An Ultra-Deep Ngs-Based Mrd Approach for Longitudinal Multi-Marker Monitoring in Aml

Document Type

Conference Proceeding

Publication Date

6-10-2026

Publication Title

HemaSphere

Keywords

biological marker, acute myeloid leukemia, adult, allele, background noise, cohort analysis, conference abstract, cumin, diagnosis, flow cytometry, follow up, high throughput sequencing, human, limit of blank, limit of detection, longitudinal study, minimal residual disease, real time polymerase chain reaction, remission, reproducibility, validity

Abstract

Background Measurable residual disease (MRD) is a validated prognostic factor in acute myeloid leukemia (AML) and guides post-remission management. MRD is assessed by longitudinal detection of leukemia-specific molecular or immunophenotypic biomarkers identified at diagnosis. Current MRD strategies rely on multiparameter flow cytometry and quantitative PCR for selected molecular targets such as mutated NPM1, while FLT3-ITD is typically detected at diagnosis by fragment analysis. Although these approaches are well established, their biomarker-specific and assay-dependent nature requires complementary technologies and may limit harmonized, integrated longitudinal monitoring across genetically heterogeneous AML. To address these methodological constraints, we developed an ultra-deep targeted DNA-based NGS assay using our proprietary CUMIN© barcoding technology to enable simultaneous multi-marker tracking while distinguishing true biological signal from technical noise. Aims In this study, we aim at establishing the analytical performance characteristics, including sensitivity, and validating the technical robustness and clinical validity of our NGS-based MRD approach. Methods A curated panel of 23 AML-associated gene biomarkers relevant for MRD monitoring was selected in accordance with European LeukemiaNet (ELN) recommendations for NGS-based MRD assessment. DNA input parameters were optimized to ensure adequate molecular coverage, enabling detection of variants at MRD-relevant sensitivity thresholds. Analytical performance was evaluated using reference cell lines and clinical samples to establish the limit of blank, limit of detection, analytical sensitivity, precision, and robustness, including inter-site reproducibility. Clinical validation assessed MRD detection in peripheral blood and bone marrow samples from AML patients at diagnosis and during longitudinal follow-ups. Results Analytical validation demonstrated high sensitivity and precision for detection of confirmed variants, including FLT3-ITD, across technical replicates and inter-site testing. The limit of blank remained below the predefined limit of detection, supporting reliable low-frequency variant identification at variant allele fractions (VAF) down to 0.01%. Robust analytical performance was confirmed through consistent results across replicates, independent runs, and inter-site evaluation. Analysis of longitudinal clinical specimens demonstrated consistent detection of pathogenic variants, including FLT3-ITD when present, without observable background signal at corresponding genomic positions, supporting effective control of technical noise. Longitudinal tracking was enabled by a dedicated bioinformatic framework (Oncoportal™ Mutation Tracker; Patent pending), allowing simultaneous monitoring of multiple diagnostic variants and computation of a composite clonal likelihood score to distinguish true residual clones from background noise. Summary/Conclusion Ultra-deep, targeted DNA-based NGS enables sensitive and reproducible multi-marker MRD assessment in AML, including FLT3-ITD, with robust longitudinal tracking and effective discrimination of biological signal from technical noise. This integrated approach supports harmonized MRD monitoring within a single workflow.

Volume

10

Find It @ Sladen

Share

COinS