Vessel skeleton-based vasodynamic optical flow for video deshaking in functional OCT angiography.

Area of Science:

• Ophthalmology
• Biomedical Engineering
• Medical Imaging

Background:

• Functional OCT angiography (fOCTA) is crucial for non-invasive detection of retinal neurovascular coupling (rNVC).
• Involuntary eye movements during fOCTA imaging compromise accurate, single-capillary level quantification of rNVC.
• Existing methods struggle to maintain vasodynamic information during image stabilization.

Purpose of the Study:

• To develop and validate a novel algorithm for stabilizing fOCTA videos.
• To preserve critical vasodynamic responses for accurate rNVC assessment.
• To enable precise detection of rNVC dysfunction at the single-capillary level.

Main Methods:

• Development of a vessel skeleton-based vasodynamic optical flow (VDO flow) algorithm.
• Application of VDO flow for fOCTA video stabilization and vasodynamic preservation.
• Evaluation of image registration speed and accuracy (vascular skeleton recombination rate).

Main Results:

• The VDO flow algorithm achieves high-speed image registration (∼0.39s per frame).
• High accuracy in image stabilization was demonstrated (63.4% vascular skeleton recombination rate).
• The method effectively preserves essential vasodynamic responses for rNVC analysis.
• Successfully detected single-capillary level rNVC dysfunction in early diabetic retinopathy.

Conclusions:

• The VDO flow algorithm significantly improves fOCTA image stabilization and vasodynamic preservation.
• This technique enables precise detection of rNVC dysfunction, even in early-stage diseases.
• The advancement facilitates broader clinical applications of fOCTA for diagnosing retinal conditions.