Doctors may soon have a powerful new tool to identify stroke patients who are at high risk of developing deadly brain swelling before obvious symptoms appear.
Researchers at Washington University School of Medicine in St. Louis have developed an artificial intelligence program that predicts this dangerous complication with very high accuracy.
The research behind the technology has led to a startup company called Beacon-Neuro.AI, and the software has now received Breakthrough Device designation from the U.S. Food and Drug Administration (FDA).
Every year, hundreds of thousands of people suffer a stroke. A stroke happens when blood flow to part of the brain is blocked or when a blood vessel bursts, causing brain cells to die.
Even after the initial stroke, patients remain at risk of serious complications, including swelling of the brain known as malignant cerebral edema.
This type of swelling affects up to 70,000 people in the United States each year. If it is not treated quickly, it can be fatal in up to 80% of cases. Doctors can perform surgery to relieve pressure inside the skull, but the treatment works best within the first 48 hours after a stroke.
One of the biggest challenges is knowing which patients will develop severe swelling. Doctors currently rely on brain scans, physical examinations, blood pressure readings, and other clinical information, but these methods are not always accurate enough to identify high-risk patients early.
The new software, called BeaconPredict, uses artificial intelligence to examine the same information that doctors already collect. It analyzes brain scans taken during the first 24 hours after a stroke together with age, blood pressure, neurological examination results, and other medical data to calculate a patient’s individual risk.
In testing involving 598 stroke patients from three medical centers, the AI system correctly identified patients who would later need surgery with an accuracy of about 99%. Current clinical methods correctly identify these patients only about 70% of the time.
The software also greatly reduced false alarms. It correctly distinguished patients who did not require surgery with much better precision than existing approaches, helping doctors avoid unnecessary procedures while still identifying those who urgently need treatment.
The FDA’s Breakthrough Device designation is reserved for promising medical technologies that could improve care for serious or life-threatening diseases. The designation allows the company to work more closely with the FDA and move through the approval process more quickly.
Although the results are encouraging, the software has not yet received full FDA approval for routine clinical use. Larger real-world studies will still be needed to confirm that using the AI tool improves patient outcomes and saves more lives.
Review and analysis: This study addresses a major unmet need in stroke care because early treatment is critical for preventing death from brain swelling.
The very high reported accuracy is impressive, but it comes from a limited validation study. Future prospective clinical trials will be important to confirm that the AI performs equally well in many hospitals and diverse patient populations.
If you care about stroke, please read studies about how to eat to prevent stroke, and diets high in flavonoids could help reduce stroke risk.
For more health information, please see recent studies about how Mediterranean diet could protect your brain health, and wild blueberries can benefit your heart and brain.
Source: Washington University School of Medicine in St. Louis.
