Scientists at Cambridge University in the UK have created an AI-powered tool capable of identifying early-stage dementia patients who are likely to either remain stable or progress to Alzheimer's disease.
Dementia poses a significant global healthcare challenge, impacting more than 55 million people and costing an estimated $820 billion annually. Cases are projected to nearly triple over the next five decades.
To develop the AI model, researchers utilised readily available, non-invasive patient data — including cognitive tests and MRI scans revealing grey matter loss — from a cohort of over 400 participants in the US. They validated the model using data from an additional 600 US participants and longitudinal data from 900 individuals in memory clinics in the UK and Singapore.
Published in a medical journal, the study found that the algorithm could distinguish between individuals with stable mild cognitive impairment and those who progressed to Alzheimer's within three years. It achieved an 82% accuracy in identifying future Alzheimer's cases and an 81% accuracy in identifying non-progressors using only cognitive tests and MRI scans.
Professor Zoe Kourtzi from Cambridge's Department of Psychology noted that the tool shows promise in predicting Alzheimer's progression, having been tested with real-world data, which enhances its reliability.
Ben Underwood, also from Cambridge, emphasised that this advancement could address significant concerns for patients and their families.