BAT Seminar: Liquid Biopsy Platform for Early Liver Cancer Detection

BAT Online Seminar 1 Topic: Liquid Biopsy Platform for Early Liver Cancer Detection Speaker: Sasimol Udomruk, Ph.D., from Chiang Mai University, Thailand The fragment size profiling of cell-free DNA (cfDNA) varies between cancer patients and healthy individuals, with shorter fragments typically present in cancer patients. Nowadays, fragmentomics analysis has become a highly prevalent method for early detection of cancer. Here, we introduce a conceptually simple approach for the HCC prediction model called “CEliver”, (CfDNA-based automated capillary Electrophoresis method for Liver cancer screening), an innovative prediction model for early hepatocellular carcinoma (HCC) in high-risk populations based on cfDNA fragmentation analysis using automated capillary electrophoresis platform-a simple technique. Through our innovation, more than 300 cfDNA features were generated from automated capillary electrophoresis data using a feature engineering process. The predictive CEliver model was developed by integrating cfDNA features with AFP based on machine learning. The CEliver model demonstrates high performance in detecting HCC in high-risk patients, achieving a sensitivity of 95%. It also performs impressively in identifying early-stage HCC (stage 0/A) with a sensitivity of 88% and a specificity of 100%. This emphasizes its predictive value. This approach offers a promising, cost-effective strategy for early HCC detection and large-scale HCC surveillance. Keywords: cell-free DNA, fragmentomics, early HCC detection Recorded on May 22nd, 2026 Organizer: Bioinformatics Academic Association in Thailand and Hub of Talents for Bioinformatics in Thailand