Mutation landscape and pathway-level disparities in Black versus White patients with acute myeloid leukemia.

Abstract

6534 Background: Racial disparities in acute myeloid leukemia (AML) outcomes are well-documented, yet race- specific genomic differences remain incompletely characterized. Current risk stratification systems, developed from European-ancestry cohorts, may inadequately classify patients (pts) of African ancestry (Stiff, Nature Genetics , 2024). We performed comprehensive clinicogenomic analysis to identify race-specific mutation patterns and their clinical impact in AML. Methods: We analyzed genomic and outcomes data on AML pts in AACR Project GENIE database. Mutation frequencies, co-mutation patterns, and pathway alterations were compared between racial groups. Survival analyses used Cox proportional hazards models adjusted for age and sex. Multiple comparison correction used Benjamini-Hochberg false discovery rate (FDR). Results: Our study included 2,359 pts (Black, n=243; White, n=2,116). Black pts demonstrated a distinct mutational landscape with fewer splicing factor ( SRSF2, U2AF1, SF3B1 ) mutations (11.1% vs 20.1%, FDR p=0.003) and tumor suppressor ( TP53, WT1 ) mutations (11.9% vs. 18.2%, FDR p=0.038). Signaling/RAS pathway mutations were nominally higher in Blacks. A notable finding was the emergence of MDS-related gene mutations in Black pts at a younger age, suggesting distinct disease biology. Among pts aged <40 years, ASXL1 mutations were more prevalent in Blacks (10.2% vs. 1.8%, p=0.006). Co-mutation architectures differed substantially by race: in Black pts, ASXL1 and RUNX1 frequently co-occurred with RAS/MAPK pathway mutations (ASXL1+RAS: 7.8% vs. 3.4%, FDR p=0.006), whereas White pts had predominant co-occurrence of ASXL1 and RUNX1 with splicing factor mutations Critical survival disparities emerged by mutation status. NPM1+FLT3 co-mutation was less common in Blacks (3.3% vs. 7.0%, FDR p=0.028) and conferred markedly inferior survival when present (median OS 6.7 vs. 14.6 months, p=0.008). Strikingly, favorable-risk mutations showed race-specific effects. NPM1-only (HR 0.79, p=0.014), IDH1-only (HR 0.68, p=0.005), and NPM1+IDH1 (HR 0.61, p=0.013) were protective in White but not in Black pts: NPM1-only (HR 0.89, p=0.76), IDH1-only (HR 0.84, p=0.71), NPM1+IDH1 (HR 0.34, p=0.13), challenging current risk stratification's universal applicability. Overall survival (OS) was similar (median OS- 18 vs. 16.7 months), though propensity-matched analysis trended toward worse outcomes in Blacks (18 vs. 25 months, p=0.091). Conclusions: Black AML pts exhibit fundamentally distinct molecular biology with differential pathway involvement, earlier MDS-related mutations, unique co-mutation patterns, and loss of prognostic value for traditionally favorable mutations. These findings demonstrate that race-blind risk stratification may systematically misclassify Black pts, supporting urgent need for ancestry-informed precision medicine approaches in AML.