Researchers at the University of California-San Francisco (UCSF) have trained a computer system to detect early signs of Alzheimer's disease in medical imagery. Their success is striking: technology has identified early signs of Alzheimer's disease faster than humans. For many people with dementia, it could soon be a way to diagnose and therefore improve therapy earlier.
Half Alcimer Accuracy
The reason for this hope is a pilot study by Jae Ho Sohn's team, now published in the journal Radiology. Conclusion in the study: under certain circumstances, computers can recognize some models better than humans. This is possible with adaptive algorithms. US researchers have been able to prove this through more than 2000 so-called brain scans for over 1,000 Alzheimer's patient data sets. With image data sets, computer systems were trained to recognize Alzheimer's disease at an early stage. And indeed, after training, the algorithm properly diagnosed Alzheimer's disease with 100% accuracy. Unmatched by even experienced professionals.
The computer recognizes the changes in the brain better
The reason for this is that even experienced physicians find it difficult to recognize and correctly classify minor changes in the brain. Even though drug imaging techniques are getting better and better, even positron emission tomography (PET) can even visualize even the worst changes in brain metabolism – even experienced doctors are often unable to adapt to the changes in the early stages of Alzheimer's disease.
Therefore, Alzheimer's patients are often diagnosed only when the first symptoms appear. Computerized systems have clear benefits here, as shown by the recent study by San Francisco: The US researchers system was able to identify Alzheimer's disease based on images taken several years before the first symptoms occurred.
Why is Alzheimer's Disease Important?
According to recent German Alzheimer's data, in Germany, 1.7 million people now live with dementia. Most of them have Alzheimer's disease. According to experts, by 2050, patients with dementia worldwide are expected to triple. That's why good therapy is important. A recent study is just a pilot study that needs to be reviewed in further investigations. However, researchers are already seeing the latest findings, which is a significant potential for better treatment for Alzheimer's patients.
"If we recognize Alzheimer's disease in the past, it will enable researchers to find new and better ways to slowly or even stop the progression of the disease." Jae Ho Sohn, co-author of the Department of Radiology and Biomedical Representation at the University of California, San Francisco
Similar studies, such as US researchers, were published in early October at McGill University's Toronto team. Mallar Chakravarty researchers also believe that artificial intelligence could soon help doctors diagnose Alzheimer's.