Cardiac MRI in Children and Adolescents Who Have Undergone Surgical Repair of Right-Sided Congenital Heart Disease: Automated Left Ventricular Volumes and Function Analysis and Effects of Different Manual Adjustments.

Abstract

To evaluate automated segmentation and the effects of different manual adjustments regarding left ventricular parameter quantification in cardiac magnetic resonance (MR) data on children and adolescents who have undergone surgical repair of right-sided congenital heart disease (CHD). Dedicated software (syngo.via, Siemens AG) was used to automatically segment and/or manually adjust the end-diastolic volume (EDV), end-systolic volume (ESV), stroke volume (SV), myocardial mass (MM) and ejection fraction (EF) before/after manual apex/base adjustment (ADJ-step 1) and after manual apex/base/myocardial contour adjustment (ADJ-step 2; reference standard). MR data of 40 patients (13.1 ± 3. y, 4 - 17 y) with repaired CHD with decreased pulmonary blood flow (CHD-DPBF) were evaluated. Intra- and inter-rater reliability was determined for 10 randomly selected patients. The software correctly detected the left ventricle in 38/40 (95%) patients. EDV after automated segmentation: 119.1 ± 44.0 ml; after ADJ-step 1: 115.8 ± 9.5 ml; after ADJ-step 2: 116.2 ± 39.4 ml. The corresponding results for ESV were 52.0 ± 18.5/49.6 ± 6.9/49.7 ± 16.4 ml; for SV 67.1 ± 28.5/66.2 ± 25.4/66.5 ± 25. ml; for EF 55.5 ± 7.3/56.7 ± 6.6/56.7 ± 6.3%; for MM 83.7 ± 35.9/76.2 ± 8.3/74.6 ± 27.2 g. Significant differences were found for ESV/MM/EF comparing the automated segmentation results with these after ADJ-step 1 and ADJ-step 2. No significant differences were found when comparing all results of ADJ-step 1 and ADJ-step 2 or when comparing EDV/SV results. Intra- and inter-rater reliability was excellent. The mean time effort was 63.4 ± 6.9 s for the automated segmentation, 74.2 ± 8.9 s for ADJ-step 1 and 269.5 ± 39.4 s for ADJ-step 2. Automated left ventricular volumes and function analysis in children and adolescents with surgically treated CHD proved to be feasible with excellent intra- and inter-rater reliability. Automated segmentation with manual apex/base adjustment provided clinically acceptable results. Automated left ventricular volume and function analysis in children and adolescents with surgically treated right-sided heart disease is feasible with excellent intra- and inter-rater reliability. Automated segmentation with manual apex/base adjustment provides clinically acceptable results. Additional manual myocardial contour adjustment does not significantly improve the results.