AI recognizes Alzheimer's already six years before diagnosis by doctors

Allows an AI to breakthrough in the diagnosis of Alzheimer's?

Researchers from California have created an artificial intelligence that can detect early on the first signs of Alzheimer's disease. The algorithm is able to detect first signs of Alzheimer's on average six years before diagnosis by a physician.

Researchers at the University of California found that newly developed artificial intelligence can detect Alzheimer's disease years before it can be diagnosed by a physician. The experts published the results of their study in the English-language journal "Radiology".

Ki detected Alzheimer's on average 75.8 months earlier

The AI ​​was able to scan images of patients' brains and detect Alzheimer's on average 75.8 months before the actual diagnosis. In the modern diagnostic method called F-FDG-PET (or Fluorine 18 (18F) fluorodeoxyglucose positron emission tomography), a radioactive glucose dye is passed through the brain and photographed. Afterwards, experts examine and interpret the images with the naked eye for the first signs of Alzheimer's disease. Because this method relies on pattern recognition, researchers saw the opportunity to significantly improve their performance through the use of a self-learning AI algorithm.

AI recognizes special features or patterns in the images

The scientists hypothesized that the deep-learning algorithm might detect features or patterns that are not apparent in the standard clinical examination of images, thereby improving the final diagnosis classification of individuals. Physicians wanted to find out if a profound learning algorithm can be trained to predict the final clinical diagnosis in patients who have undergone F-FDG-PET.

More than 1,000 subjects participated in the study

The study analyzed 2,109 images of 1,002 patients. The researchers found that the algorithm Alzheimer could already recognize in the images, although these were taken on average more than six years before the diagnosis. These results are the latest in a series of studies demonstrating the potential of AI in preventive health care and diagnosis. For example, the Francis Crick Institute reported in September that an AI in a study learned to predict cardiac mortality rates in patients with a higher level of accuracy than trained physicians and experts.

AI recognizes Alzheimer's with high accuracy

Despite some limiting factors, including a small sample size, the Californian researchers conclude in the current investigation that they have developed an algorithm that can predict Alzheimer's with high accuracy and reliability. By accessing a much larger data volume and the ability to calibrate the model, the developed algorithm could later be directly integrated into the workflow of medical professionals, say the experts. (As)