Mayo Clinic has developed an AI tool called StateViewer that can identify brain activity patterns associated with nine different types of dementia, including Alzheimer's disease. This is achieved through analysis of a single, widely available fluorodeoxyglucose positron emission tomography (FDG-PET) scan. The AI was trained and tested using over 3,600 scans, including those from patients with dementia and individuals without cognitive impairment.
StateViewer demonstrated an 88% accuracy rate in identifying the specific type of dementia. Furthermore, it enabled clinicians to interpret brain scans with nearly twice the speed and up to three times greater accuracy compared to standard methods. The tool works by comparing a patient's brain scan to a database of scans from individuals with confirmed dementia diagnoses, pinpointing patterns that align with specific dementia types or combinations. The AI displays these patterns via colour-coded brain maps, offering clinicians a visual representation of its diagnostic process. This innovation aims to improve early and accurate dementia diagnosis, even in clinics lacking specialised neurology expertise.
The AI has the potential to help match patients with appropriate care and emerging treatments, and could be especially valuable in areas with limited resources. Researchers plan to broaden the tool's application and continue assessing its performance across diverse clinical settings.
