Myelodysplastic/Myeloproliferative Neoplasms

Etiology and Pathogenesis

The cause of the MDS/MPN entities is unknown in most cases, and there is no clear understanding of their pathogenesis. In a minority of cases, the disease is related to prior cytotoxic therapy, and these cases should be classified as therapy-related myeloid neoplasms. For the remaining cases, there are no currently recognized cytogenetic or molecular genetic abnormalities specific for any MDS/MPN subtype, although data have accumulated that suggest similar molecular alterations may be shared among them.

In the past, most attention was focused on point mutations of genes encoding proteins involved in the RAS/RAF/MAPK pathway of signal transduction. More recently, remarkable progress in molecular technologies, including single-nucleotide polymorphism array karyotyping, array comparative genomic hybridization, and direct sequencing of candidate genes by sensitive next-generation sequencing technologies, has uncovered an unexpectedly high frequency of acquired uniparental isodisomy (aUPD) in MDS/MPN and identified recurrent alterations of genes previously not suspected of being involved in these neoplasms. In addition, those studies uncovered considerable overlap among these neoplasms and have revealed their unexpected complexity.

Conventional cytogenetics and single-nucleotide polymorphism array (SNP-A) demonstrate chromosome abnormalities in 70% of MDS/MPN patients. Most of these are aneuploidies (trisomy 8, monosomy 7) or deletions (del 7q, del 13q, del 20q), whereas a minority of cases have reciprocal translocations involving diverse tyrosine kinase fusion genes.

Most mutant genes in MDS/MPN fall into four functional classes: signaling, epigenetic, splicing, and transcription.

Signaling mutations result in aberrant activation of proliferative and anti-apoptotic pathways normally induced by growth factors. In addition to tyrosine kinase gene fusions, mutations have been described in colony-stimulating factor 3 receptor ( CSF3R ), downstream cytokine receptor signaling intermediates ( JAK2 , NRAS , KRAS ), and negative regulators of signaling pathways ( PTPN11 , CBL , NF1 ). Mutations involving RAS pathway are present in approximately 85% of JMML and emerged as a defining feature of this myeloid neoplasm. Signaling mutations are detectable in approximately 50% of CMML (including activating mutations of NRAS or KRAS , which are detected in 20% to 60% of patients with CMML and aCML ) and correlate with a myeloproliferative phenotype and enhanced in vitro sensitivity to granulocyte-macrophage colony-stimulating factor (GM-CSF). Up to 80% of patients with MDS/MPN-RS-T have JAK2 V617F mutation and activated JAK-STAT signaling or mutations in MPL . In addition, novel somatic inactivating NOTCH pathway mutations were recently identified in a small fraction of CMML patients. In mice, inactivation of Notch signaling in hematopoietic stem cells results in aberrant accumulation of granulocyte-monocyte progenitors, extramedullary hematopoiesis, and induction of CMML-like disease, which suggests human disease relevance.

Mutations in genes encoding epigenetic regulators and spliceosome machinery are common in MDS/MPN. The most frequently mutated genes are TET2 and ASXL1 , followed by SRSF2 , IDH1/2 , EZH2 , SUZ12 , EED , and UTX . About 50% of CMML patients have mutations involving SRSF2 and a further with 20% exhibiting mutations in other splicing complex genes, SF3B1 , U2AF35 , U2AF65 , and SF3A1 . Furthermore, SF3B1 mutations are present in 72% of patients with MDS/MPN-RS-T. These SF3B1 mutations are not always mutually exclusive and may be accompanied by DNMT3A , JAK2 , ASXL1 , and TET2 mutations. Functionally, disruption of SF3B1 function leads to the formation of ring sideroblasts, although its exact role in malignant transformation remains unclear.

Studies on mutant U2AF35 in model systems indicate global impairment of splicing induction of mRNA surveillance pathway and impairment of growth. Nevertheless, it is unknown if the critical effect of such mutations is indeed global or whether they have an impact on only a small subset of genes.

The RUNX1 gene is mutated in 15% to 40% of CMML patients, often in association with mutated RAS , leading to the notion that alterations in the RAS signal transduction pathway drive the myeloproliferation, whereas mutated RUNX1 or similar abnormalities result in the abnormal cellular development and dysplasia. RUNX1 encodes core binding factor alpha ( CBFa ), which plays a fundamental role for definite commitment of hematopoiesis.

NPM1 and TP53 mutations are uncommon in MDS/MPN. When NPM1 mutations are detected in CMML, usually in CMML-2, an alternative diagnosis of AML with monocytic differentiation should be excluded. Cases confirmed as CMML with mutated NPM1 appear to have a high probability of progression to AML and require aggressive clinical intervention, especially in patients with a high mutation burden.

SET binding protein 1 ( SETBP1 ) was recently identified as a novel oncogene mutated in 25% of aCML cases and less frequently in other MDS/MPNs, including CMML and JMML, and is associated with adverse prognosis . SETBP1 is located on chromosome 18q21.1 and encodes SET binding protein 1. Identical nucleotide alterations resulting in protein changes affecting predominantly residues 858 to 871 have been reported in Schinzel-Giedion syndrome, a rare congenital disorder characterized by mental retardation, distinctive facial features, and multiple congenital malformations. Overexpression of SETBP1 is also associated with aberration of chromosome 7 and poor prognosis in AML and was shown to confer self-renewal capabilities to murine myeloid progenitors through activation of Hoxa9 and Hoxa10.

A small number of patients with CMML reportedly harbor the FLT3 internal tandem duplication mutation ; mutations of PTPN11 , a gene mutated in a substantial proportion of cases of JMML (see later), have been reported as well. Although the JAK2 V617F mutation has been reported in 3% to 13% of cases of CMML and in 0% to 19% of aCML, it is not clear that all the reported cases fulfilled the WHO diagnostic criteria for these disorders.

Although JMML is far less common than CMML, the molecular events that contribute to its pathogenesis are better understood. The major insights into JMML pathogenesis have been largely facilitated by discovery of gene mutations associated with a group of genetic syndromes resulting from germline mutations affecting the RAS/RAF/MAPK pathway. These mutations induce pathologic activation of the pathway, and thus these disorders are grouped together as neuro-cardio-facio-cutaneous syndromes or RASopathies and share common clinical features, including propensity to development of myeloid neoplasms . The most common and well known of those disorders are neurofibromatosis type 1 (NF1), associated with NF1 mutations, and Noonan's syndrome (NS), caused by mutations of PTPN11 . An important early clue to the pathogenesis of JMML was the recognition that JMML hematopoietic progenitor cells show marked hypersensitivity to GM-CSF but not to other growth factors. Because there was no evidence that the GM-CSF receptor is abnormal in JMML, it was reasoned that the pathways of signal transduction from the GM-CSF receptor to the nucleus were likely deregulated, and attention was focused on the downstream pathways activated by the binding of the receptor with its ligand. Knowledge of this hypersensitivity to GM-CSF and the recognized association between some cases of JMML and NF1 converged to direct attention specifically to the RAS signaling pathway. Children with NF1 have an increased predilection (up to 500 times the risk of children without NF1) for development of JMML or other myeloid disorders. For unknown reasons, this risk diminishes with age and disappears as the patient reaches adulthood. Patients with inherited NF1 are deficient in one of the two alleles of the gene NF1 , which encodes neurofibromin, a guanosine triphosphatase–activating protein that downregulates the RAS pathway by hydrolyzing active RAS–guanosine triphosphate to inactive RAS–guanosine diphosphate ( Fig. 48-1 ). Approximately 10% to 25% of children with JMML, many of whom lack the clinical phenotype of NF1, acquire a second hit to the alternative wild-type NF1 allele, which invariably involves deletion of the normal parental allele in familial cases with duplication of the mutant gene through aUPD and results in total loss of neurofibromin and the inability to turn down the RAS signaling pathway. Studies have revealed that up to an additional 25% of patients with JMML have activating point mutations of NRAS or KRAS that lead to an increase in active RAS–guanosine triphosphate. These mutations are largely mutually exclusive of NF1 abnormalities . Although only a small percentage of patients with NS have JMML, somatic mutations of PTPN11 are the most frequent molecular lesion in JMML and occur in children who have no clinical features of NS, a situation similar to that observed with NF1 abnormalities. The gene associated with this disorder is PTPN11 , which encodes the protein tyrosine phosphatase SHP2, another protein that is important in regulating the RAS pathway ( Fig. 48-1 ). The PTPN11 mutation is largely mutually exclusive of mutated NF1 , KRAS , and NRAS.

More recently, homozygous mutations of Casitas B-lineage lymphoma ( CBL ) mutations were identified in an additional 10% to 17% of JMML cases. The initial CBL lesions in the majority of JMML patients occur as a germline event, either inherited in an autosomal dominant fashion or arising spontaneously, similar to NF1 and NS . In nearly all patients with JMML and CBL mutations, the mutant allele is duplicated through aUPD, resulting in both the loss of the wild-type tumor suppressor allele and the gain of the oncogenic mutation.

CBL is an E3 ubiquitin ligase that is known to mark activated receptor and non-receptor tyrosine kinases and other proteins for degradation by ubiquitination, but it also retains important adaptor functions. Multiple mechanisms of Cbl-driven oncogenesis have been proposed that imply a dominant negative function of mutant Cbl and other Cbl proteins as well as a gain of function through activated receptor and non-receptor tyrosine kinases that can no longer be ubiquitinated. A striking and important phenomenon observed in patients with JMML and homozygous CBL mutations that arise as a germline event is the high rate of spontaneous resolution of disease.

Mutations in NF1 , NRAS , KRAS , PTPN11 , and CBL (RAS pathway mutations) allow molecular diagnosis in most JMML patients and have led to recognition of aberrant RAS signaling as linking NF, NS, and other RASopathies to JMML and transient JMML-like disease. It is now generally accepted that JMML is fundamentally a disease of hyperactive RAS signaling, with somatic mutations (superimposed on germline lesions in some instances) in the NF1 , NRAS , KRAS , PTPN11 , and CBL genes found in more than 85% of cases .

Although RAS pathway lesions have traditionally been thought to represent largely mutually exclusive events, a recent whole exome sequencing study identified coexisting mutations in NRAS , KRAS , PTPN11 , CBL , and NF1 in a significant proportion of JMML patients (11.0%) .

Despite the central role of aberrant RAS pathway signaling, a small subset (≈15% of JMML cases) shows no evidence of RAS pathway mutations.

In contrast to CMML, JMML is characterized by a paucity of gene mutations. Recently, whole exome sequencing studies have identified secondary mutations in SETBP1 and JAK3 in approximately 15% of JMML patients. These mutations are presumed to be involved in progression of disease and are associated with poor clinical outcome. However, a subsequent study with droplet digital polymerase chain reaction detected SETBP1 mutations in subclones present at diagnosis in a large proportion of JMML patients who relapse, but they were below the limits of detection for conventional deep sequencing platforms, providing evidence of their presence early in disease evolution and confirmed their association with a dismal clinical outcome. More recently, large-scale whole exome sequencing study detected additional genetic mutations in the RAS pathway apart from the canonical NF1 , KRAS , NRAS , PTPN11 , and CBL alterations and aside from the RAS/MAPK pathway, including recurrent mutations in genes involved in signal transduction ( RRAS , RRAS2 , SH2B3 , SETBP1 ), transcription factors ( GATA2 , RUNX1 , ASXL1 ), splicing machinery ( ZRSR2 ), polycomb repressive complex 2 ( PRC2 ), and methylation ( DNMT3A ), expanding our knowledge of the mutation spectrum of JMML .

MDS/MPN-RS-T, which is characterized by megakaryocytic proliferation with thrombocytosis and by anemia with ring sideroblasts, was considered a provisional entity, designated RARS-T, within the MDS/MPN category in the 2008 fourth edition of the WHO classification. However, more recently, particularly after discovery of its strong association with SF3B1 mutations and often concurrent JAK2 V617F, MPL , or CALR mutations, the current updated WHO classification recognizes this entity as a distinct “overlap” MDS/MPN and to emphasize its overlapping features that distinguish it from the different myelodysplastic syndrome refractory anemia with ring sideroblasts (RARS) changed the name to myelodysplastic/myeloproliferative neoplasm with ring sideroblasts and thrombocytosis (MDS/MPN-RS-T).

Up to 50% of cases of MDS/MPN-RS-T are associated with mutated JAK2 and less common MPL or CALR mutations. Thus, abnormalities that can explain the proliferative component of MDS/MPN-RS-T have, in some cases, been discovered. However, in the few cases of MDS/MPN-RS-T that have been studied to date with in vivo culture systems, the cells do not form endogenous, proliferative colonies, as would be expected in an MPN with either mutated JAK2 or MPL W515L. Instead, they form small colonies similar to those found in MDS, indicating that one or more other genetic defects are present that lead to abnormal cellular development. As mentioned, more recently, SF3B1 mutations were detected in 72% of patients with MDS/MPN-RS-T. Recent study suggests that SF3B1 plays an important role in the regulation of hematopoietic stem cells, but SF3B1 haploinsufficiency itself is not associated with the myelodysplastic phenotype with ring sideroblasts .

In summary, recent advances in molecular technologies and discovery of recurrent somatic genetic mutations in new gene classes in myeloid neoplasms have led to significant insights into pathogenesis of MDS/MPN and dysregulated pathways that could be responsible for a combined or overlap myelodysplastic/myeloproliferative phenotype. Abnormalities in signal transduction pathways ( NRAS and KRAS , CBL , JAK2 , and CSF3R ) could be responsible for proliferative features, whereas MDS-like mutations such as those in the spliceosome complex or transcription factor RUNX1 could be responsible for myelodysplastic features and impaired maturation. However, other combinations of gene mutations, such as that of SRSF2 and TET2 , highly associated with CMML, are not intuitively associated with myeloproliferation. Likewise, JMML is largely associated with alterations of the RAS pathway with only recently identified mutations aside from the RAS pathway, in epigenetic, splicing, polycomb repressive complex 2, and transcription genes; yet JMML is characterized by both myeloproliferative and myelodysplastic features. This suggests that although some genetic combinations seen in MDS/MPN are likely to explain myelodysplastic/myeloproliferative features of those neoplasms, genetic alterations are unlikely to explain the major convergence of the MDS/MPN phenotype in isolation. Deeper insights into the downstream consequences secondary to mutational combinations and their interactions with the bone marrow microenvironment as well as epigenetic mechanisms will be critical to understand the pathogenesis and unique clinical behavior of this group of myeloid neoplasms.

Chronic Myelomonocytic Leukemia

Absolute monocytosis (≥1 × 10 ⁹ /L), with monocytes accounting for 10% or more of the peripheral blood WBCs, is the major defining feature of CMML ( Box 48-2 ). Blasts, including promonocytes (which are considered blast equivalents), account for less than 20% of the cells in the peripheral blood and less than 20% of the nucleated cells in the bone marrow. Conditions with reactive monocytosis must be ruled out, and myeloid proliferations associated with the BCR-ABL1 fusion gene or with rearrangements of PDGFRA , PDGFRB , or FGFR1 and PCM1-JAK2 are excluded from the diagnosis of CMML. Although absolute monocytosis in the blood is required for diagnosis, there is remarkable variation in the other hematologic parameters. Some patients have elevated WBC counts because of prominent neutrophilia in addition to monocytosis, whereas others have normal WBC counts or even leukopenia because of neutropenia. In some cases, myelodysplasia is minimal; in others, all the myeloid lineages have prominent dysplastic changes. However, if convincing myelodysplasia is not present, the diagnosis of CMML can still be made if the other requirements are met and acquired, clonal cytogenetic, or molecular genetic abnormality is present in hematopoietic cells or if the monocytosis has persisted for at least 3 months and other causes of monocytosis have been excluded. Rare patients with a previous diagnosis of MDS or MPN show evolution to a CMML-like phenotype.

The clinical and morphologic features of CMML are heterogeneous and vary along a spectrum from predominantly myelodysplastic to mainly myeloproliferative. Some authorities suggest that CMML can be divided into two subtypes according to WBC count: myelodysplastic CMML, in which the WBC count is less than 13 × 10 ⁹ /L; and myeloproliferative CMML, in which the WBC count is 13 × 10 ⁹ /L or greater. Myelodysplastic CMML may progress to proliferative with progression of disease and gaining of mutations that drive proliferation. Although considered controversial in the past, the subdivision of CMML into dysplastic (WBC <13 × 10 ⁹ /L) and proliferative (WBC ≥13 × 10 ⁹ /L) groups seems to be warranted in light of accumulated clinical and molecular differences between the subtypes, particularly those related to aberrancies in the RAS/MAPK pathway.

However, the WHO does not classify CMML on the basis of leukocyte counts. Because the percentage of blasts including promonocytes does influence prognosis, the 2008 WHO classification divided CMML into two categories, CMML-1 and CMML-2, according to the percentage of blasts plus promonocytes in the blood and bone marrow ( Box 48-2 ). However, a recent study on a large cohort of CMML patients has shown that within the CMML-1 group, patients with less than 5% medullary blasts and less than 2% peripheral blood blasts have a higher median survival and lower risk of progression to AML than in CMML-1 and CMML-2, justifying separation of the CMML-0 subgroup with less than 5% medullary blasts.

The revised WHO classification has adopted those findings. The previously used category of CMML-1 was split into two new subcategories, CMML-0 and CMML-1. Thus, it is currently recommended that CMML be further divided into three subcategories, depending on the number of blasts (plus promonocytes) found in the blood and bone marrow, as follows:

• CMML-0: <2% blasts in the blood; <5% blasts in the bone marrow

• CMML-1: 2% to 4% blasts in the blood; 5% to 9% blasts in the bone marrow

• CMML-2: 5% to 19% blasts in the blood; 10% to 19% in the bone marrow; or Auer rods are present irrespective of the blasts plus promonocytes count

Moreover, some cases of CMML exhibit eosinophilia. Such cases should always be studied for evidence of abnormalities of PDGFRA , PDGFRB , FGFR1 , or PCM1-JAK2 . If any of these abnormalities are found, the case should be reassigned to the subgroup of myeloid neoplasms associated with eosinophilia and rearrangement of one of these specific genes. If none of these abnormalities is found and eosinophils are 1.5 × 10 ⁹ /L or greater, the diagnosis of CMML with eosinophilia is appropriate.

Laboratory Findings

Blood

A review of WBC counts reported in the literature for patients with CMML underscores its variability. The WBC count may range from 2 to 500 × 10 ⁹ /L, with median values usually between 10 and 20 × 10 ⁹ /L. Patients usually have modest thrombocytopenia (80 to 100 × 10 ⁹ /L), but values from 1 to 700 × 10 ⁹ /L are reported. Anemia is usually mild, but hemoglobin values as low as 5 g/dL can occur.

By definition, monocytosis is present in all cases. The reported range of absolute monocytosis is impressive, varying from 1 to greater than 200 × 10 ⁹ /L, but in the majority of patients, monocytes are less than 5 × 10 ⁹ /L. Monocytes account for 10% or more of the WBCs. This percentage is important because in a number of diseases with elevated WBC counts, such as BCR-ABL1 –positive chronic myeloid leukemia (CML), only 1% to 2% monocytes in the leukocyte differential might result in a significant absolute monocytosis. In CMML, the monocytes in the peripheral blood are typically mature, with minimal morphologic abnormality ( Fig. 48-2 ); however, they can exhibit abnormal granulation, unusual nuclear lobation, and delicate nuclear chromatin. When these latter features are present, the cells are best termed abnormal monocytes—monocytes that are atypical and somewhat immature ( Fig. 48-3 ) but lack the features of promonocytes. Promonocytes, which are considered blast equivalents in CMML, are cells with more delicately folded nuclei, fine chromatin, small and indistinct nucleoli, and finely granular cytoplasm. An increase in blasts in CMML may be due to myeloblasts or monoblasts, or both. Monoblasts are large cells with abundant cytoplasm that may contain a few vacuoles or fine granules; they have lacy, delicate nuclear chromatin and one or more nucleoli ( Fig. 48-3 ). Monoblasts form a morphologic continuum with promonocytes, from which it may be difficult to distinguish them, but both cell types are considered together in tallying the number of blasts for classification purposes. Distinguishing myeloblasts, monoblasts, and promonocytes from the more mature “abnormal” monocytes and from normal monocytes is extremely important to distinguish CMML from AML. When blasts (myeloblasts, monoblasts, and promonocytes) account for 5% to 19% of the WBCs in the blood or 10% to 19% of the nucleated cells in the bone marrow, the diagnosis is CMML-2; if 20% or more are present in either location, the diagnosis is AML. The finding of Auer rods in blood or marrow cells also prompts the diagnosis of CMML-2 if blasts in the blood and marrow are less than 20%.

Neutrophils may range from less than 0.5 to nearly 200 × 10 ⁹ /L but are usually normal or only moderately increased in number. Neutrophil precursors (promyelocytes, myelocytes, and metamyelocytes) typically account for less than 10% of the WBCs in the blood at diagnosis. Dysgranulopoiesis, including neutrophils with hypolobated or abnormally segmented nuclei or hypogranular cytoplasm, is usually present in the peripheral blood, but it may be minimal, if present at all, in a substantial minority of cases. It is commonly believed that patients with higher WBC counts have less dyspoiesis than those with lower counts, but some authors have reported that there is no significant relationship between severity of dysplasia and the leukocyte count. Mild basophilia, usually less than 2%, may be present. Eosinophilia may also be seen, and if the eosinophil count is persistently more than 1.5 × 10 ⁹ /L, the diagnosis of CMML with eosinophilia can be made. In all such cases, however, studies for rearrangements of PDGFRA , PDGFRB , FGFR1 , and PCM1-JAK2 should be done; the finding of any one of these rearrangements excludes the diagnosis of CMML and places the case in the subcategory of myeloid neoplasms associated with eosinophilia and abnormalities of PDGFRA , PDGFRB , FGFR1 , or PCM1-JAK2 .

Bone Marrow

The bone marrow biopsy is hypercellular in more than 75% of cases ( Fig. 48-4 ), but normally cellular or even hypocellular specimens may be encountered. Granulocytic proliferation is often the most prominent feature in the biopsy, with a significant increase in the myeloid-to-erythroid ratio ( Fig. 48-4 ); however, erythroid precursors are usually readily identified and, in some cases, even increased in number. The number of megakaryocytes may be increased, normal, or decreased. Up to 75% of patients are reported to have micro-megakaryocytes or megakaryocytes with abnormal nuclear lobation, although in some cases, enlarged megakaryocytes can be found as well. Clustering of megakaryocytes is unusual in CMML.

The number of monocytes required in the bone marrow for the diagnosis of CMML has never been established, and the percentages reported in the literature vary widely. When the biopsy has been well fixed, thinly sectioned, and nicely stained, a proliferation of monocytes may be appreciated ( Fig. 48-4, B ). Immunohistochemical stains performed on the biopsy, such as CD14, CD68R, and CD163, may aid in their identification ( Fig. 48-4 ), although cytochemical stains for non-specific esterase performed on blood and aspirate smears are, in our experience, more reliable ( Figs. 48-5 and 48-6 ). An increase in blasts can often be appreciated in the biopsy. Staining of the biopsy specimen for CD34 may be useful in estimating the blast percentage ( Fig. 48-7 ), but CD34 may not be expressed by monoblasts or promonocytes; therefore, undue reliance should not be placed on the number of CD34 ⁺ cells, and careful morphologic inspection is necessary. Reportedly, in up to 20% of cases (and in an even greater percentage of cases of CMML-2), variably sized nodules of differentiated plasmacytoid dendritic cells, which strongly express CD123, can be found in the biopsy (see Fig. 48-4 ). A mild increase in reticulin fibers has been reported in most cases of CMML, and 30% to 60% of cases (particularly CMML-2) can demonstrate a substantial increase. Increased numbers of lymphocytes and lymphoid nodules may be observed as well.

Cellular bone marrow aspirate smears provide the best material for assessing the number of myeloblasts, monoblasts, promonocytes, and monocytes and for appreciating dysplasia in the various lineages. Cytochemical staining for alpha-naphthyl acetate esterase or alpha-naphthyl butyrate esterase to detect monocytes—either alone or in combination with naphthol AS-D chloroacetate esterase (CAE), which stains primarily neutrophils—is strongly recommended when the diagnosis of CMML is being considered (see Fig. 48-6 ). Dysgranulopoiesis, which is usually present, is more often appreciated in aspirate smears than in the peripheral blood. Dyserythropoiesis, particularly megaloblastoid changes or ring sideroblasts, is reported in about 25% of cases. The abnormal megakaryocyte morphology described in the biopsy can be appreciated in the aspirate as well.

Extramedullary Tissues

Splenic enlargement is frequent and is due to leukemic infiltration of primarily the red pulp by myelomonocytic cells ( Fig. 48-8 ). Trilineage extramedullary hematopoiesis has been reported in some splenectomy specimens from patients with CMML, and numerous foamy macrophages may be seen, particularly when the spleen has been removed as a therapeutic maneuver to relieve thrombocytopenia. Some authors report high mortality and morbidity rates associated with splenectomy in patients with CMML. Lymphadenopathy is seen in a minority of patients, and a biopsy is recommended in such cases because it may indicate extramedullary transformation to acute leukemia. In rare patients with CMML, tumoral proliferations of plasmacytoid dendritic cells, identical to those described in the bone marrow, may be seen in splenectomy or lymph node specimens.

Immunophenotype

By flow cytometric analysis, the leukemic cells express myelomonocytic antigens such as CD33 and CD13, with variable expression of CD14, CD36, and CD64. The monocytes in CMML often exhibit aberrant expression of two or more antigens, including overexpression of CD56; aberrant expression of CD2; and decreased expression of HLA-DR, CD14, CD11c, CD13, CD15, CD64, or CD36. Some of these phenotypic abnormalities, such as decreased expression of CD14, may reflect immaturity of the monocytes. A characteristic increase in the fraction of circulating classical CD14 ⁺ /CD16 ⁻ monocytes has been recently reported in patients with CMML. The maturing neutrophils may also show aberrant phenotypic features, such as asynchronous expression of maturation-associated antigens or aberrant light scatter properties. An increased number of CD34 ⁺ cells or an emerging blast population with an aberrant immunophenotype may herald the onset of transformation to AML; however, as noted previously, the immature monocytic component may not express CD34.

Currently, an 8- to 10-color multiparametric flow cytometry approach with carefully designed antibody panels allows comprehensive characterization of monocyte lineage maturation stages from early monocytic commitment of CD34 ⁺ precursors to late mature monocytes and of the neutrophil maturation pathway from myeloid blasts to mature polynuclear neutrophilic granulocytes. Moreover, abnormal antigen expression in blast, monocytic, and granulocytic compartments, similar to those seen in MDS patients, has been detected in many CMML patients. Comprehensive antibody panels may therefore be helpful in the diagnosis of CMML and in follow-up by detection of phenotypic aberrancies associated with disease progression and regression of aberrancies in response to therapy.

Immunohistochemistry on tissue sections of bone marrow biopsies may facilitate the assessment of cellular components in their architectural context and be helpful in distinguishing CMML from other MPNs and reactive conditions (see Fig. 48-7 ). Both granulocytes and monocytes, including immature forms and blasts, express CD33, which may be demonstrated in paraffin-embedded specimens. Immunostaining for lysozyme may help highlight granulocytic and monocytic components, but neither CD33 nor lysozyme can discern between them. The combined use of CD33 or lysozyme immunohistochemistry and cytochemistry for CAE may facilitate the identification of monocytic cells, which are CD33 and lysozyme positive but CAE negative; in contrast, granulocytic cells are positive for CD33, lysozyme, and CAE. Other markers, such as CD68 (KP1), CD68R (PG-M1), CD11b, CD11c, CD14, CD16, CD56, CD117, CD163, and HLA-DR, are reportedly helpful in assessing the granulocytic and monocytic components of CMML, and some authors have suggested that when a number of these markers are used in combination, the staining pattern may be useful in the differential diagnosis of CMML, aCML, and CML. Clusters of plasmacytoid dendritic cells associated with CMML (particularly CMML-2) can be identified with CD123. In addition, those nodules are positive for antigens normally expressed by reactive plasmacytoid dendritic cells, such as CD2AP, CD4, CD43, CD45RA, CD68/CD68R, CD303, BCL11, and granzyme B. In rare cases, plasmacytoid dendritic cells may aberrantly express other antigens, such as CD2, CD5, CD7, CD10, CD13, CD14, CD15, CD33, or CD56. Consistently, they show a low Ki67 proliferation index. Immunostaining for CD61 or CD42b highlights abnormal megakaryocytes. Staining for glycophorin C by immunohistochemistry may also be helpful in demonstrating erythroid precursors.

Cytogenetics and Genetics

No specific cytogenetic or genetic abnormalities have been identified in CMML. Abnormal karyotypes are reported in only 20% to 40% of cases, and the recurring abnormalities most frequently reported include +8, −7, −5, del(12p), del(20q), and complex karyotypes. Abnormalities involving KMT2A ( MLL ) at 11q23.3 are unusual in CMML; if present, special care should be taken to rule out AML.

Some authors have suggested that patients whose leukemic cells carry an isochromosome 17q have a unique form of MDS/MPN characterized by marked Pelger-Huët–like nuclei and peripheral cytoplasmic vacuolization of neutrophils ( Fig. 48-9 ), often associated with marrow fibrosis, dysmegakaryocytopoiesis, and usually a poor prognosis. However, almost all patients have absolute monocytosis and meet the criteria for CMML, although rare patients may fit into the aCML or MDS/MPN, unclassifiable category. Several currently used prognostic systems include cytogenetic results.

Somatic mutations are frequently found in CMML patients, the most frequent ones being in ASXL1 (35%-40%), TET2 (50%-60%), SRSF2 (40%-50%), RUNX1 (15%), NRAS (11%), and CBL (10%). Other mutations, including JAK2 V617F, EZH2 , and SETBP1 , occur at lower frequency in less than 10% of cases. Overall, about 90% of CMML patients exhibit one or more mutations, and concurrent mutations of TET2 and SRSF2 appear to be highly specific for CMML.

ASXL1 nonsense or frameshift (but not missense) pathogenic variant mutations are associated with aggressive disease behavior and have been incorporated into the prognostic scoring system along with karyotype and clinicopathologic parameters.

SETBP1 mutations in CMML patients are associated with elevated WBC counts, more frequent extramedullary disease, more frequent ASXL1 mutations and less frequent TET2 mutations, and an adverse prognosis, suggesting distinct cooperative molecular events in SETBP1 -mutated CMML.

Detection of pathogenic mutations can be diagnostically helpful in difficult cases, particularly in CMML with a normal karyotype, but should always be used in conjunction with morphologic and immunophenotypic findings and in the appropriate clinical context. Mutations cannot be used as a sole evidence of neoplasia because some of these mutations can occur in healthy patients, with appreciable increase in frequency with age, as so-called clonal hematopoiesis of indeterminate potential.

Other Laboratory Findings

Serum lysozyme levels are usually elevated and parallel the degree of monocytosis in the blood. Polyclonal hypergammaglobulinemia has been reported in 50% to 60% of patients, and rarely, monoclonal proteins may be detected. A positive Coombs test result in the face of no prior transfusion history was reported in almost 20% of patients evaluated in one study.

Differential Diagnosis

The diagnosis of CMML is sometimes difficult, particularly when dysplasia is minimal, the degree of monocytosis is slight, no cytogenetic abnormalities are present, and the duration of the monocytosis is unknown. Other disorders to consider in the differential diagnosis of CMML are listed in Box 48-3 and briefly discussed here.

Box 48-3

Differential Diagnosis of Chronic Myelomonocytic Leukemia

• Reactive monocytosis associated with the following:

  • Infection (tuberculosis, syphilis, subacute bacterial endocarditis)

  • Autoimmune disease (rheumatoid arthritis, systemic lupus erythematosus, ulcerative colitis, polyarteritis)

  • Sarcoidosis

  • Malignant disease (Hodgkin's lymphoma, B- and T-cell lymphomas, carcinoma)

• Chronic myeloid leukemia, BCR-ABL1 positive

• Myeloid/lymphoid neoplasm with rearrangements of PDGFRA , PDGFRB , or FGFR1 or with PCM1-JAK2

• Atypical chronic myeloid leukemia, BCR-ABL1 negative

• Acute myelomonocytic and monocytic leukemia

• BCR-ABL1 negative myeloproliferative neoplasms (e.g., primary myelofibrosis with monocytosis)

• Myelodysplastic syndrome

Reactive Monocytosis

The hallmark of CMML—absolute monocytosis—is a non-specific finding associated with a wide variety of inflammatory and hematopoietic and non-hematopoietic neoplasms, all of which should be considered and excluded before a diagnosis of CMML is rendered. Viral, fungal, protozoal, rickettsial, and mycobacterial infections are commonly accompanied by monocytosis, as are autoimmune diseases and other chronic inflammatory disorders. Monocytosis is also common in patients with lymphoma, particularly Hodgkin's lymphoma, but may be found in other hematopoietic and non-hematopoietic malignant neoplasms as well. The most important steps in distinguishing reactive monocytosis from CMML are a careful review of the clinical history for evidence of an underlying inflammatory or neoplastic disorder; physical examination to determine whether organomegaly is present (which would favor CMML); and inspection of the blood smear for evidence of dysplasia and morphologically abnormal or immature monocytes and for the absence of findings, such as lymphoma cells, that would support a different diagnosis. If, after these steps have been taken, the diagnosis of CMML is still being considered, a bone marrow specimen with appropriate genetic studies should be obtained to corroborate the diagnosis. Flow cytometric studies of the peripheral blood or bone marrow monocytes may provide useful information. The finding of multiple aberrancies, such as overexpression of CD56 and underexpression of myeloid antigens on the same cells, supports the diagnosis of CMML. Nevertheless, reactive monocytes may also show aberrant phenotypes, and additional studies are needed to determine whether specific combinations of immunophenotypic abnormalities are specific for CMML. A prudent approach in all cases is to recognize that reactive monocytosis is more common than CMML. If dysplasia is lacking or minimal in the blood and bone marrow and there is no myeloid-related karyotypic abnormality or other genetic abnormality that clearly defines the process, it is best to give a descriptive diagnosis and to defer a definitive diagnosis until after an observation period of 3 to 6 months to ascertain that the monocytosis is persistent and that no underlying cause has been discovered.

Acute Myelomonocytic and Acute Monocytic Leukemia

Acute leukemia must always be considered in the differential diagnosis of CMML, particularly CMML-2. A bone marrow aspirate and biopsy are crucial in distinguishing between these entities because blasts and promonocytes are usually more prominent in the bone marrow than in the blood. Even in the marrow specimens, the blasts may be irregularly distributed, and inspection of the biopsy and aspirate together yields the most useful information. Moreover, the distinction between monocytes, abnormal (immature) monocytes, promonocytes, and blasts is sometimes difficult, and distinguishing some cases of AML from CMML-2 can be challenging. When the number of blasts plus promonocytes is 20% or more in the blood or bone marrow, the diagnosis is AML rather than CMML. A more difficult issue is the finding of mutated NPM1 in a case in which the diagnosis of CMML is being considered. This occasion is in the setting of CMML-2, and in such cases, close follow-up and aggressive clinical intervention are recommended as mutated NPM1 is generally regarded as an AML-related mutation.

Chronic Myeloid Leukemia, BCR-ABL1 Positive

The distinction between CML, BCR-ABL1 positive and CMML is made on the basis of morphology combined with cytogenetics and molecular genetic studies; the BCR-ABL1 fusion gene is always present in CML and never present in CMML. In rare cases of CML, the BCR breakpoint is in the minor breakpoint cluster region, which leads to the production of the p190 fusion protein, which is smaller than the p210 protein found in almost all cases of CML. The p190 protein is usually associated with Philadelphia (Ph) chromosome–positive acute lymphoblastic leukemia, but rare cases with this breakpoint may initially have CML with a chronic phase that exhibits increased numbers of monocytes, mimicking CMML ( Fig. 48-10 ). Therefore, cytogenetic and genetic testing for the BCR-ABL1 fusion gene is strongly recommended whenever the diagnosis of CMML is considered.

Myeloid Neoplasms Associated With Eosinophilia and Rearrangements of PDGFRB

The initial cases of rearrangement of PDGFRB as well as some subsequently reported had features of CMML with eosinophilia. However, in the revised WHO classification, the finding of this rearrangement as well as of rearrangements of PDGFRA or FGFR1 or PCM1-JAK1 excludes the diagnosis of CMML; such cases are classified according to the specific gene involved.

Atypical Chronic Myeloid Leukemia, BCR-ABL1 Negative

This entity is discussed in detail later in this chapter. Briefly, aCML has many similarities to CMML, but it can be distinguished by the percentage of monocytes in the peripheral blood (rarely >2% to 4% and always <10% in aCML) and by the more severe granulocytic dysplasia in aCML. The use of cytochemistry for non-specific esterase to detect monocytes in the blood and bone marrow can be invaluable in difficult cases. Because the prognosis for aCML is particularly poor, its distinction from CMML is of practical importance. The presence of a BCR-ABL1 fusion gene excludes the diagnosis of both aCML and CMML.

BCR-ABL1 –Negative Myeloproliferative Neoplasms Associated With Monocytosis

Monocytosis has been reported in up to 15% of patients with primary myelofibrosis; in one study, it was an independent variable that predicted a worse outcome, particularly in younger patients. Because CMML may be associated with prominent reticulin fibrosis, its differentiation from primary myelofibrosis and other MPNs can be difficult. In such cases, the bone marrow biopsy finding of clusters of pleomorphic, bizarre megakaryocytes that range in size from small to large with abnormal, bulky nuclei may be the most helpful distinguishing feature; tight clusters of such bizarre megakaryocytes are rarely observed in CMML.

Myelodysplastic Syndrome

Cases of CMML with normal or even low WBC counts and prominent dysplasia may be difficult to distinguish from MDS, particularly MDS with multilineage dysplasia or MDS with excess blasts. As noted previously, some authorities refer to such cases as the myelodysplastic type of CMML, although this designation has not proved to be clinically useful in some studies. Despite the similarities with MDS, if the WHO criteria are carefully applied, the finding of monocytosis of at least 1 × 10 ⁹ /L is sufficient for a diagnosis of CMML rather than MDS.

Atypical Chronic Myeloid Leukemia, BCR-ABL1 Negative

It could be argued that no myeloid disease has a more inappropriate name. Atypical chronic myeloid leukemia implies that the disease is merely an atypical form of CML, BCR-ABL1 positive, but there are significant differences between these disorders ( Table 48-1 ). Importantly, aCML does not have a Ph chromosome or BCR-ABL1 fusion gene, and patients with aCML do not respond to imatinib. Furthermore, although aCML does have myeloproliferative features, including leukocytosis and splenomegaly, it is characterized by remarkable granulocytic and often multilineage dysplasia and, in most series reported to date, an aggressive clinical course. Because it has myeloproliferative and myelodysplastic features, it is appropriate to classify aCML in the broad category of MDS/MPN. The defining features of aCML are listed in Box 48-4 .