No abstract

No abstract

Funding Acknowledgements Type of funding sources: Public Institution(s). Main funding source(s): Norwegian University of Science and Technology. Background Left ventricular (LV) volumes and ejection fraction (EF) are the most used and studied parameters in echocardiography, but they are hampered with the tedious nature and limited reproducibility of manual measurements. We developed an artificial intelligence (AI) decision-support software that automatically calculates LV volumes and EF by identifying the view, timing, and endocardial border from echo images. Purpose Our aim was to investigate the impact of using the AI decision-support software for automatic measurements of LV volumes and EF compared to standard care in real-time and large databases. Methods We compared biplane three-cycle averaged AI measurements of LV volumes and EF to manual references in multiple samples: 1) Real-time AI-support during scanning of 50 consecutive patients compared to standard care (test-retest) by two of three experts, 2) comparison of test-retest variability in inter- and intra-observer scenarios in 40 subjects with test-retest datasets by separate observers and measurements by AI and four experienced observers, 3) an internal database of 1881 subjects with manual references by Simpson’s method, and 4) an external population of 849 subjects with manual references by semi-automatic tracking-based software. The influence on time consumption was tested in Sample 1. The influence on test-retest variability was tested in Sample 1 and 2, while the other samples were used to test the agreement with echocardiographic references. Results Used in real-time the AI measurements reduced total time-consumption by median (95% CI) 5.3 (4.8–6.5) minutes per patient (p <0.001). The test-retest variability for AI-measurements were superior to inter-observer scenarios (p <0.05) and non-inferior to intra-observer scenarios (p <0.025, one-sided, delta=46% increased variance). The agreement with echocardiographic references was good (Figure). The biases (AI minus reference) ranged −11.6 to 6.5 mL for end-diastolic LV volume and −5.5 to 0.3 %-points for EF. Conclusion Fully automatic AI measurements of LV volumes and EF reduced time-consumption with 5 minutes, reduced test-retest variability compared to inter-observer analyses, and showed acceptable agreement with manual measurements in a wide range of scenarios. Thus, implementation of real-time fully automatic measurements of LV volumes and EF during scanning may improve the everyday workflow and quality in the everyday clinic.

No abstract

<p>Measurements of cardiac function such as left ventricular ejection fraction and myocardial strain are typically based on 2D ultrasound imaging. The reliability of these measurements strongly depends on the correct pose of the transducer such that the 2D imaging plane properly aligns with the heart for standard measurement views, and is thus dependent on the operator's skills. In this work, we propose a deep learning-based tool that provides real-time feedback on how to move the transducer to obtain the required views. We believe this method can aid less-experienced users to acquire recordings of better quality for measurements and diagnosis, and to improve standardization of images for more experienced users. Training data was generated by slicing 3D ultrasound volumes, which permits to simulate movements of a transducer and 2D imaging plane. Each slice was labelled with an anatomical reference obtained through a semi-automatic annotation procedure, which allowed us to generate substantial amounts of training data. The method was validated and tested on 2D images from several datasets representative of a prospective clinical setting. We proposed a new metric to score the correctness of the transducer movement feedback according to several given criteria, and achieved a success rate of 75% for all models and 95% for the rotational movement. A real-time prototype application was developed based on data streaming from a clinical ultrasound system, which demonstrated the ability of the method to robustly predict the apical rotation and tilt of the 2D ultrasound image plane relative to the heart. </p>