Myeloproliferative neoplasms and personalized medicine: the perfect match?

For decades the Philadelphia-negative myeloproliferative neoplasms (MPNs) have been treated with a handful of non-specific cytoreductive drugs like hydroxyurea (HU), pipobroman, or busulfan.1 Anagrelide, with its specific action on thrombocytosis, and interferon alpha (IFNa) opened new perspectives for a more personalized approach. For example, the PT-1 study showed for the first time in MPNs that molecular lesions could have an impact on response to therapy.2 In that randomized study, patients with essential thrombocythemia (ET) and JAK2V617F mutation had better outcomes when treated with HU compared to anagrelide, a difference that was not significant in patients without the mutation.3 In addition to this possible difference in terms of clinical and hematologic response to a given drug, the introduction of biomarkers in cancer management also reached the MPN area, initially in chronic myeloid leukemia (CML) where BCR-ABL1 is now a clear marker of treatment efficacy.54 In classical MPNs, first possible evidence for such a biomarker came from studies with IFNa showing that this drug could drastically reduce JAK2V617F mutant clones, an effect that can be monitored in peripheral blood by quantification of the mutant allele burden over time.6 However, response to IFNa may depend on the mutations found in the MPN clones, since it was also shown that clones harboring mutations in the TET2 gene could be less sensitive to this drug than JAK2-mutated clones.7 Among MPNs, myelofibrosis is the disease with usually more complex mutational patterns, patients often harboring more than one mutation.8 Indeed, previous studies showed that some mutations (in ASXL1, EZH2, SRSF2, IDH1-2 genes for example) conferred a high risk of poor outcome.9 The next step for development of personalized medicine in MPNs will, therefore, probably come from the development of next generation sequencing (NGS) techniques. When widely available (and if affordable), NGS will allow for the precise determination of molecular patterns of patients with MPNs.10 These new data will hopefully help performing prospective validation of the prognostic and therapeutic impact of molecular lesions. Indeed, such data may also allow us to reclassify the individual subcategories of diseases since currently our diagnostic categories and treatment targets are also still based upon clinical descriptions of entities described over a century ago.1311 One could argue perhaps that more progress has been made in the rarer MPN entities such as the chronic eosinophilias and recently chronic neutrophilic leukemia, where the entities have been identified on a molecular basis and targeted treatment, e.g. with BCR/ABL kinase inhibitors or, recently, ruxolitinib for CSF3R mutation positive chronic neutrophilic leukemias.1514

However, personalization of MPN therapy could also rely on tools other than molecular techniques. Another possible clue to help in the treatment choice in ET, for example, could be bone marrow morphology. After PT-1, a second randomized study compared HU and anagrelide in high-risk ET patients, the Anahydret study.16 One of several differences with PT-1 consisted of the diagnostic criteria used for ET: WHO in Anahydret versus PVSG criteria in PT-1. This difference probably explains in part the different conclusions of these two studies (non-inferiority of anagrelide in Anahydret, compared to superiority of HU in PT-1). Whatever the possible debate about these results, one can consider that it could be possible to determine a profile of ET patients more likely to clearly benefit from HU, e.g. having the JAK2V617F mutation, more predisposition to arterial events, some degree of fibrosis in the bone marrow, etc. Or, indeed more radically as alluded to earlier, we could completely reclassify these diseases using scientifically robust tools into specific “personalized” categories.

In addition to these current attempts to better classify patients in order to better tailor their therapy, the discovery of the JAK2V617F mutation in MPNs also prompted the development of so-called targeted therapy, i.e. firstly intensive clinical research using JAK inhibitors. These clinical studies led to the approval of the top-of-the-class JAK1-JAK2 inhibitor, ruxolitinib, in myelofibrosis and, more recently, for the treatment of PV patients resistant to or intolerant of HU.1917 Other JAK inhibitors are currently evaluated in phase III studies, like pacritinib or momelotinib, that could provide other new drugs for MPN patients.2120 Although ruxolitinib is a potent agent to treat splenomegaly and symptoms, some characteristics of other JAK-inhibitors found in early phase studies could be of particular interest in subsets of patients that currently do not benefit from ruxolitinib therapy; e.g. pacritinib did not worsen thrombocytopenia in patients with MF and low platelet counts.22 On the other hand, momelotinib was found to induce interesting responses on anemia, specially in transfusion-dependent MF patients.23 If these findings are confirmed in the ongoing phase III studies, along with a significant efficacy on splenomegaly and symptoms, they could help increase the armamentarium to treat MF patients and address specific therapeutic needs according to the main clinical need.

All these changes in the management of MPNs will clearly change the economic impact of these not so rare diseases. Diagnostic and prognostic work up costs will necessarily increase if a bone marrow biopsy is mandatory to diagnose all the three MPN subtypes as suggested in a recent proposal for up-dating the WHO diagnostic criteria.24 To date, this biopsy is not routinely done in many centers to make a diagnosis of PV for example. In terms of prognosis, current classification of ET and PV patients relies on very simple markers: age and history of vascular events.1 If one considers that assessment for mutations associated with a poor outcome is necessary for the proper management of younger patients, the cost of NGS techniques will further increase the cost of MPN base-line assessment. Finally, the cost of therapies is also increasing rapidly when considering the development of alternatives to HU, including IFNa and JAK inhibitors. Therefore, a real benefit for patients should be clearly evidenced for supporting these new therapies. IFNa is currently being tested head-to-head against HU in different clinical studies in PV and ET patients, and this will hopefully soon provide clear evidence of its exact role in the management of these diseases (clinicaltrials.gov identifiers: 01259856 and 01949805). Ruxolitinib, the only approved JAK inhibitor to date in MPNs, provided clear benefit for many MF patients compared to other available therapies, as demonstrated in the COMFORT-2 study.17 The recent RESPONSE study also showed its efficacy as second-line therapy after HU in PV,19 a setting in which other drugs, usually much less expensive, can be used. Further studies in PV are needed to clarify the benefit of ruxolitinib in altering the natural history of PV in reducing risks of thrombosis and transformation. The burden of MPN therapy may, therefore, clearly and significantly increase in the coming years, and national and international societies will certainly have to propose evidence-based, fair and balanced guidelines for the management of MPNs and these will increasingly be based upon a personalized approach.25

Footnotes

• Financial and other disclosures provided by the author using the ICMJE (www.icmje.org) Uniform Format for Disclosure of Competing Interests are available with the full text of this paper at www.haematologica.org.