Duke Health researchers have developed a machine learning model that can identify mild cognitive impairment (MCI) by analyzing retinal images from the eye. The model, described in the journal Ophthalmology Science, demonstrates the potential for a non-invasive and cost-effective method of detecting early signs of cognitive impairment that may progress to Alzheimer’s disease.
In previous models, differentiating mild cognitive impairment from normal cognition has been challenging. However, this new machine learning model has shown promising results in distinguishing between the two. The senior author of the study, Sharon Fekrat, M.D., explains, 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 one step closer to detecting cognitive impairment earlier before it progresses to Alzheimer’s dementia.
The researchers built upon their previous model, which successfully identified patients with a known Alzheimer’s diagnosis using retinal scans and other data. The current study focused on detecting mild cognitive impairment, which often precedes Alzheimer’s disease. By utilizing machine learning techniques, the model identified specific features in the retinal images, such as changes in the neurosensory retina and its microvasculature, that indicate the presence of cognitive impairment. Additionally, patient data, including age, sex, visual acuity, and educational background, were incorporated into the model.
The results of the study showed that the model achieved a sensitivity of 79% and a specificity of 83% in differentiating individuals with normal cognition from those with a diagnosis of mild cognitive impairment.
The study’s co-first author, C. Ellis Wisely, M.D., noted the significance of having a non-invasive and cost-effective means of identifying patients with mild cognitive impairment, particularly as potential new therapies for Alzheimer’s disease emerge. Wisely stated, Having a non-invasive and less expensive means to reliably identify these patients is increasingly important, particularly as new therapies for Alzheimer’s disease may become available.
The retina has been likened to a window to the brain, and leveraging non-invasive and cost-effective retinal imaging with machine learning algorithms can be a powerful tool for screening patients at scale. Alexander Richardson, a student in the Eye Multimodal Imaging in Neurodegenerative Disease lab at Duke and co-lead author of the study, emphasized the potential of this approach, stating, 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.
Overall, the development of this machine learning model represents a significant advancement in the early detection of cognitive impairment. As the model continues to be refined and tested, it holds promise for identifying individuals at risk of developing Alzheimer’s disease, leading to earlier interventions and potential improvements in patient outcomes.
