A Computational System for Detecting Early Recurrence of Hepatocellular Carcinoma (HCC)

Background
Hepatocellular carcinoma (HCC) is one of the most common and deadly malignant tumors worldwide. Various treatment strategies can be employed depending on the stage of the cancer, including liver resection, transplantation, ablation, embolization, targeted therapy, and immunotherapy. Liver resection and transplantation are common treatments for early-stage HCC involving small tumors. Although advancements in medical technology have reduced mortality rates from liver cancer, the recurrence rate after tumor resection remains high, with up to 70% recurrence within five years. Recurrence within one year is classified as early recurrence. Notably, 60–70% of HCC cases fall into this category and are strongly associated with poor survival outcomes.

Research Collaboration
A total of 64 HCC patients who underwent surgical resection were recruited from National Cheng Kung University Hospital. Among them, 54 patients did not have cirrhosis (27 with early recurrence and 27 with late recurrence), and 10 patients had cirrhosis. Hierarchical clustering analysis using a microarray of three methylation gene groups was conducted, identifying hundreds of differentially methylated genes associated with early HCC recurrence. Functional interaction analysis was then performed to identify key enriched genes, including BDNF, FGFR1, and FGFR2.

Technical Approach
A non-invasive liquid biopsy using blood samples was utilized. After bisulfite conversion, methylation signals were detected using quantitative Methylation-Specific Polymerase Chain Reaction (qMSP). A computational system was developed to perform qualitative and quantitative analysis of multiple differentially methylated genes. This system runs a methylation-based prediction model to detect early recurrence of liver cancer and to assess its prognosis.