Automatic estimation of corneal nerves focused tortuosities

Area of Science:

• Ophthalmology
• Biomedical Engineering
• Medical Imaging

Background:

• Corneal nerve tortuosity correlates with various pathologies.
• Current manual assessment methods lack reproducibility due to undefined tortuosity metrics.
• Two distinct forms of corneal nerve tortuosity (short-range and long-range) have been recently identified.

Purpose of the Study:

• To develop and validate automatic measurements for corneal nerve tortuosity.
• To differentiate and quantify short-range and long-range corneal nerve directional changes.
• To assess the benefit of combining multiple mathematical tortuosity metrics.

Main Methods:

• Development of novel algorithms for automatic corneal nerve segmentation and analysis.
• Implementation of mathematical models to quantify both short-range and long-range tortuosity.
• Comparative analysis of automatic measurements against manual assessments and established definitions.

Main Results:

• Successfully developed automatic measurements that accurately represent two distinct definitions of corneal nerve tortuosity.
• Demonstrated that the developed metrics independently capture short-range and long-range directional changes.
• Showed that combining multiple mathematical tortuosity measurements yields superior results compared to single metrics.

Conclusions:

• Automatic quantification of corneal nerve tortuosity is feasible and improves reproducibility.
• The developed methods can reliably distinguish between different types of nerve tortuosity.
• A combined approach using multiple mathematical metrics offers a more comprehensive assessment of corneal nerve tortuosity.