Azacitidine Plus Venetoclax Maintenance in Acute Myeloid Leukemia Produces Sustained Remissions with Low Toxicity: Results of a Phase 2 Study

Background: Most patients diagnosed with acute myeloid leukemia (AML) achieve remission following induction therapy. However, disease relapse remains a major clinical concern. The QUAZAR AML-001 study demonstrated improved overall and relapse-free survival with oral azacitidine (CC-486) maintenance. We designed a phase 2 study (NCT04062266) to evaluate low doses of azacitidine plus venetoclax as maintenance therapy following both intensive and low-intensity induction in AML. We present updated results of this study with close to 2 years of median follow-up.

Methods: Patients ≥ 18 years of age with non-APL AML in first complete remission (CR) or CR with incomplete count recovery (CRi) not immediately eligible for allogeneic stem cell transplantation (SCT) were eligible. Cohort 1 enrolled patients who had received at least 2 cycles of intensive therapy (defined as containing intermediate to high-dose cytarabine). Cohort 2 enrolled patients who had received at least 3 cycles of low-intensity therapy (defined as hypomethylating agent or low-dose cytarabine-based). Patients with positive measurable residual disease (MRD) by flow cytometry were eligible in CR1 or beyond. Other inclusion requirements were ECOG ≤ 3, adequate hepatic/renal function, and adequate bone marrow reserve (defined as absolute neutrophil count > 0.5 x 10⁹/L and platelets > 30 x 10⁹/L). Treatment consisted of azacitidine 50 mg/m² SQ/IV on D1-5 and venetoclax 400 mg PO (adjusted for concomitant CYP3A inhibitors) on D1-14 in 28-day cycles, up to 24 cycles. Venetoclax duration could be reduced to D1-7 per physician discretion to mitigate myelosuppression and dose modifications during subsequent cycles were allowed. Antibacterial, antifungal, and antiviral prophylaxis were recommended in all patients. The primary objective was relapse-free survival (RFS, defined as time from enrollment to progression or death). Secondary objectives included overall survival (OS), safety/toxicity, and MRD clearance. Patients subsequently becoming eligible for SCT were taken off study and censored for analyses at the time of SCT.

Results: 35 patients have been enrolled (25 in cohort 1 and 10 in cohort 2). The median age was 47 years in cohort 1 and 71 years in cohort 2. The baseline characteristics are shown in table 1. The median number of cycles given was 9 (range 1-24). 22/35 (63%) patients had 7 days of venetoclax during cycle 1. Median follow-up is currently 20.9 months (m). For the full cohort, the median RFS was not reached (NR; 2-year 60%) and the median OS was NR (2-year 75%). In cohort 1, the median RFS and OS were NR (2-year 71%) and NR (2-year 77%), respectively. In cohort 2, the median RFS and OS were 23.5 m (2-year 34%) and NR (2-year 67%), respectively. When stratified by ELN 2017, patients who were favorable (n=16), intermediate (n=11), and adverse (n=8) had a median RFS of NR (2-year 85%), NR (2-year 64%), and 4.0 m (2-year 0%), respectively. Corresponding median OS were NR (2-year 91%), NR (2-year 76%), and 14.9 m (2-year 27%), respectively. 4/8 (50%) of the ELN adverse patients had complex cytogenetics and/or TP53 mutations. Patients who were MRD-positive at enrollment (n=7, 3 with prior venetoclax exposure) had significantly shorter RFS compared to those who were MRD-negative (median 4.0 m versus NR, p<0.01). Only 2/7 (29%) MRD-positive patients converted to MRD-negative while on maintenance. As mutations in NPM1, IDH1, or IDH2 may sensitize AML to venetoclax, we analyzed RFS based on the presence or absence of these mutations (figure 1). Patients with sensitizing mutations had a trend towards longer RFS (median RFS NR versus 23.5 m, 2-year RFS 79% vs 36%, p=0.08).

The most common grade 3/4 adverse events were infections (26%), neutropenia (17%), thrombocytopenia (17%), hypotension (6%), and neutropenic fever (6%). 4 (11%) patients required cycle 2 venetoclax dose reductions. 8 (23%) of patients went off protocol to receive SCT. All observed deaths (n=8) occurred following AML relapse (n=7) or SCT-related complications (n=1).

Conclusions: Lower dose azacitidine plus venetoclax is a feasible and safe maintenance strategy in AML. This approach yields encouraging results in MRD-negative patients with non-adverse ELN risk (favorable or intermediate risk). A genetic-guided strategy may help select patients who benefit the most from this combination (i.e. those with NPM1, IDH1, or IDH2 mutations).