Loosely coupled level sets for simultaneous 3D retinal layer segmentation in optical coherence tomography.

Area of Science:

• Ophthalmology
• Biomedical Imaging
• Medical Image Analysis

Background:

• Accurate segmentation of retinal layers in optical coherence tomography (OCT) is crucial for diagnosing eye conditions.
• Existing segmentation methods may lack robustness or require manual intervention.

Purpose of the Study:

• To develop and validate a novel, automated method for segmenting retinal layers using OCT data.
• To assess the accuracy, reproducibility, and robustness of the proposed segmentation technique.

Main Methods:

• The method utilizes attenuation coefficients derived from in-vivo human retinal OCT data.
• A flexible coupling approach is employed for simultaneous segmentation, leveraging the anatomical order of retinal layers.
• Evaluation involved manual segmentation by a medical doctor for comparison.

Main Results:

• The automated method demonstrated very good agreement with manual segmentation across various data types.
• Mean absolute deviation for all interfaces ranged from 1.9 to 8.5 µm (0.5-2.2 pixels).
• Reproducibility of the automated method was comparable to manual segmentation.

Conclusions:

• The novel segmentation method provides accurate and reproducible retinal layer segmentation from OCT images.
• This technique shows robustness across different patient groups (healthy, glaucoma), scan locations (peripapillary, macular), and OCT devices.
• The method has potential for improving clinical assessment and research in ophthalmology.