Loading summary
Transcript17 lines
- [00:02]
A
Welcome to the American Society of Hematology Conversations with Blood Authors. This Blood Podcast episode is hosted by Dr. James Griffin from the Dana Farber Cancer Institute in Boston. He discusses rapid clonal selection within early hematopoietic cell compartments, presage's outcome to ivocitinib combination therapy with Dr. Paresh Vyas, and impact of modernizing eligibility criteria on enrollment and representation in AML clinical trials with Dr. Andrew hantle.
- [00:32]
B
Welcome to this edition of Blood Podcast. We have two really interesting topics to discuss that are related to how we treat and understand the acute myeloblastic leukemia. The first paper is going to be presented by Dr. Paris Bias. It's entitled Rapid Clonal Selection within Early Hematopoietic Cell compartments presages outrunta ivositinib combination therapy. Dr. Vyas is in the Hematology Unit at the Wetherald Institute of Medicine, University of Oxford in the UK.
- [01:06]
C
Dr. Vyas thank you Dr. Griffin. It's a pleasure to be on this podcast. This paper is about understanding mechanisms of response and resistance to a subgroup of ML patients who have an acquired mutation in the gene isocitrate dehydrogenase 1, or IDH1. It constitutes about 8% of AML. These patients have a targeted inhibitor therapy available widely. It's called ivicitinib. It targets specifically the mutant protein. This inhibitor has now been combined with other treatments used for acute myeloid leukemia, both intensive chemotherapy and venetoclax and azacytidine. The first triplet combinations of this drug were carried out by Dr. Courtney Donardo at the MD Anderson Cancer center, and they showed an extremely high initial response rate of over 90%. And they showed in a heterogeneous group of patients who had relapsed refractory AML and newly diagnosed aml, that this response was far in excess of the single agent therapy. But it was unclear at the time how durable these responses would be, and it was also unclear at the time what the mechanisms of resistance to combination therapies would be. This is important because because they're currently phase 3 clinical trials either completed or in progress with combination therapies compared to doublet therapies of either Sidenote asacitidine or intensive chemotherapy alone. So the study of these patients is important for us to understand what these phase three trials may give us. It's also important for us to understand what the mechanisms of resistance might be. So we studied a cohort of patients. We studied patients who had longitudinal bone marrow sampling from trial entry through initial response either into patients who had relapsed or patients who remain disease free years after starting treatment, three to five years after starting treatment. And we conducted deep single cell analysis not only of the leukemic bone marrow trial entry, but subsequent during treatment. We employed a method that we call Target SEQ that combines genotyping with transcriptome analysis. The genotyping is very high fidelity. It's over 90% successful for multiple mutations in a single cell. And what we were able to show very clearly was that within three months of starting treatment, the response in the bone marrow across the stem, progenitor and mature cell compartments really presaged the long term outcome of these patients. So in patients who remain disease free years later, they had cleared all their leukemic mutations very early on or had clonal hematopoiesis, that is single mutations in DNMT3A or TET2. In those patients who relapsed, the relapsed clone was selected within the first three months. Often it was not apparent at trial entry. In some cases it was. But the relapsed clone clearly had a selective advantage very early on. And if you think about it, this makes complete sense because selection has to operate as soon as a patient receives drugs. But you could see this most beautifully in the compartments in the bone marrow that have leukemic stem cell potential. We were able to show that there was a common mechanism despite the heterogeneity of mutations involved in the relapse clone. So we found that there were common mechanisms of resistance across these patients, despite the fact they had heterogeneous mutational profile in the relapse clone. And it came as a complete surprise to find that these patients had common signatures of menin gene activation despite the fact that they didn't necessarily have an NPM1 mutation or a KMT2A rearrangement. This would argue that menin inhibitors may well be useful in patients outside their current clinical indication. There are additional features of the resistance phenotype that was seen in the transcriptome which you can read in the paper. But I think the general messages from this paper are as follows. Firstly, there is very rapid selection of either leukemic clones that lead to resistance or clearance of leukemic clones if you look deeply enough at a single cell level in the stem progenitor compartments. This is a general approach that can be used across not only acute myeloid leukemia, but all cancers. Secondly, it shows that there may well be common mechanisms of resistance and this may well include mechanisms where we really need to think again about what drugs could be useful? Clearly there's a lot of additional work that will follow from this and we look forward to future advances, not only from our lab, but other labs building on this work.
- [07:27]
B
Thank you very much, Dr. Vias, for that interesting presentation. A couple of questions for you. One is, as you mentioned in your paper, the technology that you use to follow the development of these clones and expansion of these clones is not really scalable to clinical practice. And yet the questions they're asking would be very much of interest to the patients and the physicians taking care of them. How do you go from these observations towards a clinical test that might predict who's going to relapse and who's going to going to go into remission?
- [08:04]
C
That's an excellent question. This will require a partnership between academic labs like ourselves and industry. It will require a partnership with diagnostic labs, and we will need to make simpler, sensitive, reproducible tests that we can apply early during the course of disease. These conversations are ongoing, but I think it'll take a little bit of time before we have scalable tests that we can initially test in clinical trials and if validated, that then can be rolled out to the community.
- [08:47]
B
Also, you showed in some patients that although a resistant clone developed quite quickly, they didn't actually relapse for sometimes years after the detection of that clone. Do you have evidence that these new clones evolve further over time, or are they just suppressed by some patients, perhaps through an immunological mechanism? Any thoughts?
- [09:13]
C
Yes. We have no evidence of clonal evolution beyond what we identified. We, we performed hull exome sequencing at the time of relapse just to be sure that we hadn't missed coding mutations in these individuals. And my best guess is that the dynamics of clonal expansion are indeed a lot more complex than we understand. And I suspect there are at least two, if not more mechanisms at play. Play? You've hinted at one, which is the immunological mechanism. And I genuinely do believe there is a reason why certain leukaemias are good risk with respect to treatment. There are often translocation positive leukemias in intensive chemotherapy and for example, the NPM1 mutation is often associated with good risk. These are immunogenic neoantigens and I suspect we have should be looking for autologous immune responses against them. This is something the field needs to pay attention to. Secondly, I think there is clonal competition and it's quite interesting that if you have an antecedent, for example, dysplastic clone that gains an advantage under AML therapy, this may retard the relapsing ML clone, and we have some preliminary evidence for that in other studies, so there's a lot to learn. Dr. Griffin here and thank you for that question.
- [10:49]
B
Thank you very much for that excellent presentation. Our next podcast is going to be presented by my colleague in the Leukemia Program at the Dana Farber Cancer Institute, Andrew Handtel. The title of the paper is Impact of Modernizing Eligibility Criteria on Enrollment and Representation in Acute Myeloid leukemia clinical trials. Dr. Hantel.
- [11:13]
D
Thanks, Dr. McGriffin for having me. I'm excited to talk about our work. So, at a high level, our study started with a fairly simple question of our eligibility criteria in AML clinical trials actually doing what we think they're doing. These criteria are really meant to protect patient safety and define a study population. But there's been growing concern that in practice they may be excluding more patients than are necessary and then doing so unevenly across different groups of patients. And this concern isn't new. Over the past several years, the FDA and professional societies like ASH and ASCO have issued guidance and recommendations encouraging investigators to modernize their eligibility criteria, which means to specifically base them on known or anticipated drug safety signals rather than on general precedent or excessive caution. And the motivation behind that guidance is the recognition that many eligibility criteria, at least anecdotally, had lacked a clear safety basis and instead kind of selected an artificially healthier trial population that might not actually reflect the patients who would ultimately receive the drug once it's approved. And this issue is particularly important in AML because these trials frequently rely on single arm phase two designs where efficacy and safety are often interpreted using historical clinical controls or cross trial comparisons rather than randomized comparators. And in that setting, eligibility criteria can effectively define the benchmark for success. The trial population is unusually healthy, outcomes can appear better than what we see later in real world practice, which has implications for efficacy estimates and safety expectations. Despite this guidance and these concerns, it hasn't been clear how closely real world AML trials actually have followed this safety based approach to eligibility or, or what the population level impact of doing so would be. And that's really the gap we're trying to address here. What makes our approach here different is both, I would say in the scale and kind of the particular methods that we used. So we had previously examined 190 phase 2 and 3 frontline AML trials and essentially abstracted all of the eligibility criteria written verbatim in the trial protocols. What we're calling the trial based criteria in the manuscript and Then we, we used kind of this existing FDA and professional society guidance and all of the data that was used on the drug to that point and that was known about the safety of the drug at the time the study was conceived and used those data to create, according to the guidance, kind of a safety based set of eligibility criteria for each of those trials. And we had previously just analyzed differences in those two sets of criteria for each of those 190 trials. And then in this manuscript we are applying those to a large multicenter cohort of over 2,200 patients with newly diagnosed AML in the United States. And this allowed us to ask a very practical kind of counterfactual of what would eligibility have looked like if these trials had followed their initial approach versus the trial approach versus the safety based approach when patients were being approached to enroll. And the results were fairly striking. Using the eligibility criteria as written in the protocol, you know, patient, about half of patients were eligible for a trial for a given trial. And then when we applied the safety based criteria, the eligibility increased up to in some cases over about 80%. And that represents a dramatic expansion of trial access, which is not achieved by really lowering safety standards, but more aligning them, aligning this eligibility with evidence. We did note that because in aml, age plays I would say an outsized role compared to other trials in terms of being in eligibility criteria, we did additional analyses kind of removing any differences related to the age based criteria and still saw double digit absolute increases in eligibility due to the kind of common organ function and prior medical history related criteria that's often used and often seems to be over restrictive relative to the actual safety data that's present. And we also looked beyond kind of change, general inclusion in terms of representation by the downstream demographics that have been noticed as being different between non Hispanic white patients and black, Asian and Hispanic patients. And we saw consistent differences between those groups. And when we applied kind of these safety based criteria, we actually saw that compared to the trial based criteria, not only was eligibility increased for everybody, but it also decreased the differences in eligibility between the groups, suggesting that this modernization can both improve, improve overall inclusion and decrease differences that we've seen previously. I would say one of the most actionable findings was really identifying which criteria appeared to matter most. Several stood out as being both kind of exclusionary generally and then associated with differences between demographic groups and those related to cardiac disease, liver disease or liver function testing, upper age limits and prior malignancy. And these are also criteria that vary one widely across the trials, suggesting that it's often the variability itself, meaning that they're not really grounded in safety data as far as we can see. So our key takeaway is really that eligibility criteria should not be loosened indiscriminately, but that they should be intentional, justified and evidence based. You know, they're really design choices. And our data show that design choices that are aligned with existing federal and professional society guidance can really substantially improve access as well as representation and in the end, generalizability to the patient in front of you. And while AML is a particularly stark example because of this reliance on single arm trials, historical controls and potentially the age criteria, I think our approach is also highly transportable to both other hematologic malignancies and potentially other solid tumors and beyond. I hope that listeners take away that trial access, patient safety and scientific validity are really deeply connected here. And by aligning eligibility criteria with drug safety rather than habit or precedent, we can can really design trials to better reflect the patients that we actually treat and generate evidence that's more reliable and more useful in real world practice. Thanks again for giving me the opportunity to discuss our work.
- [17:23]
B
Thanks very much for that Excellent presentation, Dr. Hantel. And the increase in potential eligibility for clinical trials that you report, if it is indeed possible to do that, is really impressive. So what's next? How do you prove that these new eligibility criteria for AML patients in particular will in fact be safe and not increase the toxicity, particularly in patients who are elderly or who have other comorbidities, the ones that patients that are often excluded from trials like this.
- [17:56]
D
Two points in there. One is kind of what are the implementation paths forward for something like this? And then the monitoring. So I think that for the first one, there are two ways that we're moving this forward. The first is to be more, more intentional with the trial investigators and sponsors in terms of how their trial criteria are actually developed. And so we have ongoing projects where we are working with those, I would say, types of individuals to kind of be thoughtful about their criteria, both in terms of the initial selection in the earlier phase one and phase two studies, and then how we actually change those based on the toxicity and the things that we're seeing as a result of, of those trials when we're designing the larger phase three. And then we also have the opportunity to look more at the review process itself. And in that case, we have an ongoing study looking at how training institutional review boards and other regulatory bodies around the FDA and professional society guidance would actually change the comments and conditions they would be placing on Studies as they're in development such that they become more aligned with this guidance and the safety basis. I think that there is potentially for the second question, a little bit of a disconnect between knowing that there will be outcomes for patients who have more comorbidities at baseline or who are older will very likely lead to, I would say, survival outcomes that are less than they would be if those patients were not involved because they are older and potentially sicker at baseline. But I don't think that necessarily means that those patients should be excluded if there is an expectation that, you know, those patients would generally be receiving the medication post approval and that we need to kind of be able to study the populations that we're intending to treat in the trials that we're using to generate the evidence for the FDA approvals. And so I think there's not a, you know, one size fits all answer for should X eligibility criteria be exactly this for this trial? I think it's more that we need to have a reasonable expectation that they should not be copied and pasted, that they should be based on safety based expectations and guidance that is out there. And it's more that there seems to be a large variability in how that guidance is used and also in just kind of what the baselines should be. And I think those are both places that we could move to where we can generate evidence that is good for kind of all patients, not, not just the more artificially healthy and selected patients that we often include in our trials.
- [20:34]
B
We know in AML and other hem malignancies that as new therapies are moved out of clinical trials and into the real world population, the results are often not quite as good. Will this help to address that issue?
- [20:48]
D
Yes, that is the hope and expectation for this. What we're doing in some of the projects that I mentioned previously is not only are we actually following to see, let's say if one criteria is changed either when it's developed initially or during the course of development due to, you know, IRB review. But what happens when you kind of have those additional patients included? Like first, does the actual change lead to X patient who has, you know, well controlled hiv? Getting into the trial, were there any patients like that that actually got on because of the change in the eligibility criteria? And then downstream looking at those patients relative to everybody else in the the trials and saying, were they always on one side of the survival curves or were they kind of distributed more evenly throughout it? And so those are things that we're looking at both proximate and downstream.
- [21:37]
B
Thank you for that really elegant discussion and good luck with getting this into practice around the world. It strikes me that these same issues occur in solid tumors, as you mentioned. Quickly, are they similar efforts to improve eligibility for clinical trials in solid tumors?
- [21:56]
D
Yes, we actually have a lot of these ongoing studies are across different cancer types. We have particular, I would say, focuses on hemelignancies. And there are particular things like neutrophil counts, which are very much not applicable to aml, that are a large issue for people with duffenyl phenotype that I know Ash has been taking a leading role in trying to address inequities related to that. And so I think there are large kind of overlaps in terms of implementation. And so we're trying not to be exclusionary on that front, but make sure that we also acknowledges the baseline differences in the diseases and the drug development pipelines.
- [22:33]
B
Well, thank you very much to both of the presenters for excellent discussions of these papers.
- [22:40]
A
Thank you for listening to this Blood podcast of conversations with blood authors. To read the articles, visit bloodjournal.org this episode is copyrighted by the American Society of Hematology.