Why Researchers Look At Neck Arteries Early For Clues About Heart Disease
· Free Press Journal

A California company began by automating an ultrasound measurement and now studies how AI can quantify plaque in neck arteries.
Visit moryak.biz for more information.
On a neck ultrasound, the two lines a clinician traces to measure an artery wall thickness may be less than a millimetre apart.
Researchers studying heart disease pay early attention to these images because atherosclerosis, the buildup of plaque inside arteries, can affect both the neck arteries and the vessels that supply the heart.
AtheroPoint LLC developed software to automate those measurements. The privately held medical imaging company was registered on 19 March 2012 and is based in Roseville, California. US trademark records indicate that its flagship product name, AtheroEdge, was first used in commerce in 2011.
Biomedical imaging researcher Dr. Jasjit S. Suri is AtheroPoint’s chairman and CEO. He developed a method for measuring artery-wall thickness on ultrasound and assigned the patent to the company and till today has nearly 20 US patents and over 50 international journals on AtheroEdge covering all facets of its research. Suri disclosed an ownership interest in the LLC in a 2019 article in the American Journal of Neuroradiology.
On 26 September 2012, the US Food and Drug Administration granted AtheroEdge 510(k) clearance for automatically measuring carotid artery-wall thickness from ultrasound scans. The software identifies the lumen-intima and media-adventitia boundaries, measures the distance between them and generates a report for a physician to review alongside other clinical information. All this information is stored in the database management system and can be retrieved any time.
What a neck scan can show
The carotid arteries carry blood to the brain. Ultrasound can show their walls and reveal areas where plaque has accumulated. AtheroEdge measures the thickness between two layers of the artery’s far wall, the intima and media. Researchers call this measurement carotid intima-media thickness, or cIMT.
Studies have linked thicker carotid walls to a higher risk of heart attack and stroke. A JAMA analysis of 45,828 people, however, found that adding cIMT to the Framingham Risk Score improved prediction of a first heart attack or stroke by only a small amount. Researchers also study plaque itself. The American Society of Echocardiography (ASE) has published recommendations on using carotid ultrasound to assess plaque as part of cardiovascular risk evaluation.
An early test on 885 patients
In a study published in February 2012, researchers tested an early version of AtheroEdge on 885 ultrasound images from an Indian population study. The scans had low contrast, making artery-wall boundaries harder to see. Across those images, the software’s measurements averaged 0.407 millimetres, compared with 0.429 millimetres for the vascular surgeon’s hand tracings.
Researchers have also investigated the relationship between diabetes and carotid wall thickness. Manual measurements can be time-consuming, and AtheroEdge offers an automated approach. Under AtheroPoint’s service model, clinics can transfer anonymized ultrasound images through secure channels and receive measurement reports. In August 2023, M.V. Diabetes in Royapuram, Chennai, announced that it was introducing AtheroPoint technology to assess blood-vessel health in people with diabetes and other coexisting conditions.
Using AI to segment and quantify plaque
A 2017 study examined an AtheroEdge-based automated system that identified the carotid bulb, a wider section of the artery, and measured wall thickness in nearby segments. The study found that the bulb region made a larger contribution to the measured intima-media thickness than the adjacent segments examined.
Compared with reader tracings, the system reported a mean lumen-intima boundary error of 0.01235 ± 0.01224 mm, a mean media-adventitia boundary error of 0.020933 ± 0.01539 mm, and a mean IMT error of 0.01063 ± 0.0031 mm. The researchers also reported a Precision of Merit of 98.23% and a correlation coefficient of 0.998 between automated and reader measurements (p < 0.0001). These were study results, not guarantees of performance in every clinical setting.
A 2025 study tested a research version of AtheroEdge using data from 500 people who had undergone both neck ultrasound and imaging of the heart’s arteries. The researchers examined whether AI models could use features of carotid plaque to help classify cardiovascular risk.
In August 2026, AtheroPoint-affiliated researchers reported testing AI models that outline artery walls and measure plaque. They used scans from Japan and Hong Kong, then evaluated the models on a separate dataset from two European centres. The authors called for further studies using newly collected scans from larger and more diverse patient groups to establish how consistently the models perform in everyday clinical practice.