Toward a harmonized assessment of MRD by NGS: blinded pilot study of NGS assays by MPAACT consortium.

Abstract

Background: Measurable residual disease (MRD) is a key prognostic factor in acute myeloid leukemia (AML). The Measurable residual disease Partnership and Alliance in Acute Myeloid Leukemia Clinical Treatment (MPAACT) Consortium aims to accelerate AML therapy development by validating MRD as an early clinical trial endpoint within the regulatory framework. This blinded pilot study evaluated commercial next-generation sequencing (NGS) MRD assays to identify those meeting technical, regulatory, and quality criteria for development.

Methods: Of 19 vendors meeting eligibility criteria, eight responded and seven assays from six vendors proceeded to blinded testing. Using 45 contrived DNA samples with variant allele frequencies (VAFs) from 0.1% to 15%, MPAACT evaluated sensitivity, specificity, reproducibility, accuracy, and dynamic range at 50 ng and 200 ng DNA inputs from vendor-submitted raw and annotated data.

Results: Two assays were eliminated due to inadequate variant detection or gene coverage. Among the remaining five, sensitivity varied and none reliably detected VAFs <0.5%. Increased DNA input improved low-VAF detection but elevated false positives. Assays 3, 4, and 7 showed the strongest performance; Assay 7 had the lowest detection limit but higher false positives.

Conclusions: Commercial NGS assay performance varied widely, particularly at low VAFs. Selected assays merit further optimization for MRD assessments.