Stephen T. Oh, MD, PhD: As far as risk assessment criteria for myelofibrosis [MF], there are a variety of prognostic scoring schemes that have been developed and utilized over the years. These range from the DIPSS [Dynamic International Prognostic Scoring System] or DIPSS Plus, to more recently, the MIPSS70 [Mutation-Enhanced International Prognostic Scoring System 70] or MIPSS70 Plus. There are a variety of different schemes that can be used, but they all share a number of prognostic factors, including anemia. We know that anemia is a significant adverse risk factor, thrombocytopenia as well; presence of blasts, increasing blasts in the peripheral blood is an adverse factor. With the more recent iterations of these stratification schemes, there is the incorporation of molecular markers or so-called molecular high-risk mutations that are associated with worse overall survival and increased risk of transformation to AML [acute myeloid leukemia].

Pankit Vachhani, MD: After the diagnosis of myelofibrosis, I think one of the most important things to do for any patient is appropriate risk stratification. While we don’t stage our patients, we definitely do risk stratification for our patients. This is important for a variety of things. It is important for treatment selection in the upfront setting, but also it helps guide whether someone should be sent for transplant early or later in their disease course. In fact, risk stratification is also considered by the NCCN [National Comprehensive Cancer Network] guidelines as one of the key important points in decision-making. Now, in general, the risk stratification points include items like age, anemia, leukocytosis, thrombocytopenia, circulating blasts, bone marrow fibrosis, symptoms, as well as the need for transfusions. But if we combine all the various risk stratifications, primary myelofibrosis has an overall median survival of about 6 years. This is one of the least favorable survival outcomes of all the classical Philadelphia chromosome-negative myeloproliferative neoplasms. But what the 6-year median survival does not capture is the differences in median survival for different risk categories. For example, by DIPSS, or Dynamic International Prognostic Scoring System classification, patients with intermediate-2 risk myelofibrosis have a median survival of 4 years, while high-risk patients have about 1 and a half years. Now, DIPSS and IPSS [International Prognostic Scoring System] and many of the other risk stratification schemas were developed for patients with primary myelofibrosis.

There’s also the MYSEC-PM [Myelofibrosis Secondary to PV and ET-Prognostic Model] risk categorization scoring system, which is something that was started especially for patients who have post-polycythemia vera [PV] myelofibrosis or post ET [essential thrombocythemia] myelofibrosis. I would definitely encourage using that for patients who fall into that category. I think this is also important to note because the survival outcomes are different between primary myelofibrosis versus post-PV or post-ET myelofibrosis. As an example, one institution’s data showed that patients with post-ET myelofibrosis had a median survival of 73 months, while that for primary myelofibrosis or post-polycythemia vera myelofibrosis was much less, it was 45 months and 48 months. Clearly, there is a distinction in the survival outcomes.

I would also encourage that for every patient with myelofibrosis we perform a next-generation sequencing [NGS] panel that looks into the more common myeloid gene mutations. This has been known for almost 10, 11 years at this point, that patients who have mutations in ASXL1, EZH2, IDH1 or IDH2, SRSF2, or U2AF1, they tend to have a higher risk for premature death or leukemic transformation. Thus, checking that NGS panel in every patient is important. One could use a molecularly enhanced prognostic scoring system or use one of the more commonly used DIPSS systems, but take into account the genetics as well of any individual patient. Last but not least, I do want to state that the blast phase of myelofibrosis, corresponding to 20% or more peripheral or bone marrow blasts, patients who fall into that category have one of the worst outcomes of all patients with myelofibrosis. Those patients are treated with various AML-like regimens, some of which may incorporate a JAK1 or JAK2 inhibitor, but the prognosis there is at present somewhere around 2 and a half to 6 months. Thus, an individual decision on the best time to transplant, who to transplant, and what treatment to deliver is very important there.

Ruben Mesa, MD: Genetic information from patients with myelofibrosis is very important in terms of us understanding the prognosis of the disease, both the risk of mortality, as well as the risk of progression to acute leukemia. Either one of which can be very relevant, in that either one can be catastrophic if they occur. Historically, this was first cytogenetics. Long before we had molecular testing, we would do chromosome analysis that could be done on the bone marrow or the peripheral blood, with many of those chromosomal changes that had adverse prognostic features mirroring those that we see that have adverse prognostic features with other myeloid disorders: complex cytogenetics, multiple mutations, mutations in chromosome 1, among others that are “high risk.” Additionally, we always obtained this information because we were also trying to exclude the presence of an alternative myeloid disorder like chronic myeloid leukemia. The advent of additional somatic mutations I think has helped us to refine prognosis to a much greater degree and is very complementary, some would say preferable, to cytogenetics. However, if one can obtain both, great. I just ordered a bone marrow from a patient I saw in the clinic right before we started this discussion. For this individual with MF, we will, as a baseline, obtain both cytogenetics and the next-generation sequencing myeloid panel because the presence or absence of certain mutations have an adverse prognostic significance. First, there are the driver mutations of JAK2, CALR, and MPL. CALR type 1 mutations may have a more favorable prognosis.

Likewise, the absence of any of those 3 driver mutations, or being triple-negative, may have a more adverse prognosis as they plug in these values into the MIPSS70 or the subsequent iterations of that prognostic score. Additional somatic mutations that are found on panels, depending upon the vendor that you use, are anywhere from 40 to 80 genes of potentially other mutated myeloid genes, but ones that have been particularly prognostically adverse are ASXL1, EZH1/2, multiple additional somatic mutations. TP53 can have individual prognostic significance that gets plugged into these prognostic scores that can predict a shorter survival. At our center, typically, if we’re getting the bone marrow, we’ll get cytogenetics and NGS. If a patient doesn’t have a bone marrow, we can obtain NGS on peripheral blood. I will repeat it at a certain frequency, in particular in the individual for which we’re still considering stem cell transplantation as a therapy in the future if there is disease progression. When we think about the adverse prognostic implications in MF and how that might accelerate your decision-making toward definitive intervention like transplantation, NGS can be helpful. If it’s a younger patient for whom we’re on the fence about whether we do a transplant, I may repeat that NGS every year, every 2 years. Most certainly [we would perform NGS] at the time there’s any evidence of clinical progression with change in the counts, or if we were to initiate a new therapy or a clinical trial.

Transcript edited for clarity.