Optimization of induction chemotherapy in pediatric patients with Acute Myeloid Leukemia: Real-world evidence to support removal of etoposide from induction 1

Abstract

Abstract Introduction: Induction chemotherapy for pediatric AML has historically used cytarabine, daunorubicin and etoposide (ADE). However, etoposide likely increases toxicity, and adult data suggest that ADE does not improve survival outcomes over daunorubicin and cytarabine (DA) alone. This may be particularly true in the modern era in which therapeutic intensity can be achieved by adding gemtuzumab-ozogamycin (GO). This analysis sought to compare clinical outcomes of DA to ADE for pediatric AML induction. Methods: We used data from REAL–AML, a real-world, retrospective cohort of pediatric AML patients aged ≤18 at diagnosis treated at 17 US institutions since 2011. Patients with de novo AML who received induction 1 chemotherapy with DA or ADE were included. The primary outcome was 2-year event free survival (EFS). Secondary outcomes were overall survival (OS), induction mortality, and end induction 1 clinical remission (EOI CR) and measurable residual disease (MRD) negative CR. Covariates considered were demographic and disease characteristics, organ dysfunction at presentation (defined via a merge with the Pediatric Health Information System database) and receipt of targeted therapies (FLT3 inhibitors, GO). In regression analyses, we adjusted for covariates that differed by regimen (difference >10% or p<0.1). Survival analyses used Kaplan-Meier methods and multivariable Cox models to estimate adjusted hazard ratios (aHR). EOI response analyses used modified Poisson regression to calculate adjusted prevalence ratios (aPR). Results: The analytic cohort included 795 patients; 109 received DA and 686 received ADE in first induction. Almost all patients given DA were diagnosed after 2020 whereas ADE was used across all years. Compared to those who received ADE, DA patients were older (median age 11 vs 8 years, p=.02), more commonly CNS1 (67% vs 57%, p<.001), less likely to require ICU care at diagnosis (22% vs 36%, p=.017), and more likely to have received induction 1 on a clinical trial (65% vs 30%, p<.001). Patients who received DA were more likely to have identified low- and high-risk cytomolecular features (35% and 23% vs 27% and 18%, respectively, p=0.03) and to have received GO (95% vs 16%, p< .001) and dexrazoxane (97% vs 49%, p<.001). With a median follow-up of 21m (DA) and 65m (ADE), 2y EFS was equivalent: 57% vs 56% in patients who received DA and ADE, respectively. After adjustment, the aHR was 0.90 (95%CI 0.63-1.29, p=0.56). Similar results were seen with regards to OS: 2y OS was 77% vs 73% in DA vs ADE, aHR=0.83 (95%CI 0.50-1.37, p=0.47). EOI response assessment slightly favored DA: EOI CR 84% vs 65%, MRD negative CR 73% v 62%, for DA and ADE, respectively; EOI CR aPR=1.10 (95%CI 1.00-1.21, p=0.05), MRD negative CR, aPR=1.14 (95%CI 0.97-1.33, p=0.11). TRM during induction 1 was similar by arm (3% in both arms). For the majority, induction 2 was the same as induction 1: 61% of ADE and 77% on DA. Relatedly, most patients treated on study for cycle 1, stayed on for cycle 2: ADE 84% and DA 94%. We conducted two sensitivity analyses: one limited to patients diagnosed after 2020 to account for temporal differences in pediatric AML including expanded cytomolecular and improved MRD testing, enhanced supportive care (e.g. dexrazoxane) and additional salvage options, and one restricted to patients who received GO in induction 1. When limited to post-2020 (n=205), the point estimates for EFS and OS favored ADE, but these results were not significant: EFS aHR 1.46 (95%CI 0.80-2.66, p=0.22) and OS aHR 1.49 (95%CI 0.68-3.23, p=0.32). In contrast, EOI response was comparable by arm: CR aPR 1.06 (95%CI 0.92-1.22) and MRD negative CR aPR 1.01 (95%CI 0.78-1.31). Among patients who received GO (n=215), results demonstrated equivalence across arms by all measures: EFS aHR=1.09 (95%CI 0.54-2.18, p=0.81), OS aHR 1.02 (95%CI 0.44-2.37, p=0.97), CR aPR 1.07 (0.93-1.22, p=0.32), MRD negative CR aPR=1.06 (95%CI 0.83-1.34, p=0.65).Conclusion: These real-world data convincingly demonstrate comparable outcomes with ADE and DA when used in induction 1, particularly in patients who receive GO. These data suggest that DA can be safely used as standard backbone chemotherapy for pediatric patients in this cycle, which may reduce etoposide-related long term side effects.