Automated echocardiographic left ventricular dimension assessment in dogs using artificial intelligence: Development and validation.

Collapse authors list.
Stowell, Catherine C ORCID: 0000-0003-3005-8636, Kallassy, Valeria, Lane, Beth ORCID: 0009-0000-4307-9597, Abbott, Jonathan ORCID: 0000-0001-6981-8968, Borgeat, Kieran ORCID: 0000-0001-9199-4941, Connolly, David, Domenech, Oriol, Dukes-McEwan, Joanna ORCID: 0000-0002-0326-8251, Ferasin, Luca ORCID: 0000-0002-6898-7900, Del Palacio, Josefa Fernández et al (show 9 more authors) , Linney, Chris ORCID: 0000-0002-9126-8605, Matos, Jose Novo ORCID: 0000-0001-8128-8111, Spalla, Ilaria ORCID: 0000-0003-0039-1438, Summerfield, Nuala, Vezzosi, Tommaso ORCID: 0000-0001-8301-6582, Howard, James P, Shun-Shin, Matthew J ORCID: 0000-0002-1179-0867, Francis, Darrel P ORCID: 0000-0002-3410-0814 and Fuentes, Virginia Luis ORCID: 0000-0001-8076-3806 (2024) Automated echocardiographic left ventricular dimension assessment in dogs using artificial intelligence: Development and validation. Journal of veterinary internal medicine, 38 (2). pp. 922-930. ISSN 0891-6640, 1939-1676

PDF Automated echocardiographic left ventricular dimension assessment in dogs AI Stowell JVIM 2024.pdf - Open Access published version Download (4MB) | Preview

Abstract

<h4>Background</h4>Artificial intelligence (AI) could improve accuracy and reproducibility of echocardiographic measurements in dogs.<h4>Hypothesis</h4>A neural network can be trained to measure echocardiographic left ventricular (LV) linear dimensions in dogs.<h4>Animals</h4>Training dataset: 1398 frames from 461 canine echocardiograms from a single specialist center.<h4>Validation</h4>50 additional echocardiograms from the same center.<h4>Methods</h4>Training dataset: a right parasternal 4-chamber long axis frame from each study, labeled by 1 of 18 echocardiographers, marking anterior and posterior points of the septum and free wall.<h4>Validation dataset</h4>End-diastolic and end-systolic frames from 50 studies, annotated twice (blindly) by 13 experts, producing 26 measurements of each site from each frame. The neural network also made these measurements. We quantified its accuracy as the deviation from the expert consensus, using the individual-expert deviation from consensus as context for acceptable variation. The deviation of the AI measurement away from the expert consensus was assessed on each individual frame and compared with the root-mean-square-variation of the individual expert opinions away from that consensus.<h4>Results</h4>For the septum in end-diastole, individual expert opinions deviated by 0.12 cm from the consensus, while the AI deviated by 0.11 cm (P = .61). For LVD, the corresponding values were 0.20 cm for experts and 0.13 cm for AI (P = .65); for the free wall, experts 0.20 cm, AI 0.13 cm (P < .01). In end-systole, there were no differences between individual expert and AI performances.<h4>Conclusions and clinical importance</h4>An artificial intelligence network can be trained to adequately measure linear LV dimensions, with performance indistinguishable from that of experts.

Item Type:	Article
Uncontrolled Keywords:	Heart Ventricles, Animals, Dogs, Echocardiography, Reproducibility of Results, Diastole, Artificial Intelligence
Depositing User:	Symplectic Admin
Date Deposited:	01 Jul 2024 10:54
Last Modified:	17 Jan 2026 11:08
DOI:	10.1111/jvim.17012
Related Websites:	Author Publisher
URI:	https://livrepository.liverpool.ac.uk/id/eprint/3182483
Disclaimer:	The University of Liverpool is not responsible for content contained on other websites from links within repository metadata. Please contact us if you notice anything that appears incorrect or inappropriate.