Researchers at the University of Cambridge have developed an artificial intelligence (AI) model that could reshape Alzheimer’s clinical trials by pinpointing patient subgroups with different rates of disease progression.
Study: AI-guided patient stratification improves outcomes and efficiency in the AMARANTH Alzheimer’s Disease clinical trial. Image Credit: Orawan Pattarawimonchai/Shutterstock.com
When applied retrospectively to a previous trial, the AI uncovered a striking 46 % slower cognitive decline in early-stage, slow-progressing patients treated with an amyloid-clearing drug—a benefit that was overlooked in the original analysis. This precision-based method could help design more targeted trials, potentially lowering costs and improving success rates in dementia research.
Why Alzheimer’s Trials Keep Failing
Alzheimer’s is notoriously difficult to treat, and clinical trial failures are more the norm than the exception. More than 95 % of Alzheimer’s trials have failed, despite over $43 billion invested globally. One major reason for this is that not all patients decline at the same pace or respond to treatments in the same way. When trials group together a broad mix of participants, those differences can mask a drug’s true effect.
The Cambridge team aimed to tackle this issue head-on. They created an AI model designed to predict how fast a patient’s condition will progress—something traditional methods, like brain scans or memory tests, often fail to do reliably.
What the AI Found
To test their model, researchers applied it retrospectively to data from a major Alzheimer’s trial that had originally been considered unsuccessful. The trial tested an amyloid-clearing drug, but the results didn’t show a clear benefit overall.
When the AI model re-analyzed the data, it grouped participants based on how fast their disease was likely to progress. The results were striking: for early-stage patients with slow-progressing Alzheimer’s, the drug appeared to slow cognitive decline by 46 %. This subgroup-level effect had been lost in the noise of the broader trial population.
In short, the treatment may have worked, but only for a specific group that wasn’t clearly identified during the original study.
Smarter Tools for Smarter Trials
The AI model makes its predictions by combining multiple data sources—cognitive tests, biomarkers, and brain scans—to estimate individual trajectories of decline. This trial's forecasts were three times more accurate than those of standard clinical methods.
This level of precision opens up new possibilities for designing trials. Rather than enrolling broad, mixed populations, future studies could focus on more defined subgroups—those most likely to benefit from a given treatment. Researchers estimate this approach could cut trial sizes in half and speed up timelines, making drug development for dementia faster and more cost-effective.
While the AI model was tested on clinical trial data, its real-world potential could be even broader. Doctors could eventually use it to forecast how an individual’s condition is likely to evolve and tailor treatment plans accordingly, much like how cardiologists use risk calculators to guide decisions about statins.
The model is already being explored for clinical use by Health Innovation East England. If it performs well in larger, more diverse populations, it could support more personalized care and improve outcomes for people in the early stages of Alzheimer’s.
A Step Toward Precision in Dementia Treatment
Although questions remain—such as how to ethically handle fast-progressing patients who may be excluded from certain treatments—this work marks a promising step toward more personalized and efficient Alzheimer’s drug development. As senior author Professor Zoe Kourtzi put it, the model “brings us closer to precision medicine for dementia.”
With global dementia cases expected to triple by 2050, finding smarter ways to identify who will benefit from which treatments is no longer optional—it’s urgent.
Journal Reference
Vaghari, D., Mohankumar, G., Tan, K., Lowe, A., Shering, C., Tino, P., & Kourtzi, Z. (2025). AI-guided patient stratification improves outcomes and efficiency in the AMARANTH Alzheimer’s Disease clinical trial. Nature Communications, 16(1), 1–12. DOI:10.1038/s41467-025-61355-3. https://www.nature.com/articles/s41467-025-61355-3
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.