Allogeneic Hematopoietic Stem Cell Transplantation Overcomes the Adverse Prognostic Impact of Minimal Residual Disease in Acute Myeloid Leukemia

The identification of minimal residual disease (MRD) can predict impending relapse in acute myeloid leukemia (AML) patients. Little data exists evaluating the prognostic impact of MRD, as determined by multiparametric flow cytometry (MFC), at the time of allogeneic (allo-HCT). Although disease outcomes may be worse for MRD positive (MRD+) patients, MRD is often associated with other adverse risk factors leaving it unclear whether MRD is an independent risk factor for relapse or a surrogate marker for underlying poor risk disease features.

We retrospectively analyzed 97 consecutive AML patients in complete morphological remission (CR) who underwent allo-HCT with a matched related donor (MRD, n=30, 31%), matched unrelated donor (MUD, n=4, 4%) or umbilical cord blood (UCB, n=63, 65%) at the University of Minnesota between January 2005 and June 2009. Presence of MRD at allo-HCT was determined by MFC. Analyses were done separately for myeloablative (MAC) and reduced intensity conditioning (RIC) patients, testing the impact of MRD along with conditioning intensity, age, donor type and disease status on allo-HCT outcomes.

Of 97 patients, 41 (42%) had MAC; 57 (58%) had RIC. Sixty-six were in first CR (CR1) with the rest in CR2 or later remission (CR2+). MRD at allo-HCT was detected in 7 patients after MAC (17%) and 7 after RIC (12.5%); and this frequency was similar in patients in CR1 and CR2+ (13% vs. 16%, p=0.7). MRD+ and MRD- patients had similar median age (40 vs. 44 yrs), gender, donor source (MSD or MUD vs. UCB), CMV serostatus and diagnostic cytogenetic risk group. Six of 40 (15%) intermediate and 5 of 39 (13%) high risk cytogenetics patients had MRD+ (p=0.8). Only 3 of 53 patients with a cytogenetic abnormality at diagnosis had it detected prior allo-HCT and 1 of 3 had MRD.

The median follow-up of survivors was 25 months. Two-year probabilities for MAC and RIC patients were similar: Overall survival (OS), 48% and 47 % and leukemia free survival (LFS) 43% and 41% respectively. When disease outcomes were analyzed separately by MRD status (table), OS and LFS were markedly worse in MRD+ patients receiving RIC, but this difference was not statistically significant. In multivariate analysis, MRD+ was not an independent prognostic factor for OS and LFS. Although we identified no adverse prognostic factors for MAC patients, patient with RIC in CR2+ had worse OS and LFS vs. CR1 (HR 2.6, p=0.04 and HR 2.7, p=0.04 respectively).

The negative prognostic impact of MRD was overcome by allo-HCT with MAC, but outcomes with MRD+ were suggestively inferior after RIC. However due to limited sample size, MRD in patients with RIC should be further investigated.

Weisdorf:Genzyme: Consultancy, Research Funding.