Automatic measurement of epidermal thickness from optical coherence tomography images using a new algorithm

Area of Science:

• Biomedical Imaging
• Dermatology
• Medical Image Analysis

Background:

• Optical coherence tomography (OCT) allows in vivo epidermal thickness (ET) measurement.
• Automatic dermo-epidermal junction (DEJ) detection is crucial for large-scale OCT clinical studies.
• Image noise, such as optical speckle, complicates accurate DEJ detection.

Purpose of the Study:

• To present an algorithm for automatic DEJ boundary detection in OCT images.
• To evaluate the performance of the algorithm in measuring ET.
• To compare algorithm performance against expert variability.

Main Methods:

• Acquired 29 OCT skin images from five body sites of healthy subjects.
• Seven experts manually traced the DEJ on each image.
• Developed and applied a novel image processing algorithm for DEJ detection and ET measurement.

Main Results:

• Between-expert variability in DEJ tracing was low (mean SD 3.4 μm).
• The algorithm achieved automated ET measurement precision of < 5 μm across most sites.
• Algorithm performance was verified, showing adequate results on all images.

Conclusions:

• The developed image processing method significantly reduces ET calculation time and eliminates user intervention.
• Automated ET measurement using OCT enhances data analysis speed and reproducibility in clinical trials.
• The algorithm provides a reliable tool for objective ET assessment.