Machine Learning for Noninvasive Diagnosis of Neurodegenerative Diseases Using Retinal and Optic Nerve Imaging: A Comprehensive Review.
Authors
Affiliations (7)
Affiliations (7)
- Medical Image and Signal Processing Research Center, Isfahan University of Medical Sciences, Isfahan, Iran.
- Department of Neurosurgery, School of Medicine, Isfahan University of Medical Sciences, Isfahan, Iran.
- Department of Computer Science, Durham University, Durham, United Kingdom.
- Newcastle Eye Centre, Royal Victoria Infirmary, Newcastle Upon Tyne, United Kingdom.
- Farabi Eye Hospital, Tehran University of Medical Science, Tehran, Iran.
- Experimental and Clinical Research Center, Max Delbrueck Center for Molecular Medicine and Charité - Universitaetsmedizin Berlin, Berlin, Germany.
- Department of Engineering, Durham University, Durham, United Kingdom; email: [email protected].
Abstract
Neurodegenerative disorders, including Alzheimer's disease, Parkinson's disease, and multiple sclerosis, encompass a wide range of chronic conditions with irreversible damage to the central nervous system. Current diagnostic workups of these disorders rely on invasive, time-consuming, and costly tests, such as magnetic resonance imaging and cerebrospinal fluid analysis, preventing accurate decision-making and timely therapeutic interventions. The retina is an extension of the central nervous system; thus, retinal imaging, which is a noninvasive and easily accessible tool, provides a unique window to study brain pathologies. There is a great body of evidence suggesting that neurodegenerative disorders are associated with various structural and vascular problems within the retina. Notably, training machine learning models with retinal images has yielded high levels of accuracy in classifying neurodegenerative diseases, encouraging a new era for early and automated diagnosis of these disorders. This article reviews studies that use such models for classifying these disorders.