Outcomes of Children and Adolescents with Acute Myeloid Leukemia Given a Low-Versus Standard-Dose Chemotherapy Regimen for Remission Induction (CALSIII-AML18): A Multicenter, Phase 3, Randomized, Noninferiority Trial

Background: Intensive chemotherapy is recommended for remission induction of pediatric acute myeloid leukemia (AML). However, life-threatening complications are common especially in vulnerable populations. We investigated whether a low-dose chemotherapy (LDC) regimen would be noninferior to a standard-dose chemotherapy (SDC) regimen for induction remission.This study also aimed to assess whether LDC is not inferior to SDC in terms of survival outcomes.

Methods: We conducted a randomized, noninferiority study in children with AML who were admitted to 14 medical centers in China. Patients were randomly assigned to receive two remission-induction courses of either low-dose chemotherapy (LDC) (cytarabine 10 mg/m², subcutaneous, q12 hours, 20 doses; mitoxantrone or idarubicin 5 mg/m², intravenous, days 1, 3 and 5; and G-CSF 5 mcg/kg, subcutaneous, daily, 10 doses) (n = 246) or standard-dose chemotherapy (SDC) (cytarabine 100 mg/m², intravenous, q12 hours, 20 doses; daunomycin 50 mg/m², intravenous, days 1, 3, 5; and etoposide 100 mg², intravenous, days 1 to 5) (n = 251). Depending on the risk of relapse, patients in complete remission (CR) in both arms received two to three cycles of intensive consolidation chemotherapy and/or underwent hematopoietic stem cell transplantation (HSCT). The primary endpoint was to compare CR rates in patients randomized to the LDC or SDC regimens. The secondary endpoints were to determine the safety of the LDC regimen and to compare the time to platelet and neutrophil count recovery in the two groups. Finally, in a multivariate analysis, we determined the impact of the type of remission induction on the outcome. This trial is registered at chictr.org as # ChiCTR-18000015883 and has been completed.

Results: From June 2018 to June 2022,497 participants aged <18 years with de novo AML were randomized. Patients with Down syndrome, promyelocytic leukemia, or megakaryoblastic leukemia were excluded. There were no significant differences in age, sex, or genotype between the two arms. Complete morphologic remission with or without platelet recovery (CR/CRi) was attained in 72.8% and 70.3% of patients assigned to the LDC and SDC arms, respectively, after Induction I and in 95.5% and 95.7%, respectively, after Induction II (P = .545 for Induction I and .898 for Induction II). Residual disease < 0.1%, as measured by flow cytometry, was observed in 54.8% and 58.9% of patients in the LDC and SDC arms, respectively, after Induction I and in 85.9% and 84.4%, respectively, after Induction II (P = .229 for Induction I and .827 for induction II). Median time to neutrophil count recovery was 15 days for patients in the LDC arm vs. 22 days for those in the SCD arm after Induction I (P < .001) and 12 days vs. 18 days for patients in the respective arms after Induction II (P < .001). Median time to platelet count recovery was 13 days for patients in the LDC arm vs. 19 days for those in the SCD arm after Induction I (P < .001) and 3 days vs. 8 days for patients in the respective arms after Induction II (P < .001). Grade 3 and 4 toxicities were significantly lower in patients in the LDC arm than those in the SDC arm. In Induction I, febrile neutropenia occurred in 75.2% and 93.2% (P < .001), colitis in 3.7% and 10.7% (P = .003), and gastrointestinal hemorrhage in 0.4% and 4.4% (P = .006) in the LDC and SDC arms, respectively. In Induction II, febrile neutropenia occurred in 34.4% and 66.9% (P < .001), lung or sinus infection in 2.8% and 9.3% (P = .008), and sepsis in 3.3% and 10.7% (P = .004), in the LDC and SDC arms, respectively. There was no significant difference in treatment-related mortality between the two treatment arms (P = 204).Compared to patients in the SDC arm, patients in the LDC arm had noninferior 3-year EFS of 61.9% (95% CI: 55.7 to 68.8) vs. 62.2% (95% CI: 56.1 to 68.8) (P = .994) (Figure 1A), and 3-year OS of 81.0% (95% CI: 75.7 to 86.6) vs. 83.2% (95% CI: 78.2 to 88.5) (P = .485) (Figure 1B). In a multivariate analysis considering known prognostic indicators in pediatric AML, treatment arm was not associated with outcome.

Conclusion: In children with AML, remission induction with a low-dose chemotherapy regimen with concurrent G-CSF administration was well tolerated and was associated with CR, EFS, and OS rates that were not inferior to those of patients treated with a standard chemotherapy regimen.