RUNX1 mutations and outcomes in Acute Myeloid Leukemia: A public data-based analysis framing racial disparities in genomic risk

Abstract

Abstract Background: RUNX1 mutations are established high-risk features in acute myeloid leukemia (AML), associated with poor response to induction therapy and reduced overall survival (OS). Current prognostic models, including the ELN 2022 classification, incorporate RUNX1 mutation status but are based largely on cohorts of European ancestry. This raises concerns about their generalizability to underrepresented populations, particularly African American patients, who continue to experience worse outcomes in AML across treatment settings. Methods: We analyzed publicly available datasets, including TCGA-LAML, Beat AML, and a recent multi-omic study of 100 African American AML patients (Stiff et al., Nat Genet 2024). We evaluated RUNX1 mutation frequency, OS outcomes, and co-mutation patterns, stratified by race. To contextualize survival differences by ancestry, we performed a survival analysis of adult AML patients using the SEER 17 registries (2010–2020), accessed through SEER*Stat. Kaplan-Meier estimates and median OS were compared by race using disease-specific survival. Results: RUNX1 mutations were present in approximately 10–15% of AML cases and consistently associated with inferior OS (median 16–24 months) in both de novo and AML-MRC subsets. Co-mutations in DNMT3A and TET2 further worsened the prognosis. In the African American cohort described by Stiff et al., the frequency of RUNX1 mutations was similar to that in White patients; however, African American patients showed enrichment for IDH2, RAS, and WT1 mutations and lower incidence of NPM1 mutations. Despite comparable mutation rates, African American patients demonstrated significantly worse OS, even under standardized treatment. SEER 17 survival analysis of 5,932 AML cases (683 African American) confirmed shorter median OS in African American patients compared to non-African American patients (10.4 vs. 15.7 months, p<0.01).Conclusions: RUNX1 mutations confer adverse prognosis in AML, but current risk models may not adequately reflect risk in African American patients due to differing co-mutation patterns and underrepresentation in genomic datasets. These findings highlight the need for ancestry-aware risk stratification and inclusive genomic profiling to reduce disparities in AML outcomes. References: Lew-Derivry et al. Leukemia. 2023 Jalili et al. PMCID: PMC5980915 Stiff et al. Nat Genet. 2024 Bhatnagar et al. PMCID: PMC7933110 Tang et al. Blood. 2009 Mendler et al. JCO. 2012 He et al. Scand J Clin Lab Invest. 2020 Döhner et al. Blood. 2022. PMID: 35926292