Prognostic Impact of Conditioning Intensity on Outcomes after Allogeneic Transplantation for MDS with Low Blasts

Introduction

Allogeneic hematopoietic stem cell transplantation (allo-HSCT) provides curative outcomes in patients with myelodysplastic syndromes (MDS). The blast percentage in MDS is one of the important factors to utilize allo-HSCT, and the other disease-related factors, such as transfusion dependency and chromosomal risk, are crucial to consider whether patients harboring MDS with less than 5% of marrow blasts (MDS-Lo) undergo to allo-HSCT. However, the prognostic impacts of conditioning intensity in MDS-Lo patients are still unclear. We here conducted a nationwide retrospective study to clarify the prognostic factor and investigate the optimal conditioning intensity for MDS-Lo.

Methods

In this study, patients diagnosed as the other than refractory anemia with excess blasts (RAEB)-1 and -2 according to the WHO classification (marrow blasts <5% and peripheral blood blasts <1%) were considered to be MDS-Lo. The clinical data of 1,229 de novo MDS-Lo patients (i) who had stayed MDS-Lo from diagnosis to transplantation and (ii) who underwent their initial allo-HSCT between January 2001 and December 2020 were collected from the Transplant Registry Unified Management Program of the Japan Society for Hematopoietic Cell Transplantation. Cox proportional hazards regression models were used to evaluate variables potentially affecting overall survival (OS), chronic graft-versus-host disease (GVHD)-free survival (CRFS), and GVHD- and relapse-free survival (GRFS). Fine and Gray proportional hazards models were used to evaluate variables potentially affecting cumulative incidences of relapse (CIR) and non-relapse mortality (NRM), using cumulative incidence curves to accommodate competing events.

Results

In a total of 1,229 patients, 651, 397, and 181 patients received myeloablative conditioning (MAC), reduced-intensity conditioning (RIC), and non-myeloablative conditioning (NMAC) groups, respectively. The estimated 3-year OS rates were 63.3%, 61.0%, and 69.7% in MAC, RIC, and NMAC groups, respectively; the estimated 3-year CRFS rates were 41.1%, 43.9%, and 47.2% in MAC, RIC, and NMAC groups, respectively; the estimated 3-year GRFS rates were 27.5%, 33.4%, and 35.4% in MAC, RIC, and NMAC groups, respectively; the estimated 3-year CIR were 11.9%, 14.2%, and 16.0% in MAC, RIC, and NMAC groups, respectively; and the estimated 3-year NRM were 29.8%, 31.0%, and 24.7% in MAC, RIC, and NMAC groups, respectively.

The multivariate analyses revealed four factors affecting lower OS: older recipient's age (Hazard ratio (HR) [95% confidential interval], 1.70 [1.28-2.24]; P<0.001 for 40-59 years)(HR, 2.32 [1.71-3.15]; P<0.001 for 60 years or older), performance status 2-4 at HSCT (HR, 1.84 [1.32-2.57]; P<0.001), the history of infection before HSCT (HR, 2.12 [1.55-2.89]; P<0.001), and high number of red blood cell (RBC) transfusion before HSCT (HR, 1.65 [1.35-2.02]; P<0.001). These factors also negatively affected CRFS, GRFS, and NRM. In addition, poor cytogenetic risk group was associated with worse OS (HR, 1.66 [1.32-2.09]; P<0.001), CRFS (HR, 1.48 [1.22-1.80]; P<0.001), and GRFS (HR, 1.36 [1.14-1.64]; P=0.001) but not CIR. There was no significant difference of OS, CRFS, GRFS, CIR, and NRM in either RIC or NMAC group compared to MAC group.

We next performed the subgroups analysis based on the prognostic factors. The subgroup analysis for the comparison between MAC and RIC regimens showed the better OS of MAC regimen among patients with low number of RBC transfusion (HR, 1.33 [1.01-1,75]; P=0.039). The subgroup analysis for the comparison between RIC and NMAC regimens revealed the better OS of NAMC regimen among patients aged <40 years (HR, 0.43 [0.19-0.96]; P=0.038), no history of infection (HR, 0.67 [0.50-0.91]; P=0.009), and good cytogenetic risk (HR, 0.54 [0.38-0.83]; P=0.004).

Conclusion

This study showed the prognostic factors on the post-transplant outcomes among MDS-Lo patients, and no significant difference of conditioning intensity in the entire population. Furthermore, we found the factors (the number of RBC transfusion, patient's age, the history of infection, and cytogenetic risk) potentially affecting the selection of conditioning intensity. These findings would improve the management for transplant candidates with MDS-Lo.