Intermediate Vs High Dose Busulfan-Based Conditioning for Allogeneic Cell Transplantation in Patients with Acute Leukemia or Myelodysplastic Syndromes from HLA Matched Related or Unrelated Donors: Achieving the Same with Less

Introduction:

With the improvement in transplant practices during recent years, non-relapse mortality (NRM) following allogeneic hematopoietic cell transplantation (allo-HCT) has significantly improved. The use of reduced intensity conditioning has extended allogeneic transplants to older patients and those with comorbidities that preclude traditional myeloablative conditioning (MAC) albeit at the cost of higher rates of relapse. Busulfan (BU) has been used for approximatively four decades as a major component of chemotherapy-based conditioning before allo-HCT. At our center, IV-BU dose in MAC is 12.8 mg/kg (BU12.8) and in those not eligible for MAC, it is 9.6 mg/kg (BU9.6). The goal of BU9.6 is to improve tolerability without increasing disease recurrence. The choice of BU12.8 or 9.6 is determined by consensus based on patient age, Karnofsky score and comorbidities.

Aims:

The aim of this study was to compare the impact of BU9.6 to BU12.8 on outcomes in patients with acute leukemia (AL) or myelodysplastic syndromes (MDS) who received allo-HCT after BU-based conditioning from matched related or unrelated donors.

Patients and methods:

This study included 181 patients: 134 AL [122 (67%) AML, 10 (6%) ALL and 2 (1%) biphenotypic AL] and 47 (26%) MDS patients who received allo-HCT between Jan 2012 and Dec 2018 at the Ottawa Hospital. Median age at allo-HCT was 58 years (range: 18-73), 104 (57%) were males. Median time from diagnosis to transplant was 5 months. At time of conditioning, among AL patients, 99 (74%) were in CR1, 19 (14%) in CR2 and 16 (12%) not in CR, and among MDS patients, 22 (47%) were not treated and 25 (53%) had received azacytidine. All patients received PBSC from HLA matched related [N=68 (38%)] or unrelated [N=113 (62%)] donors. Patients were classified according to the refined Disease Risk Index (DRI): 94 (52%) intermediate, 81 (45%) high and 6 (3%) very high. Karnofsky score was <80% [N=22 (12%)], 80% [N=100 (55%)], >80% [N=59 (33%)]; HCT comorbidity index was 0 [N=70 (39%)], 1 [N=25 (14%)], 2 [N=24 (13%)] and >2 [N=62 (34%)]. GVHD prophylaxis consisted of tacrolimus and short course methotrexate for all patients. Rabbit antithymocyte globulin (rATG, Thymoglobulin) was used in all transplants from unrelated donors (n=113) and in BU12.8 transplants from related donors (n=28). rATG dose was 2.5 mg/kg in 85 (47%) patients and 4.5 mg/kg in 56 (31%). Bu dose was 9.6 mg/kg for 74 (41%) patients [Bu/Flu (N=73) and Bu/Flu/TBI (N=1)], and 12.8 mg/kg for 107 (59%) [Bu/Cy (N=79), Bu/Flu (N=19) and Bu/Flu/TBI (N=9)].

Results:

Ninety-eight % of the patients engrafted; ANC>500/µL was achieved at a median of 16 days (range: 7-79) post-transplant and did not differ between BU9.6 and BU12.8. GVHD requiring systemic immunosuppressive therapy was observed in 26/74 (35%) patients (4 acute, 22 chronic) after BU9.6, and in 45/107 (42%) patients (13 acute, 32 chronic) after BU12.8. With a median follow-up of 22 months (range: 5-74) for survivors, no significant difference in survival, relapse incidence (REL) or NRM was observed between BU9.6 and BU12.8. Two-year OS, REL and NRM for the whole population were 65% (95%CI: 57-73), 29% (95%CI: 22-37) and 10.8% (95%CI: 6-16) respectively. When stratified according to DRI: intermediate, high and very high scores had 2-year OS of 76% (95%CI: 65-84), 53% (95%CI: 38-66) and 0% respectively, p=0.001; and 1-year REL rates of 12% (95%CI: 6-20), 31% (95%CI: 21-42) and 67% (95%CI: 12-92) respectively, p<0.001. In multivariable analysis adjusted for conditioning (BU9.6 vs BU12.8), DRI, the use of rATG and rATG dose, the only factors that significantly impacted OS were DRI [high-very high vs intermediate: HR=1.98 (95%CI: 1.15-3.4), p=0.013] and the use of ATG [No ATG vs ATG 4.5mg/kg: HR=2.2 (95% CI: 1.06-4.6), p=0.033], while only DRI [high-very high vs intermediate: HR=3.56 (95%CI: 1.86-6.8), p=0.0001] impacted on REL.

Conclusions:

Our results validate the efficacy of a reduced 9.6 mg/kg BU-based conditioning regimen for patients not suitable for traditional 12.8mg/kg dose. Relapse was not increased, unlike what has been reported with BU6.4mg/kg. While NRM was not decreased, the BU9.6 group was older and had more comorbidities. In addition, we confirm the value of DRI in prognosticating both relapse and survival. We recommend the use of BU9.6 in patients not suitable to receive BU12.8. Further studies are needed to determine whether the BU9.6 can replace BU12.8 for all patients.