Which cardiovascular magnetic resonance planes and sequences provide accurate measurements of branch pulmonary artery size in children with right ventricular outflow tract obstruction?

Children with right ventricular outflow tract obstructive (RVOTO) lesions require precise quantification of pulmonary artery (PA) size for proper management of branch PA stenosis. We aimed to determine which cardiovascular magnetic resonance (CMR) sequences and planes correlated best with cardiac catheterization and surgical measurements of branch PA size. Fifty-five children with RVOTO lesions and biventricular circulation underwent CMR prior to; either cardiac catheterization (n = 30) or surgery (n = 25) within a 6 month time frame. CMR sequences included axial black blood, axial, coronal oblique and sagittal oblique cine balanced steady-state free precession (bSSFP), and contrast-enhanced magnetic resonance angiography (MRA) with multiplanar reformatting in axial, coronal oblique, sagittal oblique, and cross-sectional planes. Maximal branch PA and stenosis (if present) diameter were measured. Comparisons of PA size on CMR were made to reference methods: (1) catheterization measurements performed in the anteroposterior plane at maximal expansion, and (2) surgical measurement obtained from a maximal diameter sound which could pass through the lumen. The mean differences (Δ) and intra class correlation (ICC) were used to determine agreement between different modalities. CMR branch PA measurements were compared to the corresponding cardiac catheterization measurements in 30 children (7.6 ± 5.6 years). Reformatted MRA showed better agreement for branch PA measurement (ICC > 0.8) than black blood (ICC 0.4-0.6) and cine sequences (ICC 0.6-0.8). Coronal oblique MRA and maximal cross sectional MRA provided the best correlation of right PA (RPA) size with ICC of 0.9 (Δ -0.1 ± 2.1 mm and Δ 0.5 ± 2.1 mm). Maximal cross sectional MRA and sagittal oblique MRA provided the best correlate of left PA (LPA) size (Δ 0.1 ± 2.4 and Δ -0.7 ± 2.4 mm). For stenoses, the best correlations were from coronal oblique MRA of right pulmonary artery (RPA) (Δ -0.2 ± 0.8 mm, ICC 0.9) and sagittal oblique MRA of left pulmonary artery (LPA) (Δ 0.2 ± 1.1 mm, ICC 0.9). CMR PA measurements were compared to surgical measurements in 25 children (5.4 ± 4.8 years). All MRI sequences demonstrated good agreement (ICC > 0.8) with the best (ICC 0.9) from axial cine bSSFP for both RPA and LPA. Maximal cross sectional and angulated oblique reformatted MRA provide the best correlation to catheterization for measurement of branch PA's and stenosis diameter. This is likely due to similar angiographic methods based on reformatting techniques that transect the central axis of the arteries. Axial cine bSSFP CMR was the best surgically measured correlate of PA branch size due to this being a measure of stretched diameter. Knowledge of these differences provides more precise PA measurements and may aid catheter or surgical interventions for RVOTO lesions.