Researchers from McGill University have developed and for the first time used an algorithm capable of effectively detecting signs of dementia 2 years before the onset of the first clinical manifestations of the disease using an artificial intelligence system. The results of their research are published in the journal “Neurobiology of Aging”.
Study leader Pedro Rosa-Neto said the new diagnostic method combines the analysis of positron emission tomography (PET) images and the assessment of biomarkers characteristic of Alzheimer’s disease.
Scientists have long known that a protein called amyloid accumulates in the brain of patients with Mild Cognitive Impairment (MCI), often resulting in dementia. However, this biomarker should be evaluated with caution, since amyloid accumulation can begin decades before the onset of the first symptoms of dementia and, in addition, not all patients with MCI develop Alzheimer’s disease.
The developed algorithm, after analyzing several hundred PET images of healthy people and patients with Alzheimer’s disease, allowed the computer program to correctly determine the first signs of dementia in 84% of cases in patients who do not yet have clinical symptoms.
Researchers continue to search for new methods for diagnosing Alzheimer’s disease for higher algorithm accuracy. Currently, a team of scientists is conducting additional testing of the algorithm on different groups of patients with common neurocirculatory diseases, such as cerebral stroke. This will allow specialists to quickly certify the algorithm and widely use the development in clinical practice.