Abstract 2023: AccuScan: an ultra-sensitive and robust MRD detection using WGS with single-read error correction

Abstract Whole-genome sequencing (WGS) of cell-free DNA (cfDNA) holds tremendous potential for detecting molecular residual disease (MRD), but its accuracy has been constrained by errors introduced during library preparation and sequencing. To overcome these limitations, we introduced AccuScan, a novel and robust cfDNA WGS technology that enables genome-wide error correction at the single-read level, achieving an error rate of 4.2 × 10-7, which is orders of magnitude improvement over conventional read-centric de-noising approaches. AccuScan has demonstrated analytical sensitivity down to parts per million (ppm) circulating variant allele frequency (cVAF), while maintaining 99% specificity at the sample level. By combining the convenience of WGS with the sensitivity and specificity of tumor-informed approaches, AccuScan enables precise detection of MRD in a timely manner, addressing critical unmet needs in colorectal cancer (CRC) adjuvant therapy decision-making. In the current clinical study involving 56 CRC patients who went through curative intended surgery (stages I: 7, 12.5%; stage II: 17, 30%; stage III: 26, 46.4%; stage IV: 6, 10.7% ), AccuScan achieved ∼88% sensitivity for MRD detection from a single sample collected within six weeks post-surgery. The test showed a strong correlation with recurrence risks over a three-year follow-up period while delivering 100% specificity. Notably, AccuScan captured MRD at <100 ppm cVAF, demonstrating its ultra-sensitive detection capabilities at the landmark timepoint when adjuvant therapy decisions need to be made. With a simple and distributable workflow, AccuScan demonstrated high performance and a high success rate in processing cfDNA samples across a broad range of input DNA quantities. Its current performance is powered by a straightforward statistical model that leverages single nucleotide variant (SNV) information. Future advancements in performance are anticipated with expanded cfDNA WGS datasets and the incorporation of cutting-edge machine learning strategies. Citation Format: Xinxing Li, Lui Ng, George Yeung, Yingyu Wang, Paul Tang, Tobias Wittkop, Malek Faham, Li Weng, William Cho, Dominic, Chi-chung Foo, Zhiqian Hu. AccuScan: an ultra-sensitive and robust MRD detection using WGS with single-read error correction [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 2023.