Minimal Residual Disease–Directed Therapy for Childhood Acute Myeloid Leukemia: Results of the AML02 Multicenter Trial.

Abstract

Abstract 16With improvements in risk-directed therapy and supportive care, cure rates for children with acute lymphoblastic leukemia are now approaching 90%. In contrast, little progress has been made for children with acute myeloid leukemia (AML). We designed a multicenter treatment protocol for children with newly diagnosed AML that used sequential minimal residual disease (MRD) assessment for risk assignment.Patients (n = 230) with AML were randomized to receive daunorubicin and etoposide plus high-dose cytarabine (HD-ADE, n = 113) or low-dose cytarabine (LD-ADE, n = 117) as Induction I. Induction II consisted of LD-ADE with or without gemtuzumab ozogamicin (GO), depending on the level of MRD after Induction I. The remaining treatment was based on presenting features and MRD levels, and included hematopoietic stem cell transplantation (HSCT) for high-risk patients. The protocol underwent 2 treatment changes. First, LD-ADE + GO was initially given only to patients with MRD more than 25%, but was later extended to patients with MRD more than 1% because the combination was found to be safe and effective at reducing MRD. Of the 30 patients who received LD-ADE + GO, 27 (90%) had a reduction in MRD and 13 (43%) became MRD negative, with a median overall MRD reduction of 98%. CNS-directed therapy initially consisted of intrathecal (IT) cytarabine alone. However, because 3 of the first 33 patients enrolled on AML02 developed CNS relapse, subsequent patients were treated with triple IT therapy consisting of methotrexate, hydrocortisone, and cytarabine. The cumulative incidence of CNS relapse decreased from 9% to 2% after this treatment change (P = 0.004). Of the 212 (92%) evaluable patients, 80% became MRD negative and 95% achieved clinical complete remission (CR) after 2 courses of therapy; overall 3-year event-free survival (EFS) and overall survival (OS) rates were 60% ± 5% and 70% ± 4%, respectively. Patients on the HD-ADE and LD-ADE treatment arms had similar MRD-negative rates after 1 (66% vs. 58%, P = 0.17) and 2 courses of therapy, as well as similar CR, EFS (59% vs. 61%, P = 0.79), and OS rates (68% vs. 73%, P = 0.41). Multipredictor logistic regression modeling indicated that the odds of having detectable MRD after 1 course of therapy was significantly lower for patients with core binding factor (CBF) leukemia [t(8;21) or inv(16)] (P <0.0001), and significantly greater for those with FLT3-ITD-positive leukemia (P = 0.0007). Despite treatment intensification for MRD-positive patients, presence of MRD after Induction I remained a significant adverse predictor: 3-year EFS rates were 72% for MRD-negative patients and 41% for MRD-positive patients (P<0.0001). We also observed significant differences in 3-year EFS rates between patients with and without FLT3-ITD (37% vs. 65%, P = 0.003) and between patients with and without CBF leukemia (82% vs. 53%, P = 0.0003). By Cox proportional hazard modeling, EFS was significantly associated with MRD on day 22 (P = 0.009) and CBF leukemia (P=0.04). Only 8 of the 230 patients enrolled had infection-related deaths: 1 patient died during induction, 4 in CR during or after completing chemotherapy, and 3 in CR after HSCT. As previously reported, use of the prophylactic vancomycin, ciprofloxacin, and voriconazole contributed to this low rate of toxic death.In conclusion, risk-adapted therapy, incorporation of GO, and excellent supportive care contributed to a CR rate of 95%, low mortality, and an overall 3-year OS rate of 70% for children with AML enrolled on AML02. Although this is substantially higher than rates reported in our previous studies, it should be possible to further increase cure rates by additional refinements in risk assignment, as well as by introducing novel agents that can bypass drug resistance. To this end, our current AML08 protocol has incorporated newly discovered molecular predictors, highly sensitive MRD tests, and natural killer cell therapy. Disclosures:No relevant conflicts of interest to declare.