Computer-aided detection and abnormality score for the outer retinal layer in optical coherence tomography

Area of Science:

• Ophthalmology
• Medical Imaging
• Artificial Intelligence

Background:

• Outer retinal layers (ORL) are crucial for retinal health.
• Abnormalities in ORL are associated with conditions like choroidal neovascularisation (CNV) and retinitis pigmentosa (RP).
• Accurate detection of ORL abnormalities is vital for timely diagnosis and treatment.

Purpose of the Study:

• To develop a computer-aided detection (CADe) system for identifying outer retinal layer (ORL) abnormalities using optical coherence tomography (OCT).
• To segment and classify normal versus abnormal ORL structures in OCT images.
• To evaluate the diagnostic performance of the developed CADe system.

Main Methods:

• A retrospective study included healthy individuals and patients with CNV or RP.
• A deep learning (DL) algorithm was developed for automatic segmentation of three ORL.
• A binary classifier was trained on segmented images to differentiate normal from abnormal ORL.

Main Results:

• The DL-based CADe achieved high diagnostic accuracy with an Area Under the Curve (AUC) of 0.984 on an internal test set.
• External validation demonstrated strong performance with AUCs of 0.957 and 0.978 on two independent datasets.
• The system effectively highlighted normal ORL and omitted abnormal areas in CNV and RP cases.

Conclusions:

• The developed CADe system accurately segments ORL and differentiates between normal and abnormal retinal structures from OCT images.
• The CADe system's classification performance shows potential for aiding physicians in diagnosis.
• This technology may offer future clinical applications for retinal disease management.