Unmanipulated haplo-identical donor transplantation compared with identical sibling donor had better anti-leukemia effect for refractory/relapsed acute myeloid leukemia not in remission status.

Prior studies have suggested that for leukemia patients with high-risk features, haplo-identical-hematopoietic stem cell transplantation (HID-HSCT) has a stronger anti-leukemia effect compared with HSCT using an identical sibling donor (ISD-HSCT). However, it is unclear whether an HID-HSC transplant also augments the graft-versus-leukemia (GVL) effect among refractory/relapsed (R/R) acute myeloid leukemia (AML) patients who are not in remission (NR). We conducted a retrospective analysis of 124 R/R AML patients with NR status who underwent HID-HSCT between April 2012 and December 2016 and compared these to 27 R/R AML patients who underwent an ISD-HSCT within the same timeframe. Among all of the patients, 68 (45.0%) had primary induction failure (PIF) and 83 (54.9%) were relapsed and had failed to respond to at least one cycle of salvage combination chemotherapy. Myeloablative conditioning regimens were administered to all patients. Here, we present a retrospective multivariate analysis of pre-transplantation risk factors and characteristics of all 151 patients and developed a predictive scoring system to predict patient survival. The median period of follow-up was 46 months for all patients. The HID cohort had a higher 5-year overall survival (OS) compared with the ISD cohort (48.6% ± 4.6% vs 25.9% ± 8.4, respectively; P = 0.017) and higher LFS (leukemia-free survival) (41.6% ± 7.5% vs 25.9% ± 8.4%, respectively; P = 0.019). There was no difference in the 5-year cumulative incidence of non-relapse mortality (NRM) (18.0% ± 3.8% and 34.9% ± 12.6%, respectively; P = 0.212) between the two group. However, the 5-year cumulative incidence of relapse (CIRs) was lower in the HID group compared with the ISD group (55.4% ± 8.9% vs 67.3% ± 9.9%, respectively; P = 0.021). Multivariate analysis showed three risk factors associated with OS and LFS: (1) ISD-HSCT, (2) use of a standardized conditioning regimen, and (3) less than 50% proportional reduction of blast cells in the bone marrow (BM). Based on these three risk factors, we developed a predictive scoring system for R/R AML patients undergoing HSCT. Patients who had a predictive score of 0 and 1 had a 66.6% ± 4.5% and 44.1% ± 3.6% OS rate at 5 years, respectively. Patients with a score ≥ 2 had only a 4.4 ± 0.2% OS rate at 5 years. An HID-HSCT had a better anti-leukemia effect among R/R AML patients with an NR status compared with an ISD-HSCT. We also identified pre-transplantation risk factors to delineate subgroups that could derive maximal benefit from HSCT.