Venetoclax in combination with hypomethylating agents shows promising activity in Acute Myeloid Leukemia with late relapse post allogeneic stem cell transplant: From the EBMT acute leukemia working party Free

Introduction Allogeneic hematopoietic stem cell transplantation (HSCT) remains a mainstay in the treatment of acute myeloid leukemia (AML), with post HSCT relapse being a challenge. Hypomethylating agents (HMA) [5-azacitidine (AZA) or decitabine (DEC)] have shown efficacy and accepted safety when used alone or in combination with venetoclax (VEN) as treatment for newly diagnosed or relapsed AML. Post HSCT, the potential role of HMA in augmenting the graft-versus-leukemia effect led to their adoption as maintenance, preemptive, and salvage options. Their optimal combinations, particularly with VEN, and dose or duration of therapy in this setting, remains undetermined. Methods Aregistry-based multicenter analysis from the European Society for Blood and Marrow Transplantation with approval of the Acute Leukemia Working Party, evaluating the utilization of HMA [single agents or in combination with VEN (HMA/VEN)] as treatment of relapsed AML post HSCT, including adult patients who received their first HSCT in first complete remission (CR1) 2010-2022, with subsequent relapse treated with HMA or HMA/VEN within 2 months (m) post relapse, comparing their outcomes. Results: A total of 224 patients relapsed at a median of 181 days (d) from HSCT. Karyotype was adverse (adv) in 51% and intermediate in 44%; 27% had secondary AML. NPM1 and FLT3-ITD mutations were present in 17% and 18%, respectively. Most received peripheral blood stem cells (97%) from a matched identical sibling (28%), unrelated (39%), or haploidentical donor (13%). 83 patients received HMA/VEN (75 AZA/VEN), and 141 received HMA alone (128 AZA). The median time to relapse post HSCT was longer in the HMA/VEN group [269 d; IQR 142-528 versus (vs.) 157; IQR 97-356; p=0.002], with a difference in year (y) of transplant [median 2020 vs. 2016 respectively; p<0.001). Pretransplant AZA (21% vs. 8%; p=0.006) or VEN (11% vs. 2%; p=0.01), myeloablative conditioning (59% vs. 41%; p=0.01) and post HSCT cyclophosphamide (34% vs. 21%; p=0.03) were more frequently used in the HMA/VEN group. There was no significant difference with respect to gender, female donor to male recipient, type of AML (de novo in 75% vs. 72%; p=0.59), cytogenetic (CG) risk category ELN2022 (adv risk in 43% vs. 56%;p=0.13), age at HSCT (55 vs. 57 yrs; p=0.25), time from diagnosis to CR1 (1.9 vs. 1.7 m; p=0.29), to HSCT (4.6 vs. 4.4 m; p=0.47), or from relapse to salvage (13 vs. 12 d) between HMA/VEN and HMA respectively. The median follow up was longer for the HMA group (6.8 vs. 2.6 y). The 1-y cumulative incidence of CR post salvage was higher in the HMA/VEN group [44% (95% CI 33-55) vs. 26% (95% CI 19-33); p=0.003]. The 2-y overall survival (OS) from salvage was not significantly different [24% (IQR 14-35) in the HMA/VEN group vs. 16% (IQR 11-23); p=0.18]. The 2-y incidence of second (2^nd) HSCT was 27% (95% CI 17-38) in the HMA/VEN group vs. 12% (7-18) of the HMA group. Late relapse (>6 m post HSCT) and the absence of adv CG positively affected the 1-y cumulative incidence of CR post-salvage and 2-y OS, while female gender improved 2-y OS. In multivariable analysis, the absence of adv CG (hazard ratio [HR] 0.68, 95% CI 0.5-0.9; p=0.01) and late relapse were predictive of better OS [HR for early relapse 1.64, (95% CI 1.2-2.2); p=0.002]. Regarding the impact factor of cumulative incidence of CR, an interaction was found between salvage group and time of relapse. In the late relapse group, the use of HMA/VEN was associated with improved cumulative incidence of CR (HR 3.52, 95% CI 1.9-6.7; p<0.001), while in the early relapse group there was no difference between the salvage groups (HR 0.49, 95% CI 0.3-1.9; p=0.61). For the 73 patients who reached CR following post-HSCT relapse (2-y OS from CR was 47%), the median time from salvage to CR was shorter for 36 patients who received HMA/VEN compared to 37 patients who received HMA (62 vs. 115 d; p=0.008) with more patients receiving a 2^nd HSCT in the HMA/VEN group (2^nd HSCT within 2 y from CR: 46% vs. 8%) and more patients receiving donor lymphocyte infusion in the HMA group (within 2 y from CR: 22% vs. 6%). Conclusion: HMA and HMA/VEN are potential salvage therapies for post-HSCT relapse allowing long term survival of around 20% of patients, particularly those with late relapse and no adv CG. The addition of VEN to HMA appears to be specifically beneficial for those with late relapse, leading to significantly improved CR, and a higher chance of a 2^nd HSCT.