Proposed Criteria for Response Assessment in Patients Treated in Clinical Trials for Myeloproliferative Neoplasms in Blast Phase (MPN-BP); Formal Recommendations from the Post-Myeloproliferative Neoplasm Acute Myeloid Leukemia Consortium (Post-MPN AML Consortium)

Abstract

Leukemic transformation (LT) of a myeloproliferative neoplasm (MPN) is associated with a dismal prognosis and no medical therapies have shown a survival improvement in patients with MPN in blast phase (MPN-BP). Effective therapies for the treatment of MPN-BP are a serious unmet need. Consensus response criteria do not exist for the treatment of patients with MPN-BP and this is necessary for the uniformed reporting of treatment response in clinical trials. We have identified relevant MPN and MPN-BP features in order to define treatment response categories that reflect hematological, clinical, pathological, cytogenetic and molecular changes after therapeutic intervention. We plan to validate these proposed response criteria within multi-centered clinical trials.

Introduction

The chronic myeloproliferative neoplasms (MPNs) are a heterogeneous group of malignancies of hematopoietic stem cell origin which are associated with increased risk for progression to acute myeloid leukemia (AML). Essential thrombocythemia (ET), polycythemia vera (PV), and primary myelofibrosis (PMF) are classified by the World Health Organization as Philadelphia chromosome negative MPNs [1]. The International Working Group for Myelofibrosis Research and Treatment (IWG-MRT) has proposed nomenclature for patients with ET, PV and PMF with peripheral or bone marrow blast percentages of 11–19% as accelerated phase disease (Post PV/ET AP, PMF-AP, respectively), or with ≥ 20% blasts as blast phase disease (Post-PV/ET MF-BP and PMF-BP, respectively) [2]. Post-MPN acute leukemia termed MPN-BP will collectively be referred to as ET, PV, PMF, or Post PV/ET in blast phase. The risk of leukemic transformation (LT) for patients with ET, PV and MF is approximately 1–3%, 5–15%, 10–20% respectively, and the risk for LT is increased by exposure to cytotoxic chemotherapy [3–8]. The median survival for patients with MPN-BP is approximately 3–5 months from time of diagnosis of LT [5]. There is no established standard therapeutic approach for these transformed leukemias and induction chemotherapy has not been shown to improve survival in patients with MPN-BP who are fit for intensive therapy [5]. Presently, there are limited published prospective therapeutic trials in patients with MPN-BP, and only a few retrospective reviews of institutional experience [5, 9–12]. There is a serious and urgent unmet need for well designed clinical trials with innovative approaches for the treatment of patients with MPN-BP.

Several sets of response assessment based on consensus criteria are used for the evaluation of therapeutic response in MPN trials and separate response criteria have been adopted for AML [13–16]. However, no validated consensus criteria exist for the assessment of treatment response in MPN-BP and there are several reasons to suspect that the existing response criteria for MPN and AML are not well-suited to assess therapeutic response of MPN-BP. While the recent MPN response criteria proposed by the European LeukemiaNet have incorporated molecular genetics as a response measure in addition to previously described clinical measures such as improvements in peripheral blood counts, splenomegaly and marrow fibrosis, no existing MPN response assessment addresses the elevated blast percentage associated with MPN-BP [13–15]. Conversely, the established response criteria for AML target the elevated blasts that are the hallmark of the disease, but do not address any of the clinical or unique cytogenetic or molecular genetic aspects of the underlying MPN [16]. Since effective therapy of MPN-BP might well be expected to lead to clinical improvement in both the leukemic and the MPN components of the disease, a more comprehensive and specific response assessment is needed for MPN-BP. In order to standardize response assessment in the treatment of MPN-BP, we have set out to propose response criteria for the purpose of evaluating therapeutic strategies in the setting of clinical trials that would allow for cross-study comparisons.

Methods

In 2011, a group of clinical and laboratory investigators from Memorial Sloan-Kettering Cancer Center, Mount Sinai School of Medicine, University of Pennsylvania, Weill-Cornell Medical College, and Yale University School of Medicine assembled for a series of meetings to discuss the current understanding of the pathogenesis and treatment of LT in the setting of an antecedent MPN. From these meetings, a newly formed “Post-MPN AML Consortium” was developed in order to establish close collaborations across laboratories working together to characterize molecular pathogenetic mechanisms driving LT in MPN, and to consolidate the efforts of multiple institutions in evaluating experimental therapies within clinical trials. It was quickly realized and decided that the group would need to establish viable and clinically relevant response criteria for the purpose of adequately and uniformly determining therapeutic response across the participating institutions and ultimately for the purpose of reporting our findings.

A modified version of the Delphi method was employed in the construction of these treatment response criteria. The investigators/authors are recognized experts in either laboratory or clinical investigation in the field of MPN and/or AML and are members of the Post-MPN AML Consortium. Each investigator had an opportunity to propose criterion known to be critical to response assessment from existing established response criteria for AML and MPN [13–16]. These were collectively reviewed by the group in terms of clinico-pathologic relevance, feasibility, and practicality within the context of a clinical trial for MPN-BP. This process took three rounds of review and revision with the final goal to construct response categories that were believed to hold prognostic significance and could be validated in anticipated clinical trials. Ultimately, remaining differences were reviewed by conference call and solidarity of belief in the criteria was reached by the group.

We identified five clinically relevant aspects of MPN and the associated transformed leukemia that should be considered in response assessment. These include changes in the hematologic profile, bone marrow morphology and degree of fibrosis, spleen size, karyotype and disease associated molecular markers. Aspects of the IWG criteria for AML, MDS and MF have been incorporated into the development of these response criteria [13, 16, 17]. Reports of acute leukemia remission with “re-setting” of chronic phase MPN bone marrow and peripheral blood features have been reported [5, 18]. These response categories take into account that “remission” can occur in terms of the acute leukemia and (but not necessarily) the underlying MPN. In addition, as with the MPNs, lack of progression may be associated with clinical benefit, and stable disease (SD) has been incorporated into the response assessment [13–15].

The response categories are listed in Table 1 and include complete molecular response (CMR) which implies that a complete remission in all five aspects of response assessment are achieved for both the MPN and acute leukemia; complete cytogenetic response (CCR) is the same as CMR except the MPN or leukemia associated molecular markers remain detectable; acute leukemia response-complete (ALR-C) indicates a complete remission of leukemia with residual MPN features such as splenomegaly and MPN-associated cytogenetic and molecular abnormalities; acute leukemia response-partial (ALR-P) indicates a decrease in leukemic burden but with residual blasts in the bone marrow or peripheral blood; progressive disease (PD) is defined as progression of the leukemia or MPN as defined by an increase in peripheral or bone marrow blast count or increase in spleen size and the acquisition of new cytogenetic or molecular abnormalities alone does not signify progressive disease; stable disease (SD) means failure to achieve at least ALR-P and no evidence of PD in either the MPN or leukemia.

Special attention should be given to the fact that patients with MPN or MPN-BP often do not have a bone marrow that can be aspirated and, therefore, the peripheral blood often becomes the only available source of hematopoietic cells for cytogenetic and molecular studies.

Peripheral blood

Evaluation of the hematologic profile includes the complete blood count (CBC), manual review of the peripheral blood smear with attention to the peripheral blast count, leukocyte differential, calculation of the absolute neutrophil count (ANC), and documentation of abnormal erythrocyte morphology.

Spleen

Evaluation of splenomegaly and the response to therapeutic intervention can be measured by palpation on physical exam with measurement in centimeters (cm) below the mid-clavicular line at the left costal border and/or three dimensional splenic volume by cubic centimeter (cc) preferably by magnetic resonance imaging (MRI) or if not available or if contraindicated, by computed tomography (CT). Change in spleen length/volume is represented by percent change from baseline (e.g. at screening) and can be assessed by palpation (length) or imaging (volume).

Bone marrow

Trephine bone marrow biopsy and aspiration are typically performed at the posterior iliac crest and interpretation of the specimens should include overall cellularity, a description of the myeloid, erythroid and megakaryocytic lineages with attention to dysplastic features; trichrome stain for collagen and reticulin stain for reticulin. In cases where an adequate aspirate is obtained, myeloblast percentage will preferably be determined by differential count by manual review of an aspirate smear. In the absence of an adequate aspirate, myeloblast percentage as determined by immunohistochemical staining (IHC) of the marrow biopsy specimen (i.e. by correlating discernable cytologic features with expression of CD34 and/or c-Kit/CD117) will be substituted. The degree of marrow fibrosis (collagen and reticulin) should be reported in terms of the European consensus criteria (grade 1–4) and/or the modified Bauermeister scale (grade 0–4) [19, 20].

Karyotype

Conventional cytogenetics and fluorescence in situ hybridization (FISH) should preferably be obtained from a bone marrow aspirate sample, but can reliably be substituted by peripheral blood due to high concordance of these two tissue samples in patients with MF/MF-BP. In 170 unique MF patients where peripheral blood and bone marrow samples were available, a concordance rate of 92% between bone marrow and peripheral blood cytogenetics and FISH was observed (V. Najfeld unpublished results). The MPN interphase FISH panel should include probes for the 12 most common MPN cytogenetic abnormalities: trisomy 1q (1q21or 1q25 loci), deletion 5q (EGR1), monosomy 7 (D7Z1)/deletion 7q (D7S25 ), trisomy 8(D8Z2), trisomy 9p (9p21), trisomy 9 (CEP9), deletion 11 (ATM at 11q22.3), deletion of 17p13.1 (p53), deletion 13q (RB1), deletion 20q(D20S108) and alterations involving 5q32-3 (PDGFRB at 5q32–33).

Molecular

Recent studies have delineated the spectrum of genetic abnormalities that occur in LT after MPN, including recurrent somatic mutations and copy number alterations [21–23]. The most common mutations seen in patients with LT after MPN include JAK2V617F mutations, IDH1/2 mutations, ASXL1 mutations, TET2 mutations, TP53 mutations, and SRSF2 mutations; however with the possible exception of TP53 mutations these lesions are not specific for the transition from chronic phase MPN to MPN-BP [24]. As a result, these molecular markers cannot currently be used to assess differential molecular response of the leukemic clone and of the chronic-phase MPN to therapeutic interventions, and rather can serve as a marker of clonal disease manifesting as residual MPN, residual LT, or both entities. Notably, recent studies suggest that a subset of these mutations, including in SRSF2, predict for adverse outcome in LT after MPN, suggesting that MPN-BP represents a molecularly heterogeneous entity with disparate biological and clinical features in patients with different genotypes [21].

Discussion

The creation of consensus criteria for the assessment of clinical response in patients with myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML) has allowed for more standardized reporting and has facilitated the evaluation of clinical, hematological, pathological and cytogenetic responses in clinical trials with agents of varying mechanisms of action [16, 17]. We set out to better define meaningful responses in the treatment of MPN-BP in the same manner and here propose response criteria that can be used in assessment of new therapies within the context of clinical trials.

The genetic evolution of MPN-BP is most likely a dynamic process that involves the acquisition of multiple mutations and epigenetic alterations that allow for clonal selection. The role of continued JAK2V617F-mediated constitutive signaling in the transformation of chronic MPN to MPN-BP is not clear given the variable maintenance of the JAK2 mutation in patients with LT from a JAK2-positive MPN; moreover recent studies suggest the risk of LT may be related to the mutational burden in a given individual [3, 8, 25–30]. Specifically, the leukemic clone in MPN-BP is in some cases JAK2V617F negative in patients with an antecedent JAK2V617F-positve MPN [31, 32]. More recently, additional genetic and epigenetic lesions have been identified in chronic and blast phase MPN and are being evaluated in the context of molecular pathogenesis [23, 24, 33–35]. These include and are not limited to mutations in JAK2, MPL, TET2, SRSF2, LNK, ASXL1, IDH1/2, IKZF1, p53, and epigenetic alterations affecting DNA methylation, histone modifications and micro RNA (MiR) expression. The exact role of these molecular markers on the progression from chronic phase MPN to blast phase has yet to be determined. The proposed response criteria presented here account for the depth of therapeutic response (CMR versus CCR) and will potentially allow investigators to better define a population of patients more likely to obtain long term disease free survival as defined by the loss of minimal residual disease, for example, as can be seen after hematopoietic stem cell transplant (HSCT). It will be imperative to incorporate these novel biomarkers within correlative studies in future clinical trials to determine not only the prognostic significance of these lesions in LT, but to also assess their potential to predict response to various therapies including tyrosine kinase inhibitors and/or chromatin modifying agents

There are limitations of our proposed criteria. Chromosomal and molecular diagnostics prior to LT will not always be readily available, and it will therefore not be possible to distinguish between complete remission of the leukemic clone and the presence of the residual MPN clone in these cases. Karyotypic abnormalities are identified in 25–30% in PV, ~50% in PMF and 8–10% in ET at diagnosis and have been shown to have prognostic significance in PMF and Post-PV/ET MF and can predict risk for LT [36–38]. The prognostic significance of specific karyotypic abnormalities to predict response to therapy in MPN-BP is not known. Splenomegaly in MPN patients can result from any of multiple causes including long standing portal hypertension, and the persistence of splenomegaly may confound accurate response assessment. Also, despite effective eradication of peripheral blood and marrow blasts associated with the leukemic clone, the spleen may remain enlarged secondary to extramedullary hematopoiesis related to the underlying MPN. The ALR-C and ALR-P response categories allow for the aspect of residual splenomegaly in the setting of clear acute leukemia responses.

HSCT is the only recognized therapeutic option for MPN/MPN-BP that can potentially achieve a CMR/CCR by our criteria, and previous studies have shown it can take many months to document remission by bone marrow histopathology alone [39]. Importantly, patients with LT after MPN who are treated with standard AML chemotherapy regimens almost invariably relapse, and the true clinical significance of achieving remission of the acute leukemia with a residual MPN background is not entirely known. Whether the loss of all molecular markers associated with the leukemic clone and/or MPN clone are required to ensure improved or prolonged survival is not presently known and will require detailed evaluation in clinical trials.

These proposed criteria will require modification in the near future as our understanding of the molecular pathogenesis of MPN-BP matures. Whether achieving a CMR versus a CCR versus ALR-C will predict for different survival outcomes (overall survival or leukemia-free survival) will also have to be validated in prospective trials with agents which show significant efficacy in this difficult to treat malignancy. These response criteria can be used in the setting of experimental therapeutics as well as clinical trials evaluating the role of HSCT in the treatment of MPN-BP. The intent of our proposed treatment response criteria is to improve reporting and derive consistent treatment response criteria within clinical trials in order to optimize evaluation of novel therapeutic approaches in MPN-BP.

Table I.

Myeloproliferative Neoplasm in Blast Phase (MPN-BP) Response Categories

Complete Molecular Response (CMR)
Description	Complete remission of both leukemia and MPN without detectable molecular markers associated with either leukemia or MPN
Hematologic profile	ANC>1000 Hemoglobin>10g/dL Platelets >100 X109/L Absence of leukoerythroblastosis1
Spleen	Non-palpable
Bone Marrow	Cellularity appropriate for age Resolution of abnormal morphology Blasts ≤5%2 ≤ grade 1 marrow fibrosis
Cytogenetics	Normal karyotype3
Molecular markers	Loss of any any previously documented markers associated with either the leukemic or MPN clone4

Complete Cytogenetic Response (CCR)
Description	Complete remission of both leukemia and MPN with detectable molecular markers associated with either leukemia or MPN
Hematologic profile	ANC>1000 Hemoglobin>10g/dL Platelets >100 X109/L Absence of leukoerythroblastosis1
Spleen	Non-palpable
Bone Marrow	Cellularity appropriate for age Resolution of abnormal morphology Blasts ≤5%2 ≤ grade 1 marrow fibrosis
Cytogenetics	Normal karyotype3
Molecular markers	Residual expression of MPN/leukemia associated gene mutations (e.g. JAK2V617F, MPL515L/K)4

Acute Leukemia Response-Complete (ALR-C)
Description	Complete remission of leukemia with residual MPN features
Hematologic profile	Absence of blasts 1
Spleen	<25% increase in spleen size by palpation or imaging if baseline spleen <10cm or <50% if baseline spleen ≥ 10cm
Bone Marrow	Blasts ≤5%2
Cytogenetics	Loss of cytogenetic abnormality associated with leukemic clone, may have persistent abnormality associated with MPN
Molecular markers	Loss of any previously identified markers in leukemic clone, may have persistent molecular markers associated with MPN4

Acute Leukemia Response-Partial (ALR-P)
Description	Decrease in leukemic burden but without resolution of peripheral blood or bone marrow blasts and residual MPN features
Hematologic profile	>50% reduction in blasts
Spleen	<25% increase in spleen size by palpation or imaging if baseline spleen <10cm or <50% if baseline spleen ≥ 10cm
Bone Marrow	>50% reduction in blasts
Cytogenetics	No new abnormalities
Molecular markers	No new abnormalities

Stable Disease (SD)
Description	Failure to achieve at least LR-P, but no evidence of progression for at least 8 weeks.

Progressive Disease (PD)
Description	Progression of leukemia and/or background MPN
Hematologic profile	For patients with 10–20% blasts: ≥50% increase to >20% blasts For patients with >20% blasts: ≥50% increase to > 30% blasts
Spleen	>25% increase in spleen size by palpation or imaging if baseline spleen <10cm and >50% if baseline spleen ≥ 10cm
Bone Marrow	For patients with 5–10% blasts: ≥50% increase to >10% blasts For patients with 10–20% blasts: ≥50% increase to >20% blasts For patients with >20% blasts: ≥50% increase to > 30% blasts
Cytogenetics	Does not apply
Molecular markers	Does not apply

¹absence of peripheral blood blasts by morphologic review of the peripheral smear on two occasions separated by at least 2 weeks

²blast percentage can be assessed by morphologic review of aspirate and in cases of inaspirate marrows, immunohistochemical staining of the marrow for CD34⁺, CD117⁺ is acceptable

³normal karyotype by conventional cytogenetics in peripheral blood or bone marrow aspirate, if a cytogenetic abnormality is detected prior to treatment it must not be identified at time of assessment; if an abnormality is detected at baseline by FISH it must be absent by FISH at time of assessment

⁴absence or loss of evidence of mRNA transcript by quantitative PCR assay performed in a validated laboratory, this will also include any exploratory biomarkers determined to be positive prior to therapy.