Optimizing Conditioning Intensity to Disease Burden in Adults with Acute Myeloid Leukemia Undergoing Transplantation: Results of a Prospective Multicenter Study

Background: Myeloablative conditioning (MAC) is standard in fit adults with acute myeloid leukemia (AML) undergoing allogeneic transplantation (alloHCT), albeit reduced intensity conditioning (RIC) has led to similar outcomes to MAC in patients without genetic evidence of residual disease in retrospective studies. Methods: This was a prospective multicenter study including consecutive adults and adolescents with AML from 2015-2024 who received alloHCT using peripheral blood hematopoietic cells from matched siblings (MSD) or haploidentical donors, performed in a limited-resource setting. Before 2019 patients received MAC based on cyclophosphamide 350 mg/m2, fludarabine 25 mg/m2 for 3 days (CyFlu) plus melphalan 200 mg/m2 or busulfan 9-12 mg/kg if considered fit. After 2019, those with undetectable MRD centralized and measured with Euroflow protocols could proceed to RIC regardless of age, comorbidities or fitness, using CyFlu plus melphalan 140 mg/m2 with or without total body irradiation 2Gy to reduce short-term complications. Patients with detectable (MRDpos) or active disease (>5% blasts) could proceed to sequential conditioning, MAC or RIC according to fitness. Graft-versus-host disease (GVHD) prophylaxis was based on post-transplant cyclophosphamide (PTCy) or calcineurin inhibitor (CNI) plus methotrexate. RIC-MRDneg, MAC-MRDneg, MRDpos and active disease groups were compared. The primary outcome was overall survival (OS). Secondary outcomes included, event-free survival (EFS), non-relapse mortality (NRM) and GVHD-relapse-free survival (GRFS). Results: 92 patients were analyzed. Median age was 41 years (range 15-75), n=51 were men (55.4%), 41 women (44.6%), with a median of 2 treatment lines (range 1-4), and 8.7x106 CD34 cells/kg (range 1.8-23) infused. N=54 received haploidentical grafts (58.7%) and n=38 MSD (41.3%), n=82 received PTCy (89.1%). N=39 were MRDneg and received RIC (42.4%), n=29 MRDneg and received MAC (25%), n=9 were MRDpos (9.8%) and n=15 had active disease (16.3%). Similar sex, donor type, HCT-CI and GVHD prophylaxis were observed between groups. Patients with MRDpos or active disease had ≥2 treatment lines (86.7% and 100%, respectively) than RIC or MAC-MRDneg (51.3% and 41.4%, respectively) (p=.001). Patients that received RIC-MRDneg were older than MAC-MRDneg (median of 46 vs 29 years). Statistically similar OS, EFS, NRM and GRFS rates were observed between RIC-MRDneg vs MAC-MRDneg. 2-year OS was 60.3% in RIC MRDneg and 60.5% in MAC-MRDneg, 51.9% in MRDpos and 30% in active disease. 2-year NRM was 9.6% in RIC-MRDneg vs. 11% in MAC-MRDneg, 36% in MRDpos and 31% in active disease. 2-year EFS was 38.6% in RIC-MRDneg vs. 49.3% in MAC-MRDneg, 41.7% in MRDpos and 13.3% in active disease. 2-year GRFS was 27.8% for RIC-MRDneg, 34.8% in MAC-MRDneg, 20.8% for MRDpos and 7% in active disease. GVHD rates and severities were similar. MAC vs RIC was not associated with OS or EFS in the univariable analysis. Treatment lines and disease status were the only variables associated with EFS, but not OS. In the multivariable model only MRD negativity was associated with EFS HR 0.44 (95%CI 0.21-0.95).