Three-dimensional reconstruction of ventricular septal defects: validation studies and in vivo feasibility.

ABSTRACT

OBJECTIVES: The purpose of this study was to demonstrate the feasibility of in vivo three-dimensional reconstruction of ventricular septal defects and to validate its quantitative accuracy for defect localization in excised hearts (used to permit comparison of three-dimensional and direct measurements without cardiac contraction). BACKGROUND: Appreciating the three-dimensional spatial relations of ventricular septal defects could be useful in planning surgical and catheter approaches. Currently, however, echocardiography provides only two-dimensional views, requiring mental integration. A recently developed system automatically combines two-dimensional echocardiographic images with their spatial locations to produce a three-dimensional construct. METHODS: Surgically created ventricular septal defects of varying size and location were imaged and reconstructed, along with the left and right ventricles, in the beating heart of six dogs to demonstrate the in vivo feasibility of producing a coherent image of the defect that portrays its relation to surrounding structures. Two additional gel-filled excised hearts with defects were completely reconstructed. Quantitative localization of the defects relative to other structures (ventricular apexes and valve insertions) was then validated for seven defects in excised hearts. The right septal margins of the exposed defects were also traced and compared with their reconstructed areas and circumferences. RESULTS: The three-dimensional images provided coherent images and correct spatial appreciation of the defects (two inlet, two trabecular, one outlet and one membranous Gerbode in vivo; one inlet and one apical in excised hearts). The distances between defects and other structures in the excised hearts agreed well with direct measures (y = 1.05x-0.18, r = 0.98, SEE = 0.30 cm), as did reconstructed areas (y = 1.0x-0.23, r = 0.98, SEE = 0.21 cm2) and circumferences (y = 0.97x + 0.13, r = 0.97, SEE = 0.3 cm). CONCLUSIONS: Three-dimensional reconstruction of ventricular septal defects can be achieved in the beating heart and provides an accurate appreciation of defect size and location that could be of value in planning interventions.