SVM-Based Optical Detection of Retinal Ganglion Cell Apoptosis

Area of Science:

• Ophthalmology
• Neuroscience
• Biomedical Imaging

Background:

• Retinal ganglion cell (RGC) loss is a key factor in eye diseases such as glaucoma.
• Axon damage and dendritic degeneration precede RGC death, correlating with neuronal degeneration.
• These early degenerative changes are detectable within the resolution limits of optical coherence tomography (OCT).

Purpose of the Study:

• To evaluate the optical changes associated with early retinal degeneration.
• To assess the potential of OCT in detecting early signs of neuronal damage in the retina.

Main Methods:

• Optical coherence tomography (OCT) imaging was performed on axotomised retinal explants from C57BL/6J mice up to 120 minutes post-enucleation.
• Grey-level co-occurrence-based texture analysis of the inner plexiform layer (IPL) was used to monitor optical speckle changes.
• Principal component analysis (PCA) and support vector machine (SVM) were employed for data analysis, with modulation transfer functions (MTFs) used to confirm retinal transparency.

Main Results:

• Quantitative analysis of MTFs confirmed stable optical transparency of retinal explants during the imaging period (MTF (fx) = 0.267 ± 0.02).
• Textural features extracted from the IPL successfully discriminated optical signals indicative of RGC degeneration.
• Support vector machine (SVM) classification achieved a mean accuracy of 86.3% in detecting degeneration, with combined PCA and SVM yielding 81.9% accuracy.

Conclusions:

• High-resolution OCT can detect optical changes in the inner plexiform layer (IPL) subsequent to RGC axotomy.
• OCT imaging serves as a valuable tool for indexing retinal neuronal integrity and degeneration.
• This approach holds promise for early diagnosis and monitoring of retinal neurodegenerative diseases.