Efficacy and Molecular Predictors of Response and Survival for Venetoclax/Azacitidine Therapy in Relapsed or Refractory Acute Myeloid Leukemia

Background: Despite advances in frontline treatment for acute myeloid leukemia (AML), relapsed or refractory (R/R) AML remains a significant therapeutic challenge, with poor outcomes and limited treatment options. One recently approved treatment combination – the BCL-2 inhibitor venetoclax and a hypomethylating agent – has transformed outcomes for patients with newly diagnosed AML who are ineligible for intensive chemotherapy. This combination leverages venetoclax’s ability to induce leukemic cell apoptosis with a hypomethylating agent’s capacity for direct cytotoxicity and ability to restore the differentiation and proliferation of benign hematopoietic cells. However, the efficacy of this regimen has not been extensively studied in the R/R AML setting. We sought to evaluate outcomes and molecular predictors of response and survival for patients with R/R AML treated with venetoclax combined with a hypomethylating agent in a single center experience.

Methods: We retrospectively reviewed and analyzed 46 patients with R/R AML treated with venetoclax (VEN) and azacitidine (AZA) between December 2017 and December 2023 at University of Colorado hospital. Patients were included if they were refractory to or relapsed after at least one previous cycle of therapy and were not previously treated with VEN. VEN was escalated to 400 mg per the package insert; AZA was dosed at 75 mg/m²/day on days one to seven of a 28-day treatment cycle. Molecular analysis was performed by next-generation sequencing on bone marrow biopsies at diagnosis and immediately prior to cycle one of VEN/AZA.

Results: The median age of patients was 62 years old (range 27-91) and 54% were male. 87% of patients were ECOG 0-1 and 57% had ELN 2022 adverse risk AML. The median bone marrow blasts was 35% (range 6-99%), median number of prior therapies was 2 (range 1-5), and 13 patients (28%) had relapsed after a hematopoietic stem cell transplant (HSCT). The overall response rate was 63% (29/46), with 17 (37%) patients achieving CR, 4 (9%) patients achieving CRi, and 8 (17%) patients achieving MLFS. After a median follow up of 38.0 months, the median overall survival (OS) was 11.1 months (95% CI 6.2-16.0). Median number of cycles of VEN/AZA was 1 (range 1-11) and 23 patients (50%) went on to HSCT after VEN/AZA therapy. Mutations in EZH2 were associated with worse OS (HR 2.81, 95% CI 1.07-7.36, p=0.036), worse progression-free survival (PFS) (HR 5.99, 95% CI 2.08-17.2, p<0.001), and trended towards significantly lower response rates (p=0.055). IDH2 (HR 2.64, 95% CI 0.97-7.19, p=0.058) and STAG2 (HR 3.00, 95% CI 0.86-10.5, p=0.086) also trended towards worse PFS.

Conclusion: VEN/AZA is an effective treatment option for patients with R/R AML. These outcomes compare favorably to both high-dose cytarabine-based salvage regimens and single-agent targeted therapies. With half of patients in this cohort proceeding to HSCT following therapy, VEN/AZA could be a viable regimen to bridge R/R AML patients to transplantation. Molecular analysis revealed that EZH2 is associated with worse survival outcomes; IDH2 did not seem to confer favorable outcomes, distinct from its prognostic associations in the front-line setting, but further studies are needed to confirm the significance of these and other mutations. These findings support the continued investigation of VEN/AZA in the R/R AML setting.