Diagnostic accuracy of point-of-care ultrasound with artificial intelligence-assisted assessment of left ventricular ejection fraction

Focused cardiac ultrasound (FoCUS) is becoming standard practice in a wide spectrum of clinical settings. There is limited data evaluating the real-world use of FoCUS with artificial intelligence (AI). Our objective was to determine the accuracy of FoCUS AI-assisted left ventricular ejection fraction (LVEF) assessment and compare its accuracy between novice and experienced users. In this prospective, multicentre study, participants requiring a transthoracic echocardiogram (TTE) were recruited to have a FoCUS performed by a novice or experienced user. The AI-assisted device calculated LVEF at the bedside, which was subsequently compared to TTE. A total of 449 participants were enrolled with 424 studies included in the final analysis. The overall intraclass correlation coefficient (ICC) was 0.904; ICC was 0.921 in the novice (n = 208) and 0.845 in the experienced (n = 216) cohorts. There was a significant bias of 0.73% towards TTE (p = 0.005) with a level of agreement of 11.12%. Categorical grading of LVEF severity had excellent agreement with TTE (weighted kappa = 0.83). The area under the curve (AUC) was 0.98 for identifying abnormal LVEF (<50%) with a sensitivity of ~92%, negative predictive value (NPV) of 0.97 and positive predictive value (PPV) of 0.83. For severe dysfunction (<30%), AUC was 0.99 with sensitivity 78.1%, specificity 98.0%, NPV 0.98 and PPV 0.76. FoCUS AI-assisted LVEF assessments provide highly reproducible estimations compared to TTE, consistent among novice and experienced users, suggesting applicability in diverse clinical settings.

Pouya Motazedian, Jeffrey A. Marbach, Graeme Prosperi-Porta, Simon Parlow, Pietro Di Santo, Omar Abdel-Razek, Richard Jung, William B. Bradford, Miranda Tsang, Michael Hyno, Stefano Pacifici, Sharanya Mohanty, F. Daniel Ramirez, Gordon S. Huggins, Trevor Simard, Stephanie Hon, Benjamin Hibbert