It was created by scientists from Lithuania Researchers from Kaunas University of Technology told Medical Express about a new useful development. Discuss WHO states that Alzheimer's disease is one of the most common causes of dementia (occurs in 70 percent of cases). Alzheimer's has already been identified in about 24 million people in the world, and experts expect this number to double every 20 years. Aggravating the situation and aging of the population.
One of the likely symptoms of Alzheimer's disease, the author says, is mild cognitive impairment. It represents the stage between cognitive decline in normal aging and dementia. According to Ritis Maskeliunas, a researcher at Kaunas University of Technology, it is possible to predict the occurrence of Alzheimer using MRI. This also applies to the early stages of the disease.
At the same time, such an analysis of the results of MRI, associated with the determination of signs of Alzheimer's disease, takes a lot of time. Therefore, artificial intelligence can speed up this process. Although, the journalist notes, mild cognitive impairment does not necessarily indicate the development of Alzheimer's in the future. However, this will greatly help the doctor assess the patient's condition, all possible diseases and make an accurate diagnosis.
The algorithm helped to analyze the MRI scans of 138 people and divide them into 6 categories depending on the state of the brain: from normal to signs of mild cognitive impairment leading to Alzheimer's. Overall, 51,443 and 27,310 images were used to train and test artificial intelligence. The model identified early signs of moderate cognitive impairment in 99.99 percent of cases, late signs of moderate cognitive impairment in 99.95 percent of cases, and moderate cognitive impairment in 99.95 percent of cases.
The main achievement, according to Maskeliunas, is the accuracy of the analysis performed by artificial intelligence. We are still working with medical institutions to improve the technology.
At the same time, the specialist suggests, it will be possible to apply such an algorithm when analyzing the results of patients at risk (persons over 65 years of age with a history of brain injury, suffering from high blood pressure etc.). In this way, he will inform healthcare providers about the symptoms associated with the early development of Alzheimer's disease.