Harry P. Erba, MD, PhD: Let’s move on and let’s talk about myelofibrosis. Jamile, let me turn to you as our resident pathologist. Please talk about not only the signs or symptoms of myelofibrosis, but also how we make this diagnosis.

Jamile M. Shammo, MD, FASCP, FACP: I think like other MPNs, it should be asymptomatic, but obviously to a much lesser extent. About 25% of patients with MF may be totally asymptomatic. On the other hand, patients with MF are more likely to have splenomegaly, so that would probably be one cornerstone of their symptomatology. They will have single-agent symptoms. Most will either present with anemia or end up developing anemia. And then, some might complain of bone pain, night sweats, weight loss, or early satiety. But I think it’s so important that when we talk about symptoms and signs, a lot of it depends on how hard you are thinking when you’re talking to a patient. Often, they may not tell you that they have bone pain. Why should my hematologist know about my arthritis? But that’s why I think assessing symptomatology at the time of diagnosis is extremely important.

How is myelofibrosis diagnosed? Truly, the only way that you can confirm the diagnosis, or make it for that matter, is by performing a bone marrow biopsy. We can all have our clinical suspicion that perhaps this patient may have it, but the proof is in the pudding and in doing a bone marrow biopsy. Obviously, the WHO 2016 criteria outlined what is prerequisite for making that diagnosis, and that is a cluster of atypical megakaryocytes with 2 to 3 plus reticulin fibrosis; demonstration for hematopoiesis with JAK2, MPL, or CALR mutation; and a ruling out of many other entities, such as MDS with fibrosis or CML with fibrosis. You can safely say that this isn’t an alternate myeloid neoplasm. And then, of course, you have to demonstrate other minor criteria. I think there are 4 or 5 of them, such as a leukoblastic picture in the peripheral smear, high LDH, anemia, leukocytosis, and—as we talked about—splenomegaly and constitutional symptoms.

Ruben A. Mesa, MD, FACP: One factor here is that—I say this all the time for some reason—whenever there is a sample of fibrosis in a patient, they always say, “Oh, it must be myelofibrosis.” Fibrosis does not mean myelofibrosis as a disease entity. We have MDS with fibrosis that we see quite often in consultation. This is what we end up with when a patient is referred with myelofibrosis. There is some confusion about it, particularly for us, but it’s pretty simple: fibrosis does not equal myelofibrosis as a disease. It needs to be worked out.

Harry P. Erba, MD, PhD: The pathologic hallmark, as Jamile said, is this megakaryocytic morphology, the hyperplasia, the clustering, and the atypia. All of that is more important. In fact, in prefibrotic myelofibrosis, that’s what’s actually left: just the megakaryocytic changes. As opposed to PV, where we look at risk factors or risk stratification based on thrombotic events, there have been a number of risk stratifications in myelofibrosis that are related to mortality and overall survival that have evolved over the years. I think the first ones that I used were the international prognostic scoring system, Cervantes’ criteria.

Some of you were probably involved in that analysis of about a thousand patients with myelofibrosis, where 5 factors came out as being very important in determining prognosis: age over 65, constitutional symptoms, hemoglobin under 10 g/dl, white blood cell count over 25,000, and circulating blast above 1%. What’s not on the list? Splenomegaly’s not on the list, which is interesting. But if you have none of those risk factors, you have low-risk disease. That was only about 10% of the patients. Most of the patients have intermediate or high-risk disease. High-risk disease is 3 or more of those factors, and the median survival’s 2 years. With intermediate-2 disease, you have 2 risk factors, and that has a median survival of about 4 year. It’s about 8 years median survival, or 6 to 8 years, for intermediate-1 disease.

Now, there have been some modifications of that over time. The Cervantes criteria looked at criteria diagnosis. There was the dynamic international prognostic scoring system, which basically came up with the same 5 criteria, but it double-weighted the anemia as being very important. Finally, the most recent iteration that I know of is the revised dynamic international prognostic score, which added thrombocytopenia, cytogenetic changes, and transfusion dependence. But I think the important point here is that the majority of patients we see in clinic are patients with intermediate to high-risk disease, and those patients have quite a limited life expectancy and need to be managed appropriately.

Ruben A. Mesa, MD, FACP: As I share with my trainees, one, there are more prognostic scores than we can keep track of, and I think that they’re important in terms of what they inform us about regarding the disease. There’s even one more iteration on this theme of prognostic scores involving some of the new molecular mutations.

Harry P. Erba, MD, PhD: That was going to be my question to you.

Ruben A. Mesa, MD, FACP: Presented at the ASH meeting, Professor Vannucchi—he and his group have been very active—found that a marriage of some of these clinical features and some of these molecular features has been very helpful. They did an analysis in transplant-age patients, under the age of 70, who they call the MIPSS70. What I show trainees is that they are confronted with prognostic scores from colon cancer all the way through MF. Don’t memorize any of them, but know where to find them when you need to. Be mindful of the lessons that they tell us, but also know that they have limitations.

One of the limitations with these is, why is splenomegaly not in there? In large part it is because most of the retrospective data are not really good quality data, regarding the size of the spleen, to be able to make good conclusions.

I had a patient who had ascites, and it was filled with circulating megakaryocytes in the ascites. That clearly was a huge red flag for me. That’s not on any of the prognostic scores, so it’s always important to have a part of common sense. Without question, it certainly has advanced the discussion in heterogeneous disease along with a good amount of clinical common sense. The patients you’re really worried about are progressing along that myelofibrosis phenotype with worsening debilitation, big spleen, and death from the consequence of that; or really, they’re heading toward AML. Those are 2 different phenotypes with different implications in terms of therapy.

Srdan Verstovsek, MD, PhD: The other thing is, why are we prognosticating? We’re prognosticating to select patients.

Harry P. Erba, MD, PhD: I think it’s to publish papers.

Srdan Verstovsek, MD, PhD: That’s one reason, maybe. But why is it practical, right? Practice. You’re prognosticating to select patients who you would refer to transplant. Then the patient is seen by the doctor who does the transplant, and he will look at other factors in addition to your prognostic scoring system. Comorbidities. If you have a heart problem, if you have pulmonary hypertension, or if you have some other kidney problem, you may be qualified to go for a transplant based on our IPSS or revised IPSS or many others. You are not going to make it to the transplant because you have comorbidities. There are other factors that need to be accounted for in that transplant setting other than what we are trying to do in our own practice. We are improving. Let’s say that there are so many of them—we have lots of papers, even lots of online systems to calculate risk—and none of them are perfect. It has to be individualized at the end.

Rami Komrokji, MD: I’d like to add to that clinical point. Even with MDS, some of those models would incorporate things like age. I always say when making a decision, there is prognostic information from the models where you can tell a patient that the expected survival for 100 patients is X. But then, we are also using them as the therapeutic tools to decide who we are taking to transplant. If I label somebody as higher risk just by the fact that they are above the age of 65, I give them a point that takes them from intermediate-1 to intermediate-2 disease. I’m not sure that I’m going to send that patient to transplant. I look at disease-related risk factors to decide on transplant, for example. But from a prognostic standpoint, there is no doubt that somebody who’s older will have a worse outcome, unfortunately, than young patients.

The other thing that adds to the complexity are those somatic mutations. We know that certain mutations that several groups have reported on have independent prognostic value, like the presence of ESL-1 or SRSF2 in myelofibrosis or what we call triple-negative disease now, the absence of the phenotypic driver mutations. I think some of the mutations may predict a higher chance of progression to AML. Yet we don’t know; I think we are starting to learn about what the impact of something like transplant is on those mutations. Are we doing a patient a favor? If they’ve already had a mutation and we send them to transplant, does the transplant overcome that or not? I think we are starting to get that information, but we are not yet there. Sometimes I struggle with that. I upstaged the patient, now what do I do with that?

Transcript Edited for Clarity