2024-07-14 13:12:35
Researchers at Boston University have developed a new artificial intelligence tool that can diagnose 10 different types of dementia, including vascular dementia, Lewy body dementia, and frontotemporal dementia.
The researchers explained, in the results published on Monday in the journal Nature Medicine, that the tool can diagnose all types of dementia, even if these types are concurrent.
Dementia is a general term that refers to a severe decline in mental ability, affecting daily life and leading to a range of symptoms, most notably deterioration of memory and thinking.
Every year, 10 million new cases of dementia are diagnosed worldwide, but the diversity of dementia types and the overlapping symptoms complicate the process of diagnosis and effective treatment. Alzheimer’s disease is the most common cause of dementia in the elderly.
Doctors face major challenges in diagnosing dementia, due to its different types and the similarity of symptoms between these types.
The researchers used a multimodal machine learning system to accurately identify the diseases causing dementia, using common clinical data, such as demographic information, patient and family medical history, medication use, neurological and neuropsychological examination results, and neuroimaging data, such as magnetic resonance imaging.
During the study, the new tool was trained on data from more than 50,000 people, from 9 different global datasets, to enhance its ability to distinguish between different types of dementia.
The accuracy scores for distinguishing between types of dementia range from 0 to 1, with a score of 0.5 indicating a random guess and a score of 1 indicating perfect performance.
The new machine learning model achieved a score of 0.96 in distinguishing between different types of dementia, a near-perfect performance.
The team compared the performance of neurologists and neuroradiologists alone, and with the AI tool, and found that AI could boost doctors’ accuracy in diagnosing dementia by more than 26 percent across all 10 types of dementia.
In an experiment involving 100 cases, 12 neurologists were asked to diagnose and rate a confidence score between 0 and 100, and then this score was averaged with the probability score, which was obtained by the AI tool, to obtain the AI-enhanced score.
“The AI tool enables dementia diagnosis using routinely collected clinical data, demonstrating its potential as a scalable diagnostic tool for Alzheimer’s disease and related dementias,” said lead researcher of the study at Boston University, Dr. Vijaya Kulachalama.
“It has become increasingly important to develop a diagnostic tool using routine clinical data, given the significant challenges in accessing accurate diagnostic tests, not only in remote and economically developing areas, but also in urban healthcare centres,” he added via the university’s website.
Last updated: July 10, 2024 – 16:37
Suggest a correction
2024-07-14 13:12:35