Neural Network Enables High Accuracy for Hepatitis B Surface Antigen Detection with a Plasmonic Platform

Area of Science:

• Biomedical Engineering
• Nanotechnology
• Infectious Disease Diagnostics

Background:

• Hepatitis B virus (HBV) infection diagnosis relies on detecting hepatitis B surface antigen (HBsAg).
• Current detection methods exhibit inaccuracies, potentially causing treatment delays or unnecessary interventions.

Purpose of the Study:

• To develop a highly accurate and sensitive label-free biosensing platform for HBsAg detection.
• To integrate deep learning for enhanced data analysis and improved diagnostic performance.

Main Methods:

• A label-free plasmonic biosensing approach utilizing thickness-sensitive plasmonic coupling.
• Supervised deep learning (DL) with neural networks for processing sensor output data.

Main Results:

• The DL-enhanced sensor achieved a significantly improved accuracy, reaching 99%-99.6% from 93.1%-97.4%.
• The platform demonstrated high sensitivity and rapid assay time (approximately 30 minutes).
• DL integration simplified readout procedures and reduced overall processing time.

Conclusions:

• The developed plasmonic biosensing and DL method offers a promising tool for high-precision molecular detection.
• This approach is suitable for point-of-care (POC) applications, improving hepatitis B diagnostics.