Clearance of RAS pathway mutations prior to allogeneic transplant in Acute Myeloid Leukemia are associated with favorable overall and leukemia free survival

Abstract Background: The prognostic impact of RAS-pathway mutations (RASm) in acute myeloid leukemia (AML) on post-allogeneic stem cell transplant (alloHCT) outcomes remains to be fully elucidated. RASm have been associated with inferior overall survival (OS) after treatment with lower-intensity therapies, including venetoclax. However, RASm are often cleared following treatment with intensive chemotherapy induction and consolidation (Ball et al Am J Hematol 2022). Among patients with MDS undergoing alloHCT, RASm detected in pre-conditioning blood samples confer a higher risk of relapse for those receiving reduced intensity conditioning. The prognostic impact of RASm clearance and persistence on outcomes post-alloHCT in AML has yet to be reported. Hence, we performed a retrospective analysis of patients with AML with RASm detected pre-HCT receiving alloHCT to describe the post-alloHCT outcomes and to identify prognostic risk factors for this population. Methods: This analysis includes adult patients with RASm AML (NRAS, KRAS, PTPN11, CBL, NF1, RIT1, FLT3, KIT gene mutations) who underwent alloHCT at City of Hope between 2018-2024. The objectives were to evaluate OS, leukemia free survival (LFS), graft versus host disease (GVHD) incidence, relapse incidence, and non-relapse mortality (NRM). Next generation sequencing (NGS) was used for the detection of RASm at diagnosis, pre-alloHCT, and post-alloHCT when available. University of Washington Flow Cytometry (UWFC) was used for Measurable Residual Disease (MRD) detection with sensitivity of 0.1%. Results: A total of 115 eligible patients were identified who underwent alloHCT during the study period. Median age at alloHCT was 54 years (range: 21-79), 51.3% of patients were male, 72% of patients had KPS >80%, and stem cell source was PBSC in 99% patients. Donor types were matched unrelated (MUD=47.8%), matched related (MRD=28.7%), haploidentical (Haplo=15.7%) and mismatched unrelated (mMUD=7.8%). Myeloablative conditioning (MAC) regimens were used in 52.2% of patients, and reduced intensity conditioning (RIC) was used in 47.8%. The MAC regimens were all total body irradiation (TBI) based. Fludarabine/melphalan was the most common RIC regimen (n=52; 45.8%). The GVHD prophylaxis was predominantly using calcineurin-based regimens (70.4%) followed by post-transplant cyclophosphamide (PTCy; 29.6%). Median duration of follow up was 3.2 years (range: 0.02-6.03). Prior to HCT the median lines of therapy was 1 (range: 1-6), and ELN 2022 AML risk stratification was 22.6% favorable, 20% intermediate, and 57.4% adverse risk category. Majority of patients underwent HCT in CR (82%) and 16% of patients underwent HCT with active disease. Pre-HCT, 66% of patients were MRD negative pre-HCT. In 30% of patients RASm were cleared pre-HCT, and 36% of patients had detectable RASm pre-HCT by NGS testing. The 2- and 5-year estimated OS was 56.7% and 47.7% and the 2- and 5-year LFS was 53.5% and 48.4% respectively. The estimated 2- and 5-year relapse was 23.3% and 26.6% respectively. Adverse risk per ELN was associated with significantly decreased OS (HR 2.92; p=0.012) and LFS (HR 3.18; p=0.010). Patients who did not clear RASm pre-HCT had significantly worse OS (HR 2.12; p=0.007) and LFS (HR 2.72; p = <0.001), and a higher incidence of relapse (HR 7.98; p=0.003). When comparing NRAS- and KRAS-positive patients, no significant differences in outcomes were observed. Additionally, there was no difference by conditioning regimen (MAC vs. RIC) in 2- and 5-year estimates of OS/LFS in patients who had MRD by NGS or UWFC. In univariate analysis the following risk factors were associated with a significant increase in relapse incidence: ELN adverse risk (HR 4.12; p=0.0497), presence of RASm pre-HCT (HR 7.99; p=0.005), or post-HCT (HR 11.8; p=0.0004) and active disease at time of HCT (HR 3.28; p=0.002) had an increased risk of relapse. In multivariate analysis, the presence of pre-HCT RASm was associated with increase in relapse post HCT (HR 4.80; p=0.039). Conclusion: In patients with RASm, alloHCT results in 5-year OS/LFS of 47% and 48% respectively. Persistence of RASm pre-HCT was associated with decreased LFS, decreased OS, and increased relapse rates. TBI based MAC regimens were unable to overcome the adverse prognosis conferred by persistence of RASm pre-HCT. Better eradication of RASm may be essential to improve outcomes.