Background: According to the revised WHO classification of 2016, AML with mutated RUNX1 constitutes a new provisional entity. We previously reported that the subgroup of RUNX1-mutated AML with RUNX1 wild-type (WT) loss is associated with a distinct pattern of cytogenetic and molecular genetic abnormalities and with an adverse prognosis. However, the impact of multiple RUNX1 mutations is unclear yet.

Aims: Evaluation of associated genetic alterations and prognosis in AML with >1 RUNX1 mutation as compared to those with (1) RUNX1 WT loss and (2) one RUNX1 mutation.

Patient cohorts and methods: The total cohort comprised 467 AML cases with RUNX1 mutations (mut) (296 male, 171 female). Median age was 72 years (range: 18-91 years). All patients were investigated using chromosome banding analysis (CBA) and amplicon sequencing of RUNX1. The cohort was split into the subgroups with RUNX1 WT loss (UPD or RUNX1 deletion detected by genomic arrays) (n=50), cases with >1 RUNX1mut (n = 94) and cases with 1 RUNX1mut and conservation of the RUNX1 WT allele (n = 323). Of these, 50, 55 and 58 cases, respectively, were selected for further mutation analyses of ASXL1, BCOR, CBL, CEBPA, DNMT3A, ETV6, EZH2, FLT3-ITD, FLT3-TKD, GATA2, IDH1, IDH2, KIT, KRAS, MLL-PTD, NPM, NRAS, SETBP1, SF3B1, SRSF2, TET2, TP53, U2AF1 and WT1. Variants of unknown significance were excluded from statistical analysis.

Results: In the total cohort of 467 cases, CBA revealed a normal karyotype (NK) in 53% of patients, 18% harbored trisomies, 3% showed a complex karyotype (>3 abnormalities), 26% other aberrations. The proportion of cases with trisomies was largest in cases with RUNX1 WT loss (26%), followed by >1 RUNX1mut (19%) and 1 RUNX1mut (16%). In more detail, in the total cohort, 56% of cases with trisomies harbored +8, in 30% +13 was found. A similar pattern was observed for 1 RUNX1mut, whereas in cases with WT loss +13 was the most abundant trisomy (+8: 66% in 1 RUNX1mut vs. 31% in WT loss, p=0.022; +13: 15% vs. 62%, p<0.001). Cases with >1 RUNX1mut showed an intermediate distribution (+8: 44%, +13: 50%).

Missense mutations were the most abundant mutation type (53%) inWT loss cases, followed by frameshift mutations (28%). By contrast, in cases with 1 RUNX1mut, frameshift mutations were found more frequent (45%, p=0.016), whereas missense mutations were detected at a frequency of 31% (p=0.006). In cases with >1 RUNX1mut, both were observed at similar frequencies (missense: 36%, frameshift: 38%).

Mutation analyses of 163 selected cases revealed SRSF2 (39%), ASXL1 (36%), DNMT3A (19%), BCOR (18%), IDH2 (17%), SF3B1 (17%) and TET2 (17%) mutations as most frequently mutated genes in the total cohort. Cases with RUNX1 WT loss showed a higher frequency of ASXL1mut compared to the other cases (50% vs. 29%, p=0.013), while U2AF1mut were absent from this group (0% vs. 15%, p=0.019). Differences between cases with and without RUNX1 WT loss were also detected for DNMT3A, TET2, SF3B1 (more abundant in WT loss) and IDH2, WT1 (less abundant in WT loss), although these were not statistically significant. For many genes, the group of cases with >1 RUNX1mut showed an intermediate abundance pattern, or mutation frequencies similar to cases with 1 RUNX1mut were observed. Mutations in spliceosome genes (SF3B1, SRSF2, U2AF1, ZRSR2) were very abundant in all subgroups but less frequent (although not statistically significant) in cases with 1 RUNX1mut.

Median overall survival (OS) in the total cohort was 14 months. WT loss (OS: 5 months) and >1 RUNX1mut (14 months) showed an adverse impact on prognosis compared to 1 RUNX1mut (22 months; p=0.002 and p=0.048, respectively). Mutations in ASXL1 and KRAS showed a negative impact on OS in the total cohort (10 vs. 18 months, p=0.028; 1 vs. 15 months, p<0.001), whereas DNMT3A mutations negatively affected OS in WT loss only (in WT loss: 1 vs. 7 months, p=0.005).

Conclusion: 1) RUNX1 mutated AML cases with WT loss show a high frequency of +13, RUNX1 missense mutations and accompanying ASXL1 mutations. 2) They clearly separate from cases with 1 RUNX1mut, which depict a predominance of +8, RUNX1 frameshift mutations and mutations in IDH2 and WT1. 3) Patients with >1 RUNX1mut show a pattern intermediate between the former two subgroups. 4) Loss of RUNX1 WT and > 1 RUNX1mut showed an adverse impact on OS. Thus, not only the presence and number of RUNX1mut but also the conservation of an intact RUNX1 allele is biologically and clinically relevant.

Disclosures

Stengel:MLL Munich Leukemia Laboratory: Employment. Kern:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Meggendorfer:MLL Munich Leukemia Laboratory: Employment. Perglerová:MLL2 s.r.o.: Employment. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution