Characterization of baseline hemodynamics after the Fontan procedure: a retrospective cohort study on the comparison of 4D Flow MRI and computational fluid dynamics

Abstract

Introduction: The aim of this study was to characterize the hemodynamics of Fontan patients using both four-dimensional flow magnetic resonance imaging (4D Flow MRI) and computational fluid dynamics (CFD).

Methods: Twenty-nine patients (3.5 ± 0.5 years) who had undergone the Fontan procedure were enrolled, and the superior vena cava (SVC), left pulmonary artery (LPA), right pulmonary artery (RPA), and conduit were segmented based on 4D Flow MRI images. Velocity fields from 4D Flow MRI were used as boundary conditions for CFD simulations. Hemodynamic parameters such as peak velocity (Vmax), pulmonary flow distribution (PFD), kinetic energy (KE), and viscous dissipation (VD) were estimated and compared between the two modalities.

Results and discussion: The Vmax, KE, VD, PFDTotal to LPA, and PFDTotal to RPA of the Fontan circulation were 0.61 ± 0.18 m/s, 0.15 ± 0.04 mJ, 0.14 ± 0.04 mW, 41.3 ± 15.7%, and 58.7 ± 15.7% from 4D Flow MRI; and 0.42 ± 0.20 m/s, 0.12 ± 0.05 mJ, 0.59 ± 0.30 mW, 40.2 ± 16.4%, and 59.8 ± 16.4% from CFD, respectively. The overall velocity field, KE, and PFD from the SVC were in agreement between modalities. However, PFD from the conduit and VD showed a large discrepancy between 4D Flow MRI and CFD, most likely due to spatial resolution and data noise. This study highlights the necessity for careful consideration when analyzing hemodynamic data from different modalities in Fontan patients.