A Decision Analysis of Reduced-Intensity Conditioning Allogeneic Hematopoietic Stem Cell Transplantation for Older Patients with De-Novo Myelodysplastic Syndrome (MDS): Early Transplantation Offers Survival Benefit in Higher-Risk MDS

Abstract 115

Reduced-intensity conditioning (RIC) allogeneic hematopoietic stem cell transplantation is a curative approach for older patients with myelodysplastic syndrome (MDS), but is associated with early mortality risk. We previously used decision modeling to define the role of myeloablative transplantation in younger MDS patients aged <60 years, stratified by IPSS risk (Cutler et al, Blood 2004). We thereafter undertook a similar analysis to define the role of RIC transplantation in older MDS patients.

We constructed a Markov decision model in de-novo MDS patients aged 60–70 years to evaluate life expectancy after a strategy of early T-replete RIC transplantation versus standard-of-care non-transplantation approaches. Patients with chronic myelomonocytic leukemia, isolated 5q-, unclassifiable and therapy-related MDS were excluded. Patients undergoing transplantation with anti-thymocyte globulin or alemtuzumab-based conditioning or HLA mismatched or umbilical cord donors were also excluded. We undertook adjustments for quality-of-life (QoL) using standard utility estimates for different MDS and transplantation states (0.84: RBC transfusion-independent low/intermediate-1 MDS; 0.6: RBC transfusion-dependent low/intermediate-1 MDS; 0.53: intermediate-2/high MDS; 0.92: post transplantation-overall; 0.6: post transplantation-GVHD)

We evaluated survival of 92 patients after HLA matched RIC transplantation (Dana-Farber, Fred Hutchinson and CIBMTR datasets) stratified by IPSS risk state, versus survival with: best supportive care for 183 non-anemic low/intermediate-1 IPSS patients (IMRAW and Pavia datasets); hematopoietic growth factors for 78 anemic low/intermediate-1 IPSS patients (Nordic MDS Group and GFM datasets); and hypomethylating agents for 160 intermediate-2/high IPSS risk patients (Celgene AZA-001, GFM compassionate use azacytidine and M. D. Anderson decitabine datasets).

For older patients with low/intermediate-1 IPSS risk, early RIC transplantation impaired life expectancy compared to non-transplantation approaches (Table). Excluding patients transplanted beyond 12 months from diagnosis did not alter the conclusion. QoL adjustment narrowed the gap in quality adjusted life expectancy (QALE), but sensitivity analysis did not support RIC transplantation as the preferred strategy across a broad range of possible utility estimates (0.5–1), indicating that MDS-associated morbidity (e.g. RBC transfusion-dependence) does not favor early transplantation unless its QoL impact is substantial (state utility <0.4).

For older patients with intermediate-2/high IPSS risk, early RIC transplantation improved life expectancy compared with hypomethylating agents (Table). Excluding patients transplanted beyond 12 months from diagnosis did not alter the conclusion. However, transplantation is associated with early mortality such that the life expectancy benefit was apparent only after modeling survival beyond 5 years. Importantly, QoL adjusted survival benefit of transplantation was apparent both at 5 years and beyond, highlighting the morbidity of higher-risk MDS. In sensitivity analyses, QoL adjustment across the range of plausible transplantation-associated utility estimates (0.5–1) did not change the conclusion of QALE benefit.

We conclude that for de-novo MDS patients aged 60–70 years with low/intermediate-1 disease, early transplantation is not the preferred strategy unless MDS-associated QoL impairment is substantial. For intermediate-2/high IPSS risk, early RIC transplantation offers a life expectancy benefit, with quality adjusted survival benefit detectable earlier.