Dual MRD assessment improves relapse prediction in AML

Combining multiparameter flow cytometry (FACS) and error-corrected next-generation sequencing (NGS) for measurable residual disease (MRD) assessment improved relapse prediction in patients with acute myeloid leukaemia (AML). This dual-assessment strategy may also help identify appropriate candidates for early intervention following allogeneic hematopoietic stem cell transplantation (allo-HCT).

Accurate MRD assessment is critical in AML management to guide treatment decisions before and after allo-HCT. FACS-MRD and NGS-MRD are considered complementary modalities. Study presenter Dr Evgeny Klyuchnikov (University Medical Centre Hamburg-Eppendorf, Germany) and colleagues aimed to find possible advantages of combining them in an observational, single-centre study [1].

The analysis included 107 adult patients with AML undergoing their first allo-HCT. FACS-MRD assessment was performed with a sensitivity of 10^-4 to 10^-5, and an MRD positivity cut-off of 0.1%. In parallel, error-free NGS-MRD analysis was conducted using the Oncomine Myeloid MRD (RUO) Assay, with a limit of detection of approximately 0.1 % variant allele frequency.

The results demonstrated that the combined FACS and NGS was approximately 10% more accurate at predicting relapse. Before transplant, MRD was detected by FACS in 37% of patients and by NGS in 45%, with moderate-to-substantial concordance between both methods (Cohen’s kappa 0.60). Notably, among patients classified as MRD-negative by FACS, error-corrected NGS identified additional subclonal variants, including FLT3-ITD mutations, in 26% of cases, indicating discordant findings between the 2 methods and suggesting that NGS may detect residual molecular disease missed by FACS alone.

MRD positivity by either FACS-MRD or NGS-MRD was significantly associated with an increased 5-year relapse risk and lower overall survival (OS). Using FACS, the 5-year OS was 39% (95% CI, 25–55%) in MRD-positive patients versus 79% (95% CI, 68–87%) in MRD-negative patients (P<0.001). Using NGS-MRD, the corresponding 5-year OS rates were 49% (95% CI, 39–63%) versus 76% (95% CI, 64–85%; P=0.006). Patients with discordant MRD findings demonstrated an intermediate-risk profile, whereas patients who were MRD-positive by both FACS and NGS had significantly higher relapse risk and lower OS (P<0.001). In multivariate analysis, dual MRD positivity remained the strongest independent predictor of relapse (HR 6.8; 95% CI 3.0–15.0; P<0.001) and inferior OS (HR 3.2; 95% CI 1.6–6.1; P=0.001).

Dr Klyuchnikov concluded that combined MRD assessment using FACS and NGS reduces the frequency of false positives and improves relapse prediction, although it does not significantly improve OS prediction.

Klyuchnikov E, et al. Combined pre-transplant MRD detection by FACS and NGS may enhance relapse prediction in AML patients receiving allo-SCT in complete remission. B293, EBMT congress 2026, 22-25 March, Madrid, Spain.