Artificial Intelligence now allows for the prediction of Alzheimer's disease

A new breakthrough in artificial intelligence now allows for the prediction of Alzheimer's disease up to seven years in advance. While there is no cure for Alzheimer's yet, early detection could enable preparations and potentially preventative measures.

Researchers from the University of California, San Francisco (UCSF) and Stanford University utilized machine learning techniques on over 5 million health records to train an AI model to identify patterns linking Alzheimer's with other health conditions.

Although not flawless, the AI system demonstrated promising results when tested against records of individuals who later developed Alzheimer's, accurately predicting its onset 72 percent of the time, sometimes up to seven years before symptoms appeared.

The AI's predictive ability lies in its capacity to analyze various risk factors and calculate the likelihood of Alzheimer's development. This approach could deepen our understanding of the disease's origins and identify those most susceptible to it.

"This is a significant step toward leveraging AI on regular clinical data to not only identify risk as early as possible but also comprehend its underlying biology," stated bioengineer Alice Tang from UCSF.

The model highlighted several conditions linked to Alzheimer's risk, such as high blood pressure, high cholesterol, vitamin D deficiency, depression, erectile dysfunction, and an enlarged prostate in men, along with osteoporosis in women.

While these conditions don't guarantee dementia development, the AI considers them as important predictive factors. Researchers hope similar machine learning techniques could be used to identify risk factors for other challenging-to-diagnose diseases.

"It is the combination of diseases that allows our model to predict Alzheimer's onset," Tang explained. "Our finding that osteoporosis is one predictive factor for females highlights the biological interplay between bone health and dementia risk."

The study also investigated the biological connections behind some of these links, revealing a connection between osteoporosis, Alzheimer's in women, and a variant in the gene MS4A6A, offering new avenues for studying the disorder's progression.

"This is a great example of how we can utilize patient data with machine learning to predict which patients are more likely to develop Alzheimer's and also to understand the reasons why," said Marina Sirota, a computational health scientist at UCSF.

The research has been published in Nature Aging.