Commentary
Video
Using Artificial Intelligence and Retinal Imaging to Detect Neurodegenerative Diseases Early: Sharon Fekrat, MD
Author(s):
The professor of ophthalmology and neurology at Duke University School of Medicine talked about a recent published study on using artificial intelligence and retinal scans to detect mild cognitive impairment in neurodegenerative diseases including Alzheimer disease. [WATCH TIME: 7 minutes]
WATCH TIME: 7 minutes
"One of the challenges that we have is the ability to translate or roll out this machine learning model into the real-world scenario. We need a very large, diverse, studied population in order to really get it out there. Right now, in the iMIND group, we're focusing on patients that don't have diabetes or don't have glaucoma because those 2 common conditions can cause some similar changes in the imaging. Our population that we're currently studying is very simplified and clean, so we don't have anything confusing us. We hope in the future we can then incorporate all these other conditions.”
Mild cognitive impairment (MCI), an intermediate state where cognitive impairment is present but the ability to perform activities of daily living is maintained, is often considered the clinical precursor to Alzheimer disease (AD).1 Therefore, early identification of MCI is critical for effective intervention especially as more novel treatments emerge. In recent research, machine learning models have been developed to provide researchers in the field of neurology with adjunctive measures to establish clinical diagnoses for both AD and MCI.
Recently published in the journal of Ophthalmology Science, research on a developed machine learning model by investigators at Duke Health showed an ability to differentiate normal cognition from MCI using retinal images.2 The model analyzed 236 eyes of 129 controls and 154 eyes of 80 patients with MCI from retinal pictures and images as well as quantitative data to differentiate the 2 types of patients. The model’s ability to perform the MCI diagnosis reported sensitivity of 79% and specificity of 83% while achieving an area under the curve (AUC) of 0.809 when applied to an independent test set (95% CI, 0.681-0.937).
Senior author Sharon Fekrat, MD, professor of ophthalmology and neurology, and associate professor of surgery at Duke University School of Medicine, recently sat down in an interview with NeurologyLive® to discuss how her research group, the iMIND study, plans to utilize longitudinal studies to determine the causal relationship between retinal changes and neurodegenerative diseases. She also talked about the challenges in implementing these machine learning models for real-world use, and how they can be addressed. Additionally, Fekrat spoke about how this technology eventually could replace traditional cognitive assessments, and the implications it might have for early disease detection and prevention.
