Outcome of Patients with Primary Refractory Acute Myeloid Leukemia (AML) Undergoing Myeloablative Allogeneic Stem-Cell Transplantation (alloSCT) from HLA-Identical Sibling: Favorable Outcome in a Subset of Patients.

The prognosis of patients with acute myeloid leukemia (AML) failing to standard induction chemotherapy is very poor, and only a minority of such patients will be ultimately cured after salvage chemotherapy. The precise role of allogeneic stem cell transplantation (alloSCT) in primary refractory AML has not been extensively assessed and it is not commonly indicated in this setting. In this regard, we analyzed the outcome and eventual prognostic factors in a series of adult patients who received an undepleted allograft from an HLA-identical sibling using a myeloablative conditioning for a primary refractory “de novo” AML (i.e., patients who never achieved complete response (CR) before transplantation). Overall, 361 patients (median age: 39; range: 17–71; 60% male) fulfilling these criteria and registered to the EBMT during the period 1990–2004 were included in the study. Median interval from diagnosis to transplant was 132 days (24–360) and the median number of induction courses was 2 (1–5). Percentage of bone marrow (BM) blasts at time of SCT was 20% (0–93). Most patients with available information (n=126, 35%) harbored an intermediate-risk cytogenetics (67%). Stem-cell source was peripheral blood (PB) in 61% of cases, and conditioning regimen did not contain TBI in 59% of the procedures. Following alloSCT, 218 patients (60%) achieved CR and 143 failed to respond. After a median follow-up of 26 months, 3 and 5-year overall survival was 24±2% and 19±3%, respectively. Of note, among the subset of patients achieving CR after alloSCT, overall survival and leukemia-free survival at 3-year was 37±3% and 33±3%, respectively, whereas none of those patients failing to alloSCT survived further than 10 months after transplant. The comparison of main characteristics between subgroups of patients according to response attained after alloSCT only disclosed a shorter interval from diagnosis (126 vs. 143 days, p=0.01) in the subset of responding patients. Moreover, a lower WBC count at diagnosis, inferior to median value (3-yr OS: 27±6% vs. 15±5%; RR: 2.08, 95% CI: 1.05–4.17; p=0.03), and a BM involvement of blast cells <20% at time of transplantation (3-yr OS: 35±8% vs. 10±5%; RR: 2.13, 95% CI: 1.12–4; p=0.02) were the only variables associated to a longer survival. In conclusion, allogeneic SCT is a reasonable alternative for a subset of patients with AML failing to primary induction chemotherapy who have an available HLA-identical sibling. Thus, a low WBC count at diagnosis and a low degree of BM infiltration at transplant were predictive of a more favorable outcome in the subgroup of 56 patients with this information available. Confirmation of this finding in a larger proportion of patients would be helpful to identify those patients with AML who could benefit from an allogeneic transplantation in a refractory status.