AI-guided therapy reduces mortality in stroke patients

Ischemic stroke survivors who received care recommendations from an artificial intelligence (AI)-based system received less stroke treatment within three months compared to those who did not receive stroke treatment from an AI tool, according to preliminary new study results. They had fewer recurrent strokes, heart attacks, and vascular deaths. The science was presented today at the American Stroke Association's 2024 International Stroke Conference. The conference, to be held in person in Phoenix from February 7-9, 2024, is the world's first conference for researchers and clinicians dedicated to the science of stroke and brain health.

This study showed that an artificial intelligence-based clinical decision support system for stroke treatment is effective and feasible in Chinese clinical settings and improves patient outcomes. This type of technology helps neurologists by combining the strengths of humans and AI to facilitate information sharing between the two. ”

Zixiao Li, MD, Ph.D., senior study author, attending physician, professor, and vice director of the Department of Neurology, Beijing Tiantan Hospital, Capital Medical University, Beijing, China

According to Li, ischemic stroke is the leading cause of death in China. Timely assessment and decisions regarding stroke diagnosis and treatment are critical to restoring blood flow and minimizing the amount of damage to the brain. According to the latest data from the American Heart Association's 2024 Heart Disease and Stroke Statistics: U.S. and Global Data Report, there will be 7.44 million stroke deaths worldwide in 2021, of which approximately half will be due to ischemic stroke. It was a stroke. In the United States, 87% of strokes are ischemic strokes. Ischemic stroke occurs when blood vessels in the brain become narrowed or clogged with plaque, cutting off blood flow to the brain.

In a clinical trial called GOLDEN BRIDGE II, 77 hospitals in China will be randomly selected to provide diagnosis and treatment for ischemic stroke patients based on recommendations from AI technology systems or assessments and recommendations by the hospital's stroke treatment team. Assigned. The AI ​​system integrated AI-interpreted brain imaging scans of participants with established clinical knowledge about stroke diagnosis, stroke classification, guideline-recommended treatments, and strategies to prevent secondary strokes.

The researchers then measured the number of vascular events in the study's more than 20,000 participants. Death due to ischemic stroke, hemorrhagic stroke, heart attack, or vascular event – included all study participants during the 3-month follow-up period after the first ischemic stroke.

The analysis revealed the following:

• Using an AI-based clinical decision support system, the likelihood of a new vascular event was reduced by 25.6% in the three months following the initial stroke, and patients were more likely to be treated with guideline-directed drug therapy. and the quality of stroke care has improved.
• At 3 months, participants treated in the hospital with AI support had fewer total vascular events than those receiving standard post-stroke assessment and treatment ( 2.9% vs. 3.9%).
• There was no statistically significant difference in disability levels between patients in the AI-guided or standard treatment groups at 3 months, as assessed using the modified Rankin Scale score. A tool used to determine the level of disability in people who have experienced a stroke.

“The reduction in new vascular events is an important finding as it shows that AI has the potential to bring real changes to stroke treatment and benefit this large number of stroke survivors.” said Lee, who is also a clinical university professor. Research Center for Neurological Diseases; Artificial Intelligence Research Unit in Cerebrovascular Diseases, Chinese Academy of Medical Sciences. And the China Brain Research Institute is all located in Beijing.

“In the future, we expect to see more AI applications validated through clinical research, and we hope that clinical decision support systems will help improve stroke care, including reperfusion therapy, long-term secondary prevention, and rehabilitation. We hope that it can be expanded to include many aspects, and at the same time, we hope that the application of AI can be expanded to apply to other health conditions as well.”

Research details and background:

• The study was conducted in 77 hospitals across China from January 2021 to June 2023.
• A total of 21,603 adults hospitalized with acute ischemic stroke participated in the study. Approximately one-third were women, and the average age of participants was 67 years.
• The number of participants who received each treatment approach was similar: 11,054 received AI evaluation and treatment, and 10,549 received usual care and treatment based on the hospital's neurologist's assessment and recommendations.
• Almost all participants (21,579) completed the 3-month follow-up and were therefore included in the final data assessment.
• Neurologists at the hospital testing the AI ​​technology completed training on the stroke AI clinical decision support system before the study began.
• Stroke Quality of Care is an internationally recognized composite of evidence-based performance measures of quality of acute ischemic stroke care, including eight measures at the beginning of admission and five measures at discharge. It was measured by a score, Lee said.

Limitations of the study include that hospitals were randomly assigned to the AI-based strategy or standard of care rather than individual patients. Additionally, differences in patterns of care and outcomes between hospitals and subsequent outpatient treatment may have influenced the results. Furthermore, whether improvements in care and outcomes can be sustained requires further evaluation, and the capabilities of stroke AI-based clinical decision support systems need to be constantly updated with revised evidence-based clinical guidelines. There may be. More widespread and sustainable clinical application models of stroke AI-based clinical decision support systems in other health conditions and other countries should be considered.