Abstract 17258: Ventricular Shape in TGA Patients Differs Significantly From Controls
Introduction: In classic transposition of the great arteries (TGA), little is known about how the condition impacts ventricular structure, and if/how remodeling occurs as a consequence of the original condition and after the arterial switch operation (ASO) in the early neonatal period. We hypothesized that structural differences may be present in ASO TGA patients compared to controls, and sought to quantify these differences to determine if subjects can be defined as healthy or TGA based on ventricular shape information alone.
Methods: We studied population of 12 TGA patients (6 male, age 12±1.5 years) and 12 age-matched controls (4 male, age 13±1.7 years). Ventricular shape was segmented from short-axis CMR images at the end-diastole (ED) phase of the cardiac cycle and ED volume (EDV) was computed from the segmentation. Statistical shape analysis was applied to compute the most common ventricular shape features in the TGA population, after indexing for BSA. Each subject was characterized quantitatively as a combination of these shape features for the control- and TGA-populations. The sum of these values was used as a shape measure. A statistical classification method was applied to this measure to predict the patient group.
Results: Seven shape features captured 95% of the variance of shape in the TGA population and are summarized in the figure (a) with the % of variance of each shape feature given in brackets. The sum of the shape features for both populations is shown in the figure (b), as well as the line that divides the groups (in black). The classification yielded 100% accuracy (i.e. subjects were correctly classified). The resulting receiver operating characteristic curve is shown in the figure (c). There were no EDV differences between groups.
Conclusions: We found significant shape differences between the TGA patients and the healthy controls from the quantified measures of ventricular shape, which was highly accurate in discriminating between the two groups.
- Computer modeling
- Congenital heart disease
- Cardiac MRI
- Ventricular remodeling
- Congenital heart surgery, pediatric
Author Disclosures: K. McLeod: None. K. Rydén Suther: None. H. Brun: None. B. Smevik: None. A. Eltvedt Fiane: None. H.L. Lindberg: None. E. Hopp: None. C. de Lange: None.
- © 2016 by American Heart Association, Inc.