Alzheimer's Early Retinal Detection NeuroLens: Predicting Disease Risk Through Eye Blood Vessel Patterns

NeuroLens, jointly released by the UCL Institute of Ophthalmology and startup NeuroVista Diagnostics, is a technology that predicts Alzheimer's disease risk by analyzing retinal blood vessel patterns. This non-invasive detection method can identify high-risk individuals ten years before symptoms appear, buying valuable time for early intervention.

The project grew out of long-running observations of the close connection between the brain and ocular vascular systems. UCL Institute of Ophthalmology's longitudinal studies found that the hallmark pathological change of Alzheimer's disease, beta-amyloid deposition, occurs not only in the brain but also in the retinal blood vessel walls. The retina, as an extension of the brain, provides a direct window into cerebrovascular pathology without the need for expensive PET scans or lumbar punctures.

The core technology is a deep learning vascular analysis model. NeuroLens uses a non-mydriatic fundus camera to capture high-resolution images of the patient's retina. The model automatically identifies the retinal microvascular network (with diameters below 100 micrometers) and quantifies morphological features such as vessel tortuosity, branch angle, and density distribution. The model was trained on fundus images and long-term health tracking data from 120,000 people in the UK Biobank.

In the longitudinal study, the UCL team tracked 2,400 middle-aged and older adults aged 50 to 70 for eight years. Participants were tested with NeuroLens at enrollment and underwent cognitive function assessments every two years thereafter. Results show that NeuroLens predicted Alzheimer's disease onset ten years before symptoms with 87 percent accuracy. By comparison, traditional cognitive scale screening had only 54 percent accuracy five years before symptoms appeared.

For applications, NeuroLens is particularly suited to large-scale population screening. Traditional Alzheimer's diagnosis requires PET scans (about 3,000 pounds) and cerebrospinal fluid examination (an invasive procedure), making it difficult to scale up. NeuroLens costs only 89 pounds per test, takes just two minutes, and can be performed at a community clinic or even an optician. The UK National Health Service has announced that NeuroLens will be included in routine physical exams for people over 50 starting in 2027.

On the commercialization path, NeuroVista Diagnostics will operate in a B2B2C model. The company sells NeuroLens detection devices (24,000 pounds each) and a companion AI analysis subscription service (free for up to 500 tests per month, 49 pounds per test thereafter) to hospitals, physical examination centers, and optical chains. NeuroVista has secured 42 million pounds in Series B funding from Index Ventures, valuing the company at 310 million pounds.

Critics point out that NeuroLens's prediction of "high risk" does not necessarily mean actual onset, and longer-term tracking is needed for validation. Professor Francesca Cordeiro, who heads the UCL Institute of Ophthalmology, responded that the team is collaborating with the Mayo Clinic in the US on a 50,000-person multi-center longitudinal study, expected to complete in 2028. By then, more comprehensive prediction accuracy data will be available. She also emphasized that the goal of NeuroLens is not diagnosis, but to help high-risk individuals take preventive measures early. Healthy diet, exercise, and adequate sleep are known to significantly reduce the risk of Alzheimer's disease.