AI Predicts Alzheimer’s Risk Years in Advance Using Retinal Images in Hong Kong Breakthrough
Hong Kong's Humansa Medical Center uses AI to predict Alzheimer's risk by analyzing retinal images, detecting changes up to 15 years before symptoms. This non-invasive, accurate method offers early screening with 80-92% accuracy.
AI Predicts Alzheimer’s Risk Through Retinal Imaging in Hong Kong
A healthcare centre in Hong Kong has introduced an AI-based system that forecasts Alzheimer’s disease risk by analysing retinal images. This approach is non-invasive and detects early alterations in the retina’s blood vessels and nerves, potentially identifying individuals at risk years before symptoms arise.
The Humansa Medical Center, focused on longevity, collaborated with i-Cognitio Sciences, an ophthalmic tech company from the Chinese University of Hong Kong (CUHK), to develop this technology. The AI was trained and validated on nearly 13,000 retinal images from 648 Alzheimer’s patients and over 3,000 cognitively normal individuals.
Accuracy and Validation
A 2022 study published in Lancet Digital Health reported the AI’s accuracy between 80% and 92% across multi-ethnic populations in various countries. This level of performance makes it a promising tool for early detection in diverse communities.
The Need for Early Detection
- Dementia affects about one-third of people over 85 in the Asia-Pacific region.
- In Hong Kong, 10% of those over 70 have dementia, with Alzheimer’s making up more than half of cases.
- Scientific research indicates up to 45% of dementia cases could be delayed or prevented with early intervention.
Current diagnostic methods, such as cognitive tests or brain imaging, have limitations in accuracy or accessibility. More advanced tests like amyloid PET scans and cerebrospinal fluid analysis are invasive and less available to the general public.
A Practical Solution for Community Settings
The AI model offers a simple, low-cost, and low-effort option for screening in community healthcare environments. By analysing retinal scans, it provides sufficient accuracy and sensitivity to flag individuals who might benefit from further evaluation.
The Retina as a Window to the Brain
Vincent Mok, director and founder of i-Cognitio, highlights the retina’s importance in brain diagnostics. He explains that changes in retinal blood vessels and nerves linked to Alzheimer’s can appear 10 to 15 years before clinical symptoms emerge. Detecting these changes opens a critical window for early treatment strategies.
This innovation underscores how AI can support healthcare professionals by enhancing early detection methods for neurodegenerative diseases without invasive procedures.
