Non-T-Depleted Haploidentical Transplantation Compared to Allogeneic Transplantation from Unrelated and Matched Sibling Donors in Patients with Secondary Acute Myeloid Leukemia in First Complete Remission: A Study from the ALWP/EBMT

Background: Allogeneic hematopoietic stem cell transplantation (HSCT) is a curative therapeutic option for secondary acute myeloid leukemia (sAML. Non-T-cell depleted haploidentical stem cell transplantation (HaploSCT) may serve as the preferred therapeutic option (J Hematol Oncol 2023, 16(1):106).

Methods: The study aimed to compare outcomes of HaploSCT to those of HSCT from a matched sibling donor (MSD) or matched (9-10/10) unrelated donor (MUD) performed between 2010 and 2022 in patients (pts) with sAML in first complete remission (CR1). Statistical tests included a multivariate analysis (MVA) using a Cox proportional-hazards regression model for main outcomes.

Results: A total of 3862 pts were included, HaploSCT-643, MSD-715, and MUD-2504. Median follow-up was 3.3 (interquartile range [IQR], 3.1-3.5) years. HaploSCT were performed more recent - 2019 (range, 2010-2022) compared to the other two groups, both 2017 (range, 2010-2022), (p<0.001). Median pt age was 61 (range, 19.1-75.7), 58.9 (range,18.3-73.8), and 60.9 (range, 18.5-79.5) years (p<0.001), and 60%, 57%, and 55% were male, (p<0.001). In 90.7%, 87%, and 88.6% of pts, respectively, the antecedent hematological disease was myelodysplastic/myeloproliferative disorder. The cytogenetic risk was categorized as intermediate (62%; 62%; 65%), adverse (37%; 36%; 34%), and favorable (1%; 2.3%; 1%), respectively (p=0.066) (data missing for 1011 pts). Karnofsky performance status (KPS) did not differ between the three pt groups. The incidence of female donor to male pt combination was 21%, 27%, and 11%, respectively (p<0.001). Median time from diagnosis to transplant was 5.1 (range, 0.6-17.2), 4.3 (range, 0.9-16.1), and 4.8 (range, 0.3-17.5) months, respectively, (p<0.001). Higher percentages of HaploSCT were performed with bone marrow (BM) grafts 21%, 3.2%, and 5.1%, (p<0.001), and following myeloablative conditioning 54%, 41% and 45%, (p<0.001). Post-transplant cyclophosphamide (PTCy) was the main graft-versus-host disease (GVHD) prophylaxis in HaploSCT (90%) vs. 12%, and 9.7%, respectively (p<0.001), while anti-thymocyte globulin was administered to 88% of the MSD and 90% of the MUD HSCTs. Cumulative incidence of absolute neutrophil count >0.5 x 10⁹/L at day 60 was 94% (91.8% - 95.6%), 98.9% (97.7% - 99.4%), and 97.8% (97.1% - 98.3%), respectively. Day 180 incidence of acute (a) GVHD II-IV and III-IV was 29.7%, 20.1%, 27%, and 10.6%, 5.7%, 9.1%, respectively. The 2-year (y) total and extensive chronic (c) GVHD rates were 30.9%, 33.4%, 32.5%, and 10.7%, 13.4%, 12.9%. In MVA, aGVHD II-IV and III-IV were lower in HaploSCT compared to MSD with hazard ratios (HRs) of 0.54 (IQR 0.4-0.72, p<0.001) and 0.5 (IQR 0.3-0.83, p=0.007), respectively, while total and extensive cGVHD did not differ significantly. Both aGVHD and cGVHD did not differ significantly between HaploSCT and MUD. The 2-y non-relapse mortality (NRM) was higher while relapse incidence (RI) was lower with HaploSCT compared to MSD with HRs of 0.32 (IQR 0.22-0.46, p<0.001), and 1.64 (IQR 1.24-2.17, p=0.000), respectively. The 2-y overall survival (OS) was lower with HaploSCT compared to MSD, with an HR of 0.76 (IQR 0.6-0.95, p=0.018), while leukemia-free survival (LFS) and GVHD-free/relapse-free survival (GRFS) did not differ significantly. Notably, all transplantation outcomes including RI, LFS, OS, and GRFS did not differ significantly between HaploSCT and MUD except that 2-y NRM was higher in HaploSCT compared to MUD HR=0.63 (IQR 0.47-0.85, p=0.002). Lower KPS and higher age (per 10y) were poor prognostic factors for NRM, OS, and LFS. Lower KPS was also a poor prognostic factor for aGVHD II-IV and III-IV. Adverse-risk cytogenetics was a poor prognostic factor for RI, LFS, OS and GRFS. Low-intensity conditioning was a poor prognostic factor for RI. In addition, time from diagnosis to transplant was a poor prognostic factor for aGVHD III-IV, as was female donor to male recipient for total cGVHD. Finally, the type of GVHD prophylaxis was a prognostic factor for NRM.

Conclusions: In this retrospective analysis comparing different donor types for HSCT in sAML at CR1, the best transplantation outcomes were achieved with MSD. Similar transplantation outcomes (other than a higher incidence of NRM) were observed in HaploSCT compared to MUD. Notably, HaploSCT was able to rescue ~60% of the pts with sAML. In the absence of an MSD, both Haplo and MUD are potential alternatives.