PURPOSE Patients with acute myeloid leukemia (AML) in remission remain at risk for relapse even after allogeneic hematopoietic cell transplantation (alloHCT). AML measurable residual disease (MRD) status before alloHCT has been shown to be prognostic. Whether modulation of the intensity of the alloHCT conditioning regimen in patients with AML who test positive for MRD can prevent relapse and improve survival is unknown. METHODS Ultra-deep, error-corrected sequencing for 13 commonly mutated genes in AML was performed on preconditioning blood from patients treated in a phase III clinical trial that randomly assigned adult patients with myeloid malignancy in morphologic complete remission to myeloablative conditioning (MAC) or reduced-intensity conditioning (RIC). RESULTS No mutations were detected in 32% of MAC and 37% of RIC recipients; these groups had similar survival (3-year overall survival [OS], 56% v 63%; P = .96). In patients with a detectable mutation (next-generation sequencing [NGS] positive), relapse (3-year cumulative incidence, 19% v 67%; P < .001) and survival (3-year OS, 61% v 43%; P = .02) was significantly different between the MAC and RIC arms, respectively. In multivariable analysis for NGS-positive patients, adjusting for disease risk and donor group, RIC was significantly associated with increased relapse (hazard ratio [HR], 6.38; 95% CI, 3.37 to 12.10; P < .001), decreased relapse-free survival (HR, 2.94; 95% CI, 1.84 to 4.69; P < .001), and decreased OS (HR, 1.97; 95% CI, 1.17 to 3.30; P = .01) compared with MAC. Models of AML MRD also showed benefit for MAC over RIC for those who tested positive. CONCLUSION This study provides evidence that MAC rather than RIC in patients with AML with genomic evidence of MRD before alloHCT can result in improved survival.

Acute myeloid leukaemia (AML) is a blood cancer characterized by acquired genetic mutations. There is great interest in accurately establishing measurable residual disease (MRD) burden in AML patients in remission after treatment but at risk of relapse. However, inter-and intrapatient genetic diversity means that, unlike in the chronic myeloid and acute promyelocytic leukaemias, no single genetic abnormality is pathognomonic for all cases of AML MRD. Next-generation sequencing offers the opportunity to test broadly and deeply for potential genetic evidence of residual AML, and while not currently accepted for such use clinically, is likely to be increasingly used for AML MRD testing in the future.

No abstract