Quantitative assessment of right ventricular size and function with multiple parameters from artificial intelligence‐based three‐dimensional echocardiography: A comparative study with cardiac magnetic resonance. Issue 2 (15th January 2022)

• Record Type: Journal Article
• Authors: Zhu, Ying; Bao, Yuwei; Zheng, Kangchao; Zhou, Wei; Zhang, Jun; Sun, Ruiying; Deng, Youbin; Xia, Liming; Liu, Yani
• Is Part Of: Echocardiography. Volume 39:Issue 2(2022)
• Page Start: 223
• Page End: 232
• Publication Date: 2022-01-15
• Subjects: artificial intelligence -- multiple parameters -- right ventricle -- three‐dimensional echocardiography; Echocardiography -- Periodicals; Echocardiography -- Periodicals; 616.1207543
• DOI: 10.1111/echo.15292
• Languages: English
• ISSNs: 0742-2822

Abstract

Abstract: Aims: This study aimed to explore the validation and the diagnostic value of multiple right ventricle (RV) volumes and functional parameters derived from a novel artificial intelligence (AI)‐based three‐dimensional echocardiography (3DE) algorithm compared to cardiac magnetic resonance (CMR). Methods and results: A total of 51 patients with a broad spectrum of clinical diagnoses were finally included in this study. AI‐based RV 3DE was performed in a single‐beat HeartModel mode within 24 hours after CMR. In the entire population, RV volumes and right ventricular ejection fraction (RVEF) measured by AI‐based 3DE showed statistically significant correlations with the corresponding CMR analysis ( p < 0.05 for all). However, the Bland‐Altman plots indicated that these parameters were slightly underestimated by AI‐based 3DE. Based on CMR derived RVEF < 45% as RV dysfunction, end‐systolic volume (ESV), end‐systolic volume index (ESVi), stroke volume (SV), and RVEF showed great diagnostic performance in identifying RV dysfunction, as well as some non‐volumetric parameters, including tricuspid annular systolic excursion (TAPSE), fractional area change (FAC), and free‐wall longitudinal strains (LS) ( p < 0.05 for all). The cutoff value was 43% for RVEF with a sensitivity of 94% and specificity of 67%. Conclusion: AI‐based 3DE could provide rapid and accurate quantitation of the RV volumes and function with multiple parameters. Both volumetric and non‐volumetric measurements derived from AI‐based 3DE contributed to the identification of the RV dysfunction.