Conference Abstracts - 2026 National Summit on Hematologic Cancers
Vol. 6, Issue Supplement 2, 2026 · S1-2
Context-Dependent Prognostic Significance of RAS Mutations in AML and Implications for Treatment Strategy
Melanie Naranjo, MD,Jamila Mammadova, MD,Sarah Profitt, PharmD,Callee Brooks, PharmD,Kathryn Kennedy, NP,Ashwin Kishtagari, MD,Somedeb Ball, MD,Sanjay Mohan, MD,Kateryna Fedorov, MD
Submission received: 2026-08-16 / Accepted: 2026-08-25 / Published: 2026-09-25
Abstract
Background
RAS oncogene mutations (NRAS, KRAS, HRAS) are seen in 15-20% of AML. Emerging evidence suggests RAS mutations (RASmut) confer different risk depending on the therapeutic context, yet prognostic significance remains incompletely defined.
Methods
We retrospectively analyzed adults with newly diagnosed AML receiving first-line intensive chemotherapy (IC) or hypomethylating agent plus venetoclax (HMA/Ven) at Vanderbilt University Medical Center (2015 – 2025). Responses were assessed by ELN 2022 criteria; survival was estimated by Kaplan–Meier methods and compared using log-rank testing.
Results
Of 234 eligible patients, 48 (20.5%) harbored RAS mutations: 56.3% NRAS, 33.3% KRAS, co-mutation in 10.4%. RASmut patients were younger than RAS wild-type (RASwt) (57.2 vs 64.9 years, p=0.003) and more often had ECOG3 (18.8% vs 5.9%, p=0.005). Favorable-risk disease was more common in RASmut patients (25.0% vs 9.7%, p=0.005), while adverse-risk disease predominated in RASwt (68.3% vs 37.5%, p=0.0002). RASmut patients more often received IC over HMA/Ven (68.8% vs 31.3%, p=0.03). Outcomes did not differ by RAS variant type (NRAS vs KRAS), mutation hotspot, or variant allele fraction. Among IC-treated patients, median overall survival (mOS) was comparable between RASmut and RASwt (19.9 vs 23.6 mo, p=0.8); however, RASmut with adverse-risk disease was associated with inferior mOS (6.4 vs 16.9 mo, p=0.04). Among HMA/Ven-treated patients, RASmut were associated with shorter mOS (4.7 vs 9.3 mo, p=0.03), with outcomes inferior to all molecular subgroups except TP53-mutated. In secondary/therapy related AML (s/t-AML), RASmut status did not affect mOS (6.2 vs 7.1 mo, p=0.7). Yet within RASmut cohort, s/t-AML was associated with shorter OS compared to de-novo AML (7.1 vs 31.4 mo, p=0.003).
Conclusion
These data establish RASmut as a context-dependent prognostic marker, shaped by ELN risk and therapy. RASmut patients with adverse-risk disease and/or s/t-AML represent a high-risk population with limited benefit from current therapies, supporting novel therapeutic strategies and early transplant referral.