New Method of Early RRMS Diagnosis Using OCT-Assessed Structural Retinal Data and Explainable Artificial Intelligence

Area of Science:

• Ophthalmology
• Neurology
• Medical Imaging

Background:

• Multiple sclerosis (MS) is a neurodegenerative disease affecting the central nervous system.
• Early diagnosis of MS is crucial for effective management and treatment.
• Optical coherence tomography (OCT) allows for non-invasive imaging of retinal structures.

Purpose of the Study:

• To develop a method for classifying OCT-assessed retinal data for automatic diagnosis of early-stage MS.
• To provide decision explanations for the diagnostic classification.

Main Methods:

• Analysis of macular retinal nerve fiber layer (mRNFL), ganglion cell layer (mGCL), inner plexiform layer (mIPL), and inner retinal complex (mIRL) thicknesses in 79 early-stage MS patients and 69 controls.
• Utilized recursive feature extraction (RFE) and Shapley additive explanations (SHAP) with a support vector machine (SVM) classifier.
• Identified 20 features from an initial 48 that maximized classifier accuracy.

Main Results:

• Achieved a classifier accuracy of 0.9257 for diagnosing early-stage MS.
• SHAP values highlighted average thickness as more relevant than inter-eye difference.
• Identified mGCL and mRNFL as the most influential structures, with peripheral papillomacular bundle and supero-temporal quadrant being the most affected zones.

Conclusions:

• The developed method significantly improves the success rate of automatic diagnosis for early-stage relapsing-remitting MS (RRMS).
• Enhances the transparency of clinical decision-making in MS diagnosis.
• OCT-based retinal structure assessment offers a non-invasive diagnostic approach for early MS, implementable in healthcare settings.