Feasibility and Outcome of Allogeneic Hematopoietic Stem-Cell Transplantation in High-Risk AML: Real-Life Perspective from a Single Institution

Introduction: Allogeneic stem cell transplant (alloHSCT) is the treatment of choice in patients with high-risk acute myeloid leukemia (AML). However, it is uncertain exactly how many of these patients actually undergo alloHSCT as planned initially. Moreover, the real efficacy of this therapeutic option remains controversial. Since few studies have addressed these questions, to shed light on these issues, we have analyzed the results of a policy of early donor search, the proportion of patients in which initially planned alloHSCT was performed, and the outcome of alloHSCT in our center.

Patients and Methods: We included in the study 246 AML patients (median age, 51 years, 16-70; 51% males) considered fit for intensive chemotherapy, who were treated according to 3 consecutive protocols of the Spanish CETLAM cooperative group (CETLAM-99, CETLAM-2003, and CETLAM-2012). Patients were diagnosed between 2003 and 2013 in our center or referred to our center for alloHSCT; median follow-up of alive patients was 70.4 months (12.5-143.8). Indication of allloHSCT in CR1 was established according to the stratification risk of each protocol, based on AML origin (de novo vs. secondary), cytogenetics, number of courses to achieve CR, NPM1/FLT3-ITD/CEBPA mutational profile, and MRD status at end of consolidation. Statistical analyses were performed using R v3.1 and SPSS v20. The effect of alloHSCT was analyzed by Mantel-Byar test with SCT as a time-dependent covariable.

Results: AlloSCT was planned in 167 (68%) patients deemed of high-risk and it was actually performed in 130 out of these high-risk AML patients (78%). Types of donor were: matched related donor (MRD), n=63; 7/8 or 8/8 matched unrelated donor (MUD), n=51 (HLA matching: 8/8 or 10/10= 32; 7/8 and 9/10, n=19); unrelated cord blood (UCB) n=14; and family haploidentical donor, n= 2. Status at SCT was first CR (CR1) in 63 patients (48%), ≥CR2 n=23 (18%), and non-CR AML in 44 patients (34%). The type of conditioning regimen was myeloablative in 66 (51%) of the patients.

Ninety-eight patients out of 167 patients with an indication for alloHSCT, lacked a MRD (59%). An unrelated donor search was performed in 80 (82%) out of these 98 patients, and was successful in 54 (67.5%). An UCB unit was selected in 14 patients without a MUD in a timely fashion. The main reasons for not initiating an unrelated donor search were poor performance status (n=5), refractory disease (n=4), and age (n=4). Median time from start of search to finding of an adequate donor was 50 days, and median time from start of search to SCT was 131 days.

The overall outcome of these 167 patients with an indication of alloHSCT was 2-year OS: 48±8% and 5-year LFS: 36%±8% with a markedly better outcome among patients in whom alloHSCT was finally performed (2-year OS: 57±8% vs. 11%±10%; Mantel Byar, p<0.0001). The beneficial effect of alloHSCT was maintained in the subset of patients in CR1 for whom alloHSCT was planned (2-year OS: 60±12%; vs. 33±16%; Mantel Byar p=0.001). Intere stingly, the outcome of patients with an indication of alloHSCT did not differ between patients with an HLA-identical sibling and patients in whom a donor search was started (2-year OS: 50±12%; vs. 52±11%; p=0.97), in an intention-to-treat analysis. Moreover, the outcome of alloHSCT did not vary among different donor sources (MRD, MUD, and alternative donors) (2-year OS from alloHSCT: 59%±14% vs. 61%±18% vs. 44%±30, p= 0.6).

Conclusions: A policy of early search of MUD enabled alloHSCT to be performed in the majority of patients for whom it was planned. Performing of alloHSCT in patients improved outcome of this cohort of high-risk AML patients, compared to patients not achieving alloHSCT. An adequate MUD was identified in a two thirds of patients, and a similar outcome was found after alloHSCT for all donor sources (MRD, MUD, and alternative source) in this study.