Azacitidine and Venetoclax with or without Donor Lymphocyte Infusion in Relapsed AML Post Allo-HSCT: A Single-Institution Retrospective Study

Background: Azacitidine and Venetoclax (Aza/Ven) is a standard frontline treatment regimen for newly diagnosed older/unfit AML patients (pts) and is commonly used for pts with relapsed/refractory (R/R) AML, particularly those who are not candidates for multi-agent salvage chemotherapy. However, there is limited data on safety and clinical activity of Aza/Ven in pts with relapsed AML post-allogeneic hematopoietic stem cell transplantation (alloHSCT) where overall outcomes are poor. We investigated the clinical outcomes of pts treated with Aza/Ven +/- donor lymphocyte infusions (DLI) in relapsed AML pts post-alloHSCT. Methods: We conducted a retrospective analysis of pts treated with Aza/Ven for relapsed AML post-alloHSCT between January 2015 to June 2024. Eligibility criteria included pts with relapsed AML (>5% blasts) prior to treatment with Aza/Ven. Clinical activity assessments included overall response rates (ORR: CR+CRi+CRh+MLFS), composite complete remission rates (CRc: CR+CRi+CRh), and complete remission (CR) rates based on 2022 ELN criteria. Event Free Survival (EFS) and Overall Survival (OS) were calculated from the first date of Aza/Ven treatment to treatment failure, relapse, and/or death alone, respectively. Duration of response (DOR) was measured from time of overall response to relapse or death. Results: Overall, 26 pts were treated with Aza/Ven for relapsed AML post-alloHSCT. 61.5% of pts were female. Prior to alloHSCT, 25/26 pts received induction therapy versus 1 patient treated with Aza/Ven. Allograft types were matched unrelated (n=12), matched related (n=7), and haploidentical (n = 7) with a majority receiving reduced intensity conditioning (n = 18; 69%). The median age at the time of post-alloHSCT relapse was 62 years (Range: 27-72 years) and median time to post-HSCT relapse was 12.6 months (Range: 2.5 - 66.2 months). The median donor chimerism at time of post-HSCT relapse was 82.5% (Range: 47 to >95%). Nineteen pts (73%) received Aza/Ven as the first line (1L) of treatment post-HSCT versus 2L (n=3), 3L (n=3), and 4L (n=1). At time of treatment initiation, 96.2% of pts received full dose azacitidine (75mg/m2 x 7 days) and full dose venetoclax (100-400mg x 28 days). Median number of cycles of Aza/Ven was 3 (Range: 1-27). Dose reduction of Aza and/or Ven was required in 42.% (n=11) of pts due to cytopenias (n=6), and illness/frailty (n =5). The ORR, cCR and CR rates were 69.2%, 57.7%, and 38.5%, respectively. The median time to best response was 1.9 months. A total of 11 pts (42.3%) achieved MRD negativity by flow cytometry (<0.01%) after a median 1 cycle (Range: 1 - 4). The median DOR was 6.6 months (Range: 0.3 - 32.9 ). The medianEFS and OS from Aza/Ven day 1 were 4.8 months and 15.0 months, respectively, for the entire cohort. When stratified by line of treatment, median OS was 17.4 months (1L), 5.3 months (2L), 4.4 months (3L), and 5.8 months (4L). Sixteen (61.5%) pts received DLI in this cohort; 10/16 (56.3%) pts received DLI prior to Aza/Ven. At the time of post-HSCT relapse, the median donor chimerism in this subgroup was 76% (Range: 50 to >95%). Pts receiving DLI before Aza/Ven had a median OS of 38.9 months and a median EFS of 4.0 months. Of pts receiving DLI after Aza/Ven, 4/6 (66.7%) remain in CR with the DOR ranging from 4.9 - 32.4 months. There was a numerically higher ORR (80% vs.100%) and cCR (60 vs 83.3) in pts who received DLI after Aza/Ven. Independent of DLI, Graft vs Host Disease (GVHD) of any severity developed in 17 (65.4%) pts in the cohort. Of pts receiving any DLI, 4/16 (25%) were diagnosed with GVHD after receiving . Only 1 patient required systemic (oral steroids) treatment for acute skin GVHD occurring 11 months from his 3rd DLI. As of July 2024, 7 pts remain in ongoing cCR after Aza/Ven including one pt with ongoing CR after 33.6 months (30 cycles) of Aza/Ven (Ven was discontinued after cycle 17 due to pt preference). Conclusions: Outcomes are dismal in pts with relapsed AML post-alloHSCT. This single-institution retrospective analysis reveals favorable clinical outcomes of relapsed AML post-alloHSCT treated with Aza/Ven. These limited findings suggest that early DLI intervention in addition to Aza/Ven may improve overall outcomes. Further data is warranted to investigate the optimal sequence and duration of Aza/Ven in relapsed AML post-alloHSCT.