Comparable Outcomes After Alternative and Matched Sibling Donor Hematopoietic Stem Cell Transplantation and the Role of Molecular Measurable Residual Disease for Acute Myeloid Leukemia in Elderly Patients.

Allogeneic hematopoietic stem cell transplantation (HSCT) remains the most effective postremission therapy conferring the chance of cure for acute myeloid leukemia (AML), including elderly patients. Although the number of transplantations for elderly patients with AML (eAML) is increasing owing to greater availability of various graft sources together with the adoption of advanced supportive care and reduced-intensity conditioning (RIC) regimen, there are relatively limited data on the impact of donor type in eAML compared to younger patients. In addition, few studies have evaluated the role of pretransplantation measurable residual disease (MRD) in the elderly population. Given the lack of prospective comparative study, we retrospectively compared transplantation outcomes of elderly patient with AML receiving allo-HSCT from matched sibling donor (MSD-HSCT), matched unrelated donor (MUD-HSCT) or haploidentical related donor (Haplo-HSCT), or autologous HSCT (Auto-HSCT). A total of 154 patients with a median age of 63 years (range 60-74) underwent MSD-HSCT (n = 41), MUD-HSCT (n = 36), Haplo-HSCT (n = 55), or Auto-HSCT (n = 22) for AML. RIC regimens were used in the majority of patients. In Haplo-HSCT, T-cell-replete peripheral blood stem cells with unique RIC regimens using anti-thymocyte globulin (ATG)-based GVHD prophylaxis was used. In the analysis, adjustment for MRD status at the time of transplantation was performed. MRD was measured by the quantitative molecular assays of the targets, including RUNX1-RUNX1T1, CBFB-MYH11, and NPM1, or WT1 in the absence of abnormalities in the aforementioned targets. At a median follow-up of 48 months, survival rates were similar between different donor types, whereas nonrelapse mortality (NRM) was lower in MUD-HSCT compared to MSD-HSCT (P = .002). MSD-HSCT, in which the majority of patients received a conditioning regimen not including ATG, showed more frequent severe chronic graft-versus-host disease (cGVHD). The major causes of non-relapse deaths in MSD-HSCT were related to cGVHD (71%), whereas infectious complications were mainly related to NRM in Haplo-HSCT (50%) or Auto-HSCT (100%). In the MUD-HSCT, GVHD (57%) and infection (43%) contributed similarly to non-relapse death. Cytomegalovirus infection was more frequent in Haplo-HSCT. In multivariate models, pre-transplant MRD-positivity was an independent risk factor for relapse (P = .001), whereas older age (P = .002) and the hematopoietic cell transplantation-comorbidity index (P = .009) were useful in predicting NRM. The current study demonstrated comparable outcomes after alternative and matched sibling donor HSCT in eAML aged 60 years or older, and the results also suggest the necessity for more sophisticated strategies to reduce NRM or relapse according to each donor type. The usefulness of molecular MRD assays demonstrated herein will facilitate trials for MRD-driven decision-making or risk-adaptive approaches in eAML.