Evaluation of knowledge-based reconstruction for magnetic resonance volumetry of the right ventricle after arterial switch operation for dextro-transposition of the great arteries.

Right ventricular (RV) volume and function evaluation is essential in the follow-up of patients after arterial switch operation (ASO) for dextro-transposition of the great arteries (d-TGA). Cardiac magnetic resonance (CMR) imaging using the Simpson's method is the gold-standard for measuring these parameters. However, this method can be challenging and time-consuming, especially in congenital heart disease. Knowledge-based reconstruction (KBR) is an alternative method to derive volumes from CMR datasets. It is based on the identification of a finite number of anatomical RV landmarks in various planes, followed by computer-based reconstruction of the endocardial contours by matching these landmarks with a reference library of representative RV shapes. The purpose of this study was to evaluate the feasibility, accuracy, reproducibility and labor intensity of KBR for RV volumetry in patients after ASO for d-TGA. The CMR datasets of 17 children and adolescents (males 11, median age 15) were studied for RV volumetry using both KBR and Simpson's method. The intraobserver, interobserver and intermethod variabilities were assessed using Bland-Altman analyses. Good correlation between KBR and Simpson's method was noted. Intraobserver and interobserver variability for KBR showed excellent agreement. Volume and function assessment using KBR was faster when compared with the Simpson's method (5.1 ± 0.6 vs. 6.7 ± 0.9 min, p < 0.001). KBR is a feasible, accurate, reproducible and fast method for measuring RV volumes and function derived from CMR in patients after ASO for d-TGA.