Mass General Brigham leading research into using AI to detect early signs of Alzheimer’s

A normal brain compared to the brain of a person with Alzheimer’s

Lightly edited from a report by The New England Council

BOSTON

“Mass General Brigham is leading research into using artificial intelligence to identify early warning signs of Alzheimer’s disease, a development that could significantly improve outcomes for the estimated 7 million Americans living with the condition. Dr. Lidia Moura, director of population health in MGB’s neurology department, co-authored a study examining how AI can scan electronic medical records from routine patient visits across specialties, from primary care to ophthalmology, for subtle indicators of cognitive decline that clinicians might otherwise miss. 

“The study, recently published in the journal npj Digital Medicine, analyzed 3,300 clinical notes from 200 anonymized patients. The AI tool accurately detected early signs of possible cognitive problems 88 percent of the time. The system works by deploying a team of AI ‘agents’ that check and refine their analysis of clinical notes — flagging signals as routine as a missed appointment, a family member’s comment about forgetfulness, or difficulty managing prescriptions. MGB is currently seeking philanthropic funding to launch a pilot program within three to four months. 

“The research is accompanied by a parallel study from Worcester Polytechnic Institute, in which scientists used machine learning to analyze MRI brain scans for early structural changes associated with Alzheimer’s, achieving nearly 93 percent predictive accuracy. Together, the two efforts reflect a broader push to close a significant diagnostic gap: Currently, 90 percent of people in Alzheimer’s earliest phase, mild cognitive impairment,  go undiagnosed in the United States.’’