Rapid Three Dimensional Echocardiography Assessment of Right Ventricular Volume and Function in Congenital Heart Disease: Comparison with Cardiac Magnetic Resonance Imaging

  • Dr Nee Scze Khoo, Stollery Children's Hospital, Canada
  • Dr Alistair Young, Department of Anatomy with Radiology, University of Auckland, New Zealand
  • Dr Christopher Occleshaw, Department of Radiology, Auckland Hospital, New Zealand
  • Dr Brett Cowan, Centre for Advanced MRI, University of Auckland, New Zealand
  • Ms Irene Zeng, Department of Cardiology, Auckland Hospital, New Zealand
  • Dr Thomas Gentles, Green Lane Paediatric & Congenital Cardiac Services, Starship Children's Hospital, New Zealand

    • Background: Quantitative assessment of right ventricular (RV) volumes with 2-dimensional echocardiography is unreliable. Three-dimensional echocardiography (3DE) is promising but limited by time consuming analysis. Utility of 3DE for RV volume estimation is uncertain in congenital heart disease where RV dilatation and dysfunction is common. This study compares 3DE measurements of RV volumes and ejection fraction (EF) obtained by manual tracing methods; disc summation (DS) and 3DE apical rotation (AR), with automated border detection (ABD) algorithm and with manual adjustment (MABD) in patients with congenital heart disease. 3DE measurements were compared with cardiac magnetic resonance imaging (MR).
    Method and Results: Twenty-eight (52%) of 54 consecutively recruited patients (median age, 17 years) with congenital heart disease had adequate 3DE imaging. All MRI images were adequate. Analysis were performed by separate observers for the different modalities. There was a wide range of RV size (MR EDVi 143±43ml/m2) and function (MR EF 48±10%). RT3DE systematically underestimated EDV (p<0.001) with greater underestimation in dilated RVs (r=0.60, p<0.001). There was no significant difference for EF measurement in all methods except ABD (-2.6%±6, p=0.03). No significant difference in inter and intra-observer reproducibility for EF between 3DE and MR. Mean analysis time for ABD was 5 compared with 13-19 minutes for MABD, DS and AR (p<0.0001).
    Conclusion: Approximately half of patients with congenital heart disease had adequate 3DE images. 3DE accurately estimates EF but underestimates volumes, particularly when the RV is dilated. ABD minimally underestimates EF but offers significant analysis time reduction.

    Variables EDVi [%difference], EF [absolute difference] mean, SD Difference (3DE-MR) vs. MR, p value 3DE vs. MR, r value
    DS EDVi -19.3, 14 <0.001 0.91
    AR EDVi -19.1, 13 <0.001 0.92
    ABD EDVi -21.6, 13 <0.001 0.89
    MABD EDVi -22.1, 12 <0.001 0.91
    DS EF -0.5%, 6.9 0.72 0.78
    AR EF -0.7%, 6.4 0.52 0.79
    ABD EF -2.6%, 6.0 0.03 0.83
    MABD EF +1.4%, 6.9 0.31 0.78