Clinico-genomic characterization of RAS-mutant acute myeloid leukemia

Somatic mutations within NRAS or KRAS are recurrent in acute myeloid leukemia (AML) and often arise as obligatory late events regarding AML ontogeny. RAS mutations have implications in solid cancers and in AML; however, their prognostic significance and codon-level characteristics are poorly understood, especially regarding response to intensive chemotherapy. There is an unmet need for targeting the RAS pathway. Herein, we performed clinico-genomic profiling of 89 patients with RAS-mutant AML, alongside 99 patients with RAS-wild-type AML. Median overall survival (OS) for RAS-mutant AML was shorter compared to RAS-wild-type AML (19.2 vs. 63.3 months, p = 0.05). For patients receiving cytarabine-based front-line chemotherapy, those with RAS mutations had shorter median OS compared to RAS-wild-type AML (27.1 vs. 122.2 months, p < 0.001). Within the RAS-mutant AML group, cytarabine-based front-line therapy resulted in longer median OS compared to front-line hypomethylating agents (27.1 vs. 13.2 months, p = 0.04). Hematopoietic cell transplantation (HCT) for RAS-mutant AML conferred longer median OS compared to no HCT (45.1 vs. 13.2 months, p = 0.004). There was no difference in survival among heterogeneous RAS-mutant subgroups (NRAS vs. KRAS vs. double-mutant) or among hotspot codons. Analysis of variant allele frequencies suggested that NRAS/KRAS mutations were subclonal. Although 80 (66.1%) of 121 total RAS mutations were found in codons G12 or G13, most substitutions were G12D or G13D, which are not targetable by commercial RAS^G12C inhibitors. This study sheds light on prognostic implications of RAS mutations and may inform extension of the therapeutic reach of RAS inhibitors to AML.