Although Alzheimer's disease affects tens of millions of people worldwide, it is still difficult to detect early. However, researchers investigating the possibilities of artificial intelligence in medicine have discovered that technology can help early detection of treacherous disease. The California team recently published a report on the study of the Radiology magazine and showed how once the neural network was trained, it was able to correctly diagnose Alzheimer's disease in a limited number of patients based on brain imaging visualizations made years ago. diagnosed by a doctor.
The team uses brain imaging (FDG-PET imaging) to train and test neural networks. In FDG, images of the patient's blood circulation are injected with a radioactive type of glucose, and then the body tissue, including the brain, pushes towards the surface. Scientists and doctors can then use the PET scan to detect the metabolic activity of this tissue, depending on how long the FDG is received.
The FDG-PET method is used to diagnose Alzheimer's; Patients with this disease often exhibit lower levels of metabolic activity in certain parts of the brain. However, experts have to analyze this image to find evidence of the disease, and this is very difficult because it may lead to similar results in screening for moderate cognitive impairment and Alzheimer's disease.
Therefore, the team uses 2,109 FDG-PET images from 1002 patients, trains the neural network by 90%, and the remaining 10% are tested. It also performs tests with a single test of 40 patients screened between 2006 and 2016, then compares the findings of artificial intelligence with a group of experts analyzing the same data.
With a separate set of test data, Artificial Intelligence can diagnose Alzheimer's patients with 100% accuracy and 82% accuracy in patients who do not suffer from a treacherous disease. It can also make average predictions over the next six years. In comparison, a group of doctors looking at the same scanned images identified patients with Alzheimer's disease in 57% of patients and without disease – 91%. it is not.
The researchers state that their research has several limitations, including a small amount of test data and limited educational data.