Ventricular contouring of cardiac magnetic resonance imaging is the gold standard for volumetric analysis for repaired tetralogy of Fallot (rTOF), but can be time-consuming and subject to variability. A convolutional neural network (CNN) ventricular contouring algorithm was developed to generate contours for mostly structural normal hearts. We aimed to improve this algorithm for use in rTOF and propose a more comprehensive method of evaluating algorithm performance. We evaluated the performance of a ventricular contouring CNN, that was trained on mostly structurally normal hearts, on rTOF patients. We then created an updated CNN by adding rTOF training cases and evaluated the new algorithm’s performance generating contours for both the left and right ventricles (LV and RV) on new testing data. Algorithm performance was evaluated with spatial metrics (Dice Similarity Coefficient (DSC), Hausdorff distance, and average Hausdorff distance) and volumetric comparisons (e.g., differences in RV volumes). The original Mostly Structurally Normal (MSN) algorithm was better at contouring the LV than the RV in patients with rTOF. After retraining the algorithm, the new MSN + rTOF algorithm showed improvements for LV epicardial and RV endocardial contours on testing data to which it was naïve (N = 30; e.g., DSC 0.883 vs. 0.905 for LV epicardium at end diastole, p < 0.0001) and improvements in RV end-diastolic volumetrics (median %error 8.1 vs 11.4, p = 0.0022). Even with a small number of cases, CNN-based contouring for rTOF can be improved. This work should be extended to other forms of congenital heart disease with more extreme structural abnormalities. Aspects of this work have already been implemented in clinical practice, representing rapid clinical translation. The combined use of both spatial and volumetric comparisons yielded insights into algorithm errors.

Purpose: Four-dimensional time-resolved phase-contrast cardiovascular magnetic resonance imaging (4D flow MRI) enables blood flow quantification in multiple vessels, which is crucial for patients with congenital heart disease (CHD). We investigated net flow volumes in the ascending aorta and pulmonary arteries by four different postprocessing software packages for 4D flow MRI in comparison with 2D cine phase-contrast measurements (2D PC). Material and methods: 4D flow and 2D PC datasets of 47 patients with biventricular CHD (median age 16, range 0.6–52 years) were acquired at 1.5 T. Net flow volumes in the ascending aorta, the main, right, and left pulmonary arteries were measured using four different postprocessing software applications and compared to offset-corrected 2D PC data. Reliability of 4D flow postprocessing software was assessed by Bland–Altman analysis and intraclass correlation coefficient (ICC). Linear regression of internal flow controls was calculated. Interobserver reproducibility was evaluated in 25 patients. Results: Correlation and agreement of flow volumes were very good for all software compared to 2D PC (ICC ≥ 0.94; bias ≤ 5%). Internal controls were excellent for 2D PC (r ≥ 0.95, p < 0.001) and 4D flow (r ≥ 0.94, p < 0.001) without significant difference of correlation coefficients between methods. Interobserver reliability was good for all vendors (ICC ≥ 0.94, agreement bias < 8%). Conclusion: Haemodynamic information from 4D flow in the large thoracic arteries assessed by four commercially available postprocessing applications matches routinely performed 2D PC values. Therefore, we consider 4D flow MRI-derived data ready for clinical use in patients with CHD.

Volumetric black-blood cardiovascular magnetic resonance (CMR) has been hampered by long scan times and flow sensitivity. The purpose of this study was to assess the feasibility of black-blood, electrocardiogram (ECG)-triggered and respiratory-navigated 3D fast spin echo (3D FSE) for the visualization of the whole heart and great vessels. The implemented 3D FSE technique used slice-selective excitation and non-selective refocusing pulses with variable flip angles to achieve constant echo signal for tissue with T1 (880 ms) and T2 (40 ms) similar to the vessel wall. Ten healthy subjects and 21 patients with congenital heart disease (CHD) underwent 3D FSE and conventional 3D balanced steady-state free precession (bSSFP). The sequences were compared in terms of ability to perform segmental assessment, local signal-to-noise ratio (SNRl) and local contrast-to-noise ratio (CNRl). In both healthy subjects and patients with CHD, 3D FSE showed superior pulmonary vein but inferior coronary artery origin visualisation compared to 3D bSFFP. However, in patients with CHD the combination of 3D bSSFP and 3D FSE whole-heart imaging improves the success rate of cardiac morphological diagnosis to 100% compared to either technique in isolation (3D FSE, 23.8% success rate, 3D bSSFP, 5% success rate). In the healthy subjects SNRl for 3D bSSFP was greater than for 3D FSE (30.1 ± 7.3 vs 20.9 ± 5.3; P = 0.002) whereas the CNRl was comparable (17.3 ± 5.6 vs 17.4 ± 4.9; P = 0.91) between the two scans. The feasibility of 3D FSE for whole-heart black-blood CMR imaging has been demonstrated. Due to their high success rate for segmental assessment, the combination of 3D bSSFP and 3D FSE may be an attractive alternative to gadolinium contrast enhanced morphological CMR in patients with CHD.

Introduction: Patients with Tetralogy of Fallot (TOF) and pulmonary insufficiency and a dilated right ventricle (RV) may suffer from a reduction in left ventricular (LV) performance. It is not clear whether timely pulmonary valve replacement (PVR) preserves LV mechanics. Methods: Ten TOF patients who underwent PVR were identified from hospital records, and pre- and postoperative CMR images were post-processed with a semi-automatic tissue tracking software. LV circumferential strain, time to peak strain, and torsion were compared before and after PVR. A control group of ten age-matched normal volunteers was assessed as a comparison. Results: LV circumferential strain did not change before versus after PVR (basal -18.3 +/- 3.7 % vs. -20.5 +/- 3 %, p 0.082; mid-ventricular -18.4 +/- 3.6 % vs. -19.1 +/- 2 %, p 0.571; apical -22.7 +/- 5.2 % vs. -22.1 +/- 4 %; p 0.703). There was also no difference seen between the baseline strain and normal controls (control basal -18.2 +/- 3.3 %, p 0.937; mid -18 +/- 3.2 %, p 0.798; apex -24.1 +/- 5 %, p 0.552). LV torsion remained unchanged from baseline to post PVR (systolic 2.75 (1.23 – 9.51) °/cm vs. 2.3 +/- 1.2 °/cm, p 0.285; maximum 5.5 +/- 3.5 °/cm vs. 2.34 (1.37 – 8.07) °/cm, p 0.083). There was no difference in time to measured peak LV circumferential strain before versus after PVR (basal 0.44 +/- 0.1 vs. 0.43 +/- 0.05, p 0.912; mid-ventricular 0.42 +/- 0.08 vs. 0.38 +/- 0.06, p 0.186; apical 0.40 +/- 0.08 vs. 0.40 +/- 0.06, p 0.995). At the same time, pulmonary regurgitation and RV end-diastolic and end-systolic volume indices decreased and LV end-diastolic volume increased after PVR. RV and LV ejection fractions remained constant. Conclusion: PVR allows for favorable remodeling of both ventricular volumes for TOF patients with significant pulmonary regurgitation. In this cohort, LV myocardial functional parameters such as circumferential strain, time to peak strain, and LV torsion were normal at baseline and remain unchanged after PVR.