MDS: The Current Understanding in Molecular Testing and Biomarkers

Transcript:James M. Foran, MD, FRCPC: I wanted to finish just by asking you about the role of molecular medicine; in particular, deep sequencing or next-generation sequencing in patients with MDS. Is that something that you’re performing or is that something you’re acting upon? How do you incorporate that in your care of your patients?

Azra Raza, MD: So, I want to answer that by taking you through really what happens to a patient who is now an older individual, around 70 years of age or older, and suddenly presents with a cytopenia. One of the first things that you think about after you’ve ruled out all the other nutritional, and drug-related, and infection-related causes, is then you start to worry that could this be a primary bone marrow disorder. You go ahead and do a bone marrow, and the bone marrow is unremarkable. I mean, it doesn’t show; it doesn’t meet the minimal diagnostic criteria for MDS, and that minimal diagnostic criteria is at least 10% cells have to look dysplastic, or there has to be a cytogenetic abnormality, or the blast count has to be between 5% and 19%. That’s the minimum. Patient’s bone marrow doesn’t meet any of those, so now we label the patient as having ICUS, idiopathic cytopenia of undetermined significance.

Now, in this day and age of molecular profiling, we go ahead and send a test for genetic profiling. It comes back as showing, let’s say, an ASXL1 mutation in 32% of the cells. Now, what do you call it? Is it MDS now? No, of course not, because the MDS minimal criteria are not met. But the patient’s disease gets upgraded from ICUS to CHIP, and CHIP stands for clonal hematopoiesis of indeterminate potential. So, now you know there is a clone of cells because of the presence of this mutation, but yet you can’t call it MDS, which I find surprising because if there was a cytogenetic abnormality, we could call it MDS. But because there’s a genetic abnormality, we don’t call it.

Anyway, so now we have ICUS to CHIP, and then what is the potential of this to develop into MDS? Three thousand patients followed with ICUS at Mayo Clinic, 60% of them ended up developing MDS, 20% developed a non-MDS—related neoplasm, and the other 20%, really nothing. So, it’s not a good thing to have ICUS, because 80% will end up with either MDS or another malignancy. Having CHIP, what does that mean for developing MDS? We really don’t know right now. Now, let’s say you have a patient who’s fully diagnosed with MDS, and you do a molecular profile, and now you are looking for the 40 or 50 or 100 most commonly myeloid disease–related genes, and you find one, two, three, or 10 mutations. What do you do with this information? Does it affect your treatment?

So, in 2011, we published a paper in the New England Journal of Medicine, which showed that if there is any mutation in one of five genes (p53, EZH2, ASXL1, ETV6, and RUNX1), then survival is actually compromised. However, with time, we have learned more that, for example, if there is a patient who has ringed sideroblasts but who now has an SF3B1 mutation, then that survival of that patient is likely to be better than having ringed sideroblasts and no SF3B1 mutation. Then we know that other splicing factors like SRSF2 or U2AF1, these are related with a compromised survival no matter what type of MDS they have. So, those are important. Then we have certain targetable mutations. For example, 5% to 10% patients with MDS can have a mutation in IDH1 or IDH2, and for those, we have oral inhibitors available. For those patients, in last 6 months, James, I must have had 10 patients diagnosed, who were showing IDH1 or IDH2 mutations that we were able to treat with the drug.

I think that’s an important thing. Then those patients who have isolated TET2 mutations are likely to respond better to hypomethylating agents. Their response rate is 85% compared to 50% for the rest of the population. So, I think p53 is the most important. Most commonly, the p53 mutation is present in association with a complex cytogenetic. If multiple chromosomes are damaged, p53 is mutated, median survival is 6 months. Everyone is dead by 2 years in our study. That is not a good combination to have. On the other hand, if a patient has a multiply damaged chromosome but not p53 mutation, their median survival is over 3 years, and many of them are living up to 8 years and 9 years.

In other words, genetic trumps cytogenetics here, because p53 seems to be much more important. So, my answer to your question is about genetic profiling and its role in MDS, that MDS is not one disease. We go from ICUS to CHIP to MDS, which is all these different types, and all types can have one or more of the mutations. It’s a very important time for us to follow these mutations to measure their clonal velocity. At what rate is each mutation likely to progress, either in isolation or in association with other mutations? All these data need to be collected and catalogued over the next few years so we can start incorporating it into developing more accurate prognostic classification systems and as guides for future therapies.

James M. Foran, MD, FRCPC: That’s an excellent answer, and we’ve had a more cautious approach to them. We know that some mutations, like ASXL1, may predict for a shorter response duration with azacitidine, for instance. But it doesn’t stop us from using azacitidine. We still think you do better on it than off it, so I would caution other practitioners not to use it. Sometimes it will guide therapy. Sometimes it will move you more to a transplant strategy. I think it teaches us a lot about the pathogenesis or pathobiology of the disease. But how to exploit it in individual patients will require better therapies and more therapies. And I like that you’re using it to try to identify the small subset. It could be targeted with inhibitors for that small subset with IDH1 or IDH2 mutations, as example.

Azra Raza, MD: I did want to say one other thing, James, which I forgot to say. The importance of p53 mutation at the other end of the spectrum. So, I only talked about, so far, the importance of this mutation in the presence of a complex karyotype. At the other end of the spectrum are patients who present with an isolated deletion 5q. These patients are supposed to have, generally, a more benign disease, low incidence of transformation to acute myeloid leukemia. Yet, we find that a good 18% of them can have a p53 mutation. And if that mutation is present, then their chance of rapid transformation to AML is high. Interestingly, not the response to lenalidomide, which is the same as non-p53—mutated, but it doesn’t last that long and can progress to acute leukemia. And for that patient, you need to consider a transplant option earlier if they are good candidates.

So, what I am saying is, when I say that genetics trumps cytogenetics, I’m saying that that’s true for both ends of the spectrum, for a disease that looks dreadful with multiply damaged chromosome and for a disease that you could be lulled into a complacency and tell the patient everything is hunky-dory, come back in 6 months. No, because if with that isolated deletion 5q they have a p53 mutation, you may have to start planning for alternate therapies.

James M. Foran, MD, FRCPC: It’s a tricky area. You have these abbreviations of ICUS and CHIP, and then we get to MDS. And then there’s another one called ARCH, the age-related clonal hematopoiesis, in well people with very low variance allele frequency mutations that predicts a high risk for disease, and yet most people still don’t get disease. Until we have enough information on how to really act upon these and more targeted therapies, I think it has to be incorporated into the standardized diagnostic testing and algorithms.

Transcript Edited for Clarity