Gemtuzumab Ozogamicin (GO) In Children With De Novo Acute Myeloid Leukemia (AML) Improves Event-Free Survival (EFS) By Reducing Relapse Risk – Results From The Randomized Phase III Children’s Oncology Group (COG) Trial, AAML0531

Escalation of treatment intensity for childhood acute myeloid leukemia has improved EFS but with excessive toxicity and toxic mortality (TM), compelling a search for more targeted less toxic alternatives.

To determine if the addition of GO to standard chemotherapy improves EFS in pediatric AML.

Patients < 30 years of age were randomly assigned to receive standard therapy alone (noGO) or to receive two doses of GO 3 mg/m²/dose on day 6 of Induction I (IndI) and on day 7 of Intensification II (IntII) in a 5 cycle chemotherapy backbone previously piloted by COG (Cooper et al, Cancer 2012). Use of stem cell transplant (SCT) was stratified by overall risk group assignment (based on cytogenetics, FLT3-ITD high allelic ratio (HAR), and IndI response) in which high risk patients were allocated to best allogeneic donor SCT after Int I, low risk (LR) patients received chemotherapy only, and intermediate risk (IR) were assigned to SCT if there was a matched family donor (MFD).

Between 2006-2010, 1022 of 1070 enrolled non-DS pts were eligible for analysis; characteristics are presented in Table 1. Median follow-up was 3.6 (range 0-6.4) years for those alive.

Protocol therapy was well tolerated with a TM of 2% in induction and 5% overall with no difference by study arm. VOD was observed in 3% (severe in 0.6%) with no difference by study arm. From time of enrollment, GO was significantly associated with better overall EFS (hazard ratio (HzR) 0.83 (95% CI: 0.7-0.99; p=.04) and relapse-free survival (HzR 0.74 (0.6-0.93; p=.01)) whereas OS was not significantly improved (HzR 0.91 (0.74-1.13)). At 3 yr, noGO v GO EFS was 47 v 53%, p=.05 and OS was 65 v 69%, p=.18. In multivariate analyses, GO was significantly associated with improved EFS than standard therapy after adjustment for significant adverse risk factors: age <2yr, initial WBC>100,000x10⁹/L, and black race. GO was not associated with significantly better induction complete remission (CR) when compared to standard therapy (88% v 85%; p=NS). Table 2 illustrates the overall results by randomized arm and by overall risk group from time of CR showing a consistent reduction in relapse risk in all risk groups. Further risk group analysis found several unique results. In the LR group, relapse rates (RR) trended lower in the GO arm but the benefit was reduced by TM during Int 2 & 3 that was significantly worse in the GO arm (3 v 10%, p=.02). In the IR group, EFS, RR, and OS trended towards improvement with GO. However when pts were censored at the time of SCT (as treated), the outcomes were not significantly different between arms (log-rank p=.14 EFS and p=.81 OS). Table 1 shows an imbalance for SCT received in assigned IR pts with fewer actually receiving SCT in the noGO arm. As such, Table 2 shows outcomes by intent-to-treat for IR pts with MFD censored and by as treated for IR pts receiving MFD SCT. Noteworthy in the IR group alone who did receive a SCT, GO arm pts had a significantly better DFS than the noGO arm (intent to treat: p=.022; as treated: p=.044). However, for the IR pts randomized to the noGO arm, SCT failed to provide benefit. For HR pts achieving CR they as well trended towards better survival & RR in the GO arm.

GO improves EFS in children, adolescents and young adults with AML by reducing the risk of relapse among those achieving remission.

No relevant conflicts of interest to declare.