A Fluorescence Imaging- and Deep Learning-Based Approach for Detecting Hepatitis B Virus Integration into Host...

Area of Science:

• Hepatology
• Virology
• Genomics
• Artificial Intelligence

Background:

• Hepatitis B virus (HBV) infection is a major cause of hepatocellular carcinoma.
• HBV integration into the host genome is implicated in liver carcinogenesis.
• Current detection methods for HBV integration have limitations in sensitivity, site identification, or cost-effectiveness.

Purpose of the Study:

• To develop a novel method for detecting HBV DNA integration at the single-cell level.
• To evaluate the efficacy of fluorescence in situ hybridization combined with deep learning for this purpose.
• To establish a proof-of-concept for a new diagnostic tool for HBV integration.

Main Methods:

• Development of a fluorescence in situ hybridization (FISH) approach for visualizing HBV integration in individual liver cells.
• Application of a deep learning model for analyzing FISH images and identifying integrated HBV DNA.
• Validation using hepatoma cell lines with and without known HBV DNA integration.

Main Results:

• The developed FISH approach successfully visualized HBV integration in liver cells.
• The deep learning model achieved over 90% accuracy in identifying positive and negative control cells.
• This study demonstrates the feasibility of using fluorescence imaging and AI for single-cell HBV integration detection.

Conclusions:

• Fluorescence imaging coupled with deep learning provides a sensitive and accurate method for detecting HBV integration at the single-cell level.
• This approach overcomes limitations of existing methods, offering potential for clinical applications.
• This work represents a significant advancement in understanding and diagnosing HBV-related liver diseases.