Myelodysplastic Syndrome - Myeloproliferative Neoplasm Overlap Syndromes

Myelodysplastic syndrome/myeloproliferative neoplasm (MDS/MPN) overlap syndromes—disorders that include features of both myelodysplastic syndrome (MDS) and myeloproliferative neoplasm (MPN)—are entities whose diagnosis and management have proved challenging. The MDS/MPN overlap syndromes include the following[1] :

• Myelodysplastic syndrome/myeloproliferative neoplasm –  with SF3B1 mutation and thrombocytosis
• Chronic myelomonocytic leukemia (CMML)
• Juvenile myelomonocytic leukemia (JMML)
• Myelodysplastic syndrome/myeloproliferative neoplasm with neutrophilia
• Myelodysplastic syndrome/myeloproliferative neoplasm, not otherwise specified

The overlap of MDS and MPN features in these syndromes is characterized by presence of cytopenias (due to dysplasia) and increased blood cell counts (due to myeloproliferation)—either or both of which may be present in the same patient. Morphology is an important tool in their diagnosis, but its subjective nature limits its use to an extent. This problem is further compounded by the fact that even with classic MPNs, the morphology evolves over time and may include dysplastic features at the time of progression.

These impediments necessitate the discovery of more objective diagnostic tools—tests for molecular and cytogenetic abnormalities that drive the pathogenesis of these syndromes. Although disease-defining abnormalities have not yet been found, the present knowledge of these aberrations offers better understanding of these neoplasms and can supplement the morphologic and immunophenotypic diagnostic features. The identification of the major genetic pathways involved not only may aid in diagnosis, but also may guide the future development of targeted molecular therapy as well as prognostic markers.

For example, loss of function mutations of the EZH2 gene are seen in around 10% of MDS/MPN cases and are associated with poor prognosis. EZH2 encodes the catalytic subunit of the polycomb repressive complex 2 (PRC2), a histone-modifying enzyme.[2] Histones are proteins that both provide structural support for DNA (DNA wraps around a core of histones to form nucleosomes) and are involved in the regulation of gene expression. Loss-of-function mutations of ASXL1, which encodes a protein that recruits the PRC2 complex to the histones, is a driver event in some cases of MDS/MPN. In fact, ASXL1 mutations are the most common mutations in CMML, seen in around 40% of cases.[3]

The true incidence of somatic mutations in MDS/MPN overlap syndromes remains uncertain, since these syndromes were previously under-diagnosed. MDS/MPN with SF3B1 mutation and thrombocytosis as well as CMML occur with relatively greater frequency than the other overlap neoplasms; consequently, more is known about them. The more rare MDS/MPN overlap syndromes include MDS/MPN with neutrophilia and MDS/MPN, not otherwise specified. For full discussion of MDS/MPN, not otherwise specified, see Pathology of Unclassifiable Myelodysplastic Syndromes.

[4, 5, 6, 7] Myelodysplastic syndrome/myeloproliferative neoplasm with ring sideroblasts and thrombocytosis (MDS/MPN-RS-T), a full entity under the 2016 World Health Organization (WHO) classification,[8] was renamed in the 2022 WHO classification to MDS/MPN with SF3B1 mutation and thrombocytosis, due to evolving understanding of disease biology.[1]

Pathophysiology

SF3B1 mutations are found in 28% of MDS cases overall and in over 80% of cases of with increased ring sideroblasts.  In MDS/MPN with SF3B1 mutation and thrombocytosis, the SF3B1 mutation commonly coexists with a JAK2 V617F mutation (50% to 65% of cases), or less commonly a CALR or MPL mutation (< 10%). The co-occurrence of SF3B1 with an MPN driver mutation strongly supports this diagnosis and likely accounts for its mixed MDS/MPN phenotype.[9]  

SF3B1 is a splicing factor gene. Splicing factor mutations alter splicing in different ways and affect the expression of different genes involved in RNA splicing, protein synthesis, and mitochondrial function, suggesting common mechanisms of action in MDS. Many of these altered regulatory pathways and cellular processes can be linked to the known disease pathophysiology associated with splicing factor mutations in MDS. Dolatshad et al identified aberrantly spliced events associated with clinical variables, such as isoforms that independently predict survival in MDS.[10] The aberrantly spliced target genes associated with SF3B1 and SRSF2 mutations lead to impaired erythroid cell growth and differentiation, which explains the resulting anemia in these patients.

A well-recognized candidate gene for MDS with the ring sideroblast phenotype is the iron transporter gene ABCB7. Marked down-regulation of ABCB7 was first reported in MDS with ring sideroblasts. Conditional gene targeting in mice has shown that ABCB7 is essential for hematopoiesis.[9]

Altered RNA splicing has been suggested as the mechanism underlying the observed phenotypic changes in MDS/MPN syndromes with splicing factor gene mutations, including SF3B1. Thus, the identification of aberrantly spliced target gene mutations, including SF3B1, and the identification of aberrantly spliced target genes in the hematopoietic cells of SF3B1-mutant MDS cases, is important.

Cryptic splicing of genes involved in iron homeostasis and/or hemoglobin synthesis could have a role in the ineffective erythropoiesis in patients with MDS and MDS/MPN and ring sideroblasts who have the SF3B1 mutation, where ABCB7 is often down-regulated. In an integrative analysis of MDS, the strongest association found was between the presence of SF3B1 mutations and marked down-regulation of ABCB7.[10]  Given the strong correlation between SF3B1 mutations and the presence of RS, the data suggested a three-way association among SF3B1 mutation, ABCB7 down-regulation, and the occurrence of ring sideroblasts. Therefore, the marked down-regulation of ABCB7 observed in MDS and MDS/MPN patients with ring sideroblasts has been recognized as an important finding for several years; however, the exact mechanism remains a mystery.

Diagnosis

The WHO diagnostic criteria for MDS/MPN with SF3B1 mutation and thrombocytosis include the following[1] :

• Anemia associated with erythroid lineage dysplasia with or without multilineage dysplasia, ≥15% ring sideroblasts, < 1% blasts in peripheral blood and < 5% blasts in bone marrow

• Persistent thrombocytosis with platelet count ≥450 × 109/L

• Presence of an SF3B1 mutation or, in the absence of SF3B1 mutation, no history of recent cytotoxic or growth factor therapy that could explain the myelodysplastic/myeloproliferative features; the presence of an SF3B1 mutation together with a mutation in JAK2 V617F, CALR, or MPL genes strongly supports the diagnosis of MDS/MPN-RS-T 

• No BCR-ABL1 fusion gene, no rearrangement of PDGFRA, PDGFRB, or FGFR1; or PCM1-JAK2; no (3;3)(q21;q26), inv(3)(q21q26) or del(5q)

• No preceding history of MPN, MDS (except MDS-SF3B1), or other type of MDS/MPN

Treatment

There are no formal guidelines for the management of MDS/MPN with SF3B1 mutation and thrombocytosis. Current management strategies, including response criteria, are extrapolated from related diseases such as low-risk MDS-RS and MPN (essential thrombocytosis). Patnaik and Tefferi offer an algorithm for risk-adapted therapy. Initial management includes supportive care with the use of erythropoiesis-stimulating agents (ESA), transfusions, and iron chelation therapy when necessary.[11]

SF3B1 mutations cause production of tumor growth factor beta (TGFβ) superfamily ligands called aberrant Smad2/3, which alter the receptor for erythropoietin (EPO), rendering it unresponsive to EPO. Thus, these patients do not frequently respond to treatment with ESAs. Luspatercept, a first-in-class recombinant fusion protein, binds to select TGFβ superfamily ligands to reduce aberrant Smad2/3 signaling and enhance late-stage erythropoiesis. By interfering with the signals that suppress red blood cell (RBC) production, the agent improves patients’ ability to manufacture their own RBCs, thereby reducing the need for transfusions.[12, 13]  

ESAs are a standard treatment for patients with MDS and anemia; however, only around one third of patients have an erythroid response to ESAs, suggesting that improving erythropoiesis in MDS might be best achieved by targeting downstream processes independent of EPO regulation. In the PACE-MDS trial, an open-label phase 2 dose-finding study of luspatercept in lower-risk MDS patients, a higher response rate occurred in patients with baseline serum EPO < 200 IU/L, but International Working Group–defined erythroid hematologic improvement was also observed in 58% of patients with EPO of 200 to 500 IU/L and in 43% of patients with EPO > 500 IU/L.[12] Efficacy in patients with higher EPO levels is important, as those patients typically have a poor response to ESAs.

The initial PACE-MDS trial of luspatercept enrolled adult patients with lower-risk MDS with or without RBC transfusion support whose anemia was refractory to ESAs or who were ineligible to receive ESAs. In an extension study, 13 patients with low transfusion burden (< 4 RBC units every 8 weeks) who received higher-dose concentrations of luspatercept "showed sustained increases from baseline in mean hemoglobin for at least 15 months,” the authors report, adding that 11 of these patients (85%) achieved improved RBC levels for a median duration of 8.3 months. Patients with high transfusion burden also had high rates of durable response; 79% achieved resolution of anemia with a median duration of response of 11.6 months,[12]

The MEDALIST trial of luspatercept enrolled 229 adults with low-, very-low-, or intermediate-risk MDS with ring sideroblasts. All the patients had not responded to available anemia drugs or were ineligible for treatment with such drugs, and they required RBC transfusions at least every 1 to 2 months. Most of the patients (95.2%) had previously received ESA therapy, and 206 (90.0%) had an SF3B1 mutation. Overall, 53% of those receiving luspatercept experienced a significant reduction in the number of transfusions they required or an increase in hemoglobin levels even without transfusions, compared with 12% of those receiving placebo. The most common reported adverse effects included fatigue and muscle pain, although it is difficult to determine whether these effects were related to anemia or to the drug itself.[13]

The availability of an effective new drug—particularly relevant to those harboring SF3B1 mutations—is an exciting development and is likely to offer meaningful improvements in quality of life. An analysis of the MEDALIST trial showed statistically significant increases in transfusion independence, hematologic improvement, and clinical benefit, even in a small (n=23) randomized subset  of patients with MDS/MPN-RS-T (according to the WHO 2016 classification).[14] For that reason, MDS/MPN-RS-T was added to the FDA indication for luspatercept.

Limited evidence from case reports and small case series describes the use of lenalidomide to control the elevated platelet counts in patients with MDS/MPN-RS-T. This use of lenalidomide is not FDA approved because of the lack of clinical trials, and unfortunately the rarity of this disease would make prospective trials difficult to perform.[15]

Deregulated Janus kinase 2 (JAK2) activation is central to the pathogenesis of most MPNs, of which essential thrombocytosis (ET) is the most common entity. The disease course is often long and therapy intolerance is not infrequent. Ruxolitinib, a JAK1/JAK2 inhibitor, has demonstrated efficacy in both myelofibrosis and polycythemia vera and is well tolerated. Side effects include predictable cytopenias and an augmented risk of infections. Ruxolitinib has been shown to improve symptom and splenomegaly, and—in a small proportion of cases (13.2%)—to significantly reduce platelet counts in ET patients who were refractory to or intolerant of hydroxyurea.[16] However, the the failure to reduce platelet counts in many cases and the side effect of exacerbating anemia preclude ruxolitinib as a routine choice for most patients with MDS/MPN with mutation and thrombocytosis outside of clinical trials.

Chronic myelomonocytic leukemia (CMML) is a clonal disorder of a bone marrow stem cells that exhibits heterogeneous clinical, hematological, and morphologic features, varying from predominantly dysplasia to predominantly proliferative bone marrow. Monocytosis is a major defining feature.

For decades—beginning with the 1976 and 1982 French-American-British MDS classifications—CMML was treated as a form of MDS, despite the universal recognition that these patients sometimes displayed highly proliferative features that were atypical of MDS.  Many clinical trials of therapies designed for MDS included only small subsets of CMML patients. Inclusion of CMML under the umbrella of MDS for so long has hindered a deeper understanding of the disease that might have resulted from studying it as a separate entity.

CMML is a rare disease, with an estimated annual incidence rate of 4 per million population in the United States. It occurs more commonly in men and rarely in young people: 90% of patients diagnosed with CMML are age 60 or older, very similar to MDS patients. CMML frequently progresses to acute myeloid leukemia (AML), which portends a particularly poor prognosis.

Diagnosis

The 2001 World Health Organization (WHO) classification of leukemias and other hematopoietic neoplasms first separated CMML from MDS and created a distinct category of MDS/MPN overlap syndromes. In 2008, the WHO divided CMML into three subcategories, on the basis of blast proportion: CMML-0 (< 5% blasts), CMML-1 (5-9% blasts), and CMML-2 (10-19% blasts).

The WHO continues to revise its classification, including the recommendation that CMML be classified into two subtypes: myeloproliferative (MP-CMML) with white blood cell count (WBC) ≥ 13 x 109/L and myelodysplastic (MD-CMML) with WBC < 13 x 109/L[1] .  In addition, investigators have identified certain genetic mutations (eg, in SRSF2, ASXL1, CBL, SETBP1, and JAK2) that are not exclusive to CMML but are more commonly found in patients with CMML than in MDS without proliferative features, which can aid diagnosis.

WHO diagnostic criteria for CMML are listed below.[1]

Prerequisite criteria are as follows:

1. Persistent monocytosis ≥0.5 × 10⁹/L and monocytes ≥10% of WBCs in peripheral blood
2. < 20% blasts in peripheral blood and bone marrow aspirate
3. Not meeting criteria for chronic myeloid leukemia or other myeloproliferative neoplasms
4. Not meeting criteria for myeloid/lymphoid neoplasms with tyrosine kinase fusions

Supporting criteria are as follows:

1. Dysplasia in ≥1 myeloid lineage
2. Acquired clonal cytogenetic or molecular abnormality in hematopoietic cells
3. Abnormal partitioning of peripheral blood monocyte subsets

Requirements for diagnosis are as follows:

• Prerequisite criteria must be present
• If monocytosis is ≥ 1 x 10 ⁹/L, one of the supporting criteria must be present
• If monocytosis is ≥ 0.5 x 10 ⁹/L and < 1 x 10 ⁹/L, supporting criteria 1 and 2 must be present

WHO subgrouping criteia are basd on percentage of marrow and blood blasts and promonocytes[1] :

• CMML-1: < 5% blasts in peripheral blood and < 10% blasts in bone marrow
• CMML-2: 5-19% blasts in peripheral blood, 10-19% in bone marrow

In 2018, the European Hematology Association and the European LeukemiaNet released new guidelines for adult CMML diagnosis and management that update prior recommendations from the MDS International Working Group (IWG) in 2000 and 2006.[17] Diagnostic recommendations include the following:

• Complete blood count (CBC), marrow aspirate with routine staining using May-Grünwald or Wright-Giemsa staining, and iron staining is mandatory, while bone marrow biopsy with hematoxylin-eosin and/or Gomori's Silver staining is strongly recommended. Immunohistochemistry can be added including CD34 and monocytic markers such as CD68, CD163, CD14, and CD16. A final bone marrow report should be made that includes both cytology and histology.
• Analysis of peripheral blood monocyte subset distribution by a multiparameter flow cytometry assay to distinguish CMML from reactive monocytosis is recommended, although its diagnostic robustness remains to be validated.
• Cytogenetic analysis is mandatory, with analysis of at least 20 mitoses. If an insufficient number of mitoses is obtained, or if only 1 or 2 metaphases with +8 or −7 are seen, fluorescence in situ hybridization (FISH) analysis with centromeric probes for chromosomes 7 and 8 is recommended.
• Analysis of the  ASXL1, NRAS, RUNX1, and SETBP1 genes is mandatory for risk assessment in patients eligible for transplant. Analysis of a minimum of 20 genes is recommended for patients being considered for active treatments, including transplant, and is suggested in patients  eligible only for palliative therapies and supportive care, to inform prognosis and/or reveal actionable targets. The list includes IDH1, IDH2, NPM1, and FLT3. Mutations of those genes are present in less than 5% of CMML cases but have practical therapeutic implications.
• BCR-ABL1 rearrangement should be excluded in all cases. The very rare cases of monocytosis with eosinophilia should be tested for the cytogenetically cryptic  FIP1L1-PDGFRA rearrangement and other very rare tyrosine kinase fusion genes should be investigated if cytogenetic analysis suggests a relevant rearrangement. Monocytic variants of AML are also important to rule out when CMML2 is suspected. Borderline cases must be analyzed with next-generation sequencing (including NPM1 gene status, which must be urgently assessed when AML is suspected) and flow cytometry.
• CMML should be classified according to the WHO criteria, as revised in 2016, including MDS/MPN-CMML categorization

The European guidelines recommend that all patients have a detailed risk stratification assessment with any of the following CMML-specific models incorporating mutational analysis:

• GFM CMML model ^[18]
• CPSS-mol ^[5]
• Mayo Molecular Model ^[6]  

If mutational profiling is not available, the guidelines recommend any of the clinical CMML-specific scores, including CPSS[7] or MD Anderson Prognostic Score[4] .

Treatment

Allogeneic HCT is the only curative treatment for CMML. Unfortunately, only a minority of CMML patients qualify for this procedure. Complicating matters further, no studies have defined when a transplant is the most appropriate option for CMML. We usually recommend that all patients with high-risk disease who are young and fit enough to be considered transplant candidates be referred for consideration of an HCT.

There are no drugs specifically approved for treatment of CMML. The hypomethylating agents azacitidine and decitabine are approved for MDS, for which the FDA indication included CMML. However, the pivotal trials included few patients with CMML and response rates are low, especially in proliferative forms of the disease. Essentially, all the trials that included patients with CMML used disease response criteria designed for MDS, and there is clear evidence that treatment with hypomethylating agents does not alter the disease biology. In fact, in Europe, the use of these drugs in proliferative CMML is not approved.

Instead, medical treatment of CMML typically targets specific symptoms of the disease, such as cytopenias, splenomegaly, and infections with transfusions, blood cell growth factors, and antibiotics. While these can improve patients’ quality of life, they barely modify disease evolution. Improved understanding of the pathophysiology of CMML will hopefully lead to the exploration of novel targets that potentially would be curative.

Major impediments to conducting CMML-specific drug trials has been a lack of uniform response criteria and the rarity of the entity. Defining response criteria is a challenge because the natural history and prognosis of CMML are poorly understood. About 10 prognostic scoring systems have been proposed for CMML and there are no good data on determining the best timing for transplantation. To fill in the knowledge gaps, a new project has been launched to sequence nearly 1,000 patients with CMML both before and after treatment to try to confirm the prognostic significance of specific mutations and to identify which prognostic scoring system is most reliable.

The European Hematology Association and the European LeukemiaNet guidelines for adult CMML include therapeutic algorithm for lower-risk and higher-risk CMML patients, along with the following treatment recommendations[17] :

• CMML patients without excess of marrow blasts and without (or with only mild asymptomatic) cytopenias or major signs of myeloproliferation may observed without treatment. HCT is recommended for patients younger than age 70 with a donor and no major contraindication to transplant who have higher-risk CMML. Transplant may be considered in selected lower-risk patients with poor prognostic factors that include severe cytopenias, and several poor-risk somatic mutations (especially ASXL1, RUNX1, SETBP1, NRAS). Cytoreductive treatment before HSCT is generally advocated in CMML-2
• Intensive chemotherapy is not recommended in patients with CMML, except in CMML-2 patients as a bridge to HSCT, or when the disease appears to be very close to AML (presence of Auer rods, NPM1 mutation).
• In patients ineligible for transplant, azacitidine should be considered according to its label in MD-CMML with more than 10% blasts. Use of azacitidine in patients with a blast excess >5% and significant cytopenias or myeloproliferation should be considered, preferably within clinical trials. In patients eligible for transplant, it is unclear when hypomethylating agents should be used before transplant, especially compared with intensive chemotherapy.
• Hydroxyurea is recommended in proliferative CMML, in the absence of major cytopenias or excess of marrow blasts. No single WBC count threshold or spleen size can be recommended as being the optimal level to introduce treatment; instead, the decision should be based on the patient's symptoms and comorbidity.
• For thrombocytopenia, a short steroid course may be attempted when a peripheral immune component is suspected, notably in severe thrombocytopenia contrasting with the absence of anemia in the absence of an excess of marrow blasts

Recommendations regarding transfusion are as follows[17] :

• Best practice currently individualizes hemoglobin (Hb) thresholds based on a combination of patient comorbidities (cardiac, respiratory), symptoms at a given Hb concentration, observed symptomatic benefit from a (series of) transfusion episodes and patient preference. 
• Hb < 80 g/L is typically symptomatic and should be considered to trigger RBC transfusion, but no consistent single Hb threshold can be recommended.
• No consistent single Hb target value can be recommended.
• Transfusion frequency should reflect the duration of symptomatic benefit between transfusion episodes. 
• If transfusion frequency and number of units per transfusion episode is steadily increasing, consideration should be given to other causes of anemia (eg, hypersplenism, bleeding, hemolysis) or to disease progression.
• Iron chelation therapy may be considered for RBC-transfusion–dependent patients with CMML in lower-risk categories of specific CMML scoring systems who have a serum ferritin level higher than 1000 ng/mL after transfusion of approximately 25 RBC units, in the absence of patient-related (non-MDS) factors anticipated to reduce life expectancy to < 3 years.

Prognosis

Expected survival for patients diagnosed with CMML is variable. Within 3 to 5 years, 15-30% of unselected CMML patients will progress to AML, at which point survival rates drop to 4.7 months without a hematopoietic cell transplantation (HCT) and 14.3 months with an HCT.

One known prognostic factor for survival is CMML subtype. The dysplastic and proliferative subtypes affect patients very differently: Patients with the dysplastic subtype have low blood counts and their natural history and clinical problems related to marrow failure are more similar to patients with MDS; those with the proliferative subtype have high blood counts and often have constitutional symptoms or symptoms related to an enlarged spleen. Patients with proliferative forms of the disease also have shorter survival and a higher risk of transformation to AML.

Though as the EHA guidelines recognize, a number of prognostic models have been developed and are recommended[4, 5, 6] , Itzykson et al created and validated a relatively simple prognostic score for CMML based on the following[18] :

• ASXL1 mutations - two points
• Age older than 65 years - two points
• WBC count greater than 15 ×10 ⁹/L - 3 points
• Platelet count less than 100 ×10 ⁹/L - two points
• Anemia (hemoglobin < 10 g/dL in female patients, < 11g/dL in male patients) - two points

The point total is used to categorize patients into three risk groups:

• Low risk (0 to 4 points) - Median overall survival (OS) not reached
• Intermediate risk (5 to 7 points) - Median OS 38.5 months
• High risk (8 to 12 points) - Median OS 14.4 months

Median AML-free survival rates for the three groups were 56.0, 27.4, and 9.2 months, respectively.[18]

Juvenile myelomonocytic leukemia (JMML) has historically been categorized as an MDS/MPN overlap syndrome whose molecular pathogenesis has been studied extensively. Recent recognition of its unique disease biology and rare occurrence of dysplastic morphologic features resulted in the WHO reclassifying JMML as a myeloproliferative neoplasm.[1]

Hyper-signalling of the RAS pathway is fundamental to the development of JMML. This phenomenon is caused by somatic or germline mutations in NRAS, KRAS, NF1, PTPN11, or CBL genes in around 90% cases of JMML.[19, 20] These mutations are mutually exclusive.

Hyperactivation of RAS has been seen in several autosomal dominant developmental disorders with germline mutations of these genes, termed RASopathies, in which there is also an increased risk of JMML. Study of these disorders has helped shed light on the pathogenesis of JMML. In 25% of JMML patients, RAS protein is involved in intracellular signal transduction on ligand binding to several receptors, in particular the granulocyte-macrophage colony-stimulating factor (GM-CSF) receptor, thereby mediating cell proliferation.

CBL mutations are seen in 17% cases of JMML. CBL is a ubiquitin ligase that causes proteosomal degradation of several tyrosine kinases, including those in the RAS signalling pathways. The mutations cause loss of function of CBL and may be somatic or germline. A characteristic feature of JMML caused by germline mutations of the CBL gene is the high rate of spontaneous resolution, which supports expectant management in such cases.[21]

Diagnosis

Diagnostic criteria for JMML include clinical and hematologic features plus oncogenetic study results.[8, 1]  The presence of all of the following clinical and hematologic features is mandatory:

• Peripheral blood monocyte count > 1 × 10 ⁹/L
• Blast percentage in peripheral blood and bone marrow < 20%
• Splenomegaly (not always apparent at diagnosis)
• Absence of Philadelphia chromosome ( BCR/ABL rearrangement)
• Absence of KMT2A rearrangement

One of the following genetic findings indicating RAS pathway activation is necessary to make the diagnosis:

• Somatic mutation in PTPN11, KRAS, or NRAS
• Clinical diagnosis of NF1 or germline NF1 mutation
• Germline CBL mutation and loss of heterozygosity of CBL

Approximately 10% of patients have the clinical and hematologic features of JMML but lack an oncogenetic criterion. For diagnosis in those cases, at least 2 of the following criteria must be fulfilled:

• A chromosomal abnormality other than monosomy 7
• Hemoglobin F level increased for age
• Myeloid precursors in peripheral blood
• Spontaneous growth or granulocyte-macrophage colony-stimulating factor hypersensitivity in colony assay
• Hyperphosphorylation of STAT5

Treatment

No consistently effective therapy is available for JMML. Historically, more than 90% of patients have died despite the use of chemotherapy. Advances in the understanding of the molecular mechanisms in JMML have led to consideration of novel therapeutic agents; for example, the DNA methylation inhibitor azacitidine has shown benefit, and is undergoing clinical trials.[22, 23]

Allogeneic hematopoietic stem cell transplantation (HSCT) remains the most effective therapy for JMML, resulting in cure in more than 50% of patients. Relapse represents the main cause of treatment failure. Prompt HSCT is recommended In all cases of JMML with NF1, somatic PTPN11 and KRAS mutations, and for most children with somatic NRAS mutations. However, spontaneous resolution has been reported in some cases of JMML with germline CBL mutations and specific somatic NRAS mutation, and in patients with Noonan syndrome, so a 'watch and wait' strategy is appropriate in such cases.[22]