AI Matches Human Accuracy in Echocardiograms, Cuts Cardiologist Time by Half

New independent, preliminary late-breaking science presented at the American Heart Association’s Scientific Sessions 2026 in Chicago suggests that artificial intelligence (AI) can interpret heart ultrasound exams, known as echocardiograms, with accuracy comparable to human experts while significantly reducing cardiologists’ workload. The randomized trial, the first to test an AI tool for crafting a full echocardiogram report including interpretations and measurements, involved 1,200 echocardiograms performed across 21 medical centers within Kaiser Permanente Northern California. An AI tool, sonographers, and cardiologists each prepared preliminary reports for the same scans. These reports were then randomly assigned to cardiologists for review without knowing the source. Cardiologists made substantial changes to 50.6% of AI-generated preliminary interpretations, compared with 49.7% of those prepared by sonographers and cardiologists, meeting the study’s criteria for non-inferiority. Notably, AI assistance cut the time cardiologists spent interpreting each scan by nearly half, from 8.5 to 4.6 minutes, and AI-generated reports were more consistent with historical clinical reports. Cardiologists could not identify whether the preliminary interpretation was prepared by the AI tool, a sonographer, or another cardiologist.

Co-senior author Keane K. Lee, M.D., M.S., a cardiologist at Kaiser Permanente Santa Clara Homestead Medical Center, said, “We designed this study to capture even small differences between the preliminary reports and the cardiologists’ final interpretations, allowing us to take a careful look at how well AI performed in a rigorous scientific manner.” He added that the time savings surprised the research team and could lead to faster turnaround of results and more access to care. However, Lee emphasized that “AI cannot, however, replace the human team. The cardiologist remains ultimately responsible for interpreting the echocardiogram and certifying the accuracy of the final report.” The study’s limitations include that it did not test whether AI-assisted interpretation improves long-term patient health and that cardiologists knew they were part of the study. More research is needed in real-world clinical settings with larger volumes. The findings come amid projections of a national shortfall of about 7,000 cardiologists by 2036, according to a July 2026 study published in the Journal of the American Heart Association.