Allogeneic Hematopoietic Cell Transplantation for Acute Myeloid Leukemia: Is It Time to Move Toward a Minimal Residual Disease-Based Definition of Complete Remission?

Background: Treatment algorithms for allogeneic hematopoietic cell transplantation (HCT) typically consider patients with acute myeloid leukemia (AML) in morphologic complete remission (CR) separately from those with active disease (i.e. ≥5% marrow blasts by morphology), implying distinct outcomes for these two groups. However, it is well recognized that the presence of minimal residual disease (MRD) at the time of transplantation is associated with adverse post-HCT outcomes for patients in morphologic CR. This well established effect of pre-HCT MRD prompted us to compare outcomes in patients in MRD^pos CR to those with active AML who underwent myeloablative allogeneic HCT at our institution.

Patients and Methods: We retrospectively studied 359 consecutive adults with AML who underwent myeloablative allogeneic HCT from a peripheral blood or bone marrow donor between 2006 and 2014. Pre-HCT disease staging included 10-color multiparametric flow cytometry (MFC) on bone marrow aspirates in all patients. MRD was identified as a cell population showing deviation from normal antigen expression patterns compared with normal or regenerating marrow. Any level of residual disease was considered MRD^pos.

Results: Three hundred and eleven patients (87%) were in morphologic CR at the time of transplantation, with 76 (21%) in MRD^pos CR and 235 (66%) in MRD^neg CR. 48 patients (13%) had active disease (7 untreated newly diagnosed AML, 16 untreated relapsed AML, and 25 refractory or relapsed AML who failed salvage therapies). Patients with MRD^pos CR or active AML more often had adverse-risk cytogenetics (P=0.001) and secondary leukemias (P<0.001) than MRD^neg CR patients. Patients with active AML also more often had incomplete blood count recovery before HCT than patients in morphologic CR (P<0.001). Three-year relapse estimates were 67% in MRD^pos morphologic CR patients and 65% in patients with active AML, contrasted to 22% in MRD^neg CR patients. Three-year overall survival estimates were 26%, 23%, and 73% in these three groups, respectively. After multivariable adjustment for age, cytogenetic risk, type of AML (de novo vs. secondary AML), pre-HCT karyotype (normalized vs. not), and pre-HCT peripheral blood counts (recovered vs. not), MRD^neg CR status remained statistically significantly associated with longer overall and progression-free survival as well as lower risk of relapse compared to being in MRD^pos morphologic CR or having active disease, with very similar outcomes between the latter two groups. Specifically, compared to MRD^neg CR patients, the hazard ratios (95% confidence interval) for MRD^pos CR patients and those with active disease were 3.68 (2.51-5.40) and 4.39 (2.56-7.53) (both P<0.001) for overall survival; for progression-free survival, corresponding hazard ratios were 4.37 (3.02-6.30) and 5.29 (3.18-8.80) (both P<0.001), whereas for risk of relapse, these estimates were 4.16 (2.68-6.44) and 4.86 (2.49-9.49) (both P<0.001), respectively.

Conclusion: Outcomes for adults transplanted with morphologically detectable disease closely resemble those of MRD^pos CR patients, with a cumulative relapse risk of ~65% and survival estimates of 20-25% at 3 years. This similarity held up after accounting for numerous other prognostic covariates. The resemblance in outcomes between patients with MRD^pos morphologic CR and those with active disease at the time of HCT support the use of treatment algorithms that use MRD-based rather than morphology-based disease assessments.