Implications of RAS Mutational Status in Subsets of Patients with Newly Diagnosed Acute Myeloid Leukemia (AML) across Therapy Groups

Introduction
Activating mutations in RAS have been reported in about 10-15% of patients with AML. Previous studies have not identified a prognostic significance for RAS mutations in AML treated with conventional chemotherapy. However, RAS mutations have emerged as a potential mechanism of resistance to treatment with small molecule inhibitors of FLT3, IDH, and BCL2. With broader availability of next generation sequencing, and the identification of wider heterogeneity in AML, we aimed to better characterize the genomic landscape of RAS mutated AML and outcomes within various subsets.

Methods
We conducted a retrospective analysis in 1410 consecutive patients with newly diagnosed (ND) AML) treated at out institution from 12/2011 to 01/2020, who had a baseline genomic testing available. Patients were treated with a variety of therapies, classified as: HMA-alone, HMA+Venetoclax (Ven), Intensive chemotherapy (HiDAC based), Intensive+Ven, low intensity double nucleoside analogue (cladribine+LDAC; Clad/LDAC), and Clad/LDAC + Ven. Response rates and outcomes were analyzed by age, cytogenetic prognostic subgroup (favorable [FAV], diploid [DIP], other intermediate [INT], adverse [ADV]), and type of therapy in the context of mutated (RAS-mut) or wild type RAS (RAS-WT). We sought to characterize the impact of RAS mutations on AML outcomes in the era of targeted therapy.

Results
Baseline patient and disease characteristics are detailed in table 1. Among 1410 patients with ND AML, 273 (20%) were RAS-mut; of these, 196 (72%) had NRAS-mut, 47(17%) had KRAS-mut, and 29(11%) had both. Patients with RAS-mut were younger(median age 62 vs. 66 years; P=0.001) and had higher a median WBC (P=0.02) and platelet count (P=0.01) at diagnosis compared to those with RAS-WT. Among the cytogenetic groups: FAV, INT, DIPLOID, ADV, RAS-mut were present in 39%, 20%, 16%, and 6%, respectively. Patients with RAS-mut were more likely to have concomitant mutations in ASXL1 (13%; P=0.03), , RUNX1 (10%; P=0.03), and less likely to have concomitant mutations in JAK2 (1%; P=0.03) and TP53 (8%; P<0.01). Outcomes by subset and by therapy group are summarized in Table 2. Although there was a trend in favor of RAS-mut, there was no significant difference in OS among pts younger or older than 60 yrs of age by RAS mutation. There was no difference in OS by NRAS or KRAS mutation (median OS 16 vs. 16.5 m; P=0.45). Pts with RAS-mut AML classified as non-secondary/de novo AML (P=0.003), therapy-related AML (P=0.02), and those who received intensive therapy had a significantly better OS compared to RAS-WT. The addition of Ven to each therapy group was associated with higher response rates and median OS but were not statistically significant (Figure 1). Response rates among patients with or without RAS mutations were similar, with a trend for better responses with araC-based therapy, compared to non-araC based. Compared to RAS-WT, RAS-mut was associated higher response among pts with diploid karyotype, de novo AML, and those treated with intensive therapy. Among pts with MLL-rearrangements (P=0.09) and MECOM translocation (P=0.06), RAS-mut was associated with lower response rates compared to RAS-WT. Among patients with TP53 mutated AML, co-mutation with RAS was associated with worse outcomes compared to RAS-WT (CR rates of 4% vs. 16% P<0.001; median OS 3.4 vs. 5.8 months, P=0.77).

Conclusion
RAS mutations are present across AML subsets, with higher representation among pts with FAV cytogenetics and relative underrepresentation among pts with ADV karyotype and TP53-mutations. Among patients with DIP cytogenetics, de novo AML, and those treated with intensive therapy, RAS-mut are associated with higher response rates and more favorable outcomes. Early observations suggest modest improvements in ORR with the addition of Ven to chemotherapy but no OS benefit. Targeting the RAS pathway among specific AML subsets remains an important area of further research.