We present the molecular landscape of pediatric acute myeloid leukemia (AML) and characterize nearly 1,000 participants in Children’s Oncology Group (COG) AML trials. The COG–National Cancer Institute (NCI) TARGET AML initiative assessed cases by whole-genome, targeted DNA, mRNA and microRNA sequencing and CpG methylation profiling. Validated DNA variants corresponded to diverse, infrequent mutations, with fewer than 40 genes mutated in >2% of cases. In contrast, somatic structural variants, including new gene fusions and focal deletions of MBNL1, ZEB2 and ELF1, were disproportionately prevalent in young individuals as compared to adults. Conversely, mutations in DNMT3A and TP53, which were common in adults, were conspicuously absent from virtually all pediatric cases. New mutations in GATA2, FLT3 and CBL and recurrent mutations in MYC-ITD, NRAS, KRAS and WT1 were frequent in pediatric AML. Deletions, mutations and promoter DNA hypermethylation convergently impacted Wnt signaling, Polycomb repression, innate immune cell interactions and a cluster of zinc finger–encoding genes associated with KMT2A rearrangements. These results highlight the need for and facilitate the development of age-tailored targeted therapies for the treatment of pediatric AML.

However, the relevance of these findings to childhood AML remains unclear, since several of the most 53 common adult mutations appear far less prevalent in pediatric AML 6,7 . 54To date, no comprehensive characterization of pediatric AML has been described. Here, we report the 55 initial results of the TARGET (Therapeutically Applicable Research to Generate Effective Treatments) 56 AML initiative, a collaborative COG/NCI project to comprehensively characterize the mutational, 57 transcriptional, and epigenetic landscapes of a large, well-annotated cohort of pediatric AML. 58Comparing AML molecular profiles across age groups, we show that stark differences in mutations,d 59 structural variants and DNA methylation distinguish AML in infants, children, adolescents, and adults. 60 Results 61 Overview of cohort characteristics 62A total of 1023 children enrolled in COG studies are included in the TARGET AML dataset. 63Comprehensive clinical data, including clinical outcomes and test results for common sequence 64 aberrations (outlined in We carried out analyses of microRNA, mRNA, and/or DNA methylation in 412 subjects. A summary of 94 the assays performed and case-assay overlap is presented in Fig. S3. We compared our verified variants 95 to those of 177 adult AML cases from The Cancer Genome Atlas (TCGA) project 3 , stratified by the age 96 groupings outlined in Fig. 1a. The TARGET and TCGA discovery cohorts both contained numerous AYA 97 patients (Table S3). Importantly, our conclusions regarding the molecular characteristics of this age 98 group are identical when analyzing either or both cohorts (Fig. S4). 99 Somatic gene mutations in pediatric AML 100Like adult AML, pediatric AML has one of the lowest rates of mutation among molecularly well-101 characterized cancers (Fig. S5), with < 1 somatic, protein-coding change per megabase in most cases. 102However, the landscape of somatic variants in pediatric AML is markedly different from that reported in 103 adults 3,4 (Figs. 2b, S6-S7, Table S4). RAS, KIT, and FLT3 alterations, including novel, pediatric-specific 104 FLT3 mutations (FLT3.N), are more common in children. Mutational burden increases with age, yet older 105 patients have relatively fewer recurrent cytogenetic alterations. Indeed, the number of coding SNVs, 106 within and across cohorts, is best predicted by age (Fig. 2c, p<10 -15 ) and by cytogenetic subgroup. In 107 contradistinction to the higher prevalence of small sequence variants in older patients, recurrent 108 structural alterations, fusions, and focal copy number aberrations are more common in younger patients 109 (Figs. 2d-e, p<10 -3 , see below). Patients with CBFA2T3-GLIS2, KMT2A, or NUP98 fusions tend to have 110 . CC-BY 4.0 International license peer-reviewed) is the author/funder. It is made available under a The copyright holder for this preprint (which was not . http://dx.doi.org/10.1101/125609 doi: bioRxiv preprint first posted online Jun. 13, 2017; fewer mutations (p<10 -9 ), with subgroups demonstrating inferior clinical outcome...

To study the prognostic relevance of rare genetic aberrations in acute myeloid leukemia (AML), such as t(16;21), international collaboration is required. Two different types of t(16;21) translocations can be distinguished: t(16;21)(p11;q22), resulting in the fusion gene; and t(16;21)(q24;q22), resulting in RUNX1-core binding factor (). We collected data on clinical and biological characteristics of 54 pediatric AML cases with t(16;21) rearrangements from 14 international collaborative study groups participating in the international Berlin-Frankfurt-Münster (I-BFM) AML study group. The AML-BFM cohort diagnosed between 1997 and 2013 was used as a reference cohort. (n = 23) had significantly lower median white blood cell count (12.5 × 10/L, = .03) compared with the reference cohort. rearranged AML (n = 31) had no predominant French-American-British (FAB) type, whereas 76% of had an M1/M2 FAB type (M1, M2), significantly different from the reference cohort ( = .004). Four-year event-free survival (EFS) of patients with was 7% (standard error [SE] = 5%), significantly lower compared with the reference cohort (51%, SE = 1%, < .001). Four-year EFS of was 77% (SE = 8%, = .06), significantly higher compared with the reference cohort. Cumulative incidence of relapse was 74% (SE = 8%) in , 0% (SE = 0%) in compared with 32% (SE = 1%) in the reference cohort ( < .001). Multivariate analysis identified both and as independent risk factors with hazard ratios of 1.9 ( < .0001) and 0.3 ( = .025), respectively. These results describe 2 clinically relevant distinct subtypes of pediatric AML. Similarly to other core-binding factor AMLs, patients with RUNX1-CBFA2T3 rearranged AML may benefit from stratification in the standard risk treatment, whereas patients with FUS-ERG rearranged AML should be considered high-risk.

Genomic characterization of pediatric patients with acute myeloid leukemia (AML) has led to the discovery of somatic mutations with prognostic implications. Although gene-expression profiling can differentiate subsets of pediatric AML, its clinical utility in risk stratification remains limited. Here, we evaluate gene expression, pathogenic somatic mutations, and outcome in a cohort of 435 pediatric patients with a spectrum of pediatric myeloid-related acute leukemias for biological subtype discovery. This analysis revealed 63 patients with varying immunophenotypes that span a T-lineage and myeloid continuum designated as acute myeloid/T-lymphoblastic leukemia (AMTL). Within AMTL, two patient subgroups distinguished by FLT3-ITD and PRC2 mutations have different outcomes, demonstrating the impact of mutational composition on survival. Across the cohort, variability in outcomes of patients within isomutational subsets is influenced by transcriptional identity and the presence of a stem cell–like gene-expression signature. Integration of gene expression and somatic mutations leads to improved risk stratification. Significance: Immunophenotype and somatic mutations play a significant role in treatment approach and risk stratification of acute leukemia. We conducted an integrated genomic analysis of pediatric myeloid malignancies and found that a combination of genetic and transcriptional readouts was superior to immunophenotype and genomic mutations in identifying biological subtypes and predicting outcomes. This article is highlighted in the In This Issue feature, p. 549

Supplemental material Supplemental material and methods PatientsThe study comprised COG and European patients. The COG cohort included 2003 de novo patients from COG trials AAML03P1, AAML0531 and AAML1031 1 . Patients with FLT3/ITD high allelic ratio from AAML1031 were excluded due to potentially enrolling onto the Phase I sorafenib treatment arm, which remained under the purview of the COG Data Safety Monitoring Committee and thus, were not analyzed. The European cohort consisted of 343 pediatric AML patients from AIEOP (Associazione Italiana di Ematologia e Oncologia Pediatrica), BFM (Berlin-Frankfurt-Münster), CPH (Czech Pediatric Hematology Working Group), DCOG (Dutch Childhood Oncology Group) and LAME (Leucémie Aiquë Myéloblastique Enfant), previously described by Balgobind et al 2 . Patients included in this study were diagnosed between 1995 and 2017. NUP98-KDM5A rearrangements, and other rearrangements or mutations, were detected either by paired-end RNA sequencing, whole genome sequencing or RT-PCR, as previously described 3 . Patients with FAB M3 AML were excluded from this study. Flow cytometry-based minimal residual disease (MRD) data was obtained for patients in the COG trials only. Gene expression analysisData of 1035 patients from the COG AAML1031 trial were utilized for gene expression analysis using RNAseq data. Fractional counts were normalized to trimmed mean of m-values and counts per million mapped reads (CPM). The normalized counts were log2 transformed and filtered for genes with at least 1 CPM in 5% of samples. For hierarchical clustering, the 2 relative level of expression per gene in each sample was determined by mean centering the expression values, using the geometric mean. Pearson correlation coefficients were employed as a measure of dissimilarity with the ward.D2 linkage algorithm was implemented in the R statistical programming environment (R v.3.4.0) 4 . Differential expression analysis was completed using Limma v3.32.5 R package. Genes with absolute log2 fold-change >1 and Benjamini-Hochberg adjusted p-values <0.05 were retained. Principal coordinate analysis (classical multidimensional scaling) was completed with variance stabilized transformed counts.The transformed counts were then used in a linear metric multidimensional scaling with a Bray-Curtis dissimilarity matrix using the Vegan v.2.4-3 R package.Gene-set enrichment analysis was completed using the GAGE v2.30.0 R package with gene sets from MSigDB (http://software.broadinstitute.org/gsea/msigdb) and the KEGG pathway databases. Connectivity mapping was performed using the online webtool from Broad institute, https://clue.io 5 . A query was performed using the 90 th percentile absolute log fold change of the differentially expressed genes in either NUP98-KDM5A or NUP98-NSD1 rearranged AML. Cell line and primary materialThe AMKL cell line, CHRF-288-11, carrying the NUP98-KDM5A translocation, was a kind gift from Dr. Gruber's lab, St. Jude Children's Hospital (Memphis, TN, USA) (suppl. figure 1). Cells were maintained in RPMI...