The Optimal Alternative Donor Transplant for Pediatric Patients with Acute Leukemia: A Comparison between Alfa-Beta T-Cell and B-Cell Depleted Haplo-SCT and UCBT

Children with acute leukemia (AL) in need of an allograft and lacking an HLA-identical sibling and an HLA-matched unrelated donor, could benefit either from unrelated cord blood transplantation (UCBT) or T-cell depleted haploidentical (haplo) stem cell transplantation (SCT).

Different techniques of T-cell depletion have been developed for improving immune reconstitution and for enhancing the graft-versus-leukemia (GvL) effect. Locatelli et al. (Blood 2017) reported remarkable results for children with AL given haplo-SCT after a novel method of graft manipulation based on selective, negative depletion of alfa-beta T and B cells.

On behalf of Eurocord, CTIWP, and PDWP of EBMT, we compared outcomes of patients < 18 years with AL undergoing either single-UCBT (n=444) or haplo-SCT after alfa-beta T-cell and B-cell-depletion (n=99) from 2012 to 2016 using a myeloablative conditioning regimen (MAC).

Median age at SCT was 6 (range 0.3-18) and 8 (range 0.6-18) years, and median follow-up was 36 and 14 months (p<0.05) for UCBT and haplo-SCT recipients, respectively. Compared to UCBT, haplo-SCT were performed more recently (median year 2016 vs 2012, p<0.05). Acute lymphoblastic leukemia (ALL) was the most frequent diagnosis (62% versus 55%), while 38% and 45% of patients had acute myeloid leukemia (AML) in the UCBT and haplo-SCT group, respectively. Cytogenetics was mainly at intermediate risk for UCBT (54%) and haplo-SCT (73%, p<0.05); 49% and 54% of patients were in CR1 at time of the allograft in UCBT and haplo-SCT group, respectively (p=0.45). In vivo T-cell depletion with ATG was more frequently adopted in haplo-SCT recipients (90% versus 81%, p=0.05). For UCBT recipients, HLA compatibility was 6/6 matched for 19% of patients, 5/6 for 58% and 4/6 for 23%. Cord blood graft contained a median TNC count at cryopreservation of 7.02x10e7/Kg (range 2.63-43) and at infusion of 5.3x10e7/Kg (range 0.3-39).

Conditioning regimen varied according to the type of transplant: for haplo-SCT recipients Fludarabine+TBI (28%) and Thiotepa+Busulfan+fludarabine (TBF) (24%) were the most frequent MAC. In UCBT, Busulfan+Cyclophosphamide (Cy), or Cy+Fludarabine+TBI or TBF were used in 19%, 15% and 13% of patients, respectively.

Haplo-SCT recipients did not receive any additional pharmacological post-transplant graft-versus-host disease (GvHD) prophylaxis, while CyclosporineA (CSA)+ steroids and CSA+mycophenolate mofetil were used in 38% and 32% of UCBT recipients.

The 60-day probability of neutrophil engraftment was 91% vs 93% (p<0.05) for UCBT and haplo-SCT. The cumulative incidence (CI) of grade II-IV and grade III-IV acute (a) GvHD was 38% vs 12% (p<0.05) and 17% vs 4% (p<0.05) in UCBT and haplo-SCT recipients, respectively. The 2-year CI of relapse (RI) was similar following from UCBT and haplo-SCT (22% vs 25%, p=0.75), as well as non-relapse mortality (NRM) (18% vs 10%, p=0.07), respectively. Disease recurrence was the most common causes of death in both groups, with infections and GvHD being the most frequent transplant-related fatalities.

According to donor type, no differences were found in overall survival (OS) (65% vs 68%, p=0.17) and leukemia-free survival (LFS) (61% vs 65%, p=0.24), while a lower GvHD-free, relapse-free survival (GRFS) was observed in UCBT compared to haplo-SCT (49% vs 60%, p<0.05). OS was 67% and 58% for patients with AML and ALL respectively, p=0.11.

In multivariate analysis, UCBT was associated with higher risk of NRM (HR 2.36, p<0.05), while RI was not associated with type of allograft. Higher risk of grade II-IV aGVHD was observed in UCBT (HR 2.81, p<0.05), with no difference for cGVHD among the two groups (HR 0.68, p=0.31). OS (HR 1.52, p=0.10) and LFS (HR 1.28, p=0.28) were comparable between UCBT and haplo-SCT, while GRFS resulted lower for UCBT (HR 1.58, p<0.05). No others factors were associated with outcomes.

Our results showed that RI, LFS, OS and cGVHD were comparable between UCBT and haplo-SCT recipients, while better NRM, aGVHD and GRFS were observed in patients transplanted from a family donor. These data confirm that alfa-beta T-cell and B-cell depleted haplo-SCT is a valid alternative for patients in need of an urgent allograft. Costs of manipulation in the haplo-SCT setting and costs of cord blood procurement should be further investigated. Technical expertise with this refined approach of graft manipulation and center experience are important to optimize patient's outcome.