Etoposide can be safely removed from induction chemotherapy without impacting survival for pediatric acute myeloid leukemia – a report from the Children's oncology group study AAML1831

Abstract Introduction: Cytarabine, daunorubicin and etoposide have been the standard backbone of induction chemotherapy in Children's Oncology Group (COG) frontline trials for pediatric acute myeloid leukemia (AML). Due to increased secondary malignancy risk and unclear efficacy benefit over daunorubicin and cytarabine (DA) alone, etoposide is not part of induction therapy in adult AML. Etoposide was thus removed from induction therapy in AAML1831 (NCT04293562), the current COG frontline trial for pediatric AML. Methods: To assess the impact of etoposide during induction, we compared response and 2-year survival rates for patients treated on AAML1831 Arm A (N=335) with those treated on prior Phase 3 studies AAML0531 Arm B (N=256) and AAML1031 (Arms A & B, N=878). Outcome measures were rates of flow cytometric minimal residual disease positivity (MRD+; >0.05%) at the end of Induction I (EOI1), relapse risk (RR), disease-free survival (DFS) and overall survival (OS). Outcomes were reported based on risk group assignments using AAML1831-defined criteria: low-risk 1 (LR1; favorable genetics and EOI1 MRD negative), low-risk 2 (LR2; not in LR1 or HR) and high risk (HR; any high-risk genetics or EOI1 MRD+ without favorable genetics) and retrospectively applied to patients on AAML0531 and AAML1031. Patients with FLT3-ITD or FLT3 activating mutations were excluded, as well as HR patients who did not undergo hematopoietic stem cell transplant (HSCT) in first remission to help control for variability across studies in HSCT utilization of HSCT for HR cohorts. Results: To evaluate the effect of etoposide in induction, we compared the rates of EOI1 MRD+ between AAML1831 and AAML0531 stratified by prognostic AML genetics. Rates of EOI1 MRD+ were similar (Favorable Genetics: 9% vs 9.4%, p=0.91; Neutral Genetics: 25% vs 21%, p=0.56; Unfavorable Genetics: 38% vs 27%, p=0.13). The use of gemtuzumab ozogamicin (GO) in Induction 1 of AAML1831 but not AAML1031 may be a confounding variable, though we have previously shown in AAML0531 that GO does not influence EOI1 MRD. Two-year survival outcomes were compared between AAML1831 vs AAML0531 or AAML1031. The effect of omitting etoposide is most accurately evaluated by comparing outcomes for LR2 patients between AAML1831 Arm A (no etoposide) and AAML0531 Arm B (etoposide) because both trials delivered 5 cycles and included GO. Comparison of RR, DFS and OS between AAML1831 and AAML0531 did not yield significant differences (RR: 28.3% vs 36.5%, p=0.120; DFS: 69% vs 59.2%, p=0.073; OS: 83.5% vs 78%, p=0.192). Comparison of outcomes between LR2 patients on AAML1831 (5 cycles; no etoposide; GO) and AAML1031(4 cycles; etoposide; no GO) achieved statistical significance in RR and DFS (RR: 28.3% vs 48.2%, p<0.001; DFS: 69% vs 48.9%, p<0.001), although OS did not achieve statistical significance (83.5% vs 76.8%, p=0.166). Comparing etoposide effect between trials for LR1 patients is confounded by total number of cycles (AAML1831 – 4 cycles; AAML0531 – 5 cycles) and inclusion of GO during induction (AAML1831 – GO; AAML1031 – No GO). Even so, no benefit of etoposide in induction was detected, as LR1 patients had similar RR and DFS on AAML1831 vs AAML0531 and vs. AAML1031 (RR: 18% vs 16.3%, p=0.971; and 21.9%, p=0.269; DFS: 80% vs 77.2%, p=0.383; and 73.9%, p=0.146). OS was superior for AAML1831 vs AAML0531 (96.8% vs 89.4%, p=0.040), but did not achieve statistical significance vs AAML1031 (96.8% vs 91.2%, p=0.081). HR patients on AAML1831 vs AAML0531 received similar therapy prior to HSCT, while AAML1031 did not include GO. Again, no benefit of induction etoposide was seen, as rates of RR, DFS and OS were similar on AAML1831 vs AAML0531 and vs AAML1031 (RR: 32% vs 39%%, p=0.72; and 37%, p=0.53; DFS: 63% vs 54%, p=0.55; and 45%, p=0.07; OS: 71% vs 61%, p=0.76; and 51% p=0.1). Conclusion: The removal of etoposide from induction did not diminish EOI1 MRD response or negatively impact RR, DFS or OS in pediatric AML when compared to historical data. While this observation may be influenced by improved supportive care, greater availability of salvage regimens and HSCT for CR2, we believe that these findings could set a new precedent for DA/GO induction therapy as the standard of care treatment for pediatric AML given the preserved efficacy and potential to reduce long-term toxicity. Astellas Pharma Inc. provided funding & clinical supply for AAML1831. Jazz Pharmaceuticals provided funding and a medical review of this abstract.