Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Acute Myeloid Leukemia in Children
Acute myeloid leukemia (AML) is a complex and heterogeneous group of malignancies in which genetic and epigenetic alterations lead to differentiation arrest and/or uncontrolled expansion of myeloid cell precursors. Pediatric AML is characterized by a wide array of genetic aberrations that are often not seen or differ in frequency when compared to AML diagnosed in adults and mainly comprise different fusion genes. Though the mutational landscape of childhood AML is now well-described, and despite the recent approvals of various targeted compounds for use in adults with AML, there are no new targeted therapies approved for use in children. This is due to several factors, such as delays to implement clinical studies in children, relevance of some targets in children, and/or class-related toxicity concerns (e.g., Hedgehog pathway inhibitors that may induce skeletal toxicity).
Improvements in survival have been largely achieved through the successful intensification of broadly active chemotherapy (cytarabine and anthracyclines), the effective use of hematopoietic stem cell transplantation (HSCT), improved methods for risk stratification including early response assessments, advanced diagnostics to detect genetic aberrations, improvements in supportive care, and increasing retrieval rates at relapse.
Overall survival (OS) rates for children with AML who are treated on contemporary clinical trials are now greater than 70% overall, and for some genotypes, OS rates are above 80%. However event-free survival (EFS) rates still fall behind, ranging from 50% to 65% (summarized in Table 63.1 ), and the OS for other genotypes are unacceptably low despite maximal intensification of standard chemotherapy. Genomic and biologic insights into the mechanisms of leukemogenesis have both helped to identify patients a priori who will have a poor prognosis, and may potentially benefit from an HSCT as well, and identified potential targets for new therapies. Survivors of childhood AML are at significant risk for serious long-term side effects including cardiomyopathy, infertility, and second malignancies.
Table 63.1
Results of Recent Clinical Trials for Pediatric Acute Myeloid Leukemia
| Type of AML | Study | Years of Enrollment | Eligible Age (Years) | Number of Patients | CR (%) | EFS (%) and (Median Follow-Up) | OS (%) and (Median Follow-Up) | Reference |
|---|---|---|---|---|---|---|---|---|
| AML | AIEOP AML 2002/01 | 2002–2011 | ≤18 | 482 | 87 | 55 (8-year) | 68 (8-year) | |
| AML-BFM 2004 | 2004–2010 | <18 | 611 (521 randomized) | 89 |
|
|
||
| AML-BFM 2012 registry | 2012–2017 | <18 | 324 | 90 | 62 (5-year) | 80 (5-year) | ||
| COG AAML0531 | 2006–2010 | <30 | 1022 | 87 |
|
|
||
| COG AAML 1031 | 2011–2016 | <30 | 1097 | 90 | No Bort: 45; Bort: 47 (3-year) | No Bort: 64; Bort 67 (3-year) | ||
| JPLSG AML05 | 2006–2010 | ≤18 | 443 | 54 (3-year) | 73 (3-year) | |||
| JACLS | 2003–2006 | <18 | 146 | 96 | 67 (5-year) | 78 (5-year) | ||
| MRC AML12 | 1995–2002 | <16 | 529 | 92 | 54 (10-year) | 63 (10-year) | ||
| NOPHO AML 2004 | 2004–2009 | ≤18 | 151 | 92 | 57 (3-years) | 69 (3-years) | ||
| SJCRH AML08 | 2008–2017 | ≤21 | 285 | 92 | Clo-AraC 53 & HD-ADE 52 (3-year) | Clo-AraC 75 & HD-ADE 65 (3-year) | ||
| DB AML01 | 2010–2014 | <17 | 112 | 94 | 53 (3-year) | 74 (3-year) | ||
| ELAM02 | 2005–2011 | <18 | 438 | 89 | 57 (4-years) | 73 (4-years) | ||
| ML DS | AML D05 study | 2008–2010 | <18 | 72 | 96 | 83 (3-year) | 88 (3-year) | |
| ML DS 2006 | 2007–2015 | <4 | 170 | 87 (5-year) | 89 (5-year) | |||
| COG AAML 0431 | 2007–2011 | <4 | 204 | 90 (5-year) | 93 (5-year) | |||
| TMD | COG A2971 | 1999–2004 | neonates | 135 | 57 (3-year) | 77 (3-year) | ||
| TMD prevention 2007 study | 2007–2015 | neonates | 102 | 72 (5-year) | 91 (5-year) | |||
| APL | AML P05 study (ATRA/chemo) | 2006–2011 | <18 | 43 | 86 | 84 (3-year) | 91 (3-year) | |
| AML-BFM 93/-98/-2004 (ATRA/chemo) | 1993–2007 | <18 | 81 | 93 | 73 (5-year) | 89 (5-year) | ||
| AML-BFM Standard-Risk APL (ATRA/ATO) | 2013–2016 | <18 | 11 | 100 | 100 (2.4-year) | 100 (2.4-year) | ||
| ICC-APL-01 (ATRA/chemo) | 2008–2015 | <21 | 258 | 97 | 80 (5-year) | 95 (5-year) | ||
| GIMEMA/AIEOP AIDA 2000 ATRA/chemo) | 2000–2008 | <18 | 127 | 96 | 85 (10-year) | 94 (10-year) | ||
| CALGB 9710 (ATRA/chemo/ATO) | 1999–2005 | <18 | 83 | 83 | 54 (5-year) | 82 (5-year) | ||
| COG AAML 0631 (ATRA/chemo/ATO) | 2009–2012 | <22 | 101 | 91 (3-year) | 94 (3-year) |
AML , Acute myeloid leukemia; APL , acute promyelocytic leukemia; AIEOP , Associazione Italiana di Ematologia e Oncologia Pediatrica; ATO , arsenic trioxide; ATRA , all-trans-retinoic acid; BFM , Berlin–Frankfurt–Münster Study Group; Bort , bortezomib arm; CALGB , Cancer and Leukemia Group B; COG , Children’s Oncology Group; CR , complete remission; DB , Dutch Belgian; EFS , event-free survival; ELAM , Enfants Leucémie Aiguë Myéloïde; ICC-APL , International Consortium for Childhood APL; Ida , idarubicin arm; JACLS , Japan Association of Childhood Leukemia Study; JPLSG , Japanese Pediatric Leukemia/Lymphoma Study Group; L-DNR , liposomal daunorubicin arm; ML DS , myeloid leukemia of Down syndrome; MRC , Medical Research Council; NOPHO , Nordic Society of Paediatric Haematology and Oncology; OS , overall survival; SJCRH , St Jude Children's Research Hospital; TMD , transient myeloproliferative disease.
Future advances in childhood AML must include a better understanding of whether targeted therapies approved in older adults can be effective in children with AML by determining the dependency of childhood leukemia cells on targetable oncogene products and understanding the compensatory mechanism of resistance to conventional and targeted therapies. This should lead to higher cure rates with less long-term toxicity.
Epidemiology
AML accounts for approximately 20% of cases of acute leukemia in children and adolescents younger than 20 years of age. The incidence rates have remained constant over the past 40 years, are similar between boys and girls, and are highest during the first 2 years of life before the incidence rises again in adolescents/young adults. However, the age distribution may vary between subtypes. For example, acute promyelocytic leukemia (APL) and core-binding factor (CBF) leukemia are rare in children younger than 3 years of age, whereas the incidence of acute megakaryoblastic leukemia (AMKL) and “myeloid leukemia of Down syndrome” (ML DS) is highest in infants and young children, yet rare among teenagers. The distribution of AML subtypes may also vary among ethnic groups, with some studies suggesting a higher incidence of APL among Hispanic/Mediterranean populations.
Constitutional syndromes that are associated with an increased predisposition to AML include Down syndrome (DS), Fanconi anemia, Bloom syndrome, neurofibromatosis, Noonan syndrome, congenital neutropenia, and germline haploinsufficiency of the specific oncogenes (e.g., the Runt-related transcription factor-1 or RUNX1 gene), though these constitute only a small subset of AML diagnosed in children. The World Health Organization (WHO) 2016 classification also provides a listing of myeloid malignancies that can develop in patients with germline predispositions ( Table 63.2 ). Exposure to ionizing radiation, alkylating agents, topoisomerase II inhibitors, and benzene are among the few environmental factors proven to increase the risk of AML. In the majority of cases of childhood AML, neither a genetic nor environmental cause can be identified.
Table 63.2
The 2016 World Health Organization Classification of Acute Myeloid Leukemia and Related Neoplasms
Modified from Arber DA. The 2016 WHO classification of acute myeloid leukemia: what the practicing clinician needs to know. Semin Hematol . 2019;56:90–95.
| WHO 2016 Classification | Classification of Myeloid Neoplasms With Germline Predisposition |
|---|---|
|
|
AML , Acute myeloid leukemia; APL , acute promyelocytic leukemia; NMP1 , nucleophosmin 1; WHO , World Health Organization.
Pathobiology
The French–American–British (FAB) classification system uses morphologic and histochemical features of the leukemic cells to sub-classify AML, though with limited prognostic and therapeutic significance. The WHO 2016 classification system (see Table 63.2 ) uses specific cytogenetic alterations and molecular genetic lesions to classify AML. Recent genomic approaches have validated the distinct genomic heterogeneity within AML. In addition, these approaches have identified new genetic subtypes as well as lesions that are enriched within leukemias that have normal karyotypes, that is, genetic changes not identifiably by standard karyotype analysis or miscellaneous chromosomal alterations ( Fig. 63.1 ). This chapter discusses many of the known molecular subtypes and their associated prognosis.
Recent DNA and RNA sequencing efforts have significantly expanded the list of molecular lesions that underlie AML. Pediatric AML is characterized by very few somatic mutations per cell, though there are many distinct mutations described. Data support models for childhood AML where single and catastrophic translocation events, that result in the generation of an entirely novel gene product such as RUNX1/RUNX1T1 , initiate leukemogenesis in early hematopoietic stem and progenitor cells. Additional acquired leukemogenic mutational “hits” ultimately lead to a malignant transformation. Common initiating mutations described for adults with AML include isocitrate dehydrogenase1/2 ( IDH1/2 ) and DNA methyltransferase 3 alpha (DNMT3A) that alter gene function. In contrast to chromosomal rearrangements, mutations in these genes are exceedingly rare in children. There may also be a developmental context that mediates the potency of these chromosomal alterations and their requirement for secondary mutations. Hematopoietic progenitor cells in children are fundamentally different than those of older adults. Because of their much higher proliferative and self-renewal capacity, the fetal, infant and pediatric hematopoietic stem and progenitor cell compartment provides a much more supportive environment for initiating genetic lesions than the adult one. It should be no surprise that stem cells within different developmental phases will require different oncogenic events needed for a malignant transformation.
It is useful to conceptualize mutations in AML as falling into two broad classes: class I mutations confer a proliferative or survival advantage; and class II mutations block differentiation and promote self-renewal. In general, class I and class II mutations frequently cooperate in leukemogenesis. Interestingly, many of the class II mutations arise from translocation events that lead to chimeric transcription factors with oncogenic properties. Moreover, the relationship between these class I and class II mutations is non-random, for example, KIT (class I) mutations are mostly found in CBF AML (class II), and fms-like tyrosine kinase 3 ( FLT3) mutations (class I) are mostly found in AML with nucleoporin 98 (NUP98) fusions (class II) or nucleophosmin (NPM1) mutations (class II). However, although more amenable to conventional drug design than class II mutations, class I mutations are generally secondary events and, thus, frequently subclonal.
In this chapter, we discuss the major subtypes of pediatric AML, our current understanding of the biologic processes that mediate disease, and how these major subtypes may impact survival following conventional therapy. The following section highlights some of the major genetic driver events in leukemogenesis in children. Chapter 59 provides a more detailed overview of the pathobiology of AML with a focus on genetics in adult AML.
Fusion Genes Resulting From Chromosomal Rearrangements
About 60% of all pediatric AML cases are associated with six genetic loci targeting five different protein complexes: the transcription factor CBF with 21q22 ( RUNX1 ) and 16q22 ( CBFB ) rearrangements, the epigenetic modifier KMT2A with 11q23 ( KMT2A , previously known as MLL ) translocations, the transcription factor retinoic acid receptor alpha with 17q21 ( RARA ) translocations, and the nuclear pore component nucleoporin 98 and 96 precursor with 11p15 ( NUP98 ).
Chromosomal Rearrangements Affecting Core-Binding Factor
The CBF transcription complex is a heterodimeric complex composed of a DNA-binding RUNX member (either RUNX1, 2, or 3) and CBFβ. The RUNX1:CBFβ complex functions as a master regulator of the definitive hematopoiesis. Mutations in the genes encoding the RUNX1/CBFβ transcription factor complex are one of the most common lesions seen in de novo AML, occurring in approximately 25% of cases.
The translocation t(8;21)(q22;q22) is the most common translocation found in pediatric AML and fuses the DNA-binding domain of RUNX1 to almost the entire coding region of RUNX1T1. The closely related translocation t(16;21) (q24;q22) fuses RUNX1 to CBF2A3 , a paralog of RUNX1T1 . Following the discovery of RUNX1, the inv(16)(p13;q22) and the less common t(16;16)(p13;q22) result in a fusion between CBFβ and MYH11, frequently seen in acute monoblastic leukemias with eosinophilia (FAB-M4Eo). The CBFβ/MYH11 fusion protein retains the RUNX1 binding domain and therefore its ability to interact with wild-type RUNX1. Both fusion proteins interfere with normal RUNX1 functions but exert also RUNX1-independent functions. In particular, RUNX1/RUNX1T1 can initiate leukemogenesis but requires secondary mutations for establishing AML. Consistent with this, RUNX1/RUNX1T1 expressing progenitors can persist in the bone marrow for years without expanding into overt leukemia. Nevertheless, AMLs are addicted to these fusion proteins and their transcriptional programs. Cooperating mutations include both classic class I mutations including FLT3 , KIT , and RAS members and epigenetic regulators such as ASXL1 and 2 or EZH2 .
Patients with CBF mutations tend to have an improved survival with conventional therapy and even improved survival following relapse, although this is more apparent for inv(16) than for t(8;21) rearranged cases. As a result, in some collaborative group protocols, patients with CBF AML receive fewer courses of intensive chemotherapy.
KMT2A Gene Rearrangements
The KMT2A gene, located on chromosome 11 band q23, encodes the MLL histone methylase. Upon proteolytic cleavage, it is part of a multiprotein complex involved in chromatin remodeling. Knockout of KMT2A inhibits myeloid and lymphoid differentiation and rapidly leads to bone marrow failure in mice. These results suggest that the MLL protein is not only required for definitive hematopoiesis but is also required for maintenance of HSCs in postnatal hematopoiesis.
Chromosomal translocations involving the KMT2A gene are associated with both AML and ALL leukemias. There are more than 100 different fusion partners identified, although nine specific genes account for more than 90% of the fusion events. KMT2A rearrangements are found in 60% to 70% of infants with leukemia (predominantly KMT2A/AFF1 ), regardless of immunophenotype, but are less common in older children and adults. The distribution of translocation partners varies depending on the age of the patient and the immunophenotype of leukemia, with KMT2A/MLLT10 being more frequent in infant and pediatric AML. The majority of active KMT2A fusion proteins contain the N-terminus of KMT2A and the C-terminus of the fusion partner gene, while the function of reciprocal fusion proteins harboring the KMT2A C-terminus are not well understood. The prognostic significance of KMT2A mutations is often dictated by the fusion partner and can range from favorable to highly unfavorable. Next-generation sequencing of KMT2A -rearranged leukemias demonstrated a paucity of cooperating mutations, although approximately 50% carry an activating tyrosine kinase mutation. However, those mutations are often absent at relapse, confirming they may represent secondary events in irrelevant subclones. KMT2A fusion genes alone are sufficient for leukemogenesis, and the targeting of cooperating mutations may not result in a therapeutic benefit.
RARA Gene Rearrangements
The hallmark of APL is chromosomal translocations involving the RARA gene at 17q22 coding for the retinoic acid receptor alpha. The binding of retinoid acid to a heterodimer of RARA and retinoid X receptor (RXR) leads to a conformational change and converts a transcriptional repressor to an activator that drives granulopoiesis. About 98% of all APL cases harbor the translocation t(15;17)(q24;q21) fusing the PML gene to RARA . The remaining 2% consist of eight additional translocations with the fusion proteins ZBTB16/RARA (also known as PLZF/RARA) and STAT5B/RARA, being resistant to retinoid acids including all-trans-retinoic acid (ATRA). The most frequently observed secondary mutations with 25% to 60% of all cases are internal tandem duplications of FLT3 .
PML/RARA represses RARA-RXR-driven gene expression and also alters the localization of wild-type PML away from nuclear bodies, thereby interfering with processes such as DNA damage response and telomere maintenance ultimately inhibiting senescence. Importantly, PML/RARA-expressing APL is highly sensitive to ATRA, inducing a conformational change that relieves the transcriptional repression and drives terminal granulocytic differentiation associated with apoptosis. At higher doses, ATRA also induces degradation of the fusion protein. ATRA alone achieves remission very efficiently but has a high relapse rate, whereas combinations of ATRA with combination chemotherapy yield cure rates of more than 90%. Arsenic trioxide (ATO) has also been shown to induce remissions in patients who had failed ATRA and conventional chemotherapy. Similar to ATRA, ATO triggers degradation of PML/RARA and induces apoptosis of APL, while low doses promote maturation and differentiation. The combination of ATRA with ATO achieves similar cure rates as ATRA with standard chemotherapy but avoids the side effects associated with anthracycline-based combination chemotherapy. This regimen may convert APL into the first leukemia that can be cured by targeted therapy alone.
NUP98 Rearrangements
The NUP98 gene at chromosome 11p15 encodes a 186 kDa precursor protein that by autoproteolytic cleavage matures in two nucleoporins of 98 kDa 96 kDa. Both proteins are components of the nuclear pore complex. In addition, NUP98 binds to promoter regions and activates gene expression by recruiting histone methylases that methylate lysine 4 of histone 3 (H3K4). Initially shown to be a fusion partner with HOXA9 in the t(7;11)(p15;p15) translocation, it is similar to KMT2A in that multiply other fusion partners have been subsequently identified. Most NUP98 fusions are the result of inversions or balanced translocations. Due to their cryptic nature—NUP98 rearrangements are not visible by standard karyotype analysis—their frequency has initially been underestimated. Recent next-generation sequencing analyses established that 5% to 9% of infant and pediatric AMLs harbor NUP98 rearrangements, an incidence significantly higher than in adult AML. Some standard fluorescent in situ hybridization (FISH) panels now include probes capable of identifying NUP98 fusions.
The most common NUP98 translocation in childhood AML is t(5;11)(q35;p15) fusing NUP98 to NSD1 , encoding a histone methylase with a preference for methylating lysine residues 36 and 20 in histones 3 and 4 (H3K36 and H4K20), respectively, thereby modulating gene expression. The NUP98/NSD1 contains the NSD1 SET domain, harboring the methylase activity. Thus, NUP98/NSD1 activates and maintains gene expression by methylating H3K36 of its target genes. Another prominent NUP98 fusion is t(11;12)(p15;p13), NUP98/KDM5A , that occurs mainly in AMKL, where it constitutes approximately 8% of pediatric non–DS-AMKL cases.
AMLs with NUP98 rearrangements are characterized by high expression of both the HOXA and the HOXB cluster, thereby driving leukemic self-renewal. H3K4 methylation by MLL has been shown to direct NUP98/JARID1 binding to and activation of HOXA expression. Similarly, NSD1 binds to methylated H3K4 and H3K9 by its PHD domain, which is also kept in the NUP98/NSD1 fusion protein. Furthermore, NUP98/KDM5A has also been demonstrated to bind to H3K4me3 nucleosomes via a PHD finger. This dependency of NUP98 fusion proteins on a functional MLL complex suggests that this group of AMLs may also benefit from drugs developed for KMT2A mutated AML.
Interestingly, the majority of NUP98/NSD1-positive AMLs also harbor the FLT3-ITD mutation and, to a lesser extent, Wilms tumor 1 (WT1) mutations. Such patients have a dismal outcome with current chemotherapy regimens. In this context, FLT3 inhibitors may offer another targeted option for this hard-to-treat AML cohort.
Chromosomal Rearrangements Involving E26 Transformation Specific Family Members
Several members of the E26 transformation specific (ETS) transcription factor family including ERG, FLI1, and SPI1 (also known as PU.1) are crucial regulators of hematopoiesis. They support both normal and malignant hematopoiesis by cooperating with CBF and its fusion proteins RUNX1/RUNX1T1 and CBFB/MYH11. Moreover, the expression of ERG is directly regulated by these leukemic fusion proteins.
In particular, ERG (21q22) and FLI1 (11q24) are targets of chromosomal rearrangements leading to a fusion with FUS (16p11) or its paralog EWSR1 (22q12). These translocations are prominent in Ewing sarcoma but are also observed mainly in older children and young adults with AML.
The FUS protein orchestrates both transcription and splicing by binding to RNA polymerase II with its N-terminus while interacting with RNA and the U1 snRNP via C-terminal domains. The translocations fuse the N-terminal regions of FUS and EWSR1 to the C-terminal ETS DNA-binding domains of ERG or FLI1. The resultant fusion protein dysregulates gene expression both by aberrantly activating gene expression and by interfering with RNA splicing. Interestingly, FUS/ERG co-occupies genomic binding sites with the RXR:RARA nuclear receptor complex. In line with this observation, treatment of a FUS/ERG-positive AML cell line with ATRA has been shown to block proliferation and induce differentiation. However, in particular, FUS/ERG -positive AML has a very poor prognosis with current treatment regimens.
Another member of the ETS genes is ETV6, whose fusion to RUNX1 is found in 25% of all pediatric ALL cases. The translocation t(7;12)(q36;p13) fuses ETV6 to the homeobox gene MNX1 . This translocation generates a very heterogeneous fusion; the corresponding fusion transcript is only detected in 50% of all cases and it has remained unclear if the ETV6 moiety is expressed at all, suggesting that high expression of MNX alone may be the leukemic driver. The translocation is with 30% incidence one of the most frequent rearrangements in infant AML. AMLs with this translocation have an unfavorable prognosis.
DEK/NUP214
The translocation t(6;9)(p23;q34), identified in less than 2% of all pediatric AML, is a rare location that is associated with a poor clinical outcome. Similar to NUP98, NUP214 is part of the nuclear pore complex; by interaction with transporter proteins such as CRM1, XPO1, and NXF1, it is involved in protein and mRNA transport across the nuclear membrane. DEK recruits histone deacetylases to DNA and is an epigenetic regulator of transcription and splicing. Its transcriptional signature includes the activation of HOXA and HOXB expression. Another similarity with NUP98 fusions is the more than 50% incidence of the FLT3-ITD mutation. DEK/NUP214 predicts a poor clinical outcome with lower rates of complete remission (CR) and OS.
KAT6A Fusions
The KAT6A gene at chromosome 8p11 encodes the histone acetyl transferase (HAT) MOZ (also known as MYST) that is a coactivator for transcription factors including RUNX1 and is required for maintaining the hematopoietic stem cell pool. The two main translocations involve with CREBBP and EP300 at 16p13 and 22q13, respectively, another two closely related HATs. Interestingly, each fusion protein contains both HAT domains from the fusion partners and leads to increased HOXA expression. Both fusions are with less than 1% of all AML cases rare and are mainly associated with monocytic AML of FAB subtype 4 and 5 in infants and young children. Although occasional spontaneous regressions have been reported for KAT6A/CREBBP , especially in young children, both translocations predict an unfavorable clinical outcome.
RBM15/MRTFA
The RBM15 gene at chromosome 1p13 encodes an RNA-binding protein that controls the methylation of RNA at the N6 of adenosines and the splicing of genes relevant of megakaryopoiesis including MPL. MPL is the receptor for thrombopoietin and an inducer of HOX-gene expression. About 10% of all pediatric AMKL cases harbor the translocation t(1;22)(p13;q13), which fuses RBM15 to MRTFA (also known as MKL). MRTFA is a regulator of myogenesis and cell migration. By interacting with serum response factor, it integrates RAC, RHO, and CDC42 GTPase signaling by binding to G-acting and activates the expression of cytoskeletal genes. Notably, RBM15/MRTFA retains all functional domains of each partner protein. This translocation is associated with an intermediate clinical outcome.
CBF2A3/GLIS2
Until recently, with the exception of the RBM15/MRTFA fusion, the genetic etiology of non-DS-AMKL had remained elusive. Transcriptome sequencing of a small cohort identified a cryptic inversion on chromosome 16 [inv(16)(p13.3q24.3)] in half of the patients that resulted in the joining of CBFA2T3 , a member of the ETO family of nuclear corepressors also involved in RUNX1 translocations, to GLIS2 , a member of the GLI family of transcription factors. This translocation is almost exclusively found in children younger than 4 years. CBFA2T3/GLIS2 AMKL has a distinct gene expression signature: the transcription factor GATA1 required for megakaryopoiesis, and which is mutated in DS-associated AMKL, is downregulated, while ERG as a promoter of self-renewal is upregulated. Patients harboring this mutation have an extremely poor prognosis.
You're Reading a Preview
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here