The acute myeloid leukemias (AML) represent a heterogeneous group of malignancies derived from the pluripotent hematopoietic stem cell. These leukemias are generally characterized by genetic lesions that result in a combination of defects causing unregulated proliferation of cells and defects in cellular maturation (Gilliland and Griffin 2002). AML accounts for approximately 15–20% of acute leukemia in children. In contrast to acute lymphoblastic leukemia (ALL) in childhood, for which an age-related peak incidence in children is associated with unique genetics, biology, and response to therapy, AML in children is very heterogeneous with large subsets representing disease that is generally similar to that in adults. Pediatric AML does not exhibit a dramatic peak in childhood other than for infants with disease involving translocation of the mixed lineage leukemia (MLL) gene. In both children and adults, AML is a relatively drug-resistant disease. Progress in improving outcome over the past 40 years has been associated with the use of pulses of high-dose, high systemic exposure intensive chemotherapy approaches. Refinements in chemotherapy regimens and major improvements in supportive care practices have resulted in the ability to achieve complete remission (CR) in 80–90% of pediatric patients and long-term event free survival (EFS) in 40–60% of patients. While high-dose chemotherapy Consolidation with hematopoietic stem cell transplantation (HSCT) once represented the predominant treatment approach, recent studies have revealed that large subgroups of patients characterized by specific cytogenetic and molecular features do not require transplantation as initial therapy. Conversely, other molecular analyses have identified very high risk subgroups at the time of initial diagnosis that possess highly resistant stem cell disease and are likely to benefit from stem cell transplantation. The acute myeloid leukemias (AML) represent a heterogeneous group of malignancies derived from the pluripotent hematopoietic stem cell. These leukemias are generally characterized by genetic lesions that result in a combination of defects causing unregulated proliferation of cells and defects in cellular maturation (Gilliland and Griffin 2002). AML accounts for approximately 15–20% of acute leukemia in children. In contrast to acute lymphoblastic leukemia (ALL) in childhood, for which an age-related peak incidence in children is associated with unique genetics, biology, and response to therapy, AML in children is very heterogeneous with large subsets representing disease that is generally similar to that in adults. Pediatric AML does not exhibit a dramatic peak in childhood other than for infants with disease involving translocation of the mixed lineage leukemia (MLL) gene. In both children and adults, AML is a relatively drug-resistant disease. Progress in improving outcome over the past 40 years has been associated with the use of pulses of high-dose, high systemic exposure intensive chemotherapy approaches. Refinements in chemotherapy regimens and major improvements in supportive care practices have resulted in the ability to achieve complete remission (CR) in 80–90% of pediatric patients and long-term event free survival (EFS) in 40–60% of patients. While high-dose chemotherapy Consolidation with hematopoietic stem cell transplantation (HSCT) once represented the predominant treatment approach, recent studies have revealed that large subgroups of patients characterized by specific cytogenetic and molecular features do not require transplantation as initial therapy. Conversely, other molecular analyses have identified very high risk subgroups at the time of initial diagnosis that possess highly resistant stem cell disease and are likely to benefit from stem cell transplantation. The key recent advances in treatment of AML have reflected the recent identification of molecular and cytogenetic risk groups and the development of new risk-adapted therapy approaches. In accordance with these findings, clinical and morphological classifications for AML in children and adults have now been largely replaced by cytogenetic and molecular risk classification schemes, such as that outlined by the World Health Organization (WHO) in collaboration with the Society for Hematopathology and the European Association of Haematopathology (Vardiman et al. 2009). These new classification approaches have not only effectively identified patients who benefit from different chemotherapy approaches, but they also hold promise to help identify subsets of patients who may potentially benefit from targeted molecular therapies. The clinical presentation of AML most commonly reflects the consequences of diffuse leukemic infiltration of the bone marrow. The suppression of normal hematopoiesis results in laboratory findings of anemia and thrombocytopenia with leukopenia or leukocytosis and variable numbers of blasts in the peripheral blood. Associated symptoms include anemia, fevers, and bleeding, with relatively frequent findings of organ infiltration including hepatosplenomegaly, skin involvement, and thromboses. Common clinical features are summarized in Table 5.1. Hepatosplenomegaly is noted in approximately 15–30% of AML patients (Hurwitz et al. 1993; Lange et al. 2008). Approximately 5–10% of patients present with disease involving the central nervous system (CNS) (Hurwitz et al. 1993; Lie et al. 1996; Hann et al. 1997; Woods et al. 2001; Lie et al. 2003). Monocytic leukemias are characterized by skin involvement as well as gingival swelling. Leukemias with the t(8;21) translocation and FAB-M2 morphology can present with extramedullary mass disease in the orbits, skull and other bones, skin, lymph nodes, gastrointestinal and genitourinary tracts, and brain (Tallman et al. 1993; Krishnan et al. 1994; Pui et al. 1994). Table 5.1 Clinical features of Down syndrome and non-Down syndrome AML. Patients enrolled onto CCG 2891(1989–1999) (From Gamis et al. 2003) Down syndrome Non–Down syndrome p-value Patients enrolled (No.) 161 947 Median age (years) 1.8 7.5 <0.001 Enlarged liver (%) 50.6 40.2 0.015 Enlarged spleen (%) 49.4 41.2 0.057 Enlarged nodes (%) 25.6 45.2 <0.001 CNS disease (+CSF) 4.5 20.1 <0.001 WBC Median/mm3 6,800 19,900 <0.001 <20,000 (%) 80 50.1 <0.001 20–100,000 (%) 19.4 33.8 >100,000 (%) 0.6 16.2 BM blast (%) 0–5 2.6 3.5 <0.001 5–30 31.6 13.2 >30 65.8 83.2 Median platelets/mm3 26,000 53,000 <0.001 The development of risk-adapted therapies has provided extraordinary improvements in the outcome for children and adults with AML. The accurate diagnosis and classification of AML is essential for the selection of appropriate effective therapy regimens and interpretation of results from clinical trials. Including clinical and laboratory experts from the United States and Europe, the European LeukemiaNet recently issued expert opinion-based recommendations for the use and interpretation of laboratory studies in the diagnosis and management of AML (Dohner et al. 2010). A bone marrow aspirate and biopsy should generally be a central part of the routine diagnostic evaluation of potential AML unless the patient’s clinical status precludes the procedure. Marrows with extensive leukemic infiltration or fibrosis often have technically suboptimal aspirate specimens which are either unable to be successfully aspirated or diluted by peripheral blood. Bone marrow biopsies and touch imprints have particular utility in these circumstances and are often critical for establishing a diagnosis. Standard evaluation should be conducted using a May-Grunwald-Giemsa or a Wright-Giemsa stain with evaluation and counting of a minimum of 200 leukocytes on the blood smears and 500 nucleated cells on marrow smears (Dohner et al. 2010). In addition, a lumbar puncture should be conducted for the assessment of CNS involvement and presymptomatic intrathecal therapy. Some patients may have clinical complications (e.g., coagulopathy or hypotension) that necessitate a delay in evaluation of the CNS. As discussed above, AML is characterized by dysregulated proliferation and arrest of differentiation along pathways normally found in hematopoietic development. The French–American–British (FAB) classification system was developed in 1972 and provided the primary morphologic classification of acute myeloid leukemia (Bennett et al. 1976; Bennett et al. 1985). The FAB classification system is based on cytochemical and morphologic analyses of blasts and includes ten different subtypes (Table 5.2). Use of the FAB classification in conjunction with early therapies helped to identify key therapeutic differences within AML subgroups and particularly between AML and ALL. Immunophenotyping has largely superseded the use of cytochemical staining techniques emphasized in the FAB classification, and the WHO classification (see below) has superseded use of the FAB because it incorporates cytogenetic and other classification features that guide application of contemporary risk-adapted therapies for AML. Table 5.2 FAB classification of AML (From Bennett et al. 1985) FAB AML subtype Common cytogenetic alteration M0 Acute myeloblastic leukemia minimally differentiated M1 Acute myeloblastic leukemia without maturation M2 Acute myeloblastic leukemia with maturation t(8;21)(q22;q22), t(6;9) M3 Hypergranular promyelocytic leukemia Microgranular variant t(15;17) M4 Acute myelomonocytic leukemia inv(16)(p13q22), del(16q) M4eo Acute myelomonocytic leukemia with bone marrow eosinophilia inv(16), t(16;16) M5a M5b Acute monoblastic/acute monocytic leukemia Poorly differentiated (M5A) Differentiated (M5B) del (11q), t(9;11), t(11;19) M6a M6b Acute erythroid leukemia (erythroid/myeloid and pure erythroleukemia) M7 Acute megakaryoblastic leukemia t(1;22) M8 Acute basophilic leukemia Expression cell surface and cytoplasmic markers aid in the classification of acute myeloid leukemia by definition of precursor stage, as well as markers of granulocytic, monocytic, megakaryocytic, and erythroid pathway maturation. Lineage-specific AML cell surface antigens include CD33, CD13, CD14, CDw41 (platelet IIB/IIIA), CD15, CD11b, CD36, and glycophorin A (Dinndorf et al. 1986; Kuerbitz et al. 1992; Smith et al. 1992). B-lymphocytic antigens, CD10, and T-lymphocyte antigens are present on a proportion of AML, but have not clearly had prognostic significance. However, specific antigen expression can be associated with certain AML subgroups. The expression of glycoprotein IB, glycoprotein IIB/IIIA, and factor VIII antigen is associated with megakaryocytic leukemias (Creutzig et al. 1990). In addition, the expression of CD34 and CD15 with heterogeneous expression of CD13, along with rare expression of HLA-DR, is highly characteristic of the presence of PML-RAR alpha gene rearrangements (Orfao et al. 1999). Though there is no uniform consensus regarding interpretation of blast expression levels, generally when 20% or more of leukemic cells express a marker, it is considered to be positive, though select markers, including cytoplasmic CD3, myeloperoxidase, TDT, CD34, and CD117, may be considered positive at lower levels. European LeukemiaNet guidelines outline criteria for determining lineage involvement for de novo AML (Table 5.3), and also for the approach to diagnostic characterization of mixed phenotype acute leukemias (MPAL). Notably, sophisticated multicolor flow cytometry analyses are able to detect heterogeneity in blast cell populations that may reflect prognostic significance. Table 5.3 European LeukemiaNet Guidelines for determining lineage involvement in leukemia (From Döhner et al. 2010) Expression of markers for diagnoses Diagnosis of acute myeloid leukemia (AML)a Precursor stage CD34, CD38, CD117, CD133, HLA-DR Granulocytic markers CD13, CD15, CD16, CD33, CD65, cytoplasmic myeloperoxidase (cMPO) Monocytic markers Nonspecific esterase (NSE), CD11c, CD14, CD64, lysozyme, CD4, CD11b, CD36, NG2 homologue Megakaryocytic markers CD41 (glycoprotein IIb/IIIa), CD61 (glycoprotein IIIa), CD42 (glycoprotein 1b) Erythroid marker CD235a (glycophorin A) Diagnosis of mixed phenotype acute leukemia (MPAL) Myeloid lineage MPO or evidence of monocytic differentiation (at least two of the following: NSE, CD11c, CD14, CD64, lysozyme) B-lineage CD19 (strong) with at least one of the following: CD79a, cCD22, CD10, or CD19 (weak) with at least two of the following: CD79a, cCD22, CD10 aFor the diagnosis of AML, the table provides a list of selected markers rather than a mandatory marker panel Requirements for assigning more than one lineage to a single blast population adopted from the WHO classification.3 Note that the requirement for assigning myeloid lineage in MPAL is more stringent than for establishing a diagnosis of AML. Note also that MPAL can be diagnosed if there are separate populations of lymphoid and myeloid blasts Most cases with 11q23 abnormalities express the NG2 homologue (encoded by CSPG4) reacting with the monoclonal antibody 7.1 Conventional cytogenetics is a requisite investigation for the contemporary classification and therapeutic planning for AML. In parallel, fluorescence in situ hybridization (FISH) is a powerful tool to complement cytogenetics in the detection of key gene rearrangements. The identification of recurrent balanced translocations and inversions and other recurrent genetic abnormalities is the core element for the current WHO Classification system for acute myeloid leukemia. In general, a minimum of 20 metaphase cells is required to exclude abnormal karyotypes. Notably, cytogenetic analyses from peripheral blood specimens may also be helpful when the marrow specimen is inadequate or unable to be obtained. Alternatively, if the patient’s clinical condition permits, a repeat bone marrow aspirate for cytogenetic and molecular analyses should be obtained if the initial sample was inadequate. Molecular genetic analyses of blasts for key leukemogenic mutations are a critical aspect of contemporary AML diagnosis (see below). Screening for somatically acquired mutations in the FLT3, NPM1, and CEBPA genes has potential prognostic and therapeutic implications, and is strongly encouraged by the European LeukemiaNet expert panel in the initial evaluation of patients with AML. In this regard, the recently updated WHO Classification includes AML with mutations in NPM1 or CEBPA as new provisional entities. Current investigations are analyzing the role of somatic mutations in the MLL, NRAS, WT1, KIT, RUNX1, TET2, TP53 and IDH1 genes, and gene expression alteration in ERG, MN1, EVI1, and BAALC. The presence of recurring cytogenetic abnormalities is the most prognostically significant finding in acute myeloid leukemias. The World Health Organization (WHO) Classification of the Myeloid Neoplasms incorporated diagnostic cytogenetic information into an AML classification system in 2002 (Vardiman et al. 2002). Two key differences between the WHO and previous FAB classification included: (1) a lower blast threshold (from 30% to 20% blasts in the blood or marrow) for the diagnosis of AML, and (2) the primary use of recurring clonal cytogenetic abnormalities as a basis for categorizing AML cases with unique clinical and biological subgroups. Notably, the presence of key clonal recurring cytogenetic abnormalities including t(8;21)(q22;q22), inv(16)(p13;q22), or t(16;16)(p13;q22), and t(15;17)(q22;q12) were considered to be AML regardless of blast percentage. In 2008, the WHO classification of myeloid neoplasms in acute leukemia was revised in the light of new genetic and biological information (Table 5.4) (Vardiman et al. 2009). This diagnostic classification system included new provisional entities, including AML with mutated NPM1 and AML with mutated CEBPA. The 2008 revision of the WHO also included a new category incorporating transient abnormal myelopoiesis as well as MDS and AML that is Down syndrome-related. MDS and AML related to Down syndrome are biologically identical and thus are considered together as “Myeloid leukemia associated with Down syndrome” (see Table 5.4). Incorporation of the WHO revised classification into contemporary treatment regimens permits sophisticated correlation and identification of activity of conventional chemotherapy and novel targeted agents in defined biological subgroups. Table 5.4 The 2008 revision of the World Health Organization (WHO) classification of myeloid neoplasms. (From Vardiman et al. 2009) Acute myeloid leukemia and related neoplasms Acute myeloid leukemia with recurrent genetic abnormalities AML with t(8;21)(q22;q22); RUNX1-RUNX1T1 AML with inv(16)(p13.1q22) or t(16;16)(p13.1;q22); CBFB-MYH11 APL with t(15;17)(q22;q12); PML-RARA AML with t(9;11)(p22;q23); MLLT3-MLL AML with t(6;9)(p23;q34); DEK-NUP214 AML with inv(3)(q21q26.2) or t(3;3)(q21;q26.2);RPN1-EVI1 AML (megakaryoblastic) with t(1;22)(p13;q13); RBM15-MKL1 Provisional entity: AML with mutated NPM1 Provisional entity: AML with mutated CEBPA Acute myeloid leukemia with myelodysplasia-related changes Therapy-related myeloid neoplasms Acute myeloid leukemia, not otherwise specified AML with minimal differentiation AML without maturation AML with maturation Acute myelomonocytic leukemia Acute monoblastic/monocytic leukemia Acute erythroid leukemia Pure erythroid leukemia Erythroleukemia, erythroid/myeloid 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 Acute leukemias of ambiguous lineage Acute undifferentiated leukemia Mixed phenotype acute leukemia with t(9;22)(q34;q11.2); BCR-ABL1 Mixed phenotype acute leukemia with t(v;11q23); MLL rearranged Mixed phenotype acute leukemia, B-myeloid, NOS Mixed phenotype acute leukemia, T-myeloid, NOS Provisional entity: natural killer (NK) cell lymphoblastic leukemia/lymphoma The main relevance of prognostic factors in AML is to predict the likelihood of achieving remission and/or the risk of relapse, and thereby to identify patients for whom conventional combination chemotherapy may be inadequate and HSCT in first complete remission (CR1) is indicated. It is suggested that allogeneic HSCT may only benefit patients with a relapse risk of greater than 35% (Cornelissen et al. 2007). While this figure must be dependent on the anticipated procedure-related mortality, patients at low risk of relapse are unlikely to benefit. Predictors of relapse can include pre- or post-treatment characteristics, but will always be protocol dependent. The major pre-treatment characteristics are cytogenetics, age, white blood cell (WBC) count, secondary leukemia and FLT3/ITD, NPM1 and CEBPA mutations. The major post-treatment prognostic predictor is the speed and depth of response to treatment, which may be assessed morphologically, immunologically or molecularly. While cytogenetics and response to treatment are the main prognostic factors that drive risk stratification in current treatment protocols, risk group stratification is becoming increasingly complicated. There are many potential future predictors of outcome, but most have not undergone sufficient assessment or validation in children, or indeed adults, to be reliably used to stratify treatment at present. Amongst these are WT1, MLL–PTD, RUNX1, TET2, IDH1 and TP53 mutations; BAALC, ERG, EVI1 and MN1 gene expression levels, microRNA and gene expression profiles and resistance proteins. While new markers identify targets for novel therapies, combinations of markers may interact to further refine risk group stratification. Age and WBC count at presentation are recognized as prognostically important but are not routinely used to risk-stratify treatment in children. Increasing age in both children and adults is independently associated with a worse outcome, with children greater than 10 years doing less well than younger children (Razzouk et al. 2006). Within the pediatric age range, a higher white cell count at presentation is associated with a lower CR rate, higher relapse risk, and worse overall survival. As outlined above, cytogenetics and use of the WHO classification have largely replaced FAB classification. Notably, the previously reported inferior outcome for acute megakaryoblastic leukemia (AMKL) (M7) in non-Down syndrome (non-DS) AML (Athale et al. 2001; Dastugue et al. 2002; Barnard et al. 2007) and the benefit for HSCT (Athale et al. 2001; Garderet et al. 2005) have recently been challenged (Reinhardt et al. 2005). The Berlin-Frankfurt-Münster (BFM) study group reported an improved outcome for non-DS patients with megakaryoblastic leukemia, treated on recent more intensive trials and found no benefit for HSCT (Reinhardt et al. 2005). Non-DS M7 AML may be a heterogenous disease, and a better outcome has recently been reported for non-DS M7 AML in association with t(1;22) compared to non-DS M7 AML without this karyotypic abnormality (Dastugue et al. 2002). Both acute promyelocytic leukemia (APL) and DS AML are recognized as distinct subtypes of AML in the 2008 WHO classification, with unique features and favorable outcomes when treated on appropriate protocols that recognize their individual chemotherapy sensitivities, including anthracyclines and differentiating agents in APL and high-dose cytarabine in DS AML. Each of the entities is discussed in detail below. The leukemia blast karyotype is a key determinant of outcome in AML (Grimwade et al. 1998; Raimondi et al. 1999). However, the usefulness of the diagnostic karyotype is limited by conflicting data on the prognostic significance of some primary cytogenetic aberrations, including t(8,21), various translocations involving the MLL locus on 11q23, and the definition of a complex karyotype. The main limiting factor is that approximately 25% of children (Harrison et al. 2010, Raimondi et al. 1999) (and approximately 40% of adults) with AML have a normal karyotype (cytogenetically normal: CN AML), although molecularly heterogenous disease. Based on multivariable analysis of results from the Medical Research Council (MRC) AML 10 trial, a cytogenetically based risk group classification (Grimwade et al. 1998; Wheatley et al. 1999) has been validated and adopted by a number of cooperative groups. Differences in outcome among these groups are probably explained by small numbers or by differences in the treatment delivered. This cytogenetic risk score classifies good-risk patients as those with t(15,17), t(8,21), inv(16)/t(16,16); poor risk as those with those with −5, del(5q), monosomy 7 (−7), abnormalities of 3q and a complex karyotype (5 or >); and all others as intermediate-risk (Grimwade et al. 1998; Wheatley et al. 1999). The successor trial MRC AML 12 adopted this cytogenetic risk score and reported children (455 with a cytogenetic result) aged less than 15 years old and with favorable (excluding APL), normal, and other intermediate and adverse karyotypes to have a 10-year overall survival (OS) of 64%, 60% and 43%, respectively. Based on 5,876 adult cases (age 16–59 years), the original MRC cytogenetic classification has been further refined to include rarer abnormalities not previously considered separately. The only karyotypes associated with a favorable outcome remain t(15,17), t(8,21), and inv(16)/t(16,16), but the adverse cytogenetic risk group has been extended to include abn(3q) (excluding t(3;5)), inv(3)/t(3,3), −5/del(5q)/add(5q), −7/add(7q), t(6,11), t(10,11), t(9,22), −17/abn(17p) with other changes and complex (>3 unrelated abnormalities) (Grimwade et al. 2009 a). Analysis of 729 children treated on MRC AML 10 and 12 reported the interesting new findings that 12p abnormalities and t(6;9) (p23;q34) predict a poor outcome. Additionally, abnormalities of 3q and complex karyotypes do not appear to have the same degree of adverse outcome as reported for adults (Harrison et al. 2010). Recently an adverse prognosis has been reported in an adult study for autosomal monosomy in conjunction with at least one other autosomal monosomy or structural abnormality (Breems et al. 2008). The core binding factor leukemias (CBF), t(8;21) and inv(16), are heterogeneous (Marcucci et al. 2005; Appelbaum et al. 2006) and adult studies suggest that the OS may be significantly inferior for patients with t(8;21) compared to those with inv(16)/t(16,16), predominantly because of a lower salvage rate after relapse (Marcucci et al. 2005). Outcome is treatment dependent and patients with CBF leukemias may do particularly well when they receive high-dose cytarabine (Bloomfield et al. 1998; Byrd et al. 1999; Appelbaum et al. 2006). Gemtuzumab ozogamicin (Mylotarg), a recombinant humanized anti-CD33 monoclonal antibody linked to the potent anti-tumor intercalating agent, calicheamicin, has recently been reported to preferentially benefit CBF leukemias (Burnett et al. 2006). Rearrangements of the MLL gene are the most common abnormalities in children with AML, although the frequency varies among studies (10–20%) (Rubnitz et al. 2002; Harrison 2010). A recent retrospective international study of 756 children with 11q23-rearranged AML demonstrated significant heterogeneity within this abnormality (Balgobind et al. 2009). This study failed to confirm the previously reported favorable prognosis for t(9;11)(p22;q23) (Rubnitz et al. 2002), but multivariate analysis identified t(1;11)(q21;q23) to be an independent predictor of a favorable outcome while t(10;11)(p12;q23), t(10;11)(p11.2;q23) and t(6;11)(q27;q23) independently predicted an inferior outcome (Balgobind et al. 2009). Patients in this study were treated on different chemotherapy protocols, which may in part explain differences in outcome between groups across MLL-rearranged subtypes. An international collaborative study (Hasle et al. 2007) of 258 children with AML and −7 (n = 172) or del(7q) (n = 86), both with or without other cytogenetic changes, demonstrated heterogeneity within chromosome 7 abnormalities. Monosomy 7 was associated with a lower CR rate (61% vs 89%, p < 0.001) and inferior 5-year survival (30% vs 51%, p < 0.01) compared to patients with del(7q). Additional cytogenetic abnormalities impacted outcome. Favorable risk cytogenetics were strongly associated with del(7q) and retained their favorable 5-year survival compared to del(7q) without additional favorable cytogenetics (75% vs 46%, p = 0.03). By contrast, patients with −7 and adverse cytogenetics (inv(3), −5/del(5q), or +21) had a dismal outcome (5-year OS 5%), which was not improved by HSCT. Patients with AML and a cytogenetically normal karyotype (CN-AML) are classified within the intermediate-risk group, despite having molecularly heterogenous AML. CN–AML and those within cytogenetically defined subgroups may be further risk group stratified by prognostically significant molecular characteristics. Class I mutations (RAS, FLT 3, c-KIT) activate signaling pathways, increasing proliferation and inhibiting apoptosis while class II mutations (CEBPA, MLL-PTD and probably NPM1) affect transcriptional processes leading to impaired differentiation. In addition to these mutations regulating signaling and transcription pathways, the expression levels of a number of genes (BAALC, ERG, EVI1) may have independent prognostic significance in AML. Many patients with AML harbor more than one mutation, and these mutations interact with each other and do not influence outcome in isolation. The complexity of these interactions is illustrated in Fig. 5.1. Open image in new window Fig. 5.1 Pie chart illustrating the molecular heterogeneity of cytogenetically normal AML based on mutations in the NPM1, CEBPA, MLL, FLT3 (internal tandem duplication [ITD] and tyrosine kinase domain [TKD] mutations at codons D835 and I836], NRAS, and WT1 genes. The bluish colors denote NPM1 mutated subsets, the orange/red colors CEBPA mutated subsets, and the yellow/green colors MLL mutated subsets; the grey colors depict subsets without hypothetical class II mutations, and the white sector shows the subset without any mutation in the above mentioned genes. Data are derived from mutational analysis of 485 younger adult patients with cytogenetically normal AML from AMLSG (From Dohner et al. 2010. On behalf of the European LeukemiaNet. Blood 2009; Oct 30. Epub) FLT3 is a tyrosine kinase receptor and therefore of both prognostic and clinical importance because of the potential for targeted therapy with FLT 3 inhibitors. Normal FLT3 signaling is involved in regulation of hematopoietic stem cell development. Internal mutations in the FLT3 receptor have been identified in a large subset of AML. The FLT3 receptor is a tyrosine kinase expressed on the surface of hematopoietic stem cells. Two predominant types of mutations are found in FLT3 and AML, including internal tandem duplications (ITD) mapping to the juxta-membrane portion of the receptor, and activating mutations in the kinase domain. Mutations in either the ITD or kinase domain both constitutively activate FLT3 signaling. There is an age-associated increase in incidence of AML with FLT3/ITD mutations, accounting for 12–17% of AML in children and adolescents, and rising to approximately 34% of adults (Meshinchi et al. 2001; Zwaan et al. 2003a; Meshinchi et al. 2006). Approximately 10% of patients exhibit leukemic blasts with point mutations primarily in the kinase domain at aspartic acid 835 or isoleucine 836. Tandem duplications of FLT3 (FLT3/ITD) are associated with an increased relapse risk and inferior survival in both children and adults (Abu-Duhier et al. 2000; Kottaridis et al. 2001; Meshinchi et al. 2001; Zwaan et al. 2003a; Meshinchi et al. 2006) and in two pediatric studies were the strongest predictor of relapse in multivariate analysis (Meshinchi et al. 2001; Zwaan et al. 2003a). The prognostic impact of FLT3/ITD for both adults and children is reported to be dependent on the mutant/wild type allele ratio (Whitman et al. 2001; Thiede et al. 2002; Zwaan et al. 2003; Gale et al. 2008) with high FLT3/ITD allelic ratios, indicative of homozygous mutations generated by acquired uniparental disomy, being associated with a particularly poor outcome. The impact of FLT 3 mutations in the tyrosine kinase domain (FLT3–TKD) are variably reported (Mead et al. 2007). Current clinical trials are examining the potential activity of small molecule FLT3 receptor tyrosine kinase inhibitors in conjunction with chemotherapy, and it is yet unclear if these regimens can obviate the role for allogeneic HSCT in these patients. The nucleophosmin member 1 (NPM1) gene product functions as a likely chaperone molecule between the nucleus and cytoplasm. NPM1 functions in ribosome biogenesis, maintenance and duplication of centrosomes, and regulation of genomic stability proteins including p53 and ARF. Mutation of the NPM1 gene is the most common mutation in adult AML (up to 35%) and is present in about 50% of adult CN-AML (Falini et al. 2005; Thiede et al. 2006), but only occurs in about 8% of children (Cazzaniga et al. 2005; Brown et al. 2007) and then mainly in older children. NPM1 mutation alters the distribution of the protein and causes predominantly a cytoplasmic localization. NPM1 mutations interact prognostically with FLT3 ITD and in the absence of FLT3/ITD predict a better response to Induction chemotherapy and a favorable overall survival (Dohner et al. 2005; Falini et al. 2005; Thiede et al. 2006; Brown et al. 2007; Mrozek et al. 2007; Schlenk et al. 2008), but this advantage is lost in the presence of FLT3/ITD mutation. NPM1 mutations commonly co-exist with FLT3/ITD in CN-AML. Recently it has been suggested that patients with wild type FLT3/ITD, and NPM1 mutations have a favorable disease, should be classified as good-risk, and should not be candidates for HSCT in CR1 (Schlenk et al. 2008). CEBPA mutations are found in about 5–15% of adult AML (Schlenk et al. 2008; Wouters et al. 2009) and generally in association with CN-AML. CEBPA mutations carry a favorable outcome when biallelic mutations are present (Wouters et al. 2009). There is growing evidence that NPM1 and CEPBA mutations are class II mutations and primary genetic lesions which impair hematopoietic differentiation. c-kit mutations are detected in 3–11% of pediatric AML cases (Meshinchi et al. 2003; Goemans et al. 2005) and are strongly associated with CBF leukemias. c-kit codon 816 mutations negatively impact adult patients with t(8;21) compared to those with wild type c-kit (Boissel et al. 2006; Paschka et al. 2006; Schnittger et al. 2006) and may explain the heterogeneity in outcome reported for this cytogenetic group. c-kit mutations are likely to have the same negative impact in patients with inv(16), although the evidence is less clear (Boissel et al. 2006; Paschka et al. 2006). The prognostic significance of c-kit mutations in children with CBF leukemia is limited by study size. A BFM study found a c-kit mutation in 31% of 16 patients with t(8,21) AML and reported no difference in outcome between patients with and without the mutation (Goemans et al. 2005), while the Japanese Childhood AML Cooperative Study Group reported an inferior outcome for 8 of 46 (17%) patients with t(8;21) and c-kit mutation (Shimada et al. 2006). c-kit is a tyrosine kinase receptor and offers the potential for treatment with a tyrosine kinase inhibitor. Partial tandem duplication of the mixed lineage leukemia MLL (MLL-PTD) gene occurs in 5–10% of adult CN-AML and is reported to be associated with an inferior disease free survival (DFS) (Schlenk et al. 2008). MLL-PTD mutations usually coexist with FLT3/ITD (30–40%) (Schlenk et al. 2008) and the impact of co-existence on outcome is not clear. MLL-PTD AML may respond to demethylating agents. High level BAALC (brain and acute leukemia cytoplasmic protein), EVI1 and ERG expression independently negatively impact on outcome in CN AML. BAALC interacts with FLT3/ITD, and the presence of both abnormalities confers a particularly poor prognosis (Baldus et al. 2006). RAS mutations are more common in CBF inv(16) leukemias, but their prognostic impact is unclear. They may be targets for farnesyl transferase inhibitors. WT1 gene mutations have been reported to confer an adverse prognosis (Paschka et al. 2008), although not consistently. Molecular genetic aberrations and their clinical implications have recently been the subject of extensive review (Schlenk and Dohner 2009; Scholl et al. 2009). The inter-relationship between mutations/cytogenetics/clinical features and outcome is complex. The LeukemiaNet international expert panel has recently recommended a new standardized reporting for the correlation of cytogenetics and molecular genetic data with clinical data in AML (Table 5.5) (Dohner et al. 2010). This re-defines favorable risk disease as CBF leukemias, CN-AML NPM1 mutated/FLT3 negative and CN-AML CEPBA mutated. These patients are at low risk of relapse and unlikely to benefit from HSCT. CN-AML continues to be classified as intermediate-risk but is subclassified by NPM1 and FLT3/ITD status. A risk classification schema for the management of adult AML patients with normal karyotypes that incorporates the status of FLT3, NPM1, BAALC, MLL-PTD and CEBPA has been proposed (Mrozek et al. 2007). Others propose the incorporation of additional poor prognostic factors including WT1, ERG and EVI-1 expression (Scholl et al. 2009). However, many of these molecular markers have not been extensively studied in children. Before new prognostic markers can be used to direct treatment in children with AML, their prognostic/predictive value must be validated in children and their benefit tested within the setting of a clinical trial. Table 5.5 European LeukemiaNet Guidelines for standardized reporting for correlation of cytogenetic and molecular data in AML with clinical data. Dohner et al, 2010. Standardized reporting for correlation of cytogenetic and molecular genetic data in AML with clinical dataa (From Dohner et al. 2010. On behalf of the European LeukemiaNet. Blood 2009; Oct 30. Epub) Genetic group Subsets Favorable t(8;21)(q22;q22); RUNX1-RUNX1T1 inv(16)(p13.1q22) or t(16;16)(p13.1; q22); CBFB-MYH11 Mutated NPM1 without FLT3-ITD (normal karyotype) Mutated CEBPA (normal karyotype) Intermediate-I Mutated NPM1 and FLT3-ITD (normal karyotype) Wild type NPM1 and FLT3-ITD (normal karyotype) Wild type NPM1 without FLT3-ITD (normal karyotype) Intermediate-II t(9;11)(p22;q23); MLLT3-MLL Cytogenetic abnormalities notclassified as favorable or adverse Adverse inv(3)(q21q26.2) or t(3;3)(q21;q26.2); RPN1-EVI1 t(6;9)(p23;q34); DEK-NUP214 t(v;11)(v;q23); MLL rearranged −5 or del(5q); −7; abnl(17p); complex karyotypeb aFrequencies, response rates and outcome measures should be reported by genetic group, and, if sufficient numbers are available, by specific subsets indicated; excluding cases of acute promyelocytic leukemia bThree or more chromosome abnormalities in the absence of one of the WHO designated recurring translocations or inversions, i.e., t(15;17), t(8;21), inv(16) or t(16;16), t(9;11), t(v;11)(v;q23), t(6;9), inv(3)/t(3;3); indicate how many complex karyotype cases have involvement of chromosome arms 5q, 7q, and 17p Current risk group stratification based on cytogenetic and molecular characterization fails to identify all patients at risk of relapse within cytogenetic/molecularly characterized risk groups. Early response to therapy is a powerful predictor of outcome, but morphological assessment is relatively crude. The EFS at 5 years for children treated on BFM 83 and 87 was inferior for those with more than 5% blasts on day 15 compared to those with less than 5% blasts (56% vs 27%, p = 0.0001 and 61% vs 40%, p = 0.001) (Creutzig et al. 1999). In the MRC trial, patients are assigned to one of three risk groups based on diagnostic cytogenetics and morphological response after course 1 of chemotherapy. Based on 467 patients treated on AML 12, patients in CR (<5% blasts), partial response (PR) (5–15% blasts) and residual disease (RD) (>15% blasts) after course 1 had 10 year survival rates of 76%, 55% and 25% and relapse rates of 31%, 42%, and N/A (because patients with refractory disease were unevaluable for relapse), respectively. When combined with cytogenetics, the OS at 10 years for good, standard and poor-risk patients was 83%, 70% and 39%, respectively with relapse risks of 21%, 36% and 53% respectively. While discriminatory, the MRC risk score unfortunately only identifies about 20% of the patients destined to relapse, and about 60% of relapses come from standard-risk and 20% from good-risk groups. Minimal residual disease (MRD) assays provide sensitive measurements for detecting low levels of leukemic cells that cannot be detected morphologically (Campana 2003). MRD can be assessed molecularly by real-time quantitative polymerase chain reaction (RT PCR) of leukemic specific targets (gene fusions, e.g., AML-ETO; gene mutations, e.g., NPM1; and overexpressed genes, e.g., WT1) or immunologically by multiparameter flow cytometry identifying an aberrant leukemia-associated immunophenotype (LAIP). The advantage of the latter is that an immunophenotypic AML “signature” can be identified in more than 90% of AML cases, while the molecular heterogeneity of AML is quite extensive and not easily amenable to the development of specific RT PCR assays in most cases. The sensitivity of flow cytometry is at least a log less than RT PCR assays at 103–104 compared to RT PCR 103–106, but the sensitivity of flow cytometry may improve with six to eight color laser technology. The usefulness of MRD measurement includes early assessment of response to therapy to improve risk stratification and to guide Consolidation therapy, and subsequent monitoring to detect impending relapse and guide pre-emptive therapy. The kinetics of AML1-ETO and CBFb-MYH11 fusion transcripts monitored by RT PCR, including the transcript level at diagnosis, extent of reduction after Induction chemotherapy and any increase in transcript level after attainment of CR, are thought to be predictive of outcome. WT1 transcripts and NPM1 may be useful for monitoring levels of disease and predicting outcome. There is a paucity of prospective studies of the usefulness of MRD measurement. Pediatric studies on the predictive value of minimal residual disease have reported conflicting results. Sixteen percent of 252 children who achieved morphological remission on the Children’s Cancer Group (CCG)-2961 trial had detectable MRD (defined as greater than 0.5% blasts with an aberrant phenotype) after Induction therapy, and on multivariate analysis this was shown to be the most powerful independent prognostic factor associated with poor outcome (Sievers et al. 2003). A study from St. Jude Children’s Research Hospital mirrored these findings and reported a 2 year OS of 33% for patients with detectable MRD at the end of Induction therapy compared to 72% for MRD-negative patients (p = 0.022) (Coustan-Smith et al. 2003). However, multivariate analysis of 150 children treated on AML BFM 98 study failed to show additional predictive value for MRD measured by flow cytometry compared to other routinely employed risk factors (Langebrake et al. 2006). Current approaches to risk group stratification are unrefined and confined to cytogenetics, response to treatment (assessed by different criteria and methods), and limited molecular characterization for some groups. While able to discriminate risk groups, these parameters fail to identify the majority of patients who will relapse. AML is a very heterogenous disease both within CN-AML and within cytogenetically-defined risk groups, and molecular characterization may further refine these risk groups and identify targets for therapy. Systematic evaluation of MRD assessment in conjunction with refined molecular/cytogenetic risk groups will further inform post-remission therapy. While it is important to establish a hierarchy of prognostic factors, their prognostic and predictive value needs to be validated in children with AML and not assumed from adult data, and must be tested within the context of a specific protocol or treatment. At present, refinement of risk stratification helps define the need for HSCT, but increasingly molecular characterization will offer the potential of novel targeted therapy. New prognostic/predictive indicators can only be adopted into clinical practice when properly validated for childhood as well as adult AML. Clinical trials for the therapy of pediatric acute myeloid leukemia have exhibited steady improvement in initial remission induction rates, disease-free survival, event-free survival, and overall survival over the past 5 decades. The key advances in treatment have been as a result of identification and optimization of the dosing and schedule of active drugs (predominantly cytarabine and anthracyclines, as outlined above), the identification of key subgroups of patients benefiting from specialized treatment approaches, and improved supportive care for infections and other complications. Contemporary treatment approaches incorporate four to six courses of intensive chemotherapy, and most groups do not use lower dose “Maintenance” therapy approaches. The use of myeloablative high-dose chemotherapy Consolidation with allogeneic HSCT in CR1 is controversial and under investigation. It is generally not employed in patients with favorable-risk and standard-risk cytogenetics, but it is used in some groups for patients with intermediate-risk disease and a matched related donor. Alternative donor HSCT approaches are generally reserved for patients with high risk cytogenetic or molecular features, or with resistant or residual disease. Refinements in the treatment of pediatric AML have largely resulted from the development of biologic and risk-stratified approaches to therapy. Early cancer clinical trials in the 1950s and 1960s for children with leukemia included patients with ALL and AML on the same regimen, and identified early major differences in the response of these two diseases to different chemotherapy drugs. Subsequent clinical trials treated patients with AML on separate regimens and identified unique distinguishing biology and response in patients with APL, Down syndrome leukemia, and various cytogenetic subgroups of AML. These observations were paralleled by the development of active risk-adapted regimens for different subgroups, and have in turn provided a foundation for improvements in the overall cure rate of pediatric AML as the overall survival was increased in each subgroup. Intensive anthracycline- and cytosine-based chemotherapy combined with advances in supportive care have resulted in a dramatic improvement in the survival of children with AML treated on United Kingdom (UK) MRC trials. For the past 2 decades, children (n = 1,053) and adults have been treated on the same or very similar protocols within three trials: MRC AML 10 (1988–1995), MRC AML 12 (1995–2002) and MRC AML 15 (2002–2008). Primary areas of investigation in MRC studies have been the comparison of anthracyclines/anthracenediones and testing the benefit of the addition of active agents to the intensive cytarabine and anthracycline core regimen. Successive MRC trials have failed to show any significant superiority in CR rate or OS for any of the four tested intensive Induction regimens. MRC AML 10 tested the comparative benefits of thioguanine with etoposide by comparing DAT (ara-C, daunorubicin, thioguanine) with ADE (ara-C, daunorubicin, etoposide). There was no significant difference in CR rate (DAT 90% vs ADE 93%, p = 0.3, odds ratio (OR) = 1.58 (0.68,3.51)), EFS (DAT 48% vs ADE 45%, p = 0.5, hazard ratio(HR) = 0.89 (0.65,1.23)) or OS (DAT 57% vs ADE 51%, p = 0.3, HR = 0.83 (0.59,1.17)), but there was a non-significant excess of deaths in CR in the ADE arm (8% vs 15% p = 0.09) (Stevens et al. 1998). There was no evidence that children with monocytic involvement (FAB type M4 or M5) in particular benefited from an etoposide-containing regimen (p = 0.9). ADE was brought forward into MRC AML 12 in the belief that the observed lack of advantage for ADE was in part due to the excess of deaths in CR, which could be negated by improved supportive care. MRC AML 12 found mitoxantrone (MAE) and daunorubicin (ADE) to be of equal efficacy in Induction when combined with similar doses of cytarabine and etoposide. No significant difference was noted in CR rates (ADE 92% vs MAE 90%, p = 0.3, OR = 1.30(0.70, 2.4)) or deaths in CR or EFS. Induction deaths were not significantly higher for MAE (MAE 6% vs ADE 3%, p = 0.2) (Gibson et al. 2005). There was a possibility that MAE was superior for both DFS (p = 0.03, HR = 0.72 (0.54, 0.96)) and relapse risk (RR) (p = 0.05, HR = 0.73 (0.54, 1.00)), but OS was not significantly different (ADE 61% vs MAE 65%, p = 0.2, HR = 0.84 (0.63, 1.12)). AML 15 compared ADE with FLAG-Ida (fludarabine, cytarabine, granulocyte colony stimulating factor (G-CSF), and idarubicin) with or without the addition of gemtuzumab ozogamicin at 3 mg/m2 in Induction. The data are unpublished and premature. However, in adult patients there appears to be no difference in the CR rate; an apparent reduction in the relapse risk in the FLAG–Ida arm fails to translate into an improvement in OS because of an increase in deaths in CR for patients receiving FLAG-Ida (Burnett et al. 2009). Gemtuzumab ozogamicin at 3 mg/m2 combined with Induction chemotherapy appears to benefit predominantly favorable prognosis CBF leukemias. In summary, no intensive regimen has resulted in a superior CR rate or OS, and any advantage of one regimen over another in terms of a reduction in relapse rate has been balanced by an increase in deaths in Induction or CR. MRC AML 12 failed to show an advantage for an additional course of post-remission chemotherapy (four versus five courses of treatment in total) in any risk group of patients already receiving intensive treatment, irrespective of whether HSCT was included, although confidence intervals were wide: deaths in CR, four courses 2% versus five courses 1% (p = 0.6, HR = 0.64 (0.11, 3.7)); relapse risk, four courses 37% versus five courses 37% (p = 1.0, HR = 1.01 (0.67, 1.50)); OS, four courses 74% versus five courses 74% (p = 1.0, HR = 1.01 (0.63,1.62)). This suggests that the benefit from conventional post-remission chemotherapy may reach a plateau and that further improvement may only be achieved by alternative or targeted therapies. MRC AML trials have traditionally delivered high cumulative doses of anthracycline and anthracenedione therapy (AML 10: 550 mg/m2; AML 12: 300–610 mg/m2 calculated on a conversion factor, with mitoxantrone 1 mg/m2 being equivalent to daunorubicin 5 mg/m2). AML 15 tested whether the cumulative dose could be safely reduced and high-dose (HD) Ara-C substituted without loss of anti-leukemic efficacy by comparing two blocks of anthracycline-driven MRC Consolidation with two blocks of HD Ara-C (550 vs 300 mg/m2). Follow-up is too short to answer this important question, but preliminary data suggest that there may be no detriment to substituting anthracyclines with HD Ara-C in Consolidation (Burnett et al. 2009). Gemtuzumab ozogamicin 3 mg/m2 has also been tested in combination with chemotherapy in Consolidation. Preliminary results suggest no benefit. The combined data from MRC AML 10 and AML 12 suggest no survival benefit for allogeneic HSCT in first CR for any risk group in children. While allogeneic HSCT was associated with a significant reduction in RR for good and standard risk patients (p = 0.02), but not for poor risk patients, this did not translate into a survival advantage for any risk group (p = 0.3) because the reduced relapse risk was counterbalanced by an increase in procedure-related deaths (p = 0.001). It may be argued that these data date from 1998 to 2002 and that more recent improvements in transplant-related mortality may allow an advantage for HSCT to emerge. However, the MRC data failed to show a reduction in relapse risk for poor-risk patients, without which a survival advantage cannot be achieved. In MRC AML 10, autologous HSCT in CR 1 reduced relapse (p = 0.03) and improved DFS (p = 0.02) but did not significantly improve long-term survival (p = 0.2), not because of a counter-balancing effect on procedural mortality, which was low (3%), but because of an inferior survival after relapse (autologous HSCT 7% vs Stop 27%, p = 0.05). MRC studies thus suggest that the role of allogeneic HSCT is limited in first CR and that the ceiling of benefit for chemotherapy may have been reached. Successor studies aim to improve outcome by molecular characterization and the use of targeted therapies in appropriate risk groups or biological subtypes. This will include exploring dosing and scheduling in children. The measurement of MRD by flow cytometry and molecular markers where possible may improve risk stratification and offer the potential for treatment intensification, which may include targeted therapies or stem cell transplantation. Alternatively, MRD may suggest that allogeneic HSCT can appropriately be restricted, reducing the risk of associated procedure-related mortality and morbidity. Although international collaboration is being expanded, there are still several groups in the mainland of Europe (exclusive of the UK) that conduct independent clinical trials for newly diagnosed pediatric AML. From North to South these groups include the NOPHO (Nordic Society for Pediatric Hematology/Oncology), the BFM group, LAME (the French Leucémies Aiguës Myéloblastiques de l’Enfant group), and AIEOP (the Italian Associazione Italiana Ematologia Oncologia Pediatrica). Worldwide, other than the Children’s Oncology Group (COG) and the St. Jude consortium, Japan also has an active clinical trial for newly diagnosed AML. In other countries, investigators either participate in COG or MRC protocols, or have adapted protocols from the BFM-AML group. The Netherlands and Belgium are currently establishing a pilot study, with collaboration with the NOPHO/Hong Kong consortium in the near future. The main aspects of pediatric AML trials conducted by these groups will be discussed, focusing on Induction therapy, Consolidation therapy, the use of HSCT, and Maintenance therapy. In many cases, the hypotheses tested in the European and Japanese studies are linked to results from investigations by other pediatric and adult groups. The backbone of treatment to induce remission is two courses of intensive chemotherapy. The classic combination is 3 days of daunorubicin and 7 days of cytarabine. Usually, a third drug like etoposide or thioguanine is added to this combination. NOPHO applies even four drugs, by adding both etoposide and thioguanine to cytarabine and idarubicin. More than 80% of patients achieve CR with a 3- (or 4-) drug regimen and adequate supportive care (Yates et al. 1973; Weinstein et al. 1980; Yates et al. 1982; Pui 1995; Hann et al. 1997; Lowenberg et al. 1999; Pui et al. 2004; Zwaan and Kaspers 2004; Kaspers and Ravindranath 2005). Various strategies have been used to increase CR rates and subsequent survival: substitution of daunorubicin with idarubicin or mitoxantrone, increased dosages of cytarabine, and reduced intervals between the initial cycles of chemotherapy (intensive timing) (Ravindranath et al. 1996; Woods et al. 1996; Creutzig et al. 2001 b; O’Brien et al. 2002; Perel et al. 2002; Pui et al. 2004). The general conclusion that can be drawn from these studies is that a more intensive remission induction chemotherapy improves the quality of remission, which favorably affects the likelihood of long-term disease-free survival. In particular, a comparison of different schedules in relation to the rate of resistant disease and prognosis suggested that a cumulative anthracycline dose of less than 100 mg/m2 after one Induction course is associated with a worse outcome (Kaspers and Ravindranath 2005). However, it has to be kept in mind that death due to toxicity during Induction can offset the expected benefits of some of these strategies. In that perspective, it is interesting that rates for CR and overall survival are by and large similar for MRC, COG and NOPHO trials, when MRC does not advocate intensively timed Induction chemo therapy but COG and NOPHO protocols have incorporated this strategy. The addition of a third drug, thioguanine or etoposide, to intensive anthracycline and cytarabine Induction platforms has been tested by the BFM and the MRC groups, similar to approaches in the North American Pediatric Oncology Group (POG) and CCG groups. Yet, the contribution of a third drug in Induction chemotherapy is not clear (Creutzig et al. 1987; Steuber et al. 1990; Ravindranath et al. 1991; Woods et al. 1996; Stevens et al. 1998; Kaspers and Ravindranath 2005). A direct comparison of etoposide and thioguanine, added to cytarabine and daunorubicin, did not reveal statistically significant differences in remission rates, event-free or overall survival (Stevens et al. 1998). The MRC group has recently tested the efficacy and safety of gemtuzumab ozogamicin as a third drug in Induction therapy in adult AML (Kell et al. 2003; Zwaan and Kaspers 2004). The clinical efficacy of this drug in childhood relapsed and refractory AML does suggest its usefulness, and so do the preliminary results from the MRC trial, which, however, mainly concerned adults (Zwaan et al. 2003a, b). It remains to be proven if long-term outcome will improve with the implementation of this drug in up-front combination chemotherapy in children, but COG is studying that (Zwaan et al. 2003a, b). Currently, the best approach is not apparent, and neither CR rates nor EFS or OS rates differ significantly among study groups. A goal for current European studies it to enhance drug exposure to potentially result in high quality complete remissions with as low as possible amounts of MRD. Underlying this goal is a strategy to combine drugs with different mechanisms of action, as long as toxicity remains manageable. Supporting strategies to intensify high-dose cytarabine in pediatric AML Induction therapy are adult studies demonstrating superior post-remission outcome with this approach (Weick et al. 1996; Bishop et al. 1998). The recently complete MRC AML15 study may also provide information on this issue, including information on pediatric AML, since it did include a comparison of standard low-dose and high-dose cytarabine at Induction and enrolled children as well. Unfortunately, other differences in that Induction therapy may confound the results. Most groups now use multiple courses of intensive chemotherapy as Consolidation therapy. Higher-risk patients theoretically may benefit from an additional (fifth) course of chemotherapy, but such a risk-group-adapted therapy has not been proven to be better in pediatric AML. Including two courses for Induction, the minimum total number is four as applied by the MRC group, and results of that group are among the best in the world. Study MRC AML 12 did not show a statistically significant benefit for five total courses as compared to four, but that question continued to be studied in AML15, and pediatric results have not yet been reported. The NOPHO is studying the role of two doses of gemtuzumab ozogamicin as post-Consolidation therapy, which one might consider to be an extra course of chemotherapy. The Japanese group reported that they could safely reduce the number of Consolidation courses from eight to six to five for standard-risk patients and to six for higher-risk patients (Tomizawa et al. 2007; Tsukimoto et al. 2009). One may conclude that between four and six courses of chemotherapy in total are usually needed in pediatric AML (i.e., two to four courses of Consolidation therapy). Depending on the content of each course, the number of courses will also influence the cumulative dose of anthracyclines or related drugs. It seems fair to state that the optimal cumulative dose is unknown but that cumulative doses at Induction should not be less than 100 mg/m2. On the other hand, it is quite clear that cumulative doses of anthracyclines above 300 mg/m2 are associated with increasing risk of cardiotoxicity (Kremer et al. 2001). Most studies have included several courses of intermediate- or high-dose cytarabine in their post-remission chemotherapy strategy. The POG 8821 study, for example, used a six-dose schedule, and the CCG and Nordic trials used an eight-dose timed split-dose schedule (Capizzi and Powell 1987; Capizzi et al. 1988; Lie et al. 1996, 2003; Woods et al. 1996; Kaspers and Ravindranath 2005). Some studies combined cytarabine with non-cross-resistant agents. The MRC group, for instance, included the combination of amsacrine, cytarabine, and etoposide (MACE), mitoxantrone/cytarabine, and cytarabine/L-asparaginase (CLASP). The LAME 89/91 and the BFM93 trials combined high-dose cytarabine with mitoxantrone (HAM) (Wells et al. 1994a; Michel et al. 1996; Dahl et al. 2000; Creutzig et al. 2001a, b). A majority of European groups now assume that high-dose cytarabine should be incorporated into treatment, although there are no recent randomized studies in pediatric AML that have been reported yet to support that policy. An older study by the POG with overall inferior outcome as compared to currently achievable survival rates did not reveal an improvement with higher doses of cytarabine as part of Consolidation therapy (Ravindranath et al. 1991). However, in adult AML it has been shown that high-dose cytarabine at Induction results in better post-remission outcome (Weick et al. 1996; Bishop et al. 1998). High-dose cytarabine as post-remission strategy especially benefited adults with core binding factor or normal karyotype AML (Bloomfield et al. 1998). Knowing that pediatric AML on average includes better risk and thus more chemosensitive patients than adult AML, it seems reasonable to extrapolate these data and to indeed incorporate higher doses of cytarabine in pediatric protocols. Moreover, preclinical studies have shown that AML cells from some patients are more resistant to cytarabine and that in some patients the active transport of cytarabine into the AML cells is limited, a mechanism which is necessary in the case of lower concentrations of cytarabine (Zwaan and Kaspers 2004; Hubeek et al. 2005). Both mechanisms of resistance theoretically can be overcome by higher-dose cytarabine. However, these studies do not identify the optimal moment to use high-dose cytarabine. All European groups except the BFM-AML group still recommend allogeneic HSCT for certain risk groups. In contrast, all groups except AIEOP do not recommend autologous HSCT (stem cell reinfusion would be more appropriate) anymore, in view of several randomized studies that did not demonstrate a benefit from autologous stem cell reinfusion as compared to chemotherapy (Ravindranath et al. 1996; Woods et al. 2001). Several studies seem in favor of using allogeneic HSCT in AML subgroups. However, with contemporary chemotherapy that results in DFS of more than 50%, the number of studies with appropriate data-analysis (correcting at least for time-to-transplant) that demonstrated significantly improved OS with allogeneic HSCT is actually limited to two studies reported by the LAME (Perel et al. 2005) and the European Organisation for Research and Treatment of Cancer (EORTC) (Entz-Werle et al. 2005) groups. These two studies both have a major confounder, which was the Maintenance therapy given only to patients that were not transplanted. Such Maintenance therapy was proven to have an adverse effect on outcome (Perel et al. 2005; Wells et al. 1994a, b), thus negatively influencing the outcome for chemotherapy-only patients but not for patients that were treated with allogeneic HSCT. The BFM-AML group reported no improved OS with allogeneic HSCT in either studies −87 and −93 analyzed together or in study AML-BFM-98 (Langebrake et al. 2005; Creutzig et al. 2006; Reinhardt et al. 2006). Therefore, since 2006 this group has not advocated the use of allogeneic HSCT in AML patients in CR1, while other groups still recommend allogeneic HSCT for higher-risk patients. With respect to the risk of relapse, some European studies have suggested that allogeneic HSCT may offer a benefit. The ability to retrieve relapsed AML appears to be lower after allogeneic HSCT in CR1 than that of relapsed AML after chemotherapy only. This observation was reported by the NOPHO group (Lie et al. 2005; Abrahamsson et al. 2007). The relative benefit of allogeneic HSCT in CR1 depends on the effectiveness and type of chemotherapy. For patients in CR1, allogeneic HSCT does not consistently exhibit a significant benefit in overall survival, it is costly, and it is associated with increased late effects. These issues must be carefully balanced against the reported reduced risk of relapse with allogeneic HSCT. Maintenance chemotherapy has been used in the past by several groups and is still being used by the BFM-AML group. However, most of the other studies in pediatric AML have not observed improved outcome with Maintenance therapy in the context of contemporary high-dose intensive Induction and Consolidation treatment regimens. Two studies have suggested an adverse impact on OS, apparently because of reduced salvagability after relapse (Wells et al. 1994a, b; Perel et al. 2005). Therefore, the use of Maintenance chemotherapy in pediatric AML cannot be generally recommended. An obvious exception is the treatment of APL, in which Maintenance treatment with all-trans retinoic acid (ATRA) and low-dose chemotherapy including methotrexate and mercaptopurine does provide benefit for patients (Testi et al. 2005). It cannot be excluded that specific subgroups benefit from Maintenance chemotherapy, as well as the possibility that other types of Maintenance therapy are beneficial. For example, the NOPHO is studying the value of two doses of gemtuzumab ozogamicin as post-Consolidation therapy in a randomized study, while LAME is studying the effect of interleukin-2. Based upon single agent efficacy data in adults and children, early studies demonstrated significant activity for combinations of cytarabine and anthracyclines in pediatric AML remission induction. The CCG 241 study (1975–1979) incorporated cytarabine and daunomycin into multi-agent (5-azacytidine, vincristine and prednisone) therapy for newly diagnosed pediatric AML (Chard et al. 1978). The POG 8101 (1981–1986) study provided pivotal data demonstrating that a “3+7+7 DAT” combination regimen of 3 days of 45 mg/m2/day daunorubicin, 7 days of continuous infusional low-dose (100 mg/m2/day) cytarabine and 7 days of 6-thioguanine provided superior outcomes for children with AML when compared to a vincristine, dexamethasone and cytarabine regimen originally designed for ALL (Steuber et al. 1991). In this study, the DAT regimen was established as a standard element in pediatric AML therapy with a complete remission rate of 82% (compared to 61% for the other regimen; p = 0.02). Combinations of anthracycline and cytarabine remain at the core of AML Induction regimens, and the CCG 213 study (1986–1989) did not demonstrate benefit with the addition of etoposide, thioguanine and dexamethasone (Wells et al. 1994a, b). Similarly, the CCG-2961 study (1996–1999) was not able to identify a benefit for the addition of idarubicin and fludarabine into an intensive timing remission Induction/Consolidation regimen (Lange et al. 2008). Building upon observations suggesting benefit with enhanced cytarabine exposure, the POG 8498 study (1984–1988) tested the effect of dose intensification during the second Induction course using additional high-dose 3 g/m2 q12 h × 6 doses of cytarabine followed by L-asparaginase. POG 8498 showed a trend toward higher EFS (34% vs 29%) and DFS (42% vs 34%) when compared with standard dose DAT platforms (Ravindranath et al. 1991). The POG 9421 study (1995–1999) examined incorporation of high-dose cytarabine as part of the first Induction course and the use of the multidrug resistance modulator cyclosporine. However, intensifying Induction with high-dose DAT and the addition of cyclosporine to Consolidation chemotherapy did not prolong the duration of remission or improve overall survival (Becton et al. 2006). In combination with data from MRC and BFM studies (see below), these observations have helped define effective cytarabine schedules (e.g., 7–10 days of dosing) or high-dose cytarabine exposure in second Induction course, which remains at the core of contemporary Induction regimens. Although these early improvements in Induction regimens resulted in significant increases in remission induction rates, the impact on EFS and OS was less dramatic. These observations suggested that a small population of residual blasts or leukemia stem cells survives after Induction and later cause relapse. The mechanisms of drug-resistance in AML are complex and multifactorial; however, preclinical data suggest that the leukemic stem cells found in the endosteal region of the marrow are cell cycle quiescent and escape the effects of chemotherapy (Mikkola et al. 2010; Saito et al. 2010). Cytarabine generally requires active cell division to exert full anti-leukemic activity, and in mouse models, it appears that the use of hematopoietic growth factors mobilizes quiescent AML stem cells and stimulates proliferation, rendering them sensitive to the cytotoxic effects of cytarabine (Saito et al. 2010). Exposure to chemotherapy has been hypothesized to exert similar effects on cell cycle quiescent leukemia stem cells (i.e., recruitment into the cell cycle and increased sensitivity to subsequent chemotherapy, particularly S phase active agents such as cytarabine [Vaughan et al. 1984]). In the pivotal CCG 2891 study (1989–1995), Woods and colleagues tested the hypothesis that intensively timed sequential chemotherapy with concurrent G-CSF in Induction would overcome resistance based upon leukemia cell cycle quiescence. CCG 2891 employed a core regimen of a 4-day cycle of dexamethasone, cytarabine, 6-thioguanine, etoposide and daunorubicin (DCTER) and enrolled 589 patients. The randomized study design compared two Induction approaches involving either (1) “intensive timing” in which the second cycle of chemotherapy was administered after a 6-day rest regardless of low or dropping blood counts, or (2) “standard timing” in which the second cycle of chemotherapy was administered at least 14 days or more from the beginning of the first cycle, depending on bone marrow status and recovery of hematopoiesis (Woods et al. 1996). This study demonstrated that intensively timed Induction therapy resulted in a marked improvement in actuarial EFS at 3 years of 42% versus 27% for patients on the standard timing arm (p = 0.0005). In addition, patients receiving intensively timed Induction therapy had superior DFS results at 3 years from the end of Induction, 55% versus 37% for standard timing patients (p = 0.0002). In post-Induction therapy, cycles of non-cross resistant drugs or high-dose exposure to limited agents (e.g., high-dose cytarabine) are administered to eradicate minimal residual disease and increase potential for cure. The biological heterogeneity of AML poses difficulties in identifying straightforward conclusions regarding the intensity or number of post-remission cycles required for high cure rates. The CCG 213P study (1983–1985) investigated the utility and timing of cycles of very high-dose cytarabine in therapy intensification in children with AML who had achieved remission after Induction therapy. This study suggested that Maintenance therapy did not appear to benefit patients who had received intensively timed high-dose cytarabine Intensification therapy, though patients receiving non-aggressive Intensification appeared to have improved outcome with Maintenance therapy. Patients who received two intensive courses of high-dose cytarabine and asparaginase administered at 7-day intervals exhibited a significant increase in DFS when compared with patients receiving the same therapy administered at 28-day intervals (OS 58% vs 41% from the end of Induction, p < 0.04) (Woods et al. 1990). Using a randomized design, the successor CCG 213 study (1986–1989) also demonstrated no benefit for Maintenance therapy after intensive cytarabine exposure in post-Induction/Consolidation therapy (Wells et al. 1994a, b). The hypothesis that immunologic interventions could initiate an anti-leukemia response in the post-remission setting and eliminate MRD was tested in the CCG-2961 study. However, the use of interleukin-2 in Maintenance did not show a benefit in reduction of relapse (Lange et al. 2008). The majority of randomized North American studies incorporating analysis by intent to treat have consistently indicated efficacy for myeloablative high-dose chemotherapy Consolidation therapy with stem cell rescue from an HLA-matched family donor (MFD) (Dahl et al. 1990; Feig et al. 1993; Nesbit et al. 1994; Dinndorf and Bunin 1995; Appelbaum 1997; Cassileth et al. 1998). However, improvements in contemporary chemotherapy approaches have suggested that large subsets of children with AML in CR1 have equivalent outcomes without need for stem cell transplantation and its increased risk for immediate and long-term toxicities (Creutzig and Reinhardt 2002). The relatively modest number of children receiving HSCT as part of AML therapy has created difficulties in the clear identification of risk factors and patients who will likely benefit from the procedure. A recent large multi-cooperative group analysis of factors impacting transplant outcome by Horan and colleagues combined data on 1,373 patients drawn from four cooperative group Phase III trials: POG 8821, CCG 2891, CCG 2961, and MRC 10 (Ravindranath et al. 1996; Stevens et al. 1998; Woods et al. 2001; Horan et al. 2008; Lange et al. 2008). The data set aggregated from the four studies compared 893 patients assigned to chemotherapy alone with 480 assigned to allogeneic HSCT. The majority of the 480 allogeneic HSCT patients included in this intent-to-treat analysis were from the CCG studies (n = 334), with additional patients from the POG (n = 70) and MRC (n = 76) studies. This analysis confirmed that OS and DFS were superior for the HSCT group. At 8 years, estimates for OS for HSCT and chemotherapy groups were 63% and 57% (HR = 0.77; p < 0.007), respectively, and DFS estimates were 56% and 46% (HR = 0.70; p < 0.001), respectively. The incidence of relapse was much lower in patients assigned to HSCT compared with those receiving chemotherapy alone (28% vs 47% (HR = 0.51; p <0.001)). However, patients assigned to HSCT had a higher rate of treatment-related mortality compared to those receiving chemotherapy alone (16% vs 7% (HR = 1.97; p < 0.001)). The Horan study also stratified into favorable-, intermediate-, and poor-risk disease groups. Favorable-risk disease included patients with inv(16) and t(8;21). Patients with monosomy 7, monosomy 5, deletions of 5q, or more than 15% blasts after the first course of chemotherapy were considered to have high-risk disease. All other patients were classified as having intermediate-risk disease. In addition, a small number of MRC patients with abnormalities of 3q and patients with five or more cytogenetic abnormalities were classified in the high-risk disease subgroup. The risk-stratified analysis showed a significant benefit in OS for patients who received a transplant in CR1. No significant overall benefit was seen in patients with favorable-risk or poor-risk disease. In the intermediate-risk group, the estimated OS at 8 years was 51% for the patients assigned to chemotherapy alone, compared with 62% for the patients assigned to HSCT (HR = 0.69; p <0 .006). As might be expected, the predominant beneficial effect of allogeneic transplant was in reducing the risk of relapse in this subgroup of patients; the relapse incidence was 26% in the HSCT patients compared to 54% in the chemotherapy-only patients (HR = 0.42; p < 0.001). In the favorable-risk disease group, a higher incidence of HSCT treatment-related mortality negated the beneficial effects of a modest reduction in the risk for relapse. A review of five consecutive St. Jude AML studies clearly identified challenges for future directions in pediatric AML therapy. Similar to other studies, incremental progress was achieved by identification of active agents and refinements in dose and schedule of administration. However, these investigations elegantly identified limits in the benefit of increasing the intensity of chemotherapy (AML-87) or prolongation of post-remission therapy by adding sequential myeloablative (AML-80) or nonmyeloablative (AML- 83) chemotherapy cycles. A powerful result of these studies is that risk-adapted and risk-targeted approaches will be necessary to further improve approaches for pediatric AML (Ribeiro et al. 2005). The COG-AAML0531 study for children with de novo AML is investigating whether the addition of gemtuzumab ozogamicin to contemporary therapy will improve EFS over treatment with chemotherapy alone. AAML0531 introduces gemtuzumab ozogamicin into the first Induction course and in Intensification/Consolidation courses of a regimen similar to MRC AML12 therapy with five courses of total treatment. In AAML0531, high-risk patients receive a HSCT if a matched family donor (MFD) or alternative donor is available. Intermediate-risk patients receive a HSCT only if a MFD is available. High and intermediate-risk patients without an appropriate HSCT donor and all low-risk patients receive additional chemotherapy with or without gemtuzumab ozogamicin consistent with the initial randomization. Recent MRC, European, Japanese and North American studies parallel findings in other groups and observe overall survival rates of 45–65% for most children with AML with contemporary treatment approaches. These studies have highlighted key elements of current approaches for AML therapy, including: (a) the benefit of dose-intensification early in therapy, including timed-sequential therapy approaches, (b) the need for intensive Consolidation therapy, and (c) the key importance of stratification approaches with cytogenetics/molecular genetics and response to therapy as the key predictors of response. In parallel, these approaches to risk stratification have assisted in the development of tailored less intensive but highly effective therapies for patients with Down syndrome, and the use of tailored therapy for patients with acute promyelocytic leukemia (see below). In addition, these investigations are generating improved methods to predict which patients will benefit from allogeneic transplantation in CR1, and a rationale for the use of novel targeted therapy approaches for patients with an extremely poor prognosis, including FLT3 high allelic ratio internal tandem duplications, and patients with leukemic cells with high-risk cytogenetic features. The incidence of AML cells in the cerebrospinal fluid at initial diagnosis differs between studies, but is in the range of 5–30%. Moreover, CNS involvement at relapse is observed in up to 10% of patients, while it was up to 20% in studies in which CNS prophylaxis had not been used (Pui et al. 2004; Kaspers et al. 2008). Therefore, there are good arguments to administer prophylaxis for CNS relapse. Most groups successfully use intrathecal (IT) chemotherapy given four to six times, in that their CNS relapse rates or cumulative incidence of any relapse do not seem to be higher than that reported by the BFM-AML group, which still uses cranial irradiation. The latter policy is based on a single randomized study, and cranial irradiation provided a statistically significant benefit only when randomized and non-randomized patients were pooled. Of interest, the main benefit concerned less bone marrow relapses (Creutzig et al. 1993). Indeed, a St. Jude study showed that even CNS leukemia at initial diagnosis can be successfully treated with chemotherapy only (Abbott et al. 2003). In conclusion, prophylaxis of CNS relapse is indicated using intrathecal chemotherapy, either single-agent cytarabine or cytarabine in combination with prednisolone and methotrexate, while there seems to be no place for prophylactic cranial irradiation in the setting of current intensive chemotherapy. There are no data in favor of either single-agent cytarabine or triple chemotherapy. Children with CNS disease at diagnosis have traditionally received intrathecal chemotherapy and cranial radiotherapy. The CCG-2891 study showed that children with CNS leukemia at diagnosis can be cured without the use of cranial irradiation (Woods et al. 1996) Acute myeloid leukemia can be modulated by the graft versus leukemia effect of allogeneic HSCT, and both allogeneic and autologous HSCT have been reported to result in a reduction in relapse risk in children (Stevens et al. 1998; Burnett et al. 2002). However, the benefit of allogeneic transplantation in CR1 is increasingly challenged by a DFS approximating 60% reported for chemotherapy alone. Importantly, many studies fail to translate a reduction in relapse rate into an improvement in OS (Stevens et al. 1998; Burnett et al. 2002; Reinhardt et al. 2006), and others only demonstrate improved survival for specific risk groups (Horan et al. 2008). Differences in reported benefits for allogeneic transplantation may be explained by differences in the intensity, effectiveness and outcome of chemotherapy (Creutzig et al. 2001a, b; Woods et al. 2001; Entz-Werle et al. 2005; Perel et al. 2005) and the method of data analysis (Wheatley 2002). The lack of benefit found by the MRC for allogeneic HSCT in CR1 mirrors the experience of the BFM, who also find no improvement in OS, but is contradictory to experience reported by the COG (Woods et al. 2001). Cytogenetics and speed of response to treatment, variably assessed, are powerful indicators of outcome in AML and may be used to target the most intensive therapy to those patents at greatest risk of relapse. However, it is questionable whether allogeneic transplantation benefits even high-risk patients. The MRC report a reduction in the relapse risk for good and standard-risk patients only undergoing allogeneic HSCT (although no benefit in OS), but not for poor-risk disease patients. Without a reduction in relapse rate, allogeneic HSCT cannot benefit patients. Interestingly, the COG reported a significant improvement in DFS and OS for HLA-matched related HSCT in intermediate-risk patients only but not good or poor risk patients (Horan et al. 2008). Allogeneic HSCT only provides an advantage for patients with AML if the observed reduction in relapse rate is sufficiently great to absorb the procedure-related mortality and the relapse rate post-transplant. Supporters of allogeneic HSCT in CR1 cite the high procedure-related mortality associated with older trials used to assess the benefit of allogeneic HSCT as unfavorably compromising HSCT outcomes, and argue that improvements in current HSCT practice might alter the balance in favor of HSCT. There is some evidence that current low procedure-related mortality, improved transplant procedures and more effective rescue of patients relapsing after transplant have improved HSCT outcomes (Vicente et al. 2007), but it remains to be proven that these will be significantly superior to chemotherapy, which is also improving over time. While allogeneic HSCT can reduce the relapse risk, at least in some risk groups, benefit in overall survival will require a low procedure-related mortality and innovative modulation of the graft versus leukemia effect. Some studies report both a reduction in relapse rate and an improvement in DFS. The failure to translate this benefit into an improvement in OS is due to improved salvage, particularly in those patients not previously transplanted, and raises the question of whether HSCT might best be reserved for second CR. HSCT may benefit some risk groups more than others, and improved risk group stratification by molecular characterization and MRD may help identify these groups. Most studies have involved the use of matched family or volunteer donors and the benefit of haplo-identical donors and cord stem cells remains untested. The role of any improvement in outcome must also be balanced against the long-term morbidity in children and must be cost effective. APL is a rare subtype of AML characterized by distinctive morphology (M3 or M3v), the t(15;17) translocation, a potentially fatal coagulopathy at presentation and unique sensitivity to anthracycline-containing chemotherapy and differentiating agents all-trans retinoic acid (ATRA) and arsenic trioxide (ATO). Children with APL treated with a combination of anthracycline-based chemotherapy and ATRA are reported to have a 5 year OS of 87–90%, DFS of 78–82% and EFS of 71–77% (Mann et al. 2001; de Botton et al. 2004; Ortega et al. 2005; Testi et al. 2005). Unpublished data suggest that outcomes have improved further with the increased use of ATRA, adherence to supportive care guidelines and the use of ATO in salvage therapy. However, these results are achieved with high cumulative anthracycline doses, with their potential for cardiotoxicity in children. ATRA and ATO target the APL-associated PML-RARα oncoprotein and offer a low or non-chemotherapy treatment approach, which is being increasingly tested in adult patients with this subtype of AML. While theoretically an attractive approach in children, there is limited information on the use of ATO in frontline therapy or on its long-term toxicity in this age group. APL represents approximately 4–8% of pediatric AML (Gregory and Feusner 2003), although a higher incidence has been reported in children of Hispanic and Mediterranean origin, which may suggest a genetic predisposition and/or exposure to environmental factor(s) in the causation. The median age at presentation has been reported between 12 and 15 years (de Botton et al. 2004; Ortega et al. 2005; Testi et al. 2005). It rarely occurs in the first year of life. Therapy-related APL has been described in children previously exposed to epipodophyllotoxins (Detourmignies et al. 1992). APL is characterized by rearrangements of the retinoic acid receptor-α (RARα) gene on chromosome 17q21. Several partner genes have been identified, but RARα is most commonly (>95 %) fused to the PML gene (Promyelocytic Leukemia) on chromosome 15q22 as a result of the t(15,17) (q22;q21) translocation. Identification of the underlying molecular lesion is critical to the management of APL because it determines sensitivity to molecularly targeted therapies, ATRA and ATO. PML- RARα / t(15,17) (q22;q21) APL is retinoid sensitive as are the rarer APL subtypes seen in children of NPM1-RARα / t(5;17) (q35;q21) and NuMA-RARα / t(11;17) (q13;q21) translocations. The sensitivity of these rarer subtypes to ATO has not been established. PLZF-RARα / t(11;17) (q23;q21) and STAT5b-RARα / t(17;17) (q11.2;q21) fusions are retinoid resistant (Grimwade et al. 2000; Mistry et al. 2003). PRKAR1A – RARα / t (17;17) (q21;q24) and FIP1L1-RARα/ t(4;17) (q12;q21) have also been described. Secondary chromosomal changes can be present at diagnosis. RARα is a ligand-dependent transcriptional activator that binds to specific DNA sequences found in the promoters of retinoic acid-responsive genes, which are involved in normal myeloid development. In APL, the chimeric oncoprotein (i.e., PML-RAR α, NPM1-RARα, etc.) is capable of DNA binding but prevents transcription of RARα target genes and myeloid cell differentiation. ATRA functions by binding to the RARα moiety of PML-RARα, causing degradation of the fusion protein and thus allowing terminal differentiation of leukemia promyelocytes. APL is classified amongst Acute Myeloid Leukemia with recurrent genetic abnormalities by the WHO classification and as AML M3 or M3v by the FAB classification. Morphology: The morphology of APL is distinctive, with two subtypes, M3 and M3v (15–30%): classical hypergranular M3 is characterized by hypergranular promyelocytes containing heavy azurophilic granules, bundles of Auer rods (faggots) and bilobed nuclei, while hypogranular/microgranular variant M3v blasts have an apparent paucity of granules. Immunophenotype: APL has a distinctive, but not diagnostic, phenotype of positivity for CD33, CD13, CD9 and rare expression of HLA-DR, CD34, CD7, CD11b and CD14. Antigen profiles (CD2 and CD56) may correlate with inferior outcome. Immunohistochemistry: Immunohistochemical staining provides a rapid diagnosis. The characteristic (microspeckled) nuclear PML distribution pattern of APL blasts is readily distinguishable from the wild-type (speckled) PML nuclear staining of normal cells. Cytogenetics: Conventional cytogenetics or fluorescence hybridization identifies both common and rarer RARα translocations. Conventional cytogenetics fail to identify the t(15;17) in approximately 10% of patients with PML-RARα transcripts, including those in whom the fusion gene is generated through insertion events in which chromosomes 15 and 17 may appear normal (Grimwade et al. 2000). Molecular analysis: Molecular diagnosis by reverse transcriptase polymerase chain reaction is important in cases lacking the t(15;17) by conventional cytogenetics and is critical to determine PML-RARα isoform type (bcr1,bcr2,bcr3) to allow subsequent, MRD monitoring of fusion gene transcripts by RT PCR assays (Sanz et al. 2009). The strongest predictors of outcome are the persistence of MRD, followed by the WBC count (Grimwade and Lo Coco 2002; Grimwade et al. 2009a, b). M3v morphology, FLT 3 length mutations, bcr3 PML breakpoint and higher PML-RARα expression ratios have been variably reported to be of independent significance but are interrelated and often found in association with a high WBC count. MRD: Persistence or recurrence of PML–RARα or RARα-PML fusion transcripts detected by RT PCR monitoring is the strongest independent predictor of clinical relapse and exceeds WBC count (p = 0.02) on multivariate analysis (Grimwade et al. 2009a, b). WBC count at diagnosis: The presenting WBC count strongly predicts for the risk of Induction death and relapse (Burnett et al. 1999; Ortega et al. 2005; Testi et al. 2005). The Sanz criteria classify APL into low, intermediate and high-risk based on the WBC count and platelet count (Sanz et al. 2000), but other groups have not found the platelet count to be of predictive value. In the GIMEMA-AIEOP study, the EFS was 83% versus 59% at 10 years (p = 0.7) for standard (WBC < 10 × 109/L) versus high-risk (WBC =/> 10 × 109/L) children respectively, and all Induction deaths (4%) were in high-risk patients (Testi et al. 2005). Children more commonly have high-risk disease (35–40%) than adults (de Botton et al. 2004; Ortega et al. 2005; Testi et al. 2005). Children with APL treated on anthracycline- and ATRA-based protocols have a CR rate of around 95% and an EFS of 70–80% (Mann et al. 2001; de Botton et al. 2004; Ortega et al. 2005; Testi et al. 2005). These results have been achieved with heterogeneous regimens. APL is uniquely sensitive to anthracyclines, ATRA and ATO. This presents two challenges for the treatment of children with APL. Firstly, high cumulative anthracycline doses (650 mg/m2) (Testi et al. 2005) are commonly used, and reducing cardiotoxicity should be a priority in a leukemia with such a favorable outcome. Secondly, there are limited data on the use of ATO in frontline therapy in children both in terms of efficacy and late toxicities. Induction therapy: The diagnosis of APL is generally considered an oncologic emergency, and therapy with ATRA should be initiated immediately in patients presenting with leukemia and characteristic peripheral blood blast morphology, particularly with evidence of coagulopathy. Patients are at high risk for early fulminant fatal coagulopathy, with deaths predominantly occurring within the first 5–10 days of presentation, and amenable to treatment with ATRA. The simultaneous use of ATRA and anthracycline monotherapy is currently considered optimal Induction chemotherapy (Mandelli et al. 1997; Fenaux et al. 1999; Sanz et al. 2003, 2009). In two separate trials employing simultaneous anthracycline (idarubicin) and ATRA as Induction therapy, the CR and Induction death rates for children were reported at 92% and 96% and 4% and 7.5% respectively (Ortega et al. 2005; Testi et al. 2005). Primary resistance to ATRA and idarubicin is extremely rare. However, children who do not have retinoid sensitive APL should receive standard intensive AML type therapy with anthracyclines and cytarabine. Consolidation therapy: Comparable results are reported for anthracycline monotherapy (anthracyclines and ATRA) and combination chemotherapy (anthracyclines, ATRA, cytarabine with or without intercalating agents). Any advantage for the addition of any drug to anthracycline monotherapy is protocol dependent and may relate to the anthracycline and to its cumulative dose, cumulative doses of the other drugs and to the risk group. Whilst anthracycline monotherapy in Consolidation has the advantage of reduced morbidity and mortality, there were no remission deaths in four reported pediatric studies (Mann et al. 2001; de Botton et al. 2004; Ortega et al. 2005; Testi et al. 2005) using variable Consolidation courses. The benefit of cytarabine in Consolidation is uncertain. Combined analysis of the GIMEMA and PETHEMA APL trials, which differed only in the inclusion or exclusion of drugs other than anthracyclines during Consolidation, showed that the omission of non-anthracycline drugs was not associated with an inferior anti-leukemia effect (Sanz et al. 2000). In contrast, the European APL 2000 study reported benefit for cytarabine in combination with daunorubicin (495 mg/m2) and ATRA in newly diagnosed patients with APL (Ades et al. 2006). Cytarabine was given in both Induction and Consolidation. The benefit of cytarabine may be dependent on the anthracycline and its cumulative dose. A comparison of PETHEMA LPA 99 (high cumulative dose of idarubicin and mitoxantrone, no cytarabine) and APL 2000 (lower cumulative dose of daunorubicin and cytarabine) found benefit for cytarabine in high-risk patients with a WBC count of >10 × 109/L (p = 0.026) (Ades et al. 2006). Patients with high-risk disease may benefit from more intensive Consolidation. The PETHEMA LPA 99 study employing anthracycline monotherapy in Consolidation reported a reduction in the relapse rate in high and intermediate-risk patients from 20.1% to 8.7% (p = 0.004) when ATRA was added to Consolidation and the anthracycline dose intensified (750 mg/m2 daunorubicin equivalent) (Sanz et al. 2004). High risk patients treated on AIDA 2000 received identical Induction and Consolidation to those treated on AIDA 0493 with the addition of ATRA to each Consolidation block. The relapse rate for high-risk patients in AIDA 2000 was 2% at a median follow-up of 2 years compared to 29% for AIDA 0493 at a median follow-up of 4.5 years (p = 0.0004) (Lo Coco et al. 2004). Maintenance treatment: The benefit of Maintenance therapy with intermittent ATRA (15 days every 3 months) in combination with 6-mercaptopurine and methotrexate has been variably reported (Tallman et al. 1997; Fenaux et al. 1999; Testi et al. 2005). It is unclear if ATRA in Maintenance is of benefit to patients who receive prolonged exposure in earlier phases of treatment or to those who are in molecular remission at the end of Consolidation. Molecular monitoring in APL: Serial molecular MRD monitoring by RT PCR during and after therapy can detect persistent or recurrent molecular disease and guide pre-emptive therapy. The persistence of PML-RARA transcripts at the end of Consolidation (but not Induction) predicts for the risk of hematological relapse as does the recurrence of PCR positivity; both situations may be averted by pre-emptive therapy (Diverio et al. 1998; Burnett et al. 1999). The results from two pediatric studies suggest that less than 5% of children are MRD positive at the end of Consolidation (Ortega et al. 2005; Testi et al. 2005). Hematopoietic stem cell transplantation: The only role for HSCT in frontline therapy is for the small number of patients with persistent or recurrence of MRD who may benefit from either allogeneic or autologous transplantation dependent on their molecular status after salvage therapy. All-trans retinoic acid: An ATRA dose of 25 mg/m2 appears to be effective in children (de Botton et al. 2004; Ortega et al. 2005; Testi et al. 2005), which is considerably lower than that of 45 mg/m2 routinely employed in adults. Three studies using an ATRA dose of 25 mg/m2 report EFS of 71–82% and OS of 76–90%, suggesting efficacy. Hemorrhage: Death from hemorrhage, particularly intracranial hemorrhage (ICH), is the most common cause of failure to achieve CR, and children with a WBC count >10 × 109/L and the M3 variant are at particular risk. This classifies APL as a medical emergency. The coagulopathy can be initiated or exacerbated by chemotherapy, and ATRA, which improves the coagulopathy, must be started immediately on morphological suspicion and not delayed until the diagnosis is molecularly confirmed. Patients with a high WBC count should not undergo leukopheresis as this can exacerbate the coagulopathy. The incidence of death from ICH in children is reported at 3–4% despite increased awareness, pre-emptive therapy and the use of ATRA (de Botton et al. 2004; Ortega et al. 2005; Testi et al. 2005). Fibrinogen levels should be maintained above 1.5–2 g/L with fresh frozen plasma (FFP) or cryoprecipitate and the platelet count above 50 × 109/L until the coagulopathy resolves (Milligan et al. 2006). There is no proven benefit for the use of heparin, tranexamic acid or anti-fibrinolytic drugs. NovoSeven may be considered in life-threatening hemorrhage unresponsive to platelets, FFP and cryoprecipitate. APL differentiation syndrome: APL differentiation syndrome is associated with high mortality if not promptly treated (De Botton et al. 1998; de Botton et al. 2003). It is thought to be related to surface adhesion molecule modulation and cytokine release following ATRA- or ATO-induced differentiation of APL cells (Frankel et al. 1992). Dexamethasone should be introduced at the earliest signs of APL differentiation syndrome, and if the condition progresses, ATRA/ATO should be temporarily discontinued. Although there is no proven benefit for prophylactic dexamethasone, it is usually given simultaneously with chemotherapy to patients with high WBC count who are at particular risk (De Botton et al. 1998, 2003). The incidence of APL differentiating syndrome is similar in children and adults (definite 3–5%, indeterminate 5–15%) with a mortality rate of less than 1% (Mann et al. 2001; de Botton et al. 2004; Ortega et al. 2005; Testi et al. 2005). Pseudotumour cerebri: Children have a higher incidence of ATRA-associated headache and pseudotumour cerebri than adults: 13–30% and 5–9%, respectively for children receiving ATRA at a dose of 25 mg/m2 (Mann et al. 2001; Ortega et al. 2005; Testi et al. 2005) and 39% and 16%, respectively for those receiving 45 mg/m2 (de Botton et al. 2004). In addition, these studies suggest that there is no relative benefit in terms of disease control in treating children with doses higher than 25 mg/m2. Arsenic trioxide (ATO): Preliminary data from adult studies suggests at least comparability between ATO and single agent ATRA in terms of the achievement of CR (>90%) and that combined ATO and ATRA is superior to ATO or ATRA alone (Niu et al. 1999; Soignet et al. 2001; Shen et al. 2004). A small single center pediatric study using monotherapy with ATO reported a hematological and molecular remission rate of 91% with a relapse-free survival (RFS) and OS of 81% and 91%, respectively at 30 months (George et al. 2004). ATO and ATRA with or without the addition of gemtuzumab ozogamicin have shown favorable results in adults (Estey et al. 2006; Ravandi et al. 2009). The addition of ATO to Consolidation (Consolidation blocks 1 and 2: ATO 0.15 mg/kg/day for 5 days/week for 5 weeks) following remission induction with ATRA and chemotherapy has been reported to significantly improve the EFS and OS in adults with newly diagnosed APL (Powell 2007). ATO is usually well tolerated, although its use in Induction and hematological relapse is associated with hyperleukocytosis and APL differentiation syndrome. Other adverse effects include prolongation of the QT interval, reversible periphe-ral neuropathy and skin hyperpigmentation. Little is known of late ATO cardiac and neurological toxicity, but hyperpigmentation, palmar keratosis, distal neuropathy and muscular atrophy have been observed with long-term use. There is a need to define toxicity in children and to assess the role and best strategy for the use of ATO in this age group. Approximately 3–5% of children have persistent molecular disease at the end of Consolidation and 20–25% have either a molecular or hematological relapse. The outcome following molecular relapse with pre-emptive therapy is superior to that following hematological relapse. Approximately 3–5% of APL patients develop extramedullary (EM) relapse, usually involving the CNS or the skin. ATO has emerged as the single most active agent in patients with relapsed APL, with CR rates of approximately 80–90% and most patients achieving molecular remission after two cycles (Ghavamzadeh et al. 2006; Lo Coco et al. 2007; Tallman 2007). There appears to be no advantage for combining ATRA with ATO in the relapsed setting. The best Consolidation strategy after ATO-induced second remission is unknown, but options include additional cycles of ATO, ATRA, gemtuzumab ozogamicin, and standard chemotherapy and HSCT. Gemtuzumab ozogamicin offers the combination of calicheamicin – a cytotoxic agent with similarities to anthracyclines, and an anti-CD33 antibody; CD33 is homogenously expressed in virtually 100% of APL. Data from the European Bone Marrow Transplant (EBMT) Registry report a 5-year cumulative incidence of leukemia-free survival (LFS) of 51% for 195 patients autografted in CR2 and a 5-year LFS of 59% for the 137 patients allografted in CR2 (Sanz et al. 2007). The appropriate procedure is guided by the molecular status, age, availability of an HLA-identical donor and the time from diagnosis to transplant. Extramedullary relapse has been increasingly reported since the introduction of ATRA in the treatment of APL (Evans and Grimwade 1999; Specchia et al. 2001; Breccia et al. 2003; de Botton et al. 2006). This may be due to longer survival increasing the number of patients at risk, although the role of ATRA in mediating increased expression of adhesion molecules has been questioned. It has also been suggested that APL differentiation syndrome and intracerebral bleeding during Induction might be associated with an increased risk of extramedullary relapse (Evans and Grimwade 1999; Specchia et al. 2001; Breccia et al. 2003; de Botton et al. 2006). EM relapse usually occurs in the CNS and is typically accompanied by overt or molecular bone marrow relapse. EM relapse in APL should be regarded as a systemic relapse and both CNS directed and systemic therapy given (Sanz et al. 2009). Irrespective of the intensity of treatment, patients with an EM relapse have an OS comparable with those with a hematological relapse. Although ATO crosses the blood–brain barrier, the concentrations in cerebral spinal fluid (CSF) achieved by intravenous infusion are probably insufficient for the treatment of meningeal leukemia (Knipp et al. 2007). This combined with the potential for increased ATO neurological toxicity in individuals receiving concomitant intrathecal chemotherapy and cranial irradiation may limit its usefulness in CNS relapse. This argues in favor of the use of CNS intrathecal prophylaxis in the frontline protocols (starting with Consolidation), at least for patients with unfavorable presenting features associated with a higher risk of CNS relapse, expecting that this might result in a significant proportion of patients being over-treated. FLT-3 inhibitors: FLT-3 inhibitors offer the possibility of using combination therapy that targets both mutations contributing to the pathogenesis of APL: FLT-3 inhibitors for FLT3/ITD and ATRA for the PML/RARα fusion. However, FLT3 mutations may only exist in subclones or be lost at relapse, and FLT3 inhibitors and ATRA may not work synergistically in vivo (Gale et al. 2005). Histone deacetylase (HDAC) inhibitors (Fazi et al. 2005) may have a potential role in the treatment of APL in the future. The outcome for children with APL is likely to continue to improve as novel approaches such as the addition of ATO and gemtuzumab ozogamicin to conventional frontline therapy and the possibilities of low- or non-chemotherapy approaches are explored. The anti-leukemia effect of reducing the cumulative anthracycline dose, or substituting lipsomal anthracyclines, should be tested within a clinical trial. The benefit, content and duration of Maintenance merit clarification. Risk stratification may be more accurately defined by molecular monitoring than by the WBC count, and targeting treatment to the risk of relapse may reduce unnecessary toxicity to standard-risk patients. The ideal frequency of molecular monitoring, the best pre-emptive therapy, and the role of novel agents should be established. Finally cardiotoxicity and treatment-related myelodysplasia deserve careful monitoring, as do the toxicities of new agents as they are introduced. Children with Down syndrome are at an increased risk for developing both acute myeloid and acute lymphoblastic leukemia. The increased incidence of acute myeloid leukemia in children with Down syndrome was first described in 1957 by Krivit and Good, and is predominately observed during the first 4 years of life (Krivit and Good 1957). The risk for the development of hematologic abnormalities including leukemia is relatively high in early childhood though it decreases with age. Hasle and co-workers conducted a Danish population-based study of 2,814 individuals with Down syndrome and found that the risk for leukemia decreased dramatically as patients became older, with a standardized incidence ratio (SIR) of 56 for children aged between 0 and 4 years, decreasing to 10 for individuals between 5 and 29 years of age. Significantly, no cases of leukemia were observed after 29 years of age. The cumulative risk of leukemia is 2.7% by age 30 (Hasle et al. 2000). Among Down syndrome leukemia patients, malignancies derived from primitive megakaryocytic and erythroid blood progenitors predominate (FAB-M7 subtype), with a 600-fold increase in relative risk for acute megakaryocytic leukemia (AMLK) as compared to other children (Zipursky et al. 1994). The biological and clinical features of AMKL in children with Down syndrome are distinctly different from other forms of AMKL, and Down syndrome AMKL is now considered a specific sub-type of AML in DS in the World Health Organization (WHO) classification (see above). In addition, the risk for myelodysplastic syndrome (MDS) with marrow fibrosis is also markedly increased in children with Down syndrome, with an estimated relative risk of 175, and in many cases the MDS precedes the development of leukemia (Creutzig et al. 1996; Lange et al. 1998). Newborns with Down syndrome are also at risk for development of an abnormal myeloproliferative disorder, and approximately 5–10% may exhibit leukocytosis with circulating blast-like cells (Zipursky 2003). This is often accompanied by hepatosplenomegaly, reflecting a proliferation of a hematopoietic progenitor from the liver. Most cases of the myeloproliferation resolve spontaneously, and the main thrust of therapy is to protect the patient from life-threatening complications of hyperleukocytosis, liver injury, or related metabolic complications. This transient myeloproliferative disorder (TMD) almost exclusively presents in infants less than 2–3 months of age, and is associated with an estimated 10–20% risk for the development of subsequent characteristic acute myeloid leukemia of infants with Down syndrome in 3–4 years (Massey et al. 2006; Klusmann et al. 2008; Muramatsu et al. 2008). While all children with Down syndrome share the additional genes from chromosome 21, the reason why only a small subset develops acute leukemia or exhibits neonatal transient myeloproliferative disorder is unknown. Human chromosome 21 contains over 300 genes and other regulatory transcripts, including the Down syndrome critical region at 21q22, which is associated with the predominance of phenotypic characteristics of Down syndrome. Multiple lines of evidence clearly implicate additional copies of this region of chromosome 21 as at least partially responsible for the initiation of Down syndrome leukemia. Notably, in individuals who are mosaic for Down syndrome, if a leukemia or transient myeloproliferative disorder occurs, it is always found to harbor additional copies of chromosome 21. Chromosome 21, and particularly the 21q22 region, includes a variety of genes implicated in leukemia. These include oncogenes such as ERG, ETS2, RUNX1, and SON, as well as other genes critical in folate/single carbon metabolism and oxidant stress. Limited detailed examination of oncogenes found on chromosome 21q22 has, as of yet, not identified consistent frequent patterns of mutation leading to pathologically constitutive activation. In addition, various mouse models with trisomy of syntenic regions of chromosome 21 variably exhibit hematologic disorders such as anemia, macrocytosis and platelet abnormalities, but they are not consistently associated with spontaneous leukemia. These observations have led to the search for other associated genetic mutations that may be responsible for or contribute to the evolution of leukemia in patients with Down syndrome. In the context of AMKL, a key observation has been the identification of somatic mutations in the megakaryocyte-erythroid transcription factor, GATA1, located at chromosome Xp11.23 (Wechsler et al. 2002). The blasts in the vast majority of patients with Down syndrome and transient myeloproliferative disorder and AMKL, possess N-terminal truncating mutations in GATA1 that produce a shortened version of the protein (GATA1s), and result in a block in differentiation and proliferation of megakaryocytic precursors (Wechsler et al. 2002; Groet et al. 2003; Hitzler et al. 2003; Mundschau et al. 2003; Rainis et al. 2003; Xu et al. 2003; Shimada et al. 2004). The reason for the particular propensity of GATA1s mutations to be found in Down syndrome, TMD and AMKL is unknown, and it appears likely that additional cancer-causing mutations, or genetic “hits,” are also necessary for the full development of AMKL. In the absence of Down syndrome, an inherited mutation of GATA1 results in the expression of only the short isoform and causes anemia and neutropenia, but not leukemia (Hollanda et al. 2006). A variety of mutations in other oncogenes (e.g., JAK3 genes, FLT3, and RAS) have been described as occurring at variable frequencies in the AMKL of Down syndrome patients, as have alterations of tumor suppressor genes including TP53 (Malinge et al. 2009). Recently, the chromosome 21 micro-RNA, miR-125b-2, has been identified as increasing proliferation and enhancing self-renewal of megakaryocytic and megakaryocytic/erythroid progenitors, and it appears to interact with the GATA1s in preclinical Down syndrome AMKL models. The target genes of miR-125b are down-regulated in DS-AMKL, and miR-125b-2 has been proposed as an onco-miR involved in the pathogenesis of Down syndrome AMKL (Klusmann et al. 2010). These additional mutations and genetic alterations may provide clues to other pathways that specifically cooperate with trisomy 21q22 and GATA1s mutations in the development of TMD and AMKL. This emerging data suggests a stochastic pathway for the development of AMKL in children with Down syndrome, with additional copies of chromosome region 21q22 representing a “first hit,” the subsequent acquisition of GATA1s mutations representing a “second hit,” and transformation to leukemia also requiring altered expression or mutations in other pathways. The leukemias in children with Down syndrome may share common early “hits” with a variety of subsequent parallel common pathways to leukemia. The acquisition of late leukemogenic mutations may be influenced by other features associated with increased genotoxicity in Down syndrome (e.g., alterations in oxidant and single carbon/folate metabolism, etc.), and various models for leukemogenesis have been proposed (Cabelof et al. 2009; Malinge et al. 2009; Roy et al. 2009). The unique biology of myeloid leukemia in children with Down syndrome may also offer insight into the superior clinical response of these patients to cytarabine and anthracycline regimens, and parallel in vitro studies that demonstrated marked sensitivity of DS leukemia blasts to chemotherapy (Taub et al. 1999; Frost et al. 2000; Zwaan et al. 2002; Taub and Ge 2005). Leukemia cells from patients with Down syndrome also exhibit marked sensitivity to anthracyclines (2–7-fold), mitoxantrone (ninefold), amsacrine (16-fold), etoposide (20-fold), 6-thioguanine (threefold), busulfan (fivefold), vincristine (23-fold), and prednisolone (more than 1.1-fold), when compared to leukemia cells from patients without Down syndrome (Zwaan et al. 2002). Elegant studies from the laboratory of Taub and co-workers also demonstrate that AMKL blasts generate significantly higher levels of the active intracellular cytarabine metabolite, ara-CTP, when compared to blasts from AML patients without Down syndrome. This group has developed the hypothesis that mutated GATA1s reduces expression of cytidine deaminase (CDA), the enzyme that deaminates and inactivates cytarabine, rendering cells more sensitive to cytotoxicity. Supporting this notion, they have demonstrated significantly lower CDA expression in megakaryoblasts from patients with Down syndrome when compared to blasts from other AML patients. In addition, transfection of the full-length GATA1 cDNA resulted in significantly increased CDA expression and reduced cytarabine sensitivity in a DS AMKL cell line (Ge et al. 2005). Related investigations have linked alterations in single carbon metabolism with increased expression of chromosome 21q22.3 genes, including cystathionine-β-synthase (CBS). CBS catalyzes an intermediate step in the synthesis of cysteine, and increased activity in patients with Down syndrome is associated with significantly lower levels of homocysteine, methionine and S-adenosylmethionine, and potentially altered folate metabolism. These observations may also explain in part sensitivity to cytarabine as well as methotrexate (Taub et al. 1999; Taub and Ge 2005). Overexpression of chromosome 21-localized genes, including superoxide dismutase, may be linked to increased chemotherapy drug sensitivity. In TMD, immature megakaryoblasts accumulate in the liver predominantly and to a lesser extent in the bone marrow and peripheral blood; this results in leukocytosis, which can be severe. Transient myeloprolifera-tive disorder is most commonly a disorder of hepatic hematopoiesis and, thus, a prenatal presentation or in utero diagnosis is not uncommon. TMD exhibits a variable clinical presentation and course, and it is likely that most patients are asymptomatic and do not come to medical attention. However, a subset of patients may develop life-threatening leukocytosis, hydrops fetalis, liver dysfunction and bleeding diathesis, or pericardial and/or pleural effusions, as well as hepatosplenomegaly, lymphadenopathy or leukemia cutis. Approximately 70–80% of patients with TMD exhibit a spontaneous remission. Causes of death in TMD are largely associated with organ failure secondary to hyperviscosity affecting the heart, lungs or kidneys, bowel ischemia and necrotizing enterocolitis, or liver failure secondary to fibrosis. Liver abnormalities in patients with TMD can often present insidiously and manifest only with a persistently and progressively increasing conjugated bilirubinemia. Immunophenotypic evaluation of TMD blasts reflect largely myeloid and megakaryoblastic differentiation with expression of CD33, CD38, CD117, CD34, CD7, CD56, CD36, CD71, CD42b, and receptors for thrombopoietin and erythropoietin (Yumura-Yagi et al. 1992; Girodon et al. 2000; Karandikar et al. 2001; Langebrake et al. 2005). Leukemia-associated recurring chromosomal abnormalities are uncommon in TMD (Forestier et al. 2008). TMD generally resolves slowly over several months, and is most commonly managed with supportive treatment including low-dose chemotherapy, exchange transfusion or leukapheresis. Patients with severe leukocytosis or symptomatic disease, including those with hepatic dysfunction and evolving fibrosis, have been successfully managed with low-dose cytarabine. Several low-dose cytarabine regimens have been employed by the cooperative cancer groups and include: (1) POG 9481 cytarabine dosing of 10 mg/m2 per dose twice a day for 7 days (Massey et al. 2006), (2) COG 2971 dosing of 3.33 mg/kg/24 h as continuous intravenous infusion for 5 days (Gamis 2005), and the (3) AML-BFM 93, 98 and 04 series dosing of 0.5–1.5 mg/kg for 3–12 days (Klusmann et al. 2008). In most circumstances, there is a prompt response to cytarabine with eradication of peripheral blasts in the first 1–2 weeks of therapy. Mortality from symptomatic TMD is approximately 15–20%, and multivariate analysis has revealed severe leukocytosis (WBC > 109/L), preterm delivery, ascites, and bleeding diathesis as risk factors for poor outcome (Zipursky 2003; Klusmann et al. 2008; Muramatsu et al. 2008). Approximately 20% of patients with TMD will progress or recur with acute megakaryoblastic leukemia. After successful recovery from an episode of TMD, patients require close monitoring for residual complications, but also for the potential increased risk for the development of leukemia in the next 2–4 years of life. General recommendations include monthly hemogram with white cell differential and platelet count for the first year and, if normal, every 2 months for the subsequent years. AMKL in patients with Down syndrome has significantly different clinical features when compared with AML in other children, and in the CCG 2861/2891 study of 118 patients with Down syndrome and 1,088 patients without Down syndrome, some of the characteristic differences were: a younger median age of presentation (1.8 vs 7.5 years), lower presenting white blood cell counts (7.6 × 109/L vs 19.9 × 109/L), lower platelet count at diagnosis (29.0 × 103/μL vs 52.0 × 103/μL), a greater frequency of antecedent MDS (20% vs 8%), and an under-representation of characteristic t(8;21); t(15;17); 16q22 AML chromosomal translocations (Lange et al. 1998). Down syndrome AMKL represents a unique biology and is now recognized as a distinct diagnostic entity in the WHO classification of myeloid leukemias. Central nervous system involvement is also less frequent, and the bone marrow morphology can reveal accompanying varying levels of dysplasia and fibrosis with abnormalities in megakaryocytes (Gamis et al. 2003; Creutzig et al. 2005; Zeller et al. 2005; Rao et al. 2006). Paralleling the biology of AMKL, immunophenotypic characterization of blasts generally reflects the differentiation along pathways in early myeloid differentiation (expression of CD33, CD38, CD117, CD34, and CD7), megakaryocytic lineage markers (expression of CD42b and CD41), and erythroid lineage markers (expression of CD36 and Glycophorin A) (Yumura-Yagi et al. 1992; Girodon et al. 2000; Karandikar et al. 2001; Langebrake et al. 2005). Consistent with the unique biology of this disease, the characteristic translocations t(1;22) and t(1;3) that occur in other forms of AMKL are generally not observed in Down syndrome AMKL (Lange 2000; Forestier et al. 2008). Compared to other pediatric patients with AML, children with Down syndrome and AMKL exhibit significantly increased frequencies of dup(1q), del(6q), del(7p), dup(7q), +8, +11, del(16q), and +21 (Forestier et al. 2008). Recent clinical findings indicate that AMKL in younger children with Down syndrome often responds well to multi-agent chemotherapy regimens of modest intensity, with cure rates in excess of those observed in children with acute myeloid leukemia without Down syndrome. Ravindranath was the first to observe superior outcomes for patients with Down syndrome and AML who were treated with high-dose cytarabine on POG 8498 (Ravindranath et al. 1992). This finding paralleled biological investigations of cytarabine and anthracycline sensitivity in AMKL blasts from patients with Down syndrome (see above), and Down syndrome patients consistently exhibited relatively low relapse rates in multiple subsequent small and large-scale clinical studies in pediatric AML (Table 5.6). However, these studies also indicated that children with Down syndrome have an increased risk of side effects with chemotherapy, and high rates of regimen-related mortality may compromise overall outcome. This phenomenon was most clearly identified in the CCG 2891 study, which was a randomized trial for children with AML comparing intensive timing versus standard timing of the same five-drug cytarabine and anthracycline-based Induction chemotherapy regimen. The intensive timing Induction delivered the two Induction chemotherapy courses separated by a 6-day rest interval, regardless of marrow recovery, while in the standard timing Induction, the second course was administered after marrow recovery if the day-14 marrow had less than 5% blasts or on days 14–17 if significant residual leukemia was present. The CCG 2861 and 2891 studies enrolled 1,206 children with AML, including 118 children with Down syndrome and 1,088 children without Down syndrome, and the results revealed a marked excess of regimen-related mortality in the children with Down syndrome treated on the intensive timing arm (32%) when compared to children with Down syndrome treated with standard timing (2%), or compared with other children with AML but without Down syndrome (5% mortality in standard timing; 11% with intensive timing) (Lange et al. 1998). The optimal chemotherapy regimen for children with Down syndrome and AMKL remains under investigation, and most cooperative groups have developed separate strata within larger AML studies or independent clinical studies for patients with Down syndrome and AMKL. As indicated in Table 5.6, contemporary treatment regimens exhibit approximately 74–85% EFS for children with Down syndrome and AMKL. The intensity of therapy has been successfully significantly reduced in clinical regimens for children with Down syndrome and AMKL without compromising relapse rates. Table 5.6 Down syndrome-AML clinical studies: outcomes and chemotherapy drug exposure. (From Ravindranath et al. 1992; Creutzig et al. 1996; Ravindranath et al. 1996; Lange et al. 1998; Craze et al. 1999; Kojima et al. 2000; Creutzig et al. 2005; Abildgaard et al. 2006; Al-Ahmari et al. 2006; O’Brien et al. 2008) Regimen Number of patients Cytarabine dose (g/m2) Anthracycline dose (mg/m2) EFS (%) Treatment-related mortality (%) Refractory (%) Relapse (%) Japanese Consortium (Kojima et al. 2000) 33 4.2 300 80 9 0 9 HSC-Toronto (Al-Ahmari et al. 2006) 18 6.7 0 67 NR – 33 MRC AML 10 (Craze et al. 1999) 32 10.6 650 59 34 0 8 MRC AML 10 Modified (Craze et al. 1999) 13 ND ND 46 8 8 38 CCG 2861/2891 STD timing (Lange et al. 1998) 85 15.2 320 74 2 2 ND CCG 2861/2891 INT timing (Lange et al. 1998) 25 15.2 320 52 32 4 ND POG 9421 (O’Brien et al. 2008) 57 20.7 135/80 77 4 – 14 BFM-98 (Creutzig et al. 2005) 66 23–29 220–240 89 5 0 6 BFM-87/93 (Creutzig et al. 1996) 21 23–43 220–400 48 33 0 14 POG 8498 (Ravindranath et al. 1992) 12 40.7 230 100 0 0 0 POG 8821 (Ravindranath et al. 1996) 34 48.1 350 68 18 – 17 NOPHO-AML93 (Abildgaard et al. 2006) 41 49.6 150/30 85 5 5 7 NOPHO-AML88 (Abildgaard et al. 2006) 15 50.1 450 47 33 13 7 Abildgaard et al. (2006) Ann Hematol 85:275–280 Al-Ahmari et al. (2006) Br J Haematol 133:646–648 Craze et al. (1999) Arch Dis Child 81:32–37 Creutzig et al. (2005) Leukemia 19:1355–1360 Creutzig et al. (1996) Leukemia 10:1677–1686 Kojima et al. (2000) Leukemia 14:786–791 Lange et al. (1998) Blood 91:608–615 O’Brien et al. (2008) J Clin Oncol 26:414–420 Ravindranath et al. (1992) Blood 80:2210–2214 Ravindranath et al. (1996) N Engl J Med 334:1428–1434 In addition to infectious complications and mortality, children with Down syndrome appear to have increased risk for clinical cardiotoxicity. The POG 9421 study included 57 patients with Down syndrome and AMKL and treated patients with five cycles of chemotherapy, including daunorubicin 135 mg/m2 and mitoxantrone 80 mg/m2. Symptomatic cardiomyopathy was seen in ten patients (17.5%) relatively soon after completion of treatment, and three died as a result of congestive heart failure. It is postulated that increased dosage of genes located on chromosome 21, including carbonyl reductase and superoxide dismutase, predisposes to risk of anthracycline-related cardiomyopathy (O’Brien et al. 2008). AML in the older child and young adult with Down syndrome appears to have a somewhat more resistant biology, paralleling that seen in other children without Down syndrome, and current therapy recommendations include treatment on standard acute myeloid leukemia regimens. This may be reflected in part in recent observations that Down syndrome patients who had TMD and subsequently developed AMKL exhibited superior EFS and fewer relapses when compared with Down syndrome patients with AMKL and no previous history of TMD (91% vs 70%; log-rank p = 0.039) (Klusmann et al. 2008). The results of the CCG 2891 study indicated inferior outcomes with increasing age at diagnosis of Down syndrome-AML, particularly for children older than 4 years; this was mirrored in the CCG 2971 study (Gamis et al. 2003). These data continue to support the notions that tailoring therapy to the unique biology of Down syndrome-AML in the younger patient may hold great promise to preserve excellent relapse-free survival with markedly reduced toxicity. Current COG approaches treat older children (>4 years of age) with Down syndrome leukemia on the same pediatric AML regimens with other children without Down syndrome. About 5% of children with newly diagnosed AML still have ≥5% blasts in their bone marrow after two courses of combination chemotherapy, which is called resistant (or refractory) disease. Moreover, about 20% of patients conventionally defined as achieving CR have minimal residual leukemia (0.1–4% blasts) at the end of two cycles of Induction therapy. The prognosis of these children with resistant, MRD-positive or relapsed AML is relatively poor (Webb et al. 1999; Pui et al. 2004; Kaspers and Ravindranath 2005). The most frequent event for newly diagnosed patients with AML is a relapse, occurring in about one in three children. Bone marrow usually is involved, and the CNS in up to 10% of cases (including combined relapses). About 50% of patients suffer an early relapse, arbitrarily defined as within 1 year from initial diagnosis, and 50% suffer a late relapse, after 1 year. Outcome from relapse is significantly better in case of a longer duration of CR1 (Stahnke et al. 1998; Webb et al. 1999; Kaspers et al. 2009). However, the international study Relapsed AML 2001/01 showed early treatment response to be an even more important prognostic factor (Kaspers et al. 2009). In that study, early response was determined by morphological examination of the bone marrow obtained at day 28 from start of re-Induction, and defined as good in case of less than 20% AML blasts, and poor in case of 20% or more of such blasts. Day 28 in practice is between days 28 and 42 from the start of re-Induction therapy. In addition to treatment response, favorable cytogenetics t(8;21) and inv(16), as initially identified in studies with newly diagnosed children, also seem to be a favorable prognostic factor at relapsed AML (Webb et al. 1999; Kaspers et al. 2009). Study Relapsed AML 2001/01 and several other recently published studies report a probability of long-term survival from relapsed AML of more than 30%, as summarized by Goemans et al. (2008). Therefore, re-Induction chemotherapy should be offered to all children and adolescents with relapsed AML who can tolerate intensive treatment. However, patients that respond poorly to the first course of re-Induction chemotherapy, patients that do not achieve a second complete remission, and patients that relapse again could be offered more experimental therapy in that setting of a very dismal prognosis. Fortunately, innovative treatment seems achievable within 10–20 years from now. Whether or not treatment of a patient with relapsed AML with curative intention should be attempted may depend on many factors, such as age, duration of first remission, cytogenetic findings, the treatment modalities the patient received before relapse, and especially the physical and psychological condition of the patient and his/her parents. There are only a few studies on the treatment of relatively unselected groups of children with relapsed AML. High-dose chemotherapy with two or more of the agents cytarabine, mitoxantrone, etoposide, idarubicin, and fludarabine, can induce complete remission in 50–80% of the relapsed AML cases. Patients with AML that express CD33 can be treated with gemtuzumab ozogamicin alone or in combination with other agents, although responses in CD33-negative cases have been reported. Children who achieve subsequent CR should receive HSCT including autografts or allografts from related or unrelated HLA-matched donors. The survival rate from relapsed AML is about 20%, but is significantly higher in late relapses (Lowenberg et al. 1999; Pui et al. 2004; Kaspers and Ravindranath 2005). In the setting of the so-called “International Pediatric AML Group,” and coordinated by the AML committee of the International BFM Study Group, a large intergroup randomized study in pediatric relapsed and refractory AML is ongoing. For the first re-Induction course, the FLAG regimen is being compared with FLAG plus liposomal daunorubicin, followed by another course of FLAG. Stem cell transplantation is intended in all patients, either from a matched sibling donor (preferred) or an unrelated matched donor. Results are encouraging and survival is demonstrated in both early and late relapses. Moreover, the study proves the possibility to run an international study in pediatric AML with more than ten groups participating worldwide. Another group with unfavorable prognosis is the group of patients with minimal residual disease (MRD). These patients have persistent measurable AML blast cells after one to two courses of intensive chemotherapy (Coustan-Smith et al. 2003; Pui et al. 2004). MRD can be detected by polymerase chain reaction (PCR) amplification of antigen receptor genes and/or of fusion transcripts, and by flowcytometric detection of ectopic or aberrant immunophenotypes. Flowcytometry is applicable in the majority of patients with AML. There is a strong correlation between MRD levels detected during clinical remission and treatment outcome (Sievers et al. 2003; Campana and Coustan-Smith 2004). Monitoring of MRD will probably allow the identification of patients at high risk for a subsequent relapse and to adjust further treatment (Campana and Coustan-Smith 2004; Kaspers and Creutzig 2005). However, the prediction of a subgroup with an excellent prognosis based on MRD measurements still is difficult, probably because of the limited sensitivity of flow cytometry in comparison to PCR-based techniques (which unfortunately are useful in only a minority of AML patients). Significant improvements have already been achieved with conventional drugs that have been available for the treatment of AML for decades. Therefore, it seems realistic to expect major advances with the development of new drugs and especially novel treatment modalities. Discussion of all of these challenges is beyond the scope of this chapter, but they have been reviewed elsewhere (Kaspers and Zwaan 2007). In brief, innovative treatment will consist of at least four different modalities. First, the development of novel but conventional drugs. A good example might be clofarabine, a relatively novel nucleoside analogue that is being studied in both children and adults with AML. Second, the design of drugs with novel mechanisms of action, often targeted at leukemia-specific abnormalities. Typical examples are monoclonal-antibody mediated treatment such as with gemtuzumab ozogamicin (targeting CD33-positive cells) and with tyrosine kinase inhibitors (targeting FLT3- or kit-mutated AML cells). Third, improving the graft-versus-leukemia effect of donor cells without an increase in graft-versus-host-disease. And finally, the development of vaccination and other immunotherapy approaches. Innovative treatment will become more and more tailored and personalized, which will necessitate large-scale and thus international collaboration. Such collaborative efforts will be the only way to enroll a sufficient number of patients with a specific subtype of AML in a given study. Despite such an international setting, randomized studies in large cohorts of patients will more and more become impossible because of limited patient numbers eligible for subgroup-directed therapy. Thus, innovative trial designs will also be necessary. Fortunately, international large-scale collaboration has been realized in the past 10–20 years, especially in the setting of the International BFM Study Group, the so-called International Pediatric AML Group, and the International Consortium for Childhood APL. These collaborations together with the efforts of many professionals in pediatric AML will undoubtedly further improve the outcome of this disease in terms of cure, side-effects and late effects. Advances in cytogenetics and molecular biology have yielded significant insight into the biological heterogeneity of acute myeloid leukemia, and the response of different biological subgroups to specific chemotherapy approaches. These insights have led to the development of targeted therapies (e.g., all-trans-retinoic acid and arsenic trioxide for acute promyelocytic leukemia with the PML/RAR translocation), or modified conventional chemotherapy treatment regimens tailored for host and disease biology (e.g., Down syndrome acute megakaryoblastic leukemia and Fanconi anemia with AML). However, for the majority of patients with AML, chemotherapy dose-exposure appears to have reached a maximum level now limited by significant toxicity, and for many patients, treatment outcomes remain unsatisfactory. The core chemotherapeutic regimens were developed over 3 decades ago and are based largely upon anthracyclines and cytarabine. Significant incremental progress in improving outcomes has been based upon refinements in chemotherapy dose and scheduling and parallel advances in supportive care with reduction in the risk of complications. Primary recent advances in pediatric AML therapy include: 1. Molecularly targeted therapy for acute promyelocytic leukemia 2. Therapies tailored to the unique disease biology and host sensitivity of patients with Down syndrome and AML 3. Improved outcomes with dose exposure and intensity in Induction 4. Better characterization of patients who do and do not derive benefit from allogenic transplantation 5. Importance of cytogenetics and molecular classification to predict response 6. Early insights into the number of post-remission cycles of Intensification needed for curative approaches Molecularly targeted therapy for acute promyelocytic leukemia Therapies tailored to the unique disease biology and host sensitivity of patients with Down syndrome and AML Improved outcomes with dose exposure and intensity in Induction Better characterization of patients who do and do not derive benefit from allogenic transplantation Importance of cytogenetics and molecular classification to predict response Early insights into the number of post-remission cycles of Intensification needed for curative approaches

In low-income countries (LIC), the treatment of children with cancer is influenced by unique circumstances that restrict improvement in cure rates. The management of acute myeloid leukemia (AML) demands a multidisciplinary teamwork approach as the treatment intensity required to decrease the disease-associated mortality may actually contribute to increased mortality because of toxicity. Assuring rapid access to medical facilities, providing appropriate medications and blood products, developing a core of trained health care professionals and laboratory personnel, improving diagnostic methods, and balancing the intensity of treatment and the quality of supportive care are all necessary for improved treatment results. 6-Mercaptopurine 6-Thioguanine Cytarabine + daunorubicin + etoposide Central America Pediatric Hematology Oncology Association Cytarabine + idarubicin + etoposide Acute megakaryoblastic leukemia Acute myeloid leukemia Acute promyelocytic leukemia Acute promyelocytic leukemia differentiation syndrome Cytarabine Arsenic trioxide all-trans-retinoic acid The Berlin/Frankfurt/Muenster Study Group Bone marrow Central nervous system Low number of blasts in CSF (<5) Blasts in LCR ≥5 Complete remission Initial complete remission Cerebrospinal fluid Daunorubicin + cytarabine + 6-thioguanine Daunorubicin Down syndrome Event-free survival French-American-British Gruppo Italiano Malattie Ematologiche High-dose cytarabine + mitoxantrone High-income countries Hematopoietic stem cell transplantation Intensive care unit Idarubicin Low-income countries Hypogranular or microgranular variant form of M3 myeloid leukemia Mitoxantrone + cytarabine + etoposide Myeloid leukemia Myeloid leukemia of Down syndrome The Medical Research Council Mitoxantrone Nordic Society for Pediatric Hematology and Oncology Overall survival Peripheral blood Programa Español de Tratamientos en Hematologia Transient myeloproliferative disorder Treatment-related mortality Etoposide White blood cells World Health Organization

Despite major advances in our understanding of the biology of childhood acute myeloid leukemia (AML) and the development of new cytotoxic drugs, the prognosis of long-term survival is still only 60-65 %. In the present research, we studied the pharmacokinetics of drugs used in the induction therapy of childhood AML and performed in vitro drug sensitivity testing of leukemic cells from children with AML. The aims of the studies were to correlate the results of the analysis to biological and clinical parameters and to identify subgroups of AML with specific drug sensitivity profiles in order to better understand why treatment fails in some patients and how therapy may be improved. Blood samples were analysed to study the pharmacokinetics of doxorubicin (n=41), etoposide (n=45) and 6-thioguanine (n=50). Doxorubicin plasma concentration and total body clearance were correlated to the effect of induction therapy, and doxorubicin plasma concentration was an independent factor for complete remission, both in univariate and multivariate analysis including sex, age, and white blood cell count at diagnosis. For etoposide and 6-thioguanine no correlation was found between pharmacokinetics and clinical effect. Children with Down syndrome (DS) tended to reach higher blood concentrations of etoposide and thioguanine nucleotides, indicating that dose reduction may be reasonable to reach the same drug exposure as in children without DS. Leukemic cells from 201 children with newly diagnosed AML, 15 of whom had DS, were successfully analysed for in vitro drug sensitivity by the fluorometric microculture cytotoxicity assay (FMCA). We found that samples from children with DS were highly sensitive to most drugs used in AML treatment. In non-DS children, the t(9;11) samples were significantly more sensitive to cytarabine (p=0.03) and doxorubicin (p=0.035) than other samples. The findings might explain the very favorable outcome reported in children with DS and t(9;11)-positive AML. A specific drug resistance profile was found for several other genetic subgroups as well. A detailed study of MLL-rearranged leukemia showed that cellular drug sensitivity is correlated both to partner genes and cell lineage, findings that support the strategy of contemporary protocols to include high-dose cytarabine in the treatment of patients with MLL-rearrangement, both in AML and acute lymphoblastic leukemia (ALL). Our results indicate that drug resistance and pharmacokinetic studies may yield important information regarding drug response in different sub-groups of childhood AML, helping us to optimize future chemotherapy in childhood AML.

Background: During the last four decades the prognosis of childhood acute myeloid leukemia (AML) has been substantially improved due to an increase in complete remission (CR) rates, event-free survival (EFS) and reduced early mortality. The relapsed AML still remains a therapeutic challenge.

1 online resource (252 pages) This book has been created for parents who have decided to make education and research an integral part of the treatment process. Although it also gives information useful to doctors, caregivers and other health professionals, it tells parents where and how to look for information covering virtually all topics related to childhood acute myeloid leukemia (also Acute granulocytic leukemia; Acute myeloid leukemia; Acute nonlymphoblastic leukemia; Acute nonlymphocytic leukemia; child Acute myeloid leukemia; child Acute nonlymphocytic leukemia), from the essentials to the most advanced areas of res Print version record Introduction; PART I: THE ESSENTIALS; PART II: ADDITIONAL RESOURCES AND ADVANCED MATERIAL; PART III. APPENDICES; ONLINE GLOSSARIES; CHILDHOOD ACUTE MYELOID LEUKEMIA GLOSSARY; INDEX

Children aged 0–2 years (i.e. infants) with acute myeloid leukemia (AML) are a peculiar subgroup of patients in the childhood AML scenario. They present with distinctive biological and clinical characteristics, including a high prevalence of prognostically unfavorable risk factors and an increased susceptibility to therapy-related toxicity. Remarkable improvements have been achieved over the last 2 decades in the treatment of these patients and their outcome is becoming superimposable to that of the older age groups. In this review, we will focus on peculiarities of this young subgroup of children with AML, describing their clinical presentation, the biology of disease and factors influencing outcome. Treatment results and toxicity data reported by major collaborative groups are also summarized and compared.

Acute myeloid leukemia (AML) accounts for 15–20 % of the acute leukemias in children and 80 % of acute leukemia in adults [1, 2]. Significant advances as the result of multigroup cooperative trials have transformed this once uniformly fatal disease into one with a cure rate in children approaching 60 % [3], although the results are still significantly worse in adults. The difference in clinical outcome between childhood and adult AML is in part reflected by differences in the underlying biology. Recent advances in our understanding of stem cell physiology and the molecular basis of AML have elucidated the heterogeneity of AML and are leading to the identification of novel targets for therapy [4, 5]. The improvement in overall survival for children and young adults with AML can be attributed to better supportive care and advances in the application of chemotherapy and bone marrow transplantation. Future advancements in the treatment of children with AML will likely come about from the use of targeted therapies in conjunction with current chemotherapy regimens.

Objective: Leukemia is the most common pediatric malignancy and a major cause of morbidity and mortality in children. Among all subtypes, a lack of consensus exists regarding the diagnosis and treatment of acute myeloid leukemia (AML). Patient survival rates have remained modest for the past three decades in AML. Recently, targeted therapy has emerged as a promising treatment. Data Sources: We searched the PubMed database for recently published research papers on diagnostic development, target therapy, and other novel therapies of AML. Clinical trial information was obtained from ClinicalTrials.gov. For the major purpose of this review that is to outline the latest therapeutic development of AML, we only listed the ongoing clinical trials for reference. However, the published results of complete clinical trials were also mentioned. Study Selection: This article reviewed the latest developments related to the diagnosis and treatment of AML. In the first portion, we provided some novel insights on the molecular basis of AML, as well as provided an update on the classification of AML. In the second portion, we summarized the results of research on potential molecular therapeutic agents including monoclonal antibodies, tyrosine kinase/Fms-like tyrosine kinase 3 (FLT3) inhibitors, epigenetic/demethylating agents, and cellular therapeutic agents. We will also highlight ongoing research and clinical trials in pediatric AML. Results: We described clonal evolution and how it changes our view on leukemogenesis, treatment responses, and disease relapse. Pediatric-specific genomic mapping was discussed with a novel diagnostic method highlighted. In the later portion of this review, we summarized the researches on potential molecular therapeutic agents including monoclonal antibodies, tyrosine kinase/FLT3 inhibitors, epigenetic/demethylating agents, and cellular therapeutic agents. Conclusion: Gene sequencing techniques should set the basis for next-generation diagnostic methods of AML, and target therapy should be the focus of future clinical research in the exploration of therapeutic possibilities.

Acute myeloid leukemia (AML) accounts for only 15–20% of acute leukemias diagnosed in children. Pediatric and adult AML differ significantly in molecular characteristics and clinical features. These differences translate to overall better outcomes in children than adults. In the past 30 years, marked improvements in outcome have been achieved, especially for children with favorable AML subtypes. Progress is attributable to intensified treatment, optimized supportive care, and advances in stem cell transplantation. However, much more progress is needed to improve outcome and to decrease the burden of treatment for all patients.

We present the results of a retrospective study of acute myeloid leukemia (AML) treatment in Serbia in the period 2000–2012. Treatment was performed in four centers, two of which were located in Belgrade, one in Nis, and one in Novi Sad. Children affected by non-acute promyelocytic leukemia (non-APL) subtypes received treatment regimens derived from protocols of Berlin Frankfurt Munster AML study group, whereas children with APL were treated according to the AIEOP/GIMEMA ATRA plus Idarubicin “AIDA” protocol. Total number of patients was 106, out of whom 48 were girls (45.3 %) and 58 boys (54.7 %); median age at diagnosis was 9.0 years (range 1 month–17.9 years). In total, 82.1 % of patients achieved complete remission after induction treatment. Twelve patients (11.3 %) died during induction, before achieving complete remission; there were nine deaths during remission (10.5 %) and 20 patients relapsed (23.2 %). Median time of follow-up was 54 months. Two patients (1.9 %) were lost to follow-up. Event-free survival was 50.3 % and overall survival was 58.9 %.

AML is a heterogeneous and complex disease that is the culmination of the interaction between genetic and epigenetic alterations in the hematopoietic progenitors, leading to dysregulation of multiple critical signal transduction pathways, resulting in hematopoietic insufficiency due to the accumulation of immature myeloid progenitors. Despite identification of numerous cytogenetic, molecular and epigenetic alterations in AML, attempts at defining a unifying disease causing event in AML have failed. In contrast to chronic myeloid leukemia (CML), in which evolution of a single event (BCR-ABL translocation) is causally associated with disease pathogenesis, such a single step process does not appear to be responsible in AML pathogenesis. In vivo studies have demonstrated that common cytogenetic alterations such as t(8;21) or inv(16) that are highly associated with AML are not sufficient for the disease pathogenesis. These findings led to the hypothesis that evolution of AML may require multiple genetic changes and that the disease may require cooperation between two or more alterations. AML is a heterogeneous and complex disease that is the culmination of the interaction between genetic and epigenetic alterations in the hematopoietic progenitors, leading to dysregulation of multiple critical signal transduction pathways, resulting in hematopoietic insufficiency due to the accumulation of immature myeloid progenitors. Despite identification of numerous cytogenetic, molecular and epigenetic alterations in AML, attempts at defining a unifying disease causing event in AML have failed. In contrast to chronic myeloid leukemia (CML), in which evolution of a single event (BCR-ABL translocation) is causally associated with disease pathogenesis, such a single step process does not appear to be responsible in AML pathogenesis. In vivo studies have demonstrated that common cytogenetic alterations such as t(8;21) or inv(16) that are highly associated with AML are not sufficient for the disease pathogenesis. These findings led to the hypothesis that evolution of AML may require multiple genetic changes and that the disease may require cooperation between two or more alterations. One such model suggests that AML is the result of cooperation between two classes of genetic alterations, in which the initial event leads to maturation arrest (Dash and Gilliland 2001). The initial transforming events are genomic alterations such as core binding factor (CBF) translocations and mixed lineage leukemia (MLL) translocations that are referred to as type II mutations and are thought to arise in the hematopoietic stem/progenitor cells (Fig. 3.1a). Cells which are transformed by such a mechanism would have self renewal properties, but lack proliferative potential and represent a minor clone in the hematopoietic progenitor population, and may potentially persist as a minor clone for extended periods of time. Open image in new window Fig. 3.1 Acquisition of type II mutations such as core binding factor (CBF) translocations in the hematopoietic stem/progenitor population (a) leads to the transformation of the progenitor cell by causing impaired differentiation (b). This transformed clone with impaired differentiation capacity may persist as a minor population for years or decades until such time as it acquires a second “hit” that imparts proliferative advantage to the arrested clone, leading to explosive expansion of the clone, and the leukemic phenotype (c). Acquisition of additional genomic or epigenomic events in the leukemic cells may lead to clonal evolution within the leukemic population (d) Multiple studies have demonstrated that in teenage patients with t(8;21) AML, the fusion transcript was able to be detected in the patients’ neonatal Guthrie card blood spots, demonstrating intrauterine origin of the transforming event and long latency of evolution of AML (Greaves 1993; Mahmoud et al. 1995; Wiemels et al. 2002). The arrested clone would persist as a sub-clinical population until a time when an additional genetic event occurs, providing a proliferative advantage to the arrested clone. The proliferative change referred to as type I mutation includes activating mutations such FLT3/ITD and c-KIT mutations (Fig. 3.1b). Hematopoietic cells with concomitant type I and type II mutations would have both self-renewal capacity and a significant proliferative advantage (Fig. 3.1c), thus outcompeting other hematopoietic cells, resulting in AML phenotype and the resultant hematopoietic insufficiency. In vivo studies have provided support for this cooperativity model, in which transduction of individual mutations did not result in AML phenotype; however, simultaneous introduction of the two mutations led to rapid evolution of leukemia (Schessl et al. 2005). Recent technologic advances have allowed more complete interrogation of the AML transcriptome, genome and epigenome, leading to the identification of an increasing number of alterations associated with AML and suggesting that the process that leads to AML pathogenesis may prove to be even more complex. Data suggest that additional genomic and epigenomic events in leukemic cells may lead to further alteration of the leukemic clone and introduce genomic heterogeneity and clonal evolution within the leukemic cell population (Fig. 3.1d). Disease associated alterations in AML can be grouped into several broad categories as outlined below. Large structural alterations of the chromosomes identified in leukemic cells, including duplications, deletions and translocations, are identified by standard karyotype analysis in the majority of children with AML. Specific karyotypic abnormalities remain the most significant tools for diagnosis as well as for defining prognosis in AML. More recently, disease associated mutations of specific genes that mediate hematopoietic development have been identified that alter multiple signal transduction pathways and contribute to AML pathogenesis. Such mutations have come to define AML biology as they are prevalent, provide prognostic information and act as potential targets for directed therapy. Additional areas are emerging as major contributors to the pathogenesis of AML, including regulation of translation by non-coding RNA (miRNA) as well as epigenetic alterations due to aberrant DNA methylation and altered histone modifications. Nonrandom numerical and structural chromosome aberrations have been identified in approximately 80% of children with AML and are considered to be major contributors in an early stage of disease development. Karyotypic alterations of the leukemic blast are regarded as one of the most significant biological markers that define AML biology, and the presence of specific chromosomal abnormalities have significant prognostic implications. New WHO classification of AML has recognized this intimate association of cytogenetic alterations with diagnosis and prognosis in AML, by including the cytogenetic alterations as part of the classification schema of AML ( Chap. 5). Informative cytogenetics are generally available in >80% of children with AML and clonal abnormalities are demonstrated in 80% of those with informative cytogenetics (Kalwinsky et al. 1990; Raimondi et al. 1999). The number of chromosomal alterations identified in AML is in excess of 300, although >70% of pediatric AML cases fall into six specific cytogenetic categories, in which about 20% lack cytogenetic abnormalities, nearly 20% have CBF AML, 12% have t(15;17) and nearly 20% have rearrangements involving the MLL gene (Fig. 3.2) In addition to karyotypic alterations, disease associated mutations are identified in AML, with the highest prevalence in those with normal karyotype; >90% pediatric AML cases have at least one genomic alteration. Open image in new window Fig. 3.2 Cytogenetic abnormalities identified in childhood AML. Panel to the left demonstrates the most common karyotypic alterations identified in childhood AML; 80% of all children have disease associated genomic structural alterations. Mutation profile in those without cytogenetic abnormalities (normal karyotype) is shown in the right panel, where 76% of those in the normal karyotype population have one of the known mutations; thus >95% of all children with AML have at least one genomic alteration Acute promyelocytic leukemia (APL), which is due to the abnormal accumulation of immature granulocytes (promyelocytes), is defined by a chromosomal translocation involving the retinoic acid receptor alpha (RARα or RARA) gene and is unique from other forms of AML in its biology as well as responsiveness to targeted therapy with retinoids. Retinoids are derivatives of Vitamin A, which upon binding and stimulation of their receptors (RAR and RXR receptors), regulate embryogenesis, cell differentiation and apoptosis. In the absence of ligand, RAR-RXR dimers create a complex with transcriptional repressors that cause transcriptional repression through histone deacetylation (HDAC). Stimulation with retinoic acid (RA) causes conformational change of the complex, leading to the release of co-repressors and recruitment of co-activators, acetylation of histones associated with activation of transcription of target genes resulting in cellular differentiation. Translocations of chromosome 17 involving the RARA gene also usually involve the reciprocal translocation of PML zinc finger gene on chromosome 15. The t(15;17) translocation juxtaposes the PML and RARA gene and leads to the expression of a novel fusion product, which is a potent recruiter of transcriptional repressors that are resistant to de-repression by physiologic concentrations of RA. Leukemic cells with t(15;17) are arrested in the immature, promyelocyte stage. Identification of the underlying molecular biology of APL allowed for the utility of pharmacologic doses of RA to overcome the transcription repression induced by the PML/RARA fusion product, allowing for continued differentiation of the leukemic cells down granulocytic pathway; such findings led to APL being the first disease to be treated by a molecularly targeted therapy. With the incorporation of all trans retinoic acid (ATRA) into the therapeutic regimen, patients with APL have an excellent outcome with an overall survival of approximately 75–85% at 3 years; thus patients with APL are currently treated differently from those having other types of AML. Understanding the biologic basis of APL allowed for the development of a rational, targeted therapy for this subtype of AML. The resultant fusion transcript is also used for monitoring response to therapy and predicting relapse, as emergence of the fusion transcript can be detected in advance of morphologic relapse, allowing for initiation of treatment at the time of minimal disease, thus improving the survival for those with impending relapse. Recurrent chromosome translocations, inversions, and deletions result in structural genomic rearrangements often involving genes that encode transcription factors (Caligiuri et al. 1997; Look 1997). CBF is an alpha/beta heterodimeric transcription factor involved in the transcriptional regulation of several genes important in hematopoiesis. They are characterized by a DNA binding subunit (CBF-alpha; encoded by AML1 or RUNX-1 gene) and a non-DNA binding subunit (CBF-beta; encoded by CBFB). Disruption of the CBF subunit alpha and beta genes is involved in t(8;21)(q22;q22) and inv(16)(p13q22), respectively (Ito 1996; Mrozek et al. 1997). The RUNX1/AML1 gene, which is located on chromosome band 21q22, is disrupted in t(8;21) (q22;q22) translocation, which is present in nearly 12% of children with de novo AML. The CBF-beta subunit is encoded by CBF-beta on chromosome 16q22 and is disrupted by inv(16)(p13q22) and t(16;16)(p13;q22) rearrangements. The CBF-alpha subunit binds directly to the enhancer core DNA sequence of target genes, whereas the beta subunit, which lacks DNA binding capacity, increases the affinity and stabilizes the binding of the alpha subunit to the DNA. The precise mechanism through which the disruption of AML1 and CBF beta in the t(8;21)(q22;q22) and inv(16)(p13q22), respectively, contribute to leukemogenesis is unclear. Data from in vitro studies and transgenic animal models suggest a dominant negative role for the fusion genes created by these rearrangements that lead to the expression of a leukemic phenotype through a possible common pathway. This balanced translocation between chromosomes 8 and 21 results in the creation of a novel fusion transcript that juxtaposes the AML1 gene on chromosome 21 to the ETO gene on chromosome 8. The AML1 N-terminus contains the DNA binding motif, whereas the C-terminus contains a transcriptional activation domain (Ito 1996). The ETO gene on chromosome 8 encodes a protein that functions as a transcription factor. The functional domains of the protein include two zinc finger domains as well as a proline rich region (Miyoshi et al. 1993). The chimeric AML1/ETO fusion gene that is created as a result of t(8;21) translocation contains virtually the entire ETO coding region that is fused to the DNA binding domain of the AML1 gene minus the activation domain. Over-expression of this AML1/ETO chimeric protein, which retains the DNA binding function of the AML1 gene but without its activation function, acts as a dominant negative regulator of AML1 function, thus preventing trans-activation of CBF targets and contributing to leukemic transformation. Two distinct translocations, inv(16)(p13q22) and t(16;16)(p13;q22), collectively referred to as inv(16), result in the fusion of the CBF-beta gene from chromosome 16q22 with the MYH11 gene from chromosome 16p13. The MYH11 gene encodes a myosin heavy chain (van der Reijden et al. 1993). A number of variants of the CBFbeta/MYH11 fusion transcripts have been identified in those with inv(16) AML, with the most common variant detected in approximately 85% of patients with inv(16). It is not known whether the different fusion transcripts have any clinical or biological significance. The exact mechanism whereby the CBFbeta/MYH11 fusion gene contributes to malignant transformation is unclear; however, data from murine models suggest that CBFbeta/MYH11 interferes with normal hematopoiesis in a dominant-negative manner, resulting from the N-terminal portion of the CBF-beta fusing to the C-terminal domain of the MYH11. Such an interaction directly represses the AML1-mediated transcriptional activation by sequestering AML1 into a functionally inactive complex, resulting in a similar leukemogenic paradigm as in t(8;21). Cumulatively, CBF AML (inv(16) and t(8;21)) account for nearly a quarter of all childhood AML cases, and the prevalence of these translocations declines with age; in contrast, the prevalence is 10–15% in adults and <5% in AML patients older than 60 years. The presence of t(8;21) or inv(16) is associated with clinical outcome. Patients with either of these mutations appear to have a more rapid response to chemotherapy, a lower rate of relapse and improved survival compared to those without these translocations (Lange et al. 2008). Although the exact mechanism of responsiveness in this subtype is not clear, there are data that support the conclusion that patients with CBF AML benefit from intensification of cytarabine arabinoside as part of their Consolidation therapy. MLL is a large protein with multiple functional domains that is ubiquitously expressed in hematopoietic cells and is involved in the regulation of hematopoietic differentiation and proliferation. The MLL gene is located on chromosome 11, band 23; it is a methyltransferase that is involved in the regulation of Hox gene expression and methylation of histone residue H3K4. In vitro data have demonstrated that MLL fusion proteins can transform hematopoietic precursors to leukemia initiating cells, implicating this gene in stem cell maintenance and hematopoietic regulation. (Krivtsov et al. 2006; Krivtsov et al. 2009). Structural alterations involving the MLL gene are common in AML and appear to have an age dependent prevalence. There is a high prevalence of MLL rearrangements in AML patients less than 2 years of age (Chessells et al. 1995), and this prevalence decreases with increasing age (Balgobind et al. 2009); the prevalence of MLL rearrangements in adults with AML is <5% (Fig. 3.3a) (Grimwade et al. 1998; Balgobind et al. 2009). In MLL rearrangements, the N terminus of MLL can be juxtaposed to the C terminus of over 50 different partners (Fig. 3.3b), resulting in the loss of the methyltransferase domain. A major group of MLL fusion partners appears to interact with the methyltransferase that positively regulates transcription by methylation of specific histone lysine residues. Studies using gene expression profiling in MLL patients with AML, ALL or biphenotypic leukemia have demonstrated that those with MLL translocations have a unique gene expression profile regardless of the morphologic diagnosis of their disease, suggesting a common set of regulated pathways that could be therapeutically exploited. Open image in new window Fig. 3.3 (a) Prevalence of mixed lineage leukemia (MLL) translocations in different age groups. MLL is highly prevalent (27%) in children < 2 years of age. The prevalence decreases to 13% and 8% in those 2–10 and >10, respectively. Prevalence of MLL in adults is 3%. (b) Prevalence of most common MLL translocation partners In addition to the above mentioned alterations that are present in nearly 70% of all childhood AML, less common structural alterations are described with variable biology and prognostic significance. Notable examples are monosomy 7 (−7), monosomy 5 and deletion 5q. These alterations cumulatively account for 2–4% of cases of childhood AML, and >10% of adult AML. Although specific genes in these chromosomal regions that contribute to disease pathogenesis have not been identified, the presence of these alterations is associated with resistance to chemotherapy and extremely poor survival (Hasle et al. 2007). Gross cytogenetic abnormalities involving large chromosomal regions can be detected by conventional cytogenetics, and the frequency and implication of these changes have been described above. Karyotypic analysis in AML is limited by the ability to grow leukemic cells in vitro prior to analysis as well as the size of the region involved. Small structural alterations, including deletions or duplications, can escape detection by conventional cytogenetics and may require more specialized techniques such as fluorescence in situ hybridization (FISH). Loss of heterozygosity (LOH) that is generally associated with deletions leads to haplo-insufficiency that may contribute to malignant transformation. New technology using single nucleotide polymorphism (SNP) genotyping and copy number evaluation of the leukemic DNA can identify genomic alterations not amenable to detection by conventional means. The new techniques have identified regions of copy-neutral LOH (CN-LOH), also referred to as acquired uniparental disomy (aUPD) (Raghavan et al. 2005, 2008; Gupta et al. 2008). aUPD is a homologous recombination mediated process by which a homozygous state is achieved after an initial acquisition of a heterozygous mutation. In this DNA repair process, a mutant allele is used as a template, and the wild-type allele is converted to that of the mutant one, resulting in a homozygous state. Such an event leading to acquisition of LOH can alter the gene dosage of a particular mutation and may have significant clinical implications. aUPD has been described in specific populations of AML including FLT3/ITD and CEBPA mutations with varying prevalence in different studies (Radtke et al. 2009; Walter et al. 2009). Chromosomal translocations described above are considered to cause differentiation arrest. A broad body of data has demonstrated that such alterations are not sufficient for frank leukemic transformation, and additional events are required for the evolution of leukemic phenotype. Activating mutations of the kinase-mediated RAS/MAP signal transduction pathways have been demonstrated to provide the proliferative signal for the arrested clone and lead to leukemic progression. Regulation of cellular processes is initiated by extracellular signals (cytokines) whose actions are mediated through various signal transduction pathways, resulting in cellular differentiation, proliferation and survival. Stimulation of specific membrane bound receptors by their native ligand leads to the activation of their intracellular kinase moieties. The activated receptor recruits a number of proteins in the cytoplasm to form a com-plex of protein–protein interactions in the intracellular domain. SHC proteins, GRB2, GRB2-associated binder 2 (GAB2), SHIP, CBL, and CBLB (CBLB-related protein) are a few of the many adaptor proteins that interact with the activated receptor (Dosil et al. 1993; Zhang et al. 2000; Lavagna-Sevenier et al. 1998a, b). As each protein binds to the complex, it becomes activated in turn, resulting in a cascade of phosphorylation reactions that culminates in activation of a number of secondary mediators, including MAP kinase, STAT, and AKT/PI3 kinase signal transduction pathways. These activated mediators are chaperoned to the nuclear interphase by heat shock protein 90 (HSP90), in which the message is translocated to the nucleus. In the nucleus, these transcriptional mediators trigger a series of events culminating in regulation of cell differentiation, proliferation, apoptosis, and cell survival. As integrity of cellular processes are highly dependent on regulation of signal transduction pathways, alterations of pathway stimulation, either upstream (by receptor mutations) or downstream (by signal transducers), can alter the balance of proliferation, differentiation and apoptosis. Mutations in the above mentioned signaling cascades are common in AML and their prevalence and biologic significance is described below. FMS-like tyrosine kinase 3 (FLT3), a receptor tyrosine kinase (RTK), is a membrane-bound receptor with an intrinsic tyrosine kinase activity that is highly expressed in hematopoietic stem/progenitor cells and regulates hematopoietic differentiation and proliferation. FLT3 is expressed at variable levels in all myeloid malignancies. Mutations in the FLT3 gene have been identified in AML that lead to constitutive activation of the receptor kinase activity. Internal tandem duplication of the juxtamembrane domain coding sequence in FLT3 (FLT3/ITD) was initially identified in a high proportion of patients with AML (Nakao et al. 1996). Additional studies identified missense mutations in the activation loop domain of the tyrosine kinase moiety of the FLT3 receptor (FLT3 Activation loop mutation, FLT3/ALM) (Yamamoto et al. 2001). These mutations are the result of a segmental duplication of a fragment of the juxtamembrane domain coding region (encoded by exons 14 and 15) of FLT3 and are highly prevalent in hematologic malignancies, occurring in CML (5–10%), MDS (5–10%), and AML (15–35%) patients (Nakao et al. 1996; Stirewalt et al. 2001; Meshinchi et al. 2006; Kottaridis et al. 2001; Gale et al. 2007). Prevalence of FLT3/ITD is highly age dependent; it is rare in infant AML and the prevalence increases in a step-wise manner to 5–10% in age 5–10, 20% in young adults and >35% in AML patients older than 55 years (Meshinchi et al. 2006). There is also considerable variability in the size of the duplicated segment (ranging from 3 to >400 base pairs) and this variability may have biologic and clinical significance (Nakao et al. 1996; Stirewalt et al. 2001; Meshinchi et al. 2006). FLT3/ITDs lead to constitutive activation of the FLT3 receptor and promote cellular proliferation via the activation of multiple signaling pathways including RAS/MAPK, STAT and the AKT/PI3 kinase pathways. Studies have demonstrated that cytokine independent cellular proliferation of FLT3/ITD-transduced cells was mediated by RAS and STAT5 pathways (Mizuki et al. 2000). Demonstration of STAT5 activation in FLT3/ITD highlighted that some of the effects of FLT3/ITDs are unique to the mutated receptor, where in contrast to FLT3/ITD, ligand-induced FLT3-WT activation does not lead to STAT5 activation and no STAT5 DNA binding (Mizuki et al. 2000). Other studies have demonstrated that in addition to RAS and STAT pathways, FLT3/ITDs constitutively phosphorylate and activate AKT, and that ligand independent growth requires this aberrant AKT activation by FLT3/ITDs, such that the ligand independent growth is reversed if AKT activation is blocked. Presence of FLT3/ITD is associated with laboratory and clinical characteristics and outcome. Patients with FLT3/ITD have higher diagnostic white blood count, higher marrow blast percentage and more likely to have normal karyotypes. Furthermore, presence of FLT3/ITD is highly associated with clinical outcome, where those with FLT3/ITD have a higher relapse rate and poor outcome. Large studies have demonstrated the prognostic significance of FLT3/ITD in adult and pediatric AML trials (Thiede et al. 2002; Meshinchi et al. 2006). In these studies, survival of patients with FLT3/ITD was 20–30% compared to 50% for those without FLT3/ITD. In addition, it was demonstrated that the mutant to wild type allelic ratio observed in patients with FLT3/ITD appeared to influence outcome (Meshinchi et al. 2006; Thiede et al. 2002) Evaluation of various thresholds of FLT3/ITD allelic ratio (ITD-AR) established an allelic ratio threshold that demarcates patients with FLT3/ITD at high risk of relapse from those expected to have a more favorable outcome, demonstrating that it is not merely the presence of FLT3/ITD, but the level of mutant allele that determines clinical outcome (Meshinchi et al. 2006). Additional reports have demonstrated similar difference in clinical outcome for patients with AML having different mutant to wild type allelic ratios (Thiede et al. 2007). Currently, AML trials conducted by the Children’s Oncology Group (COG) utilize a FLT3/ITD allelic ratio of ≥0.4 to identify FLT3/ITD positive patients who are at high risk of relapse; these patients are referred for stem cell transplantation using the most suitable available donor. FLT3/ALM are missense point mutations within the activation loop of the FLT3 kinase domain that promote autonomous and constitutive activation of the receptor kinase, similar to that of FLT3/ITD (Yamamoto et al. 2001; Thiede et al. 2002). However, there are significant biological differences between the two types of FLT3 mutations (Choudhary et al. 2005; Grundler et al. 2005; Rocnik et al. 2006; Vempati et al. 2007), as FLT3/ITDs and FLT3/ALMs appear to promote activation of different downstream effectors (Choudhary et al. 2005) and different biological responses. Animal studies have demonstrated that mice harboring FLT3/ITDs primarily develop an oligoclonal myeloproliferative disorder, while mice harboring FLT3/ALM are more likely to develop oligoclonal lymphoid disorders (Grundler et al. 2005). In addition, despite their biologic similarity, patients who harbor the two different mutations appear to have drastically different clinical outcome; those with FLT3/ALM differ from those with FLT3/ITD as their outcome is not different from those without FLT3 mutations. Prevalence of FLT3/ALM is 6–8%, regardless of age, and the presence of FLT3/ALM and FLT3/ITD appears to be mutually exclusive (Meshinchi et al. 2006). These studies further demonstrated that the presence of FLT3/ALM is not associated with leukocytosis, and in contrast to FLT3/ITD, patients with FLT3/ALM AML do not have adverse outcomes. c-KIT is a proto-oncogene located on chromosome band 4q11–12 which, similar to FLT3, encodes a transmembrane protein that is a member of the type III receptor tyrosine kinase family (Paschka et al. 2006; Renneville et al. 2008). Ligand induced activation leads to c-KIT dimerization and transphosphorylation, thereby activating downstream signaling pathways (Renneville et al. 2008). Preferential activation of alternative pathways leading to cytokine independent growth and increased drug resistance have been reported to be linked to specific c-KIT mutations (Ning et al. 2001a, b, c). Mutations of c-KIT are infrequent in adult AML (2–8%) and tend to cluster within the activation loop (exon 17) and a region of the extracellular domain integral to receptor dimerization (exon 8) (Renneville et al. 2008). Their prevalence is higher, however, in CBF AML patients (6–48%) and may be associated with worse clinical outcome in adults (Care et al. 2003; Cairoli et al. 2005; Nanri et al. 2005; Schnittger et al. 2005; Paschka et al. 2006). Although adult data suggested that the presence of KIT mutations are associated with worse clinical outcome, a study conducted in pediatric cooperative trials demonstrated a prevalence of c-KIT mutations of nearly 20% but lack of clinical significance in pediatric CBF AML (Pollard et al. 2010). FMS and PDGFR are other members of class III RTK receptors that function in hematopoietic development. Activating point mutations in the kinase domains of FMS have been described in 5–10% of selected AML cases in adults (Ridge et al. 1990; Padua et al. 1998), although mutations in FMS were not identified in pediatric patients with AML (Meshinchi et al. 2003). Although expression and translocations involving PDGFR have been demonstrated in hematopoietic disorders, mutations in this gene have not been reported in AML (Yamada et al. 2006). Mutations in N-RAS gene occur in approximately 10–30% of adult AML cases (Radich et al. 1990; Stirewalt et al. 2001) and 10% of pediatric AML (Berman et al. 2009). Point mutations frequently occur in codons 12, 13, and 61 (Byrne and Marshall 1998), and prevent the hydrolysis of RAS-GTP, keeping RAS in the “on” or activated state. The coincident occurrence of both a FLT3 and RAS mutation in the same case is very rare, as might be expected, since there may be no a priori selective advantage to two mutations affecting the same pathway. However unlike FLT3-ITD mutations, RAS mutations are not prognostically important (Berman et al. 2009). With both FLT3 and RAS mutations, there may be discordance between diagnostic and relapse specimens. Thus, a patient with either a RAS or FLT3 mutation at diagnosis may relapse without the mutation (or vice versa). This suggests that these mutations are not the primary oncogenic events. NPM1 encodes a ubiquitously expressed and highly conserved phosphoprotein that is primarily localized to the nucleolus, but shuttles rapidly between the nucleus and cytoplasm in its role as a molecular chaperone. The NPM1 protein has been shown to contribute to many basic cellular processes, including biosynthesis of ribosomes, prevention of protein aggregation in the nucleolus, and regulation of centrosome duplication through cyclin E/cyclin-dependent kinase 2 phosphorylation. Abnormalities in NPM1 have been implicated in the pathogenesis of various hematopoietic malignancies, including anaplastic large cell lymphoma with t(2;5) translocation, resulting in expression of the NPM1-ALK fusion protein. Mutations in NPM1 are common in AML, with a prevalence of 30% in adult and 8–10% in pediatric AML (Dohner et al. 2005; Brown et al. 2007; Gale et al. 2007; Hollink et al. 2007). NPM1 mutations appear to be more prevalent in AML with normal karyotype, with a prevalence of nearly 40–50% in adults and 20% in pediatric AML. Disease associated mutations, characterized by four base insertions in exon 12 of the NPM gene, lead to impaired nuclear localization of the nucleophosmin protein. Presence of NPM mutations portends favorable outcome with reduced relapse risk and improved survival; those with NPM mutations have an outcome similar to those with CBF AML. There is significant overlap between NPM mutations and FLT3/ITD, as nearly one third of those with NPM mutations also have FLT3/ITD. In adult AML, it has been shown that FLT3/ITD retains its high-risk status in the presence of NPM mutations. In pediatric AML, the presence of NPM1 mutations in the presence of FLT3/ITD has been reported to partially ameliorate the poor prognosis conferred by FLT3/ITD alone (Hollink et al. 2007). CCAAT/enhancer binding protein-alpha (C/EBPa) is encoded by the CEBPA gene and functions as a transcription factor that regulates proliferation and terminal granulocytic differentiation. It is a member of the basic region leucine zipper class of transcription factors, and is expressed in a variety of tissues, including adipocytes, hepatocytes, and Type II alveolar cells. In the hematopoietic system, C/EBPa is expressed in cells of myelomonocytic origin and regulates the expression of genes that regulate granulocyte differentiation. CEBPa has been shown to inhibit E2F pathways, downregulate C-MYC and thus allow myeloid differentiation. Mutations in the CEBPA occur in 4% of childhood AML cases and the overwhelming majority of cases with these mutations occur in a bi-allelic manner, in which two distinct mutations occur, one in the N-terminal domain (NTD) and the second in the opposite allele affecting the bZip domain (Ho et al. 2008). NTD mutation results in an early termination codon, which truncates the p42 protein, allowing intact translation of only the p30 protein. In a bi-allelic model, the bZip mutation in the opposite allele disrupts the bZip domains of its protein products, resulting in loss of function. Such a bi-allelic mutation would result in the expression of p30 as the only functional protein. Kirstetter et al. demonstrated that sole expression of p30 may be sufficient for the development of AML; in their mouse model, expression of the p30 isoform in the absence of p42 resulted in a fully-penetrant AML phenotype (Kirstetter et al. 2008). This model would argue that both types of mutations are required for AML pathogenesis. COG studies have demonstrated that CEBPA mutations tend to occur in patients with normal cytogenetics and are associated with decreased relapse risk as well as improved survival compared to patients without a mutation (Ho et al. 2009). The WT1 gene, located on chromosome 11p13, encodes a zinc-finger protein that exists in multiple isoforms and functions as a transcription factor. WT1 is expressed primarily in tissues of the developing genitourinary and hematopoietic systems, and mutations in WT1 occur in both syndrome-associated and sporadic cases of nephroblastoma (Wilms Tumor), the most common childhood renal malignancy. WT1 is also expressed in CD34+ hematopoietic progenitors, and is over-expressed in a subset of acute leukemias, including AML, ALL, and CML in blast crisis (Bergmann et al. 1997). Somatic mutations of WT1 gene have been reported to have variable prognostic significance (Paschka et al. 2007). Initial studies in adults suggested that the presence of WT1 mutations was an independent prognostic factor for poor outcome. However, recent studies, including a comprehensive evaluation of pediatric AML, have identified WT mutations in 8–10% of patients; further, it was demonstrated that a strong association between WT1 mutations and FLT3/ITD existed. In that study, although the presence of WT1 mutation was associated with worse outcome, it was shown that the prognostic significance of WT1 mutation was due to FLT3/ITD; however, in patients without FLT3/ITD, the presence of a WT1 mutation did not portend poor outcome (Pollard et al. 2008). In contrast, Hollink et al. have reported that WT1 mutations represent an independent poor risk factor (Hollink et al. 2009). Furthermore, although expression of WT1 has been used as a marker of disease response and emergence of MRD (Bergmann et al. 1997; Lapillonne et al. 2006), no significant association of WT1 expression was observed in a pediatric trial study (Noronha et al. 2009). Comprehensive evaluation of WT1 expression within specific risk groups has not been performed in childhood AML and currently is not utilized in risk group stratification. Disease associated genomic alterations in TET2 gene (4q24) have recently been demonstrated in myeloid malignancies in adults. The prevalence of these mutations appears to be 20% in MDS, 10–20% in myeloproliferative disease and 12% in adult AML (Delhommeau et al. 2009; Tefferi et al. 2009). Although the prevalence of TET2 mutations is not well defined in childhood AML, early data from COG suggest that these mutations are uncommon, with a prevalence of 3% and unclear prognostic significance (Fig. 3.4). Open image in new window Fig. 3.4 Prevalence of disease-associated mutations in pediatric AML Recent advances in whole genome sequencing have enabled interrogation of the entire leukemic genome for identification of novel disease associated mutations. Initial evaluation of a patient with AML, identified numerous previously undescribed mutations; however, screening of an additional 200 patients failed to demonstrate any of these mutations in a larger cohort (Ley et al. 2008). This study highlights the need to separate the mutations that promote the leukemic process from “passenger” mutations that do not contribute to AML initiation or progression. Whole exome sequencing and whole transcriptome sequencing allow sequencing of the entire exonic and the transcribed sequences, respectively, allowing additional means of evaluating disease profile. Application of these new technologies to AML will enable identification of genomic alterations that would be impractical to identify on a gene by gene basis. As described above, AML is characterized by genomic alterations that are crucial to disease pathogenesis. An increasing body of evidence demonstrates that epigenetic modifications may contribute to the pathogenesis of leukemia. The term epigenetics refers to heritable changes in phenotype or gene expression caused by mechanisms other than alterations in the genome (DNA). Epigenetic changes that underlie the development of leukemia can be in one of two major categories: changes in the DNA methylation state and alterations in histone modification patterns. These two processes work in concert to lead to the activation or repression of gene expression. Emerging evidence shows that altered methylation and changes in histone modifications may be involved in normal and malignant hematopoiesis and contribute to myeloid leukemia and outcome (Chen et al. 2010). Evidence linking alterations in DNA methylation and malignancies has prompted comprehensive evaluations of DNA methylation patterns in AML in an attempt to identify specific methylation “signatures” that can be used for risk identification as well as identification of potential targets for epigenetically directed therapies. One study in adult AML used DNA methylation patterns to group patients into 16 different cohorts based on their methylation pattern (Figueroa et al. 2010). The methylation pattern segregated patients with specific mutations and cytogenetic groups. This study also identified a 15 gene methylation classifier predictive of overall survival. The contribution of alterations in DNA methylation patterning and its clinical significance is not yet well defined in childhood AML, but such evaluations are currently underway. Histone acetylation is associated with transcriptionally active chromatin (euchromatin) and is catalyzed by histone acetyltransferases (HATs). Several transcriptional co-activators, including CREB binding protein (CBP), have been shown to possess HAT activity. Conversely, transcriptional co-repressor complexes such as nuclear co-repressor 1 (NCOR1) have been shown to contain subunits with histone deacetylase (HDAC) activity. As acetylation and methylation status of the histone lysine residues help to determine whether chromatin is in the transcriptionally active or repressed state, altered histone modification patterns could mediate malignant transformation by altering expression of oncogenes. As alterations in methylation and acetylation status have significant contributions to myeloid transformation and proliferation, they are attractive targets for epigenetic therapy by DNA methyltransferase (DNMT), as well as histone deacetylase (HDAC) or methylase inhibitors. MicroRNAs (miRNAs) are evolutionarily conserved non-coding RNAs that contain 19–25 nucleotides and arise by cleavage from 70 to 100 nucleotide hairpin precursors (Bartel 2004, 2009; Bartel and Chen 2004; Chen et al. 2010). They hybridize to complementary messenger RNA (mRNA) targets and most commonly function by inhibiting translation of specific mRNAs. miRNAs have provided a new dimension to our understanding of complex gene regulatory networks. miRNAs can function as oncogenes or as tumor suppressor genes in leukemias, although the data for the latter are more limited. Specific miRNAs have been shown to be highly expressed in MLL leukemias (Mi et al. 2007), and miRNA expression signatures have been defined for AML cases with specific mutations or translocations. Common oncogene and tumor suppressor mutations also appear to be associated with unique miRNA signatures. For example, miR155 is up-regulated in cases with FLT3-ITD mutations (Garzon and Croce 2008). miRNA deregulation, specifically the expression pattern of 12 miRNAs, has also been associated with clinical outcome (Marcucci et al. 2008). As miRNAs can contribute to leukemogenesis by functioning as oncogenes or tumor suppressor genes, miRNAs also have the potential to be therapeutic targets or tools. They can be targeted by modulating expression level of the miRNA, or using miRNA-based cancer gene therapy to target multiple gene networks that are controlled by a single, aberrantly expressed miRNA. Increased understanding of epigenetic regulation of myeloid leukemogenesis may provide new understanding of AML biology and identify new tools for more accurate diagnosis, risk stratification and treatment. Acute myeloid leukemia is an enormously complex disease characterized by a myriad of genomic and epigenomic alterations that interact to create a wide variety of different AML subtypes having significantly different outcomes. Knowledge of the genomic alterations, whether karyotypic abnormalities or specific disease associated mutations, provides a great deal of information for clinical decision making. Patients with CBF AML (t(8;21) and inv(16)) are expected to have a favorable outcome and these patients can be spared more intensive post-Induction therapy such as hematopoietic cell transplantation. Given the growing body of evidence on the prognostic significance of CEBPA and NPM mutations, patients with these mutations are considered low-risk and are treated similar to those with CBF AML. In contrast, patients with monosomy 7 or with monosomy 5/del5q are considered to have an extremely high risk of relapse and although these alterations are rare in childhood AML, there have been no improvements in the outcome of patients with these alterations in the last decade. More complete understanding of the underlying mechanisms and the genes linked to disease pathogenesis is needed. Patients with FLT3/ITD are at high risk of relapse. As FLT3 mutation causes constitutive activation of the FLT3 receptor kinase, specific kinase inhibitors are being clinically tested. Patients with high mutant FLT3/ITD to normal allelic ratios also have high risk disease and are thus usually considered for stem cell transplantation in first complete remission. Although the presence of c-KIT mutations have not been definitively shown to be associated with adverse outcome in children, such mutant receptors may be inhibited by kinase inhibitors; these patients would be candidates for directed therapy with specific kinase inhibitors. Despite the increasing breadth of information about the genetic and epigenetic alterations in AML, a specific risk status can be identified for only 30–40% of the patients and the ability to assign risk status or specific therapy for the remaining cohort, with a survival rate of near 50%, is lacking. With the increase in the number of tools for interrogation of AML genomic and epigenomic changes, the ability to allocate patients to individualized therapy directed by the underlying biology of their disease is anticipated.

Microbial infection is common in patients with leukemia. However, aspergillosis is a rare complication of oral region. Attention to early diagnosis and multidisciplinary treatment can reduce the wide range of complications of the infection in these patients. This article reports 4-year history of a child with leukemia, who suffered from a rare aspergillosis infection in the palatal region of oral cavity during chemotherapy. Early diagnosis and treatment of bacterial and fungal infections, especially rare types such as aspergillosis, in patients with leukemia is important for reducing widespread complications. However, prosthetic treatment and attention to psychological complications of orofacial defects would also improve their quality of life significantly. These improvements can be seen more clearly in children.

Tristan Knight,1,2 Holly Edwards,3,4 Jeffrey W Taub,1–2,4 Yubin Ge2–41Division of Pediatric Hematology and Oncology, Department of Pediatrics, Children’s Hospital of Michigan, Detroit, MI, USA; 2Department of Pediatrics, Wayne State University School of Medicine, Detroit, MI, USA; 3Department of Oncology, Wayne State University School of Medicine, Detroit, MI, USA; 4Molecular Therapeutics Program, Karmanos Cancer Institute, Wayne State University School of Medicine, Detroit, MI, USAAbstract: Venetoclax (ABT-199), a BH3-mimetic and selective BCL-2 inhibitor, was recently approved by the US Food and Drug Administration (FDA) for the treatment of acute myeloid leukemia (AML) in adult patients aged 75 years or older, or otherwise unable to tolerate intensive induction chemotherapy, in combination with either hypomethylating agents or low-dose cytarabine. In this review article, we discuss venetoclax’s mechanism of action, in relation to both the BCL-2 protein family in general and BH3-mimetic activity in particular. We then outline the pharmacological advances that preceded and facilitated its development, as well as providing an overview of key preclinical and clinical studies which lead to its use first in chronic lymphoid leukemia (CLL), then in small lymphocytic leukemia (SLL), and subsequently in AML. Finally, we seek to offer an overview of the challenges and opportunities encountered as venetoclax moves into more widespread use, including its use and activity against leukemia initiating cells and oxidative phosphorylation.Keywords: ABT-199, acute myeloid leukemia, apoptosis, BCL-2 inhibitor, BH3-mimetic, venetoclax

The objectives of this review were to discuss standard and investigational treatment strategies for adolescent and young adult with acute myeloid leukemia, excluding acute promyelocytic leukemia. Acute myeloid leukemia (AML) in adolescent and young adult patients (AYAs) may need a different type of therapy than those currently used in children and older patients. As soon as AML is diagnosed, AYA patient should be offered to participate in well-designed clinical trials. The standard treatment approach for AYAs with AML is remission induction chemotherapy with an anthracycline/cytarabine combination, followed by either consolidation chemotherapy or stem cell transplantation, depending on the ability of the patient to tolerate intensive treatment and cytogenetic features. Presently, continuing progress of novel drugs targeting specific pathways in acute leukemia may bring AML treatment into a new era.

To describe symptoms, diagnostic features, treatments, and outcomes of pneumatosis intestinalis (PI) in pediatric patients being treated for acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML). This retrospective chart review included 514 patients ≤18 years of age diagnosed with ALL or AML between January 1999 and December 2007. PI episodes were identified by radiology report reviews. Twenty patients with ALL and four patients with AML presented 31 PI episodes. The median time between diagnoses of leukemia and PI was 1.0 month (interquartile range 0.8–6.4 months). Plain radiographs diagnosed all PI episodes. Computerized tomography (CT) and ultrasound were performed in 6 and 13 episodes, respectively. All CT and three ultrasounds demonstrated PI. Thirty episodes occurred exclusively in the colon, most commonly in the ascending (n = 26) and transverse (n = 18) segments. Treatment included complete bowel rest in 27 (87.1%) and intravenous broad-spectrum antibiotics in 29 (93.5%). One case required colectomy. Two episodes were untreated. There were no fatalities associated with PI. PI is uncommon in children with ALL or AML. Ultrasound is less sensitive than plain radiographs for diagnosis. PI occurred almost exclusively in the colon. With conservative management, most patients had excellent outcome.

This article is from Indian Journal of Critical Care Medicine : Peer-reviewed, Official Publication of Indian Society of Critical Care Medicine, volume 16. Abstract Severe complications that develop in the early stages in patients with acute leukemia have a mortal course. Bleeding, leukostasis, and less frequently, infections are responsible for early mortality. Hemorrhage is most common in acute leukemia and usually leads to death. Hemorrhage may occur due to chemotherapy or bone marrow transplantation in patients with acute leukemia. Leukocytosis, thrombocytopenia, sepsis, and coagulopathy increase the risk of bleeding. There may be multiple etiologic factors. Subdural or subarachnoid hemorrhage is less common than an intra-axial hemorrhage. The incidence of spontaneous subdural hematoma is higher in patients with leukemia. Although advances in the treatment of platelet transfusion and disseminated intravascular coagulation have decreased the incidence of hemorrhagic complications in patients receiving chemotherapy for acute leukemia, intracranial hemorrhage-related deaths are a significant problem. We discussed the etiology and management of chronic subdural hematoma detected in a two-year-old male patient with Acute Myeloid Leukemia and hyperleukocytosis.

The outcome for children with acute myeloid leukemia (AML) has improved significantly over the past 30 years, with complete remission and overall survival rates exceeding 90 and 60%, respectively, in recent clinical trials. However, these improvements have not been achieved by the introduction of new agents. Instead, intensification of standard chemotherapy, more precise risk classification, improvements in supportive care, and the use of minimal residual disease to monitor response to therapy have all contributed to this success. Nevertheless, novel therapies are needed, as the cure rates for many subtypes of childhood AML remain unacceptably low. Here, we briefly review advances in our understanding of the biology and genetics of AML, the results of recent clinical trials, and current recommendations for the treatment of children with AML.

Acute myelogenous leukemia; Acute nonlymphocytic leukemia; Agranular CD4+ CD56+ hematodermic neoplasm; Agranular CD4+ NK cell leukemia; Blastic NK cell lymphoma; Blastic NK leukemia/lymphoma Acute Myeloid Leukemia (AML) is a bone marrow cancer, characterized by the clonal expansion and lack of differentiation of early myeloid cells. In the majority of cases, the diagnosis of AML requires a myeloid blast count ≥20% of nucleated cells in the blood or bone marrow. Blasts can also infiltrate other organs or tissues such as skin, gums, or lymph nodes. Rarely AML can present as an isolated myeloid sarcoma in any tissue. The definition of myeloid blasts includes abnormal promyelocytes in acute promyelocytic leukemia (APL), and promonocytes in monocytic/monoblastic AML. Almost any myeloid lineage can be affected; myeloid or monocytic are relatively common, but rare basophilic, pure erythroid and megakaryocytic forms of AML exist. AML is an extremely heterogeneous disease, with many distinct subgroups defined by specific morphological, cytogenetic, molecular, and gene expression characteristics. The World Health Organisation (WHO) has classified AML according to some of the most significant clinical, morphological, cytogenetic, and molecular pathological factors (ref) (Table 1). Other subgroups include AML with myelodysplasia-related changes, therapy-related myeloid neoplasms, myeloid sarcoma, and myeloid proliferations related to Down Syndrome. The WHO classification has largely replaced the earlier French American British (FAB) classification, which was principally based on morphological characteristics. A morphological classification does remain, in the AML not otherwise specified (AML NOS) WHO subgroup, which is used if no other defining features are present. Table 1 WHO classification of AML and related neoplasms WHO subtype Morphology Immunophenotype Molecular/Other AML with recurrent genetic abnormalities AML with t(8;21)(q22;q22); RUNX1-RUNX1T1 Blasts >5%. Large blasts with granulation and auer rods. Most commonly AML with maturation (see below), often with dysplastic features in maturing cells. Occasionally absence of maturation or rarely monocytic differentiation CD34+, CD13+, CD33+, HLA-DR+, MPO+ Subpopulation with CD15+ (maturing granulocytes) Often aberrant CD19+, more rarely CD56+ Additional chromosomal abnormalities in 70%. KRAS/NRAS mutations in 30% pediatric cases KIT mutations in 20–25% 10 year survival 61% AML with inv(16)(p13;q22) or t(16;16)(p13;q22); CBFB-MYH11 Blasts >5%. Morphological features of myelomonocytic leukemia (see below). Abnormal eosinophil component in the BM (large purple-red granules in immature forms). Rarely lack eosinophilia or show myeloid or monoctyic differentiation only Multiple CD34+, CD117+ blast populations Granulocytic: CD13+, CD33+, CD15+, MPO+ Monocytic: CD14+, CD11b+, CD11c+, CD64+, CD36+ Often aberrant expression CD2 Inv(16) more common that t(16;16) Additional chromosomal abnormalities in 40%, trisomy 22 fairly specific KIT mutations in 30% 10 year survival 55% APL with PML-RARA Atypical promyelocytes >5%. Hypergranular variant with prominent red-purple granules (rare in PB). Some cells contain bundles of auer rods (faggot cells). Microgranular variant with more readily visible dumbbell or bilobed nucleus and high PB count CD13+, CD33+, MPO+, HLA-DR−, CD34− Aberrant expression CD2 in microgranular variant Additional chromosomal abnormalities in 40%, with trisomy 8 the most common Mutations in FLT3 ITD or TKD in 35–45% 10 year survival 73% AML with t(9;11)(p21;q23); MLLT3-KMT2A Blasts ≥20%. Morphological features most often monocytic or myelomonocytic (see below). Occasionally AML with or without maturation In children: CD33+, CD65+, CD4+, HLA-DR+ CD13−/+, CD34−/+, CD14 −/+ In adults: markers of monocytic differentiation – CD14+, CD4+, CD11b+, CD11c+, CD64+, CD36+. Variable expression CD34, CD117, CD56 Secondary cytogenetic abnormalities are frequent, with trisomy 8 the most common 10 year survival 39% AML with t(6;9)(p23;q34); DEK-NUP214 Blasts ≥20%. Most commonly AML with maturation or acute myelomonocytic leukemia (see below). Can have morphological features of any FAB subtype other than APL or acute megakaryoblastic leukemia. Marrow and PB basophilia in approximately 50% cases MPO+, CD13+, CD33+, HLA-DR+ CD117±, CD34±, CD15± Tdt + in 50% The t(6;9) translocation is the sole chromosomal abnormality in most cases. The fusion protein acts as an aberrant transcription factor FLT3-ITD very common 69%–78% cases 10 year survival 26% AML with inv(3)(q21;q26) or t(3;3)(q21;q26); GATA2, MECOM Blasts ≥20%. Most commonly AML without maturation, acute myelomonocytic leukemia or acute megakaryoblastic leukemia. Can have morphological features of any FAB subtype other than APL. PB may have dysplastic neutrophils, giant and hypogranular platelets. Multilineage dysplasia is common in the BM CD13+, CD33+, HLA-DR+, CD34+, CD38 + Some aberrant CD7 expression. Some with megakaryocytic marker expression (CD41 and CD61) Repositioning distal GATA2 enhancer. MECOM activation and haploinsufficiency GATA2 10 year survival 3%. AML (megakaryoblastic) with t(1;22)(p13;q13) Blasts ≥20%. Morphologically acute megakaryoblastic leukemia. Fibrosis usually present. Occurs exclusively in infant and children <3 years Megakaryocyte markers: CD41+ and/or CD61+. CD36 + Often CD34−, MPO− and HLA-DR− The t(1;22) translocation is the sole chromosomal abnormality in most cases. The fusion gene may modulate chromatin organization Provisional: AML with BCR-ABL1 Blasts ≥20%. May present as de novo disease or result from blast crisis of CML CD13+, CD33+, MPO+ Preliminary data suggest deletion of IGH, TCR, IKZF1 or CDKN2A may support a diagnosis of de novo disease vs. CML blast phase Unfavorable prognosis AML with mutated NPM1 Blasts ≥20%. Most commonly myelomonocytic or monocytic (80–90% of acute monocytic leukemias have an NPM1 mutation). NPM1 mutations are also seen in AML with and without maturation. Some have multilineage dysplasia. WCC usually high CD13+, CD33+, MPO+, CD34− Monocytic markers common: CD14+, CD11b+, CD68+ Abnormal karyotype in only 5–15%. NPM1 mutations usually mutually exclusive of balanced translocations that define AML entities FLT3-ITD in 40%. Favorable prognosis when found in isolation AML with biallelic mutations of CEBPA Blasts ≥20%. No distinctive morphological features, but most commonly AML with or without maturation. Less commonly monocytic or myelomonocytic CD13+, CD33+, CD65+, CD11b+, CD15+ HLA-DR ± and CD34± Aberrant CD7 in over 50% Abnormal karyotype in 30% FLT3-ITD in 22–33% cases Favorable prognosis. Provisional: AML with mutated RUNX1 Blasts ≥20% CD13+, CD33+, MPO+ Cases with MDS-related cytogenetic abnormalities excluded from this category Unfavorable prognosis AML with myelodysplasia-related changes AML with myelodysplasia-related changes Blasts ≥20%. Most cases have morphological evidence multilineage dysplasia. Morphological classification requires dysplasia in ≥50% cells in ≥2 lineages Other defining criteria are MDS-related cytogenetic abnormalities or a preceding diagnosis of MDS Variable findings CD34+, Tdt+, CD13+, CD33+ Aberrant CD7 and CD56 common Chromosome abnormalities are similar to those in MDS (Table 6) 10 year survival 16% Therapy-related myeloid neoplasms Therapy-related AML/MDS and AML/MDS/MPN Blasts ≥20% to diagnose AML. Often preceded by myelodysplastic phase. Multilineage dysplasia present in the majority. Distinguishing therapy-related AML, MDS and MDS/MPN may not be clinically relevant. Best considered as one clinical syndrome No specific immunophenotypic findings CD34+, CD33+, CD13+ Aberrant CD7 and CD56 common. Abnormal karyotype in >90%. Unbalanced aberrations in 70%, mainly loss of material from chromosome 5 and 7, associated with myelodysplastic phase (longer latency) In 20–30% balanced translocations are found and most often present as de novo AML (shorter latency) AML, not otherwise specified AML with minimal differentiation ≥20% blasts. No evidence of myeloid differentiation by light microscopy (absence of granulation and auer rods) CD34+, HLA-DR+, CD38+, CD13+ MPO- CD117±, CD33± Aberrant CD7 in 40% Complex karyotype most common abnormality. Unbalanced MDS-related abnormalities and RUNX1 mutations result in assignment to alternative categories AML without maturation ≥20% myeloid blasts. Blasts ≥90% of non-erythroid cells MPO+, CD13+, CD33+, CD117+ CD34+, HLA-DR+ Aberrant CD7 in 30% Aberrant CD2, CD4, CD19 or CD56 in 10–20% No association with specific recurrent chromosomal abnormalities AML with maturation ≥20% myeloid blasts ± azurophilic granules; ≥10% maturing cells of neutrophil lineage. Cells of monocytic lineage ≤20% of BM cells CD13+, CD33+, CD15+ CD34±, CD117±, HLA-DR± Aberrant CD7 in 20–30% No association with specific recurrent chromosomal abnormalities Acute myelomonocytic leukemia ≥20% myeloblasts, monoblasts, promonocytes; ≥20% monocytes and monocytic precursors; ≥20% neutrophils and granulocytic precursors Multiple blast populations CD13+, CD33+ Immature: CD34+, CD117+ Granulocytic: CD15+ Monocytic: CD14+, CD4+, CD11b+, CD11c + CD64+, CD36+ Aberrant CD7+ 30% Myeloid-associated, nonspecific cytogenetic abnormalities common e.g., trisomy 8 Acute monoblastic/monoctyic leukemia ≥80% monocytic cells; monoblastic if ≥80% are monoblasts or monocytic if <80% are monoblasts HLA-DR+, CD13+, CD33+, CD15+ ≥2 of CD14+, CD11b+, CD11c+, CD64+, CD36+ MPO−/+ more common in monocytic Aberrant CD7+/CD56+ 25–40% Myeloid associated, nonspecific cytogenetic abnormalities are present in the majority of cases Pure erythroid leukemia >80% immature erythroid precursors with ≥30% proerythroblasts CD117±, CD36+, CD235a (glycophorin A) −/+, MPO−, HLA-DR−, CD34−, CD41/61− No specific chromosomal abnormality, complex karyotype most common Unfavorable prognosis Acute megakaryoblastic leukemia ≥20% blasts with ≥50% shown to be of megakaryocyte lineage by immunophenotype CD36+, CD61+, CD41+, CD42 −/+ HLA-DR±, CD34±, CD45−/+, MPO- 3q26 abnormalities and t(1;22) result in assignment to alternative category Unfavorable prognosis Acute basophilic leukemia ≥20% blasts of basophilic lineage. Medium sized cells with oval to bilobed nucleus, nucleoli, basophilic cytoplasm and coarse basophilic granules CD13+, CD33+, CD123+, CD203+,CD11b, CD34±, HLA-DR±, CD117− Aberrant CD9+ in most No specific chromosomal abnormality Acute panmyelosis with myelofibrosis Pan myeloid proliferation with clusters of blasts. Dysplastic megakaryocytes. Marked reticulin fibrosis. Trephine diagnosis Blasts MPO−, CD13+, CD117+ Multilineage involvement – erythroid, megakaryocytic, granulocytic markers present Karyotype usually abnormal Myeloid proliferations related to Down syndrome Transient abnormal myelopoiesis (TAM) Unique disorder of Down syndrome newborns. Morphological features of acute megakaryoblastic leukemia. PB basophilia. Spontaneous remission in the majority. 20–30% develop nontransient AML 1–3 years later CD34+, CD56+, CD117+, CD13+, CD33+, CD7+, CD41+, CD42+, CD36+, CD61+ MPO−, CD15−, CD14− Trisomy 21 required and additionally acquired mutations in GATA1 are present Myeloid leukemia associated with Down syndrome Most commonly acute megakaryoblastic leukemia. No biological difference between MDS and AML therefore the term myeloid leukemia includes both morphological cases of MDS and AML CD34±, CD56±, CD117+, CD13+, CD33+, CD7+, CD41±, CD42+, CD36+, CD61+ MPO−, CD15−, CD14− Trisomy 21 required and additionally acquired mutations in GATA1 are present. Trisomy 8 common AML results in rapidly progressive bone marrow failure, hence the term “acute.” The replacement of normal hematopoiesis by proliferating myeloid blasts causes clinically significant, and often rapidly progressive, anemia, thrombocytopenia, and neutropenia. The resulting symptoms include fatigue, shortness of breath, bleeding, bruising, petechiae, pyrexia, and symptoms of infection. Disseminated intravascular coagulation can also be a presenting feature, most commonly in APL, and untreated can cause rapidly fatal bleeding. Central nervous system involvement at presentation occurs in approximately 0.5% of cases and can cause headaches and neurological symptoms or signs. The white cell count (WCC) is typically elevated, and if >150 × 109/L, symptoms of leucostasis which include headache, respiratory distress, and visual disturbance may be present. This is more common if the high WCC is comprised of monoblasts. Up to 10% of cases of AML may have an “aleukemic” presentation, with cytopenias and possible dysplastic features, but no blasts present in the peripheral blood. The majority of cases of AML occur with no known predisposing factors. However, in 10–15% of cases, prior treatment with chemotherapy or radiotherapy is implicated, and as increasing numbers of people survive their initial cancer, this is predicted to expand. AML can evolve from a prior acquired hematological condition, most often the myelodysplastic syndromes, but also myeloproliferative neoplasms. Rare environmental predisposing factors include exposure to benzene and ionizing radiation. Inherited conditions such as Down syndrome have long been known to greatly increase the risk of development of AML, particularly acute megakaryoblastic anemia. It was previously thought that the contribution of germline mutations to the development of AML was restricted to rare disorders such as Dyskeratosis Congenita and Fanconi anemia, or Bloom and Li-Fraumeni syndrome, which are characterized by defective DNA repair. However, it is increasingly recognized that up to 10% of people who develop AML have an inherited predisposition to hematological malignancies, and myeloid neoplasms with germline predisposition is a new entity in the most recent WHO classification (Table 2). The importance of a careful family history in patients with AML, and the involvement of geneticists to screen for potential germline alterations that may affect family members, is increasing. Incidence Table 2 Myeloid neoplasms with germline predisposition Myeloid neoplasm with germline predisposition without a pre-existing disorder or organ dysfunction Acute myeloid leukemia with germline CEBPA mutation Myeloid neoplasms with germline DDX41 mutation Myeloid neoplasms with germline predisposition and pre-existing platelet disorders Myeloid neoplasms with germline RUNX1 mutation Myeloid neoplasms with germline ANKRD26 mutation Myeloid neoplasms with germline ETV6 mutation Myeloid neoplasms with germline predisposition and other organ dysfunction Myeloid neoplasms with germline GATA2 mutation Myeloid neoplasms associated with bone marrow failure syndromes Myeloid neoplasms associated with telomere biology disorders Myeloid neoplasms associated with Noonan syndrome Myeloid neoplasms associated with Down syndrome Incidence The overall incidence of AML in the general population is 3/100,000. AML accounts for 90% of cases of acute leukemia in adults, but is rare in children, in whom acute lymphoblastic leukemia (ALL) predominates. Age Age The incidence of AML rises with age, from 1/100,000 at age 25 to over 15/100,000 by the age of 75. Due to the aging population, as well as increased survival post chemotherapy or radiotherapy for other cancers, the overall incidence of AML will rise over the coming decades. Sex Sex The sex ratio is 1:1 in younger people presenting with AML but there is a slight male preponderance in the over 65 s (ratio 1.3:1). The reason for this remains unclear. Treatment Treatment The backbone of treatment for AML remains cytotoxic chemotherapy; however, as the molecular basis of this disease is becoming more understood, novel targeted therapies are starting to improve survival in this disease. The mainstay of treatment for younger, fitter patients with AML is intensive induction chemotherapy consisting of an anthracycline (most often daunorubicin) and cytarabine (ELN guidelines 2017), followed by post-remission consolidation therapy. Molecularly targeted therapy, such as the FLT3 inhibitor midostaurin, or the anti-CD33 monoclonal antibody-drug conjugate gemtuzumab is now becoming increasingly used. Assessment of risk, based on genetic features, and the monitoring of minimal residual disease are critical to informing decisions regarding post remission treatment with additional cycles of intensive chemotherapy or high dose therapy followed by allogeneic transplant. The treatment of APL is significantly different to other forms of AML. Due to the coagulopathy that accompanies this form of AML, it used to be the subtype with the highest mortality. Since the 1980s, the PML-RARA fusion due to the t(15,17)(q22;q12) has been targetable directly with the vitamin A derivative all-trans retinoic acid (ATRA), which overcomes the differentiation block caused by this oncogene. ATRA may be used in combination with an anthracycline cytotoxic chemotherapy drug, resulting in complete remission rates of >90%. More recently, trials have demonstrated that an entirely chemotherapy-free treatment using ATRA with arsenic trioxide may be even more effective than the chemotherapy-containing regimes. Outcome Outcome Survival in AML is influenced primarily by age and the genetic abnormalities present. Overall, more than half of patients with AML will die from this disease. Increasing age adversely affects outcome for a number of reasons. These include inability to deliver intensive chemotherapy, higher treatment-related mortality, and higher rates of chemotherapy resistance. AML that is secondary to a prior hematological disorder such as MDS, or is therapy-related, have an adverse prognosis. The important role of genetics in risk stratification is highlighted by the European LeukemiaNet (ELN) guidelines, which have categorized genetic abnormalities into favorable, intermediate, and adverse (Table 3). Table 3 ELN 2017 risk stratification of AML by genetics Risk category Genetic abnormality Favorable t(8;21)(q22;22); RUNX1-RUNX1T1 inv(16)(p13q22) or t(16;16)(p13;q22); CBFB-MYH11 Mutated NPM1 without FLT3-ITD or with FLT3-ITDlow( Bialleleic mutated CEBPA Intermediate Mutated NPM1 and FLT3- ITDhigh Wild type NPM1 without FLT3-ITD or with FLT3-ITDlow (w/o adverse risk lesions) t(9;11)(p21;q23); MLLT3-KMT2A Cytogenetic abnormalities not classified as favorable or adverse Adverse t(6;9)(p23;q34); DEK-NUP214 t(v;11q23); KMT2A rearranged t(9;22)(q34;q11); BCR-ABL1 inv(3)(q21;q26) or t(3;3)(q21;q26); GATA2,MECOM(EVI1) −5 or del(5q); −7; −17/abn(17p) Complex karyotype, monosomal karyotype Wild type NPM1 and FLT3- ITDhigh Mutated RUNX1, Mutated ASXL1 Mutated TP53 Outcomes vary considerably between the groups: for example patients with APL, due to t(15,17) have a 73% 10-year survival, compared to a 3% 10 year survival in those with disruption of the gene MECOM due to alterations in chromosome 3. However, as these genetic alterations become therapeutic targets, their prognostic significance may change. The laboratory diagnosis of AML involves morphological and histological assessment of peripheral blood and bone marrow, in addition to molecular and cytogenetic analysis. Bone marrow is aspirated and biopsied most commonly from the posterior iliac crest. Liquid bone marrow can be spread immediately on slides to make a smear, which ideally contains particles at the end of trails. The slides are stained with Wright Giemsa or MGG (May-Grunwald Giemsa) stains. While examination of the bone marrow trephine biopsy is not essential to establish the diagnosis of AML in most instances, and should be avoided in severely coagulopathic patients due to the increased bleeding risk, it may contribute valuable information. At diagnosis, where there is significant marrow fibrosis, hypocellular samples or other reasons for an inadequate aspirate, the trephine is needed for diagnosis. The diagnosis of acute panmyelosis with fibrosis can only be made from a trephine, or increased fibrosis in addition to other histological features may suggest a preceeding myeloproliferative neoplasm. For follow-up samples post-chemotherapy, histological examination of the bone marrow can often best inform the treatment response and hence remission status. Ideally the bone marrow core should be 1.5–2.0 cm in length. In the case of a dry or aparticulate aspirate, the trephine can be rolled on a slide to produce a trephine roll, which can give a more rapid indication of an elevated blast count. If a poor liquid sample is obtained, then a trephine biopsy may be taken into saline, and disaggregated for immunophenotyping by flow cytometry. The DNA obtained from trephine samples taken into saline is usually better preserved than when DNA is extracted from formalin-fixed trephine biopsies. This may be used for molecular analysis if no liquid bone marrow is available, and where there are low numbers of blasts in the peripheral blood. The WHO classification has moved on from using morphology as the primary means for classification of AML. However, morphology is critical in identifying and enumerating blasts and in the identification of specific subgroups of AML. A myeloid blast count ≥20% in the peripheral blood or bone marrow is required for the diagnosis of AML, The exception to the requirement for >20% blasts is the presence of cytogenetic translocations t(8,21), inv(16), t(16,16), or t(15,17) which are categorized as AML provided >5% blasts are present. Blast morphology is best appreciated in the bone marrow aspirate sample and morphological features are detailed in Table 4 (Fig. 1a). Table 4 Blast morphology Myeloblasts Medium to large sized cells (14–18 μm) with a single round or oval nucleus Fine chromatin with one to several distinct nucleoli Minimal to moderate amount of basophilic cytoplasm Myeloid blasts are usually granular (minimally differentiated AMLs being one exception) Some contain Auer rods (fused granules in a rod-like form), which are pathognomic of AML Monoblasts Larger than myeloblasts (15–20 μm) Single round nucleus and visible nucleoli Voluminous blue-grey cytoplasm Often vacuolated Promonocytes Blast equivalent Appearances similar to monoblasts Lobulated nucleus with less prominent nucleoli and more prominent vacuolation Proerythroblasts Medium to round large cells with nucleoli Minimal deeply basophilic cytoplasm Cases with ≥80% immature erythroid precursors and ≥ 30% proerythroblasts are classified as pure erythroid leukemia under AML, NOS. Megakaryoblasts Medium to large cells (12–18 μm) Usually lack granules Cytoplasmic blebbing Open image in new window Fig. 1 Morphological features of AML (a) myeloblast in aspirate sample (b) the marrow is hypercellular with expansion of immature hemopoietic precursors expressing CD34 (c). (d) Faggot cell in a case of APL with an accumulation of Auer rods in its cytoplasm giving an appearance of a “bundle of sticks” (e) H&E appearances of APL in trephine biopsy. (f) Blast and dysplastic eosinophils in AML with inv(16)(p13q22) Immunocytochemistry is now rarely performed due to the universal use of flow cytometry. Historically, cytochemical stains such as Sudan black were used to confirm myeloid lineage. A toludine blue stain can be used to help confirm the diagnosis of the very rare acute basophilic leukemia, if this is suspected. In most cases, the bone marrow trephine biopsy is hypercellular with expanded myelopoiesis and virtual maturation arrest (Fig. 1b). Most of the cellularity is accounted for by CD34+ immature hemopoietic precursors with a high nuclear to cytoplasmic ratio (Fig. 1c). The other two lineages might be difficult to appreciate in the routine hematoxylin and eosin (H&E) stained sections and necessitate the use of immunohistochemistry, particularly when there is a suspicion of AML with myelodysplasia-related changes. Careful examination of bone marrow trephine sample is particularly useful in cases of hypocellular AML and ones with significant fibrosis as bone marrow aspirate in both instances may be nondiagnostic. APL is the most important subtype of AML to identify morphologically, as appropriate supportive care and ATRA must be started as soon as possible to reduce early mortality. A careful examination of the peripheral blood is essential, as APL often presents as a pancytopenia, and hence there may be limited numbers of leukocytes to assess. The atypical promyelocytes in the more common hypergranular variant are often packed with large cherry-red granules, and the nucleus is often bilobed. These are often called “dumbell,” “buttock,” or “butterfly” cells. Auer rods can be a prominent feature, and the term “faggot cell” is used to describe the pathognomonic promyelocytes containing large numbers of Auer rods that resemble bundles of sticks (Fig. 1d). The microgranular variant can be less easy to identify morphologically as the granules are submicroscopic. The clue to the diagnosis lies in the typical nuclear morphology, as in the hypergranular variant. Auer rods and more typical granular forms are often present in small numbers. The bone marrow trephine is hypercellular due to myeloid expansion (Fig. 1e). There is an accumulation of promyelocytes with abundant, often granular, cytoplasm and convoluted/bean shaped nucleli. AML with t(8,21)(q22;q22) typically has the morphological appearances of AML with maturation. Auer rods are more often present than in other subtypes of non-APL AML, these are usually present singly within myeloid blasts. In AML with inv(16)(p13q22) or t(16,16)(p13;q22), the peripheral blood often contains increased numbers of monocytes and promonocytes, with an often marked increase in eosinophils in the bone marrow (Fig. 1f). Dysplastic features are often seen in AML; however, dysplasia must be present in >50% of cells in ≥2 lineages to be classified morphologically as AML with myelodysplasia related changes (Fig. 2). In the bone marrow trephine, dyserythropoiesis is characterized by maturation synchrony within and dys-synchrony between erythroid islands. Dispersal of immature erythroid precursors and para-trabecular translocation are also often seen. Dysmegakaryopoiesis often manifests as decreased cell size and simplification of nuclear features. Forms are often small/hypolobated or micromegakaryocytes (Fig. 2). They might display dispersal of nuclear lobes. Open image in new window Fig. 2 AML with MDS-related changes. Blast are identifiable in the aspirate sample (a) as well as tissue sections (b) stained with CD34 (c). Dysplastic megakaryocytes show nuclear simplification and abnormal separation of nuclear lobes as seen in the aspirate (a), bone marrow trephine, H&E (b), and CD61 (d) stained sections This diagnosis can also be made if AML develops from a prior myelodysplastic syndrome, or in the presence of MDS-related cytogenetic abnormalities (Table 5). AML with mutated NPM1 or biallelic mutation of CEPBA are excluded from this category, even if significant dysplasia is present. They are categorized according to these mutations and the relatively favorable prognosis relating to them, compared to the adverse prognosis of AML with MDS-related changes. Table 5 Cytogenetic abnormalities sufficient to diagnose AML with MDS-related changes Cytogenetic abnormalites Complex karyotype (3 or more abnormalities) Unbalanced abnormalities Balanced abnormalities −7/del(7q) t(11;16)(q23;p13) del(5q)/t(5q) t(3;21)(q26;q22) i(17q)/t(17p) t(1;3)(p36;q21) −13/del(13q) t(2;11)(p21;q23) del(11q) t(5;12)(q32;p13) del(12p)/t(12p) t(5;7)(q32;q11) idic(X)(q13) t(5;17(q32;p13) t(5;10)(q32;q21) t(3;5)(q25;q35) Therapy-related myeloid neoplasms occur as a result of prior exposure to chemotherapy and/or radiotherapy. The most commonly implicated cytotoxic drugs are alkylating agents and topoisomerase II inhibitors. Alkylating agent and radiation induced AML usually occurs 5–10 years post therapy and is characterized by the MDS-related cytogenetic changes listed in Table 5. In contrast, topoisomerase II inhibitor induced AML has a shorter latency (1–5 years) and is characterized by balanced translocations, most commonly involving KMT2A. Therapy-related AML is most commonly morphologically myelomonocytic, with dysplastic features often present. This subgroup is reserved for cases without a defining genetic abnormality, dysplasia, or prior therapy (Fig. 3). The morphological definitions of each subtype are outlined in Table 4. Open image in new window Fig. 3 Acute myelomonocytic leukemia, NOS with (a) blast equivalent promonocytes in aspirate sample and an expansion of myeloid cells (b), some with indented nuclei and also expressing CD68/PGM1 (c) in the bone marrow Morphological examination is performed to assess response to chemotherapy, disease dynamics, and remission status. Depending on the timing of the sampling, the bone marrow is often hypocellular and displays variable ratios of hematopoietic cells lineages. The erythroid lineage is the first to recover post chemotherapy, followed by myeloid and megakaryocytic lineages. Dysplastic changes might be apparent, either due to the chemotherapy or persistence of an underlying MDS in AML with MDS-related changes. Assessment of blast percentage is often difficult in hypocellular cases due to suboptimal aspirate and flow cytometry samples. Immunohistochemistry is of particular importance in these cases. Often CD34 and CD117 are appropriate blast markers, but attention must be paid to the immunophenotype of the blasts at diagnosis, as either one or both of these markers may be negative at diagnosis. CD117 also highlights erythroid precursors which may be increased as the bone marrow recovers from chemotherapy. Repeat sampling may be required to distinguish between regenerating marrow and persistent disease, particularly where there is no leukemia-associated immunophenotype or genetic abnormalities, and the assessment of response is based on morphology alone. The marrow stroma is often edematous with mild increase in reticulin fibers and increased macrophage activity. Immunophenotyping using multi-color flow cytometry is central to establishing the diagnosis of AML. Standard panels of antigens that are informative in AML have been published in the ELN guideline 2017. These are detailed in Table 6. Different subtypes of AML have characteristic immunophenotypic profiles which can give an early indication of the disease category, although genetic assessment is always required. APL, for example, is typically CD9 positive, CD34, and HLA-DR negative, while AML with t(8,21) often aberrantly expresses the lymphoid antigen CD19. Flow cytometry will identify acute leukemias of ambiguous lineage. These may be acute undifferentiated leukemia, without adequate lineage-associated antigen expression to assign the leukemia to any lineage or mixed lymphoid/myeloid acute leukemia. The criteria for diagnosis of these rare mixed phenotype leukemias are stringent, as it is relatively common for AML blasts to aberrantly express lymphoid antigens. Table 6 Expression of cell surface and cytoplasmic markers for the diagnosis of AML Diagnosis of acute myeloid leukemia Markers of immaturity CD34, CD117, CD33, CD13, HLA-DR Granulocytic markers CD65, cytoplasmic myeloperoxidase (MPO) Monocytic markers CD14, CD36, CD64 Megakaryocytic markers CD41 (glycoprotein IIb/IIa), CD61 (glycoprotein IIIa) Erythroid markers CD234a (glycophorin A), CD36 The identification of cytogenetic and molecular abnormalities is now central to the classification of AML. The recurrent molecular features of interest can be broadly categorized into abnormalities visible at the chromosomal level and those visible only at the molecular genetic level. Genetic testing should be performed if full diagnostic, prognostic, and therapy-guiding information is required, which may not be the case in situations where patients are not fit for intensive, remission-induction chemotherapy. The exception is APL, which should almost always be confirmed genetically. Overall 60% of cases of AML have an abnormal karyotype. In children and younger adults, balanced translocations predominate, whereas in older adults unbalanced translocations as well as chromosomal losses and gains are more common. Over 100 balanced translocations have been recurrently associated with AML, although only 8 chromosomal abnormalities and their variants are at present included in the WHO classification. The majority of these translocations result in the formation of chimeric fusion proteins, which disrupt transcription factors involved in hematopoiesis. Chromosomal abnormalities can be visualized using Giemsa stained metaphase preparations to determine the karyotype. Cytogenetic analysis usually requires the examination of at least 20 metaphase cells and is best performed on a bone marrow aspirate sample taken into cytogenetic culture medium. The limitations of karyotyping include failure to obtain results due to metaphases failing to grow in culture, or failure to detect abnormalities due to the presence of small or cryptic chromosomal rearrangements (seen in up to 10% of those with canonical chimeric fusion genes). It is relatively labor intensive and takes a minimum of several days to obtain results. Fluorescence in situ hybridization (FISH) is used where rapid results are required (for example in suspected APL), or in cases where karyotyping fails. Specific FISH probes in routine use can detect PML-RARA, RUNX1-RUNX1T1, CBFB-MYH11, KMT2A fusions, BCR-ABL1, abnormalities of 3q26, complex karyotypes, and loss of TP53. Usually 100–200 interphase cells are assessed, although metaphase FISH can be used if interphase FISH is equivocal. The α and β subunits of the core binding factor (CBF) complex are affected by the t(8,21)(q22;22) and inv(16)(p13q22)/t(16,16)(p13;q22) translocations, which lead to the formation of chimeric fusion proteins RUNX1 (AML1/CBFα2)-RUNX1T1 (ETO) and CBFβ-MYH11, respectively. These translocations occur in approximately 13% of AML cases in younger adults and are associated with a more favorable prognosis. The initiating event in the vast majority of cases of APL is a chromosomal translocation juxtaposing PML (promyelocytic leukemia gene) with RARα (retinoic acid receptor alpha). The PML-RARα fusion protein causes enhanced recruitment of repressive complexes to chromatin. The PML-RARα complexes bind to many normal and novel RARα target genes and repress their expression by deacetylation and DNA/histone methylation. This enhanced repression is integral to the pathogenesis of APL and is reversed by treatment with ATRA (All-Trans Retinoic Acid). The combination of ATRA and arsenic trioxide is the first “non-chemotherapeutic” treatment regime for an acute leukaemia. The 3q26 translocations inv(3)(q21;q26) and t(3;3)(q21;26) are unusual in that they do not result in the formation of a fusion gene, but rather result in the repositioning of a distal GATA2 enhancer causing haploinsufficiency of GATA2 and activation of MECOM expression. These abnormalities confer an extremely adverse prognosis. The methyltransferase KMT2A which was previously called MLL has over 70 identified fusion gene partners to date. The most common of these are MLLT3 (AF9), MLLT1 (ENL), MLLT10 (AF10), MLLT4 (AF6), and ELL. These translocations confer a generally adverse prognosis, the exception being t(9;11)(p21;q23) MLLT3-KMT2A, which confers an intermediate prognosis. A complex karyotype in AML is defined as three or more unrelated chromosome abnormalities in the absence of one of the WHO category defining recurring translocations. Complex karyotypes, along with loss of either the whole of chromosome 5 or 7, or loss of the q arm of these chromosomes are common MDS-related abnormalies (see Table 5) which confer a poor prognosis. The molecular landscape of AML at the sub-chromosomal level has been increasingly well characterized over recent years. There are just over 20 commonly mutated genes in AML, which can be grouped into nine functional categories (Table 7). Analysis of the patterns of co-occurring mutations has identified 11 non-overlapping groups with distinct clinical phenotypes and prognostic features. In addition to the already recognized groupings in the WHO classification three new groupings have been postulated: AML with mutations in chromatin and RNA-splicing regulators, AML with TP53 mutations, and AML with IDH1 mutations. Table 7 Commonly mutated genes in AML Functional category Genes Transcription factor fusions PML-RARA, RUNX1-RUNX1T1, CBFB-MYH11 NPM1 gene NPM1 Tumor suppressor genes TP53, WT1, PHF6 DNA methylation related genes DNMT3A, TET2, IDH2, IDH1 Signaling genes FLT3, KIT, KRAS, NRAS, PTPN11 Chromatin modifying genes KMT2A (MLL) fusions, ASXL1, EZH2 Myeloid transcription factor genes RUNX1, CEBPA Cohesin complex genes SMC1A, SMC3, SMC5, RAD21 Spliceosome complex genes U2AF1, SF3B1 Molecular testing in AML should include NPM1, CEBPA, and RUNX1, which define disease categories, and FLT3 ITD/TKD and IDH1/2 as potentially targetable mutations with current therapies. In addition, mutations in several other genes such as in TP53, ASXL1, and WT1, provide further prognostic information. A variety of molecular sequencing techniques can be applied, but as information is required about increasing numbers of genes, large panel next-generation sequencing is increasingly taking the place of single gene tests. In the future whole exome or whole genome sequencing in AML will become increasingly used. Genetic alterations can be used to monitor minimal residual disease (i.e. 1 cell in a million) in follow up samples. FLT3 is a receptor tyrosine kinase expressed on hematopoietic progenitors and mutated in one third of AML cases. There are two main types of mutation that both result in the constitutive activation of the receptor. Internal tandem duplications (ITDs) occur in the juxta-membrane region of the receptor; less frequent are mutations in the tyrosine kinase domain (TKD). FLT3-ITD is an adverse prognostic indicator when present at high allelic ratios (resulting from acquired uniparental disomy), but both FLT3- mutations are now targetable by protein kinase inhibitors, improving overall survival. NPM1 is the most commonly mutated gene in AML, occurring in 35% of de novo cases of AML. NPM1 is a phosphoprotein which, whilst predominantly located in the nucleolus, constantly shuttles between the nucleus and the cytoplasm. A number of mutations have been described All result in loss of a nucleolar binding motif and commonly the gain of a nuclear export signal, leading to the aberrant cytosolic location of this protein. These mutations are not seen in the presence of balanced translocations but commonly co-occur with FLT3-ITD. NPM1 as a sole mutation confers a favorable prognosis. When co-occuring with high allelic FLT3 mutations, the prognosis of each combines to give an overall intermediate risk. CEBPA is mutated in 10% of AML cases. Mutations usually result in disruption of DNA binding/dimerations regions or the production of a truncated protein. Biallelic mutations are common, either resulting from uniparental disomy or acquisition, a second heterozygous mutation. Biallelic CEBPA mutations are generally associated with a favorable prognosis. In addition to the chromosomal translocation involving RUNX1-RUNX1T1, t(8;21), RUNX1 may also be mutated in AML and is included as a provisional entity in the 2016 WHO update. This confers an adverse prognosis. The main differential diagnosis for AML in a bone marrow aspirate or trephine is acute lymphoblastic leukemia, or an acute leukemia of ambiguous lineage. The morphology of lymphoblasts and myeloblasts can be very similar, particularly in the case of AML with minimal differentiation. Immunophenotyping or less commonly immunohistochemistry is required to confirm the diagnosis. Typically myeloblasts are larger than lymphoblasts and can possess a more irregular or convoluted nucleus and a lower nucleus:cytoplasmic ratio. Cytoplasmic granulation is a much more prominent feature of myeloblasts, but sparse granulation can occur in lymphoblasts. Auer rods are pathognomonic of AML. Other disorders that should be considered in the differential include high-grade lymphoma and blastic plasmacytoid dendritic cell neoplasm. Severe megaloblastic anemia should be excluded by checking B12 and folate, particularly before diagnosing pure erythroid leukemia. If blasts are increased but <20% and dysplasia is present, then myelodysplastic syndrome with excess blasts is the likely diagnosis. Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is an aggressive hematological malignancy derived from plasmacytoid dendritic cells (PDCs). In most patients, BPDCN presents with cutaneous lesions. These can be isolated nodules, isolated bruise-like plaques or present as disseminated nodules/papules/macules. Patients with leukemic presentation often lack skin lesions but display lymphadenopathy, splenomegaly, or signs and symptoms of bone marrow failure as in AML. Rare sites of involvement include the liver, tonsils, soft tissues, paranasal cavities, lungs, eyes, and central nervous system. While central nervous system involvement is seen 4 to 9% of patients at diagnosis, it is more frequent (17–33%) at relapse. Nearly 10%–20% of BPDCN patients either have a previous history of, or will develop another hematological malignancy, such as chronic myelomonocytic leukaemia (CMML), MDS, or AML. Incidence Incidence BPDCN is a very rare neoplasm and its exact incidence is unknown due to the low number of cases reported and no formal extensive studies. The largest study to date (379 cases) gives an incidence of 0.45/1,000,000. The variability of nomenclature further complicates assessment of the incidence. Age Age Most patients are elderly with a mean/median age of 61–67 years at diagnosis, although there are documented cases from all age groups, including children. Sex Sex About 75% of the patients are male with a male to female ratio of 2–5:1. Site Site Skin involvement is most common (64–100% of cases) with frequent bone marrow, peripheral blood (60–90%), and lymph node (40–50%) involvement. Treatment Treatment Treatment approach varies and best practice is not yet established. Surgical excision in combination with local radiotherapy and systemic steroid therapy has been assessed by a number of groups. While this might result in the resolution of the cutaneous lesion systemic relapse occurs within 6–9 months. Most conventional chemotherapy approaches consist of an AML or ALL-type acute leukemia-like treatment regimen. Intrathecal chemotherapy is often part of therapy as a pre-emptive measure in view of frequent leptomeningeal involvement. Allogeneic hematopoietic stem cell transplantation (HSCT), particularly when performed in first remission, may improve the survival. Outcome Outcome Disease outcome is poor with a clinical course similar to that of acute leukemia with high-risk features. The largest study to date of a cohort of 379 patients found a median overall survival of 23 months with an inferior survival rate of 19 months amongst patients receiving upfront HSCT as part of their treatment. Cutaneous lesions are often variable in size (few millimetres to over 10 centimentres) and purplish in color. They can be associated with erythema, hyperpigmentation, purpura, or ulceration. Histological examination of cutaneous lesions usually reveals a prominent dermal infiltrate, often extending onto the subcutaneous fat (Fig. 4a, b, and c). Epidermis and appendages are spared for the most part. Infiltrates within lymph nodes are mostly restricted to the paracortical areas and medulla with sparing of the cortex. The degree of bone marrow involvement (Fig. 4d and e) is highly variable. There may be features of associate hematological malignancies, such as myelodysplasia, most often identified by dysplastic features in the megakaryocyte lineage. Open image in new window Fig. 4 Blastic plasmacytoid dendritic cell neoplasm with cutaneous (a) and bone marrow (d) involvement. The neoplastic cells express CD123 (b – skin, e – bone marrow) and CD4 (c – skin). Cell morphology is best appreciated in the aspirate sample (f) The cellular morphology is best appreciated in aspirate (Fig. 4f) or peripheral blood samples. Blastic plasmacytoid dendritic cells are medium sized with a round or irregular nucleus with a fine or intermediate chromatin texture. They may have prominent nucleoli. They contain basophilic cytoplasm often forming a tail. The cytoplasm may contain small vacuoles under the cell membrane, which may give a monocytic appearance. BPDCs must be distinguished from acute leukemia or high-grade lymphoma by flow cytometry or immunohistochemistry, using combination antibodies targeting PDC-associated antigens such as CD123, CD303, and TCL1A. BPDCs also express CD45RA, CD4, CD56, and CD43, indeed they have previously been termed “CD4/CD56 positive cell tumors.” CD4 and CD56 are commonly included in panels assessing acute leukemias, and their dual positivity would lead to examination of the other PDC-associated antibodies that are not a standard component of most panels, even if this diagnosis had not previously been suspected by morphological assessment. Expression of other T-cell (CD7 and to a lesser extent CD2 and CD5 but not CD3), B cell associated (CD79a but not CD19 or CD20), as well as myeloid markers (CD33 and CD117 but not myeloperoxidase) are also frequent findings. About a third of the cases are positive for Tdt, and while CD34 is usually negative on tissue sections, it can often be detected by flow cytometry. CD68 shows cytoplasmic dot-positivity. Expression of BCL6, IRF4/MUM1, BCL2, and S100 can also be seen. Two-thirds of the cases harbor chromosomal abnormalities, often with a complex karyotype. The most frequently recurring, nonspecific chromosomal abnormalities involve 5q21 or 5q34, 12p13, 13q13–21, 6q23-qter, 15q, and loss of chromosome 9. The genomic landscape of BPDCNs is yet to be fully explored. Gene expression profiling has shown this neoplasm to be distinct from both myeloid and lymphoid malignancies and also revealed increased expression of genes involved with Notch signaling and activation of the NF-kappaB pathway, which may represent a promising therapeutic target. TET2 appears to be the most frequently mutated gene and recurrent mutations in ASXL1 and and RAS family genes have also been described. Mutations in genes recurrently mutated in AML, NPM1, FLT3-ITD, IDH2, TP53 and KIT have been detected in some patients. Recent studies have also described mutations in numerous other genes less commonly mutated in myeloid malignancies, such as BRAF, CDKN1B, CDKN2A. The main differential diagnoses of cutaneous lesions are cutaneous involvement by AML or benign mature plasmacytoid dendritic cell proliferations (MPDCP) accompanying myeloid neoplasia, particularly CMML. However, differentiating BPDCN from CD33+/CD4+/CD56+ AML, particularly if undifferentiated or myelomonocytoc/monoblastic with low-level CD123 expression, can be more challenging and usually relies on using additional BPDCN markers. Examination of a paired bone marrow sample is necessary with extensive flow cytometric and potentially molecular assessment, although from a clinical point of view, distinction might not be necessary for treatment planning. Distinction from neoplastic proliferation of mature plasmacytoid dendritic cells also requires bone marrow assessment. MPDCPs are often well-defined nodules of plasmacytoid dendritic cells with an immunophenotype of normal PDCs, although aberrant T-cell marker and/or myeloid marker (e.g., CD33) expression can also be seen. CD56 is usually negative or weakly positive and the proliferation index is low. Clonal genetic alterations shared with the associated myeloid malignancy can also be demonstrated.

Öz