Prognostic Factors For Children With Acute Myeloid Leukemia Who Achieve Minimal Residual Disease-Negative Status After Induction Therapy

Despite successful remission induction, a significant proportion of children with acute myeloid leukemia (AML) experience relapse. In our previous AML02 study, adjusting treatment intensity according to levels of minimal residual disease (MRD) after remission induction has contributed to improved outcome. Because relapse still occurred in a proportion of patients who achieved MRD-negative status after remission induction therapy, we searched for presenting features associated with outcome in this patient cohort.

To identify presenting features predictive of relapse among children with newly diagnosed AML who achieve MRD-negative status after remission induction therapy.

Among patients enrolled in AML02 who were MRD-negative (<0.1% leukemic cells among bone marrow mononuclear cells by flow cytometry) at the end of remission induction therapy, we evaluated the association between presenting features [age, gender, race, WBC count, 11q23 non-t(9;11), acute megakaryoblastic leukemia (AMKL) without t(1;22), FLT3-ITD, CBF rearrangement [t(8;21), inv(16)/t(16;16)], and provisional risk classification at study entry as defined by cytogenetics and AML FAB subtype] and risk of relapse. Event free survival (EFS) and overall survival (OS) were analyzed by a Cox proportional regression model. A stepwise model selection of predictors identified in the univariate analysis was then performed in the multivariate analysis.

Of the 197 patients enrolled in AML02 with evaluable bone marrow after induction I, 126 (64%) had negative MRD. Three-year EFS for this group was 75.4% and OS was 82.5%. Of the 193 patients evaluable for MRD after induction II, 155 (80.3%) had negative MRD and their 3-year EFS and OS were 71.6% and 78.1%, respectively. Patients with negative MRD had significantly better EFS and OS than those with positive MRD in each course.

In a univariate analysis, 11q23 non-t(9;11), AMKL without t(1;22), or standard risk (compared to low risk) were associated with an inferior EFS (p <0.05) among patients with negative MRD after induction I and/or induction II. In a multivariable analysis performed with the cohort of 126 patients who achieved MRD negativity after induction I, a worse outcome was associated with 11q23 non-t(9;11) (HR 3.21, p =0.021 for EFS; HR 5.84, p =0.005 for OS), AMKL without t(1;22) (HR 7.83, p =0.034 for OS), and age ≥ 10 years (HR 5.97, p<0.001 for EFS; HR 10.05, p <0.001 for OS). The same factors also emerged as significant predictors of adverse outcome in a multivariable analysis focused on patients who were MRD negative after induction II: 11q23 non-t(9;11) (HR 3.48, p =0.01 for EFS; HR 5.82, p =0.003 for OS), AMKL without t(1;22) (HR 4.43, p =0.016 for EFS; HR 9.7, p =0.002 for OS), and age ≥10 years (HR 3.83, p =0.003 for EFS; HR 4.13, p =0.014 for OS). Black race also predicted a worse EFS for patients who were MRD negative after induction II (HR 3.1, p =0.01). In contrast, patients with CBF rearrangement had significantly better outcomes: EFS (HR 0.18, p =0.003) and OS (HR 0.15, p =0.021) after induction I and EFS (HR 0.15, p =0.001) and OS (HR 0.22, p =0.025) after induction II.

Among children with AML who achieve MRD negativity after remission induction therapy, 11q23 non-t(9;11), AMKL without t(1;22), and age ≥ 10 years are associated with a higher risk of relapse. Close post-remission MRD monitoring for detection of emerging relapse and complementary molecular methods of MRD detection might be indicated in these patients.

No relevant conflicts of interest to declare.