Science Daily · Feb 10, 2026 · Collected from RSS
Researchers at the University of Michigan have created an AI system that can interpret brain MRI scans in just seconds, accurately identifying a wide range of neurological conditions and determining which cases need urgent care. Trained on hundreds of thousands of real-world scans along with patient histories, the model achieved accuracy as high as 97.5% and outperformed other advanced AI tools.
A newly developed artificial intelligence system from the University of Michigan can analyze brain MRI scans and deliver a diagnosis in a matter of seconds, according to a new study. The model identified neurological conditions with accuracy reaching 97.5% and was also able to assess how urgently patients needed medical care. Researchers say this first-of-its-kind technology has the potential to reshape how brain imaging is handled across health systems in the United States. The findings were published in Nature Biomedical Engineering. "As the global demand for MRI rises and places significant strain our physicians and health systems, our AI model has potential to reduce burden by improving diagnosis and treatment with fast, accurate information," said senior author Todd Hollon, M.D., a neurosurgeon at University of Michigan Health and assistant professor of neurosurgery at U-M Medical School. Testing the Prima AI System Hollon named the new technology Prima. Over a one-year period, his research team evaluated the system using more than 30,000 MRI studies. Across more than 50 different radiologic diagnoses involving major neurological disorders, Prima delivered stronger diagnostic performance than other advanced AI models. In addition to identifying disease, the system also proved capable of determining which cases required higher priority. Certain neurological conditions, including strokes and brain hemorrhages, demand immediate medical attention. Hollon said that in these situations, Prima can automatically alert health care providers so action can be taken quickly. The system was designed to notify the most appropriate subspecialist, such as a stroke neurologist or neurosurgeon. Feedback becomes available immediately after a patient completes imaging. "Accuracy is paramount when reading a brain MRI, but quick turnaround times are critical for timely diagnosis and improved outcomes," said Yiwei Lyu, M.S., co-first author and postdoctoral fellow of Computer Science and Engineering at U-M. "At key steps in the process, our results show how Prima can improve workflows and streamline clinical care without abandoning accuracy." What Is Prima? Prima is classified as a vision language model (VLM), a type of artificial intelligence that can process images, video, and text together in real time. While artificial intelligence has been applied to MRI analysis before, researchers say Prima takes a different approach. Earlier models were typically trained on carefully selected subsets of MRI data and designed to perform narrow tasks, such as identifying lesions or estimating dementia risk. Prima was trained on a much broader dataset. Hollon's team used every available MRI collected since radiology records were digitized at University of Michigan Health. This included more than 200,000 MRI studies and 5.6 million imaging sequences. The model also incorporated patients' clinical histories and the reasons physicians ordered each imaging study. "Prima works like a radiologist by integrating information regarding the patient's medical history and imaging data to produce a comprehensive understanding of their health," said co-first author Samir Harake, a data scientist in Hollon's Machine Learning in Neurosurgery Lab. "This enables better performance across a broad range of prediction tasks." Addressing MRI Delays and Radiology Shortages Each year, millions of MRI scans are performed worldwide, many of them focused on neurological disease. Researchers say the demand for these scans is growing faster than the availability of neuroradiology services. This imbalance has contributed to staffing shortages, diagnostic delays, and errors. Depending on where a patient receives a scan, results may take days or even longer to return. "Whether you are receiving a scan at a larger health system that is facing increasing volume or a rural hospital with limited resources, innovative technologies are needed to improve access to radiology services," said Vikas Gulani, M.D. Ph.D., co-author and chair of the Department of Radiology at U-M Health. "Our teams at University of Michigan have collaborated to develop a cutting-edge solution to this problem with tremendous, scalable potential." The Future of AI in Medical Imaging Although Prima performed strongly, researchers emphasize that the work is still in an early evaluation phase. Future research will focus on incorporating more detailed patient information and electronic medical record data to further improve diagnostic accuracy. This approach mirrors how radiologists and physicians interpret MRIs and other imaging studies in real clinical settings. While artificial intelligence is already used in health care, most existing systems are limited to narrowly defined tasks. Hollon describes Prima as "ChatGPT for medical imaging," noting that similar technology could eventually be adapted for other imaging types, including mammograms, chest X-rays and ultrasounds. "Like the way AI tools can help draft an email or provide recommendations, Prima aims to be a co-pilot for interpreting medical imaging studies," Hollon said. "We believe that Prima exemplifies the transformative potential of integrating health systems and AI-driven models to improve health care through innovation." Additional authors: Asadur Chowdury, M.S., Soumyanil Banerjee, M.S., Rachel Gologorsky, Shixuan Liu, Anna-Katharina Meissner, M.D., Akshay Rao, Chenhui Zhao, Akhil Kondepudi, Cheng Jiang, Xinhai Hou, Rushikesh S. Joshi, M.D., Volker Neuschmelting, M.D., Ashok Srinivasan, M.D., Dawn Kleindorfer, M.D., Brian Athey, Ph.D., Aditya Pandey, M.D., and Honglak Lee, Ph.D., all of University of Michigan. Funding/disclosures: This work was supported in part by the National Institute of Neurological Disorders and Stroke (K12NS080223) of the National Institutes of Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. This work was also supported by the Chan Zuckerberg Initiative (CZI), Frankel Institute for Heart and Brain Health, the Mark Trauner Brain Research Fund, the Zenkel Family Foundation, Ian's Friends Foundation and the UM Precision Health Investigators Awards grant program.