AI-Enhanced Handheld Ultrasound Improves Carotid Plaque Detection

This gap makes it harder for frontline clinicians to decide who needs confirmatory imaging, closer follow-up, or more intensive management of stroke risk factors. Researchers in Hunan Province, China, developed an artificial intelligence model that sharpens handheld ultrasound images after acquisition. The newly introduced technology aims to improve carotid plaque detection in community and primary care settings.

The team built an AI-enhanced super‑resolution approach tailored to handheld ultrasound. The method is designed for use after images are captured, without changing scanning workflows. The work was led at the Key Laboratory of Medical Imaging Precision Theranostics and Radiation Protection, College of Hunan Province, Hengyang Medical School, The Affiliated Changsha Central Hospital, University of South China, Changsha, China. 

The AI-enhanced super‑resolution handheld ultrasound system processes acquired frames to increase apparent spatial resolution and contrast. By sharpening vessel-wall boundaries and plaque interfaces, it aims to reveal lesions that are small or low in contrast on native handheld images. The approach is positioned for point-of-care use where cart-based systems or confirmatory modalities may not be immediately available.

Investigators tested the technology in a community screening program of adults aged 40 years and older. In 117 participants with 153 carotid plaques, AI-enhanced images showed 94.8% of plaques compared with 87.6% on standard handheld ultrasound images. The enhancement revealed 11 additional plaques, which were mainly small or low-contrast and showed mild vessel narrowing on reference imaging.

Assessment of plaque instability features improved as well. Artificial intelligence–enhanced images correctly flagged 63.2% of plaques that looked unstable, up from 47.4%, while correctly ruling out stable‑appearing plaques about 96% of the time. However, the enhanced images still missed more than one‑third of plaques with unstable‑appearing features, indicating the tool should not be used on its own to rule out risk.

The study was published in Annals of Family Medicine under the title “AI-Enhanced Super-Resolution Handheld Ultrasound for Carotid Plaque Detection in Community Screening.” It is accompanied by an editorial examining what health systems may need to support broader adoption of artificial intelligence-enabled point-of-care ultrasound, authored by researchers from the Department of Family Medicine at The Warren Alpert Medical School of Brown University.