Impact Of Conditioning Regimen, Donor Source, and DIPSS Score On Outcome Of Allogeneic Hematopoietic Stem Cell Transplantation For Myelofibrosis

To investigate impact of conditioning regimen, donor source, and Dynamic International Prognostic Scoring System (DIPSS) on outcome of allogeneic hematopoietic stem cell transplantation for myelofibrosis, we analyzed results in 388 patients transplanted between 1990 and 2011 in three major transplant centers (Essen, Germany: n = 63; Hamburg, Germany; n = 156; Seattle, USA: n = 169), including 216 male and 171 female patients, 18 – 79 (median 54) years old, with either primary (n = 283) or post ET/PV (n = 150) myelofibrosis who received myeloablative (n = 190), reduced intensity (n = 182), or non-myeloablative (n = 16) conditioning and were given stem cells from related (n = 156, including 152 HLA-identical siblings, 1 haplo-identical relative, and 3 identical twins ) or unrelated donors (n = 232; 175 were HLA-matched and 57 were HLA-mismatched). Stem cell source was bone marrow (n = 51) or peripheral blood (n = 337). According to DIPSS at time of transplantation patients were classified as low (n = 35), intermediate-1 (n = 106), intermediate-2 (n = 161), or high (n = 84) risk. The conditioning regimen consisted of busulfan 16 mg/kg and cyclophosphamide (n = 136), busulfan 10 mg//kg and fludarabine (n = 154), treosulfan-based (n = 26), TBI 12 Gy-based (n = 52), or low dose TB 2 Gy and fludarabine (n = 16), and melphalan 140 mg/m2 and fludarabine (n = 4). After a follow-up of 3.5 – 7.7 (median 5) years the 1-year non-relapse mortality (NRM) of the entire study cohort was 24% (95% CI: 20 – 28). In multivariate analysis significant factors for NRM were age as continuous variable (HR: 1.029; 95% CI: 1.011 – 1.048, p = 0.002), HLA mismatch (HR: 2.026; 95% CI: 1.318 – 3.113, p = 0.001), and TBI containing conditioning (HR 1.872; 95% CI: 1.198 – 2.926, p = 0.006). The cumulative incidence of relapse was 17% (95% CI: 13 – 21) at 5 years, negatively impacted by age as continuous variable (HR: 1.030; 95% CI: 1.003 – 1.057, p = 0.028), unrelated donors (HR: 0.560; 95% CI: 0.339 – 0.926, p = 0.024), and the use of ATG (HR: 2.425; 95% CI: 1.342 – 4.381, p = 0.003) or campath (HR: 3.257; 95% CI: 2.129 – 6.985, p < 0.0001 The estimated 3 and 5 year overall survival for the study group was 62% (95% CI: 56 – 68%) and 59% (95% CI: 53 – 65%), respectively. Significant factors for 5-year survival in the univariate analysis were HLA-matched vs. mismatched donor (61% vs. 47%, p = 0.05), related vs. unrelated donors (65% vs. 54%, p = 0.03), high dose conditioning vs. RIC vs. NMA (58% vs. 61% vs. 32%, p = 0.04), non TBI vs. TBI containing conditioning regimens (62% vs. 42%; p = 0.001), and DIPSS score at transplantation (low 83% vs. intermediate-1 64% vs. intermediate-2 58% vs. high risk 45%; p < 0.001). In the multivariate analysis for survival significant factors were age as continuous variable (HR: 1.027; 95% CI: 1.009 – 1.045; p = 0.004), DIPSS score: intermediate-1 (HR: 1.662, p = 0.3), intermediate-2 (HR: 2.069, p = 0.09), high (HR: 2.924, p = 0.02), and TBI-containing conditioning regimen (HR: 1.973; p < 0.001). Conclusion This analysis of results in a large cohort of patients confirms a high success rate of allogeneic hematopoietic stem cell transplantation for myelofibrosis significantly impacted by DIPSS classification at the time of transplantation. In addition the analysis confirms the risk factors of patient age and donor HLA-mismatch. Furthermore, the data show similar results between RIC and MAC, but less favorable outcome with TBI containing conditioning, regardless of the dose of TBI. Novel approaches need to be developed, particularly for patients with more advanced disease. Contribution The authors NK and MD contributed equally. Disclosures: No relevant conflicts of interest to declare.