Summary: New artificial intelligence technology will analyze clinical data, brain images, and genetic information from Alzheimer’s patients to identify novel biomarkers and patterns associated with the disease.
Source: University of Pennsylvania
With effective Alzheimer’s treatments still scarce, researchers are turning to biomarkers—early biological indicators of disease—to improve diagnosis, patient stratification, and treatment development. Large-scale data collection from tens of thousands of patients has created an opportunity but also a challenge: integrating clinical records, brain imaging, and genetic data to extract meaningful, reproducible signals has outpaced traditional analytic methods.
To tackle this challenge, the Perelman School of Medicine at the University of Pennsylvania has received a $17.8 million grant from the National Institute on Aging at the NIH to lead a five-year, multi-center effort. Working with 11 research centers and drawing on datasets from more than 60,000 people affected by Alzheimer’s disease, the project will apply advanced artificial intelligence and machine learning techniques to unify and analyze genetic, imaging, and clinical data. This initiative aims to identify more precise diagnostic biomarkers and actionable drug targets for a disease that affects nearly 50 million people worldwide.
Penn Medicine investigators Christos Davatzikos, PhD, professor of Radiology and director of the Center for Biomedical Image Computing and Analytics, and Li Shen, PhD, professor of Informatics, are two of five co-principal investigators leading the effort. The team will collaborate with co-principal investigators at the University of Southern California, the University of Pittsburgh, and Indiana University, among others, to combine expertise in neuroimaging, genetics, clinical neurology, and computational science.
“Brain aging and neurodegenerative diseases, among which Alzheimer’s is the most frequent, are highly heterogeneous,” said Davatzikos. “This is an unprecedented attempt to dissect that heterogeneity, which may help inform treatment, as well as future clinical trials.” Heterogeneity in clinical presentation, imaging patterns, and genetic risk likely contributes to mixed outcomes in clinical trials; refining how patients are classified could make trials more efficient and increase the chance of identifying therapies that help specific subgroups.
Shen emphasizes the power of machine learning to detect complex, non-obvious relationships across datasets. “We know that there are complex patterns in the brain that we may not be able to detect visually. Similarly, there may not be a single genetic marker that puts someone at high risk for Alzheimer’s, but rather a combination of genes that may form a pattern and create a perfect storm,” he said. Advanced AI models can integrate diverse data types—structural and functional brain images, longitudinal cognitive scores, and genomic profiles—to reveal multivariate biomarkers that predict diagnosis, progression, and subtype.
The project’s first major aim is to uncover relationships across the three core modalities—genetic variants, neuroimaging features, and clinical symptoms—so researchers can identify patterns that predict Alzheimer’s onset and disease course. By analyzing these multimodal signatures, the team will work to redefine subtypes of Alzheimer’s, distinguishing groups with shared biological mechanisms and trajectories. This stratified approach seeks to move beyond a one-size-fits-all disease label toward precision definitions that align with potential targeted interventions.
Following biomarker discovery, the investigators will build predictive models of cognitive decline and disease progression. These models aim to provide clinically useful tools that can forecast individual trajectories, identify patients at highest risk for rapid decline, and inform personalized treatment decisions. In practical terms, better predictive tools could guide clinicians in choosing interventions, help design adaptive clinical trials, and prioritize candidates for novel therapies.
A key data source for this undertaking is the Alzheimer’s Disease Sequencing Project, an NIH-funded consortium led at Penn by Gerard Schellenberg, PhD, and Li-San Wang, PhD, together with colleagues from around 40 institutions. That sequencing initiative focuses on discovering genomic variants that either increase risk for Alzheimer’s or confer protection, complementing imaging and clinical data to provide a richer biological context for interpretation.
The scale of the project—integrating data from more than 60,000 individuals and coordinating across a broad research network—makes it one of the most ambitious attempts to use AI for Alzheimer’s biomarker discovery. By leveraging sophisticated computational methods and large, diverse datasets, the team aims to translate complex data into robust biomarkers and predictive tools that can accelerate development of targeted therapies and improve outcomes for patients and families affected by Alzheimer’s disease.
About this artificial intelligence research news
Source: University of Pennsylvania
Contact: Lauren Ingeno – University of Pennsylvania
Image: The image is in the public domain