AI screening for Alzheimer’s disease now 75% accurate
Zhao Yitian, researcher of the intelligent medical imaging (iMED) team at the Ningbo Institute of Materials Technology and Engineering, Chinese Academy of Sciences, told an interview, “Through the collection of ophthalmic images and smart analysis of the images, the artificial intelligence (AI) developed by the team is based on The screening model has achieved an accuracy of 75% when screening people from multiple communities for Alzheimer’s disease.In-depth analysis of the relationship between structural changes in the eye and neurodegenerative disease has demonstrated neurodegeneration. It can potentially shape early detection plans for disease.” Science and Technology Daily reported.
In order to explore the relationship between structural changes in the retina and Alzheimer’s disease, the iMED team collaborated with multiple medical institutions, including Sichuan University Huaxi Hospital, Zhejiang People’s Hospital, Peking University Third Hospital, and Ningbo University People’s Hospital. In addition to collecting eye/brain data, we mainly analyzed fundus images of optical coherence tomography scan perfusion display (OCTA).
According to the IMED team, optical coherence tomography (OCT) scanning is an advanced non-invasive imaging technique that can display structures at different depths of the fundus, including the retina and choroid, and can accurately scan changes in blood flow in fundus structures, allowing OCTA Can generate images. This has important implications for research on changes in fundus blood vessels due to Alzheimer’s disease.
Using a proprietary smart analysis algorithm, the team automatically quantified the fundus structure of Alzheimer’s patients and performed cross-sectional statistics and analysis of derived biomarkers and clinical data. Analyzes showed significant associations between several quantitative measures, including vascular density, vascular fractal dimension, and vascular curvature, and the development of Alzheimer’s disease. This result is consistent with clinical a priori consensus.
Based on this, the team tested Alzheimer’s disease by designing an advanced AI model for blood flow imaging image information. By simply inputting ophthalmological images into the AI model, it is possible to quickly determine whether a subject has Alzheimer’s disease.
The team also built ophthalmic image analysis and smart diagnostic models for brain diseases such as stroke and Parkinson’s disease. As a result, some ocular biomarkers and incidence statistics are both correlated, providing a new approach for rapid and convenient screening of multiple brain diseases. (Editing YF)