3-D localization of Diabetic Macular Edema using OCT thickness maps

Area of Science:

• Ophthalmology
• Medical Imaging
• Biomedical Engineering

Background:

• Diabetic Macular Edema (DME) is a leading cause of vision loss.
• Accurate segmentation of sub-retinal layers is crucial for DME assessment.
• Current automated methods lack detailed analysis of inner sub-retinal layers in DME.

Purpose of the Study:

• To develop and validate a novel automated system for segmenting 3-D sub-retinal layer thickness maps.
• To identify and quantify structural irregularities in OCT images of patients with DME.
• To establish a new metric for tracking DME severity based on sub-retinal layer changes.

Main Methods:

• Automated segmentation of 3-D thickness maps from Optical Coherence Tomography (OCT) image stacks.
• Comparison of automated segmentation with manual segmentation to assess correlation (r > 0.7).
• Analysis of thickness maps to identify irregular regions in inner and outer nuclear layers.

Main Results:

• Automated thickness maps show high correlation with manual segmentations.
• Significant irregularities detected in inner and outer nuclear layers of DME patients.
• A novel metric derived from the combined area of irregularity strongly correlates with DME severity (r = 0.99).

Conclusions:

• The proposed automated system accurately segments sub-retinal layers in OCT images.
• It effectively identifies and quantifies structural changes associated with DME.
• The novel metric offers a robust and precise method for monitoring DME progression.