Frontiers in Cardiovascular Medicine (Oct 2021)

Feasibility and Accuracy of a Fully Automated Right Ventricular Quantification Software With Three-Dimensional Echocardiography: Comparison With Cardiac Magnetic Resonance

  • Ashfaq Ahmad,
  • Ashfaq Ahmad,
  • Ashfaq Ahmad,
  • He Li,
  • He Li,
  • He Li,
  • Xiaojing Wan,
  • Yi Zhong,
  • Yi Zhong,
  • Yi Zhong,
  • Yanting Zhang,
  • Yanting Zhang,
  • Yanting Zhang,
  • Juanjuan Liu,
  • Juanjuan Liu,
  • Juanjuan Liu,
  • Ying Gao,
  • Ying Gao,
  • Ying Gao,
  • Mingzhu Qian,
  • Mingzhu Qian,
  • Mingzhu Qian,
  • Yixia Lin,
  • Yixia Lin,
  • Yixia Lin,
  • Luyang Yi,
  • Luyang Yi,
  • Luyang Yi,
  • Li Zhang,
  • Li Zhang,
  • Li Zhang,
  • Li Zhang,
  • Yuman Li,
  • Yuman Li,
  • Yuman Li,
  • Mingxing Xie,
  • Mingxing Xie,
  • Mingxing Xie,
  • Mingxing Xie,
  • Mingxing Xie

DOI
https://doi.org/10.3389/fcvm.2021.732893
Journal volume & issue
Vol. 8

Abstract

Read online

Background: A novel, fully automated right ventricular (RV) software for three-dimensional quantification of RV volumes and function was developed. The direct comparison of the software performance with cardiac magnetic resonance (CMR) was limited. Therefore, the aim of this study was to test the feasibility, accuracy, and reproducibility of a fully automated RV quantification software against CMR imaging as a reference.Methods: A total of 170 patients who underwent both CMR and three-dimensional echocardiography were enrolled. RV end-diastolic volume (RVEDV), RV end-systolic volume (RVESV), and RV ejection fraction (RVEF) were obtained using fully automated three-dimensional RV quantification software and compared with a CMR reference. For inter-technical agreement, Spearman correlation and Bland–Altman analysis were used.Results: The fully automated RV quantification software was feasible in 149 patients. RVEDV and RVESV were underestimated, and RVEF was overestimated compared with CMR values. RV measurements obtained from the manual editing method correlated better with CMR values than that without manual editing (RVEDV, 0.924 vs. 0.794: RVESV, 0.955 vs. 0.854; RVEF, 0.941 vs. 0.781 respectively, all p < 0.0001) with less bias and narrower limit of agreement (LOA). The bias and LOA for RV volumes and EF using the automated software without and with manual editing were greater in patients with severely impaired RV function or low frame rate than those with normal and mild impaired RV function, or high frame rate. The fully automated RV three-dimensional measurements were highly reproducible.Conclusion: The novel fully automated RV software shows good feasibility and reproducibility, and the measurements had a high correlation with CMR values. These findings support the routine application of the novel 3D automated RV software in clinical practice.

Keywords