Artificial Intelligence Speeds Up Stroke Treatment: UTHealth Houston Study
Artificial intelligence (AI) software is proving to be a game-changer in treating strokes caused by large vessel occlusion (LVO), according to a recent study from UTHealth Houston. The research, published in JAMA Neurology, demonstrates that AI can significantly improve treatment times for patients with LVO strokes, ultimately leading to better outcomes.
LVO strokes occur when a major artery in the brain becomes blocked, accounting for a substantial portion of acute ischemic strokes. Quick intervention through endovascular thrombectomy, a minimally invasive procedure that removes blood clots from blocked brain arteries, can greatly enhance patient outcomes. However, the effectiveness of this treatment relies heavily on swift action.
Dr. Youngran Kim, assistant professor of management, policy, and community health at UTHealth Houston School of Public Health, explained the significance of early identification and streamlined care for LVO patients, saying, “The benefit of endovascular thrombectomy on functional recovery is time-sensitive, so early identification of patients with strokes with large vessel occlusions, and process improvements to accelerate in-hospital care, are critical.”
To improve in-hospital endovascular therapy workflows, researchers conducted a clinical trial involving 243 LVO stroke patients at four comprehensive stroke centers in the Greater Houston area. They introduced an AI-enabled solution called Viz LVO, which automatically detects LVO from computed tomography (CT) angiograms and sends real-time alerts to clinicians’ mobile phones. This technology significantly reduced the time to thrombectomy initiation by 11 minutes on average and decreased the time from CT scan initiation to the start of endovascular therapy by nearly 10 minutes.
Dr. Luca Giancardo, an associate professor at McWilliams School of Biomedical Informatics at UTHealth Houston, emphasized the potential of AI in stroke care, noting that “We are just at the beginning of automated machine-learning algorithms to benefit acute stroke care.” Future applications could include using CT angiograms to detect brain damage without advanced imaging or even utilizing retina imaging as a proxy for brain scans.
These findings are especially promising given the time-sensitive nature of stroke treatment. Dr. Sunil A. Sheth, associate professor of neurology and director of the vascular neurology program at McGovern Medical School, explained, “Nearly 2 million brain cells die every minute the blockage remains, so speeding up treatments by 10 to 15 minutes can result in substantial improvements.”
This study is part of a broader effort to enhance stroke outcomes, following a previous study led by Dr. Kim and Dr. Sheth that revealed gender disparities in routing stroke patients to comprehensive stroke centers. With the continued development of AI technology, stroke treatment is on the path to becoming faster, more efficient, and more accessible, ultimately benefiting patients’ lives.