University of Cambridge AI tool ‘three times more accurate’ at predicting patients that will progress to Alzheimer’s than current method

A tool powered by artificial intelligence capable of predicting in four cases out of five whether people with early signs of dementia will remain stable or develop Alzheimer’s disease has been created by Cambridge scientists.

They believe it could reduce the need for invasive and costly diagnostic tests and give interventions such as lifestyle changes or new medicines the best chance of success.

Alzheimer’s accounts for 60 to 80 per cent of the 55 million cases of dementia worldwide, which have an estimated healthcare cost of $820billion. The number of cases is expected to almost treble over the next half century.

There is no cure, but treatments are more effective if the disease is diagnosed early.

But early dementia diagnosis and prognosis is not always accurate without the use of invasive or expensive tests such as positron emission tomography (PET) scans or lumbar puncture - and these are not available in all memory clinics.

It means up to a third of patients may be misdiagnosed, while others are diagnosed too late for treatment to be effective.

But now a machine learning model has been developed by a team led by scientists from the Department of Psychology at the University of Cambridge which is able to predict whether and how fast an individual with mild memory and thinking problems will progress to developing Alzheimer’s disease.

Senior author Prof Zoe Kourtzi, from the Department of Psychology, said: “We’ve created a tool which, despite using only data from cognitive tests and MRI scans, is much more sensitive than current approaches at predicting whether someone will progress from mild symptoms to Alzheimer’s – and if so, whether this progress will be fast or slow.

“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.”

The model was built using routinely-collected, non-invasive, low-cost data from cognitive tests and structural MRI scans showing grey matter atrophy from more than 400 individuals who were part of a research cohort in the USA.

They tested the model using real-world patient data from a further 600 participants in the US study and longitudinal data from 900 people at memory clinics in the UK and Singapore.

The algorithm correctly identified those who went on to develop Alzheimer’s within three years in 82 per cent of cases and identified those who did not in 81 per cent of cases from cognitive tests and an MRI scan alone.

That makes it about three times more accurate at predicting the progression to Alzheimer’s than the current standard of care, which uses clinical markers such as grey matter atrophy or cognitive scores, or clinical diagnosis.

Dr Ben Underwood, honorary consultant psychiatrist at Cambridgeshire and Peterborough NHS Foundation Trusts (CPFT) and assistant professor at the Department of Psychiatry, said: “Memory problems are common as we get older. In clinic, I see how uncertainty about whether these might be the first signs of dementia can cause a lot of worry for people and their families, as well as being frustrating for doctors who would much prefer to give definitive answers.

“The fact that we might be able to reduce this uncertainty with information we already have is exciting and is likely to become even more important as new treatments emerge.”

The researchers were able to stratify people with Alzheimer’s using data from each person’s first visit at the memory clinic into three groups - those whose symptoms would remain stable (around 50 per cent of participants), those who would progress to Alzheimer’s slowly (around 35 per cent) and those who would progress more rapidly (the remaining 15 per cent).

The predictions were validated when looking at follow-up data over six years, meaning it could help identify patients at an early enough stage to benefit from new treatments, and indicating those people who need close monitoring because their condition is likely to deteriorate rapidly.

The 50 per cent of people who have symptoms such as memory loss but remain stable could then be directed to a different clinical pathway, since their symptoms have other causes than dementia, such as anxiety or depression.

Prof Kourtzi said: “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.”

The team aim to extend their model to other forms of dementia, such as vascular dementia and frontotemporal dementia, and using different types of data, such as markers from blood tests.

Prof Kourtzi added: “If we’re going to tackle the growing health challenge presented by dementia, we will need better tools for identifying and intervening at the earliest possible stage.

“Our vision is to scale up our AI tool to help clinicians assign the right person at the right time to the right diagnostic and treatment pathway. Our tool can help match the right patients to clinical trials, accelerating new drug discovery for disease modifying treatments.”

The research was published in eClinical Medicine on Friday (12 July).