A new ML model can identify mild cognitive impairment from retinal scans

A non-invasive and inexpensive method of identifying the early signs of cognitive impairment.

Follow us onFollow Tech Explorist on Google News

Previously, scientists at the Duke University Medical Center have developed a model that used retinal scans and other data to identify patients with a known Alzheimer’s diagnosis successfully. The Optical coherence tomography (OCT) and OCT angiography (OCTA) scans identified structural abnormalities in the microvasculature and neurosensory retina of Alzheimer’s patients.

New work expands on that work, in which scientists have developed a new machine learning model that can differentiate normal cognition from mild cognitive impairment using retinal images from the eye. The model examines retinal images and related data to detect people with mild cognitive impairment, identifying key traits.

The model potentially be used as a non-invasive and inexpensive method of identifying the early signs of cognitive impairment that could progress to Alzheimer’s disease.

Senior author Sharon Fekrat, M.D., professor in Duke‘s departments of Ophthalmology and Neurology and associate professor in the Department of Surgery, said, “This is particularly exciting work because we have previously been unable to differentiate mild cognitive impairment from normal cognition in previous models. This work brings us closer to detecting cognitive impairment before it progresses to Alzheimer‘s dementia.”

Along with patient information like age, sex, visual acuity, years of education, and specific markers in the OCT and OCTA images that indicate the existence of a cognitive impairment, the new model also extracts quantitative information from the images themselves.

With a sensitivity of 79% and specificity of 83%, the model assessed retinal images and pictures with quantitative data to distinguish between individuals with normal cognition and those diagnosed with mild cognitive impairment.

Co-first author C. Ellis Wisely, M.D., assistant professor in the Department of Ophthalmology, said, “This is the first study to use retinal OCT and OCTA images to distinguish people with mild cognitive impairment from individuals with normal cognition. Having a non-invasive and less expensive means to identify these patients reliably is increasingly important, particularly as new therapies for Alzheimer’s disease may become available.”

Co-lead author Alexander Richardson, a student in the Eye Multimodal Imaging in Neurodegenerative Disease lab at Duke, said, “The retina is a window to the brain, and machine learning algorithms that leverage non-invasive and cost-effective retinal imaging to assess neurological health can be a potent tool to screen patients at scale.”

Journal Reference:

  1. C. Ellis Wisly, Alexander Richardson, Ricardo Henao, et al. A convolutional neural network using multimodal retinal imaging for differentiation of mild cognitive impairment from normal cognition. Ophthalmology Science. DOI: 10.1016/j.xops.2023.100355