A new AI tool can predict whether people with mild memory and mental agility problems are likely to go on to develop Alzheimer's disease in future - without the need for invasive or costly diagnostic tests.
The tool would allow those at risk to modify their lifestyles or start new drug treatments at an early stage when they are most effective.
It would also prevent inappropriate treatment of people with cognitive problems likely to be caused by other conditions, such as anxiety and depression.
Scientists at the University of Cambridge used the artificially intelligent algorithm to analyse cognitive tests and MRI brain scans from 1,500 patients in the UK, USA and Singapore.
It was able to distinguish people with mild mental agility problems that would remain stable from those who would progress to Alzheimer's disease over the following three years.
The tool's prediction was more than 80% accurate, three times better than existing clinical methods for identifying patients likely to develop the disease, according to the study published in the journal eClinicalMedicine.
Professor Zoe Kourtzi, the study's senior author, said the AI tool could also predict whether a patient's symptoms would deteriorate slowly or more rapidly.
"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," she said.
Being able to accurately identify patients likely to develop Alzheimer's by using only routine clinical data and MRI brain scans would be game-changing for the NHS.
Currently an accurate diagnosis requires either an expensive PET brain scanner or a sample of spinal fluid taken by specially trained staff. The NHS is short of both.
The lack of resources could hamper access to new drugs that can slow the progression of symptoms - but only if patients are diagnosed in the early stages of the disease.
Dr Ben Underwood, honorary consultant psychiatrist at Cambridgeshire and Peterborough NHS Foundation Trust, said he frequently sees people with memory problems.
"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," he said.
"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."
Alison Gilderdale first started repeating herself and struggling with her memory a decade ago. But it took six years for the symptoms to become clear enough for doctors to diagnose Alzheimer's.
An earlier diagnosis would have helped her recognise what was happening to her.
"I thought I was ok and it was everyone else saying 'she's not right'," she said.
"Now I get lots of help. Things like getting dressed were difficult."