A deep-learning algorithm that interprets electrocardiogram (ECG) data can predict mortality risk after heart surgery, a new NHLBI-funded study finds.
Before undergoing surgeries, doctors often conduct a preoperative risk assessment to predict how a patient will fare after surgery, yet these assessments don’t always forecast correctly. The study authors supposed that using a deep-learning algorithm of ECGs might be able to identify hidden risk markers that can help categorize those patients who might be at a higher risk of death after surgery. They used data from 45,969 patients who had undergone an ECG 30 days before their surgeries. The algorithm, validated in three independent health-care systems, outperformed the established Revised Cardiac Risk Index score, and worked equally well for heart surgeries, non-heart surgeries, and catheterization laboratory procedures.
According to the study authors, “This algorithm can provide additional information to clinicians making the decision to perform medical procedures and stratify the risk of future complications.”
The study appeared in The Lancet Digital Health.