Additional cytogenetic/molecular abnormalities that may have prognostic significance

AML with RUNX1::CBFA2T3 (t(16;21)(q24;q22)) gene fusions. In leukemias with t(16;21)(q24;q22), the RUNX1 gene is fused with the CBFA2T3 gene, and the gene expression profile is closely related to that of AML cases with t(8;21) and RUNX1::RUNX1T1 fusions.[292] Patients present at a median age of 7 years. This cancer is rare, representing approximately 0.1% to 0.3% of pediatric AML cases. Among 23 patients with RUNX1::CBFA2T3 fusions, five presented with secondary AML, including two patients who had a primary diagnosis of Ewing sarcoma. Outcomes were favorable for the cohort of 23 patients, with a 4-year EFS rate of 77% and a cumulative incidence of relapse rate of 0%.[292]

RAS variants. Although variants in RAS have been identified in 20% to 25% of patients with AML, the prognostic significance of these variants has not been clearly shown.[249,338] Variants in NRAS are more commonly observed than variants in KRAS in pediatric AML cases.[249,339] RAS variants occur with similar frequency for all Type II alteration subtypes, with the exception of APL, for which RAS variants are seldom observed.[249]

AML with RBM15::MRTFA gene fusions. The t(1;22)(p13;q13) translocation that produces RBM15::MRTFA fusions (also known as RBM15::MKL1) is uncommon (<1% of pediatric AML) and is restricted to AMKL.[210,299,340-343] Studies have found that t(1;22)(p13;q13) is observed in 10% to 20% of children with AMKL who have evaluable cytogenetics or molecular genetics.[265,266,298,300] Most AMKL cases with t(1;22) occur in infants, with the median age at presentation (4–7 months) being younger than for other children with AMKL.[265,295,300,344] Cases with detectable RBM15::MKL1 fusion transcripts in the absence of t(1;22) have also been reported because these young patients usually have hypoplastic bone marrow.[341] An international collaborative retrospective study of 51 t(1;22) cases reported that patients with this abnormality had a 5-year EFS rate of 54.5% and an OS rate of 58.2%, similar to the rates for other children with AMKL.[265] In another international retrospective analysis of 153 cases with non–Down syndrome AMKL who had samples available for molecular analysis, the 4-year EFS rate for patients with t(1;22) was 59% and the OS rate was 70%, significantly better than for AMKL patients with other specific genetic abnormalities (CBFA2T3::GUS2 fusions, NUP98::KDM5A fusions, KMT2A rearrangements, monosomy 7).[298] Similar outcomes were seen in the COG AAML0531 and AAML1031 phase III trials (5-year OS rates, 86% ± 26% [n = 7] and 54% ± 14% [n = 14] for AAML0531 and AAML1031, respectively).[300]

HOX rearrangements. Cases with a gene fusion involving a HOX cluster gene represented 15% of pediatric AMKL in one report.[266] This report observed that these patients appear to have a relatively favorable prognosis, although the small number of cases studied limits confidence in this assessment.

GATA1 variants. GATA1-truncating variants in non–Down syndrome AMKL arise in young children (median age, 1–2 years) and are associated with amplification of the RCAN1 gene on chromosome 21.[266] These patients represented approximately 10% of non–Down syndrome AMKL and appeared to have a favorable outcome if there were no prognostically unfavorable fusion genes also present, although the number of patients studied was small (n = 8).[266]

Hypodiploidy. Hypodiploidy is defined as a modal chromosome number of less than or equal to 45. This occurs rarely in pediatric patients with AML. In a retrospective cohort analysis, the iBFM AML study group aimed to characterize hypodiploidy in pediatric patients with AML. The study excluded several patient groups, including patients with APL, Down syndrome, or loss of chromosome 7.[345] Their observations included the following:

Hypodiploidy. Hypodiploidy is defined as a modal chromosome number of less than or equal to 45. This occurs rarely in pediatric patients with AML. In a retrospective cohort analysis, the iBFM AML study group aimed to characterize hypodiploidy in pediatric patients with AML. The study excluded several patient groups, including patients with APL, Down syndrome, or loss of chromosome 7.[345] Their observations included the following:

Hypodiploidy was observed in 1.3% of children with AML. Approximately 80% of patients had a modal chromosome number of 45, and the remaining 20% of patients had a modal chromosome number of either 43 or 44.

Hypodiploidy. Hypodiploidy is defined as a modal chromosome number of less than or equal to 45. This occurs rarely in pediatric patients with AML. In a retrospective cohort analysis, the iBFM AML study group aimed to characterize hypodiploidy in pediatric patients with AML. The study excluded several patient groups, including patients with APL, Down syndrome, or loss of chromosome 7.[345] Their observations included the following:

Most patients (>80%) with a modal chromosome number of 43 or 44 also met the criteria for complex karyotype. In this study, a complex karyotype was defined as at least three independent chromosomal abnormalities, regardless of whether these were structural abnormalities or defects in chromosome number, and an absence of recurrent aberrations as defined by the WHO.

Hypodiploidy. Hypodiploidy is defined as a modal chromosome number of less than or equal to 45. This occurs rarely in pediatric patients with AML. In a retrospective cohort analysis, the iBFM AML study group aimed to characterize hypodiploidy in pediatric patients with AML. The study excluded several patient groups, including patients with APL, Down syndrome, or loss of chromosome 7.[345] Their observations included the following:

Patients with a modal chromosome number of 43 or 44 had decreased EFS rates and OS rates when compared with patients who had 45 chromosomes (EFS rate, 21% vs. 37%; P = .07; OS rate, 33% vs. 56%; P = .1).

UBTF tandem duplication. UBTF is located at chromosome 17q21.31, and it codes for a nucleolar protein that interacts with ribosomal DNA to mediate RNA polymerase 1 ribosomal RNA transcription.[346]

UBTF tandem duplication. UBTF is located at chromosome 17q21.31, and it codes for a nucleolar protein that interacts with ribosomal DNA to mediate RNA polymerase 1 ribosomal RNA transcription.[346]

UBTF tandem duplication (UBTF-TD) is mutually exclusive with other leukemia driver genomic alterations. Like other leukemogenic drivers, it is maintained at relapse.

UBTF tandem duplication. UBTF is located at chromosome 17q21.31, and it codes for a nucleolar protein that interacts with ribosomal DNA to mediate RNA polymerase 1 ribosomal RNA transcription.[346]

UBTF genomic alterations involving heterozygous somatic variants resulting in in-frame tandem duplication of UBTF exon 13 are observed in approximately 4% of pediatric AML cases.

UBTF tandem duplication. UBTF is located at chromosome 17q21.31, and it codes for a nucleolar protein that interacts with ribosomal DNA to mediate RNA polymerase 1 ribosomal RNA transcription.[346]

UBTF-TD AML in the pediatric population primarily occurs during adolescence (median age, 12–14 years). It is also observed in adults younger than 60 years, but it is uncommon among AML in older adult patients.

UBTF tandem duplication. UBTF is located at chromosome 17q21.31, and it codes for a nucleolar protein that interacts with ribosomal DNA to mediate RNA polymerase 1 ribosomal RNA transcription.[346]

FLT3 ITD is common in cases of AML with UBTF-TD. Approximately two-thirds of cases have FLT3 ITD. In addition, approximately 40% of cases with UBTF-TD AML have WT1 variants.

UBTF tandem duplication. UBTF is located at chromosome 17q21.31, and it codes for a nucleolar protein that interacts with ribosomal DNA to mediate RNA polymerase 1 ribosomal RNA transcription.[346]

In the AAML1031 clinical trial, EFS and OS rates for patients with UBTF-TD were 30% and 44%, respectively. These values were lower than those for non–UBTF-TD patients enrolled in AAML1031 (45% and 64%, respectively). Outcome for patients with UBTF-TD was similar to that for patients with KMT2A rearrangements.

UBTF tandem duplication. UBTF is located at chromosome 17q21.31, and it codes for a nucleolar protein that interacts with ribosomal DNA to mediate RNA polymerase 1 ribosomal RNA transcription.[346]

In the AAML1031 trial, co-occurrence of UBTF-TD with either FLT3 ITD or WT1 variants was associated with an inferior prognosis, compared with patients with UBTF-TD alone.

AML with CBFB::GDXY insertions. CBFB encodes the CBFB protein that is part of the multiprotein, core-binding transcription factor complex, which master regulates a gene expression program critical for hematopoiesis. CBFB is recurrently fused with MYH11 in inv(16)/t(16;16) AML.[347]

AML with CBFB::GDXY insertions. CBFB encodes the CBFB protein that is part of the multiprotein, core-binding transcription factor complex, which master regulates a gene expression program critical for hematopoiesis. CBFB is recurrently fused with MYH11 in inv(16)/t(16;16) AML.[347]

In-frame insertions in exon 3 of CBFB have been identified in about 0.4% of pediatric AML cases at diagnosis. All described insertions lead to replacement of aspartic acid at position 87 (D87) with either glycine, aspartic acid, serin, and tyrosine (GDSY) or glycine, aspartic acid, threonine, and tyrosine (GDTY).

AML with CBFB::GDXY insertions. CBFB encodes the CBFB protein that is part of the multiprotein, core-binding transcription factor complex, which master regulates a gene expression program critical for hematopoiesis. CBFB is recurrently fused with MYH11 in inv(16)/t(16;16) AML.[347]

CBFB::GDXY insertions are associated with a gene expression profile overlapping with CBFB::MYH11–expressing AML, with the exception of increased expression of stem cell genes such as HOXA cluster genes and MEIS1.

AML with CBFB::GDXY insertions. CBFB encodes the CBFB protein that is part of the multiprotein, core-binding transcription factor complex, which master regulates a gene expression program critical for hematopoiesis. CBFB is recurrently fused with MYH11 in inv(16)/t(16;16) AML.[347]

CBFB::GDXY insertions frequently co-occur with FLT3 tyrosine kinase domain (TKD) and BCOR1 variants, but lack KIT variants, which are frequently found in CBFB::MYH11 AML.

AML with CBFB::GDXY insertions. CBFB encodes the CBFB protein that is part of the multiprotein, core-binding transcription factor complex, which master regulates a gene expression program critical for hematopoiesis. CBFB is recurrently fused with MYH11 in inv(16)/t(16;16) AML.[347]

CBFB::GDXY insertions appear to be enriched among adolescents and young adults.

AML with CBFB::GDXY insertions. CBFB encodes the CBFB protein that is part of the multiprotein, core-binding transcription factor complex, which master regulates a gene expression program critical for hematopoiesis. CBFB is recurrently fused with MYH11 in inv(16)/t(16;16) AML.[347]

The impact of CBFB::GDXY insertions on patient outcomes are unclear due to a paucity of data. However, early analysis suggests that these patients may not have the same favorable outcome as patients with CBFB::MYH11 fusions.

RUNX1 variants. AML with RUNX1 variants was a provisional entity in the 2016 WHO classification. In the 5th edition of the WHO classification, it falls into the category of AML with other defined genetic alterations.[223] This subtype of AML is more common in adults than in children. In adults, the RUNX1 variant is associated with a high risk of treatment failure. A meta-analysis of outcomes for adult patients with RUNX1 variants also demonstrated high-risk disease, although this significance was lost in the context of intermediate-risk cytogenetics.[348]In a study of children with AML, RUNX1 variants were observed in 11 of 503 patients (approximately 2%). Six of 11 patients with AML and RUNX1 variants failed to achieve remission, and their 5-year EFS rate was 9%, suggesting that the RUNX1 variant confers a poor prognosis in both children and adults.[349] However, a second study in which 23 children were found to have RUNX1 variants among 488 children with AML found no significant impact of RUNX1 variants on response or outcome. Additionally, analysis identified that children with RUNX1 variants were more frequently male, adolescents, and had a greater incidence of co-occurring FLT3 ITD and other variants. However, in each of these groups, univariable and multivariable analyses found no survival differences based on the presence of RUNX1 variants.[350] Genetic variants of RUNX1 result in a familial platelet disorder with associated myeloid malignancy (FPD-MM).[223]

WT1 variants. WT1, a zinc-finger protein regulating gene transcription, is altered in approximately 10% of cytogenetically normal cases of AML in adults.[351-354] The WT1 variant has been shown in some,[351,352,354] but not all, studies [353] to be an independent predictor of worse DFS, EFS, and OS in adult patients.In children with AML, WT1 variants are observed in approximately 10% of cases.[355,356] Cases with WT1 variants are enriched among children with normal cytogenetics and FLT3 ITD but are less common among children younger than 3 years.[355,356] AML cases with NUP98::NSD1 fusions are enriched for both FLT3 ITD and WT1 variants.[308] In univariate analyses, WT1 variants are predictive of poorer outcome in pediatric patients, but the independent prognostic significance of WT1 variant status is unclear because of its strong association with FLT3 ITD and its association with NUP98::NSD1 fusions.[308,355,356] The largest study of WT1 variants in children with AML observed that children with WT1 variants in the absence of FLT3 ITD had outcomes similar to that of children without WT1 variants, while children with both WT1 variants and FLT3 ITD had survival rates less than 20%.[355]In a study of children with refractory AML, WT1 was overrepresented, compared with a cohort who did achieve remission (54% [15 of 28 patients] vs. 15%).[313]

DNMT3A variants. Variants of the DNMT3A gene have been identified in approximately 20% of adult patients with AML. These variants are uncommon in patients with favorable cytogenetics but occur in one-third of adult patients with intermediate-risk cytogenetics.[357] Variants in this gene are independently associated with poor outcome.[357-359] DNMT3A variants are virtually absent in children.[360]

IDH1 and IDH2 variants. Variants in IDH1 and IDH2, which code for isocitrate dehydrogenase, occur in approximately 20% of adults with AML,[220,361-365] and they are enriched in patients with NPM1 variants.[362,363,366] The specific variants that occur in IDH1 and IDH2 create a novel enzymatic activity that promotes conversion of alpha-ketoglutarate to 2-hydroxyglutarate.[367,368] This novel activity appears to induce a DNA hypermethylation phenotype similar to that observed in AML cases with loss-of-function variants in TET2.[366]Variants in IDH1 and IDH2 are rare in pediatric AML, occurring in 0% to 4% of cases.[220,360,369-373] There is no indication of a negative prognostic effect for IDH1 and IDH2 variants in children with AML.[220,369]

CSF3R variants. CSF3R is the gene encoding the granulocyte colony-stimulating factor (G-CSF) receptor, and activating variants in CSF3R are observed in 2% to 3% of pediatric AML cases.[374] These variants lead to enhanced signaling through the G-CSF receptor. They are primarily observed in AML with either CEBPA variants or with CBF abnormalities (RUNX1::RUNX1T1 and CBFB::MYH11 fusions).[374] In a study of 2,150 pediatric patients with AML, 35 patients (1.6%) were found to have CSF3R variants; 30 (89%) of these cases were in patients with either RUNX1::RUNX1T1 fusions (n = 18) or with CEBPA variants (n = 12).[262] Risk of relapse was significantly higher for patients with co-occurring CSF3R and CEBPA variants, compared with patients with RUNX1::RUNX1T1 fusions and CSF3R variants.[262] Although relapse rates are higher in patients with AML who have co-occurring CSF3R and CEBPA variants, OS is not adversely impacted, reflecting a high salvage rate with reinduction therapy and HSCT.[253]Activating variants in CSF3R are also observed in patients with severe congenital neutropenia. These variants are not the cause of severe congenital neutropenia, but rather arise as somatic variants and can represent an early step in the pathway to AML.[375] In one study of patients with severe congenital neutropenia, 34% of patients who had not developed a myeloid malignancy had CSF3R variants detectable in peripheral blood neutrophils and mononuclear cells, while 78% of patients who had developed a myeloid malignancy showed CSF3R variants.[375] A study of 31 patients with severe congenital neutropenia who developed AML or MDS observed CSF3R variants in approximately 80% of patients. The study also observed a high frequency of RUNX1 variants (approximately 60%), suggesting cooperation between CSF3R and RUNX1 variants for leukemia development within the context of severe congenital neutropenia.[376]