How and where to measure? Comparing contemporary noninvasive assessment of pulmonary artery size in children with congenital heart disease

  • Dr Marina Hughes, Great Ormond St Hospital for Children NHS Trust, United Kingdom
  • Dr F Bailliard, University of North Carolina Children's Hospital, United States
  • Dr Andrew Taylor, Great Ormond St Hospital for Children NHS Trust, United Kingdom
  • Dr J Marek, Great Ormond St Hospital for Children NHS Trust, United Kingdom

    • Objectives
    Knowledge of the size and nature of the proximal PAs is crucial for planning surgical management of infants with CHD. Because contemporary clinical data on their relative merits is scant, we compared PA measurements using gadolinium-enhanced 3D cardiac magnetic resonance angiography (CMRA) and 2Dechocardiography.
    Methods
    A prospective, observational study of 56 consecutive children undergoing CMR and 2Decho under GA, between 2006-2008.
    Two independent observers measured optimised 2Decho images of PAs and 2 independent observers measured cross-sectional area, shortest and orthogonal PA diameters with Siemens 3D software from isotropic CMRA images.
    Results
    The full spectrum of pre-surgical CHD was represented. 29/56 (52%) had functionally univentricular hearts. The median age was 1.1 years (range 0.2-8.5) and median weight 9.4kg (range 4.0-24.1).
    CMRA assessed the 3D form of all vessels. 2Decho measurements were lacking from the distal LPA in 39 patients.
    Inter-technique assessment showed significant differences between narrowest CMRA measurement and the 2Decho measurement of the proximal PAs, (mean difference 1.0mm, paired 2-tailed P < 0.01) but when shortest and orthogonal CMRA diameters were averaged, measurements compared well to single-plane 2Decho.
    There was reasonable correlation of 2Decho (calculated) Nakata index and measured CMRA Nakata index (r=0.54, p<0.01).
    Interobserver variability was significant (mean difference>0.3mm, P<0.01) for 2Decho measurements of the proximal LPA and distal RPA, but was insignificant for all paired CMRA measurements.
    Conclusions
    2Decho is limited and may fail to visualise the narrowest PA dimension, but when acoustic windows permit there is good correlation of 2Decho and CMRA measurements of PAs.