Accurate Cross-Section Estimation of Blood Vessels in Choroidal Haller's Layer: An Iterative Method based on 3D...

Area of Science:

• Ophthalmology
• Medical Imaging
• Biomedical Engineering

Background:

• Eye diseases such as polypoidal choroidal vasculopathy (PCV) and age-related macular degeneration (AMD) impact choroidal vasculature.
• Detecting subtle changes in the complex choroidal vasculature using 2D OCT B-scan images is challenging.

Purpose of the Study:

• To develop a novel algorithmic method for estimating vessel cross-sections in the choroidal Haller's layer.
• To improve the visualization, analysis, and quantification of choroidal vascular changes for early disease diagnosis.

Main Methods:

• Proposed a novel algorithmic approach to estimate cross-sections of choroidal vessels.
• Evaluated the method's accuracy using synthetic and clinical data.
• Compared the proposed method against a well-established tree-based method.

Main Results:

• The novel method achieved a 90% confidence score from trained optometrists.
• Demonstrated approximately a 60% improvement in accuracy compared to the tree-based method.
• Facilitates tracing and quantifying minute variations in the choroidal vessel network.

Conclusions:

• The proposed method offers a significant advancement in analyzing choroidal vasculature.
• This technique has the potential to aid in the early diagnosis of PCV, AMD, and other related eye conditions.
• Algorithmic cross-section evaluation provides a more effective approach than traditional 2D image analysis for choroidal vasculature.