Researchers have deployed an AI model to reexamine data from a completed clinical trial of an Alzheimer's disease drug. They discovered that in individuals with early-stage, slowly progressing mild cognitive impairment—often preceding Alzheimer's—the medication reduced cognitive decline by 46%.
The team, using AI, divided the trial participants into two groups based on their projected progression towards Alzheimer's: slow or rapid progressors. This allowed them to assess the drug's impact on each group.
The findings were published in Nature Communications.
Nature CommunicationsThis approach of precisely selecting trial participants could potentially lead to more efficient clinical trials, reducing development costs for new medications and accelerating the discovery process.
The AI model from University of Cambridge researchers predicts the progression rate of people in early cognitive decline developing Alzheimer's. This model provides predictions that are three times more accurate than standard clinical evaluations which rely on memory tests, MRI scans, and blood analyses.
With this patient stratification tool, data from a previously completed clinical trial—one that showed no overall efficacy for the examined drug—was reanalyzed. The results revealed that while the drug successfully removed beta amyloid protein in both groups as intended, symptom improvement was only significant among slow-progressing patients at an early stage. Beta amyloid is one of the earliest markers to appear in Alzheimer's disease.
The latest findings present considerable implications. Employing AI to categorize patients into distinct groupings, like those progressing slowly versus rapidly towards Alzheimer’s, allows researchers a better opportunity to identify who benefits from treatment and could hasten the development of new Alzheimer's drugs.
Professor Zoe Kourtzi from Cambridge University's Department of Psychology and senior author of the report stated: "New hope-filled drugs often fail when given too late to patients beyond the point they can benefit. Our AI model lets us pinpoint these patients accurately, pairing them with optimal treatments. This precision accelerates trials' speed and reduces costs, fueling a personal medicine approach for potential dementia cures."
"Our AI model produces scores predicting each patient's Alzheimer’s progression rate, allowing precise segregation into slow- or fast-progressing groups to evaluate drug effects on each group."
The NHS's innovation arm in Eastern England, Health Innovation East, is now endorsing Kourtzi's efforts to integrate this AI-assisted method into clinical practice for patient benefit.
Joanna Dempsey, Principal Advisor at Health Innovation East England commented: "This AI-based approach could appreciably reduce NHS resource pressures and dementia care costs by streamlining drug development, pinpointing which patients will most benefit from treatments, thus accelerating access to effective medicines and targeted support for those with dementia."
Similar drugs are not intended as Alzheimer's cures but aim to slow cognitive decline and prevent symptom worsening.
Dementia remains the UK’s top cause of death and a significant global mortality factor. Annually, it incurs $1.3 trillion in costs, with projected tripling incidence by 2050. Despite ongoing research investments and breakthroughs, clinical trials for dementia treatments largely fail, exhibiting success rates below 5%. This is due to variations in symptom presentation, disease progression, and treatment responses among patients.
Although recently-approved dementia drugs in the US exist, side effect risks and cost inefficiencies have hindered NHS adoption.
A crucial aspect of treating Alzheimer's involves understanding individual patient differences for tailored therapies, as existing medications don't universally apply due to their complexity. "AI can guide us towards effective patients who benefit from these medicines by timing treatment when drugs are most helpful in fighting these cruel diseases. Clinical trials guided by AI promise faster progress and cheaper, more efficient discovery of precise treatments for individuals, thus reducing healthcare burdens," Kourtzi concludes.
Adding: "Like many, I've watched as dementia stole a loved one from me. We must speed up dementia medicine advancements. Over £40 billion has been spent in the past three decades on research and development - we can't afford another 30 years of waiting."