National Cancer Institute

Expert-reviewed information summary about the treatment of childhood acute myeloid leukemia, myelodysplastic syndromes, and other myeloproliferative disorders.

This PDQ cancer information summary for health professionals provides comprehensive, peer-reviewed, evidence-based information about the treatment of childhood acute myeloid leukemia and other myeloid malignancies. It is intended as a resource to inform and assist clinicians who care for cancer patients. It does not provide formal guidelines or recommendations for making health care decisions.

This summary is reviewed regularly and updated as necessary by the PDQ Pediatric Treatment Editorial Board, which is editorially independent of the National Cancer Institute (NCI). The summary reflects an independent review of the literature and does not represent a policy statement of NCI or the National Institutes of Health (NIH).

Childhood AML Treatment

Childhood Acute Myeloid Leukemia Treatment

General Information

Fortunately, cancer in children and adolescents is rare, although the overall incidence of childhood cancer has been slowly increasing since 1975. Children and adolescents with cancer should be referred to medical centers that have a multidisciplinary team of cancer specialists with experience treating the cancers that occur during childhood and adolescence. This multidisciplinary team approach incorporates the skills of the primary care physician, pediatric surgical subspecialists, radiation oncologists, pediatric medical oncologists/hematologists, rehabilitation specialists, pediatric nurse specialists, social workers, and others to ensure that children receive treatment, supportive care, and rehabilitation that will achieve optimal survival and quality of life. (Refer to the PDQ Supportive and Palliative Care summaries for specific information about supportive care for children and adolescents with cancer.)

Guidelines for pediatric cancer centers and their role in the treatment of children with cancer have been outlined by the American Academy of Pediatrics. At these pediatric cancer centers, clinical trials are available for most types of cancer that occur in children and adolescents, and the opportunity to participate in these trials is offered to most patients/families. Clinical trials for children and adolescents with cancer are generally designed to compare potentially better therapy with therapy that is currently accepted as standard. Most of the progress made in identifying curative therapies for childhood cancers has been achieved through clinical trials. Information about ongoing clinical trials is available from the NCI website.

Dramatic improvements in survival have been achieved for children and adolescents with cancer. Between 1975 and 2010, childhood cancer mortality decreased by more than 50%. For acute myeloid leukemia, the 5-year survival rate increased over the same time from less than 20% to 68% for children younger than 15 years and from less than 20% to 57% for adolescents aged 15 to 19 years. Childhood and adolescent cancer survivors require close follow-up because cancer therapy side effects may persist or develop months or years after treatment. (Refer to the PDQ summary on Late Effects of Treatment for Childhood Cancer for specific information about the incidence, type, and monitoring of late effects in childhood and adolescent cancer survivors.)

Myeloid Leukemias in Children

Approximately 20% of childhood leukemias are of myeloid origin and they represent a spectrum of hematopoietic malignancies. The majority of myeloid leukemias are acute, and the remainder include chronic and/or subacute myeloproliferative disorders such as chronic myelogenous leukemia (CML) and juvenile myelomonocytic leukemia (JMML). Myelodysplastic syndromes (MDS) occur much less frequently in children than in adults and almost invariably represent clonal, preleukemic conditions.

Acute myeloid leukemia (AML) is defined as a clonal disorder caused by malignant transformation of a bone marrow–derived, self-renewing stem cell or progenitor, which demonstrates a decreased rate of self-destruction as well as aberrant, and usually limited, differentiation capacity. These events lead to increased accumulation in the bone marrow and other organs by these malignant myeloid cells. To be called acute, the bone marrow usually must include greater than 20% leukemic blasts, with some exceptions as noted in subsequent sections.

CML represents the most common of the chronic myeloproliferative disorders in childhood, although it accounts for only 10% to 15% of childhood myeloid leukemia. Although CML has been diagnosed in very young children, most patients are aged 6 years and older. CML is a clonal panmyelopathy that involves all hematopoietic cell lineages. While the white blood cell (WBC) count can be extremely elevated, the bone marrow does not show increased numbers of leukemic blasts during the chronic phase of this disease. CML is nearly always characterized by the presence of the Philadelphia chromosome, a translocation between chromosomes 9 and 22 (i.e., t(9;22)) resulting in fusion of the and genes. Other chronic myeloproliferative syndromes, such as polycythemia vera and essential thrombocytosis, are extremely rare in children.

JMML represents the most common myeloproliferative syndrome observed in young children. JMML occurs at a median age of 1.8 years and characteristically presents with hepatosplenomegaly, lymphadenopathy, fever, and skin rash along with an elevated WBC count and increased circulating monocytes. In addition, patients often have an elevated hemoglobin F, hypersensitivity of the leukemic cells to granulocyte-macrophage colony-stimulating factor (GM-CSF), monosomy 7, and leukemia cell mutations in a gene involved in RAS pathway signaling (e.g., , , , or ).

The transient myeloproliferative disorder (TMD) (also termed transient leukemia) observed in infants with Down syndrome represents a clonal expansion of myeloblasts that can be difficult to distinguish from AML. Most importantly, TMD spontaneously regresses in most cases within the first 3 months of life. TMD blasts most commonly have megakaryoblastic differentiation characteristics and distinctive mutations involving the gene. TMD may occur in phenotypically normal infants with genetic mosaicism in the bone marrow for trisomy 21. While TMD is generally not characterized by cytogenetic abnormalities other than trisomy 21, the presence of additional cytogenetic findings may predict an increased risk for developing subsequent AML. Approximately 20% of infants with Down syndrome and TMD eventually develop AML, with most cases diagnosed within the first 3 years of life. Early death from TMD-related complications occurs in 10% to 20% of affected children. Infants with progressive organomegaly, visceral effusions, and laboratory evidence of progressive liver dysfunction are at a particularly high risk for early mortality.

MDS in children represents a heterogeneous group of disorders characterized by ineffective hematopoiesis, impaired maturation of myeloid progenitors with dysplastic morphologic features, and cytopenias. Although the majority of patients have hypercellular bone marrows without increased numbers of leukemic blasts, some patients may present with very hypocellular bone marrow, making the distinction between severe aplastic anemia and low-blast count AML difficult.

There are genetic risks associated with the development of AML. There is a high concordance rate of AML in identical twins; however, this is not believed to be related to genetic risk, but rather to shared circulation and the inability of one twin to reject leukemic cells from the other twin during fetal development. There is an estimated twofold- to fourfold-risk of fraternal twins each developing leukemia up to about age 6 years, after which the risk is not significantly greater than that of the general population. The development of AML has also been associated with a variety of predisposition syndromes that result from chromosomal imbalances or instabilities, defects in DNA repair, altered cytokine receptor or signal transduction pathway activation, as well as altered protein synthesis.

  • Inherited syndromes
    • Chromosomal imbalances:
      • Down syndrome.
      • Familial monosomy 7.
    • Chromosomal instability syndromes:
      • Fanconi anemia.
      • Dyskeratosis congenita.
      • Bloom syndrome.
    • Syndromes of growth and cell survival signaling pathway defects:
      • Neurofibromatosis type 1 (particularly JMML development).
      • Noonan syndrome (particularly JMML development).
      • Severe congenital neutropenia (Kostmann syndrome).
      • Shwachman-Diamond syndrome.
      • Diamond-Blackfan anemia.
      • Congenital amegakaryocytic thrombocytopenia.
      • germline syndrome (particularly in JMML).
  • Acquired syndromes
    • Severe aplastic anemia.
    • Paroxysmal nocturnal hemoglobinuria.
    • Amegakaryocytic thrombocytopenia.
    • Acquired monosomy 7.
  • Familial MDS and AML syndromes
    • Familial platelet disorder with a propensity to develop AML (associated with germline mutations).
    • Familial MDS and AML syndromes with germline mutations.
    • Familial MDS and AML syndromes with germline mutations.
    • Telomere biology disorders due to a mutation in or (i.e., occult dyskeratosis congenita).

Nonsyndromic genetic susceptibility to AML is also being studied. For example, homozygosity for a specific polymorphism has been associated with an increased risk of infant AML.

Classification of Pediatric Myeloid Malignancies

French-American-British (FAB) Classification for Childhood Acute Myeloid Leukemia

The first comprehensive morphologic-histochemical classification system for acute myeloid leukemia (AML) was developed by the FAB Cooperative Group. This classification system, which has been replaced by the World Health Organization (WHO) system described below, categorized AML into the following major subtypes primarily based on morphology and immunohistochemical detection of lineage markers:

  • M0: Acute myeloblastic leukemia without differentiation. M0 AML, also referred to as minimally differentiated AML, does not express myeloperoxidase (MPO) at the light microscopy level, but may show characteristic granules by electron microscopy. M0 AML can be defined by expression of cluster determinant (CD) markers such as CD13, CD33, and CD117 (c-KIT) in the absence of lymphoid differentiation.
  • M1: Acute myeloblastic leukemia with minimal differentiation but with the expression of MPO that is detected by immunohistochemistry or flow cytometry.
  • M2: Acute myeloblastic leukemia with differentiation.
  • M3: Acute promyelocytic leukemia (APL) hypergranular type. (Refer to the Acute Promyelocytic Leukemia section of this summary for more information on treatment options under clinical evaluation.)
  • M3v: APL, microgranular variant. Cytoplasm of promyelocytes demonstrates a fine granularity, and nuclei are often folded. Same clinical, cytogenetic, and therapeutic implications as FAB M3.
  • M4: Acute myelomonocytic leukemia (AMML).
  • M4Eo: AMML with eosinophilia (abnormal eosinophils with dysplastic basophilic granules).
  • M5: Acute monocytic leukemia (AMoL).
    • M5a: AMoL without differentiation (monoblastic).
    • M5b: AMoL with differentiation.
  • M6: Acute erythroid leukemia (AEL).
    • M6a: Erythroleukemia.
    • M6b: Pure erythroid leukemia (myeloblast component not apparent).
    • M6c: Presence of myeloblasts and proerythroblasts.
  • M7: Acute megakaryocytic leukemia (AMKL).

Other extremely rare subtypes of AML include acute eosinophilic leukemia and acute basophilic leukemia.

World Health Organization (WHO) Classification System

In 2001, the WHO proposed a new classification system that incorporated diagnostic cytogenetic information and more reliably correlated with outcome. In this classification, patients with t(8;21), inv(16), t(15;17), or () translocations, which collectively constituted nearly half of the cases of childhood AML, were classified as . This classification system also decreased the bone marrow percentage of leukemic blast requirement for the diagnosis of AML from 30% to 20%; an additional clarification was made so that patients with recurrent cytogenetic abnormalities did not need to meet the minimum blast requirement to be considered AML.

In 2008, the WHO expanded the number of cytogenetic abnormalities linked to AML classification, and for the first time included specific gene mutations ( and mutations) in its classification system. Such a genetically based classification system links AML class with outcome and provides significant biologic and prognostic information. With new emerging technologies aimed at genetic, epigenetic, proteomic, and immunophenotypic classification, AML classification will likely evolve and provide informative prognostic and biologic guidelines to clinicians and researchers.

  • AML with recurrent genetic abnormalities:
    • AML with t(8;21)(q22;q22), - .
    • AML with inv(16)(p13.1;q22) or t(16;16)(p13.1;q22), .
    • APL with t(15;17)(q24;q21), .
    • AML with t(9;11)(p22;q23), .
    • AML with t(6;9)(p23;q34), .
    • AML with inv(3)(q21;q26.2) or t(3;3)(q21;q26.2), .
    • AML (megakaryoblastic) with t(1;22)(p13;q13), .
    • AML with mutated .
    • AML with mutated .
  • AML with myelodysplasia-related features.
  • Therapy-related myeloid neoplasms.
  • AML, not otherwise specified:
    • AML with minimal differentiation.
    • AML without maturation.
    • AML with maturation.
    • Acute myelomonocytic leukemia.
    • Acute monoblastic and monocytic leukemia.
    • Acute erythroid leukemia.
    • Acute megakaryoblastic leukemia.
    • Acute basophilic leukemia.
    • Acute panmyelosis with myelofibrosis.
  • Myeloid sarcoma.
  • Myeloid proliferations related to Down syndrome:
    • Transient abnormal myelopoiesis.
    • Myeloid leukemia associated with Down syndrome.
  • Blastic plasmacytoid dendritic cell neoplasm.

Histochemical Evaluation

The treatment for children with AML differs significantly from that for acute lymphoblastic leukemia (ALL). As a consequence, it is critical to distinguish AML from ALL. Special histochemical stains performed on bone marrow specimens of children with acute leukemia can be helpful to confirm their diagnosis, although such approaches have been mostly replaced by flow cytometric immunophenotyping. The stains most commonly used include myeloperoxidase, periodic acid-Schiff, Sudan Black B, and esterase. In most cases the staining pattern with these histochemical stains will distinguish AML from AMML and ALL (see below).

Immunophenotypic Evaluation

The use of monoclonal antibodies to determine cell-surface antigens of AML cells is helpful to reinforce the histologic diagnosis. Various lineage-specific monoclonal antibodies that detect antigens on AML cells should be used at the time of initial diagnostic workup, along with a battery of lineage-specific T-lymphocyte and B-lymphocyte markers to help distinguish AML from ALL and bilineal (as defined below) or biphenotypic leukemias. The expression of various cluster determinant (CD) proteins that are relatively lineage-specific for AML include CD33, CD13, CD14, CDw41 (or platelet antiglycoprotein IIb/IIIa), CD15, CD11B, CD36, and antiglycophorin A. Lineage-associated B-lymphocytic antigens CD10, CD19, CD20, CD22, and CD24 may be present in 10% to 20% of AMLs, but monoclonal surface immunoglobulin and cytoplasmic immunoglobulin heavy chains are usually absent; similarly, CD2, CD3, CD5, and CD7 lineage-associated T-lymphocytic antigens are present in 20% to 40% of AMLs. The aberrant expression of lymphoid-associated antigens by AML cells is relatively common but generally has no prognostic significance.

Immunophenotyping can also be helpful in distinguishing some FAB subtypes of AML. Testing for the presence of HLA-DR can be helpful in identifying APL. Overall, HLA-DR is expressed on 75% to 80% of AMLs but rarely expressed on APL. In addition, APL cases with PML-RARA were noted to express CD34/CD15 and demonstrate a heterogeneous pattern of CD13 expression. Testing for the presence of glycoprotein Ib, glycoprotein IIb/IIIa, or Factor VIII antigen expression is helpful in making the diagnosis of M7 (megakaryocytic leukemia). Glycophorin expression is helpful in making the diagnosis of M6 (erythroid leukemia).

Less than 5% of cases of acute leukemia in children are of ambiguous lineage, expressing features of both myeloid and lymphoid lineage. These cases are distinct from ALL with myeloid coexpression in that the predominant lineage cannot be determined by immunophenotypic and histochemical studies. The definition of leukemia of ambiguous lineage varies among studies, although most investigators now use criteria established by the European Group for the Immunological Characterization of Leukemias (EGIL) or the more stringent WHO criteria. In the WHO classification, the presence of MPO is required to establish myeloid lineage. This is not the case for the EGIL classification.

The WHO classification system is summarized in Table 2.

Leukemias of mixed phenotype comprise the following two groups of patients:

Biphenotypic cases represent the majority of mixed phenotype leukemias. B-myeloid biphenotypic leukemias lacking the fusion have a lower rate of complete remission (CR) and a significantly worse event-free survival (EFS) compared with patients with B-precursor ALL. Some studies suggest that patients with biphenotypic leukemia may fare better with a lymphoid, as opposed to a myeloid, treatment regimen, although the optimal treatment for patients remains unclear.

Cytogenetic Evaluation and Molecular Abnormalities

Chromosomal analyses of leukemia (using either conventional cytogenetic methods and/or molecular methods) should be performed on children with AML because chromosomal abnormalities are important diagnostic and prognostic markers. Clonal chromosomal abnormalities have been identified in the blasts of about 75% of children with AML and are useful in defining subtypes with particular characteristics (e.g., t(8;21), t(15;17), inv(16), 11q23 abnormalities, t(1;22)). Leukemias with the chromosomal abnormalities t(8;21) and inv(16) are called core-binding factor leukemias; core-binding factor (a transcription factor involved in hematopoietic stem cell differentiation) is disrupted by each of these abnormalities.

A unifying concept for the role of specific mutations in AML is that mutations that promote proliferation (Type I) and mutations that block normal myeloid development (Type II) are required for full conversion of hematopoietic stem/precursor cells to malignancy. Support for this conceptual construct comes from the observation that there is generally mutual exclusivity within each type of mutation, such that a single Type I and a single Type II mutation are present within each case. Further support comes from genetically engineered models of AML for which cooperative events rather than single mutations are required for leukemia development. Type I mutations are commonly in genes involved in growth factor signal transduction and include mutations in , , , , and . Type II genomic alterations include the common translocations and mutations associated with favorable prognosis (t(8;21), inv(16), t(16;16), t(15;17), , and ). () rearrangements (translocations and partial tandem duplication) are also classified as Type II mutations.

Specific recurring cytogenetic and molecular abnormalities are briefly described below. The abnormalities are listed by those in clinical use that identify patients with favorable or unfavorable prognosis, followed by other abnormalities.

  • t(8;21) (): In leukemias with t(8;21), the ) gene on chromosome 21 is fused with the () gene on chromosome 8. The t(8;21) translocation is associated with the FAB M2 subtype and with granulocytic sarcomas. Adults with t(8;21) have a more favorable prognosis than adults with other types of AML. These children have a more favorable outcome compared with children with AML characterized by normal or complex karyotypes with 5-year overall survival (OS) of 80% to 90%. The t(8;21) translocation occurs in approximately 12% of children with AML.

    Although both and fusion genes disrupt the activity of core-binding factor, which contains RUNX1 and CBFB, cases with these genomic alterations have distinctive secondary mutations. Both subtypes commonly show mutations in genes that activate receptor tyrosine kinase signaling (e.g., , and ); and are the most commonly mutated genes for both subtypes. cases additionally have frequent mutations in genes regulating chromatin conformation (e.g., and ) and genes encoding members of the cohesin complex (approximately 40% and 20% of cases, respectively). Mutations in and and mutations in members of the cohesin complex are rare in leukemias. Data exist (primarily from adults) that these secondary mutations may have prognostic significance, but further study is required to understand their prognostic significance in children. Additionally, mutations may indicate increased risk of treatment failure for patients with core-binding factor AML, although the prognostic significance of mutations may be dependent on the mutant-allele ratio (high ratio unfavorable) and/or the specific type of mutation (exon 17 mutations unfavorable). exon 17 mutations are enriched in patients with fusions when compared with patients with fusions.

  • inv(16) (): In leukemias with inv(16), the gene () at chromosome band 16q22 is fused with the gene at chromosome band 16p13. The inv(16) translocation is associated with the FAB M4Eo subtype. Inv(16) confers a favorable prognosis for both adults and children with AML with a 5-year OS of about 85%. Inv(16) occurs in 7% to 9% of children with AML. As noted above, cases with and cases with have distinctive secondary mutations; secondary mutations are primarily restricted to genes that activate receptor tyrosine kinase signaling (, , and ).
  • t(15;17) (): AML with t(15;17) is invariably associated with APL, a distinct subtype of AML that is treated differently than other types of AML because of its marked sensitivity to the differentiating effects of all- retinoic acid. The t(15;17) translocation leads to the production of a fusion protein involving the retinoid acid receptor alpha and PML. Other much less common translocations involving the retinoic acid receptor alpha can also result in APL (e.g., t(11;17)(q23;q21) involving the gene).
  • Identification of cases with the t(11;17)(q23;q21) is important because of their decreased sensitivity to all- retinoic acid. APL represents about 7% of children with AML.
  • Nucleophosmin () mutations: NPM1 is a protein that has been linked to ribosomal protein assembly and transport as well as being a molecular chaperone involved in preventing protein aggregation in the nucleolus. Immunohistochemical methods can be used to accurately identify patients with mutations by the demonstration of cytoplasmic localization of . Mutations in the NPM1 protein that diminish its nuclear localization are primarily associated with a subset of AML with a normal karyotype, absence of CD34 expression, and an improved prognosis in the absence of -internal tandem duplication () mutations in adults and younger adults.

    Studies of children with AML suggest a lower rate of occurrence of mutations in children compared with adults with normal cytogenetics. mutations occur in approximately 8% of pediatric patients with AML and are uncommon in children younger than 2 years. mutations are associated with a favorable prognosis in patients with AML characterized by a normal karyotype. For the pediatric population, conflicting reports have been published regarding the prognostic significance of an mutation when a - mutation is also present, with one study reporting that an mutation did not completely abrogate the poor prognosis associated with having a - mutation, but with other studies showing no impact of a - mutation on the favorable prognosis associated with an mutation.

  • mutations: Mutations in the gene () occur in a subset of children and adults with cytogenetically normal AML. In adults younger than 60 years, approximately 15% of cytogenetically normal AML cases have mutations in . Outcome for adults with AML with mutations appears to be relatively favorable and similar to that of patients with core-binding factor leukemias. Studies in adults with AML have demonstrated that double-mutant, but not single-allele mutant, AML was independently associated with a favorable prognosis.

    mutations occur in 5% to 8% of children with AML and have been preferentially found in the cytogenetically normal subtype of AML with FAB M1 or M2; 70% to 80% of pediatric patients have double-mutant alleles, which is predictive of a significantly improved survival and similar to the effect observed in adult studies. Although both double- and single-mutant alleles of were associated with a favorable prognosis in children with AML in one large study, a second study observed inferior outcome for patients with single mutations. However, very low numbers of children with single-allele mutants were included in these two studies (only 13 ), making a conclusion regarding the prognostic significance of single-allele mutations in children premature. In newly diagnosed patients with double-mutant AML, in addition to usual family history queries, germline screening should be considered because 5% to 10% of these patients are reported to have a germline mutation.

  • Chromosomes 5 and 7: Chromosomal abnormalities associated with poor prognosis in adults with AML include those involving chromosome 5 (monosomy 5 and del(5q)) and chromosome 7 (monosomy 7). These cytogenetic subgroups represent approximately 2% and 4% of pediatric AML cases, respectively, and are also associated with poor prognosis in children.

    In the past, patients with del(7q) were also considered to be at high risk of treatment failure and data from adults with AML support a poor prognosis for both del(7q) and monosomy 7. However, outcome for children with del(7q), but not monosomy 7, appears to be comparable to that of other children with AML. The presence of del(7q) does not abrogate the prognostic significance of favorable cytogenetic characteristics (e.g., inv(16) and t(8;21)).

    Chromosome 5 and 7 abnormalities appear to lack prognostic significance in AML patients with Down syndrome who are 4 years of age and younger.

  • Chromosome 3 (inv(3)(q21;q26) or t(3;3)(q21;q26)) and overexpression: The inv(3) and t(3;3) abnormalities involving the gene located at chromosome 3q26 are associated with poor prognosis in adults with AML, but are very uncommon in children (<1% of pediatric AML cases).
  • mutations: Presence of a - mutation appears to be associated with poor prognosis in adults with AML, particularly when both alleles are mutated or there is a high ratio of the mutant allele to the normal allele. - mutations also convey a poor prognosis in children with AML. The frequency of - mutations in children is lower than that observed in adults, especially for children younger than 10 years, for whom 5% to 10% of cases have the mutation (compared with approximately 30% for adults). The prevalence of - is increased in certain genomic subtypes of pediatric AML, including those with the fusion gene, of which 80% to 90% have -. Approximately 15% of patients with - have -, and patients with both - and - have a poorer prognosis than do patients with - and without -.

    For APL, - and point mutations occur in 30% to 40% of children and adults. Presence of the - mutation is strongly associated with the microgranular variant (M3v) of APL and with hyperleukocytosis. It remains unclear whether mutations are associated with poorer prognosis in patients with APL who are treated with modern therapy that includes all- retinoic acid and arsenic trioxide.

    Activating point mutations of have also been identified in both adults and children with AML, though the clinical significance of these mutations is not clearly defined.

  • () gene rearrangements: Translocations of chromosomal band 11q23 involving the gene, including most AMLs secondary to epipodophyllotoxin, are associated with monocytic differentiation (FAB M4 and M5). rearrangements are also reported in 5% to 10% of FAB M7 (AMKL) patients. The most common translocation, representing approximately 50% of -rearranged cases in the pediatric AML population, is t(9;11)(p22;q23) in which the gene is fused with the gene. An gene rearrangement occurs in approximately 20% of children with AML. However, more than 50 different fusion partners have been identified for the gene in patients with AML. The median age for 11q23/-rearranged cases in the pediatric AML setting is approximately 2 years, and most translocation subgroups have a median age at presentation of younger than 5 years. However, pediatric cases with t(6;11)(q27;q23) and t(11;17)(q23;q21) have significantly older median ages at presentation (12 years and 9 years, respectively).

    Outcome for patients with de novo AML and () gene rearrangement is generally reported as being similar to that for other patients with AML. However, as the gene can participate in translocations with many different fusion partners, the specific fusion partner appears to influence prognosis, as demonstrated by a large international retrospective study evaluating outcome for 756 children with 11q23- or -rearranged AML. For example, cases with t(1;11)(q21;q23), representing 3% of all 11q23/-rearranged AML, showed a highly favorable outcome with 5-year event-free survival (EFS) of 92%. While reports from single clinical trial groups have variably described more favorable prognosis for cases with t(9;11), in which the gene is fused with the gene, the international retrospective study did not confirm the favorable prognosis of the t(9;11)(p22;q23) subgroup. An international collaboration evaluating pediatric AMKL observed that the presence of t(9;11), which was seen in approximately 5% of AMKL cases, was associated with an inferior outcome compared with other AMKL cases.

    Several 11q23/ ()-rearranged AML subgroups appear to be associated with poor outcome. For example, cases with the t(10;11) translocation are a group at high risk of relapse in bone marrow and the CNS. Some cases with the t(10;11) translocation have fusion of the gene with the - at 10p12, while others have fusion of with at 10p11.2. The international retrospective study found that these cases, which present at a median age of approximately 1 year, have a 5-year EFS in the 20% to 30% range. Patients with t(6;11)(q27;q23) and with t(4;11)(q21;q23) also have a poor outcome, with a 5-year EFS of 11% and 29%, respectively, in the international retrospective study. A follow-up study by the international collaborative group demonstrated that additional cytogenetic abnormalities further influenced outcome of children with translocations, with complex karyotypes and trisomy 19 predicting poor outcome and trisomy 8 predicting a more favorable outcome.

  • t(6;9) (): t(6;9) leads to the formation of a leukemia-associated fusion protein DEK-NUP214. This subgroup of AML has been associated with a poor prognosis in adults with AML, and occurs infrequently in children (less than 1% of AML cases). The median age of children with DEK-NUP214 AML is 10 to 11 years, and approximately 40% of pediatric patients have -. t(6;9) AML appears to be associated with a high risk of treatment failure in children, particularly for those not proceeding to allogeneic stem cell transplantation.
  • t(1;22) (): The t(1;22)(p13;q13) translocation is uncommon (<1% of pediatric AML) and is restricted to acute megakaryocytic leukemia (AMKL). Studies have found that t(1;22)(p13;q13) is observed in 12% to 14% of children with AMKL and evaluable cytogenetics or molecular genetics. Most AMKL cases with t(1;22) occur in infants, with the median age at presentation (4–7 months) being younger than that for other children with AMKL. The translocation is uncommon in children with Down syndrome who develop AMKL. In leukemias with t(1;22), the () gene on chromosome 1 is fused to the () gene on chromosome 22. Cases with detectable fusion transcripts in the absence of t(1;22) have also been reported.

    Controversy exists regarding the prognostic significance of the t(1;22) in pediatric AMKL. In a report from the Berlin-Frankfurt-Münster (BFM) study group of 97 non–Down syndrome AMKL patients, presence of t(1;22) (n = 8) was associated with a significantly inferior outcome (5-year EFS, 38% vs. 53% in other AMKL patients), although all of the observed events in patients with t(1;22) were related to treatment-related mortality. An international collaborative retrospective study with a larger number of t(1;22) cases reported that patients with this abnormality had a 5-year EFS of 54.5% and an OS of 58.2%, similar to the rates for other children with AMKL. In another international retrospective analysis of 153 cases with non–Down syndrome AMKL with samples available for molecular analysis, the 4-year EFS for patients with t(1;22) was 59% and OS was 70%, significantly better than AMKL patients with other specific genetic abnormalities (, , rearrangements, monosomy7).

  • t(8;16) (): The t(8;16) translocation fuses the gene on chromosome 8p11 to on chromosome 16p13. t(8;16) AML occurs rarely in children, and in an international BFM AML study of 62 children, presence of this translocation was associated with younger age at diagnosis (median, 1.2 years), FAB M4/M5 phenotype, erythrophagocytosis, leukemia cutis, and disseminated intravascular coagulation. Outcome for children with t(8;16) AML appears similar to other types of AML. A substantial proportion of infants diagnosed with t(8;16) AML in the first month of life show spontaneous remission, although AML recurrence may occur months to years later. These observations suggest that a policy could be considered in cases of t(8;16) AML diagnosed in the neonatal period if close long-term monitoring can be ensured.
  • t(7;12)(q36;p13): The t(7;12)(q36;p13) translocation involves on chromosome 12p13 and variable breakpoints on chromosome 7q36 in the region of (). The translocation may be cryptic by conventional karyotyping and in some cases may be confirmed only by FISH. This alteration occurs virtually exclusively in children younger than 2 years, is mutually exclusive with () rearrangement, and is associated with a high risk of treatment failure.
  • gene fusions: has been reported to form leukemogenic gene fusions with more than 20 different partners. In the pediatric AML setting, the two most common fusion genes are and , with the former observed in one report in approximately 15% of cytogenetically normal pediatric AML and the latter observed in approximately 10% of pediatric AMKL. AML cases with either fusion gene show high expression of and genes, indicative of a stem cell phenotype.
    • : The fusion gene, which is often cytogenetically cryptic, results from the fusion of NUP98 (chromosome 11p15) with NSD1 (chromosome 5q35). This alteration occurs in approximately 4% to 5% of pediatric AML cases. The highest frequency in the pediatric population is in the 5- to 9-year age group (approximately 8%), with lower frequency in younger children (approximately 2% in children younger than 2 years). - cases present with high WBC count (median, 147 × 10/L in one study). Most AML cases do not show cytogenetic aberrations. A high percentage of cases (80% to 90%) have . A study that included 12 children with - AML reported that although all patients achieved CR, presence of independently predicted for poor prognosis, and children with AML had a high risk of relapse, with a resulting 4-year EFS of approximately 10%. In another study that included children (n = 38) and adults (n = 7) with - AML, presence of both - and - independently predicted for poor prognosis; patients with both lesions had a low CR rate (approximately 30%) and a low 3-year EFS rate (approximately 15%).
    • (also called ): is a recurrent cryptic translocation in pediatric AMKL, accounting for 9% to 10% of AMKL cases and having a median age at presentation of approximately 2 years. This lesion appears to confer a high risk of relapse (36% ± 14%) and poor EFS and OS (36% ± 13% for each).
  • : is a fusion resulting from a cryptic chromosome 16 inversion (inv(16)(p13.3q24.3)) that is present in approximately 2% of pediatric AML. It occurs most frequently in non–Down syndrome AMKL (~15% of patients), but also has been observed in other cytogenetically normal pediatric AML subtypes (~4% of patients). It has been associated with an inferior outcome.
  • mutations: Although mutations in have been identified in 20% to 25% of patients with AML, the prognostic significance of these mutations has not been clearly shown. Mutations in are observed more commonly than mutations in pediatric AML cases. mutations occur with similar frequency for all Type II alteration subtypes with the exception of APL, for which mutations are seldom observed.
  • mutations: Mutations in occur in approximately 5% of AML, but in 10% to 40% of AML with core-binding factor abnormalities. The presence of activating mutations in adults with this AML subtype appears to be associated with a poorer prognosis compared with core-binding factor AML without mutation. The prognostic significance of mutations occurring in pediatric core-binding factor AML remains unclear, although the largest pediatric study reported to date observed no prognostic significance for mutations.
  • mutations: mutations are present in most, if not all, Down syndrome children with either transient myeloproliferative disease or AMKL. mutations are not observed in non–Down syndrome children with AMKL or in Down syndrome children with other types of leukemia. is a transcription factor that is required for normal development of erythroid cells, megakaryocytes, eosinophils, and mast cells. mutations confer increased sensitivity to cytarabine by down-regulating cytidine deaminase expression, possibly providing an explanation for the superior outcome of children with Down syndrome and M7 AML when treated with cytarabine-containing regimens.
  • mutations: WT1, a zinc-finger protein regulating gene transcription, is mutated in approximately 10% of cytogenetically normal cases of AML in adults. The mutation has been shown in some, but not all, studies to be an independent predictor of worse disease-free, event-free, and OS of adults. In children with AML, mutations are observed in approximately 10% of cases. Cases with mutations are enriched among children with normal cytogenetics and -, but are less common among children younger than 3 years. AML cases with are enriched for both - and mutations. In univariate analyses, mutations are predictive of poorer outcome in pediatric patients, but the independent prognostic significance of mutation status is unclear because of its strong association with - and its association with . The largest study of mutations in children with AML observed that children with mutations in the absence of - had outcomes similar to that of children without mutations, while children with both mutation and - had survival rates less than 20%.
  • mutations: Mutations of the gene () have been identified in approximately 20% of adult AML patients, being virtually absent in patients with favorable cytogenetics but occurring in one-third of adult patients with intermediate-risk cytogenetics. Mutations in this gene are independently associated with poor outcome. mutations appear to be very uncommon in children.
  • and mutations: Mutations in and , which code for isocitrate dehydrogenase, occur in approximately 20% of adults with AML, and they are enriched in patients with mutations. The specific mutations that occur in and create a novel enzymatic activity that promotes conversion of alpha-ketoglutarate to 2-hydroxyglutarate. This novel activity appears to induce a DNA hypermethylation phenotype similar to that observed in AML cases with loss of function mutations in . Mutations in and are uncommon in pediatric AML, occurring in 0% to 4% of cases. There is no indication of a negative prognostic effect for and mutations in children with AML.
  • mutations: is the gene encoding the granulocyte colony-stimulating factor (G-CSF) receptor, and activating mutations in are observed in 2% to 3% of pediatric AML cases. These mutations lead to enhanced signaling through the G-CSF receptor, and they are primarily observed in AML with either mutations or with core-binding factor abnormalities ( and inversion 16). The clinical characteristics of and prognosis for patients with mutations do not seem to be significantly different from those of patients without mutations.

    Activating mutations in are also observed in patients with severe congenital neutropenia. These mutations are not the cause of severe congenital neutropenia, but rather arise as somatic mutations and can represent an early step in the pathway to AML. In one study of patients with severe congenital neutropenia, 34% of patients who had not developed a myeloid malignancy had mutations detectable in peripheral blood neutrophils and mononuclear cells, while 78% of patients who had developed a myeloid malignancy showed mutations. A study of 31 patients with severe congenital neutropenia who developed AML or MDS observed mutations in approximately 80%, and also observed a high frequency of mutations (approximately 60%), suggesting cooperation between and mutations for leukemia development within the context of severe congenital neutropenia.

  • miR-155 expression: miR-155 is a microRNA that is normally upregulated in hematopoietic cells and myeloid progenitor cells as part of an inflammatory response but when aberrantly dysregulated and highly expressed, it independently enhances survival and growth factor independence through repression of PU.1. A study of 363 adults with cytogenetically normal AML found that miR-155 was highly associated with induction failure, disease-free survival, and OS. An independent study of children with cytogenetically normal AML (N = 198) similarly found miR-155 to be an adverse factor. Induction failure (54% vs. 17%; < .001), 3-year OS (51% vs. 75%; = .002), and EFS (32% vs. 59%; < .001) were all worse in patients with high miR-155 expression. As with adults, children in this trial who had high miR-155 expression were more likely to have mutations (69%; < .001). Multivariate analyses found that miR-155 maintained an independent adverse impact on these outcome parameters when controlling for age; white blood cell count; and , , and mutations.

Classification of Myelodysplastic Syndromes in Children

The FAB classification of myelodysplastic syndromes (MDS) was not completely applicable to children. Traditionally, MDS classification systems have been divided into several distinct categories based on the presence of the following:

  • Myelodysplasia.
  • Types of cytopenia.
  • Specific chromosomal abnormalities.
  • Percentage of myeloblasts.

A modified classification schema for MDS and myeloproliferative disorders (MPDs) was published by the WHO in 2008 and included subsections that focused on pediatric MDS and MPD. The bone marrow and peripheral blood findings for the myelodysplastic syndromes according to the 2008 WHO classification schema are summarized in Tables 3 and 4.

A pediatric approach to the WHO classification of myelodysplastic and myeloproliferative diseases was published in 2003. A retrospective comparison of the WHO classification to the Category, Cytology, and Cytogenetics system (CCC) and to a Pediatric WHO adaptation for MDS/MPD, has shown that the latter two systems appear to effectively classify childhood MDS better than the more general WHO system. For instance, while refractory anemia with ring sideroblasts is rare in children, refractory anemia and refractory anemia with excess blasts are more common. When such refractory cytopenias with excess blasts (5%–20%) are associated with recurrent cytogenetic abnormalities usually associated with AML, a diagnosis of the latter should be made and treated accordingly.

The WHO classification schema has a subgroup that includes JMML (formerly juvenile chronic myeloid leukemia), CMML, and Ph chromosome–negative CML. This group can show mixed myeloproliferative and sometimes myelodysplastic features. JMML shares some characteristics with adult CMML but is a distinct syndrome (see below). A subgroup of children younger than 4 years at diagnosis with JMML associated with monosomy 7, are considered to have a subtype of JMML characterized by lower WBC, higher percentage of circulating monocytes, higher mean cell volume for red blood cells, a lower bone marrow myeloid to erythroid ratio and often, normal to moderately increased fetal hemoglobin.

The International Prognostic Scoring System is used to determine the risk of progression to AML and the outcome in adult patients with MDS. When this system was applied to children with MDS or JMML, only a blast count of less than 5% and a platelet count of more than 100 x 10/L were associated with a better survival in MDS, and a platelet count of more than 40 x 10/L predicted a better outcome in JMML. These results suggest that MDS and JMML in children may be significantly different disorders than adult-type MDS.

MDS in children with monosomy 7 and high-grade MDS behaves more like MDS in adults and are best classified as adult MDS, as well as treated with allogeneic hematopoietic stem cell transplantation. The risk group or grade of MDS is defined according to International Prognostic Scoring System guidelines.

The diagnostic criteria for childhood myelodysplastic syndrome (refractory cytopenia of childhood [RCC]–provisional entry) include the following:

  • Persistent cytopenia with <5% bone marrow blasts and <2% peripheral blood blasts.
  • Dysplastic changes should be present.

Stage Information

There is presently no therapeutically or prognostically meaningful staging system for these myeloid malignancies. Leukemia is considered to be disseminated in the hematopoietic system at diagnosis, even in children with AML who present with isolated chloromas (also called granulocytic sarcomas). If these children do not receive systemic chemotherapy, they invariably develop AML in months or years. AML may invade nonhematopoietic tissue such as meninges, brain parenchyma, testes or ovaries, or skin (leukemia cutis). Extramedullary leukemia is more common in infants than in older children with AML.

Newly diagnosed

Childhood AML is diagnosed when bone marrow has greater than 20% blasts. The blasts have the morphologic and histochemical characteristics of one of the FAB subtypes of AML. It can also be diagnosed by biopsy of a chloroma. For treatment purposes, patients with clonal cytogenetic abnormalities typically associated with AML, such as t(8:21) (), inv(16)(), t(9;11)() or t(15;17)() and who have less than 20% bone marrow blasts, are considered to have AML rather than myelodysplastic syndrome.

In remission

Remission is defined in the United States as peripheral blood counts (WBC count, differential, and platelet count) rising toward normal, a mildly hypocellular to normal cellular marrow with fewer than 5% blasts, and no clinical signs or symptoms of the disease, including in the CNS or at other extramedullary sites. Achieving a hypoplastic bone marrow is usually the first step in obtaining remission in AML with the exception of the M3 (acute promyelocytic leukemia [APL]); a hypoplastic marrow phase is often not necessary before the achievement of remission in APL. Additionally, early recovery marrows in any of the subtypes of AML may be difficult to distinguish from persistent leukemia, although the application of flow cytometric immunophenotyping and cytogenetic/molecular testing have made this less problematic. Correlation with blood cell counts and clinical status is imperative in passing final judgment on the results of early bone marrow findings in AML. If the findings are in doubt, the bone marrow aspirate should be repeated in 1 to 2 weeks.

Treatment Overview for Acute Myeloid Leukemia (AML)

The mainstay of the therapeutic approach is systemically administered combination chemotherapy. Future approaches involving risk-group stratification and biologically targeted therapies are being tested to improve antileukemic treatment while sparing normal tissues. Optimal treatment of acute myeloid leukemia (AML) requires control of bone marrow and systemic disease. Treatment of the central nervous system (CNS), usually with intrathecal medication, is a component of most pediatric AML protocols but has not yet been shown to contribute directly to an improvement in survival. CNS irradiation is not necessary in patients either as prophylaxis or for those presenting with cerebrospinal fluid leukemia that clears with intrathecal and systemic chemotherapy.

Treatment is ordinarily divided into two phases: (1) induction (to attain remission), and (2) postremission consolidation/intensification. Postremission therapy may consist of varying numbers of courses of intensive chemotherapy and/or allogeneic hematopoietic stem cell transplantation (HSCT). For example, ongoing trials of the Children’s Oncology Group (COG) and the United Kingdom Medical Research Council (MRC) utilize similar chemotherapy regimens consisting of two courses of induction chemotherapy followed by two additional courses of intensification chemotherapy.

Maintenance therapy is not part of most pediatric AML protocols as two randomized clinical trials failed to show a benefit for maintenance chemotherapy. The exception to this generalization is acute promyelocytic leukemia (APL), for which maintenance therapy has been shown to improve event-free survival and overall survival (OS).

Treatment approaches currently used for AML are usually associated with severe and protracted myelosuppression along with other associated complications. Treatment with hematopoietic growth factors (granulocyte-macrophage colony-stimulating factor [GM-CSF] and granulocyte colony-stimulating factor [G-CSF]) has been used in an attempt to reduce the toxic effects associated with severe myelosuppression but does not influence ultimate outcome. Virtually all randomized trials of hematopoietic growth factors (GM-CSF and G-CSF) in adults with AML have demonstrated significant reduction in the time to neutrophil recovery, but varying degrees of reduction in morbidity and little, if any, effect on mortality. (Refer to the Treatment Option Overview for AML section in the PDQ summary on Adult Acute Myeloid Leukemia Treatment for more information.) The BFM 98 study confirmed a lack of benefit for the use of G-CSF in a randomized pediatric AML trial.

Because of the intensity of therapy utilized to treat AML, children with this disease should have their care coordinated by specialists in pediatric oncology and be treated in cancer centers or hospitals with the necessary supportive care facilities (e.g., to administer specialized blood products; to manage infectious complications; to provide pediatric intensive care; and to provide emotional and developmental support). Approximately one-half of the remission induction failures are due to resistant disease and the other half are due to toxic deaths. For example, in the MRC 10 and 12 AML trials, there was a 4% resistant disease rate in addition to a 4% induction death rate. With increasing rates of survival for children treated for AML comes an increased awareness of long-term sequelae of various treatments. For children who receive intensive chemotherapy, including anthracyclines, continued monitoring of cardiac function is critical. Periodic renal and auditory examinations are also suggested. In addition, total-body irradiation before HSCT increases the risk of growth failure, gonadal and thyroid dysfunction, and cataract formation.

Prognostic Factors in Childhood AML

Prognostic factors in childhood AML have been identified and can be categorized as follows:

  • Age: Several reports published since 2000 have identified older age as being an adverse prognostic factor. The age effect is not large, but there is consistency in the observation that adolescents have a somewhat poorer outcome than younger children.

    While outcome for infants with ALL remains inferior to that of older children, outcome for infants with AML is similar to that of older children when they are treated with standard AML regimens. Infants have been reported to have a 5-year survival of 60% to 70%, although with increased treatment-associated toxicity.

  • Race/Ethnicity: In both the Children's Cancer Group (CCG) CCG-2891 and (CCG-2961) studies, Caucasian children had higher OS rates than African American and Hispanic children. A trend for lower survival rates for African American children compared with Caucasian children was also observed in children treated on St. Jude Children’s Research Hospital AML clinical trials.
  • Down syndrome: For children with Down syndrome who develop AML, outcome is generally favorable. The prognosis is particularly good (event-free survival exceeding 80%) in children aged 4 years or younger at diagnosis, the age group that accounts for the vast majority of Down syndrome patients with AML.

    A large study of children with AML and Down syndrome confirmed the prognostic significance of younger age, and it identified the absence of cytogenetic abnormalities (other than trisomy 21), representing approximately 30% of cases, as an independent predictor of inferior OS and EFS.

  • Body mass index: In the (CCG-2961) study, obesity (body mass index more than 95th percentile for age) was predictive of inferior survival. Inferior survival was attributable to early treatment-related mortality that was primarily due to infectious complications. Obesity has been associated with inferior survival in children with AML, primarily caused by a higher rate of fatal infections during the early phases of treatment.
  • White blood cell (WBC) count: WBC count at diagnosis has been consistently noted to be inversely related to survival. Patients with high presenting leukocyte counts have a higher risk of developing pulmonary and CNS complications and have a higher risk of induction death.
  • FAB subtype: Associations between FAB subtype and prognosis have been more variable.
    • M3 subtype. The M3 (APL) subtype has a favorable outcome in studies utilizing all- retinoic acid in combination with chemotherapy.
    • M7 subtype. Some studies have indicated a relatively poor outcome for M7 (megakaryocytic leukemia) in patients without Down syndrome, although reports suggest an intermediate prognosis for this group of patients when contemporary treatment approaches are used. In a retrospective study of non–Down syndrome M7 patients with samples available for molecular analysis, the presence of specific genetic abnormalities ( [cryptic inv(16)(p13q24)], [also called (t(11;12)(p15;p13)], [] rearrangements, monosomy 7) was associated with a significantly worse outcome than for other M7 patients.
    • M0 subtype. The M0, or minimally differentiated subtype, has been associated with a poor outcome.
  • CNS disease: CNS involvement at diagnosis is categorized based on the presence or absence of blasts in cerebrospinal fluid (CSF), as follows:
    • CSF negative for blasts on cytospin, regardless of CSF WBC count.
    • CSF with fewer than five WBC/μL and cytospin positive for blasts.
    • CSF with five or more WBC/μL and cytospin positive for blasts.

    CNS2 disease has been observed in approximately 13% of children with AML and CNS3 disease in 11% to 17% of children with AML. In another study, patients with CNS3 were younger and had a higher incidence of t(9;11), t(8;21) or inv(16).

    The presence of CNS disease (CNS2 and/or CNS3) at diagnosis has not been shown to affect OS; however, it may be associated with an increased risk of isolated CNS relapse.

  • Cytogenetic and molecular characteristics: Cytogenetic and molecular characteristics are also associated with prognosis. (Refer to the Cytogenetic evaluation and molecular abnormalities section in the Classification of Pediatric Myeloid Malignancies subsection of this summary for detailed information.) Cytogenetic and molecular characteristics that are currently used in clinical trials for treatment assignment include the following:
    • : inv(16)/t(16;16) and t(8;21), t(15;17), biallelic mutations, and mutations.
    • : monosomy 7, monosomy 5/del(5q), 3q abnormalities, and - with high-allelic ratio.
  • Response to therapy/minimal residual disease (MRD): Early response to therapy, generally measured after the first course of induction therapy, is predictive of outcome and can be assessed by standard morphologic examination of bone marrow, by cytogenetic analysis, by fluorescence hybridization, or by more sophisticated techniques to identify MRD. Multiple groups have shown that the level of MRD after one course of induction therapy is an independent predictor of prognosis.

    Molecular approaches to assessing MRD in AML (e.g., using quantitative reverse transcriptase–polymerase chain reaction [RT–PCR]) have been challenging to apply because of the genomic heterogeneity of pediatric AML and the instability of some genomic alterations. Quantitative RT–PCR detection of fusion transcripts can effectively predict higher risk of relapse for patients in clinical remission. Other molecular alterations such as mutations and fusion transcripts have also been successfully employed as leukemia-specific molecular markers in MRD assays, and for these alterations the level of MRD has shown prognostic significance. The presence of - has been shown to be discordant between diagnosis and relapse, although when its presence persists (usually associated with a high allelic ratio at diagnosis), it can be useful in detecting residual leukemia.

    For APL, MRD detection at the end of induction therapy lacks prognostic significance, likely relating to the delayed clearance of differentiating leukemic cells destined to eventually die. However, the kinetics of molecular remission after completion of induction therapy is prognostic, with the persistence of minimal disease after three courses of therapy portending increased risk of relapse.

    Flow cytometric methods have been used for MRD detection and can detect leukemic blasts based on the expression of aberrant surface antigens that differ from the pattern observed in normal progenitors. A CCG study of 252 pediatric patients with AML in morphologic remission demonstrated that MRD as assessed by flow cytometry was the strongest prognostic factor predicting outcome in a multivariate analysis. Other reports have confirmed both the utility of flow cytometric methods for MRD detection in the pediatric AML setting and the prognostic significance of MRD at various time points after treatment initiation.

Risk classification systems under clinical evaluation

Risk classification for treatment assignment on the study is based on cytogenetics, molecular markers, and MRD at bone marrow recovery postinduction I, with patients being divided into a low-risk or high-risk group as follows:

The group represents about 73% of patients, has a predicted OS of approximately 75%, and is defined by the following:

  • Inv(16), t(8;21), () mutations, or mutations with any MRD status.
  • Standard-risk cytogenetics (defined by the absence of either low-risk or high-risk cytogenetic characteristics) with negative MRD at end of Induction I.

The group represents the remaining 27% of patients, has a predicted OS less than 35%, and is defined by the following:

  • High allelic ratio --positive with any MRD status.
  • Monosomy 7 with any MRD status.
  • del(5q) with any MRD status.
  • Standard-risk cytogenetics with positive MRD at end of Induction I.

The high-risk group of patients will be offered transplantation in first remission with the most appropriate available donor. Patients in the low-risk group will only be offered transplantation in second complete remission.

Treatment of Newly Diagnosed AML

The general principles of therapy for children and adolescents with acute myeloid leukemia (AML) are discussed below, followed by a more specific discussion of the treatment of children with acute promyelocytic leukemia (APL) and Down syndrome.

Overall survival (OS) rates have improved over the past three decades for children with AML, with 5-year survival rates now in the 55% to 65% range. Overall remission-induction rates are approximately 85% to 90%, and event-free survival (EFS) rates from the time of diagnosis are in the 45% to 55% range. There is, however, a wide range in outcome for different biological subtypes of AML (refer to the Cytogenetic Evaluation and Molecular Abnormalities section of this summary for more information); after taking specific biological factors of their leukemia into account, the predicted outcome for any individual patient may be much better or much worse than the overall outcome for the general population of children with AML.

Induction Chemotherapy

Contemporary pediatric AML protocols result in 85% to 90% complete remission (CR) rates. Approximately 3% of patients die during the induction phase, most commonly due to treatment-related complications. To achieve a CR, inducing profound bone marrow aplasia (with the exception of the M3 APL subtype) is usually necessary with currently used combination chemotherapy regimens. Because induction chemotherapy produces severe myelosuppression, morbidity and mortality from infection or hemorrhage during the induction period may be significant.

The two most effective drugs used to induce remission in children with AML are cytarabine and an anthracycline. Commonly used pediatric induction therapy regimens use cytarabine and an anthracycline in combination with other agents such as etoposide and/or thioguanine. The United Kingdom Medical Research Council (MRC) 10 Trial compared induction with cytarabine, daunorubicin, and etoposide (ADE) versus cytarabine and daunorubicin administered with thioguanine (DAT); the results showed no difference between the thioguanine and etoposide arms in remission rate or disease-free survival (DFS), although the thioguanine-containing regimen was associated with increased toxicity.

The anthracycline that has been most used in induction regimens for children with AML is daunorubicin, although idarubicin and the anthracenedione mitoxantrone have also been used. Randomized trials have attempted to determine whether any other anthracycline or anthracenedione is superior to daunorubicin as a component of induction therapy for children with AML.

  • The German Berlin-Frankfurt-Münster (BFM) Group AML-BFM 93 study evaluated cytarabine plus etoposide with either daunorubicin or idarubicin (ADE or AIE) and observed similar EFS and OS for both induction treatments.
  • The clinical trial studied induction with cytarabine, mitoxantrone, and etoposide (MAE) in children and adults with AML compared with a similar regimen using daunorubicin (ADE). For all patients, MAE showed a reduction in relapse risk, but the increased rate of treatment-related mortality observed for patients receiving MAE led to no significant difference in disease-free survival or OS in comparison to ADE. Similar results were noted when analyses were restricted to pediatric patients.
  • The AML-BFM 2004 clinical trial compared liposomal daunorubicin (L-DNR) to idarubicin at a higher-than-equivalent dose (80 mg/m vs. 12 mg/m per day for 3 days) during induction. Five-year results in both treatment arms were similar for both OS and EFS. Treatment-related mortality was significantly lower with L-DNR than idarubicin (2 of 257 patients vs. 10 of 264 patients).

In the absence of convincing data that another anthracycline or mitoxantrone produces superior outcome to daunorubicin when given at an equitoxic dose, daunorubicin remains the anthracycline most commonly used during induction therapy for children with AML in the United States.

The intensity of induction therapy influences the overall outcome of therapy. The CCG-2891 study demonstrated that intensively timed induction therapy (4-day treatment courses separated by only 6 days) produced better EFS than standard-timing induction therapy (4-day treatment courses separated by 2 weeks or longer). The MRC has intensified induction therapy by prolonging the duration of cytarabine treatment to 10 days. Another way of intensifying induction therapy is by the use of high-dose cytarabine. While studies in nonelderly adults suggest an advantage for intensifying induction therapy with high-dose cytarabine (2–3 g/m/dose) compared with standard-dose cytarabine, a benefit for the use of high-dose cytarabine compared with standard-dose cytarabine in children was not observed using a cytarabine dose of 1 g/m given twice daily for 7 days with daunorubicin and thioguanine. A second pediatric study also failed to detect a benefit for high-dose cytarabine over standard-dose cytarabine when used during induction therapy.

Because further intensification of induction regimens has increased toxicity with little improvement in EFS or OS, alternative approaches have been examined. The Children's Oncology Group (COG) recently completed a series of trials—, a pilot study, and , a randomized trial—that examined incorporation of the anti-CD33 conjugated antibody gemtuzumab ozogamicin into induction therapy. With the use of gemtuzumab ozogamicin during induction cycle one, dosed at 3 mg/m on day 6, the randomized trial identified an improved EFS but not OS; this was because of a reduction in postremission relapse overall and specifically in distinct subsets of patients. These subsets included patients with low-risk cytogenetics, patients with intermediate-risk AML who went on to receive stem cell transplantation (SCT) from a matched-related donor, and patients with high-risk AML ( high allelic ratio, >0.4) who then received a SCT from any donor. The expression intensity of CD33 on leukemic cells appeared to predict which patients benefited from gemtuzumab ozogamicin on the COG AAML0531 clinical trial.[] Patients whose CD33 intensity fell into the highest three population quartiles benefited from gemtuzumab ozogamicin (improved relapse risk, DFS, and EFS), whereas those in the lowest quartile had no reduction in relapse risk, EFS, or OS. This impact was seen for low-, intermediate-, and high-risk patients.

A meta-analysis of five randomized clinical trials that evaluated gemtuzumab ozogamicin for adults with AML observed the greatest OS benefit for patients with low-risk cytogenetics (t(8;21)(q22;q22) and inv(16)(p13q22)/t(16;16)(p13;q22)). Adult AML patients with intermediate-risk cytogenetics who received gemtuzumab ozogamicin had a significant but more modest improvement in OS, while there was no evidence of benefit for patients with adverse cytogenetics. Gemtuzumab ozogamicin is currently not available in the United States, except when approved for compassionate use.

In children with high-risk AML, the estimated incidence of severe bacterial infections is 50% to 60%, and the estimated incidence of invasive fungal infections is 7.0% to 12.5%. Several approaches have been examined in terms of reducing the morbidity and mortality from infection in children with AML.

Hematopoietic growth factors such as granulocyte-macrophage colony-stimulating factor (GM-CSF) or granulocyte colony-stimulating factor (G-CSF) during AML induction therapy have been evaluated in multiple placebo-controlled studies in adults with AML in attempts to reduce the toxicity associated with prolonged myelosuppression. These studies have generally shown a reduction of several days in the duration of neutropenia with the use of either G-CSF or GM-CSF but have not shown significant effects on treatment-related mortality or OS. (Refer to the Treatment Option Overview for AML section in the PDQ summary on Adult Acute Myeloid Leukemia Treatment for more information.) A randomized study in children with AML that evaluated G-CSF administered after induction chemotherapy showed a reduction in duration of neutropenia but no difference in infectious complications or mortality. A higher relapse rate has been reported for children with AML expressing the differentiation defective G-CSF receptor isoform IV. Thus, routine prophylactic use of hematopoietic growth factors is not recommended for children with AML.

The use of antibacterial prophylaxis in children undergoing treatment for AML has been supported by several studies. A retrospective study from St. Jude Children's Research Hospital (SJCRH) in patients with AML reported that the use of intravenous cefepime or vancomycin in conjunction with oral ciprofloxacin or a cephalosporin significantly reduced the incidence of bacterial infection and sepsis compared with patients receiving only oral or no antibiotic prophylaxis. A retrospective report from the trial demonstrated significant reductions in sterile-site bacterial infection and particularly gram-positive, sterile-site infections were both associated with the use of antibacterial prophylaxis. Of note, this study also reported that prophylactic use of G-CSF reduced bacterial and infections. In a study that compared the percentage of bloodstream infections or invasive fungal infections in children with ALL or AML who underwent chemotherapy and received antibacterial and antifungal prophylaxis, a significant reduction in both variables was observed compared with a historical control group that did not receive any prophylaxis. While such studies suggest a benefit to the use of antibiotic prophylaxis, prospective randomized trials are needed in this pediatric group of patients.

Similarly, the role of antifungal prophylaxis has not been studied in children with AML using randomized, prospective studies. Nevertheless, two meta-analysis reports have suggested that antifungal prophylaxis in pediatric patients with AML during treatment-induced neutropenia or during bone marrow transplantation does reduce the frequency of invasive fungal infections and in some instances nonrelapse mortality. However, another study that analyzed 1,024 patients with AML treated on the clinical trial reported no benefit of antifungal prophylaxis on fungal infections or nonrelapse mortality. Several randomized trials in adults with AML, however, have reported significant benefit in reducing invasive fungal infection with the use of antifungal prophylaxis. Such studies have also balanced cost with adverse side effects; when effectiveness at reducing invasive fungal infection is balanced with these other factors, posaconazole, voriconazole, caspofungin, and micafungin are considered reasonable choices.

Induction failure is seen in 10% to 15% of children with AML. Subsequent outcomes for patients with induction failure are similar to those for patients with AML who relapse early (<12 months after remission). Like patients with relapsed AML, induction failure patients are typically guided towards SCT once they attain a remission, because recent studies suggest a better EFS than in patients treated with chemotherapy only (31.2% vs. 5%, < .0001). Attainment of morphologic CR for these patients is a significant prognostic factor for DFS after SCT (46% vs. 0%; = .02), with failure primarily resulting from relapse (relapse risk, 53.9% vs. 88.9%; = .02).

Treatment options under clinical evaluation

The following are examples of national and/or institutional clinical trials that are currently being conducted. Information about ongoing clinical trials is available from the NCI website.

  • (Clofarabine Plus Cytarabine Versus Conventional Induction Therapy and a Study of Natural Killer Cell Transplantation in Newly Diagnosed AML): SJCRH is conducting a randomized trial for children with newly diagnosed AML. This trial compares two induction regimens: cytarabine/daunorubicin/etoposide (ADE) versus clofarabine/cytarabine. Responses are assessed via morphology and flow cytometry (minimal residual disease) at the end of the induction phase.
  • (Bortezomib and Sorafenib Tosylate in Patients With Newly Diagnosed AML With or Without Mutations): COG-AAML1031 uses an ADE induction therapy backbone. For patients without -–positive AML, the study is using a randomized design to evaluate whether the addition of bortezomib throughout the course of therapy improves EFS and OS. For patients with high allelic ratio -–positive AML, the primary objective is to evaluate the feasibility of combining sorafenib (a small molecule inhibitor) with standard chemotherapy. A secondary objective for this patient population is to determine the antileukemic activity of sorafenib for -–positive AML.

Central Nervous System (CNS) Prophylaxis for AML

Although the presence of CNS leukemia at diagnosis (i.e., clinical neurologic features and/or leukemic cells in cerebral spinal fluid on cytocentrifuge preparation) is more common in childhood AML than in childhood acute lymphoblastic leukemia (ALL), survival is not adversely affected. This finding is perhaps related to both the higher doses of chemotherapy used in AML (with potential crossover to the CNS) and the fact that marrow disease has not yet been as effectively brought under long-term control in AML as in ALL. Children with M4 and M5 AML have the highest incidence of CNS leukemia (especially those with inv(16) or 11q23 chromosomal abnormalities). The use of some form of intrathecal chemotherapy as CNS-directed treatment is now incorporated into most protocols for the treatment of childhood AML and is considered a standard part of the treatment for AML. Cranial radiation is no longer routinely employed in the treatment of children with AML.

Granulocytic Sarcoma/Chloroma

Granulocytic sarcoma (chloroma) describes extramedullary collections of leukemia cells. These collections can occur, albeit rarely, as the sole evidence of leukemia. In a review of three AML studies conducted by the former Children's Cancer Group, fewer than 1% of patients had isolated granulocytic sarcoma, and 11% had granulocytic sarcoma along with marrow disease at the time of diagnosis. Importantly, the patient who presents with an isolated tumor, without evidence of marrow involvement, must be treated as if there is systemic disease. Patients with isolated granulocytic sarcoma have a good prognosis if treated with current AML therapy.

In a study of 1,459 children with newly diagnosed AML, patients with orbital granulocytic sarcoma and CNS granulocytic sarcoma had better survival than patients with marrow disease and granulocytic sarcoma at other sites and AML patients without any extramedullary disease. The majority of patients with orbital granulocytic sarcoma have a t(8;21) abnormality, which has been associated with a favorable prognosis. The use of radiation therapy does not improve survival in patients with granulocytic sarcoma who have a complete response to chemotherapy, but may be necessary if the site(s) of granulocytic sarcoma do not show complete response to chemotherapy or for disease that recurs locally.

Current Clinical Trials

Check the list of NCI-supported cancer clinical trials that are now accepting patients with untreated childhood acute myeloid leukemia and other myeloid malignancies. The list of clinical trials can be further narrowed by location, drug, intervention, and other criteria.

General information about clinical trials is also available from the NCI website.

Postremission Therapy for AML

A major challenge in the treatment of children with acute myeloid leukemia (AML) is to prolong the duration of the initial remission with additional chemotherapy or hematopoietic stem cell transplantation (HSCT). In practice, most patients are treated with intensive chemotherapy after remission is achieved, as only a small subset have a matched-family donor (MFD). Such therapy includes some of the drugs used in induction while also introducing non-cross–resistant drugs and commonly high-dose cytarabine. Studies in adults with AML have demonstrated that consolidation with a high-dose cytarabine regimen improves outcome compared with consolidation with a standard-dose cytarabine regimen, particularly in patients with inv(16) and t(8;21) AML subtypes. (Refer to the Adult AML in Remission section in the PDQ summary on Adult Acute Myeloid Leukemia Treatment for more information.) Randomized studies evaluating the contribution of high-dose cytarabine to postremission therapy have not been conducted in children, but studies employing historical controls suggest that consolidation with a high-dose cytarabine regimen improves outcome compared with less intensive consolidation therapies.

The optimal number of postremission courses of therapy remains unclear, but appears to require at least three courses of intensive therapy, including the induction course. A United Kingdom Medical Research Council (MRC) study randomly assigned adult and pediatric patients to four versus five courses of intensive therapy. Five courses did not show an advantage in relapse-free and overall survival (OS).[]

The use of HSCT in first remission has been under evaluation since the late 1970s, and evidence-based appraisals concerning indications for autologous and allogeneic HSCT have been published. Prospective trials of transplantation in children with AML suggest that overall, 60% to 70% of children with HLA-matched donors available who undergo allogeneic HSCT during their first remission experience long-term remissions, with the caveat that outcome after allogeneic HSCT is dependent upon risk-classification status. In prospective trials of allogeneic HSCT compared with chemotherapy and/or autologous HSCT, a superior disease-free survival (DFS) has been observed for patients who were assigned to allogeneic transplantation based on availability of a family 6/6 or 5/6 HLA-matched donor in adults and children. However, the superiority of allogeneic HSCT over chemotherapy has not always been observed. Several large cooperative group clinical trials for children with AML have found no benefit for autologous HSCT over intensive chemotherapy.

Current application of allogeneic HSCT involves incorporation of risk classification into the determination of whether transplantation should be pursued in first remission. Because of the improved outcome in patients with favorable prognostic features receiving contemporary chemotherapy regimens and the lack of demonstrable superiority for HSCT in this patient population, it is now recommended that this group of patients receive MFD HSCT only after first relapse and the achievement of a second complete remission (CR).

There is conflicting evidence regarding the role of allogeneic HSCT in first remission for patients with intermediate-risk characteristics:

  • A study combining the results of the POG-8821, CCG-2891, , and MRC-Leuk-AML-10-Child studies identified a DFS and OS advantage for allogeneic HSCT in patients with intermediate-risk AML but not favorable-risk (inv(16) and t(8;21)) or poor-risk (del(5q), monosomy 5 or 7, or more than 15% blasts after first induction for POG/CCG studies), as well as including 3q abnormalities and complex cytogenetics in the MRC study. Weaknesses of this study include the large percentage of patients not assigned to a risk group and the relatively low EFS and OS rates for patients with intermediate risk assigned to chemotherapy compared with results observed in more recent clinical trials.
  • The AML99 clinical trial from the Japanese Childhood AML Cooperative Study Group observed a significant difference in DFS for intermediate risk patients assigned to MFD HSCT, but there was not a significant difference in OS.
  • The AML-BFM 99 clinical trial demonstrated no significant difference for intermediate risk patients in either DFS or OS for patients assigned to MFD HSCT versus those assigned to chemotherapy.

Given the improved outcome for patients with intermediate-risk AML in recent clinical trials and the burden of acute and chronic toxicities associated with allogeneic transplantation, many childhood AML treatment groups (including the Children's Oncology Group [COG]) employ chemotherapy for intermediate-risk patients in first remission and reserve allogeneic HSCT for use after potential relapse.

There are conflicting data regarding the role of allogeneic HSCT in first remission for patients with high-risk disease, complicated by the differing definitions of used by different study groups.

  • A retrospective analysis from COG and Center for International Blood and Marrow Transplant Research (CIBMTR) data on patients with AML and high-risk cytogenetics, defined as monosomy 7/del(7q), monosomy 5/del(5q), abnormalities of 3q, t(6;9), or complex karyotypes comparing chemotherapy only with minimal residual disease (MRD) donor and matched-unrelated donor (MUD) stem cell transplantation demonstrated no difference in the 5-year OS among the three treatment groups.
  • A Nordic Society for Pediatric Hematology and Oncology study reported that time-intensive reinduction therapy followed by transplant with best available donor for patients whose AML did not respond to induction therapy resulted in 70% survival at a median follow-up of 2.6 years.[]
  • A single-institution retrospective study of 50 consecutive patients (aged 0–30 years) with high-risk AML (, 11q23 [] rearrangements, presence of chromosome 5 or 7 abnormalities, induction failure, persistent disease) reported a 5-year 72% OS and a 69% leukemia-free survival; they also reported a 17% treatment-related mortality.
  • A subgroup analysis from the AML-BFM 98 clinical trial demonstrated improved survival rates for patients with 11q23 aberrations allocated to allogeneic HSCT, but not for patients without 11q23 aberrations.
  • For children with - (high allelic ratio), those who received MFD HSCT (n = 6) had higher OS than who received standard chemotherapy (n = 28); however the number of cases studied limited the ability to draw conclusions.

Many, but not all, pediatric clinical trial groups prescribe allogeneic HSCT for high-risk patients in first remission. For example, the COG frontline AML clinical trial () prescribes allogeneic HSCT in first remission only for patients with predicted high risk of treatment failure based on unfavorable cytogenetic and molecular characteristics and elevated end-of-induction MRD levels. On the other hand, the AML-BFM trials restrict allogeneic HSCT to patients in second CR and to refractory AML based on results from their AML-BFM 98 study showing no improvement in DFS or OS for high-risk patients receiving allogeneic HSCT in first CR and on the ability of HSCT in second CR to successfully treat a substantial proportion of patients. Additionally, late sequelae (e.g., cardiomyopathy, skeletal anomalies, and liver dysfunction or cirrhosis) were increased for children undergoing allogeneic HSCT in first remission on the AML-BFM 98 study.

Because definitions of high-, intermediate-, and low-risk AML are evolving due to the ongoing association of molecular characteristics of the tumor with outcome (e.g., internal tandem duplications, mutations, and mutations) and response to therapy (e.g., MRD assessments postinduction therapy), further analysis of subpopulations of patients treated with allogeneic HSCT will be an ongoing need in current and future clinical trials.

If transplant is chosen in first CR, the optimal preparative regimen and source of donor cells has not been determined, although alternative donor sources, including haploidentical donors, are being studied. Of note, there are no data that suggest total-body irradiation (TBI) is superior to busulfan-based myeloablative regimens. A randomized trial comparing busulfan plus fludarabine versus busulfan plus cyclophosphamide as a preparative regimen for AML in first CR demonstrated that the former regimen was associated with less toxicity and comparable DFS and OS. In addition, a large prospective CIBMTR cohort study of children and adults with AML, myelodysplastic syndromes (MDS), and chronic myelogenous leukemia (CML) showed superior survival of patients with “early-stage” disease (chronic-phase CML, first CR AML, and MDS-refractory anemia) with busulfan-based regimens compared with TBI.

Maintenance chemotherapy has been shown to be effective in the treatment of acute promyelocytic leukemia. In other subtypes, there are no data that demonstrate that maintenance therapy given after intensive postremission therapy significantly prolongs remission duration. Maintenance chemotherapy failed to show benefit in two randomized studies, and maintenance therapy with interleukin-2 also proved ineffective.

Treatment Options Under Clinical Evaluation

The following are examples of national and/or institutional clinical trials that are currently being conducted. Information about ongoing clinical trials is available from the NCI website.

  • (Clofarabine Plus Cytarabine Versus Conventional Induction Therapy and a Study of Natural Killer Cell Transplantation in Newly Diagnosed AML): St. Jude Children’s Research Hospital is conducting a randomized trial for children with newly diagnosed AML in which the efficacy of postchemotherapy NK cell transplantation is being assessed after five cycles of chemotherapy.
  • (Bortezomib and Sorafenib Tosylate in Patients With Newly Diagnosed AML With or Without Mutations): This is a phase III COG study designed to answer the question of whether the addition of the proteasome inhibitor bortezomib to chemotherapy during induction and postremission therapy improves outcome; in addition, this study will test whether the addition of sorafenib to chemotherapy along with HSCT for patients with high-allelic ratio -–positive AML improves outcome compared with historical controls.

Current Clinical Trials

Check the list of NCI-supported cancer clinical trials that are now accepting patients with childhood acute myeloid leukemia in remission. The list of clinical trials can be further narrowed by location, drug, intervention, and other criteria.

General information about clinical trials is also available from the NCI website.

Acute Promyelocytic Leukemia

Acute promyelocytic leukemia (APL) is a distinct subtype of acute myeloid leukemia (AML) and is treated differently than other types of AML. Optimal treatment requires rapid initiation of treatment with all- retinoic acid (ATRA) and supportive care measures. The characteristic chromosomal abnormality associated with APL is t(15;17). This translocation involves a breakpoint that includes the retinoic acid receptor and leads to production of the promyelocytic leukemia (PML)-retinoic acid receptor alpha (RARA) fusion protein. Patients with a suspected diagnosis of APL can have their diagnosis confirmed by detection of the PML-RARA fusion (e.g., through fluorescence hybridization [FISH], reverse transcriptase–polymerase chain reaction [RT–PCR], or conventional cytogenetics). An immunofluorescence method using an anti-PML monoclonal antibody can rapidly establish the presence of the PML-RARA fusion protein based on the characteristic distribution pattern of PML that occurs in the presence of the fusion protein.

Clinically, APL is characterized by severe coagulopathy that is often present at the time of diagnosis. Mortality during induction (particularly with cytotoxic agents used alone) caused by bleeding complications is more common in this subtype than in other French-American-British classifications. A lumbar puncture at diagnosis should not be performed until evidence of coagulopathy has resolved. Initiation of ATRA therapy is strongly recommended as soon as APL is suspected based on morphological and clinical presentation, because ATRA has been shown to ameliorate bleeding risk for patients with APL. A retrospective analysis identified an increase in early death resulting from hemorrhage in patients with APL in whom ATRA introduction was delayed.

APL in children is generally similar to APL in adults, though children have a higher incidence of hyperleukocytosis (defined as white blood cell [WBC] count higher than 10 × 10/L) and a higher incidence of the microgranular morphologic subtype. Similar to adults, children with WBC counts less than 10 × 10/L at diagnosis have significantly better outcome than patients with higher WBC counts. The prognostic significance of WBC count is used to define high-risk and low-risk patient populations and to assign postinduction treatment, with high-risk patients most commonly defined by WBC of 10 × 10/L or greater. mutations (either internal tandem duplications or kinase domain mutations) are observed in 40% to 50% of APL cases, with the presence of mutations correlating with higher WBC counts and the microgranular variant (M3v) subtype. mutation has been associated with an increased risk of induction death, and in some reports, an increased risk of treatment failure. Data from a combined analysis of two European trials demonstrated that children younger than 4 years with APL presented with higher WBC counts, had an increased incidence of the M3v subtype, and had a higher cumulative incidence of relapse and fatal cardiac toxicity during remission than did adolescents and adults; however, overall survival (OS) was similar.[]

The basis for current treatment programs for APL is the sensitivity of leukemia cells from patients with APL to the differentiation-inducing effects of ATRA. The dramatic efficacy of ATRA against APL results from the ability of pharmacologic doses of ATRA to overcome the repression of signaling caused by the PML-RARA fusion protein at physiologic ATRA concentrations. Restoration of signaling leads to differentiation of APL cells and then to postmaturation apoptosis. Most patients with APL achieve a complete remission (CR) when treated with ATRA, though single-agent ATRA is generally not curative. A series of randomized clinical trials defined the benefit of combining ATRA with chemotherapy during induction therapy and also the utility of using ATRA as maintenance therapy. ATRA is also commonly used as a component of postinduction consolidation therapy, with treatment regimens that include several additional courses of ATRA given with an anthracycline with or without cytarabine. Evidence for the benefit of giving ATRA with consolidation chemotherapy is derived from historical comparisons of results from adult APL clinical trials showing significant improvements in outcome for patients receiving ATRA given in conjunction with chemotherapy compared with chemotherapy alone. (Refer to the Acute Promyelocytic Leukemia section in the PDQ summary on Adult Acute Myeloid Leukemia Treatment for more information.) For children with APL, survival rates exceeding 80% are now achievable using treatment programs that prescribe the rapid initiation of ATRA and appropriate supportive care measures. For patients in CR for more than 5 years, relapse is extremely rare.[]

The standard approach to treating children with APL builds upon adult clinical trial results and begins with induction therapy using ATRA given in combination with an anthracycline administered with or without cytarabine. One regimen uses ATRA in conjunction with standard-dose cytarabine and daunorubicin, while another utilizes idarubicin and ATRA without cytarabine for remission induction. Almost all children with APL treated with one of these approaches achieves CR in the absence of coagulopathy-related mortality. Assessment of response to induction therapy in the first month of treatment using morphologic and molecular criteria may provide misleading results as delayed persistence of differentiating leukemia cells can occur in patients who will ultimately achieve CR. Alterations in planned treatment based on these early observations are not appropriate as resistance of APL to ATRA plus anthracycline-containing regimens is extremely rare.

Consolidation therapy has typically included ATRA given with an anthracycline with or without cytarabine. The role of cytarabine in consolidation therapy regimens is controversial. While a randomized study addressing the contribution of cytarabine to a daunorubicin plus ATRA regimen in adults with low-risk APL showed a benefit for the addition of cytarabine, regimens using high-dose anthracycline appear to produce as good or better results for low-risk patients. For high-risk patients (WBC ≥10 × 10/L), a historical comparison of the LPA2005 trial to the preceding PETHEMA LPA99 trial suggested that the addition of cytarabine to anthracycline-ATRA combinations can lower the relapse rate. The results of the AIDA-2000 trial confirmed that the cumulative incidence of relapse for adult patients with high-risk disease can be reduced to approximately 10% with consolidation regimens containing ATRA, anthracyclines, and cytarabine.

Maintenance therapy includes ATRA plus 6-mercaptopurine and methotrexate; this combination showed an advantage over ATRA alone in randomized trials in adults with APL. A randomized study in adults has reported that maintenance therapy does not improve event-free survival (EFS) for patients with APL who achieve a complete molecular remission at the end of consolidation. However, the utility of maintenance therapy in APL may be dependent on multiple factors (e.g., risk group, the anthracycline used during induction, the intensity of induction and consolidation therapy, etc.), and at this time maintenance therapy remains standard for children with APL. Because of the favorable outcomes observed with chemotherapy plus ATRA (EFS rates of 70%–80%), hematopoietic stem cell transplantation is not recommended in first CR.

Central nervous system (CNS) relapse is uncommon for patients with APL, particularly for those with WBC count less than 10 × 10/L. In two clinical trials enrolling over 1,400 adults with APL in which CNS prophylaxis was not administered, the cumulative incidence of CNS relapse was less than 1% for patients with WBC less than 10 × 10/L, while it was approximately 5% for those with WBC of 10 × 10/L or greater. In addition to high WBC at diagnosis, CNS hemorrhage during induction is also a risk factor for CNS relapse. A review of published cases of pediatric APL also observed low rates of CNS relapse. Because of the low incidence of CNS relapse among children with APL presenting with WBC less than 10 × 10/L, CNS surveillance and prophylactic CNS therapy may not be needed for this group of patients, although there is no consensus on this topic.

Arsenic trioxide has also been identified as an active agent in patients with APL, and there are now data for its use as induction therapy, consolidation therapy, and in the treatment of patients with relapsed APL:

  • For adults with relapsed APL, approximately 85% achieve morphologic remission after treatment with this agent. Arsenic trioxide is well tolerated in children with relapsed APL. The toxicity profile and response rates in children are similar to that observed in adults.
  • In adults with newly diagnosed APL, the addition of two consolidation courses of arsenic trioxide to a standard APL treatment regimen resulted in a significant improvement in EFS (80% vs. 63% at 3 years; < .0001) and disease-free survival (90% vs. 70% at 3 years; < .0001), although the outcome of patients who did not receive arsenic trioxide was inferior to the results obtained in the GIMEMA or PETHEMA trials.
  • The concurrent use of arsenic trioxide and ATRA in newly diagnosed patients with APL results in high rates of CR. Early experience in children with newly diagnosed APL also shows high rates of CR to arsenic trioxide, either as a single agent or given with ATRA.[] Results of a meta-analysis of seven published studies in adult APL patients suggest that the combination of arsenic trioxide and ATRA may be more effective than arsenic trioxide alone in inducing CR. The impact of arsenic induction (either alone or with ATRA) on EFS and OS has not been well characterized in children, although early results appear promising.
  • Arsenic trioxide was evaluated as a component of induction therapy with idarubicin and ATRA in the APML4 clinical trial, which enrolled both children and adults (N = 124 evaluable patients). Patients received two courses of consolidation therapy with arsenic trioxide and ATRA (but no anthracycline) and maintenance therapy with ATRA, 6-mercaptopurine, and methotrexate. The 2-year rate for freedom from relapse was 97.5%, failure-free survival (FFS) was 88.1%, and OS was 93.2%. These results are superior for FFS and freedom from relapse when compared with the predecessor clinical trial (APML3) that did not use arsenic trioxide.
  • A German and Italian phase III clinical trial compared ATRA plus chemotherapy with ATRA plus arsenic trioxide in adults with APL classified as low to intermediate risk (WBC ≤ 10 × 10/L). Patients were randomly assigned to receive either ATRA plus arsenic trioxide for induction and consolidation therapy or standard ATRA-idarubicin induction therapy followed by three cycles of consolidation therapy with ATRA plus chemotherapy and maintenance therapy with low-dose chemotherapy and ATRA.

    All patients receiving ATRA plus arsenic trioxide (n = 77) achieved CR at the end of induction therapy, while 95% of patients receiving ATRA plus chemotherapy (n = 79) achieved CR. EFS rates were 97% in the ATRA-arsenic trioxide group compared with 86% in the ATRA-chemotherapy group ( = .02). Two-year OS probability was 99% (95% confidence interval [CI], 96–100) in the ATRA-arsenic trioxide group and 91% (95% CI, 85–97) in the ATRA-chemotherapy group ( = .02). These results indicate that low- to intermediate-risk APL is curable for a high percentage of patients without conventional chemotherapy.

Because arsenic trioxide causes QT interval prolongation that can lead to life-threatening arrhythmias (e.g., torsades de pointes), it is essential to monitor electrolytes closely in patients receiving arsenic trioxide and to maintain potassium and magnesium values at midnormal ranges.

The induction and consolidation therapies currently employed result in molecular remission as measured by reverse transcriptase–polymerase chain reaction (RT–PCR) for PML-RARA in the large majority of APL patients, with 1% or fewer showing molecular evidence of disease at the end of consolidation therapy. While two negative RT-PCR assays after completion of therapy are associated with long-term remission, conversion from negative to RT-PCR positivity is highly predictive of subsequent hematologic relapse. Patients with persistent or relapsing disease based upon PML-RARA RT-PCR measurement may benefit from intervention with relapse therapies (refer to the Recurrent Acute Promyelocytic Leukemia (APL) subsection of the Recurrent Childhood Acute Myeloid Leukemia and Other Myeloid Malignancies section of this summary for more information).

Molecular Variants of APL Other than PML-RARA

Uncommon molecular variants of APL produce fusion proteins that join distinctive gene partners (e.g., , , , and ) to RARA. Recognition of these rare variants is important as they differ in their sensitivity to ATRA and to arsenic trioxide. The PLZF-RARA variant, characterized by t(11;17)(q23;q21), represents about 0.8% of APL, expresses surface CD56, and has very fine granules compared with t(15;17) APL. APL with PLZF-RARA has been associated with a poor prognosis and does not usually respond to ATRA or to arsenic trioxide. The rare APL variants with NPM-RARA (t(5;17)(q35;q21)) or with NuMA-RARA (t(11;17)(q13;q21)) translocations may still be responsive to ATRA.

Current Clinical Trials

Check the list of NCI-supported cancer clinical trials that are now accepting patients with childhood acute promyelocytic leukemia (M3). The list of clinical trials can be further narrowed by location, drug, intervention, and other criteria.

General information about clinical trials is also available from the NCI website.

Children With Down Syndrome

Children with Down syndrome have a tenfold to twentyfold increased risk of leukemia compared with children without Down syndrome; the ratio of acute lymphoblastic leukemia to acute myeloid leukemia (AML) is nevertheless typical for childhood acute leukemia. The exception is during the first 3 years of life, when AML, particularly the megakaryoblastic subtype, predominates and exhibits a distinctive biology characterized by mutations and increased sensitivity to cytarabine. Importantly, these risks appear to be similar whether a child has phenotypic characteristics of Down syndrome or whether a child has only genetic bone marrow mosaicism.

In addition to increased risk of AML during the first 3 years of life, about 10% of neonates with Down syndrome also develop a transient myeloproliferative disorder (TMD) (also termed transient leukemia). This disorder mimics congenital AML, but typically improves spontaneously within the first 3 months of life, though TMD can remit as late as 20 months. Although TMD is usually a self-resolving condition, it can be associated with significant morbidity and may be fatal in 10% to 20% of affected infants. Infants with progressive organomegaly, visceral effusions, preterm delivery (less than 37-weeks gestation), bleeding diatheses, failure of spontaneous remission, laboratory evidence of progressive liver dysfunction (elevated direct bilirubin), and very high white blood cell (WBC) count are at particularly high risk for early mortality. Death has been reported to occur in 21% of these patients with high-risk TMD. Three risk groups have been identified based on the diagnostic clinical findings of hepatomegaly with or without life-threatening symptoms: (1) low risk includes those with neither finding (38% of patients and 92% ± 8% OS); (2) intermediate risk with hepatomegaly alone (40% of patients and 77% ± 12% overall survival [OS]); and (3) high risk with both characteristics (21% of patients and 51% ± 19% OS). Therapeutic intervention is warranted in patients in whom severe hydrops or organ failure is apparent. Several treatment approaches have been used, including exchange transfusion, leukapheresis, and low-dose cytarabine.

The mean time for the development of AML in the 10% to 30% of children who have a spontaneous remission of TMD but then develop AML has been reported to be approximately 16 months, with a range of 1 to 30 months. Thus, most infants with Down syndrome and TMD who later develop AML will do so within the first 3 years of life. Patients with Down syndrome who develop AML with an antecedent TMD have superior event-free survival (EFS) (91% ± 5%) compared with such children without TMD (70% ± 4%) at 5 years, although this was not observed in another study. While TMD is generally not characterized by cytogenetic abnormalities other than trisomy 21, the presence of additional cytogenetic findings may connote an increased risk for developing subsequent AML.

Outcome is generally favorable for children with Down syndrome who develop AML. The prognosis is particularly good (EFS exceeding 80%) in children aged 4 years or younger at diagnosis, the age group that accounts for the vast majority of Down syndrome patients with AML. A large study of 451 children with AML and Down syndrome (age >6 months and <5 years) confirmed the generally favorable outcome for this patient population (7-year EFS of 78% and 7-year OS of 79%). Multivariate analyses revealed that WBC count (≥20 × 109/L) and age (>3 years) were independent predictors for lower EFS, although 7-year EFS for the older population (>3 years) and for the higher WBC-count population still exceeded 60%. Absence of leukemia cell cytogenetic abnormalities (other than trisomy 21), observed in approximately 30% of patients, independently predicted for inferior OS and EFS (7-year EFS of 65% compared with 82% for patients with aberrant karyotypes).

Appropriate therapy for younger children (aged ≤4 years) with Down syndrome and AML is less intensive than current standard childhood AML therapy, and hematopoietic stem cell transplant is not indicated in first remission.

Children with mosaicism for trisomy 21 are recommended to be treated similarly to those children with clinically evident Down syndrome. Children with Down syndrome who are older than 4 years have a significantly worse prognosis. Although an optimal treatment for these children has not been defined, they are usually treated on AML regimens designed for children without Down syndrome.

Myelodysplastic Syndromes

The myelodysplastic syndromes (MDS) and myeloproliferative syndromes (MPS), which represent between 5% and 10% of all myeloid malignancies in children, are a heterogeneous group of disorders with the former usually presenting with cytopenias and the latter with increased peripheral white blood cell, red blood cell, or platelet counts. MDS is characterized by ineffective hematopoiesis and increased cell death, while MPS is associated with increased progenitor proliferation and survival. Because they both represent disorders of very primitive, multipotential hematopoietic stem cells, curative therapeutic approaches nearly always require allogeneic hematopoietic stem cell transplantation.

Patients usually present with signs of cytopenias, including pallor, infection, or bruising. The bone marrow is usually characterized by hypercellularity and dysplastic changes in myeloid precursors. Clonal evolution eventually can lead to the development of acute myeloid leukemia (AML). The percentage of abnormal blasts is less than 20%. The less common, hypocellular MDS, can be distinguished from aplastic anemia in part by its marked dysplasia, clonal nature, and higher percentage of CD34-positive precursors.

Although the etiology of MDS has not been elucidated, clues have begun to be defined. For instance, approximately 20% of malignant myeloid disorders, including MDS, in adults have been shown to have mutations in the gene. Other genes shown to be mutated in MDS include , , , , , , , , , and . Most of these genes are key elements of epigenetic regulation of the genome and affect DNA methylation and/or histone modification. Mutations in proteins involved in RNA splicing have been described in 45% to 85% of MDS and appear to occur early in the course of the disease. MDS in both adults and children has been shown to have aberrant DNA methylation patterns and approximately one-half of cases are characterized by hypermethylation of the promoters for the and genes, both of which play roles in cell cycle regulation.

Patients with the following germline mutations or inherited disorders have a significantly increased risk of developing MDS:

  • Fanconi anemia: Caused by germline mutations in DNA repair genes.
  • Dyskeratosis congenita: Resulting from mutations in genes regulating telomere length.
  • Shwachman-Diamond syndrome, Diamond-Blackfan anemia, and other bone marrow failure syndromes: Resulting from mutations in genes encoding ribosome-associated proteins. mutations have been linked to Diamond-Blackfan anemia and MDS predisposition.
  • Severe congenital neutropenia: Caused by mutations in the gene encoding elastase. The 15-year cumulative risk of MDS in patients with severe congenital neutropenia, also known as Kostmann syndrome, has been estimated to be 15%, with an annual risk of MDS/AML of 2% to 3%. It is unclear how mutations affecting this protein and how the chronic exposure of granulocyte colony-stimulating factor (G-CSF) contribute to the development of MDS.
  • Trisomy 21 syndrome: mutations are nearly always present in the transient leukemia associated with Trisomy 21.
  • Congenital amegakaryocytic thrombocytopenia (CAMT): Inherited mutations in the or genes are associated with CAMT. Mutations in in gene are the underlying genetic cause of CAMT; there is a less than 10% risk of developing MDS/AML in patients with CAMT.
  • mutations: Germline mutations of have been reported in patients with MDS/AML in conjunction with monocytopenia, B and NK deficiency, pulmonary alveolar proteinosis, and susceptibility to opportunistic infections.
  • or mutations: Inherited mutations in the or genes are associated with familial MDS/AML.

The French-American-British (FAB) and World Health Organization (WHO) classification systems of MDS and MPS have been difficult to apply to pediatric patients. Alternative classification systems for children have been proposed, but none have been uniformly adopted, with the exception of the modified 2008 WHO classification system. The WHO system has been modified for pediatrics. Refer to Table 3 and Table 4 for the WHO classification schema and diagnostic criteria.

The refractory cytopenia subtype represents approximately 50% of all childhood cases of MDS. The presence of an isolated monosomy 7 is the most common cytogenetic abnormality, although it does not appear to portend a poor prognosis compared with its presence in overt AML. However, the presence of monosomy 7 in combination with other cytogenetic abnormalities is associated with a poor prognosis. The relatively common abnormalities of -Y, 20q- and 5q- in adults with MDS are rare in childhood MDS. The presence of cytogenetic abnormalities found in AML defines disease that should be treated as AML and not MDS.

The International Prognostic Scoring System can help to distinguish low-risk from high-risk MDS, although its utility in children with MDS is more limited than in adults as many characteristics differ between children and adults. The median survival for children with high-risk MDS remains substantially better than adults and the presence of monosomy 7 in children has not had the same adverse prognostic impact as in adults with MDS.

The optimal therapy for childhood MDS has not been established. A key issue in thinking about therapy for pediatric patients with MDS is that these disorders usually involve a primitive hematopoietic stem cell. Thus, allogeneic hematopoietic stem cell transplantation (HSCT) is considered to be the optimal approach to treatment for pediatric patients with MDS. Unresolved issues include determining the best transplant preparative regimen and source of donor cells. However, some data are important to consider when making decisions. For example, survival as high as 80% has been reported for patients with early-stage MDS proceeding to transplant within a few months of diagnosis. Further, early transplant and not receiving pretransplant chemotherapy have been associated with improved survival in children with MDS.[] Disease-free survival (DFS) has been estimated to be between 50% to 70% for pediatric patients with advanced MDS using myeloablative transplant preparative regimens. While using nonmyeloablative preparative transplant regimens are being tested in patients with MDS and AML, such regimens are still investigational for children with these disorders, but may be reasonable in the setting of a clinical trial or when a patient’s organ function is compromised in such a way that they would not tolerate a myeloablative regimen.; []

The question of whether chemotherapy should be used in high-risk MDS has been examined. The Children's Cancer Group 2891 trial accrued patients between 1989 and 1995, including children with MDS. There were 77 patients with refractory anemia (n = 2), refractory anemia with excess blasts (n = 33), refractory anemia with excess blasts in transformation (n = 26), or AML with antecedent MDS (n = 16) who were enrolled and randomly assigned to standard or intensively timed induction. Subsequently, patients were allocated to allogeneic HSCT if there was a suitable family donor, or randomly assigned to autologous HSCT or chemotherapy. Patients with refractory anemia/refractory anemia with excess blasts had a poor remission rate (45%), and those with refractory anemia with excess blasts in transformation (69%) or AML with history of MDS (81%) had similar remission rates compared with de novo AML (77%). Six-year survival was poor for those with refractory anemia/refractory anemia with excess blasts (28%) and refractory anemia with excess blasts in transformation (30%). Patients with AML and antecedent MDS had a similar outcome to those with de novo AML (50% survival compared with 45%). Allogeneic HSCT appeared to improve survival at a marginal level of significance ( = .08). Based on analysis of these data and the literature, the authors concluded that children with a history of MDS who present with AML and many of those with refractory anemia with excess blasts in transformation do as well with AML therapy at diagnosis as children with AML. An exception to this conclusion is children with AML with a precedent MDS and monosomy 7; these patients have a very poor prognosis and are usually treated with some type of allogeneic HSCT. An analysis of 37 children with MDS treated on Berlin-Frankfurt-Münster AML protocols 83, 87, and 93 confirmed the induction response of 74% for patients with refractory anemia with excess blasts in transformation and suggested that transplantation was beneficial. Another study by the same group showed that with current approaches to HSCT, survival occurred in more than 60% of children with advanced MDS, and outcomes for patients receiving unrelated donor cells were similar to those for patients who received matched-family donor (MFD) cells.

A significant issue to consider for these results is that the subtype refractory anemia with excess blasts in transformation is likely to represent patients with overt AML, while refractory anemia and refractory anemia with excess blasts represent MDS. The WHO classification has now omitted the category of refractory anemia with excess blasts in transformation, concluding that this entity was essentially AML. The optimal therapy for patients with refractory anemia/refractory anemia with excess blasts without an HLA-MFD is unknown. Some of these patients require no therapy for years and have indolent diseases. Because failure rates after HSCT are lower in this group, strong consideration should be given for transplantation, especially when a 5/6 or 6/6 HLA-MFD is available. However, alternative forms of HSCT, utilizing matched unrelated donor cord blood, should be considered when treatment is required, as is usually the case in patients with severe symptomatic cytopenias. The 8-year DFS for children with various stages of MDS transplanted with either HLA matched or mismatched unrelated donor transplants has been reported to be 65% and 40%, respectively.[] A 3-year DFS of 50% was reported with the use of unrelated cord blood donor transplants for children with MDS, when the transplants were done after 2001.[]

Because MDS in children is often associated with inherited predisposition syndromes, reports of transplantation in small numbers of patients have been documented. For example, in patients with Fanconi anemia and AML or advanced MDS, the 5-year overall survival (OS) has been reported to be 33% to 55%.[] Second transplants have also been used in pediatric patients with MDS/MPD who relapse or suffer graft failure. The 3-year OS was 33% for those retransplanted for relapse and 57% for those transplanted for initial graft failure.[]

For patients with clinically significant cytopenias, supportive care, including transfusions and prophylactic antibiotics, can be considered. In addition, the use of hematopoietic growth factors can improve the hematopoietic status, but there remains some concern that such treatment could accelerate conversion to AML. Steroid therapy has also been used, including glucocorticoids and androgens, with mixed results. Treatments directed toward scavenging free oxygen radicals with amifostine or the use of differentiation-promoting retinoids, DNA methylation inhibitors (e.g., azacytidine and decitabine), and histone deacetylase inhibitors, have all shown some response, but no definitive trials in children with MDS have been reported. Azacytidine has been approved by the U.S. Food and Drug Administration (FDA) for the treatment of MDS in adults based on randomized studies. (Refer to the Disease-Modifying Agents section in the PDQ summary on Myelodysplastic Syndromes Treatment for more information.) Agents, such as lenalidomide, an analog of thalidomide, have been tested based on findings that demonstrated increased activity in the bone marrow of patients with MDS. Lenalidomide has shown most efficacy in patients with 5q- syndrome, especially those with thrombocytosis, and is now FDA-approved for use in this group. Immunosuppression with antithymocyte globulin and/or cyclosporine has also been reported.

Treatment Options Under Clinical Evaluation

The following are examples of national and/or institutional clinical trials that are currently being conducted. Information about ongoing clinical trials is available from the NCI website.

  • The use of a variety of inhibitors of DNA methylation and histone deacetylase inhibitors, as well as other therapies designed to induce differentiation, are being studied in both young and older adults with MDS.

Current Clinical Trials

Check the list of NCI-supported cancer clinical trials that are now accepting patients with childhood myelodysplastic syndromes. The list of clinical trials can be further narrowed by location, drug, intervention, and other criteria.

General information about clinical trials is also available from the NCI website.

Therapy-Related AML/Myelodysplastic Syndromes

The development of acute myeloid leukemia (AML) or myelodysplastic syndromes (MDS) after treatment with ionizing radiation or chemotherapy, particularly alkylating agents and topoisomerase inhibitors, is termed (t-AML or t-MDS, respectively). In addition to genotoxic exposures, genetic predisposition susceptibilities (such as polymorphisms in drug detoxification and DNA repair pathway components) may contribute to the occurrence of secondary AML/MDS. The risk of t-AML/t-MDS is regimen-dependent and often related to the cumulative doses of chemotherapy agents received, and the dose and field of radiation administered. Regimens previously used that employed high cumulative doses of either epipodophyllotoxins (e.g., etoposide or teniposide) or alkylating agents (e.g., mechlorethamine, melphalan, busulfan, and cyclophosphamide) induced excessively high rates of t-AML/t-MDS that exceeded 10% in some cases. However, most current chemotherapy regimens that are used to treat childhood cancers have a cumulative incidence of t-AML/t-MDS not greater than 1% to 2%. t-AML/t-MDS resulting from epipodophyllotoxins and other topoisomerase II inhibitors (e.g., anthracyclines) usually occur within 2 years of exposure and are commonly associated with chromosome 11q23 abnormalities, although other subtypes of AML (e.g., acute promyelocytic leukemia) have been reported. t-AML that occurs after exposure to alkylating agents or ionizing radiation often presents 5 to 7 years later and is commonly associated with monosomies or deletions of chromosomes 5 and 7.

The goal of treatment is to achieve an initial complete remission (CR) using AML-directed regimens and then, usually, proceed directly to hematopoietic stem cell transplantation (HSCT) with the best available donor. However, treatment is challenging because of the following:

Accordingly, CR rates and overall survival (OS) rates are usually lower for patients with t-AML compared with patients with de novo AML. Patients with t-MDS-refractory anemia usually have not needed induction chemotherapy before transplant; the role of induction therapy before transplant is controversial in patients with refractory anemia with excess blasts-1.

Only a few reports describe the outcome of children undergoing HSCT for t-AML. One study described outcomes of 27 children with t-AML who received related and unrelated donor HSCT. Three-year OS rates were 18.5% ± 7.5% and event-free survival rates were 18.7% ± 7.5%. Poor survival was mainly the result of very high transplant-related mortality (59.6% ± 8.4%). Another study reported a second retrospective single-center experience of 14 patients transplanted for t-AML/t-MDS between 1975 and 2007. Survival was 29%, but in this review only 63% of patients diagnosed with t-AML/t-MDS underwent HSCT. A multicenter study (CCG-2891) looked at outcomes of 24 children with t-AML/t-MDS compared with other children enrolled on the study with de novo AML (n = 898) or MDS (n = 62). Children with t-AML/t-MDS were older and low-risk cytogenetics rarely occurred. Although rates of achieving CR and OS at 3 years were worse in the t-AML/t-MDS group (CR, 50% vs. 72%; = .016; OS, 26% vs. 47%; = .007), survival was similar (OS, 45% vs. 53%; = .87) if patients achieved a CR. The importance of remission to survival in these patients is further illustrated by another single-center report of 21 children undergoing HSCT for t-AML/t-MDS between 1994 and 2009. Of the 21 children, 12 had t-AML (11 in CR at the time of transplant), seven had refractory anemia (for whom induction was not done), and two had refractory anemia with excess blasts. Survival of the entire cohort was 61%; those in remission or with refractory anemia had a disease-free survival of 66%, and for the three patients with more than 5% blasts at the time of HSCT, survival was 0% ( = .015). Because t-AML is rare in children, it is not known whether the significant decrease in transplant-related mortality after unrelated donor HSCT noted over the past several years will translate to improved survival in this population. Patients should be carefully assessed for pre-HSCT morbidities caused by earlier therapies and approaches should be adapted to give adequate intensity while minimizing transplant-related mortality.

Juvenile Myelomonocytic Leukemia


Juvenile myelomonocytic leukemia (JMML) is a rare leukemia that occurs approximately ten times less frequently than acute myeloid leukemia (AML) in children, with an annual incidence of about 1 to 2 cases per 1 million people. JMML typically presents in young children (median age, approximately 1.8 years) and occurs more commonly in boys (male to female ratio, approximately 2.5:1).

Clinical Presentation and Diagnostic Criteria

Common clinical features at diagnosis include the following:

  • Hepatosplenomegaly (97%).
  • Lymphadenopathy (76%).
  • Pallor (64%).
  • Fever (54%).
  • Skin rash (36%).

In children presenting with clinical features suggestive of JMML, current criteria used for a definitive diagnosis are described in Table 5.

Pathogenesis and Related Syndromes

The pathogenesis of JMML has been closely linked to activation of the oncogene pathway, along with related syndromes (refer to Figure 1). In addition, distinctive RNA expression and DNA methylation patterns have been reported; they are correlated with clinical factors such as age and appear to be associated with prognosis.

Schematic diagram showing ligand-stimulated Ras activation, the Ras-Erk pathway, and gene mutations contributing to the neuro-cardio-facio-cutaneous congenital disorders and JMML.Figure 1. Schematic diagram showing ligand-stimulated Ras activation, the Ras-Erk pathway, and the gene mutations found to date contributing to the neuro-cardio-facio-cutaneous congenital disorders and JMML. NL/MGCL: Noonan-like/multiple giant cell lesion; CFC: cardia-facio-cutaneous; JMML: juvenile myelomonocytic leukemia. Reprinted from Leukemia Research, 33 (3), Rebecca J. Chan, Todd Cooper, Christian P. Kratz, Brian Weiss, Mignon L. Loh, Juvenile myelomonocytic leukemia: A report from the 2nd International JMML Symposium, Pages 355-62, Copyright 2009, with permission from Elsevier.

Children with neurofibromatosis type 1 (NF1) and Noonan syndrome are at increased risk of developing JMML:

  • NF1. Up to 14% of cases of JMML occur in children with NF1.
  • Noonan syndrome. Noonan syndrome is usually inherited as an autosomal dominant condition, but can also arise spontaneously. It is characterized by facial dysmorphism, short stature, webbed neck, neurocognitive abnormalities, and cardiac abnormalities. Germline mutations in are observed in Noonan syndrome children with JMML.

    Importantly, some children with Noonan syndrome have a hematologic picture indistinguishable from JMML that self-resolves during infancy, similar to what happens in children with Down syndrome and transient myeloproliferative disorder. Within a large prospective cohort of 641 patients with Noonan syndrome and a germline mutation, 36 patients (~6%) showed myeloproliferative features, with 20 patients (~3%) meeting the consensus diagnostic criteria for JMML. Of the 20 patients meeting the criteria for JMML, 12 patients had severe neonatal manifestations (e.g., life-threatening complications related to congenital heart defects, pleural effusion, leukemia infiltrates, and/or thrombocytopenia), and 10 of 20 patients died during the first month of life. Among the remaining eight patients, none required intensive therapy at diagnosis or during follow-up. All 16 patients with myeloproliferative features not meeting JMML criteria were alive, with a median follow-up of 3 years, and none received chemotherapy.

Mutations in the () gene, a E3 ubiquitin-protein ligase that is involved in targeting proteins, particularly tyrosine kinases, for proteasomal degradation occur in 10% to 15% of JMML cases, with many of these cases occurring in children with germline mutations. germline mutations result in an autosomal dominant developmental disorder that is characterized by impaired growth, developmental delay, cryptorchidism, and a predisposition to JMML. Some individuals with germline mutations experience spontaneous regression of their JMML but develop vasculitis later in life. mutations are nearly always mutually exclusive of and mutations.

Genomics of JMML

The genomic landscape of JMML is characterized by mutations in one of five genes of the Ras pathway: , , , , and . In a series of 118 consecutively diagnosed JMML cases with Ras pathway–activating mutations, was the most commonly mutated gene, accounting for 51% of cases (19% germline and 32% somatic) (refer to Figure 2). Patients with mutated accounted for 19% of cases, and patients with mutated accounted for 15% of cases. and mutations accounted for 8% and 11% of cases, respectively. Although mutations among these five genes are generally mutually exclusive, 10% to 17% of cases have mutations in two of these Ras pathway genes, a finding that is associated with poorer prognosis.

The mutation rate in JMML leukemia cells is very low, but additional mutations beyond those of the five Ras pathway genes described above are observed. Secondary genomic alterations are observed for genes of the transcriptional repressor complex PRC2 (e.g., was mutated in 7%–8% of cases). Some genes associated with myeloproliferative neoplasms in adults are also mutated at low rates in JMML (e.g., was mutated in 7%–9% of cases). mutations are also observed in a small percentage (4%–12%) of JMML cases. Cases with germline and germline mutations showed low rates of additional mutations (refer to Figure 2).

Chart showing alteration profiles in individual JMML cases.Figure 2. Alteration profiles in individual JMML cases. Germline and somatically acquired alterations with recurring hits in the RAS pathway and PRC2 network are shown for 118 patients with JMML who underwent detailed genetic analysis. Blast excess was defined as a blast count ≥10% but <20% of nucleated cells in the bone marrow at diagnosis. Blast crisis was defined as a blast count ≥20% of nucleated cells in the bone marrow. NS, Noonan syndrome. Reprinted by permission from Macmillan Publishers Ltd: Nature Genetics (Caye A, Strullu M, Guidez F, et al.: Juvenile myelomonocytic leukemia displays mutations in components of the RAS pathway and the PRC2 network. Nat Genet 47 [11]: 1334-40, 2015), copyright (2015).


Several factors affect prognosis in JMML, including the following:

  • Number of non–Ras pathway mutations. A strong predictor of prognosis for children with JMML is the number of mutations beyond the disease-defining Ras pathway mutations. The first study observed that zero or one somatic alteration (pathogenic mutation or monosomy 7) was identified in 64 patients (65.3%) at diagnosis, whereas two or more alterations were identified in 34 patients (34.7%). In multivariate analysis, mutation number (2 or more versus 0 or 1) maintained significance as a predictor of inferior event-free survival and overall survival. A higher proportion of patients diagnosed with two or more alterations were older and male, and these patients also demonstrated a higher rate of monosomy 7 or somatic mutations. Similar findings were reported in a second study that additionally observed that patients with Ras pathway double mutations (15 of 96 patients) were at the highest risk of treatment failure.
  • Age, platelet count, and fetal hemoglobin level after any treatment. Historically, more than 90% of patients with JMML died despite the use of chemotherapy, but with the application of hematopoietic stem cell transplantation (HSCT), survival rates of approximately 50% are now observed. Patients appeared to follow three distinct clinical courses: (1) rapidly progressive disease and early demise; (2) transiently stable disease followed by progression and death; and (3) clinical improvement that lasted up to 9 years before progression or, rarely, long-term survival. Favorable prognostic factors for survival after any therapy include being younger than 3 years, having a platelet count greater than 33 × 10/L, and low age-adjusted fetal hemoglobin levels. In contrast, being older than 2 years and having high blood fetal hemoglobin levels at diagnosis are predictors of poor outcome.
  • : overexpression is present in approximately one-half of children with JMML and identifies a biologically distinctive subset of JMML. LIN28B is an RNA-binding protein that regulates stem cell renewal. overexpression was positively correlated with high blood fetal hemoglobin level and age (both of which are associated with poor prognosis), and it was negatively correlated with presence of monosomy 7 (also associated with inferior prognosis). Although overexpression identifies a subset of patients with increased risk of treatment failure, it was not found to be an independent prognostic factor when other factors such as age and monosomy 7 status are considered.

Treatment of JMML

Treatment options for JMML include the following:

  • Hematopoietic stem cell transplant (HSCT).

The role of conventional antileukemia therapy in the treatment of JMML is not defined. The absence of consensus response criteria for JMML complicates determination of the role of specific agents in the treatment of JMML. Some of the agents that have shown antileukemia activity against JMML include etoposide, cytarabine, thiopurines (thioguanine and 6-mercaptopurine), isotretinoin, and farnesyl inhibitors, but none of these have been shown to improve outcome.; []

HSCT currently offers the best chance of cure for JMML.

Evidence (HSCT):

Disease recurrence is the primary cause of treatment failure for children with JMML after HSCT and occurs in 30% to 40% of cases. While the role of donor lymphocyte infusions is uncertain, reports indicate that approximately 50% of patients with relapsed JMML can be successfully treated with a second HSCT.

Chronic Myelogenous Leukemia

Chronic myelogenous leukemia (CML) accounts for less than 5% of all childhood leukemia, and in the pediatric age range, occurs most commonly in older adolescents. The cytogenetic abnormality most characteristic of CML is the Philadelphia chromosome (Ph), which represents a translocation of chromosomes 9 and 22 (t(9;22)) resulting in a BCR-ABL fusion protein. CML is characterized by a marked leukocytosis and is often associated with thrombocytosis, sometimes with abnormal platelet function. Bone marrow aspiration or biopsy reveals hypercellularity with relatively normal granulocytic maturation and no significant increase in leukemic blasts. Although reduced leukocyte alkaline phosphatase activity is seen in CML, this is not a specific finding.

CML has three clinical phases: chronic, accelerated, and blast crisis. Chronic phase, which lasts for approximately 3 years if untreated, usually presents with side effects secondary to hyperleukocytosis such as weakness, fever, night sweats, bone pain, respiratory distress, priapism, left upper quadrant pain (splenomegaly), and, rarely, hearing loss and visual disturbances. The accelerated phase is characterized by progressive splenomegaly, thrombocytopenia, and increased percentage of peripheral and bone marrow blasts, along with accumulation of karyotypic abnormalities in addition to the Ph chromosome. Blast crisis is notable for the bone marrow, showing greater than 30% blasts and a clinical picture that is indistinguishable from acute leukemia. Approximately two-thirds of blast crisis is myeloid, and the remainder is lymphoid, usually of B lineage. Patients in blast crisis will die within a few months.

Before the tyrosine kinase inhibitor (TKI) era, allogeneic hematopoietic stem cell transplantation (HSCT) was the primary treatment for children with CML. Published reports from this period described survival rates of 70% to 80% when an HLA-matched family donor (MFD) was used in the treatment of children in early chronic phase, with lower survival rates when HLA-matched unrelated donors were used. Relapse rates were low (less than 20%) when transplant was performed in chronic phase. The primary cause of death was treatment-related mortality, which was increased with HLA-matched unrelated donors compared with HLA-MFDs in most reports. High-resolution DNA matching for HLA alleles appeared to reduce rates of treatment-related mortality leading to improved outcome for HSCT using unrelated donors. Compared with transplantation in chronic phase, transplantation in accelerated phase or blast crisis and in second-chronic phase resulted in significantly reduced survival. The use of T-lymphocyte depletion to avoid graft-versus-host disease resulted in a higher relapse rate and decreased overall survival (OS), supporting the contribution of a graft-versus-leukemia effect to favorable outcome after allogeneic HSCT.

The introduction of the TKI imatinib (Gleevec) as a therapeutic drug targeted at inhibiting the BCR-ABL fusion kinase revolutionized the treatment of patients with CML, for both children and adults. As most data for the use of TKIs for CML is from adult clinical trials, the adult experience is initially described, followed by a description of the more limited experience for children.

Treatment of CML in Adults with TKIs

Imatinib is a potent inhibitor of the ABL tyrosine kinase, and also of PDGF receptors (alpha and beta) and KIT. Imatinib treatment achieves clinical, cytogenetic, and molecular remissions (as defined by the absence of BCR-ABL fusion transcripts) in a high proportion of CML patients treated in chronic phase. Imatinib replaced the use of alpha-interferon in the initial treatment of CML based on the results of a large phase III trial comparing imatinib with interferon plus cytarabine (IRIS). Patients receiving imatinib had higher complete cytogenetic response rates (76% vs. 14% at 18 months), and the rate of treatment failure diminished over time, from 3.3% and 7.5% in the first and second years of imatinib treatment, respectively, to less than 1% by the fifth year of treatment. After censoring for patients who died from causes unrelated to CML or transplantation, the overall estimated survival rate for patients randomly assigned to imatinib was 95% at 60 months.

Guidelines for imatinib treatment have been developed for adults with CML based on patient response to treatment, including the timing of achieving complete hematologic response, complete cytogenetic response, and major molecular response (defined as attainment of a 3-log reduction in /control gene ratio). The identification of kinase domain mutations at the time of failure or of suboptimal response to imatinib treatment also has clinical implications, as there are alternative BCR-ABL kinase inhibitors (e.g., dasatinib and nilotinib) that maintain their activity against some (but not all) mutations that confer resistance to imatinib. Poor adherence is a major reason for loss of complete cytogenetic response and imatinib failure for adult CML patients on long-term therapy.

Two other TKIs, dasatinib and nilotinib, have been shown to be effective in patients who have an inadequate response to imatinib, although not in patients with the mutation. Both dasatinib and nilotinib have also received regulatory approval for the treatment of newly diagnosed chronic-phase CML in adults, on the basis of the following studies:

  • Dasatinib was approved on the basis of a phase III trial that compared dasatinib (100 mg daily) with imatinib (400 mg daily). There was no significant difference in progression-free survival (PFS) or OS. However, after 12 months of treatment, dasatinib was associated with a higher rate of complete cytogenetic response (83% vs. 72%, = .001) and major molecular response (46% vs. 28%, < .0001). Responses were achieved in a shorter time with dasatinib ( < .0001).
  • Nilotinib (at a dose of either 300 mg or 400 mg twice daily) was compared with imatinib (400 mg daily) in a phase III trial. At 12 months, the rates of complete cytogenetic response were significantly higher for nilotinib (80% for the 300-mg dose and 78% for the 400-mg dose) than were the rates for imatinib (65%) ( < .001 for both comparisons). Also, nilotinib was associated with higher rates of major molecular response (44% for the 300-mg dose and 43% for the 400-mg dose compared with 22% for imatinib, < .001 for both comparisons). The 300 mg twice-daily dose of nilotinib was associated with a more favorable safety profile compared with the 400-mg dose.

Because of the superiority over imatinib in terms of complete cytogenetic response rate and major molecular response rate, both dasatinib and nilotinib are extensively used as firstline therapy in adults with CML. However, despite more rapid responses with dasatinib and nilotinib than with imatinib when used as frontline therapy, PFS and OS appear to be similar for all three agents. Additional follow-up will be required to better define the impact of these agents on long-term PFS and OS.

Bosutinib is another TKI that targets the fusion and has been approved by the U.S. Food and Drug Administration (FDA) for the treatment of all phases of CML in adults who show intolerance to or whose disease shows resistance to previous therapy with another TKI.

Ponatinib is a inhibitor that is effective against the mutation. Ponatinib induced objective responses in approximately 70% of heavily pretreated adults with chronic-phase CML, with responses observed regardless of the baseline kinase domain mutation. Development of ponatinib has been complicated by the high rate of vascular occlusion observed in patients receiving the agent, with arterial and venous thrombosis and occlusions (including myocardial infarction and stroke) occurring in more than 20% of treated patients. Neither bosutinib nor ponatinib have been studied in the pediatric population.

For adult CML patients who proceed to allogeneic HSCT, there is no evidence that pretransplant imatinib adversely impacts outcome. A retrospective study that compared 145 patients who received imatinib before transplant with a historical cohort of 231 patients showed no difference in early hepatic toxic effects or engraftment delay. In addition, OS, disease-free survival, relapse, and nonrelapse mortality were similar between the two cohorts. The only factor associated with poor outcome in the cohort that received imatinib was a poor initial response to imatinib. Further evidence for a lack of effect of pretransplant imatinib on posttransplant outcomes was supplied by a report from the Center for International Blood and Marrow Transplant Research; this report compared outcomes of 181 pediatric and adult subjects with CML in first chronic phase treated with imatinib before HSCT with that of 657 subjects who did not receive imatinib before HSCT. Among the patients in first chronic phase, imatinib therapy before HSCT was associated with better OS. A third report of imatinib followed by allogeneic HSCT supports the efficacy of this transplantation strategy in patients with imatinib failure in first chronic phase; the 3-year OS rate was 94% for this group (n = 37), with approximately 90% achieving a complete molecular remission after HSCT.

For adult patients treated with a TKI alone (without HSCT), the optimal duration of therapy remains unknown and most patients continue TKI treatment indefinitely. However, in an attempt to answer the question of length of treatment, a prospective study reported on 69 adults treated with imatinib for more than 2 years who had been in a cytogenetic major response for more than 2 years. The patients were followed monthly and restarted on imatinib if there was evidence of molecular relapse. Of this group, 61% experienced disease relapse, with about 38% still in cytogenetic major response at 24 months. Of note, all of the patients who had disease recurrence responded again to the reinitiation of imatinib. Another study reported on 40 chronic-phase CML patients who stopped treatment with imatinib after at least 2 years of sustained undetectable minimal residual disease (MRD) by polymerase chain reaction (PCR). At 24 months, the probability of sustained molecular remission for patients no longer receiving imatinib was 47.1%. Most relapses occurred within 4 months of stopping treatment with imatinib, and no relapses beyond 27 months were observed. All patients with molecular relapse demonstrated a favorable response when imatinib was restarted; with a median follow-up of 42 months, no patients had progressive disease or developed the fusion. Further research is required before cessation of imatinib or other targeted therapy for selected patients with CML in molecular remission can be recommended as a standard clinical practice.

Treatment of CML in Children

Imatinib has shown a high level of activity in children with CML that is comparable with that observed in adults. In a prospective trial of 44 pediatric patients with newly diagnosed CML treated with imatinib (260 mg/day), the PFS rate at 36 months was 98%. A complete hematologic response was achieved in 98% of the patients. The rate of complete cytogenetic response was 61% and the rate of major molecular response was 31% at 12 months, similar to the rates seen in adult chronic-phase CML patients treated with imatinib. As a result of this high level of activity, it is common to initiate treatment of children with CML with imatinib rather than proceeding immediately to allogeneic stem cell transplantation. The pharmacokinetics of imatinib in children appears consistent with prior results in adults.

Doses of imatinib used in phase II trials for children with CML have ranged from 260 mg/m to 340 mg/m, which provide comparable drug exposures as the adult flat-doses of 400 mg to 600 mg. In an Italian study of 47 pediatric chronic-phase CML patients treated with 340 mg/m per day of imatinib, complete cytogenetic response was achieved in 91.5% of patients at a median time of 6 months, and the rate of major molecular response at 12 months was 66.6%. Thus, it appears that starting with the higher dose of 340 mg/m has superior efficacy and is typically tolerable, with dose adjustment as needed for toxicity. Also, early molecular responses, such as PCR-based MRD measurement at 3 months of therapy showing up to 10% BCR-ABL1/ABL, have been reported to be associated with improved PFS, similar to early molecular response data in adults. The monitoring guidelines described above for adults with CML are reasonable to utilize in children.

Imatinib is generally well tolerated in children, with adverse effects usually being mild to moderate and quickly reversible with treatment discontinuation or dose reduction. Growth retardation occurs in some children receiving imatinib. The growth inhibitory effects of imatinib appear to be most pronounced in prepubertal children, compared with pubertal children; children receiving imatinib and experiencing growth impairment may show a return to normal growth rates when they reach puberty.

There are fewer published data regarding the efficacy and toxicities of other TKIs in children with CML. A phase I trial of dasatinib in children showed that drug disposition, tolerability, and efficacy of this agent for patients with CML was similar to that observed in adults. A safe pediatric dose of the other TKIs (nilotinib, bosutinib, ponatinib) has not yet been established.

In children who develop a hematologic or cytogenetic relapse on imatinib or who have an inadequate initial response to imatinib, determination of kinase domain mutation status should be considered to help guide subsequent therapy. Depending upon the patient’s mutation status, alternative kinase inhibitors such as dasatinib or nilotinib can be considered based on adult experience with these agents. A pediatric phase I study of dasatinib showed good tolerance for dasatinib in children at doses used to treat adults with CML, and nilotinib is under investigation in children with CML or Ph chromosome–positive acute lymphoblastic leukemia (ALL) (). These agents are active against many mutants that confer resistance to imatinib, although the agents are ineffective in patients with the mutation. In the presence of the mutation, which is resistant to all FDA-approved kinase inhibitors, strong consideration should be given to performing an allogeneic transplant.

The question of whether a pediatric patient with CML should receive an allogeneic transplant when multiple TKIs are available remains unanswered; however, recent reports suggest that PFS does not improve when utilizing HSCT, compared with the sustained use of imatinib. The potential advantages and disadvantages need to be discussed with the patient and family. While HSCT is currently the only known definitive curative therapy for CML, patients discontinuing treatment with TKIs after sustained molecular remissions, who remained in molecular remission, have been reported.

Treatment Options Under Clinical Evaluation

Based on their activity in adults with CML, other TKIs are being studied in children. Dasatinib has undergone phase I testing in children and showed drug disposition, tolerability, and efficacy for patients with CML that was similar to that observed in adults. Nilotinib is under investigation in children with CML or Ph chromosome–positive (Ph+) ALL in a clinical trial to determine the pharmacokinetics of nilotinib in children ( [CAMN107A2120]). A phase II evaluation of nilotinib in children with CML has been initiated ().

The following are examples of national and/or institutional clinical trials that are currently being conducted for patients with CML. Information about ongoing clinical trials is available from the NCI website.

  • (A Pharmacokinetic Study of Nilotinib in Pediatric Patients With Ph+ CML or ALL): A clinical trial is assessing the pharmacokinetics of nilotinib in Ph+ CML pediatric patients that are newly diagnosed or resistant or intolerant to imatinib or dasatinib or refractory or relapsed Ph+ ALL. Efficacy and safety are being evaluated as secondary objectives.
  • (Open Label, Phase II Study to Evaluate Efficacy and Safety of Oral Nilotinib in Ph+ CML Pediatric Patients): A phase II clinical trial of nilotinib is evaluating the safety and efficacy of nilotinib in the Ph+ CML in pediatric patients (aged 1 to <18 years).

Current Clinical Trials

Check the list of NCI-supported cancer clinical trials that are now accepting patients with childhood chronic myelogenous leukemia. The list of clinical trials can be further narrowed by location, drug, intervention, and other criteria.

General information about clinical trials is also available from the NCI website.

Recurrent Childhood AML and Other Myeloid Malignancies

The diagnosis of recurrent or relapsed acute myeloid leukemia (AML) according to Children's Oncology Group (COG) criteria is essentially the same as the criteria for making the diagnosis of AML. Usually this is defined as patients having more than 5% bone marrow blasts and a diagnosis of AML according to World Health Organization (WHO) classification criteria.

Despite second remission induction in over one-half of children with AML treated with drugs similar to drugs used in initial induction therapy, the prognosis for a child with recurrent or progressive AML is generally poor. Approximately 50% to 60% of relapses occur within the first year after diagnosis, with most relapses occurring by 4 years from diagnosis. The vast majority of relapses occur in the bone marrow, with central nervous system (CNS) relapse being very uncommon. Length of first remission is an important factor affecting the ability to attain a second remission; children with a first remission of less than 1 year have substantially lower rates of remission than children whose first remission is greater than 1 year (50%–60% vs. 70%–90%, respectively). Survival for children with shorter first remissions is also substantially lower (approximately 10%) than that for children with first remissions exceeding 1 year (approximately 40%). In addition, specific molecular alterations at the time of relapse have been reported to impact subsequent survival. For instance, the presence of either or mutations at first relapse were associated as independent risk factors for worse overall survival (OS) in patients achieving a second remission.

Regimens that have been successfully used to induce remission in children with recurrent AML have commonly included high-dose cytarabine given in combination with other agents, such as mitoxantrone, fludarabine and idarubicin, L-asparaginase, etoposide, and clofarabine and etoposide. Regimens built upon clofarabine have also been used;[] as have regimens of 2-chloroadenosine. The COG trial evaluated the combination of clofarabine plus high-dose cytarabine in patients with relapsed AML; the response rate was 48% and the OS rate, with 21 of 23 responders undergoing hematopoietic stem cell transplantation (HSCT), was 46%. Minimal residual disease (MRD) before HSCT was a strong predictor of survival.[] The standard-dose cytarabine regimens used in the United Kingdom Medical Research Council AML 10 study for newly diagnosed children with AML (cytarabine and daunorubicin plus either etoposide or thioguanine) have, when used in the setting of relapse, produced remission rates similar to those achieved with high-dose cytarabine regimens. In a COG phase II study, the addition of bortezomib to idarubicin plus low-dose cytarabine resulted in an overall complete remission (CR) rate of 57%, and the addition of bortezomib to etoposide and high-dose cytarabine resulted in an overall CR rate of 48%.

In a report of 379 children with AML who relapsed after initial treatment on the German Berlin-Frankfurt-Münster (BFM) group protocols, a second complete remission (CR2) rate was 63% and OS was 23%.[] The most significant prognostic factors associated with a favorable outcome after relapse included achieving CR2, a relapse greater than 12 months from initial diagnosis, no allogeneic bone marrow transplant in first remission, and favorable cytogenetics (t(8;21), t(15;17), and inv(16)). A subsequent study by the BFM group compared fludarabine, cytarabine, and granulocyte colony-stimulating factor (FLAG) with FLAG plus liposomal daunorubicin. Four-year OS was 38%, with no difference in survival for the total group; however, the addition of liposomal daunorubicin increased the likelihood of obtaining a remission and led to significant improvement in OS in patients with core-binding factor mutations (82%, FLAG plus liposomal daunorubicin vs. 58%, FLAG; = .04).[] The international (NCT00186966) trial also found that early response to salvage therapy was highly prognostic.[] The Therapeutic Advances in Childhood Leukemia and Lymphoma Consortium also identified duration of previous remission as a powerful prognostic factor, with 5-year OS rates of 54% ± 10% for patients with greater than 12 months first remission duration and 19% ± 6% for patients with shorter periods of first remission. A retrospective study of 71 patients with relapsed AML from Japan reported a 5-year OS rate of 37%. Patients who had an early relapse had a 27% second remission rate compared with 88% for patients who had a late relapse. The 5-year OS rate was higher in patients who went to HSCT after achieving a CR2 (66%) than in patients not in remission (17%).

The selection of further treatment after the achievement of a second remission depends on previous treatment as well as individual considerations. Consolidation chemotherapy followed by HSCT is conventionally recommended, though there are no controlled prospective data regarding the contribution of additional courses of therapy once CR2 is obtained. Unrelated donor HSCT has been reported to result in 5-year probabilities of leukemia-free survival of 45%, 20%, and 12% for patients with AML transplanted in CR2, overt relapse, and primary induction failure, respectively.[] The optimal type of transplant preparative regimen and source of donor cells has not been determined, although alternative donor sources, including haploidentical donors, are being studied. A number of studies, including a large, prospective Center for International Blood and Marrow Transplant Research (CIBMTR) cohort study of children and adults with myeloid diseases, have shown similar or superior survival with busulfan-based regimens compared with total-body irradiation (TBI).

There is evidence that long-term survival can be achieved in a portion of pediatric patients who undergo a second transplant subsequent to relapse after a first myeloablative transplant. Survival was associated with late relapse (>6 months from first transplant), achievement of complete response before the second procedure, and use of a TBI-based regimen (after receiving a non-TBI regimen for the first procedure). A large prospective cohort study that included children and adults with myeloid diseases showed comparable or superior outcome with busulfan-based regimens compared with TBI.

Clinical trials, including new chemotherapy and/or biologic agents and/or novel bone marrow transplant (autologous, matched or mismatched unrelated donor, cord blood) programs, are also considerations. Information about ongoing clinical trials is available from the NCI website.

Relapse in Children With Down Syndrome

A small number of publications address outcomes in children with Down syndrome who relapse after initial therapy or who have refractory AML. The Japanese Pediatric Leukemia/Lymphoma Study Group reported the outcomes of 29 Down syndrome patients with relapsed (n = 26) or refractory (n = 3) AML. As expected with Down syndrome, the children in this cohort were very young (median age, 2 years); relapses were almost all early (median 8.6 months, 80% <12 months from diagnosis); and 89% had M7 French-American-British classification. In contrast to the excellent outcomes achieved after initial therapy, only 50% of the children attained a second remission, and the 3-year OS rate was 26%.[] Approximately one-half of the children underwent allogeneic transplant, and no advantage was noted with transplant compared with chemotherapy, but numbers were small. A CIBMTR study of children with Down syndrome and AML who underwent HSCT reported a similarly poor outcome, with a 3-year OS of 19%.[] The main cause of failure after transplant was relapse, which exceeded 60%; transplant-related mortality was approximately 20%. A Japanese registry study reported better survival after transplant of children with Down Syndrome using reduced intensity conditioning regimens compared with myeloablative approaches, but numbers were very small (n = 5), and the efficacy of reduced intensity approaches in Down children with AML requires further study.[]

Isolated CNS Relapse

Isolated CNS relapse occurs in 3% to 5% of pediatric AML patients. Factors associated with an increased risk of isolated CNS relapse include the following:

  • Age younger than 2 years at initial diagnosis.
  • M5 leukemia.
  • 11q23 abnormalities.
  • CNS involvement at initial diagnosis.

The outcome of isolated CNS relapse when treated as a systemic relapse is similar to that of bone marrow relapse. In one study, the 8-year OS for a cohort of children with an isolated CNS relapse was 26% ± 16%.

Recurrent Acute Promyelocytic Leukemia (APL)

Despite the improvement in outcomes for patients with newly diagnosed APL, approximately 10% to 20% of patients relapse.

An important issue in children is the previous exposure to anthracyclines, which can range from 400 mg/m to 750 mg/m. Thus, regimens containing anthracyclines are often not optimal for children with APL who suffer relapse. For children with recurrent APL, the use of arsenic trioxide as a single agent or regimens including all- retinoic acid should be considered, depending on the therapy given during first remission. Arsenic trioxide is an active agent in patients with recurrent APL, with approximately 85% of patients achieving remission after treatment with this agent. Data are limited on the use of arsenic trioxide in children, although published reports suggest that children with relapsed APL have a response to arsenic trioxide similar to that of adults. Because arsenic trioxide causes QT-interval prolongation that can lead to life-threatening arrhythmias, it is essential to monitor electrolytes closely in patients receiving arsenic trioxide and to maintain potassium and magnesium values at midnormal ranges. The use of anti-CD33/calicheamicin monoclonal antibody (gemtuzumab ozogamicin) as a single agent resulted in 91% (9 of 11 patients) molecular remission after two doses and in 100% of patients (13 of 13) after three doses, thus demonstrating excellent activity of this agent in relapsed APL. Gemtuzumab ozogamicin is currently not available in the United States, except for compassionate-use approval.

Retrospective pediatric studies have reported 5-year event-free survival (EFS) rates after either autologous or allogeneic transplantation approaches to be similar at approximately 70%. When considering autologous transplantation, a study in adult patients demonstrated improved 7-year EFS (77% vs. 50%) when both the patient and the stem cell product had negative promyelocytic leukemia/retinoic acid receptor alpha fusion transcript by polymerase chain reaction (molecular remission) before transplant. Another study demonstrated that among seven patients undergoing autologous HSCT and whose cells were minimal residual disease (MRD)-positive, all relapsed in less than 9 months after transplantation; however, only one of eight patients whose autologous donor cells were MRD-negative relapsed. Another report demonstrated that the 5-year EFS was 83.3% for patients who underwent autologous HSCT in second molecular remission and was 34.5% for patients who received only maintenance therapy. Such data support the use of autologous transplantation in patients who are MRD-negative in second complete remission who have poorly matched allogeneic donors.

Current Clinical Trials

Check the list of NCI-supported cancer clinical trials that are now accepting patients with recurrent childhood acute myeloid leukemia. The list of clinical trials can be further narrowed by location, drug, intervention, and other criteria.

General information about clinical trials is also available from the NCI website.

Survivorship and Adverse Late Sequelae

While the issues of long-term complications of cancer and its treatment cross many disease categories, several important issues related to the treatment of myeloid malignancies are worth stressing. (Refer to the PDQ summary on Late Effects of Treatment for Childhood Cancer for more information.)

The Children’s Cancer Survivor Study examined 272 survivors of childhood acute myeloid leukemia (AML) who did not undergo a hematopoietic stem cell transplant (HSCT). This study identified second malignancies (cumulative incidence, 1.7%) and cardiac toxic effects (cumulative incidence, 4.7%) as significant long-term risks. Cardiomyopathy has been reported in 4.3% of survivors of AML based on Berlin-Frankfurt-Münster studies. Of these, 2.5% showed clinical symptoms. A retrospective study of cardiac function of children treated with United Kingdom Medical Research Council-based regimens at a median of 13 months after treatment reported a mean detrimental change in left ventricular stroke volume of 8.4% compared with baseline values. For pediatric patients, the risk of developing early toxicity was 13.7%, and the risk of developing late cardiac toxic effects (defined as 1 year after completing first-line therapy) was 17.4%. Early cardiac toxic effects was a significant predictor of late cardiac toxic effects and the development of clinical cardiomyopathy requiring long-term therapy. Retrospective analysis of a single study suggests cardiac risk may be increased in children with Down syndrome, but prospective studies are required to confirm this finding.

Renal, gastrointestinal, and hepatic late adverse effects have been reported to be rare for children undergoing chemotherapy only for treatment of AML. A Nordic Society for Pediatric Hematology and Oncology retrospective trial of children treated for AML with chemotherapy only at a median follow-up of 11 years, based on a self-reported use of health care services, demonstrated similar health care usage and marital status as their siblings. A COG study using a health-related, quality-of-life comparison reported an overall 21% of 5-year survivors having a severe or life-threatening chronic health condition; when compared by type of treatment, this percentage was 16% for the chemotherapy-only treated group, 21% for the autologous HSCT treated group, and 33% for those who received an allogeneic HSCT.

In a review from one institution, the highest frequency of adverse long-term sequelae for children treated for AML included the following incidence rates: growth abnormalities (51%), neurocognitive abnormalities (30%), transfusion-acquired hepatitis (28%), infertility (25%), endocrinopathies (16%), restrictive lung disease (20%), chronic graft-versus-host disease (20%), secondary malignancies (14%), and cataracts (12%). Most of these adverse sequelae are the consequence of myeloablative, allogeneic HSCT. Although cardiac abnormalities were reported in 8% of patients, this is an issue that may be particularly relevant with the current use of increased anthracyclines in clinical trials for children with newly diagnosed AML. Another study examined outcomes for children younger than 3 years with AML or acute lymphoblastic leukemia (ALL) who underwent HSCT. The toxicities reported include growth hormone deficiency (59%), dyslipidemias (59%), hypothyroidism (35%), osteochondromas (24%), and decreased bone mineral density (24%). Two of the 33 patients developed secondary malignancies. Survivors had average intelligence but frequent attention-deficit problems and fine-movement abnormalities, compared with population controls. In contrast, The Bone Marrow Transplant Survivor Study compared childhood AML or ALL survivors with siblings using a self-reporting questionnaire. The median follow-up was 8.4 years, and 86% of patients received total-body irradiation (TBI) as part of their preparative transplant regimen. Survivors of leukemia who received an HSCT had significantly higher frequencies of several adverse effects, including diabetes, hypothyroidism, osteoporosis, cataracts, osteonecrosis, exercise-induced shortness of breath, neurosensory impairments and problems with balance, tremor, and weakness than siblings. The overall assessment of health was significantly decreased in survivors compared with siblings (odds ratio = 2.2; = .03). Significant differences were not observed between regimens using TBI compared with chemotherapy only, which mostly included busulfan. The outcomes were similar for patients with AML and ALL, suggesting that the primary cause underlying the adverse late effects was undergoing an HSCT.

A population-based study of survivors of childhood AML who had not undergone an HSCT reported equivalent rates of educational achievement, employment, and marital status compared with siblings. AML survivors were, however, significantly more likely to be receiving prescription drugs, especially for asthma, compared with siblings (23% vs. 9%; = .03). Chronic fatigue has also been demonstrated to be a significantly more likely adverse late effect in survivors of childhood AML than in survivors of other malignancies.

New therapeutic approaches to reduce long-term adverse sequelae are needed, especially for reducing the late sequelae associated with myeloablative HSCT.

Important resources for details on follow-up and risks for survivors of cancer have been developed by the Children Oncology Group’s Long-Term Follow-Up Guidelines for Survivors of Childhood, Adolescent, and Young Adult Cancers and the National Comprehensive Cancer Network Guidelines for Acute Myeloid Leukemia. Furthermore, having access to past medical history that can be shared with subsequent medical providers has become increasingly recognized as important for cancer survivors. Different templates that address this issue are available, such as the Cancer Survivor's Treatment Record and the Cancer Survivor’s Medical Treatment Summary.

Changes to This Summary (04/04/2017)

The PDQ cancer information summaries are reviewed regularly and updated as new information becomes available. This section describes the latest changes made to this summary as of the date above.

Classification of Pediatric Myeloid Malignancies

Revised text to state that both and subtypes commonly show mutations in genes that activate receptor tyrosine kinase signaling; and are the most commonly mutated genes for both subtypes. Also added text to state that mutations may indicate increased risk of treatment failure for patients with core-binding factor AML, although the prognostic significance of mutations may be dependent on the mutant-allele ratio and/or the specific type of mutation. exon 17 mutations are enriched in patients with fusions when compared with patients with fusions (cited Faber et al. and Eisfeld et al. as references 41 and 42, respectively).

Revised text to state that cases with and cases with have distinctive secondary mutations; secondary mutations are primarily restricted to genes that activate receptor tyrosine kinase signaling.

This summary is written and maintained by the PDQ Pediatric Treatment Editorial Board, which is editorially independent of NCI. The summary reflects an independent review of the literature and does not represent a policy statement of NCI or NIH. More information about summary policies and the role of the PDQ Editorial Boards in maintaining the PDQ summaries can be found on the About This PDQ Summary and PDQ® - NCI's Comprehensive Cancer Database pages.

About This PDQ Summary

Purpose of This Summary

This PDQ cancer information summary for health professionals provides comprehensive, peer-reviewed, evidence-based information about the treatment of childhood acute myeloid leukemia and other myeloid malignancies. It is intended as a resource to inform and assist clinicians who care for cancer patients. It does not provide formal guidelines or recommendations for making health care decisions.

Reviewers and Updates

This summary is reviewed regularly and updated as necessary by the PDQ Pediatric Treatment Editorial Board, which is editorially independent of the National Cancer Institute (NCI). The summary reflects an independent review of the literature and does not represent a policy statement of NCI or the National Institutes of Health (NIH).

Board members review recently published articles each month to determine whether an article should:

  • be discussed at a meeting,
  • be cited with text, or
  • replace or update an existing article that is already cited.

Changes to the summaries are made through a consensus process in which Board members evaluate the strength of the evidence in the published articles and determine how the article should be included in the summary.

The lead reviewers for Childhood Acute Myeloid Leukemia/Other Myeloid Malignancies Treatment are:

  • Alan Scott Gamis, MD, MPH (Children's Mercy Hospital)
  • Karen J. Marcus, MD (Dana-Farber Cancer Institute/Boston Children's Hospital)
  • Michael A. Pulsipher, MD (Children's Hospital Los Angeles)
  • Lewis B. Silverman, MD (Dana-Farber Cancer Institute/Boston Children's Hospital)
  • Malcolm A. Smith, MD, PhD (National Cancer Institute)

Any comments or questions about the summary content should be submitted to through the NCI website's Email Us. Do not contact the individual Board Members with questions or comments about the summaries. Board members will not respond to individual inquiries.

Levels of Evidence

Some of the reference citations in this summary are accompanied by a level-of-evidence designation. These designations are intended to help readers assess the strength of the evidence supporting the use of specific interventions or approaches. The PDQ Pediatric Treatment Editorial Board uses a formal evidence ranking system in developing its level-of-evidence designations.

Permission to Use This Summary

PDQ is a registered trademark. Although the content of PDQ documents can be used freely as text, it cannot be identified as an NCI PDQ cancer information summary unless it is presented in its entirety and is regularly updated. However, an author would be permitted to write a sentence such as “NCI’s PDQ cancer information summary about breast cancer prevention states the risks succinctly: [include excerpt from the summary].”

The preferred citation for this PDQ summary is:

PDQ® Pediatric Treatment Editorial Board. PDQ Childhood Acute Myeloid Leukemia/Other Myeloid Malignancies Treatment. Bethesda, MD: National Cancer Institute. Updated . Available at: Accessed . [PMID: 26389454]

Images in this summary are used with permission of the author(s), artist, and/or publisher for use within the PDQ summaries only. Permission to use images outside the context of PDQ information must be obtained from the owner(s) and cannot be granted by the National Cancer Institute. Information about using the illustrations in this summary, along with many other cancer-related images, is available in Visuals Online, a collection of over 2,000 scientific images.


Based on the strength of the available evidence, treatment options may be described as either “standard” or “under clinical evaluation.” These classifications should not be used as a basis for insurance reimbursement determinations. More information on insurance coverage is available on on the Managing Cancer Care page.

Contact Us

More information about contacting us or receiving help with the website can be found on our Contact Us for Help page. Questions can also be submitted to through the website’s Email Us.


Adam Goodwin
by Bob Riter
September 21, 2015

Stay informed with the latest information from OncoLink!   Subscribe to OncoLink eNews
View our newsletter archives