Researchers at the University of California in San Francisco (UCSF) have trained a computer system to detect early signs of Alzheimer's disease in medical images. Their success is amazing: technology has identified the first signs of Alzheimer's disease much earlier than humans. For many people with dementia, this may soon be a way to get an early diagnosis and thus a better therapy.
One hundred percent accuracy in Alzheimer's disease
The reason for this hope is the pilot study conducted by the team of Jae Ho Sohn, which was published in the journal Radiology. The main conclusion from the study: Under certain conditions, computers can recognize certain patterns better than humans. This is possible thanks to adaptive algorithms. American scientists could prove it thanks to data sets containing over 2,000 so-called brain scans of over 1,000 Alzheimer's patients. Thanks to image data sets, computer systems have been trained to recognize Alzheimer's disease in the early stages. And indeed, after training, the algorithm correctly diagnosed Alzheimer's disease with 100% accuracy. Unrivaled even for experienced specialists.
The computer better recognizes changes in the brain
The reason for this is that even experienced physicians have difficulty recognizing and correctly classifying small changes in the brain. Although imaging techniques in medicine are getting better – even the smallest changes in brain metabolism can be visualized with positron emission tomography (PET) – even experienced physicians are often unable to adapt to the changes typical of the early stages of Alzheimer's disease.
Often Alzheimer's disease is diagnosed by doctors only when the first symptoms appear. Computer systems have a clear advantage here, as recent research by San Francisco researchers has shown: The American scientists' system was also able to identify Alzheimer's cases from images taken several years before the onset of symptoms.
Why is early diagnosis of Alzheimer's important?
According to the latest data from the German Alzheimer's Society, 1.7 million people now live with dementia in Germany. Most of them have Alzheimer's disease. According to experts, the number of dementia patients in the world is expected to triple by 2050. Therefore, good therapies are important. The newly published study is only a pilot study, which should be reviewed along with further research. Nevertheless, scientists already see in the newly discovered discoveries a significant potential for better treatment of patients with Alzheimer's disease.
"If we recognize Alzheimer's disease more quickly, scientists will have the opportunity to find new and better ways to slow down or even stop disease progression." Jae Ho Sohn from the Department of Radiology and Biomedical University of California San Francisco and co-author of the study
Similar research, such as American scientists, has been published by the McGill University team in Toronto in early October. Mallar Chakravarty researchers also believe that artificial intelligence may soon help doctors diagnose Alzheimer's disease.