Blast-phase myeloproliferative neoplasms: risk factors and treatment approaches

Abstract

Introduction:

The past 10 years have seen dramatic advances in the understanding of the molecular pathogenesis of BCR-ABL negative myeloproliferative neoplasms (MPNs). With this knowledge has come novel, molecularly targeted therapies such as JAK inhibitors that may decrease symptoms and improve quality of life for patients with MPNs. Despite these advances, progression of the disease to an acute leukemic (blast) phase remains difficult to predict and even more difficult to treat, with high rates of disease relapse and mortality.

Areas covered:

We performed a literature review of known risk factors for progression of MPNs towards blast phase and treatment options for transformed disease, including approved and investigational agents. Herein, we review the current literature and suggest strategies for improving outcomes in the future.

Expert commentary:

Further understanding of the biologic basis for transformation of MPNs from the chronic to blast phase is needed in order to predict, prevent, and treat these cases. Patients with MPNs in blast phase should be encouraged to participate in clinical trials whenever possible.

1. Introduction

The term myeloproliferative neoplasm (MPN) describes a group of bone marrow disorders with heterogeneous clinical manifestations and variable prognostic implications. For one MPN subtype, chronic myeloid leukemia, the discovery of a chromosomal rearrangement causing BCR-ABL1 gene fusion has led to the development of extremely effective targeted therapeutics and dramatically improved outcomes. For the BCR-ABL1-negative MPNs, namely essential thrombocythemia (ET), polycythemia vera (PV), and primary myelofibrosis (PMF), experts are continuing to unravel the molecular mechanisms of pathogenesis and disease progression. It is now clear that activation of the Janus kinase-signal transducers and activators of transcription (JAK-STAT) pathway is fundamental to the development of MPNs, and occurs via activating mutations in JAK2 [1–4], myeloproliferative leukemia (MPL) virus oncogene [5,6], and calreticulin (CALR) [7–10].

In their chronic phase, MPNs are characterized by uncontrolled cellular proliferation with intact differentiation. Over time, these disorders can precipitate profound constitutional symptoms, thrombovascular events, and progressive bone marrow failure. Perhaps, the most feared complication in the natural history of MPNs is the arrest of normal differentiation and subsequent transformation to acute myeloid leukemia, also known as blast phase (MPN-BP). The International Working Group for Myelofibrosis Research and Treatment (IWG-MRT) has standardized the nomenclature to describe transformation of BCR-ABL1-negative MPNs, such that accelerated phase (MPN-AP) refers to the presence of 11–19% blasts in the peripheral blood or marrow, and blast phase refers to ≥20% blasts in periphery or marrow [11]. Despite extensive study, the mechanisms of blast phase transformation remain somewhat unclear [12,13]. Blast phase transformation occurs at variable frequencies between MPN subtypes, with PMF being the most likely to transform. While various clinical and pathologic features have been associated with increased risk of transformation, it often remains difficult to predict which patients will progress. Once blast phase transformation occurs, outcomes are generally dismal, with median survival from the time of blast phase transformation of approximately 5 months [12,13]. Treatment options at the time of progression remain somewhat limited, though advances in the understanding of the mechanisms of transformation may pave the way for novel therapeutic approaches.

2. Risk factors for progression to blast phase

Due to the profound clinical implications of blast phase transformation, concerted efforts have been made to identify those at highest risk for progression early in the disease course. Clinical, pathologic, and most recently molecular features have been evaluated for association with progression, with variable degrees of predictive success (see Table 1). In aggregate, these factors can be used to inform risk-directed management strategies such as monitoring interval and treatment timing and intensity while still in the chronic phase of the disease.

Table 1.

Risk factors for blast phase progression.

Clinical	Reference
Laboratory parameters
Circulating blasts	[13,16]
Cutoffs range from ≥3% to ≥10%
Thrombocytopenia	[13,16]
Platelets <50,000/μL or <100,000/μL
Leukocytosis	[14,18]
WBC >30,000/μL (or ≥15,000/μL for PV)
WBC >30,000/μL	[18]
Transfusion dependence	[17]
Age >61 years (PV)	[14]
Exposures
P32/Chlorambucil	[14]
Pipobroman	[14]
Pathologic
MPN subtype (PMF)	[14,25]
Unfavorable cytogenetics	[15,26–29]
Including17p abnormalities, complex karyotype, +8, −7/7q-, −5/5q-, 12p-, inv(3), and 11q23 rearrangements
Negative for JAK2, CALR, or MPL mutations	[25]
Additional somatic mutations	[44,45,50]
Including ASXL1, SRSF2, IDH1/2, TP53, and TET2
Number of additional somatic mutations >2	[50]
Prognostic scoring systems
DIPSS score of intermediate-2 or high risk	[54]

DIPSS: Dynamic International Prognostic Scoring System; MPL: myeloproliferative leukemia; MPN: myeloproliferative neoplasm; WBC: white blood cells; PMF: primary myelofibrosis; PV: polycythemia vera.

2.1. Clinical factors associated with blast phase transformation

2.1.1. Clinical and laboratory features

Various clinical features and laboratory parameters have been associated with leukemic transformation, though conclusions may be limited by heterogeneous disease features in some studies and small patient populations in others. This may be further confounded by the observation that the clinical importance of such factors may vary between MPN subsets. In one study of >1500 patients with PV, 50 (3%) transformed to blast phase at a median time of 10.8 years. In this cohort, risk for transformation was associated with older age (>61 years) and white blood cell (WBC) ≥15,000/μL [14]. In PMF, several studies have identified different factors associated with transformation. A study of 370 consecutive patients with PMF treated at MD Anderson Cancer Center identified thrombocytopenia (<50,000/μL) and circulating blast count ≥10% to be clinical factors associated with leukemic transformation [15]. A similar retrospective study of 331 patients with PMF seen at the Mayo Clinic found the two strongest clinical predictors of leukemic transformation to be peripheral blasts ≥3% and platelets <100,000/μL [16]. As would be expected, it has been shown that progression to accelerated phase portends increased risk of further progression to blast phase and subsequent shortened survival times compared to those who remain in the chronic phase of the disease [15]. Others have shown red blood cell transfusion requirement [17] and WBC >30,000/μL [18] to be predictive in patients with PMF.

2.1.2. Drugs

Exposure to certain cytoreductive agents has been shown to increase risk of MPN-BP transformation. Agents with proven leukemogenic potential in this setting include alkylating agents, specifically chlorambucil and pipobroman, and radioactive phosphorus (P-32) [14,19,20]. As a result, these agents have fallen out of favor for the treatment of MPNs. Of note, the leukemogenic potential of hydroxyurea, which is routinely used clinically, has historically been debated; however, recent studies show no definitive evidence for its association with leukemic progression [14,16,21,22]. The largest studies to date are reported by Tefferi et al. and Finazzi et al. examining 1545 and 1638 patients with PV, respectively, and both independently found no association between hydroxyurea use and blast phase progression [14,21].

2.2. Pathologic factors associated with blast phase transformation

2.2.1. MPN subtype

It is clear that the risk of leukemic transformation differs between MPN subtypes. PMF confers the highest risk of blast phase development, with 10–20% of patients undergoing disease transformation at 10 years [13,23,24]. Those with PV have a lower incidence – one study reported 2.3% at 10 years and 7.9% at 20 years [14] – and leukemic transformation from ET is uncommon. In 826 patients with MPN followed at the Mayo Clinic over 20 years, the cumulative incidence for blast phase transformation for patients with ET, PV, and PMF was 3.8%, 6.8%, and 14.2%, respectively [25]. Even in those with the same molecular aberration (JAK2V617F), this difference persists across MPN subtypes, as demonstrated by significantly improved overall survival (OS) for those with JAK2-mutated ET when compared to those with PV and JAK2-mutated PMF (19 years, 13.4 years, and 4.8 years, respectively).

2.2.2. Cytogenetic features

In concordance with other myeloid neoplasms, certain cytogenetic features have been shown to be associated with prognosis for MPNs. Several chromosomal abnormalities are specifically associated with a risk of blast phase progression, including chromosome 17p abnormalities [15], complex karyotype, +8, −7/7q-, −5/5q-, 12p-, inv(3), or 11q23 rearrangements [26–28]. One study showed that acquisition of aberrations of chromosomes 5, 7, or 17p in granulocytes portended a six times greater risk for blast phase progression (95% CI 1.2–27.7, P = 0.006) [29]. Unfavorable karyotype has been included in the most recent prognostic model for MPNs, the Dynamic International Prognostic Scoring System (DIPSS)-plus, as described below [28].

2.2.3. Molecular features

With advances in understanding of molecular mechanisms in MPNs come new opportunity to subclassify these disorders further in order to predict disease behavior and response to therapies. We now know that molecular aberrations resulting in activation of the JAK-STAT pathway, including gene mutations in JAK2, CALR, and MPL are paramount to the development of MPNs in most cases, and that these mutations are mutually exclusive [4,25,30–32]. JAK2-activating mutations are present in >95% of PV cases, and also the most common mutation in ET and PMF [33]. CALR mutations are present in 15–32% of ET cases and 25–35% of PMF cases, and MPL mutations are seen in about 4% of ET and 8% of PMF [5,7,8,33–35].

The presence or absence of JAK2, MPL, or CALR mutations has been shown to be prognostically informative. One study evaluated survival and blast phase transformation in >1500 patients with MPNs by their identified molecular aberrations [25]. In those with ET, they found no association between mutational status and outcomes. For PMF, however, mutational status was associated with both survival and blast phase transformation. Patients with CALR mutations fared the best and those with triple-negative status fared the worst, with median survival of 15.9 years versus 2.3 years, respectively (HR 5.1, 95% CI 3.2–8.0). The incidence of blast phase transformation was significantly lower in CALR-mutated cases than triple-negative cases, with transformation rates of 6.5% and 25%, respectively (HR 7.6, 95% CI 2.8–20.2). Those with CALR mutations also had lower incidence of transformation when compared to those with JAK2 mutations (HR 2.7, 95% CI 1.1–6.6), but not when compared to those with MPL mutations (HR 1.9, 95% CI 0.5–7.7). Several other studies have confirmed the favorable prognostic impact of CALR mutations and unfavorable impact of triple-negative status, particularly in PMF [8,36]. In addition, the type of CALR mutation identified may matter. Tefferi et al. found that patients with PMF characterized by type 1 CALR mutations, which involve a 52-bp deletion in exon 9 of the gene, experienced significantly longer survival than those with type 2 mutations, which involve a 5-bp insertion in exon 9 (HR 2.5, 95% CI 1.1–5.4) [37], a finding that has since been replicated by the same group of investigators in a separate cohort of patients recruited from the Mayo Clinic and University of Florence [25]. Those with type 2 CALR mutations were found to have a significantly higher peripheral blast percentage (P = 0.04), but risk of transformation and leukemia-free survival was not assessed [37].

Others have examined the impact of mutational burden on clinical course, with discordant results. One study found that higher JAK2 mutant allele burden (>50%, as opposed to <50%) in patients with PV was associated with fibrotic transformation, but not with blast phase transformation [38]. Another group showed that those with JAK2 mutant allele frequencies in the lowest quartile had significantly shorter overall and leukemia-free survival compared to those with higher mutational burden or wild-type JAK2 [39]. In contrast, a retrospective study of patients with JAK2-mutated PMF in blast phase found that 80% had high JAK2 mutant allele levels (>50%) [40]. Another study assessed JAK2 kinetics around the time of transformation in 16 patients. In the four patients who were negative for the JAK2V617F mutation while in chronic phase, all four remained negative in blast phase. Of the 12 who were JAK2 mutation positive in chronic phase, mutation percentage at the time of transformation rose in 1 case, decreased in 7 cases, and remained stable in 4 [13]. Others have similarly shown a lack of consistent role of JAK2 presence and kinetics in blast phase transformation [41,42].

As current understanding of the mutational landscape of MPNs has progressed, this story has become more complex. We now know that numerous other gene mutations in addition to JAK2, CALR, and MPL exist in MPNs in a recurrent fashion and in patterns distinct from de novo acute myelogenous leukemia (AML), including those involving epigenetic modifiers and spliceosomal machinery. For example, those with secondary AML evolved from MPNs or myelodysplastic syndrome (MDS) are more likely to have mutations in TP53, CUX1, and 9pUPD when compared with de novo AML, where mutations in NPM1 and FLT3 are more common [43]. One study evaluated a total of 879 patients with PMF for gene mutations previously described as prognostic in AML [44], and identified 3 that predicted for decreased leukemia-free survival (ASXL1, SRSF2, and IDH1/2), and 3 that predicted for decreased OS (ASXL1, SRSF2, and EZH2). Only ASXL1 mutation maintained relevance independent of prognostic model score [45]. Other studies have identified several mutations that occur with increased frequencies in leukemic blasts from transformed MPNs, including TP53, TET2, SH2B3, and IDH1, suggesting that such mutations may play a role in leukemogenesis [46–49]. Another group used a targeted next-generation sequencing panel of 104 known cancer-related genes to follow the mutational status of 197 patients with MPNs over time. In this study, the mutational profile was relatively stable over time, with only one new somatic mutation identified per 66 patient-years. They found that mutations in TP53 and TET2 were associated with inferior survival and increased risk for blast phase transformation, and that those with heterozygous TP53 mutations remained in a chronic phase until loss of the wild-type allele occurred, at which time rapid leukemic progression was observed. The number of somatic mutations (>2) observed was found to be predictive of survival and leukemic transformation. Those with ASXL1 mutations had lower hemoglobin levels at diagnosis and those with EZH2 mutations had higher WBC counts at diagnosis, though patients with these mutations were not found to have any differences in clinical course [50].

While familial MPNs are rare, one group recently reported autosomal dominant transmission of a copy number variation on chromosome 14 that results in duplication of ATG2B and GSKIP and was associated with development of hematologic malignancies in four families [51]. In these families, two-thirds (22/33) of the individuals affected presented with an ET phenotype, which was diagnosed at a younger age compared to sporadic forms, and half progressed to myelofibrosis or AML. Overexpression of ATG2B and GSKIP was found to stabilize cells harboring mutations in JAK2, MPL, or CALR and was also associated with increased likelihood of developing additional deleterious mutations such as TET2. Further investigation into the role of these two genes in sporadic MPN cases is ongoing.

2.3. Prognostic models

Several prognostic scoring systems have been developed that use clinical features to predict survival in patients with MPNs, specifically in the setting of PMF. These include the International Prognostic Scoring System (IPSS), the DIPSS, and the DIPSS-Plus, the latter of which incorporates unfavorable cytogenetics, red blood cell transfusion requirements, and thrombocytopenia into its algorithm [28,52,53]. While initially developed in order to predict OS, the DIPSS has been shown to predict for blast phase transformation in PMF as well [54]. Under the DIPSS score, those with intermediate-2 or high-risk disease were 7.8 or 24.9 times more likely to develop blast phase disease than those with low risk [54]. In contrast, in multivariate analysis using the patient population identified for development of the newer DIPSS-plus score, DIPSS alone did not predict for leukemic transformation, but the combination of two factors – thrombocytopenia and unfavorable karyotype – was able to risk-stratify patients into a higher-risk group with risks of 5-year transformation and 10-year transformation of 18% and 31%, respectively, compared to a lower-risk group with risks of 6% and 12%, respectively [28].

3. Treatment approaches for blast phase MPNs

Once transition to MPN-BP occurs, the disease is difficult to treat and confers a dismal prognosis. Even with intensive therapy, the median survival after blast phase transformation is generally <6 months [12,13]. The Mayo Clinic reported on 91 cases of MPN-BP, and found 98% mortality among their cohort with a median survival of just 2.6 months [12]. It is clear that more effective management strategies are needed for these patients; however, clinical trials for MPN-BP are limited by relatively sample sizes and heterogeneous disease biology, as they are often included with other secondary AMLs or de novo AML. The few clinical trials that have focused on MPN-BP are detailed in Table 2. In addition, previous studies were limited by a lack of uniform, disease-specific response criteria, so results were reported using standard AML criteria [55]. This has now been addressed by the development of a defined set of response definitions for MPN-BP that specify depth of response by including complete molecular response and complete cytogenetic response classifications, and recognizes a return to chronic phase MPN, which they refer to as acute leukemia response-partial [56]. It should be noted, however, that the clinical application and relevance of these criteria await validation in the context of well-designed prospective clinical trials.

Table 2.

Summary of key clinical trials for blast phase MPNs.

Treatment(s) studied	Population	Key results	Reference
Allogeneic hematopoietic stem cell transplant or decitabine (if ineligible for transplant or without a donor)	11 patients with MPN-BP	Transplant group: at 20 months, 53% remained alive. Decitabine group: At 9 months, 67% remained alive	[72]
Azacitidine	54 patients with MPN-BP (n = 26) or MPN-AP (n = 28)	ORR 52% (CR 24%) Median response duration 9 months For accelerated phase disease CR 35%, vs. 12% for blast phase disease	[73]
Ruxolitinib	Relapsed/refractory AML, including 18 patients with MPN-BP	Among those with MPN-BP: ORR 17% (2/18 CR, 1/18 CRi)	[78]
Decitabine + ruxolitinib	MPN-BP and MPN-AP	Phase I/II study is ongoing and open to accrual. Anticipated accrual of 60 patients	NCT02076191

AML: acute myelogenous leukemia; CR: complete remission; CRi: complete remission with incomplete blood count recovery; MPN-BP: myeloproliferative neoplasm-blast phase; MPN-AP: myeloproliferative neoplasm-accelerated phase; ORR: overall response rate.

3.1. Intensive approaches – induction chemotherapy and allogeneic hematopoietic stem cell transplantation

Allogeneic hematopoietic stem cell transplantation (AHSCT) remains the only potentially curative therapy for MPN-BP. Induction chemotherapy alone, without subsequent AHSCT, may induce responses but these are not durable. In the aforementioned Mayo Clinic series of 91 patients with MPN-BP, 24 patients underwent various induction chemotherapy approaches. No complete responses were seen, but 41% had a return of their disease to the chronic phase. Only one patient in that retrospective series went on to AHSCT [12]. MD Anderson Cancer Center reported outcomes of 74 patients with MPN-BP treated by various approaches, including intensive chemotherapy (55% of patients), low-intensity chemotherapy (16%), (which included gemtuzumab ozogamicin, various hypomethylating agents as monotherapy and in combinations, and vincristine plus prednisone), and supportive care alone (26%) [13]. A total of eight patients received an AHSCT either upfront, or after responding to intensive chemotherapy. Those who were treated with supportive measures alone had a median survival of only 6 weeks. Those treated with chemotherapy alone (either intensive or less intensive) without subsequent transplant did similarly, with median survivals of 6 months and 7 months, respectively. Among this group, those who achieved a complete remission (CR) or CR with incomplete blood count recovery (CRi) had an improved median survival of 13 months compared to just 4 months in those who did not. AHSCT resulted in significantly better outcomes, as 73% remained alive at a median follow-up of 31 months.

Another group from University Health Network in Toronto devised a more uniform approach whereby all patients with MPN-BP with reasonable levels of fitness were offered curative-intent therapy first with induction chemotherapy, and those with good response and a suitable donor went on to AHSCT [57]. They reported results from 75 patients treated via this algorithm. Approximately, half (52%) received curative-intent therapy, and 2-year OS was 26% in this group as opposed to only 3% in the non-curative-intent group. Among those treated with curative intent, 46% achieved CR or CRi, and an additional 31% reverted to chronic phase of the disease, with 43% then going on to transplant. Survival of those who underwent transplant was significantly better than those who did not, with 2-year OS of 47% vs. 15%. Another study from the Societe Française de Greffe de Moelle et de Therapie Cellulaire (SFGM-TC) reported a more modest leukemia-free survival of 19% in 60 patients with MPN-BP or MDS/MPN-BP treated with AHSCT [58]. The Mayo Clinic group reported on the importance of obtaining disease response prior to transplant in a small cohort [59]. At the time of transplant, 5/8 patients were either in CR or had return of their disease to chronic phase, and all five of these patients remained alive and in CR at a median follow-up of 20.3 months after transplant. Of the three patients with persistent leukemic disease at transplant, two had relapsed and expired after transplant. While small sample sizes, heterogeneity of disease characteristics, treatment specifics, and the retrospective nature of most published reports limit the generalizability of available data thus far, intensive approaches including AHSCT remain a standard of care for those who are eligible, and represent a potentially curative approach for some. Questions yet to be answered include degree of response required prior to transplant (is return to chronic phase good enough or is CR better?), intensity of conditioning required, and role for alternative donor approaches in this setting [60].

3.2. Hypomethylating agents

Hypomethylating agents decitabine and azacitidine, both US FDA approved for MDS, have been employed as low-intensity agents to treat MPN-BP, and several small clinical studies have demonstrated modest efficacy of this approach. There are several molecular rationales supporting epigenome-modifying therapy in MPN-BP. First, epigenetic modification including hypermethylation of p15 and p16 genes has been associated with blast phase transformation of MPNs [61]. MPNs have also shown alterations in genome-wide methylation patterns, and those with aberrant DNA methylation were found to be responsive to decitabine in vitro [62]. Treatment with decitabine followed by a histone deacetylase (HDAC) inhibitor has been shown to cause reduction in CD34+ cells, reduction in JAK2 mutant hematopoietic progenitor cell proportion, and correction of abnormal trafficking of CD34+ cells in myelofibrosis cell lines and murine models [63,64]. Further, gene mutations in epigenetic modifiers such as TET2, ASXL1, SRSF2, EZH2, and IDH1/2 have been implicated in pathogenesis of MPNs, though variably associated with leukemic transformation. In MDS, improved response to hypomethylating agents has been associated with certain mutations in epigenetic modifiers including TET2, which is present in approximately 14% of MPNs [65–69]. Decitabine has been shown to be clinically effective in chronic phase myelofibrosis [70–72].

The Groupe Francophone des Myelodysplasies examined 54 patients with BCR-ABL1-negative MPNs who experienced disease progression to acute leukemia or MDS for which they were treated with azacitidine [73]. They demonstrated a 52% overall response rate (ORR) including 24% CR, and median response duration of 9 months. Responses were heterogeneous, with those transformed from ET faring the best. Those with post-ET transformation compared to post-PV transformation had ORR of 71% versus 33%, respectively, and CR rates of 43% versus 14%, respectively. Response rates were also higher in those with lesser-grade progression classified as MDS as opposed to AML, with CR 35% compared to 12% [73]. In a retrospective analysis by the MD Anderson Cancer Center group, they found that 6/21 patients (29%) with transformed MPNs responded to decitabine (3 CR, 2 CRi, 1 PR). The median response duration was 7 months and those who responded lived longer, with median survival for responders of 10.5 months compared to 4 months for non-responders. Higher response rates were seen among those with lesser-grade disease classified as accelerated phase, with 62% (8/13) ORR [74]. Another small study reported outcomes from 11 consecutive patients with MPN-BP treated with either AHSCT or decitabine if ineligible for transplant or without a donor. While limited by small cohort size, 4/6 patients who received decitabine were still alive at 9 months [72]. While the overall impact of hypomethylating agents on the natural history of transformed MPNs appears relatively modest, they may prolong survival in those unable to be treated with curative intent, and their relatively favorable tolerability profile may allow for these agents to be utilized as a bridge to transplant and as a backbone to test more efficacious therapeutic combinations in the future.

3.3. JAK inhibitors

Discovery of the role of overactive JAK-STAT signaling in the pathogenesis of MPNs paved the way for development of molecularly targeted agents directed at this pathway. The JAK1/2 inhibitor ruxolitinib has gained FDA approval for patients with PMF [75,76], and more recently as a second-line option for those with PV [77]. Ruxolitinib has proven useful in improving splenomegaly and symptoms related to chronic phase MPNs; however, the extent of its effects on the natural history of these disorders and on the underlying malignant clone remains a matter of debate [78,79]. The utility of JAK inhibitors in the setting of MPN-BP is under investigation. A single-agent phase II study of ruxolitinib in relapsed/refractory AML included 18 patients classified as MPN-BP. Subjects were treated with ruxolitinib at a starting dose of 25 mg orally twice daily. Responses were seen in 3/18 patients – two attained CR and one attained CRi [78]. Ruxolitinib was relatively well tolerated in these patients, with only 4 of 38 total enrollees experiencing grade 3 or greater toxicities.

3.4. Therapies under investigation

As understanding of molecular mechanisms driving MPNs evolves, novel therapies and therapeutic combinations can be expected to follow. In an in vitro model using bone marrow cells obtained from a JAK2V617F mutant mouse after MPN-BP development, the combination of decitabine and ruxolitinib was found to have a synergistic effect on colony suppression [30]. The MPN research consortium is conducting a phase I/II multicenter study of this combination for transformed MPNs (NCT02076191). Other approaches involving JAK inhibitors are under investigation for chronic phase MPNs, including combinations with HDAC inhibitors, immunomodulatory agents, anti-fibrosing agents, hedgehog pathway inhibitors, and phosphoinositide 3 kinase (PI3K) inhibitors (see Table 3). Interim results of several combination trials have been presented at national meetings. A phase Ib/II study combining hedgehog pathway inhibitor sonidegib with ruxolitinib reported results of 27 patients with myelofibrosis. At 24 weeks, 55.6% of patients had a reduction in palpable spleen length by ≥35%, with toxicities of note including worsening of anemia, muscle spasms, elevated creatine kinase (CK), and myalgias [80]. A phase Ib study of PI3K inhibitor buparlisib with ruxolitinib for myelofibrosis reported ≥50% reduction in palpable spleen length in 82% (18/22) of JAK-inhibitor naive patients, and 55% (11/20) of patients previously treated with JAK inhibitors. Toxicities were primarily hematologic in nature, namely worsening of anemia and thrombocytopenia [81].

Table 3.

Novel agents and combinations under investigation.

Novel agent or combination	Mechanism of action	Phase of study	Disease indication	NCT ID
Decitabine + ruxolitinib	Hypomethylating agent + JAK 1/2inhibitor	I/II	MPN-AP and MPN-BP	NCT02076191
Azacitidine + ruxolitinib	Hypomethylating agent + JAK 1/2inhibitor	II	Myelofibrosis and MDS/MPNs, in chronic phases	NCT01787487
Pabinostat + ruxolitinib	HDAC + JAK 1/2 inhibitor	I/II	Myelofibrosis in chronic and accelerated phases	NCT01693601
Lenalidomide + ruxolitinib	Immunomodulator + JAK 1/2 inhibitor	II	Myelofibrosis in chronic phase	NCT01375140
Simtuzumab + ruxolitinib	Anti-fibrosing agent + JAK 1/2 inhibitor	II	Myelofibrosis in chronic phase	NCT01369498
Sonidegib + ruxolitinib	Hedgehog pathway inhibitor + JAK 1/2 inhibitor	Ib/II	Myelofibrosis in chronic phase	NCT01787552
Buparlisib + ruxolitinib	PI3K inhibitor + JAK 1/2 inhibitor	I	Myelofibrosis in chronic phase	NCT01730248
Ruxolitinib prior to stem cell transplant	JAK 1/2 inhibitor	II	Myelofibrosis in chronic phase	NCT01790295
Momelotinib	JAK 1/2 inhibitor	III	Myelofibrosis in chronic phase	NCT01969838
Pacritinib	JAK2 inhibitor	III	Myelofibrosis in chronic phase (on clinical hold)	NCT01773187
NS-018	JAK2 inhibitor	I/II	Myelofibrosis in chronic phase	NCT01423851
Alisertib	Aurora kinase A inhibitor	I	Myelofibrosis in chronic phase	NCT01668173
PRM-151	Anti-fibrosing agent	II	Myelofibrosis in chronic phase	NCT01981850
Imetelstat	Telomerase inhibitor	II	Myelofibrosis in chronic phase	NCT02426086

HDAC: histone deacetylase; JAK: Janus kinase; MPN-BP: myeloproliferative neoplasm-blast phase; MPN-AP: myeloproliferative neoplasm-accelerated phase.

Therapeutic agents for MPNs in chronic phase may be useful after transformation as well, perhaps in combination with intensive approaches such as AHSCT. Several studies have shown a suggestion of improved outcomes for patients with higher-risk PMF treated with ruxolitinib prior to AHSCT, perhaps as a result of improvements in the underlying disease as well as patient performance status, and this is currently under prospective investigation (NCT01790295) [82,83].

Other agents in development for chronic phase MPNs may similarly prove useful in combination with intensive therapies. Novel agents under investigation as single agents for chronic phase MPNs are incorporated in Table 3. Momelotinib, a potent JAK1/2 inhibitor, demonstrated improvement in symptoms and spleen size in most patients in a phase I/II study, while improving hemoglobin as opposed to prior drugs of the same class that tend to worsen anemia [84]. Preliminary results from a phase II study of anti-fibrotic agent PRM-151 showed that 9/26 (35%) of patients with chronic phase myelofibrosis had some degree of response, and 6 patients had demonstrable improvement in bone marrow fibrosis [85]. Imetelstat, a novel telomerase inhibitor, invoked responses in 7/33 (21%) of patients with high- or intermediate-2-risk PMF [86]. Four patients exhibited complete responses, and all four of these patients had reversal of their bone marrow fibrosis.

4. Expert commentary

MPNs are relatively uncommon disorders with a variable propensity to transform to blast phase disease. Due to the relative rarity of these disorders, MPN-BP, strong predictive models of disease transformation and robust data to guide treatment approaches are lacking. This therefore represents a significant area of unmet medical need. We strongly recommend participation in a clinical trial for this population whenever feasible and appropriate.

Despite these challenges, it has become clear that certain clinical and pathological features can identify patients at high risk of undergoing disease transformation. Given the consistently dismal prognosis once blast transformation develops, early identification of those at high risk of progression is key. We advocate that patients with suspected MPNs undergo a comprehensive evaluation at the time of diagnosis including a bone marrow examination to confirm both the MPN subtype and phase of disease, along with cytogenetic and genomic profiling at baseline to identify those with high-risk features. In those at high risk for disease progression, movement toward curative-intent therapy early in the disease course should be considered if the patient’s physical condition and donor status allows. Once blast phase transformation has occurred, an individualized approach to therapy is necessary. AHSCT remains potentially curative, and is generally our goal in eligible patients. Therapies including clinical trial approaches, hypomethylating agent therapy and JAK inhibitor therapy that have the potential to elicit disease responses and improve symptoms and performance status to the point that transplant may be feasible should be considered wherever possible.

5. Five-year view

Profound advances have been made in understanding the pathogenesis and developing effective treatments for MPNs in the past decade. However, significant gaps in knowledge continue to surround the issue of disease transformation to blast phase. Over the coming years, we expect that these gaps will begin to be filled. Improved understanding of the pathogenic mechanisms driving leukemic transformation is necessary in order to develop better tools to predict progression, and thus guide early intervention to prevent the development of blast phase disease.

At the moment, multiple promising molecularly targeted therapies, notably new potent and selective JAK inhibitors, anti-fibrotic therapies, and telomerase inhibitors, novel epigenetic modulators are undergoing clinical investigation as single agents and in combination for chronic phase MPNs. Such strategies may prove effective in controlling the underlying disease and thus preventing transformation. Alternatively, these agents may prove useful when used in sequence with intensive therapies, and may help individual patients make it to transplant. AHSCT strategies should be optimized for this disease subtype, which would include evaluation of optimal conditioning intensity and donor source. Posttransplant monitoring and the role, if any of posttransplant maintenance therapy and evaluation of minimal residual disease by molecular assays should be clarified.

Last, MPNs constitute a group of malignant disorders uniquely characterized by a low somatic mutational load, and in fact are often single-gene disorders. This attribute may lend MPNs susceptible to gene editing strategies. A group at Johns Hopkins University has used MPN-patient derived induced pluripotent stem cells harboring a JAK2V617F mutation to demonstrate that this mutation can be specifically targeted using CRISPR/Cas9 [87]. A group at Brigham and Women’s Hospital/Dana-Farber Cancer Institute has used CRISPR/Cas9 to introduce mutations in exon 9 of CALR in cell lines order to clarify pathogenic mechanisms of CALR mutations, suggesting that this gene may also be targe-table for therapeutic purposes [88]. Several other groups across the country are also evaluating the possibility of gene editing for these disorders in vitro.

Key issues.

• Clinical, pathologic and molecular features can be used to identify patients with MPNs at higher risk for blast phase transformation.

• Once blast phase transformation develops, prognosis is generally dismal.

• Allogeneic stem cell transplant remains the only potentially curative treatment option.

• MPN-BP represents an area of significant unmet medical need, and affected patients should be offered participation in clinical trials whenever feasible and appropriate.

• Hypomethylating agents and JAK inhibitors have shown the potential to induce responses for MPN-BP in some patients, but responses generally are not durable. Further study of these agents in combination and with other novel agents are ongoing.