The Value of Prognostication in the Modern Management of Myelofibrosis

May 13, 2020

In an interview with Targeted Oncology, Verstovsek, medical oncologist and professor, Department of Leukemia, University of Texas MD Anderson Cancer Center, discussed mutational profiling for myelofibrosis and how the information obtained can be valuable for clinical management of the disease.

The management of myelofibrosis (MF), like most hematologic malignancies, is deviating towards the use of molecular information to guide treatment decisions. MF management has not reached the point in which therapies are chosen based on information obtained from molecular profiling, but with more research pending, the field is headed in that direction.

Currently, hematologic oncologists are testing for driver mutations, which are predominantly JAK2 V617F, MPL515, or CALR, as well as non-driver mutations. These each provide different signals as to whether a patient requires transplantation immediately or if it can be delayed, in the context of other prognostic factors.

In a paper published in the International Journal of Hematology, Srdan Verstovsek, MD, PhD, and Prithviraj Bose, MD highlighted that understanding molercular complexity may be helpful before treatment because it may be informative for deciding if a patient should receive ruxolitinib (Jakafi), the current standard of care. Patients with MF may develop resistance to ruxolitinib, and some studies have shown high rates of treatment discontinuation because of this resistance.

Non-driver mutations like ASXL1EZH2, and SRSF2, were found to be indicative of shorter survival periods, and mutations like ASXL1SRSF2IDH1, and IDH2 were found to be associated with leukemic formations.

In an interview with Targeted Oncology, Verstovsek, medical oncologist and professor, Department of Leukemia, University of Texas MD Anderson Cancer Center, discussed mutational profiling for MF and how the information obtained can be valuable for clinical management of the disease.

TARGETED ONCOLOGY: Can you first provide background on mutational profiling in myeloproliferative neoplasm (MPNs)?

Verstovsek: One can say that mutational profiling started with the discovery of JAK1 V617S mutation as the mutation associated with MPNs in 2005. This led us to recognize that there are mutations associated with hyperactivity of the JAK/STAT pathway, which is an underlying biological abnormality in all types of MPNs.

When the JAK2 V617F mutation was identified in about 60% of patients with MF specifically, then the question was, what else is responsible for the other 40% of patients that have a hyperactive JAK/STAT pathway, what other mutations. This question led to the discovery of the MPL mutation, which is the mutation in the thrombopoietin receptor. It also led to the discovery of the calreticulin mutation that fills up almost all of the other 40% of cases that are negative for JAK2 V617F mutations. However, there are still 10% of patients that have no JAK2MPL, or calreticulin mutation. This was the beginning of profiling.

Along the way, we recognized that there are many other mutations that are not associated with the JAK/STAT pathway. We call them non-driver mutations. If you do genetic profiling, you have the ability to do a wide range of testing of patients with driver mutation to see which 1 of the 3 types of mutations is present. You can then also test 40, 80, or even 400 other genes and you will find that many patients have additional non-driver mutations, and they may have significance for the outcome of certain therapies or the survival dis-advantage for certain patients. Genetic testing has an important role.

Today, when we talk about profiling, we go beyond just the driver mutations.

TARGETED ONCOLOGY: What are the clinical and prognostic implications of driver mutations?

Verstovsek: For diagnostic purposes, 1 of the major criteria for MF is the presence of one of the driver mutation. In 90% of cases, it would be 1 of the 3: JAK2 V617F, MPL515, or CALR.The consequence of either of the 3 is the hyperactivity of the JAK/STAT pathway. The difference in having1 over the other is evident in the clinical presentation in terms of their clinical picture and what the overall outcome for the patient is.

For example, patients with calreticulin-mutated MF are usually younger. They have a higher platelet count and are less likely to be anemic, thrombocytopenic, or require transfusions. They have a much better overall prognosis because the clinical scenario appears to be less aggressive. The triple-negative patients who do not have any of the three driver mutation, on the other hand, have the worse outcomes because they have more aggressive disease. 

There is a spectrum that includes patients with calreticulin on 1 end, triple-negative on the other end, and in the middle, there are patients with JAK2 or MLP mutations.

The management of these different patients is not affected by the presence or absence of these mutations, however. This goes for JAK2 inhibitors, anemia drugs, and any other drugs use to manage the symptoms and signs of the disease.

TARGETED ONCOLOGY: What information can be gathered from non-driver mutations?

Verstovsek: While driver mutations have diagnostic and subsequent prognostic value, non-driver mutations actually have much more impact on patient outcomes. As a result, prognostic scoring systems are being developed beyond the initial attempts from more than 10 years ago. We initialy developed international prognostic scoring system that includes variables like symptoms, age, and blood cell count. This simple approach has been enhanced with the ability to expand the genetic testing from the driver mutations to others.

Therefore, we now have prognostic scoring systems that include multiple different variables, including the presence or absence of different driver mutations and non-driver mutations. One of the more recent ones that have gained popularity is called MIPSS70. This is a molecular enhanced prognostic scoring system with multiple variables that can provide prognostic information for patients beyond the basic clinical assessment.

In testing for the multiple other genetic abnormalities, there is a limit to the knowledge that we can get. Some assays test 400 genes, and the more genes you test, the more abnormalities you may find. Through collaborative research, though, a set of 5 different mutations have been identified to be prognostically important. These are called high-risk molecular abnormalities and non-driver mutations and include SXL1EZH2SRSF2, and IDH1/2. This is something that has been confirmed multiple times.

The real application of the knowledge about genetic complexity is in patients with early-stage MF. If present, then we can advise patients to get to transplant sooner rather than later. That is the true practical benefit of genetic testing, identification of genetic complexity, and implementation of the prognostic scoring systems that include genetic information.

TARGETED ONCOLOGY: In what cases is mutational profiling most important for patients with MF?

Verstovsek: This is a topic of significant importance. Some may envision that most doctors are familiar with the need for prognostication with the primary goal of identifying patients that have a life expectancy of fewer than 5 years. These are typically intermediate-2 to high-risk patients (patients fit into 4 categories which are low-risk, intermediate-1, intermediate-2, and high-risk). It is the intermediate-2 to high-risk patients that we should refer to transplant. That is the primary use of prognostication.

In a case where you utilize the old-fashioned international prognostic scoring system, you may have patients in the lower-risk groups who are doing fine and delay transplant. With genetic testing, however, one can enhance ability to prognosticate, to say that the patient should go to transplant sooner, based on abnormalities found in their genetic profile. In this case, additional information from genetic testing would help us have a more aggressive approach with a patient.

TARGETED ONCOLOGY: Your stated that the overall purpose of getting this mutational information is to help refer some patients for transplant. Exactly which patients would you identify as eligible for transplant?

Verstovsek: This information is built into the National Cancer Care Network (NCCN) guidelines, which guide everyday practice. It suggests that the first step when you have a new patient with MF is to prognosticate to see what the outcome is and if they are intermediate-2 and high-risk.

These guidelines state that if you have the ability to do genetic profiling at a larger scale, then consider early intervention with transplant in patients who are otherwise intermediate-1 risk, but have the genetic complexity identified.

The guidelines do not mandate widespread genetic testing because it may be impossible to do, in terms of cost-effectiveness.

Eventually, the NCCN guidelines will switch to include more complex prognostic scoring systems that have been developed, like MIPSS70, and endorse it for widespread use. [Currently], in everyday practice we are missing pieces of information that would help us embrace MIPSS70 or similar prognostic scoring systems, as standard.

TARGETED ONCOLOGY: Can you discuss the clinical and prognostic implications of the myelodepletive phenotype and the therapeutic options and limitations for these patients?

Verstovsek: One aspect of prognostication is to identify patients that would be having a different outcome based on the age, blood cell count, as I previously mentioned, and also based on karyotype and genetics. It has been realized that there are some other factors of importance that are possibly more important for the management of these patients.

A theme among academic circles has been that different patients with MF may have a different clinical phenotype, for example, there are patients with a very low blood cell count and are symptomatic without a large spleen; and there are patients with high blood cell count and very large spleen. The clinical experience of seeing such patients has led us to conclude that there appears to be two types of patients. Myelodepletive phenotype is the 1 that is clinically more challenging to manage, and overall, management of these patients in unsatisfactory. 

Perhaps we should be looking beyond just the prognostication for risk of death and look a clinical experience to identify patients that are need of different types of therapies based on their clinical presentations.


Bose, P., Verstovsek, S. Mutational profiling in myelofibrosis: implications for management. Int J Hemato. 2020. 111, 192–199. doi: 10.1007/s12185-019-02758-z

Related Content