Summary: Scientists at the University of Cambridge have developed an AI tool using cognitive tests and MRI scans that accurately predicts Alzheimer’s progression, reducing the need for invasive tests and improving early treatment outcomes.
Key Takeaways
- AI Tool for Early Alzheimer’s Detection: Scientists at the University of Cambridge developed an AI tool that uses cognitive tests and MRI scans to predict Alzheimer’s progression with 82% accuracy, reducing the need for invasive and costly diagnostic tests.
- Significant Impact on Dementia Management: With dementia affecting over 55 million people globally and expected to triple in the next 50 years, this tool can improve early treatment outcomes and reduce misdiagnosis by being three times more accurate than current clinical methods.
- Stratification and Validation: The AI model categorizes Alzheimer’s patients into three groups—stable, slow progressors, and rapid progressors—and its predictions were validated with follow-up data over six years, aiding in personalized patient care and monitoring.
———————————————————————————————————————————————————
Scientists at the University of Cambridge in England have developed an AI tool that can predict in four out of five cases whether people with early signs of dementia will remain stable or develop Alzheimer’s disease. This new approach could reduce the need for invasive and costly diagnostic tests while improving early treatment outcomes, such as lifestyle changes or new medicines.
Dementia affects over 55 million people worldwide, with an annual cost of $820 billion. The number of cases is expected to nearly triple in the next 50 years. Alzheimer’s disease causes 60%-80% of dementia cases. Early detection is crucial for effective treatment, but current methods often require expensive or invasive tests, leading to misdiagnosis or late diagnosis.
Reducing Alzheimer’s Misdiagnosis
The team from Cambridge’s Department of Psychology developed a machine learning model using non-invasive, low-cost patient data—cognitive tests and MRI scans—from over 400 individuals in a U.S. research cohort. They tested the model on data from 600 US participants and 900 from U.K. and Singapore memory clinics.
The algorithm accurately predicted Alzheimer’s progression within three years, identifying 82% of those who would develop Alzheimer’s and 81% of those who would not, solely based on cognitive tests and MRI scans. This was three times more accurate than current clinical methods, significantly reducing misdiagnosis.
Enhancing Early Treatment for Alzheimer’s Patients
The model also stratified Alzheimer’s patients into three groups: those with stable symptoms (50%), slow progressors (35%), and rapid progressors (15%). These predictions were validated with follow-up data over six years, helping to identify patients who could benefit from early treatments or need close monitoring.
“We’ve created a tool that, using only cognitive tests and MRI scans, is much more sensitive at predicting Alzheimer’s progression,” says Cambridge professor Zoe Kourtzi, PhD. “This can significantly improve patient wellbeing by identifying those needing the closest care and reducing unnecessary tests.”