AI Outperforms Sonographers in Assessing Cardiac Function: Smidt Heart Institute's Breakthrough Study
New research from the Smidt Heart Institute has shown that artificial intelligence (AI) is more effective at assessing and diagnosing cardiac function through echocardiogram analysis than human sonographers. The first-of-its-kind blinded, randomized non-inferiority clinical trial focused on the initial assessment of left ventricular ejection fraction (LVEF) by AI versus sonographers. LVEF is a critical measurement for diagnosing cardiac function, guiding clinical decisions regarding patient eligibility for various therapies and interventions, including surgeries.
Accurate LVEF assessment is crucial for disease diagnosis, risk stratification, and treatment response assessment. However, conventional approaches to measuring LVEF are subject to heterogeneity and variance due to their reliance on manual and subjective human tracings. Clinical practice guidelines recommend repeated measurements of LVEF over multiple cardiac cycles to improve precision and account for arrhythmic or hemodynamic sources of variation. Unfortunately, repeated human measurements are rarely done in practice due to logistical constraints, leading to suboptimal detection of subtle changes in LVEF.
The study included 3,769 echocardiographic studies, with 274 excluded due to poor image quality. The primary endpoint was the change in LVEF between initial AI or sonographer assessment and final cardiologist assessment, with substantial change being defined as more than a 5% change. The proportion of studies with substantial change was 16.8% in the AI group and 27.2% in the sonographer group, demonstrating AI's superiority (P < 0.001 for non-inferiority, P < 0.001 for superiority).
The AI technology is expected to be beneficial when deployed across the Cedars-Sinai clinical system and nationwide. The study's success sets a precedent for how AI algorithms can be tested and implemented for improved patient care. The AI-guided workflow saved time for both sonographers and cardiologists, who could not distinguish between initial assessments made by AI or sonographers.
"The results have immediate implications for patients undergoing cardiac function imaging as well as broader implications for the field of cardiac imaging," said cardiologist David Ouyang, MD, principal investigator of the clinical trial and senior author of the study. "This trial offers rigorous evidence that utilizing AI in this novel way can improve the quality and effectiveness of echocardiogram imaging for many patients."
In recent years, numerous AI algorithms have been developed with the goal of automating LVEF assessment in real-world patient care settings. Although these algorithms have demonstrated improved precision on limited retrospective datasets, no cardiovascular AI technologies have been validated in blinded, randomized clinical trials until now.
"This successful clinical trial sets a superb precedent for how novel clinical AI algorithms can be discovered and tested within health systems, increasing the likelihood of seamless deployment for improved patient care," said Sumeet Chugh, MD, director of the Division of Artificial Intelligence in Medicine and the Pauline and Harold Price Chair in Cardiac Electrophysiology Research.
The integration of AI into the echocardiogram interpretation workflow can support healthcare professionals in making more accurate and timely diagnoses, ultimately leading to better patient outcomes. The use of AI in this area of healthcare has the potential to increase efficiency and accuracy in diagnosing cardiac function, while reducing the workload of sonographers and cardiologists, and maintaining high-quality care for patients.