Conventional Cytogenetics, Molecular Profiling, and Flow Cytometric Response Data Allow the Creation of a Two-Tiered Risk-Group System for Risk-Based Therapy Allocation In Childhood AML- a Report From the Children's Oncology Group

Abstract 761

Conventional cytogenetics and morphologic response to induction chemotherapy have historically provided the tools for predicting outcome in patients with acute myeloid leukemia (AML). Several mutations are predictive of clinical outcome and have provided additional tools to help predict outcome in patients with AML without other risk features. Despite these recent advances, our ability to identify specific risk groups has been limited to a subset of patients, and nearly half of the patients with AML are regarded as having standard-risk (SR) disease. We inquired whether adding response by multidimensional flow cytometry (MDF) to data from conventional cytogenetics analysis and presence/absence of genomic alterations of FLT3 (FLT3-ITD), CEBPA, and NPM would provide a more robust risk-stratification system for risk-based therapy allocation. COG AML protocol AAML03P1 collected comprehensive cytogenetics characteristics, mutation profile (FLT3/ITD, CEBPA, and NPM mutation status), and MDF data on most of the 340 eligible patients enrolled on the study. Molecular and cytogenetic data were available for 275 of the 293 (94%) patients with responsive disease at the end of induction I. Disease-free survival (DFS) from the end of induction I was determined based on a combined molecular and cytogenetic risk profile. Risk status was defined based on the presence of t(8;21), inv(16), NPMc, and CEBPA mutations (favorable risk, FR) or the presence of monosomy 7, monosomy 5/del5q, and high allelic ratio FLT3-ITD (high risk, HR). On the basis of this allocation, 88 (32%) patients had FR AML, and 26 (10%) had HR AML. The remaining 161 (59%) patients without specific risk features were considered to have SR AML. DFS at 2 years from the end of induction I was 70%±12% for the FR cohort, 55% ±9% for the SR cohort, and 17%±20% for the HR cohort (p<0.001). The prevalence and prognostic significance of minimal residual disease (MRD) were assessed in the 103 patients with SR disease; 31 (30%) had evidence of MRD by MDF. DFS at 3 years from the end of induction I was significantly worse for those with MRD than for those without it (26%±21% vs. 67%±13%, p=0.01). Corresponding relapse risk in patients with or without MRD was 69%±21% and 30%±13%, respectively (p=0.011). We assessed the clinical impact of MRD in patients with HR or FR disease. Of the 18 patients with HR AML who had MRD data, 8 (44%) had MRD. DFS at 2 years for patients with HR disease with or without MRD was not significantly different (13%±23% vs. 36%±40%; p=0.127). Of the 73 patients with FR AML, MRD was detected in 12 (16%); MRD did not significantly influence DFS at 3 years from the end of induction I in those patients (45%±33% vs. 72%±17%, p=0.138). Thus, although the presence of MRD was significantly associated with worse outcome in patients with SR AML, similar significance could not be demonstrated in the FR or HR cohorts. Clinical outcomes of risk groups were reassessed after combining the MRD data with specific cytogenetic and molecular risk groups, i.e., patients with SR AML and MRD were added to the HR cohort, and those without MRD were added to the FR cohort. In the new risk-stratification system, 57 of 217 (26%) patients were in the HR cohort, and the remaining 160 (74%) patients were in the FR cohort. DFS at 3 years from induction I was 68%±9% for the FR cohort and 20%±16% (p<0.001) for the HR cohort. Cumulative incidence of relapse at 3 years from the end of induction I for those with FR or HR disease was 27%±9% and 71%±17%, respectively (p<0.001). Cytogenetics, molecular genotyping, and post-induction MDF analysis provide a robust means of stratifying all pediatric patients with AML into 2 risk groups with significantly different outcomes. This novel risk-allocation schema will be implemented in the upcoming COG Phase III AML trial.