Classification of healthy and diseased retina using SD-OCT imaging and Random Forest algorithm

Area of Science:

• Ophthalmology
• Medical Imaging
• Artificial Intelligence

Background:

• Age-related macular degeneration (AMD) and Diabetic Macular Edema (DME) are leading causes of vision loss.
• Accurate and early diagnosis is crucial for effective treatment and management.
• Spectral Domain Optical Coherence Tomography (SD-OCT) provides detailed cross-sectional images of the retina.

Purpose of the Study:

• To develop and validate a novel automated classification model for identifying AMD and DME.
• To utilize quantitative retinal features extracted from SD-OCT images for disease detection.
• To compare the proposed model's performance against existing state-of-the-art methods.

Main Methods:

• Automated extraction of ten clinically relevant retinal features from SD-OCT images.
• Features include retinal layer thickness and pathology volumes (drusen, hyper-reflective spots).
• Classification using Random Forest algorithm with 15-fold cross-validation on 251 subjects (59 normal, 177 AMD, 15 DME).

Main Results:

• Achieved >95% accuracy for classifying DME, AMD, and normal cases.
• Two-class classification (normal vs. pathology) yielded >96% accuracy.
• Area Under the Curve (AUC) reached 0.99 for all datasets.
• Outperformed four other state-of-the-art classification methods.

Conclusions:

• The proposed classification model demonstrates high accuracy in identifying AMD and DME from SD-OCT images.
• Automated feature extraction and Random Forest classification offer a robust approach for ophthalmic disease diagnosis.
• This method holds potential for improving early detection and patient management of retinal pathologies.