AI Early Detection System Prevents Strokes by Identifying Atrial Fibrillation

Researchers have harnessed the power of artificial intelligence (AI) to identify irregular cardiac rhythms in individuals who have not yet exhibited symptoms. This AI system, which detected hidden signals in routine medical diagnostic tests, offers a promising avenue for doctors to proactively prevent strokes and other cardiovascular problems in individuals with atrial fibrillation, the most common type of heart rhythm disorder. While previous algorithms were predominantly tested on a narrow demographic, this new AI algorithm demonstrates effectiveness across various situations and patient populations, including veterans and underserved communities in the United States. The groundbreaking findings have been published in JAMA Cardiology, a reputable peer-reviewed journal.

Dr. David Ouyang, a cardiologist at Cedars-Sinai’s Smidt Heart Institute and a researcher in the Division of Artificial Intelligence in Medicine, is the senior author of this study. He emphasizes the significance of this research in identifying concealed heart conditions and promoting the development of equitable and universally applicable algorithms for all patients. According to experts, approximately one in three people with atrial fibrillation remains undiagnosed. In this condition, the heart’s electrical signals, responsible for orchestrating the flow of blood from the upper chambers to the lower chambers, become chaotic. This can result in blood pooling in the upper chambers, leading to the formation of blood clots that may travel to the brain and trigger an ischemic stroke.

The core of this groundbreaking discovery lies in an AI algorithm that was trained to recognize patterns within electrocardiogram (ECG) readings. ECG is a diagnostic test that monitors the heart’s electrical activity, typically involving the placement of electrodes on a patient’s body to capture these signals. By analyzing nearly one million ECGs, the AI model was not only able to accurately predict the onset of atrial fibrillation within 31 days but also demonstrated its effectiveness when applied to medical records of patients at Cedars-Sinai.

Dr. Sumeet Chugh, the director of the Division of Artificial Intelligence in Medicine and the medical director of the Heart Rhythm Center at the Department of Cardiology, highlights the study’s geographic and ethnic diversity in its sample of veterans. This diversity underscores the potential of this algorithm to benefit a broad spectrum of the U.S. population. The research is a testament to the innovative ways in which AI is being harnessed at the Smidt Heart Institute and the Division of Artificial Intelligence in Medicine to proactively manage complex and challenging cardiac conditions.