Azacitidine Followed By Intensive Induction/Consolidation Chemotherapy in Older Patients with Acute Myeloid Leukemia (AML): Results from the Randomized AML-AZA Trial of the Study Alliance Leukemias (SAL)

Introduction

Aberrant DNA methylation is a common feature of acute myeloid leukemia (AML) and increases with age. DNMT inhibitors such as Azacitidine (AZA) can induce meaningful responses and remissions in AML as monotherapy. The combination of AZA with standard chemotherapy (7+3) has not been tested in a randomized trial.

Patients and study design

The AML-AZA trial compared AZA directly followed by standard induction therapy, AZA followed by standard consolidation, and further Azacitdine maintenance with standard induction and consolidation without AZA in older patients with AML. All patients received standard Cytarabine (100 mg/sqm) and Daunorubicin (60 mg/sqm) induction (“7+3”) and up to two cycles of intermediate dose Cytarabine (1 g/sqm q12hr days 1, 3, 5) as consolidation therapy. AZA (75 mg/sqm for 5 days) preceded each therapy cycle in the AZA arm. In addition, AZA maintenance for up to 1 year was also scheduled for patients in the AZA arm. 105 patients were randomized to receive AZA plus Cytarabine plus Daunorubicin as induction therapy (AZA + 7+3) and 109 patients to receive 7+3 only (control group). Median age was 70 years in both treatment arms. Patient cohorts were well balanced with regard to blast counts in bone marrow, secondary versus de novo AML and molecular genetics risk group. More patients in the AZA + 7+3 arm (39/105; 37.1%) than in the control group (25/109; 22.9%) showed high risk cytogenetics (p=0.057). Event free survival (EFS) was the primary end point. Secondary endpoints were overall survival (OS), complete remission (CR) rate, toxicity and different treatment response according to molecular markers.

Results

Overall, 214 of 216 planned patients were enrolled into the AML-AZA trial. Due to a higher number of severe adverse events (SAE), AZA administration was stopped after recruitment of 214 patients whereas chemotherapy was continued as planned. Percentages of patients in the AZA arm with AZA doses as initially planned were as follows: 99% for first induction cycle, 72% for the second induction cycle. AZA as maintenance therapy for at least one cycle was delivered to 18% of patients in the AZA group. At least one SAE occurred in 51% of AZA + 7+3 patients compared to 31% of 7+3 patients (p=0.005). Cardiac disorders with CTCAE grade 3-5 occurred more frequently in the AZA + 7+3 arm (n = 15) than in the 7+3 arm (n = 6) (not significant). Leukopenia was prolonged by one day (median 23 vs 22 days) in the AZA + 7+3 group (p=0.043), whereas time of thrombocytopenia was not different. The early death rates at 30, 60 and 90 days did not differ significantly between treatment groups. Efficacy analyses were performed on an intention-to-treat basis. Median EFS as the primary endpoint was 6 months in both treatment arms (p=0.96). Median OS was 16 months for patients treated with AZA + 7+3 and 21 months for 7+3 (p=0.35). Median relapse free survival was 12 months in both treatment arms (p=0.95). 48 of 100 patients (48%) in the AZA + 7+3 arm achieved complete remission (CR) after induction therapy versus 57 of 109 patients (52%) in the 7+3 arm (p=0.58). DNMT3A exon 23 mutations were detected in 30 out of 162 analyzed patients. Exploratory analyses were performed to detect a potential interaction between AZA + 7+3 response and DNMT3A mutation status. Trends for improved EFS and OS were noted for AZA + 7+3 treatment in DNMT3A mutated patients.

Conclusion

AZA as addition prior to standard induction and consolidation chemotherapy does not prolong EFS and OS in unselected older AML patients and it is more toxic. However, a trend towards better efficacy in patients with DNMT3A mutation was observed and should be further explored.