Comparison of Matched/Mismatched Unrelated Donor Stem Cell Transplantation to Autologous Stem Cell Transplantation for Acute Myeloid Leukemia in First Complete Remission: A Study from the Acute Leukemia Working Party of the European Group for Blood and Marrow Transplantation (EBMT)

Background. Optimal post-remission strategy for patients with acute myeloid leukemia (AML) is still matter of debate. Allogeneic Stem Cell Transplant (allo-HSCT) is the most effective treatment to prevent leukemia relapse, and for patients who lack a sibling donor transplantation from a matched or mismatched unrelated donor (URD) is usually the preferred alternative. However, increase in donor-recipient HLA mismatch, patient age and comorbidity scores lead to higher non relapse mortality (NRM) rates; moreover incidence of chronic GVHD is rather high after transplant from unrelated donors. Autologous Stem Cell Transplant (ASCT) has several advantages compared to allo-HSCT including low NRM, no GVHD risk, less late effects and better quality of life. The aim of the current study was to compare the outcome of allo-HSCT from matched (10/10 URD) or mismatched unrelated donor at a single HLA-locus (9/10 URD) to ASCT in patients with AML in first CR.

Patients and methods. We performed a retrospective analysis of 2689 AML patients receiving 10/10 URD (n=1260), 9/10 URD-HSCT (n=356) or ASCT (n=1073) in first CR between 2005 and 2013 and reported to the ALWP of the EBMT.

Results. Median FU was 35, 27 and 27 months for ASCT, 10/10 and 9/10 URD, respectively (p<10^-4);  median age was 48.7, 50.8, 48.7 years, respectively (p=10^-3). Time from diagnosis to transplant was longer for URD compared to ASCT (p<10^-4); patients who received URD had more frequently poor risk cytogenetics (p<10^-4), were more likely to get a TBI-based conditioning (p<10^-4) and were transplanted more recently (p<10^-4), compared to patients who received ASCT. The 2-year cumulative incidence of relapse (RI) for ASCT, 10/10 and 9/10 URD were 46.3±3%, 24.9±3% and 27.7±5%, respectively (p<10^-5), while the 2-year NRM rates were 3.1±2%, 16.4±4% and 20.5±4%, respectively (p<10^-5).  The 2-year KM estimates of leukemia-free survival (LFS) were 50.6±3% for ASCT, 58.7±3% for 10/10 URD and 51.8±6% for 9/10 URD (p=0.002), while the 2-year overall survival (OS) rates were  68.2±3, 63.6±3% and 55.1±6%, respectively (p<10^-4). ASCT showed significantly higher RI compared to URD independently of cytogenetic risk (good risk: p<10^-4, intermediate and poor risk: p<10^-5); accordingly, 2y LFS was significantly better for URD compared to ASCT in all risk groups (good risk: p=0.034, intermediate risk: p=0.0007, poor risk: p=0.021). ASCT and URD showed similar OS in good and poor risk patients, while in intermediate risk group ASCT resulted in similar OS compared to 10/10 URD and better OS compared to 9/10 URD (66.2±4% for ASCT, 65.8±5% for 10/10 URD, 55.4±7% for 9/10 URD, p=0.012) (Fig 1). Within intermediate cytogenetic risk group, FLT3-ITD mutational status affected outcome; in patients harboring FLT3-ITD 10/10 URD showed the best LFS and OS (LFS: 36.3±11% for ASCT, 58.4±7% for 10/10 and 34±13% for 9/10 URD, p=10^-3; OS: 51.7±12%, 62.2±7% and 41.4±14%, respectively, p=0.02). Conversely, in patients with wild type FLT3-ITD URD showed better LFS compared to ASCT (51.3±8% for ASCT, 66.7±7% for 10/10 URD, 64±13%  for 9/10 URD, p=0.008), while no difference was observed in OS. Multivariate analysis confirmed significantly lower RI for 10/10 (HR 0.36, p<10^-5,  95% CI:0.29-0.44) and 9/10 URD (HR 0.43, p<10^-5, 95% CI:0.32-0.57) and higher NRM for 10/10 URD (HR 3.88, p<10^-5, 95% CI:2.37-6.33) and 9/10 URD (HR 4.89, p<10^-5, 95% CI:2.84-8.43) compared to ASCT. URD-SCT was associated with better LFS compared to ASCT (HR 0.57, p<10^-5, 95%, CI:0.47-0.67 for 10/10 URD; HR 0.69, p=0.002, 95% CI:0.55-0.87 for 9/10 URD). 10/10 URD was associated with better OS compared to ASCT (HR 0.81, p=0.031, 95% CI:0.66-0.98) but no difference in OS was observed between 9/10 URD and ASCT (HR 1.02, p=0.87, 95% CI:0.79-1.31).    

Conclusion. In AML patients lacking an HLA-matched sibling donor URD-HSCT significantly reduces relapse risk and improves LFS. 10/10 URD showed better OS compared to ASCT in MV analysis in our series, while 9/10 URD impact on LFS didn't translate in better OS. In intermediate risk patients, in the absence of an HLA fully matched sibling or unrelated donor, autologous transplant may be considered as a valid option as ASCT results seem to overlap 10/10 URD outcome and to provide better survival compared to mismatch URD. Analysis is ongoing to better define which subpopulation of patients might benefit from each approach.

Figure 1. OS in patients with intermediate risk cytogenetics.