Assessing Fuchs Corneal Endothelial Dystrophy Using Artificial Intelligence-Derived Morphometric Parameters From...

Area of Science:

• Ophthalmology
• Medical Imaging
• Artificial Intelligence

Background:

• Fuchs corneal endothelial dystrophy (FECD) is a progressive condition affecting corneal clarity.
• Accurate characterization of FECD is crucial for patient management and predicting disease progression.
• Specular microscopy is a key diagnostic tool for evaluating the corneal endothelium.

Purpose of the Study:

• To assess the effectiveness of artificial intelligence (AI) in analyzing morphometric parameters from specular microscopy images for FECD characterization.
• To compare AI-derived measurements with traditional methods and clinical grading scales.

Main Methods:

• A convolutional neural network (CNN) was employed to segment specular microscopy images and extract morphometric parameters.
• Parameters analyzed included endothelial cell density, guttae area ratio, coefficient of variation, and hexagonality.
• AI-derived data were correlated with clinical FECD classifications using the modified Krachmer grading scale.

Main Results:

• Significant differences were observed between AI-based and microscope software cell density measurements.
• The guttae area ratio demonstrated a strong correlation with FECD clinical grades, particularly in the central cornea.
• Peripheral guttae area ratio showed a weaker, yet significant, correlation in the inferior region.

Conclusions:

• Convolutional neural networks (CNNs) are valuable tools for precise FECD evaluation via specular microscopy.
• Guttae area ratio is a reliable parameter for assessing FECD severity, aligning well with clinical grading.
• These findings support AI's role in predicting corneal edema risk and monitoring FECD therapies.