Recent research indicates that artificial intelligence could be twice as accurate as current methods in determining the timing of strokes, potentially aiding doctors in making urgent care decisions. The study suggests that utilizing technology to analyze brain scans of stroke patients could allow for up to 50% more patients to receive appropriate treatment.
The software development team, comprised of researchers from Imperial College London, the Technical University of Munich, and the University of Edinburgh, claims their model is significantly more precise than existing visual assessment methods used by physicians to evaluate brain scans.
Trained on 800 brain scans with known stroke timings and tested on 2,000 patients, the study found that the model's accuracy stems from its ability to incorporate additional features from the scans, such as texture.
Patients arriving at hospitals with suspected strokes undergo CT scans to determine the cause and severity of their condition, ideally within one hour. Physicians then review the results to make treatment decisions, where timing is crucial.
Thrombolysis, a process that breaks down blood clots with medication, can be performed within 24 hours of symptom onset, while thrombectomy—surgically removing a clot to restore blood flow—is most effective within six hours after symptoms appear, although it may be performed in rare cases up to 24 hours later.
Dr. Paul Bentley, from the Department of Brain Sciences at Imperial College and a neurologist at Imperial College Healthcare NHS Trust, led the study published in the medical journal NPJ Digital Medicine. He emphasized the importance of knowing the onset time of a stroke and the potential for reversing its symptoms to guide treatment decisions.