Verstovsek Outlines Features of and Therapies for Myeloproliferative Neoplasms

July 30, 2019
Audrey Sternberg

Srdan Verstovsek, MD, PhD, discussed the treatment decisions he makes when treating a patient with a myeloproliferative neoplasm.

Srdan Verstovsek, MD, PhD

In aTargeted Oncologycase-based peer perspectives dinner presentation, Srdan Verstovsek, MD, PhD, discussed with a group of physicians the considerations he makes with the diagnosis and management of patients with myeloproliferative neoplasms (MPNs). Verstovsek, chief of the Section of Myeloproliferative Neoplasms and director of the Hanns A. Pielenz Clinical Research Center for Myeloproliferative Neoplasia at The University of Texas MD Anderson Cancer Center, explained his treatment decisions based on a case scenario of a patient with myelofibrosis.


January 2016

A 59-year-old man presented to his physician with symptoms of fatigue, night sweats, weight loss, and abdominal pain with early satiety past 6 months. On physical exam, his spleen was palpable 10 cm below left costal margin.Genetic testing revealed aJAK2V617F mutation and anASXL1mutation. A bone marrow biopsy showed megakaryocyte proliferation and atypia with evidence of reticulin fibrosis. A blood smear also revealed leukoerythroblastosis.His laboratory results were notable for red blood cells of 3.20 × 1012/L, hemoglobin of 9.7 g/dL, hematocrit of 34%, mean corpuscular volume of 93.1 fL, white blood cells of 12.1 ×109/L, variable platelets of 247 ×109/L, and peripheral blood blasts of 0%.

Targeted Oncology:What are your initial impressions of the case? Is it standard practice to have next-generation sequencing (NGS) for the purposes of detecting other mutations besides the driver, as in theAXL1mutations?

VERSTOVSEK:This is a typical case of myelofibrosis. He has fibrosis, he has aJAK2mutation, he has an enlarged spleen, and he is anemic. Testing the allele burden for driver mutations is part of the diagnostic process. Driver mutations are mandatory for diagnosis. Many institutions use a panel to diagnose muta­tions. The driver mutation would be in the JAK pathway; that is the underlying biological problem.

Usually, I am surrounded by colleagues from academic centers, and they are performing clinical studies. They always do NGS as part of the study, and they would always want to do that because that can affect eligibility for new drugs. They want to know the complexity of the person who is on the study. We do this as a standard practice; everybody who comes in the door gets NGS. You get a bone marrow biopsy, a cytogenetics panel, and NGS to see the genetic complexity of the patient, which includes the driver mutation. But that is not standard practice otherwise.

If you repeat the molecular testing, the lab is going to ask why. You have to justify it or else it is not going to be paid [for by the patient’s insurance]. It has to be justifiable, such as the patient is failing therapy and you want to know what is wrong. You cannot just do these tests every 6 months. It is very expen­sive. We would like to do that, but we cannot.

The new prognostic scoring system incorporates genetic test­ing into the algorithms to see who should be referred to [receive] transplant sooner rather than later. The standard way of look­ing at prognosis is by the patient’s age, symptoms, anemia, CBC [complete blood count], and blasts. Now, many academic centers do genetic testing, and many of the mutations—like theIDHmutation,EZH2,ASXL1,SRSF2—should be accounted for when you prognosticate to refer patients for a transplant. However, this is expensive, and it is not readily available for everybody. Not many people get to the transplant, so what is the value in doing extra testing?

You do see allele burden reported in clinical trials. Why is it relevant there?

The allele burden is how many cells in a sample have the gene mutation. We do not do that routinely in our academic centers because for our purposes, we do not have a real value for it yet. For genetic testing with the NGS panels, we can identify prog­nostic mutations that are leading to perhaps sooner referrals for transplant. But with allele burden, we do not have a real value.

In clinical trials, they do as much [testing] as possible. Investigators will do NGS and the allele burden to see if it is going down—I cannot say if the allele burden going down means anything. We want to have a drug that has biological activity to say that there is a possible change to the outcome of the patient by changing the natural history of the disease. That is why you will see theJAK2allele burden or the allele burden of any mutation, the cytogenetics, bone marrow fibrosis, and modifications to the parameters that are then supposed to tell us that whatever drug is being tested is changing something about the biology of the disease. You have to be careful about assign­ing any clinical activity to changes in the biological parameters.

A case in point is in the interferon studies where there is a significant decrease in theJAK2allele burden, but it does not correlate with a hematological response, a bone marrow response, or preventing progression.1This is all investigational at the moment; I do not use this in clinical practice.

What are important diagnostic parameters of myelofibrosis?

It is not easy to make a diagnosis of myelofibrosis. We looked at 560 patients and asked, “Is the referring diagnosis correct?” There was discordance between the referring diagnosis and our diagnosis in about 12.5% of patients who were misdiag­nosed.2The main problem is that when you do not have aJAK2mutation, you do not have [circulating] blasts and you do not have fibrosis. These are the 3 parameters that are skewing the diagnosis.

Everybody is aware of early-stage myelofibrosis, or prefibrotic myelofibrosis, which is a new entity since 2016 and is similar to ET [essential thrombocythemia]. The issue is, you have a diagnosis of prefibrotic myelofibrosis; there is no fibrosis or leukoerythroblastosis—these are things that should not be there by definition of diagnosis. They may have slight anemia, slightly elevated white blood cells, but no leukoerythro­blastosis and no fibrosis. These are the 2 parameters that are different between prefibrotic myelofibrosis and the standard, classical myelofibrosis.3

The prognosis is excellent; it is similar to ET. How did we come to recognize prefibrotic myelofibrosis? It was carved out of ET. Patients with ET had different sizes, shapes, and colors of mega­karyocytes, and they were followed retrospectively. There was a [4% to 11%] difference over 15 years of transformation to post-ET myelofibrosis and a [2.1% to 5.3%] difference over 15 years to get to post-ET acute myeloid leukemia.4We are talking about some change down the road after 15 years of small changes between ET and prefibrotic myelofibrosis. These patients have good outcomes, they are not dying, and they should not be referred to transplant. We manage these patients as having ET. You look at the thromboembolic risk like you do in ET and you follow the patient. The 2016 WHO [World Health Organization] criteria were changed to include prefibrotic myelofibrosis.3

Apart from prefibrotic myelofibrosis, you have an issue with MDS [myelodysplastic syndrome] with fibers [or] MDS with fibro­sis. You can dissect this by doing molecular testing beyond theJAK2mutation, because there is evolution from myelofibrosis or early stages of MPNs to MDS by opposition of an MDS type of mutation, a splicing gene muta­tion, that would be moving toward MDS.

Sometimes MDS is just a result of the therapy itself. Hydroxyurea [HU] can change how the cells look in the bone marrow. My hematopathologist says, “It looks like MPN, but there is dysplasia in all the lines; maybe it’s because of 10 years of HU.”

How do you treat patients with MDS/MPN? What is the best therapy?

These patients may or may not have an enlarged spleen, they may have a lot of symptoms, they are usually pancytopenic, and they typically have a complex genetic profile. The best study…so far for these overlapping syndromes is the combination of Jakafi [ruxolitinib] and Vidaza [azacitidine] that came out last year in theAmerican Journal of Hematology.5That regimen is expensive and not approved, so how are you going to use that?

In practice, we usually recommend that people use azacitidine because the MDS part [of the disease] is usually more problem­atic. You cannot use the optimal dose of ruxolitinib because the numbers are so low. The primary problems are typically pancyto­penia, anemia, and low platelets. Azacitadine would be standard practice even in our hands if we do not have a study that would support 2 drugs together because of expense.

If you have overlapped syndromes where MDS is prominent and the pathologist says it is MDS/MPN, we would be using as a standard practice more hypomethylating agents. There have been studies with azacitidine in myelofibrosis, and they did not work [because of] a short response and no value. The only value to extend that discussion is where there is the value for hypo­methylating agents in MPN, as in those patients with increased blasts. We now have accelerated phase [characterized by the presence of] 10% to 19% blasts, and ≥20% is blast phase.6If their blasts go up, then we would be adding hypomethylating agents to ruxolitinib; we would not stop. Ruxolitinib would then stay at a flat [dose], because azacitidine would be lowering the counts as it would be lowering the blasts. You set your dose of ruxoli­tinib, you do not modify it, and in that case, you would have a 50/50 chance to decrease the blasts, which would give you the opportunity to send the patient to transplant. If there is no trans­plant, then just keep going as long as you can, because there is not much you can do once they start going through the accel­erated phase [of the disease]. They will end up in blast phase sooner or later.

If you look up NCCN [National Comprehensive Cancer Network] guidelines, it is written that hypomethylating agents would be the first choice for accelerated or blast phase with or with­out ruxolitinib.6Otherwise, in the chronic phase, there is not much news.

What risk assessment models are available for prognosis?

In academic settings, we may do genetic testing and combine these with the standard age, symptoms, white blood cell count, blasts, and anemia. These new prognostic models, the MIPPS70+ [mutation-enhanced international prognostic scoring system for transplant-age patients] version 2.0 and the GIPSS [genetically inspired prognostic scoring system] are described in the TABLE.7

How does this affect the choice of therapy?

Ruxolitinib is the one that we use often in people with splenic symptoms. Genetics on its own does affect the use of ruxoli­tinib, and the number of genetic abnormalities, if there are 3 or more on whichever panel you choose, will typically shorten the duration of benefit compared with patients who are genetically simple. Clonal evolution of therapy is present in about a third of patients when you compare the initiation of therapy to when it is stopped. Genetics has tremendous value, but it is a matter of cost and what you can do about it.

What is the utility of ruxolitinib in patients like this one with myelofibrosis?

We all know that ruxolitinib can shrink the spleen; that is some­thing that you can achieve regularly. It is a matter of the dose and how you do it. Sometimes the body weight or the albumin goes up [as a result of therapy]. Some people gain too much weight, 20 to 30 pounds, and they complain about it. The body compo­sition changes.

The label of ruxolitinib changed in July 2014 to say that it prolongs overall survival [OS] because the [COMFORT-I and COMFORT-II trials] showed a difference in survival between patients treated with ruxolitinib versus those who were not.8 The concern about that claim was due to the fact that the stud­ies were not designed to look at survival itself.

How do determine which dose to give the patient?

If you start with a low dose, you can administer the agent twice per day. The patient will feel good, but the spleen does not do as well. If you go up from 5 mg all the way to 20 mg as much as you can early on in the administration, you can get the spleen to respond and improve OS. If you can increase the dose every month, that is optimal. If you start low, go up to get the spleen under control. In these clinical trials, the [size of the] spleen correlates with OS.


  1. Kuriakose E, Vandris K, Wang YL, et al. Decrease inJAK2V617F allele burden is not a prerequisite to clinical response in patients with polycythemia vera. Haematologica. 2012;97(4):538-542. doi: 10.3324/haematol.2011.053348.
  2. Yi CA, Jeyakumar G, Medina P, et al. Discrepancy in diagnosis of primary myelofi­brosis between referral and tertiary care centers.Leuk Res.2014;38(1):91-94. doi: 10.1016/j.leukres.2013.11.002.
  3. Guglielmelli P, Pacilli A, Rotunno G, et al; AGIMM Group. Presentation and outcome of patients with 2016 WHO diagnosis of prefibrotic and overt primary myelofibrosis.Blood.2017;129(24):3227-3236. doi: 10.1182/blood-2017-01-761999.
  4. Cerquozzi S, Tefferi A. Blast transformation and fibrotic progression in polycy­themia vera and essential thrombocythemia: a literature review of incidence and risk factors.Blood Cancer J.2015;5(11):e366. doi: 10.1038/bcj.2015.95.
  5. Assi R, Kantarjian HM, Garcia-Manero G, et al. A phase II trial of ruxolitinib in combination with azacytidine in myelodysplastic syndrome/myeloproliferative neoplasms.Am J Hematol.2018;93(2):277-285. doi: 10.1002/ajh.24972.
  6. NCCN Clinical Practice Guidelines in Oncology. Myeloproliferative Neoplasms, version 2.2019. National Comprehensive Cancer Network website.­sionals/physician_gls/pdf/mpn.pdf. Published October 29, 2018. Accessed June 26, 2019.
  7. Tefferi A, Guglielmelli P, Pardanani A, Vannucchi AM. Myelofibrosis treatment algo­rithm 2018.Blood Cancer J. 2018;8(8):72. doi: 10.1038/s41408-018-0109-0.
  8. Jakafi [package insert]. Wilmington, DE: Incyte Corporation; 2011. www.accessdata. Accessed June 26, 2019.