AI tool outperforms clinical tests at predicting Alzheimer’s disease

An AI tool can predict the progress of Alzheimer’s disease better than current clinical diagnostic tools, according to research published by eClinical Medicine.

The research, led by scientists from the Department of Psychology at the University of Cambridge, included a machine learning model trained on non-invasive, routinely collected patient data such as cognitive tests and structural MRI scans.

The team then tested the model using real-world patient data from 1,500 participants in the US, UK and Singapore.

According to the study, the algorithm distinguished people with stable mild cognitive impairment and those who progressed to Alzheimer’s disease within a three-year period.

It identified individuals who went on to develop Alzheimer’s in just over 80% of cases. It also correctly identified those who didn’t in a little over 80% of cases, too.

It categorised participants whose symptoms would remain stable, those who would progress slowly, and those who would progress more rapidly – as validated with follow-up data over 6 years.

The research team found that it was three times more accurate at predicting the progression of Alzheimer’s than the current standard of care.

Read: AI identifies early breast cancer that doctors missed

It claims this will help reduce misdiagnosis, improve new patient treatments, and better identify those who need closer monitoring.

It will also direct those struggling with memory loss, but predicted to remain stable, to a different clinical pathway as their symptoms may be due to other issues such as anxiety or depression.

“This has the potential to significantly improve patient wellbeing, showing us which people need closest care, while removing the anxiety for those patients we predict will remain stable.

At a time of intense pressure on healthcare resources, this will also help remove the need for unnecessary invasive and costly diagnostic tests,” said senior author, Professor Zoe Kourtzi from the Department of Psychology at the University of Cambridge.

Kourtzi added: “AI models are only as good as the data they are trained on. To make sure ours has the potential to be adopted in a healthcare setting, we trained and tested it on routinely collected data not just from research cohorts, but from patients in actual memory clinics. This shows it will be generalisable to a real-world setting.”