Multispectral Imaging-Based System for Detecting Tissue Oxygen Saturation With Wound Segmentation for Monitoring...

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Wound Healing Research

Background:

• Blood circulation is a critical factor in wound healing.
• Accurate monitoring of tissue oxygen saturation (StO2) is essential for assessing wound healing progress.
• Existing methods may lack the precision or efficiency needed for real-time wound assessment.

Purpose of the Study:

• To propose and evaluate a novel tissue oxygen saturation detecting (TOSD) system based on multispectral imaging (MSI).
• To quantify StO2 levels in cutaneous tissues for wound healing assessment.
• To develop an automated wound segmentation algorithm for improved accuracy and efficiency.

Main Methods:

• A TOSD system utilizing red and near-infrared light illumination was developed.
• A ResNet34-based U-Net model was employed for automated wound and skin segmentation.
• Animal experiments on mice were conducted to validate the system's performance over seven observation periods.
• Multispectral imaging captured tissue data, and StO2 levels were calculated.

Main Results:

• The automated wound segmentation algorithm achieved a high Dice score of 93.49%.
• The TOSD system detected significant variations in StO2 levels corresponding to different wound healing phases.
• Changes in StO2 were identified earlier than changes in blood flux measured by laser speckle contrast imaging (LSCI).
• Principal component analysis (PCA) effectively visualized wound healing phases using features from TOSD and LSCI.

Conclusions:

• The TOSD system, combined with automated segmentation, accurately distinguishes wound healing phases by monitoring StO2 levels.
• This technology offers a valuable tool for medical professionals to perform precise wound assessments.
• The findings support advancements in StO2 monitoring, wound segmentation, and healing phase classification for preclinical research.