BACKGROUND: Acute myeloid leukemia (AML) contains heterogenous leukemia cells in its morphology, phenotype, genetic and epigenetic alterations. In AML, leukemia stem cells (LSCs) are at the apex of the hierarchy of these heterogeneity, and they are enriched within CD34+CD38- fraction. They are thought to initiate the disease and relapse after treatment, leading to dismal prognosis. To treat relapsed or refractory AML patients urgently with transplantation, umbilical cord blood is an advantageous donor cell source because of its accessibility. Although several groups reported that resistant LSCs contributed to relapse after chemotherapy, relapse after allogeneic hematopoietic cell transplantation (HCT) has not been discussed up to this time. We investigated the clinical impact of the proportion of CD34+CD38- fraction on relapse after umbilical cord blood transplantation (CBT) for AML patients not in remission.

METHOD: We retrospectively investigated the outcome of AML patients who were not in remission and underwent CBT as the first allogeneic HCT from Aug 2013 to Mar 2016, in Toranomon Hospital. Before transplantation, the proportion of CD34+CD38- fraction in leukemia cells was quantified using flow cytometry in each patient. The patients whose proportion of CD34+ fraction was less than 10% of leukemia cells were excluded. The patients who were dead within 4 weeks after transplantation were also excluded from the analysis.

RESULT: A total of 58 patients were studied. They included 22 of AML, not otherwise specified (AML-NOS) and 36 of AML with myelodysplasia-related changes (AML-MRC). The median proportion of CD34+ fraction was 85.7% (range, 19.3 - 99.1) and 95.8% (range, 14.8 - 99.5) in AML-NOS and AML-MRC, respectively. The proportion of CD34+CD38- fraction was significantly higher in patients with AML-MRC patients, compared to patients with AML-NOS; 44.1% (range, 1.3 - 97.4) vs. 12.1% (0.3 - 85.0) , p <0.01. Therefore, further analysis was proceeded focused on AML-MRC patients. The median age of patients was 66 years (36 - 72), and 28 patients (78%) were males. Patients were categorized into 3 groups, according to the cytogenetic abnormalities before transplantation; favorable (n = 0), intermediate (n = 17), and adverse group (n = 19). Disease status was as follows; primary induction failure (n = 7), first relapse (n = 9), azacitidine failure (n = 4), and untreated before transplantation (n = 16). Thirty four patients (94%) used myeloablative conditioning regimen, and 35 patients (97%) used the combination of tacrolimus and mycophenolate mofetil for GVHD prevention. At 2 years after transplantation, the cumulative incidence of relapse was 31.7%. To assess the impact of pre-transplant factor on the incidence of relapse, we performed univariate and multivariate analysis. Analyzed pre-transplant factors were as follows; age (≥ 66 vs. < 66), treatment before transplantation (treated vs. untreated), karyotype (adverse vs. intermediate), the proportion of CD34+CD38- fraction (≥ 44.1% vs. < 44.1%). Although there was a trend of more relapses in patients with adverse karyotype (p = 0.09), no significant pre-transplant factor was extracted in univariate analysis. We then studied about the combinations of each pre-transplant factor, and we identified the combination of karyotype and the proportion of CD34+CD38- fraction as an only significant pre-transplant factor for relapse after transplantation, in both of univariate and multivariate analysis. The patients with adverse karyotype and higher proportion of CD34+CD38- fraction relapsed more frequently after CBT, compared to the remaining patients (66.7% vs. 12.8%, p = 0.02, in univariate analysis; hazard ratio 4.81, 95% CI 1.26 - 18.3, p = 0.02, in multivariate analysis).

CONCLUSION: The present study concluded that the combination of karyotype and the proportion of CD34+CD38- fraction predicted relapse after CBT for AML-MRC patients not in remission. LSCs in CD34+CD38- fraction might be resistant not only to chemothrapy but also to allogeneic HCT.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution