Artificial Intelligence could help diagnose dementia in 24 hours

Scientists are trialling an artificial-intelligence (AI)  system thought to be able to diagnose dementia after a single brain scan. It may also predict whether the condition will remain constant for many years, gradually deteriorate, or require immediate treatment. Currently, it can take several scans and tests to diagnose dementia.

The researchers involved say earlier diagnoses with their system could greatly improve patient outcomes. “If we intervene early, the treatments can kick in early and slow down the progression of the disease and at the same time avoid more damage,” Prof Zoe Kourtzi, of Cambridge University and a fellow of the national centre for AI and data science – The Alan Turing Institute, said. “And it’s likely that symptoms occur much later in life or may never occur.”

Dr Timothy Rittman, clinical lead of the QMIN-MC trial which uses a machine learning algorithm developed by Prof Kourtzi, says the algorithm trains itself to diagnose patients by looking at MRI brain scans to identify patterns. It then combines these finding with the results of standard memory tests.

Dr Rittman, explained: “Traditionally, when we look at patient scans we are looking for patterns to be able to help us exclude things like strokes and brain tumours. The computer can do this much more comprehensively than any human, helping to give us not only a more accurate diagnosis, but also a prognosis as well. With a better prognosis we can identify how quickly a patient is moving away from the normal pattern of the disease and amend their treatment and care accordingly.”

The Cambridge University team will now work with doctors at Addenbrooke’s Hospital in Cambridge to test the AI approach in a real-world setting, with patients coming into the memory clinic with concerns about dementia.