Impact of RAS mutations on outcomes in newly diagnosed AML patients

Abstract

Abstract Introduction: Mutations in RAS oncogenes (NRAS and KRAS) are present in 15–20% of patients (pts) with acute myeloid leukemia (AML). Until the recent development of 4-gene prognostic model (mPRS/ELN'24; Dohner et al, 2024) for pts treated with hypomethylating agent/venetoclax (HMA/Ven), RAS mutations (mut) have not been included in AML risk classification systems (i.e. ELN'22). In this retrospective analysis we evaluated the impact of RASmut on outcomes of pts treated with modern intensive chemotherapy regimens as well as those treated with HMA/Ven. Methods: Pts with newly diagnosed RASmut AML treated at our institution from March 2015 to May 2025 were identified. RAS wild type (wt) AML pts records served as internal control. Composite complete remission (CRc) was defined as CR + CR with incomplete (CRi) + CR with partial hematologic recovery (CRh). The Kaplan-Meier method and log-rank tests were used to estimate and compare the median overall survival (mOS). Results: A total of 234 newly diagnosed AML pts who had NGS testing at diagnosis were included in the analysis. 48 (20.5%) had RASmut at the time of diagnosis. The median age was 56.7 years (y), 30 (62.5%) were male. NRAS was present in 27 (56.3%) pts, most common variants were: G12D (55.5%) and G13D (22.2%). KRAS was present in 16 (33.3%) pts, most common variants were: G12D (31.2%), G13D (25%), and Q61H (18.8%). 5 (10.4%) pts had concurrent NRAS/KRAS mutations. 41.7% of patients had at least one myelodysplasia (MDS) related mutation (ELN'22); other co-mutations present in pts were: NPM1 (22.9%) and FLT3-ITD (14.6%) and TP53 (10.4%). 33 (68.8%) RASmut pts were treated with intensive chemotherapy (IC). IC-RASmut patients were younger than IC-RASwt (n=95) pts, median age 46.7y vs. 56.2y (p<0.05), respectively. RASmut pts were equally distributed between all three ELN'22 risk cohorts, 11 (33.3%) pts in each. CRc rate in the IC-RASmut cohort was 57.6%, compared to 66.3% in the RASwt cohort, (p=0.4). The mOS of IC-RASmut and RASwt cohorts was comparable, 19.9months (m) vs. 20.8m (p=0.94), respectively. NRASmut pts had numerically worse mOS compared to KRASmut and NRASmut/KRASmut (19.9m vs. 39.4m vs. 37.7m, p=0.56). There was no difference in mOS between IC-RASmut and IC-RASwt pts within individual ELN'22 categories. 15 (31.3%) RASmut pts were treated with HMA/Ven. The median age was 72.9y, similar to 72.5y in the HMA/Ven-RASwt (n=91) cohort, (p=0.51). 4 (26.7%) HMA/Ven-RASmut pts had concurrent TP53mut and thus were classified as mPRS-low benefit (LB). The CRc rate of HMA/Ven-RASmut pts was 33.3%, compared to 53.8% in HMA/Ven-RASwt cohort, (mPRS-high benefit (HB): 59.3%, FLT3-ITDmut: 58.3%, mPRS-LB: 40%), (p=0.17). Consistent with mPRS/ELN'24, the mOS of intermediate benefit (IB) cohort was significantly lower than that of HB cohort (7.5m vs. 12.5m, p=0.044), and significantly higher than that of LB cohort (7.5m vs. 4m, p=0.023). There was no mOS difference between RASmut and FLT3mut pts within the mPRS-IB. Secondary or therapy related (s/t) AML (n=18) represented 37.5% of all RASmut AML. There were numerically more s/t-AML pts in the HMA/Ven cohort than in the IC cohort, (53.3% vs. 30.3%, p=0.2). Within the IC cohort, mOS of the s/t-AML was significantly shorter than that of de-novo AML (6.7m vs. 29.8m, p=<0.01), but there was no difference in the mOS with respect to RASmut status, (RASmut 8.3m vs. RASwt 6.4m, p=0.6). However, within s/t-AML pts treated with HMA/Ven, those with RASmut had significantly worse survival than those without (2.8m vs. 10.1m, p=0.01), despite none of the pts having prior HMA exposure and equal proportions of MDS related and TP53 mutations within the RASmut and RASwt cohorts. Monocytic differentiation (MoD) was seen in 62.5% of all RASmut pts, with similar proportions within IC and HMA/Ven cohorts. The CR rates and the OS for pts with and without MoD were comparable, irrespective of RASmut status and treatment category. Conclusions: In this contemporary single institution cohort, RASmut were seen in 20% of newly diagnosed AML. A majority of pts were young and were treated with IC. In intensively treated pts, RASmut status was not predictive of outcomes. In HMA/Ven RASmut pts the mOS was consistent with the mPRS/ELN'24 model. HMA/Ven treated RASmut s/t-AML pts had worse outcomes than RASwt counterparts highlighting low efficacy of HMA/Ven in this difficult to treat population, and highlighting a need for a new standard for mPRS-LB pts.