First-Of-Its-Kind AI-Driven Brain Imaging Platform to Better Guide Stroke Treatment Options

By MedImaging International staff writers
Posted on 05 Mar 2025

Image: Composite samples of the MRI images that AStrID will draw on to advance research that promises to optimize stroke rehabilitation (Photo courtesy of GU School of Medicine)

Each year, approximately 800,000 people in the U.S. experience strokes, with marginalized and minoritized groups being disproportionately affected. Strokes vary in terms of size and location within the brain, making recovery studies particularly challenging. Each survivor faces different difficulties depending on where the stroke occurred and its severity, such as weakness on one side of the body or difficulty speaking. Some treatments designed to aid recovery may only benefit individuals with strokes in specific areas of the brain. Until now, it has been difficult to quickly identify individuals based on the size and location of their stroke in order to test these treatments. Now, a new artificial intelligence (AI)-driven brain imaging platform is expected to significantly improve treatment guidance and transform stroke recovery research.

A group of stroke physician-scientists at Georgetown University Medical Center (Washington, D.C., USA) has launched the Acute Stroke Imaging Database (AStrID), an initiative that scans MRI images to automatically identify the type and location of strokes in approximately 5,000 acute stroke cases seen annually at the MedStar Health system (Columbia, MD, USA). MedStar Health serves a diverse patient population, representing a wide range of socioeconomic, cultural, and racial backgrounds from 10 hospitals in Maryland, Washington, D.C., and Virginia. This makes it an ideal sample group to reflect the U.S. population. AStrID is part of a larger stroke registry that reviews electronic medical records of stroke patients to aid in research efforts. By allowing researchers to group stroke survivors based on the specifics of their strokes, AStrID enables the identification of the most appropriate clinical trials and treatments for each patient, thereby enhancing the reliability of stroke recovery studies.

The AStrID development process began with hundreds of training images in which the researchers manually identified the location of the strokes. They then fed this information into the AStrID learning algorithm, which, based on these training sets, can detect stroke locations in new MRI images. Using the platform’s search tool, researchers can pinpoint images in AStrID that display strokes in specific brain regions. In the next year, the researchers aim to include up to 20,000 stroke images in AStrID, potentially making it the largest stroke imaging repository globally. To protect patient confidentiality, AStrID removes all identifying details from the MRIs before processing them, storing only the digital images that mark the stroke locations, not the actual MRI scans. These images are kept separately from the electronic medical records within the broader stroke registry, with access to the two being restricted and linked only through anonymized codes.

The research team’s primary focus is on helping patients recover language abilities post-stroke. Approximately one-third of stroke survivors experience aphasia, a condition that impairs language use. After a stroke, the brain has the capacity to reorganize itself and form new connections, a process known as brain plasticity, which enables recovery of lost functions over time. Enhancing brain plasticity could improve recovery from aphasia, but the pattern of this plasticity varies based on the size and location of each stroke. With AStrID, the researchers can identify groups of patients with similar stroke profiles, helping them uncover previously hidden patterns of brain plasticity that could potentially be targeted by new treatments to enhance recovery.

“Disparities in access to services affect every phase of stroke care,” said Peter E. Turkeltaub, MD, PhD, director of the Cognitive Recovery Lab at Georgetown University and the Aphasia Clinic at MedStar National Rehabilitation Hospital. “AStrID and the broader stroke registry can help us understand how these disparities affect outcomes. We can identify groups of people who we’d expect to have similar recoveries because they have very similar strokes, and then assess if access to rehabilitation or other stroke services changes their recovery. That will help us find the people who would benefit most from better stroke services to improve equity and outcomes.”