Comparing induction strategies in acute myeloid leukemia: 10-year real-world experience with 7+3 and HMA/venetoclax.

Abstract

6524 Background: Hypomethylating agents (HMA) combined with venetoclax are the standard of care for AML in adults ineligible for intensive chemotherapy, but use is expanding to fit patients given prospective data suggesting comparable outcomes. Whether these findings are reproducible in real-world practice remains under investigation. Methods: We conducted a retrospective cohort study of adults with newly diagnosed AML treated at the University of Kentucky from 2016-2025 with either 7+3 or HMA/venetoclax. Overlap propensity score weighting balanced age, sex, ECOG, ELN risk group, prior MDS, and therapy-related AML. Primary endpoints were CR/CRi, RFS, OS, and 30-day mortality. Cox models included allogeneic HSCT as a time-varying covariate. Survival outcomes were stratified by ELN risk group and mutational risk group according to venetoclax sensitivity (favorable: NPM1 and IDH1/2 , adverse: TP53 and FLT3 ). Results: Prior to weighting, patients treated with HMA/venetoclax were older (68 vs. 50 years), had worse performance status (mean ECOG 1.2 vs. 0.7), and were more likely to have prior MDS, therapy-related AML, and less favorable ELN risk. Unweighted data demonstrated a median OS of 13.3 months for HMA/venetoclax and 19.6 months for 7+3. After overlap weighting of 379 patients (250 received 7+3 and 129 received HMA/venetoclax), acceptable covariate balance was achieved with all standardized differences < 0.01. HMA/venetoclax was associated with numerically higher odds of achieving CR/CRi compared with 7+3 (OR 1.84, p=0.09), and this reached significance in adverse-risk disease (OR 2.31, p=0.006). 30-day mortality was similar between 7+3 and HMA/venetoclax groups (8.3% vs. 8.0%, p=0.96). Among patients who achieved CR/CRi, no significant difference in relapse was noted (p=0.24). RFS and OS did not differ significantly between treatment groups (p=0.76 and 0.18, respectively). Analysis of ELN risk subgroups and mutation subgroups incorporating venetoclax sensitivity demonstrated no association with treatment group in survival outcomes (for OS, p=0.53 and 0.43, respectively), suggesting no effect modification from these variables. Allogeneic HSCT was strongly associated with improved OS (HR 0.26, p<0.001), which persisted after adjusting for ELN risk (HR 0.39, p<0.001) and mutational subgroups (HR 0.44, p=0.001). Across all models, treatment assignment did not significantly impact OS once transplant was modeled as a time-dependent variable. Conclusions: After accounting for baseline confounding and allogeneic HSCT as a time-varying covariate, survival in AML is primarily driven by molecular risk profile and transplant rather than initial induction strategy. These results support further investigation of HMA/venetoclax as a frontline alternative in fit patients with AML and the intention to proceed to transplant.