Ruben A. Mesa, MD, FACP: Elias, as we think about prognostication, there's some new areas that have been raised as possible new markers. In fact, I'm sure in large meetings such as ASH, we'll have a variety of candidate-type new prognostic ways that get vetted over time. One of these was presented at ASCO where they were actually looking at the leukemia stem cell phenotype as potentially being another factor.

They looked at three different components of the stem cell: the positivity or negativity of the CD34 in terms of the blast. A second population was CD34-positive, CD38-negative, and then ALDH-intermediate. This was the most common subtype in that group. And another group that was CD34-positive, CD38-negative, and ALDH-high. And they identified that that group that had that ALDH-high, CD34-positive, CD38-negative, really had a worse outcome. They had more association with complex cytogenetics and certainly thought that it might have some implications in terms of therapy. How do you view looking at the stem cell as a possible additional way to risk-stratify patients?

Elias Jabbour, MD: Well, definitely it’s a continuum of the spectrum. So, we're identifying a group of patients who are doing poorly from the beginning. So, you have early stage, a population destined to do poorly like in this abstract report where they have what they called high expression of aldehyde dehydrogenase. These are correlated with bad features. The question we have today: how do we treat these patients? What is the implication of identifying poor players?

Rafael mentioned a beautiful paper, and Rami quoted the p53 mutation in MDS. We say, well, if you're somebody with a bad feature, go for transplant. They still do poorly. So, these are all great to stratify our patients, but they have not yet changed our practice. I think what we need to do today, at least for the colleagues in the community, if you identify such a patient, refer them to clinical trials in order to improve their outcome. We are refining all our knowledge thanks to the biology and to the work done by colleagues like Rafael and others. But we're still not doing so well in treatment.

Ruben A. Mesa, MD, FACP: I have to agree so strongly. Our pediatric colleagues have really been very successful in creating a culture where the standard of care is the clinical trial and the percentage of pediatric patients who go on clinical trials is the vast majority. And clearly, there’s this discussion as to what that trial should be but that there clearly is a need to push that forward; they really have made some strides. In acute leukemia, I think we still struggle that we still have patients really not being treated as part of a clinical trial. There's so much for us to learn about the disease.

Rami Komrokji, MD: Right. I think looking at the stem cell compartment could be interesting in two ways. Obviously, from a therapeutic point, that would be an ideal way if we can differentiate leukemia stem cells from normal stem cells, if we can identify certain receptors or targets in the leukemic stem cells where we can eradicate those, saving their normal stem cells. Obviously we have things now that are very exciting, like with the CAR-T therapy, if you can identify the stem cells in leukemia and develop therapy for that. So, it had a huge therapeutic implication, if we can identify the stem cells and in those diseases.

The other thing could be also in terms of monitoring of the disease because many times the mutations or the cytogenetic abnormalities are harbored in those stem cells. With our simple way of looking at cytogenetics, now we really don't look at the stem cells; we are looking at probably the common myeloid progenitor compartment. And in papers in MDS, for example, you show eradication of normal sites, complete cytogenetic response that sometimes does not correlate with outcome. But if you do a sorting technique and you look at the stem cells, the stem cells still harbor the mutation. So, it could have a therapeutic implication and a monitoring implication paying attention to the stem cell compartment.

Ruben A. Mesa, MD, FACP: So, before we transition over to treatment, this has been a great discussion around biology. Rafael, your group has certainly been involved at various points in your career, really helping to distinguish that issue of progression. As you think about molecular features in a de novo AML patient and someone who potentially has progressed from MDS, how might those two be different?

Rafael Bejar MD, PhD: Well, there was a really interesting paper last year by Coleman Lindsley that examined this question. They took patients who had clear evidence of MDS or another myelo disorder who progressed to secondary AML, sequenced them to figure out what types of mutations these patients have. And, interestingly, they were many of the same mutations that you find in myelodysplastic syndromes and myeloproliferative neoplasms, things like splicing-factor mutations, ASXL1 and so on. Then, he took another population of younger AML patients and found out their mutational profile was slightly different. They had more NPM1 mutations and more things like that.

The important piece is that when we have older patients with AML, even if they didn’t have a preexisting diagnosis of a myelodysplastic syndrome, many of them have that MDS-like mutational pattern or phenotype and those are the patients that don’t do as well. And it may be one of the reasons that older AML patients have inferior outcomes. But the point is that there are older patients that don’t have those features and they may actually be a better risk than we suspect. So, identifying that population and actually thinking about the possibility of curing them with standard treatment, if they can tolerate it, has really identified a group of patients that may actually benefit from the types of treatments that we typically feel are probably less efficacious in the elderly.

Elias Jabbour, MD: I have a question: in practice today, we see the referral patterns and sometimes we see physicians ordering a bunch of molecular features. I’m not one to mention names, but there’s plenty of pharma groups or whatever that promote these tests or people don’t do anything. Patients will come to you even just with a blood sample or the CBC. Practically speaking, the classification of FLT3-ITD and PM1 mutations, C/EBP-alpha and others, are more of a concern for diploid karyotype if I have somebody in other terms with a very bad karyotype. What is the weight of the other mutations? Do we need to do them especially if I don’t have a treatment to implement?

Rafael Bejar MD, PhD: That’s a good point. So, what is the clinical utility of knowing about these mutations?

Elias Jabbour, MD: Correct.

Rafael Bejar MD, PhD: Even in those patients who have bad risk features, these mutations can still have prognostic significance. I’ll give you an example: patients who have a complex karyotype don’t all have mutations of TP53. Those patients that have that mutation really have the worst outcomes. The complex karyotype patients who lack that p53 mutation may do much better than you would have predicted based on that karyotype alone. So, you may be able to even downstage their prognostic risk to some extent.

Rami Komrokji, MD: I think you could also look at the good risk group in AML, for example, chromosome 16 or the translocation 8;21. If those patients have the second mutation, that actually brings them to an intermediate risk group. So, you could look at the two extremes.

Rafael Bejar MD, PhD: The mutation is going to define prognosis.

Rami Komrokji, MD: Right, exactly.

Rafael Bejar MD, PhD: That’s built in our existing system.

Transcript Edited for Clarity