AI model demonstrates potential for streamlining clinical trials

New research has shown that AI can improve the precision of patient selection for clinical trials. 

By choosing individuals who are most likely to benefit from treatment, the approach could help to reduce the cost of developing new medicines by streamlining clinical trials.

Scientists tested their model on data for an Alzheimer’s drug.

“Our AI model gives us a score to show how quickly each patient will progress towards Alzheimer’s disease. This allowed us to precisely split the patients on the clinical trial into two groups – slow, and fast progressing, so we could look at the effects of the drug on each group,” explained Professor Zoe Kourtzi in the University of Cambridge’s Department of Psychology, senior author of the report.

Using this patient stratification model, Vaghari et al. reanalysed results of a completed clinical trial for an Alzheimer’s drug that did not demonstrate efficacy in the total population studied.

While the drug cleared beta amyloid protein in both patient groups, changes were only seen in individuals with early stage, slow-disease progression.

Notably, their AI model demonstrated that the medicine slowed cognitive decline by 46 percent in a group of patients with early stage, slow-progressing mild cognitive impairment.

Enabling personalised Alzheimer’s drug development with AI

“… Making clinical trials faster, cheaper and better, guided by AI has strong potential to accelerate discovery of new precise treatments for individual patients, reducing side effects and costs for healthcare services,” explained Professor Kourtzi.

“This AI-enabled approach could have a significant impact on easing NHS pressure and costs in dementia care by enabling more personalised drug development – identifying which patients are most likely to benefit from treatment, resulting in faster access to effective medicines and targeted support for people living with dementia,” Joanna Dempsey, Principal Advisor at Health Innovation East England shared.

Results from the study were published in Nature Communications.