Background: High-dose cytarabine (HDAC) in initial induction therapy has been tested in several clinical trials both in children and adults with acute myeloid leukemia (AML), however, the trial results are conflicting. In addition, there is growing evidence on significance of measurable residual disease (MRD) monitoring to predict subsequent relapse, but MRD in AML is still challenging in terms of methodology and quality control. In the nationwide clinical trial AML-12 by the Japanese Pediatric Leukemia/Lymphoma Study Group (JPLSG), the hematological malignancy subgroup of the Japan Children's Cancer Group (JCCG), a role of HDAC in initial induction and feasibility and efficacy of central reference laboratory-based flow cytometric MRD (FCM-MRD) monitoring were evaluated in children with AML. Patients & Methods: JPLSG-AML-12 trial (registered as jRCTs041180128) was designed as a seamless phase II/III trial; because of a safety concern in HDAC induction, children (0-18 years old) with newly diagnosed de novo AML (excluding acute promyelocytic leukemia and AML in Down syndrome patients) registered in selected institutions were randomized to receive either standard ECM (etoposide, standard-dose cytarabine, and mitoxantrone) or experimental HD-ECM (etoposide, HDAC, and mitoxantrone) induction, while patients registered in non-selected institutions were non-randomly allocated to receive standard ECM; in the following phase III part, all the patients were randomized. FCM-MRD were measured at two time points [end of Induction-1 (EOI-1) and end of Induction-2 (EOI-2)] at either of the two central reference laboratories (Mie University or Kyoto University) but was not used to guide subsequent therapies. Following complete remission (CR) evaluation at EOI-2, patients achieving CR were allocated to either of the 3 risk groups and received post-remission chemotherapies; core-binding-factor standard risk (CBF SR), non-CBF SR, or high risk (HR) group. Only HR patients were allocated to receive allogeneic hematopoietic stem cell transplantation in first CR. Primary endpoints of the trial were 3-year EFS and positive FCM-MRD (≥0.1%) rate at EOI-1. Results: Total 359 eligible patients were registered at 102 institutions between 2014-2018. CR, 3-year EFS, and 3-year OS rates of all patients were 94.0%, 63.1% (95%CI, 57.9-67.9%), and 80.3% (95%CI, 75.8-84.1%), respectively. Among all, 324 patients were randomized (ECM, n=168; HD-ECM, n=156). Median follow-up period of these patients was 50.9 months (range, 0.2 - 84.7 months). No significant difference in outcomes were observed between the 2 arms; CR rates were 94.2% vs. 93.8% (p=0.998), 3-year EFS rates were 64.3% (95%CI, 56.5-71.0%) vs. 61.2% (53.1-68.4%) (p=0.551), 3-year OS rates were 84.4% (95%CI, 78.0-89.1%) vs. 75.3% (95%CI, 67.7-81.4%) (p=0.053), 3-year cumulative incidence of relapse were 28.6% (95%CI, 21.5-36.0%) vs. 30.9% (95%CI, 23.3-38.8%) (p=0.585), and 3-year non-relapse mortality rates were 0.9% (95%CI, 0.1-6.4%) vs. 1.8% (95%CI, 0.4-7.1%) (p=0.832). These outcome parameters were not different within each risk groups as well. FCM-MRD at EOI-1 was evaluated in 135 patients (80.4%) in the ECM arm and 127 patients (81.4%) in the HD-ECM arm; the positive FCM-MRD rates were 21.5% vs. 25.2% (p=0.575). Regarding grade 3 or higher adverse events (AEs) evaluated by CTCAE version 4 during Induction-1, disseminated intravascular coagulation was more frequent in the HD-ECM arm (1.8% vs. 6.4%, p=0.046), but no significant difference was observed for other AEs. Univariable Cox regression analysis for baseline characteristics revealed that the intermediate-risk cytogenetics and positive FLT3-ITD were significantly associated with decreased EFS rate. However, in multivariable analyses for EFS after EOI-1 and EOI-2, FCM-MRD was the sole poor prognostic factor; hazard ratios were 5.24 (95%CI, 3.19-8.63) (p<0.001) at EOI-1 and 5.56 (95%CI, 2.93-10.54) (p<0.001) at EOI-2. Conclusions: Although HDAC in initial induction did not improve the outcome of children with AML, overall outcomes were comparable to the recently completed pediatric AML trials worldwide. Central reference laboratory-based FCM-MRD measurement was feasible and MRD at both EOI-1 and EOI-2 were the most powerful prognostic factors to predict outcome of children with AML. Therefore, FCM-MRD should be included in future risk stratification for children with AML. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal