Can Alzheimer’s Progression Be Predicted? New AI Model Does It 80% of the Time

On a global scale dementia is a major health problem, affecting more than 55 million people worldwide and costing an estimated $820 billion annually. Moreover, cases are expected to nearly triple in the next 50 years. The main type is Alzheimer’s diseaseaccounting for 60 to 80% of cases. Although early detection of this disease is critical to improving the effectiveness of treatment, early diagnosis and prognosis of the disease dementia They can be inaccurate without the use of invasive or expensive tests, resulting in up to a third of patients being misdiagnosed.

To solve this problem, a team led by scientists from Department of Psychology belonging Cambridge university developed a machine learning model that can predict whether and how quickly a person with mild memory and thinking problems will develop Alzheimer’s disease. The study, published in Electronic clinical medicineshow that it is more accurate than current clinical diagnostic tools. With this new approach, the need for invasive and expensive diagnostic tests can be reduced while improving early treatment outcomes.

“We have created a tool that, despite using only cognitive test data and MRI, is much more sensitive than current approaches.”

“We have created a tool that, despite using only cognitive test and MRI data, is much more sensitive than current approaches in predicting whether someone will progress from mild symptoms to Alzheimer’s disease and if so, will this progress be fast or slow,” says the professor. Zoe Kurzibelonging Department of Psychology belonging Cambridge university and lead author. “This has the potential to significantly improve patient well-being by showing us which people need more careful care, as well as relieving anxiety in those patients we expect to remain stable. At a time of intense pressure on healthcare resources, it could also help eliminate the need for unnecessary invasive and expensive diagnostic tests.”

The team now hopes to extend their model to other forms of dementia, such as vascular dementia and frontotemporal dementia, and use different types of data, such as blood test markers. “If we are to tackle the growing public health problem of dementia, we will need better tools to detect it and intervene at the earliest possible stage. “Our tool could help recruit the right patients into clinical trials, accelerating the discovery of new drugs for disease-modifying treatments.”

HOW WAS THIS DEVELOPED?

The researchers used routinely collected, noninvasive, and inexpensive patient data, such as cognitive testing and structural MRIs showing gray matter atrophy, from more than 400 people. They then tested the model using real-life patient data from another 600 participants in the same research group in USAin addition to longitudinal data from 900 individuals from memory clinics in Great Britain And Singapore. The algorithm was able to distinguish people with stable mild cognitive impairment from those whose decline was progressing. Alzheimer’s disease within three years.

In doing so, he correctly identified the people who developed it. Alzheimer’s disease in 82% of cases and in those who did not develop it, in 81% of cases. All this is based on cognitive tests and MRI. On the other hand, the algorithm was more accurate in predicting progress towards Alzheimer’s disease than standard clinical markers such as gray matter atrophy or cognitive performance. Specifically, it was three times more accurate than other markers. At the same time, it was shown that the model can significantly reduce the number of misdiagnoses.

The algorithm was able to distinguish between people with stable mild cognitive impairment and those who developed Alzheimer’s disease within three years.

The model also allowed the researchers to stratify people with Alzheimer’s disease, using data from each person’s first visit to a memory clinic, into three groups: those whose symptoms would remain stable (about 50% of participants), those whose symptoms would slowly progress toward Alzheimer’s disease (about 35%) and those who will progress faster (the remaining 15%). The predictions were confirmed by examining follow-up data for six years.

This can help identify these people early enough to benefit from new treatments, and also identify people who need close monitoring as their condition can deteriorate quickly. It is important to note that the 50% of people who have symptoms such as memory loss but remain stable are best referred to another clinical pathway as their symptoms may be due to causes other than the disease itself. dementiasuch as anxiety or depression.

While the researchers tested the algorithm on the research group’s data, they validated it using independent data including nearly 900 people who visited memory clinics in the United States. Great Britain And Singapore. In particular, in Great Britain Patients were recruited through quantitative MRI studies in memory clinics National Health Service (QMIN-MC), under the supervision of the study’s co-author, Dr. Timothy Rittmanbelonging Cambridge University Hospitals NHS Trust and from Cambridgeshire and Peterborough NHS Foundation Trust (KPFT). According to the researchers, this shows that it should be used in clinical settings with real patients.