AUTHORS
Eric Duncavage, MD, Haley Abel, PhD2, Andrew E.O. Hughes, MD, PhD3, Rohan Kodgule, MD4, Yi-Shan Lee, MD, PhD5, Joshua Robinson6, Tianjiao Wang, PhD7, Cara Shirai, PhD8, Bijal Parikh, MD, PhD9, Sharon Heath10, Daniel C. Link, MD11, John F. DiPersio, MD12, Samuel Urrutia, MD, MS13, Ramzi Abboud, MD14, Timothy Ley, MD15, Meagan A. Jacoby, MD, PhD16, Matthew J. Walter, MD17, Geoffrey L. Uy, MD18, David H Spencer, MD, PhD19
AFFILIATIONS
1 Washington University School of Medicine, St Louis, Missouri, United States
2 Washington University School of Medicine, Saint Louis, Missouri, United States
3 Washington University School of Medicine, St Louis, Missouri, United States
4 Washington University School of Medicine, St Louis, Missouri, United States
5 Washington University School of Medicine, St Louis, Missouri, United States
6 Washington University School of Medicine, St Louis, Missouri, United States
7 Washington University School of Medicine, St Louis, Missouri, United States
8 Washington University School of Medicine, St Louis, Missouri, United States
9 Washington University School of Medicine, St Louis, Missouri, United States
10 Washington University School of Medicine, Saint Louis, Missouri, United States
11 Washington University School of Medicine, Saint Louis, Missouri, United States
12 Washington University School of Medicine, Saint Louis, Missouri, United States
13 Washington University School of Medicine, Saint Louis, Missouri, United States
14 Washington University School of Medicine, Saint Louis, Missouri, United States
15 Washington University School of Medicine, Saint Louis, Missouri, United States
16 Washington University School of Medicine, Saint Louis, Missouri, United States
17 Washington University School of Medicine, Saint Louis, Missouri, United States
18 Washington University School of Medicine, Saint Louis, Missouri, United States
19 Washington University School of Medicine, Saint Louis, Missouri, United States
ABSTRACT
Background: Measurable residual disease (MRD) is an important predictor of relapse and survival in acute myeloid leukemia (AML) and is increasingly used in higher-risk myelodysplastic syndromes (MDS). MRD is commonly evaluated using multiparameter flow cytometry (MFC) to assess for abnormal antigen expression patterns, or by PCR for recurrent gene fusions or individual mutations (i.e., insertions in NPM1 or FLT3). Tumor informed sequencing panels capable of detecting somatic mutations present at diagnosis have recently been applied to AML and MDS MRD detection. Here, we investigated the clinical performance and patient outcomes of sequencing-based MRD in a real-world cohort of AML and MDS patients who were tested with MyeloSeq MRD, a tumor-informed, error-corrected sequencing panel that simultaneously tracks multiple diagnostic mutations in peripheral blood or bone marrow.
Methods: We identified 173 patients with AML and 74 with MDS who underwent initial diagnostic testing and evaluation at follow-up timepoints for molecular MRD between May 2022 and January 2025 using MyeloSeq MRD, an error-corrected sequencing assay targeting 50 genes that performs ~10,000x coverage in consensus reads to achieve a maximum analytic limit of detection of 0.05% variant allele frequency (VAF). A subset of samples was tested with MFC using an established reference laboratory (Hematologics, Inc., Seattle, WA). MyeloSeq MRD results from follow-up specimens were classified as molecular MRD-positive when previously identified diagnostic mutations were detected above the validated limit of detection. Mutations in DNMT3A, TET2, or ASXL1 (DTA) were not used for MRD determination except in patients who had received an allogeneic hematopoietic cell transplant (alloHCT). MFC-based MRD used a difference-from-normal approach and patients with suboptimal or inadequate specimens or equivocal results were excluded from analysis. Pathologic relapse was defined by ≥5% marrow blasts or recurrent disease-associated cytogenetic abnormalities.
Results: After excluding 21 patients with fusion-driven AML suitable for established RT-qPCR monitoring, 219 of 226 patients (97%) had at least one trackable non-DTA mutation at diagnosis, with means of 3.8 mutations per patient in AML and 3.1 in MDS. The paired MFC analysis included 470 AML and 226 MDS specimens from 148 and 71 patients, respectively. Overall, 49.6% of follow-up specimens were MyeloSeq MRD-positive compared with 32.3% by MFC. Among MFC-positive specimens, 96.0% of AML and 96.5% of MDS specimens were also positive by MyeloSeq MRD. Conversely, MyeloSeq MRD detected residual mutations in 21.4% of MFC-negative AML specimens and 40.7% of MFC-negative MDS specimens. MyeloSeq MRD demonstrated 99.0% sensitivity relative to MFC when abnormal cells comprised ≥0.25% of the specimen and 97.5% sensitivity at an MFC burden of ≥0.1%. Among 116 patients with discordant results where MyeloSeq MRD was positive and MFC was negative, 47 (41%) progressed to a clinical relapse, transplantation, death, or conversion to MFC positivity by the next MRD timepoint or 60 days. MyeloSeq MRD positivity preceded pathologic relapse by longer intervals than MFC positivity, including median pre-transplant intervals of 130 versus 88 days in AML and 186 versus 83 days in MDS. Among patients with at least three evaluable MRD timepoints, each additional MyeloSeq-positive result was associated with inferior relapse-free survival (hazard ratio 1.71, P=3×10⁻⁴).
Conclusion: A streamlined, tumor-informed sequencing panel provided trackable MRD markers in nearly all evaluable patients with AML or MDS. MyeloSeq MRD identified substantially more positive specimens than MFC, retained high sensitivity for MFC-detectable disease, and frequently anticipated subsequent clinical events. Repeated sequencing-based MRD positivity further stratified relapse risk, supporting longitudinal, high-depth multigene NGS as a broadly applicable and operationally simple approach to MRD monitoring in myeloid malignancies.
DISCLOSURES
Eric Duncavage, MD: BrightDx - Current equity holder in private company; Genescopy - Consultancy (Includes expert testimony); Cofactor Genomics - Consultancy (Includes expert testimony); Blueprint Medicines - Consultancy (Includes expert testimony); BrightDx - Consultancy (Includes expert testimony); Caris Life Sciences - Consultancy (Includes expert testimony); Bruker Spatial Biology - Consultancy (Includes expert testimony); Syndax - Consultancy (Includes expert testimony)