Four-dimensional flow MRI quantification of pulmonary regurgitation in patients with repaired tetralogy of Fallot: a study on reproducibility and consistency

Abstract

Objective To investigate the feasibility and consistency of four-dimensional flow (4D flow) quantification of pulmonary regurgitation in patients with repaired Tetralogy of Fallot (ToF) by comparing with conventional two-dimensional flow (2D flow) and echocardiography. Methods Both the 4D flow and 2D flow imaging were acquired with repaired ToF (a total of 21 patients) consecutively on 3.0 T MR scanner from May 2018 to August 2018. Pulmonary flow and regurgitant fraction were measured by a commercial post processing software Circle CVI42. All patients underwent echocardiography within one week after or before MR examination. The inter/intra-observer variability by 2D/4D f1ow and agreement between the two methods were investigated by interclass correlation coefficients (ICC) and Bland-Altman analyses. The agreement between MR and echocardiography were analyzed by weighted Kappa coefficient. The correlation between pulmonary regurgitation and cardiac function was also investigated by Pearson analysis. Results All patients were included and completed the examinations successfully. Both inter-observer and intra-observer agreement by 4D flow for total forward volume (ICC=0.993, 0.996, respectively, P<0.001), total backward volume (ICC=0.994, 0.997, respectively, P<0.001) and regurgitant fraction (ICC=0.968, 0.985, respectively, P<0.001) were good. The total forward volume, total backward volume and regurgitant fraction measured by 2D flow and 4D flow reached a good agreement (ICC=0.954, 0.913,0.721,respectively,P<0.001). The consistency was good for severity of regurgitation measured by 2D flow (weighted Kappa=0.897, P<0.001) and 4D flow (weighted Kappa=0.710, P=0.001) compared with echocardiography. Significant correlation was found among right ventricular cardiac index(r=0.600, P<0.05), right ventricular end-diastolic volume index(r=0.788, P<0.05), right ventricular end-systolic volume index(r=0.683, P<0.05) and left ventricular end-diastolic volume index(r=0.578, P<0.05), left ventricular end-systolic volume index(r=0.687, P<0.05) with regurgitant fraction measured by 2D flow. Regurgitant fraction measured by 4D flow had a significant correlation with right ventricular cardiac index(r=0.606, P<0.05), right ventricular end-diastolic volume index(r=0.685,P<0.05), right ventricular end-systolic volume index(r=0.534, P<0.05) and left ventricular end-diastolic volume index(r=0.459, P<0.05), left ventricular mass index(r=0.633, P<0.05). Conclusion 4D flow MRI provides highly reproducible measurements of pulmonary flow on morphology and haemodynamics in patients with repaired ToF compared with 2D flow MRI and echocardiography. Key words: Magnetic resonance imaging; Pulmonary regurgitation; Tetralogy of Fallot