Accelerated Phase of MPN: What It Is and What to Do About It

By consensus definition, MPN-AP refers to the presence of peripheral circulating or bone marrow blasts in the 10-19% range and MPN-BP refers to 20% ≥ blasts in the peripheral blood or marrow, in an individual with a pre-existing MPN^4,5. Overall survival (OS) in both MPN-AP and MPN-BP is limited particularly in the absence of allogeneic hematopoietic stem cell transplantation (allo-HCT). Even in the current era of myeloid therapies, outcomes for MPN-AP/BP are quite poor with median OS of 9 months reported in a multicenter retrospective study⁶.

The median survival in MPN-AP is in the 16- to 18-month range. Of note however, a relatively large series from the Mayo clinic shows that patients with primary myelofibrosis (PMF) with peripheral or marrow blasts in the 5-9% range have a survival outcome that is indistinguishable from that of patients with 10 to 19% blasts⁷. Similarly, data reported by MD Anderson demonstrates that a blast range of 5-9% in PMF has similarly poor outcomes when compared to 10-19% even in the era of ruxolitinib^8,9. Overall, patients with 5% or more circulating or marrow blasts can therefore be viewed as being on a continuum towards progression to MPN-BP, which has a particularly dismal survival outcome measured in the 3-5 month range. Therefore, it is reasonable to propose that therapeutic strategies designed for particularly high risk populations such as MPN-BP should similarly be applied to MPN-AP, including those with blasts in the 5% or greater range. For those who are fit enough to undergo allo-HCT, this should be entertained early on, with a view to bridging patients to allo-HCT as soon as there is effective disease control.

Risk factors for progression to MPN AP/BP include MPN subtype and is highest for PMF. In a large contemporary series, the 20-year cumulative estimate of blast phase transformation was 3.8%, 6.2% and 14.2% for essential thrombocythemia (ET), polycythemia vera (PV) and PMF, respectively¹⁰. Other risk factors for progression to MPN-BP include clinicopathologic features, previous treatment, laboratory values, genetic features of disease, and prognostic scoring systems (Table 1). The mutational landscape of MPN -AP/BP is complex, and in addition to the canonical driver mutations in JAK2, MPL and CALR, includes acquisition of co-mutations (Table 2) which segregate along various pathways including DNA methylation, chromatin modification, DNA repair and RNA splicing11, 12, 13. The evolutionary pathway to MPN-BP is far from linear however, and in some instances, evidence has pointed towards bi-clonal disease or the existence of a pre-JAK2 ancestral clone14, 15, 16. Acquisition of high molecular risk mutations including ASXL1, EZH2, IDH1/2, SRSF2, and TP53 coupled with worsening clinical parameters including increasing circulating blasts and worsening thrombocytopenia can point towards those patients with advanced MPNs, who are on a trajectory toward MPN-AP/ BP. Early consideration of potentially curative approaches such as allo-HCT is important in these scenarios while in the chronic phase, given the paucity of effective interventions once the disease progresses to MPN AP/BP¹⁷.

Therapeutic interventions range from intensive induction chemotherapy to less intensive approaches such as hypomethylating agent (HMA) based therapies. Much of the experience regarding outcomes with these approaches are retrospective in nature and are largely derived from the treatment of patients in the blast phase. It is likely however, that the general principles and lessons learned in this regard would also be applicable to MPN-AP¹. Outcomes across a number of targeted therapy approaches are summarized in Table 3.

Given the relative chemoresistance of MPN AP/BP, with response rates in the 35-45% range and short remission durations, the benefit of intensive chemotherapy as a therapeutic strategy is limited to those situations where allo-HCT can be utilized for consolidation^1,11,18,19. HMA-based approaches are associated with CR rates in approximately 25% range in most series, with overall response rates in the 40-50% range^1,19. In reported retrospective comparisons, there has been no survival advantage to an intensive chemotherapy approach over a less intensive one, in the absence of consolidative allo-HCT^11,12. Of note, there are no randomized studies comparing these approaches. While JAK inhibition has a well-established role in the management of chronic-phase MPNs, JAK inhibitor monotherapy is largely ineffective for MPN-AP/BP²⁰ and there is no evidence yet that JAK inhibition significantly changes natural history with regard to likelihood of evolution to AP/BP²¹. There have also been several prospective studies evaluating ruxolitinib in combination with HMAs in MPN-AP/BP. Response rates have ranged from 26-45% with median OS of less than one year22, 23, 24. In the absence of randomized trials or a matched historical population, it remains unclear if the addition of ruxolitinib to an HMA offers additional benefit beyond HMA monotherapy.

While allo-HCT is considered the only curative approach for MPN-AP/BP, considerations around the appropriate time for transplantation and which patients may derive the most benefit from allo-HCT continue to evolve. Blast percentage certainly has prognostic implications in the setting of survival outcomes in MPNs, however the impact of blast burden on patients undergoing allo-HCT is less clear. A clinical/molecular score in patients with MF undergoing allo-HCT found no impact of blast percentage on post-transplant outcomes²⁵. In addition, a recent analysis of 35 patients with AP MF that underwent reduced-intensity conditioning (RIC) allo-HCT reported a 5-year OS of 65% which compared favorably to the chronic-phase MF group (n=314) that also underwent RIC allo-HCT²⁶. This suggests that in patients with MPN-AP that are eligible for transplant consideration should be given to proceeding with allo-HCT even if blast burden is not reduced. The impact of blast reduction in MPN-BP prior to allo-HCT is also not well defined^1,17. Some studies have demonstrated that while achievement of a CR prior to allo-HCT may offer some benefit, overall survival is still poor27, 28, 29, 30. Furthermore, the ability to achieve a deep remission prior to allo-HCT with current therapies is quite limited. Molecular factors also have an impact on outcomes after allo-HCT in MPN-BP; in particular the presence of TP53 mutation is associated with poor survival even in patients with blast reduction prior to allo-HCT³¹. These findings starkly highlight the need for prospective novel therapies in this patient population.

The B-cell lymphoma 2 (BCL-2) family of proteins are involved in the regulation of mitochondrial outer membrane permeability and the resultant apoptosis of cells. The family is comprised of BCL-2 along with other proteins such as BCL-xL and MCL-1³². Pre-clinical studies established the reliance of AML cells on BCL-2 for survival and mitochondrial dependence on BCL-2 as an anti-apoptotic mechanism; inhibition of BCL-2, on the other hand, was found to drive apoptosis of AML cells and inhibit growth of AML progenitor cells33, 34, 35. Pre-clinical studies of JAK2-driven hematologic malignancies have demonstrated overexpression of the BCL-2 family of proteins and that resistance to JAK inhibition can be overcome with addition of a BCL-2/BCL-xL inhibitor³⁶. The clinical benefit of the BCL-2 inhibitor venetoclax in combination with azacitidine has been confirmed in AML and the use of venetoclax in combination with intensive chemotherapy is under investigation^37,38. In addition, the use of the BCL-2/BCL-xL inhibitor navitoclax in combination with ruxolitinib for patients with myelofibrosis and lack of response to ruxolitinib has demonstrated promising results in the Phase II setting³⁹. There is much interest in ascertaining the role of venetoclax in MPN-AP/BP; the majority of data is based in retrospective analysis at this time.

There are several retrospective analyses that have investigated venetoclax-based combination therapies in MPN-AP/BP. The initial experience reported by Tremblay et al analyzed 9 patients with MPN-AP/BP (8 MPN-BP, 1 MPN-AP) that received HMA+Venetoclax. CR/CRi rate was 33% and 3 patients proceeded to allo-HCT. The median OS was 4.2 months for this cohort⁴⁰. Masarova et al reported on 31 patients with MPN-BP that received venetoclax-containing regimens at a single center. Fourteen patients were treated in the frontline setting and 17 had R/R disease; 18 patients received an HMA, 5 patients received cladribine and low-dose cytarabine (LDAC), 6 patients received intensive chemotherapy, and 2 patients received an IDH inhibitor in combination with venetoclax. CR/CRi rate was 19% with all responses occurring in the frontline population and 2 patients received an allo-HCT; median OS for the cohort was 4 months. Of note, 83% of patients developed a Grade 3+ infection and the 8-week mortality of the 32-patient cohort was 32%⁴¹. Gangat et al reported on 32 patients with MPN-BP treated with HMA+Venetoclax across 3 centers; 23 received HMA+Venetoclax in the frontline setting and 9 received HMA-Venetoclax in the relapsed/refractory (R/R setting). CR/CRi rate in the frontline patient cohort (n=23) was 44% and 6 patients went onto receive allo-HCT. There was no significant difference in median OS when comparing patients treated with HMA+Venetoclax to a historical population treated with HMA monotherapy (n=26) or intensive chemotherapy (n=69) (8 months vs 5.5 months vs 8 months, respectively; p=0.13)⁴². A multi-center retrospective analysis by King et al analyzed outcomes of 27 patients with MPN-AP/BP treated with venetoclax-containing combination therapies. Twenty-one patients had MPN-BP while 6 patients had MPN-AP at time of treatment; 24 patients received an HMA combined with venetoclax while 3 patients received LDAC. The acute leukemia response-complete (ALR-C) rate was 30% and 5 patients received an allo-HCT; the median OS was 6 months for patients with MPN-BP and median OS of 3.6 patients for those with MPN-AP⁴³.

While BCL-2 inhibition in MPN-AP/BP may hold promise, its current role in management is unclear. An analysis of 80 patients with MPN-AP/BP diagnosed and treated from 2017 onwards found no significant difference in overall survival between patients treated with HMA-Venetoclax in the frontline setting (median OS 0.59 years) when compared to those treated with other HMA-based regimens (median OS 1.1 years)⁶. Given pre-clinical work elucidating the role of BCL-xL and MCL-1 dependency in MPNs and erythroid/megakaryocytic AML, prospective evaluation of navitoclax in MPN-AP/BP is underway (NCT05455294, NCT05222984) and may provide a means of overcoming the resistance to venetoclax seen in MPN-AP/BP^36,44, 45, 46.

Isocitrate dehydrogenase (IDH)1 and 2 are essential enzymes within the tricarboxylic acid (TCA) cycle and catalyze the formation of alpha-ketoglutarate, which is necessary for appropriate epigenetic regulation. Mutations in IDH1 and IDH2 are seen across a variety of malignancies and lead to accumulation of the oncometabolite R-2-hydroxyglutarate (R-2HG)⁴⁷. Accumulation of R-2-HG inhibits alpha-ketoglutarate dependent dioxygenases such as Jumonji C-domain lysine demethylases and ten-11 translocation (TET) enzymes. In myeloid malignancies this promotes leukemogenesis via proliferation and differentiation block of immature hematopoietic cells. In MPNs, both IDH1 and IDH2 mutations are associated with inferior OS and increased risk of progression to MPN-BP¹⁷. Pre-clinical work has demonstrated the efficacy of IDH inhibition as monotherapy and in combination with JAK inhibition in the treatment of IDH-mutated MPNs⁴⁸. The incidence of IDH1/2 mutations is enriched in MPN-AP/BP and ranges from 19-26%, making IDH inhibition an intriguing targeted therapy approach11, 12, 13. Ivosidenib and enasidenib, inhibitors of the IDH1 and IDH2 mutant enzymes respectively, have provided a novel lower-intensity approach in the treatment for IDH-mutated AML49, 50, 51, 52, 53. The efficacy of this approach in AML has spurred investigation in MPN-AP/BP as well. In our single-center analysis of 8 patients with IDH2-mutated MPN-AP/BP treated with enasidenib-based regimens; 7 patients had MPN-BP while 1 had MPN-AP. Six patients were treated in the frontline setting while 2 received treatment in the R/R setting; 7 of these patients received enasidenib monotherapy while one received enasidenib in combination with azacitidine. Using 2012 MPN-BP consensus criteria⁵⁴ the response rate was 75% with an ALR-C rate of 25%. The 1-year OS was 50% and the longest response surpassed 4 years. Differentiation syndrome was seen in 25% of patients. Of note, even in patients with durable responses, there were molecular and histopathologic findings of chronic-phase MPN that persisted, which suggests that IDH inhibition did not address the underlying chronic-phase MPN². Chifotides et al retrospectively analyzed 12 patients with IDH-mutated MPN-BP at their institution (7 IDH1 mutated, 5 IDH2 mutated) that received IDH inhibitor-based therapy. A variety of IDH inhibitor-including approaches were employed including IDH inhibitor monotherapy or in combination with ruxolitinib, venetoclax, HMA therapy, or intensive chemotherapy. Seven patients were treated in the frontline setting while 5 were treated in the R/R setting; 25% of patients achieved a CR and the median OS of the entire cohort was 10 months³. A prospective study of ruxolitinib combined with enasidenib for patients with IDH2-mutated MPN-AP/BP and MF is ongoing (NCT04281498); amongst the 5 patients with MPN-AP/BP treated the CR rate was 40% with the longest duration of treatment being 13 cycles and ongoing⁵⁵. Other prospective efforts investigating IDH inhibition are include utilization of a novel IDH1/2 inhibitor (NCT04603001).