MRD conversion during oral azacitidine maintenance correlates with outcomes in predominantly favorable-risk AML.

Oral azacitidine improves survival in patients with AML in first complete remission who are not candidates for allogeneic transplantation. Most favorable-risk AML patients harbor molecular markers enabling sensitive measurable residual disease (MRD) monitoring. While MRD is a well-established prognostic factor in AML, the clinical significance of MRD dynamics during maintenance therapy remains incompletely defined. We conducted a multicenter, real-world retrospective study of 30 patients with AML, predominantly ELN favorable-risk, treated with oral azacitidine maintenance following intensive induction-consolidation. MRD was assessed longitudinally using standardized molecular and flow-based assays. Outcomes were analyzed according to MRD status at maintenance initiation and subsequent MRD conversion. At initiation of oral azacitidine, 47% of patients were MRD-positive. MRD conversion to negativity occurred in 64% of these patients during maintenance. Relapse-free survival (RFS) was comparable between patients who were MRD-negative at baseline and those who converted to MRD negativity (24-months restricted median RFS 577 vs. 638 days), whereas patients with persistent MRD positivity experienced early relapse (24-months restricted median RFS 63 days). Despite inferior RFS, overall survival did not significantly differ between MRD-defined groups. Patients relapsing after oral azacitidine retained sensitivity to subsequent therapies, including venetoclax-azacitidine-based regimens, with high rates of molecular response. Oral azacitidine was associated with high rates of MRD conversion, conferring RFS comparable to patients who were MRD-negative at treatment initiation. Persistent MRD identified patients at high-risk for early relapse, while effective salvage therapies mitigated overall survival differences. These findings support prospective evaluation of MRD-guided maintenance strategies in favorable-risk AML but should be considered hypothesis-generating given the retrospective design and limited cohort size. The trial is on behalf of the Israel Acute Leukemia Group.