Early Detection of Acute Myeloid Leukemia (AML)

How specific gene mutations can enable prediction of acute myeloid leukemia (AML) up to a decade prior to disease development.

Guest: Pinkal Desai, MD, hematologist and medical oncologist at the Weill Cornell Medicine and NewYork-Presbyterian Hospital Leukemia Program.

Host: John Leonard, MD, world-renowned hematologist and medical oncologist at Weill Cornell Medicine and NewYork-Presbyterian Hospital.
Early Detection of Acute Myeloid Leukemia (AML)
Featured Speaker:
Pinkal Desai, MD
Guest Bio
Pinkal Desai MD, MPH is an Assistant Professor of Medicine at Weill Cornell Medical College and Assistant Attending Physician at the New York-Presbyterian Hospital.

Learn more about Pinkal Desai, MD

Host Bio
John P. Leonard, MD, is a world-renowned expert in the research and treatment of lymphoma and other cancers, and is devoted to providing personalized and compassionate care to people affected by these diseases. As the Associate Dean of Clinical Research at Weill Cornell Medicine and NewYork-Presbyterian Hospital, Dr. Leonard is a leading proponent of the value of clinical trials in delivering novel therapies and cures to patients.

Learn more about Dr. John Leonard
Early Detection of Acute Myeloid Leukemia (AML)

Dr. John Leonard: Welcome to Weill Cornel Medicine Cancer Cast, conversations about new developments in medicine, cancer care, and research. I'm your host, Dr. John Leonard and today's topic is early detection of acute myeloid leukemia or AML. My guest today is Dr. Pinal Desai. Dr. Desai is a hematologist and medical oncologist at the Weill Cornell Medicine Leukemia program. She specializes in the treatment of acute leukemia, myelodysplastic syndrome, and myeloproliferative disorders. We're excited today to talk about Dr. Desai's latest study, which was recently published in Nature Medicine. This study was the first of its kind to demonstrate the relationship between specific gene mutations and the risk of developing acute myeloid leukemia. This data suggests that we may be able to uncover the potential to predict disease risks up to a decade before the onset of actually manifesting leukemia, so this is a very exciting new development. Thank you for joining us today.

Pinkal Desai, MD: Thank you for having me here.

Dr. Leonard: I would like to start off by obviously acute leukemia is a challenging disease. It’s an exciting disease scientifically but obviously clinically has with it a lot of challenges in taking care of patients and a very serious illness for many people. Tell us a little bit about how you found yourself drawn to working in the leukemia field.

Dr. Desai: When I started doing my fellowship and was getting my first exposure to oncology, what drew me to acute leukemia is the patients. I found that taking care of patients with acute leukemia was intensely satisfying because they do come very sick in the hospital and you develop a very deep relationship with them quickly. It's very satisfying to help them go through different phases of treatment, be part of a very intense close team and the science of the disease is also very fascinating. 

Dr. Leonard: Before we get into the specifics of your study which I find very exciting, I think it’s important to understand a little bit about acute myeloid leukemia or AML. There are many types of leukemia and some people in our audience may be familiar, but others less so. Just in a nutshell, if you could give a little bit of background as to what is AML, why is it so difficult to treat in many cases and the usual course for patients with AML before we get into the specifics of your study?

Dr. Desai: Acute myeloid leukemia is a cancer of the bone marrow where the basic pathology is that there is a proliferation of immature cells in the bone marrow that keeps on occupying more and more space so the normal cells that are supposed to make the red cells and white cells and platelets, they don't have a way to manufacture them and it leads to a bone marrow failure state. Patients present with more of the time low blood counts and at the same time they could have a high white cell count which is basically all of these leukemia cells or immature cells that are released into the circulation, but you certainly can have patients who have nothing seen in the blood but when we look inside the bone marrow by doing a bone marrow biopsy, we find this abnormal proliferation of cells. These cells lose their ability to manufacture normal cells and they get stuck at that immature stage causing the phenotype of the disease.

It's a very challenging disease and one that is tough to treat because, by definition, everyone's leukemia is different. In the old days, we didn't have very complicated and sophisticated ways of measuring the underlying genetics of leukemia, but now we know that the signature of one person's leukemia is very different than someone else's. A more complicated aspect of it is that it's a very heterogeneous disease to begin with so it's not that AML is completely made up of one clone of cells. There might be several at the same time and you have to try to clean all of these out in order to cure a patient. When they come into the clinic or hospital, the treatments of leukemia are also challenging. In a young patient, usually that would include pretty intensive chemotherapy and a hospitalization of about a month. A majority of the younger patients would need a bone marrow transplant in order to maximize their chances of cure. In older patients, many times they're not able to go through such transplants in which case the goal of treatment is to get them into remission and try to keep them there, but we do not cure the majority of older patients with AML. Even in younger patients, there is still the risk of relapse, even though they go through a stem cell transplant. As a disease, not only we have constant challenges of how to best target this disease with combinations of leukemias or other targeted drugs, but to prevent the relapse that is very high in a lot of patients.

Dr. Leonard: I think that’s a great explanation of a complicated topic. You alluded to the genetics of leukemia and in part I think because it’s been so easy in researching leukemia and studying leukemia that you have easy access to the cells whether it’s the blood or the bone marrow, there's been a tremendous amount of the genetics that have been studied and explained or at least correlate with the different outcomes and we’re going to get to your specific study in a second, which is largely based on those genetics in part. Tell us a little bit about the spectrum of genetics in patients with AML in the big picture, not specific genes, but just the sense of how many genetic mutations and how different patients can be in their genetic profile.

Dr. Desai: The one thing I want to clarify is that when we talk about genetics, the genomics of leukemia, these are not gene abnormalities that people are born with. There are some familial leukemias where there is something that people are born with that make them high risk of developing leukemia, but that is an exceedingly small percentage of leukemia patients. The genomics of leukemia is disease-related mutations. These are gene abnormalities that are present within the leukemic clone and not outside the leukemia clone. When we do a very deep sequencing of these leukemia cells, you can have over 200 individual genes that could be mutated in patients with leukemia. Moreover, there could be combinations of genes so you can have a patient with five genes that are mutated versus just one and it absolutely matters where these gene mutations are. When we talk about the prognosis of AML or the risk of this leukemia coming back in the future, that is completely dependent on two big factors. One is the chromosomal abnormalities which are big arrangements in chromosomes in the leukemic clone and the second is individual gene mutations that are within the DNA. Both of these in combination help us determine which leukemia patient is more likely to relapse versus not. You can generally in a ballpark way estimate the relapse risk as low intermediate or high, but it's very complicated because you have to take into consideration all of these 200 genes and various combinations to come up with some kind of a score that helps us understand what is the risk. This is important because if somebody has a low risk of relapse, then we would not transplant these patients, but if they have an intermediate or high risk of relapse, we strongly consider stem cell transplant or a bone marrow transplant in order to maximize the chance of cure for these patients.

Dr. Leonard: Tell us a little bit about your recent study, which has gotten a lot of attention and Nature Medicine is really one of the leading  translational scientific journals in the field, so I think that speaks to the importance of these findings. Tell us a little bit about what you were trying to do in these studies that were performing.

Dr. Desai: One of the things that was of major interest to myself and the leukemia program at Weill Cornell Medicine in general was could you perhaps identify patients who are at risk of developing leukemia. We’re not talking about an inborn genetic risk, but something that people acquire over time. If you consider other cancers, there are certain cancers where there exists a premalignant state for example in uterine cancer that’s the most common or breast cancer where you can have this premalignant state that increases the risk of cancer in the future. Such a thing was not known in AML. There was no known genomic state that exists many years before that can help us understand what this patient will go through. About a few years ago, there was this concept that was floated around that patient who is healthy and does not have a bone marrow disease could have an acquired genetic mutation, so they're not born with it, but some mistakes happened over time as patients age and you can detect certain mutations. Nobody knew whether these mutations are definitely tied to AML risks, so we were able to identify in the study we did a big epidemiologic study and did a genomic analysis of patients who were diagnosed with leukemia about 200 of them and we were able to get blood samples on them 10 years before the diagnosis and compare it with people who never developed leukemia and had blood drawn also at the same time.

We found that it is a difference in people who had leukemia where we found an excess of these mutations about four times higher odds of having a mutation 10 years before diagnosis compared to people who never developed leukemia. More importantly, the risk is different depending on the mutation you're looking at. Just having any mutation does increase the risk of leukemia, but within that general statement, there is a panel of what we call high-risk genes where the risk of leukemia is extremely high, sometimes 50 times higher than somebody who did not have that particular mutation. This is an important finding because it establishes that you can identify normal people who have a risk of leukemia in the future because we used to believe that leukemia just happens suddenly. You develop something, the bone marrow starts proliferating these bad cells, and within weeks, you make the diagnosis of AML. To have something that could potentially be identifiable as a risk 10 years before opens up a huge research area because now we can know who will get it and perhaps there is a way to monitor and intervene so that these people don’t get AML.

Dr. Leonard: I think this is very early, but very exciting, and I think you highlight that you can potentially identify a person who might have a significantly higher risk in the future of getting leukemia. Let's maybe play along with me a little bit and recognizing there's a lot of dots to connect and validation and interventions and we often say that just because we know there's a problem doesn't mean we have a solution to the problem. In the big picture, what does that mean for patients? If a patient had a blood test, and let's say it’s a healthy person, and their risk of leukemia in the highest risk group based on this study would be 50 times normal risk, is that the highest risk group? What does that mean in real terms? Is that 1-in-100 chance? 1-in-1000 chance?

Dr. Desai: I like to stress that leukemia is still a rare disease, so the vast majority patients do not get leukemia and on an average, there are about 20,000 new diagnoses of leukemia per year in the United States. The concept here is we certainly do not want to, based on this study, just start intervening on these patients because there are still people who will never get leukemia even though they have the mutation, so what would make sense is that certain mutations in the study just to quote a couple mutations, the TP53 and the IDH mutations carried the highest risk of leukemia and at least in our study, everybody we identified with this mutation developed the disease down the line. This needs to be validated and there was another study that looked at it similarly and also found that these mutations carry a much higher risk than the others. What this would mean for a patient if they come into the clinic right now with these mutations is I would definitely favor a close monitoring program for these patients.

We also found as part of the study that we can find these mutations 10 years ahead of time, so that is not a push to do something right at the moment where they are identified, but they have to be monitored. I think that is an increasingly common thing to do now when we find these mutations. What we found also is that if the percentage of the mutation in the blood keeps on increasing over time, there is an actual trend as to when leukemia might happen. I think that’s where for the immediate impact of the study would be impactful. In the past, we would tell them we don’t know what this means and you don’t have a disease and there was no real push or data to suggest that we should monitor these people, but I think with this data, close monitoring would be warranted as to whether you can actually do something to prevent this is a big area of research that needs to be done.

Dr. Leonard: As you know well, one of the big risk factors out there for getting AML is another cancer and treatment for other cancers. For instance, the scenario of getting chemotherapy or certain types of chemotherapy for breast cancer as an example or radiation for other conditions, those might be a risk factor for ultimately later developing an AML to envision the possibility that one might before someone treated a patient for some other medical problem or cancer that one might screen for these mutations and therefore perhaps not use radiation or not use certain chemotherapies because of a higher risk of getting it. Is there a link there or those two separate phenomena that you wouldn’t potentially connect those?

Dr. Desai: That’s a great point and I think everybody is pushing to this concept that in a normal healthy population, if you can identify people who have a higher risk of AML, if you increase the likelihood of developing leukemia in the population which would exactly be people who have had another cancer or chemotherapy, these are the people you would want to monitor more closely. I want to be cautious about saying that screening for these mutations before treatment of their other cancer that should we change the management of that primary tumor based on the presence of this mutation at this point, I would advise that that is not the way we would envision it for the moment because you have to make sure that the kind of treatments that cure that patient, we maximize that. We cannot give something that impacts their survival from that tumor in order for a future risk for AML. I do completely agree that these are the people we should be monitoring and there will be a push to do that but not change management at the moment. I think that this calls for a prospective monitoring of these patients and understanding if these mutations change over time before and after chemotherapy for the primary tumor. Once this relationship is established that perhaps giving chemotherapy for another cancer brings on a more complex mutation down the line compared to what was screened before, there may be trials that would focus then on saying can we avoid chemotherapy, but I do have to make this point that all of this has to be in conjunction with our solid tumor oncologist because you certainly don’t want their primary solid tumor to be not treated well because of a future risk.

Dr. Leonard: I’ll give you one or two other scenarios that come to mind that might be practical and I recognize that you're speculating. A bone marrow donor for a family member who had one of these mutations, do you envision screening donors and perhaps not using such individuals as donors? Where do you think that fits?

Dr. Desai: I think I do envision that because when you're thinking about a leukemia patient who is going to go through a very intense procedure like a stem cell transplant, you do want the donor cells to be in the best shape. If there is a mutation that increases risk of leukemia, you don’t want to potential use that for a patient who already isn't going to be immunocompromised because of the transplant and there's actually some data that does suggest that if the donor has any kind of these mutations, I'm not talking about the high-risk mutations, that there is delayed engraftment and potentially more low blood counts in the recipient of the stem cell transplant. I do think that is a field where this needs to be explored more and I personally would envision that such a strategy would be in use because these mutations are not that uncommon. If you look at people over 60 years old, you can find any mutation in about 10-15% of the general populations, so potentially 10% of your donors carry a mutation and you'd rather have somebody who does not.

Dr. Leonard: Finally, you emphasize that these were acquired mutations rather than inherited mutations. At this point, it seems that one wouldn’t expect this scenario of a patient with leukemia has family members, has children, that might be worried about their risk of leukemia, this would not be a strategy appropriate for them or would it be something that might potentially be useful?

Dr. Desai: That is correct. This is an acquired genetic abnormality or mutation so patients who have this or general elderly people who have this mutation does not mean that their children are going to have this. This is a normal process of aging so as people age, just statistically, there's a very high likelihood that some problem happens in cell divisions and a mutation develops and certain mutations will have a higher risk of AML while some don’t, but it is a normal process of aging and it doesn't mean that if one person has it their family members need to be immediately screened because that has no relevance to it.

Dr. Leonard: Before we wrap up, just a minute or two on where you're going from here. Where does this research lead us? We've touched on a number of different directions you and your colleagues could pursue, but what are the high priority steps from your perspective?

Dr. Desai: The first step is we need to establish a prospective monitoring strategy for patients or normal healthy people who have this mutation. You could perhaps focus it on people who have had previous chemotherapy because their risk of leukemia is higher. At some point with a lot of collaboration from multiple institutions, envision an intervention strategy, but that will require much more research on what intervention strategy and will it be successful. I think the first step is to actually do this prospectively since our study was based on a retrospective group of patients. The second thing is that there is also a lot of research that’s needed on people who have the mutation and don’t develop leukemia because something is different, not just the presence of the type of mutation, but there might be some other factors that somehow is protractive for that patient versus not. There's more research needed in that area as well to maximize the chances of us identifying which patient will and won't develop leukemia. Ultimately, it’s all about the absolute risk. If we know that someone is guaranteed to develop leukemia, then it’s easier to think that perhaps you can intervene, but if there's a 50-50 chance, you still would not want to give drugs that might have side effects in order to prevent a disease that may or may not happen.

Dr. Leonard: This has been a great discussion. I've learned a lot and I think it's a great example of the importance of patients participating in translational science. You start with a question around can you detect traces or risks of leukemia years ahead of time and then you end up with no many potentially important clinical questions and strategies that there's a lot of work to do, but a very big impact. I give you and your colleagues a lot of credit and really thank you for joining us today. It’s a great example of the translational work going on here at Weill Cornell. Thank you for joining us.

Dr. Desai: Thank you.

Dr. Leonard: I want to invite the audience to download, subscribe, rate and review Cancer Cast on Apple Podcast, Google Play music or online at weillcornell.org. We also encourage you to write to us at Cancer Cast at med.cornell.edu with questions, comments, and topics you'd like to see us cover more in-depth in the future. That's it for Cancer Cast, conversations about new developments in medicine, cancer care, and research. I'm Dr. John Leonard. Thanks for tuning in.