To assess whether artificial intelligence can enable efficient diagnosis of dangerous blood clots, researchers at the BerbeeWalsh Department of Emergency Medicine at the University of Wisconsin School of Medicine and Public Health are helping to lead a national clinical study.
The study is designed to evaluate AI-powered ultrasound technology designed to detect deep vein thrombosis (DVT).
DVT occurs when a blood clot forms in a vein, typically in the thigh, lower leg or pelvis. Severe cases can lead to life-threatening complications like pulmonary embolism. While nearly 900,000 people are affected by a blood clot every year, according to the U.S. Centers for Disease Control and Prevention, the dangerous condition is often underdiagnosed or untreated until serious health complications occur.
Traditional diagnostic workflows for DVT are often time-consuming, resource-intensive, and heavily reliant on the availability of specialists such as radiologists, which can strain health care systems and delay care. DVT can also be difficult to diagnose quickly, especially in emergency or underserved health care settings that lack ultrasound experts and equipment.
The technology being tested in the multi-site, double-blind study uses real-time AI guidance with handheld ultrasound and a software data collection and communication tool designed to collect ultrasound data. The system, which is already in clinical use in Europe, is designed to enable health care team members who do not have specialized training in DVT assessment to perform rapid bedside scans of patient veins. Results are transmitted immediately to physicians for remote review, accelerating diagnosis and potentially reducing treatment delays.
“Quick and accurate detection of DVT can greatly improve patient outcomes,” said Dr. Hani Kuttab, assistant professor of emergency medicine and site principal investigator for the study. “We need ways to safely streamline diagnosis and help care teams make critical, time-sensitive treatment decisions.”
Assistant professors and UW Health emergency physicians Dr. Matthew VandeHei and Dr. Collin Michels are collaborating on the study. UW Health is the first Midwestern and fourth U.S.-based health care system to join the study, which tests the safety and efficacy of the ThinkSono Guidance System developed by ThinkSono, a UK-based AI company.
The research team will conduct traditional ultrasound scans and ThinkSono Guidance System scans on UW Health patients with and without DVT and compare performance of both methods.
“This technology has the potential to shift how we approach vascular imaging in acute care,” said VandeHei. “Successful validation could expand access to rapid diagnostic imaging, especially to rural hospitals and other communities with limited resources.”
This research is being supported by funds from ThinkSono GmbH.