Abstract 4577: Ultra-sensitive detection of circulating tumor DNA using methylation anchor probes to enhance hepatocellular carcinoma (HCC) detection

Abstract Background: Hepatocellular carcinoma (HCC) is one of the leading causes of cancer-related death worldwide. Circulating cell-free DNA (cfDNA) methylation markers offer a promising approach for early cancer detection and therefore improve patient outcomes. Among the many CpG sites with altered methylation status, only specific methylation patterns (haplotypes) are disease-related and are often rare in clinical samples. Current cfDNA-methylation signal-based methods lack sufficient sensitivity, particularly in detecting low-abundance tumor-derived DNA with limited sequencing reads. Here, we introduce an experimental method to address this issue by using short hybrid capture probes specifically designed to bind tumor-derived DNA methylation haplotypes, while minimizing binding to wild-type fragments. This approach reduces the number of sequencing reads required and improves assay sensitivity. Methods: Methylation haplotypes were selected based on in-house whole-genome bisulfite sequencing (WGBS) data from tumor tissue and paired normal tissues adjacent to the tumors (NATs) of HCC patients. Further filtering was applied using cfDNA data from a group of healthy individuals. Short hybrid capture probes were then rationally designed to specifically target HCC-related methylation haplotypes. The panel was tested on 118 clinical samples (57 HCC patients and 61 controls). Data processing included the quantification of haplotype frequencies, followed by mutual information-based feature selection to identify the most distinctive markers in the training data to enhance classification performance. Finally, a boosted tree model (XGBoost) was applied to predict cancer. Results: All samples were divided into a training set (75%) and a test set (25%), ensuring a balanced distribution of controls and cancer stages. The top 128 features were selected for both early (Stage I & II) and late (Stage III & IV) HCC. Across all stages of HCC, the panel achieved 97.6% specificity and 97.2% sensitivity in the training set, and 94.7% specificity and 95.2% sensitivity in the test set. Conclusions: We developed a novel and cost-effective method to specifically enrich HCC-related methylation haplotypes, enabling accurate and ultra-sensitive detection of HCC. In the future, we will test more HCC cancer, precancer, and benign samples to further assess our assay performance. Also, our method is versatile and holds potential for the early detection and monitoring of other cancer types, paving the way for broader applications in precision oncology. Citation Format: Nan Lin, Ying Mei, Nina Guanyi Xie, Yangjunyi Li, Liyuan Zhao, Zhihui Xu, Chen Zhu, Yujie Chen, Yalong Zhang, Bingqiang Fu, Jinlei Song, Min Li, Xiaohui Wu, Xueguang Sun. Ultra-sensitive detection of circulating tumor DNA using methylation anchor probes to enhance hepatocellular carcinoma (HCC) detection [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 4577.