Portable MRI shows promise for detecting strokes in emergency settings, study finds

A portable, AI-powered MRI system designed to bring brain imaging directly to the bedside demonstrated strong performance in detecting strokes, including very small ischemic lesions, according to results from the largest data set yet evaluating the technology.

Hyperfine Inc. said a prospective, multi-center observational study published in the November issue of Stroke: Vascular and Interventional Neurology evaluated 95 patients and found that its next-generation Swoop portable MRI system significantly improved diagnostic accuracy and efficiency compared with earlier versions of the scanner. The data support the use of portable MRI for stroke detection in multiple clinical settings, including hospital emergency departments, where time is critical.

The study combined data from the ACTION PMR study at Massachusetts General Hospital and Buffalo General Medical Center with additional patients from Yale New Haven Hospital. Researchers assessed how well the Swoop system detected ischemic lesions using diffusion-weighted imaging, or DWI, a type of MRI sequence considered essential for identifying acute stroke. Performance of the original Swoop scanner was compared with a next-generation system using an advanced, multi-directional DWI sequence.

According to the findings, the next-generation Swoop system was able to identify lesions as small as 2.8 millimeters, or 0.15 milliliters in volume, allowing clinicians to detect very small strokes. For clinically relevant lesions larger than 1 milliliter, the system achieved 100% sensitivity and 100% specificity. Scan times were reduced by about 30%, and image quality across the brain improved, boosting diagnostic confidence.

“We previously showed that using DWI in combination with FLAIR on the portable MRI system can be used as a ‘tissue clock’ for stroke detection, similar to conventional MRI,” said Taylor Kimberly, MD, PhD, chief of the Neurocritical Care Division at Mass General Brigham. “With this study, we took the next step and evaluated the capability of ultra-low-field MRI with advanced, multi-directional DWI sequences to detect very small ischemic lesions. The results show that the next-generation portable MRI system with a multi-directional DWI sequence enables detection of very small strokes in a clinically feasible timeframe. The portable MRI system’s ability to detect clinically relevant strokes opens new possibilities for transforming stroke diagnosis and management—bringing timely evaluation to more patients and care settings than ever before.”

The Swoop system is an ultra-low-field, portable MRI scanner that can be wheeled to a patient’s bedside and plugged into a standard electrical outlet. Unlike conventional MRI machines, which are fixed installations requiring shielded rooms and patient transport, the Swoop system is designed for use in settings where a full diagnostic MRI exam may not be practical. Cleared by the U.S. Food and Drug Administration for brain imaging in patients of all ages, the system uses artificial intelligence to help reconstruct images of the brain that trained physicians can interpret to aid diagnosis.

“Stroke detection represents a critical driver of the Swoop system’s expansion into emergency departments,” said Maria Sainz, president and CEO of Hyperfine. “The results from our next-generation Swoop system, combined with our new, advanced multi-direction DWI sequence that was recently cleared by the FDA, are truly remarkable. This data gives us even greater confidence that the Swoop system can reliably detect clinically relevant strokes, streamline workflows, and further strengthen the value of integrating portable MRI into stroke diagnosis and care.”

Hyperfine said it provided portable MRI systems under sponsored research agreements but was not involved in the design or analysis of the investigator-initiated study or in the decision to publish the results.

Portable imaging and AI reshape stroke diagnosis

The findings come amid broader advances in stroke imaging and neurocritical care, where speed, access and precision increasingly determine patient outcomes. Stroke treatment is highly time dependent, with therapies such as thrombolysis and mechanical thrombectomy most effective when delivered within narrow windows. As a result, researchers and device makers have focused on technologies that can shorten the time from symptom onset to diagnosis.

One major trend is the integration of artificial intelligence into medical imaging. AI algorithms are now routinely used to accelerate image reconstruction, improve image quality from lower-field scanners, and automatically flag suspected abnormalities. In stroke care, AI tools have been deployed to detect large vessel occlusions on CT angiography, estimate infarct core size, and prioritize scans for rapid review by clinicians. Applying these techniques to ultra-low-field MRI has helped overcome historical limitations in image resolution and signal quality.

Another key development is the push to decentralize imaging. Traditional MRI scanners are expensive, immobile and often located far from emergency departments or intensive care units. Transporting critically ill or unstable patients to radiology suites can introduce delays and risks. Portable imaging devices, including mobile CT and MRI systems, aim to bring diagnostic capability directly to the patient, whether in the ED, ICU or even resource-limited settings.

Advances in MRI sequence design have also played a role. Improved diffusion-weighted imaging, faster acquisition techniques and better motion correction have made it more feasible to obtain diagnostically useful scans in shorter timeframes. For stroke evaluation, the ability to visualize small ischemic lesions and distinguish acute from older injury is particularly important for treatment decisions.

Together, these developments reflect a broader shift toward faster, more accessible neuroimaging that supports clinical decision-making at the point of care. As portable MRI systems continue to improve and accumulate clinical evidence, they may complement conventional imaging by expanding access to timely stroke diagnosis in settings where it was previously difficult or impossible.