Optimizing the Outcomes of Allogeneic Hematopoietic Stem Cell (HSC) Transplantation for Hematological Malignancies: The Perfect Combination of Conditioning Regimen, Disease Stage and Type of HSC Donor

Introduction

Allogeneic hematopoietic stem cell transplantation (allo-HSCT) remains the only curative strategy for a majority of patients with high-risk hematological malignancies. Several studies have evaluated the impact of conditioning intensity on the long-term transplantation outcomes, mainly retrospective and derived from large registries or non-randomized trials. These studies showed that more intensive conditioning (MAC) regimens are associated with a reduced risk of relapse, but do not translate into improvement of survival due to increased non-relapse mortality (NRM).  Reduced intensity/Non-myeloablative conditioning (RIC/NMA) through graft-versus-leukemia effect has  been  associated with lower NRM, but higher relapse rates leading  to  similar  overall  survival  (OS)  when  compared  to MAC. However, in daily clinical practice, these results are difficult to follow because of the combination of important impact of disease stage, the type of HSC donor and its HLA matching with the patient. In addition, the historical nature of the previous studies may lead to the observation of different results today as the experience in drugs toxicities management has changed over time.

The objective of this study is to evaluate the impact of the conditioning regimen intensity taking into account the disease stage and the type of HSC donor with its HLA matching on transplantation outcomes in a large population of patients with high-risk hematological malignancies.

Material and methods

A total of 542 patients who received allo-HSCT between January 2006 and December 2014 in our center were included, 321 (59%) were males, the median age at allo-HSCT was 49 years (range: 18-70). There was 256 (47%) acute leukemia (202 AML, 54 ALL), 61 (11%) MDS, 60 (11%) multiple myeloma, 46 (8%) NHL, 25 (5%) Hodgkin’s disease, 23 (4%) myeloproliferative neoplasms, 21 (4%) CML, 12 (2%) CLL and the rest with other hematological diseases. All patients were classified as at high-risk according to either clinical, immunophenotypic, cytogenetic or molecular markers. Conditioning regimen was classified as recently published (Gyurkocza et al. Blood 2014), therefore 282 (52%) received MAC and 260 (48%) received RIC/NMA; at allo-HSCT 320 (59%) patients were in CR and 222 (41%) in less than CR. HSC donor was identical siblings (Sib) for 199 (37%) patients (100 BM, 99 PBSC), 10/10 HLA matched unrelated (MUD) for 159 (29%) (79 BM, 80 PBSC), 6/6 HLA matched double cord blood (CB) units for 12 (2%), 9/10 HLA mismatched unrelated (MMUD) for 114 (21%) (54 BM, 60 PBSC), and the rest of 58 (11%) were 5/6 or 4/6 MM double CB units. For sex mismatching, in 119 (22%), it was female donor to a male patient; 295 (54%) were ABO compatible, 105 (20%) had minor incompatibility and 142 (26%) major incompatibility.

Results

The median follow-up for surviving patients was 29 months (range: 4-96). We conducted a cox multivariate model for OS including patient age, disease status at allo-HSCT, conditioning regimen, type of donor and HLA matching, in addition to ABO and sex mismatching, with stratification on the type of disease; this model showed a significant impact of disease status in favor of CR (HR=1.5, 95%CI: 1.2-2.0, p=0.001), conditioning regimen in favor of MAC (HR=0.68, 95%CI: 0.53-0.88, p=0.003) and type of donor in favor of  Sib (HR=0.68, 95%CI: 0.5-0.9, p=0.01). Interestingly, we were able to find an optimal association between these 3 factors leading to significantly better results in terms of OS and NRM independently of the disease type. When in CR, patients receiving MAC from Sib or from MUD had significant better OS and NRM compared to the rest of patients with 5-years rates of 71% vs 36% (p<0.0001) and 15% vs 37% (p=0.001) respectively. If not in CR, only patients who received HSC from Sib either after RIC or MAC showed significantly better OS and NRM compared to the rest of patients with 5-years rates of 50% vs 26% (p=0.001) and 22% vs 45% (p=0.008) respectively (see Figure). Considering only MMUD, patients receiving CB with RIC had better OS and NRM rates compared to 9/10 MMUD (RIC or MAC) and to MAC CB (p=0.07).

Conclusion

We provide in this large study, a practical daily clinical practice outcome preview after allo-HSCT, independently of the type of disease, for the combination of significant impacting factors namely disease status at allo-HSCT, conditioning regimen and type of HSC donor with a superiority for MAC when used in CR from Sib or MUD.