"The single most important finding of this study was that a brain disease was diagnosed with a basic picture of the eye. This is very different from traditional approaches where to find a problem with the brain you look at different brain images," Maximilian Diaz, a PhD student of biomedical engineering at the University of Florida in Gainesville, said in a news release.
He presented his research November 29 at the Radiological Society of North America (RSNA) virtual annual meeting.
Parkinson's disease is characterized by the loss of dopaminergic neurons in the substantia nigra. It is typically diagnosed based on the presence of tremors, muscle stiffness and impaired balance. But these symptoms develop after prolonged progression with significant injury to dopaminergic neurons.
"This means we are diagnosing patients late," Diaz said in the release.
Thinning of the retina walls and retinal microvasculature also occurs in PD and may represent an early non-motor sign of the disease. Diaz and colleagues trained a type of artificial intelligence called support vector machine learning to detect signs suggestive of PD in fundus eye images.
They used data from two age- and gender-matched datasets from the UK Biobank (UKB) and the University of Florida (UF) to compare fundus eye images from 310 PD patients and 266 healthy controls.
Machine-learning networks were able to accurately distinguish PD images from control images based on the retinal microvasculature, achieving a maximal accuracy between 0.698 and 0.719. Patients with PD had smaller blood vessels in the retina than healthy controls.
The findings "support the idea that changes in brain physiology can be observed in the eye," Diaz told the conference.
"It's just a simple picture of the eye, you can have it done in less than a minute, and the cost of the equipment is much less than a CT or MRI machine," he added in the news release. "If we can make this a yearly screening, then the hope is that we can catch more cases sooner, which can help us better understand the disease and find a cure or a way to slow the progression."
In an email to Reuters Health, Diaz said, "For this diagnostic test to reach the clinic, we would need to increase our subject population - a task we are currently working on but was set back due to COVID - to both improve the overall performance and generalizability of the test and also provide sufficient evidence to the (Food and Drug Administration) that our test is functional."
The study had no commercial funding and the authors have no relevant disclosures.
SOURCE: https://bit.ly/3qisCpD Radiological Society of North America annual meeting, presented November 29, 2020.
By Megan Brooks
© 2023 The Author(s). Published by Medicom Medical Publishers.
User license: Creative Commons Attribution – NonCommercial (CC BY-NC 4.0)
Posted on
site created by:
© 2023 Medicom Medical Publishers. All rights reserved. Terms and Conditions | Privacy Policy