Artificial Intelligence Empowers Novice Users to Acquire Diagnostic-Quality Echocardiography

Abstract

Background: Cardiac ultrasound exams provide real-time data to guide clinical decisions but require highly trained sonographers. Artificial intelligence (AI) that uses deep learning algorithms to guide novices in the acquisition of diagnostic echocardiographic studies may broaden access and improve care.

Objectives: The objective of this trial was to evaluate whether nurses without previous ultrasound experience (novices) could obtain diagnostic-quality acquisitions of 10 echocardiographic views using AI-based software.

Methods: This noninferiority study was prospective, international, nonrandomized, and conducted at 2 medical centers, in the United States and France, from November 2023 to August 2024. Two limited cardiac exams were performed on adult patients scheduled for a clinically indicated echocardiogram; one was conducted by a novice using AI guidance and one by an expert (experienced sonographer or cardiologist) without it. Primary endpoints were evaluated by 5 experienced cardiologists to assess whether the novice exam was of sufficient quality to visually analyze the left ventricular size and function, the right ventricle size, and the presence of nontrivial pericardial effusion. Secondary endpoints included 8 additional cardiac parameters.

Results: A total of 240 patients (mean age 62.6 years; 117 women (48.8%); mean body mass index 26.6 kg/m²) completed the study. One hundred percent of the exams performed by novices with the studied software were of sufficient quality to assess the primary endpoints. Cardiac parameters assessed in exams conducted by novices and experts were strongly correlated.

Conclusions: AI-based software provides a safe means for novices to perform diagnostic-quality cardiac ultrasounds after a short training period.

Keywords: artificial intelligence; cardiac ultrasound; deep learning; echocardiography.

Graphical abstract

Figure 1

Artificial Intelligence-Based Software User Interface and Setup for Echocardiographic Acquisition An example of the studied application screen (left) showing the actual view in real-time with the Live Guidance items (middle of the screen), the recording features (auto record, best-effort record, manual record), the 10 echocardiographic views (left of the screen), the view example and the probe positioning (right of the screen), and the setup for the apical 4-chamber view acquisition. A2C = apical 2-chamber; A3C = apical 3-chamber; A4C = apical 4-chamber; A5C = apical 5-chamber; PLAX = parasternal long-axis; PSAX-AV = parasternal short-axis at the aortic valve; PSAX-MV = parasternal short-axis at the mitral valve; PSAX-PM = parasternal short-axis at the papillary muscle; SC-4C = subcostal 4-chamber; SC-IVC = subcostal inferior vena cava.

Central Illustration

Artificial Intelligence-Based Software Empowers Novice Users to Acquire Heart Exams The AI-powered software provides real-time guidance to assist novices in acquiring diagnostic-quality echocardiographic views, with automated clip recording. The study included 240 patients who underwent 2 echocardiography exams: one by an expert and one by a novice using AI guidance. All exams were independently assessed by 5 cardiologists blinded to the operator type. Primary endpoints included the assessment of left ventricular size and function, right ventricular size, and nontrivial pericardial effusion. Results demonstrated no significant difference in assessment between novice and expert exams. Diagnostic quality was achieved by the novices on 99.7% of the clips recorded with the “auto record” feature, 93.9% of the clips recorded with the “best-effort record” feature, and 40.5% of the clips recorded with the “manual record.” This clinical trial demonstrated that novices can perform heart exams, allowing the evaluation of the most important cardiac parameters. AI = artificial intelligence.

Figure 2

Patient Flowchart The flowchart includes nurse participation, patient enrollment and participation (reasons for withdrawal are specified), and panel participation.