With AI, researchers discover new way to detect sudden cardiac death risk (news.berkeley.edu)

Researchers at UC Berkeley have developed an AI system capable of significantly enhancing the detection of sudden cardiac death risk by analyzing electrocardiograms (EKGs). Training on over 440,000 EKGs linked with death data, the model identifies specific waveform patterns that indicate a higher risk of cardiac arrest. Remarkably, it outperformed traditional clinical tests, flagging a 7% annual risk of sudden cardiac death, as opposed to the 4.6% identified by existing methods. This advancement could lead to better identification of patients who may need life-saving interventions, such as internal defibrillators. This breakthrough is particularly significant given that sudden cardiac arrest claims over 300,000 lives annually in the U.S., often without warning, affecting both high-risk individuals and seemingly healthy young athletes. The AI model's ability to tap into widely available EKG data offers a transformative approach to patient monitoring and intervention, potentially curbing the number of preventable deaths. Additionally, the study could catalyze further insights into the physiological mechanisms behind sudden cardiac events, marking a pivotal shift in cardiac health research and patient care strategies.