessary cookies to make the site work. With your permission we'd also like to use analytics, preference and marketing cookies to understand how the site is used and improve it.

Choose an option to continue. You can change it at any time on our Cookie Policy page, linked at the bottom of every page.