Treosulfan Compared to Busulfan in Allogeneic Haematopoietic Stem Cell Transplantation for Myelofibrosis: A Registry-Based Study from the Chronic Malignancies Working Party of the EBMT

Background

The optimal conditioning prior to allogeneic stem cell transplantation (HSCT) for myelofibrosis (MF) remains an ongoing matter of debate. There is a paucity of randomized trials robustly showing superiority of one over another. Registry studies comparing reduced intensity conditioning (RIC) versus myeloablative conditioning (MAC) regimens did not show significant differences in overall survival (OS). Recently, a multi-center phase 3 trial in MDS-AML reported an advantage of fludarabine-treosulfan (TREO) over fludarabine-busuflan (BU) RIC (Lancet Haematol Beelen et al 2020). The role of this large EBMT registry study was to compare outcomes of MF HSCT following TREO and BU based MAC or RIC regimens.

Methods

Patients with MF transplanted from a T-cell replete graft between 2010-2018 following BUMAC, BURIC or TREO were identified in the EBMT registry. Patients who were transformed into AML were excluded. Centers were contacted to participate in additional data collection. 63 centers included a total of 536 patients of whom 6 patients were excluded due to missing data for major endpoints (relapse or regimen). The role of TREO was tested in multi-variable models adjusted on major prognostic factors especially those who were not well balanced between groups and using multiple imputation for missing values.

Results

Most TREO (n=73) patients received it in combination with FLUDARABINE (95%) and most BU (n=457) received FLUDARABINE +/- other (93%). Median total dose received of TREO was 42 g/m2 and median dose of BU was 8mg/kg. Main differences between the 3 groups were: 1) Recipient age; patients were the oldest in the RIC group (median age 61 years, (IQR: 55-65), the youngest in BUMAC (56 years, (IQR: 51-62)), and intermediate in the TREO group (59 years, (IQR: 54-64); 2) Performance status; Karnofsky score was < 80 in 20%, 30%, 42% in the TREO, BUMAC and BURIC groups, respectively; 3) Comorbidity score (HCT-CI) was lowest in the MAC group (HCT-CT ‘low’ in 56%, 45% and 37% of BUMAC, BURIC and TREO groups); 4) Disease status; patients in the TREO group were more often transplanted with progressive disease (58% vs 30% in BUMAC and 34% in BURIC); 5) JAK inhibitor prior to HSCT; a higher proportion of patients in the TREO group were treated with a JAK inhibitor before HSCT (61% vs 50% in BUMAC and 42% in BURIC); 6) Cytomegalovirus (CMV) donor/ recipient status; CMV combinations were more frequently donor recipient -/- in the TREO cohort (41% vs 28% in MAC and 36% in RIC); 7) an HLA matched sibling donor was less frequently used in TREO group (18% vs. 28% in BURIC and 36% in BUMAC). The 3 groups were well balanced regarding period of transplantation, MF classification (primary versus secondary), splenomegaly at time of transplantation, chromosome abnormalities, MF related symptoms, DIPSS score, GVHD prophylaxis and use of ATG. Median time to neutrophil recovery was 17 days (range: 6-54), with no differences between the 3 groups. Cumulative incidence of engraftment at day 30 was 91.5% (95%CI: 89.1-93.9), with no differences between the 3 groups. Cumulative incidence of grade 2-4 acute GVHD at 120 days for the entire cohort was 23% (95%CI: 19-26%); similar in the 3 groups. Three-year OS was 62% (95% CI 58 -66), 3-year PFS was 49% (45-53). At 3 years, cumulative incidence of relapse was 26% (22-29) and NRM was 25% (21-29). Figure 1 shows OS, PFS, NRM and cumulative incidence of relapse (CIR) according to regimen. Multiple variables Cox models after selection are shown in Table 1. TREO group had a significantly better OS than BUMAC (HR: 0.61, 95%CI: 0.39-0.93), BURIC had similar OS to BUMAC (HR: 0.92, 95%CI: 0.68-1.26) and there was a trend for a better OS with TREO over BURIC (HR: 0.66, 95%CI: 0.41-1.05). The same trend was also observed for PFS with a significant advantage of TREO over BUMAC (HR: 0.57, 95%CI: 0.38-0.84) and BURIC (HR: 0.60, 95%CI: 0.39-0.91) while BURIC had a similar risk to BUMAC (HR: 1.05, 95%CI: 0.80-1.39). The improvement of OS and PFS was related to a lower NRM (HR: 0.44, 95%CI: 0.24-0.80 TREO vs BUMAC; HR: 0.54, 95%CI: 0.28-1.04 TREO vs BURIC) and same relapse risk with TREO over BU (TREO vs BUMAC, HR: 0.71, 95%CI: 0.42-1.20; TREO vs BURIC, HR: 0.63, 95%CI: 0.36-1.11).

Conclusion

TREO based regimens compared to BU based shows better PFS in MF HSCT. This advantage was also confirmed for OS and NRM when compared to BUMAC. More specific prospective studies are needed to confirm findings from this registry based study.