Levels of agreement between cardiac magnetic resonance and conductance catheter measurements of right ventricular volumes after pulmonary artery banding.

Abstract

Pressure-volume analysis is the gold standard for quantifying pump function of the right ventricle (RV); however, volume measurements based on a conductive catheter may be imprecise. The reference method for volume assessment is cardiac magnetic resonance (CMR). To determine the levels of agreement between RV volume measurements obtained by cine CMR, phase-contrast CMR (PC CMR), and a conductance catheter in an animal model. CMR was performed in 20 sheep three months after pulmonary artery banding. Ejection fraction (EF), end-diastolic (EDV), end-systolic (ESV), and stroke volumes (SV) were obtained by cine CMR and conductance catheter. Statistically significant differences between cine CMR and conductance catheter derived volume measurements were found for EDV (P < 0.001), ESV (P < 0.05), and SV (P < 0.05). Bland-Altman analysis showed very poor agreement between the two methods: EDV, bias 36.27 mL, agreement of limits 1.96-70.57 mL; ESV, bias 15.33 mL, agreement of limits -6.89-37.55 mL; and SV, bias 20.69 mL, agreement of limits 8.01-49.10 mL. Good agreement was found for SV between cine CMR and PC CMR (bias -7.0 mL, agreement of limits -24.01-9.98 mL), while SV derived from PC CMR measurements showed poor agreement with conductance catheter (bias 27.76 mL, agreement of limits -3.84-59.26 mL). Poor agreement between the conductance catheter and CMR RV volume measurements was found. PC CMR and cine CMR measurements of SV agreed well.