Back to Annual Meeting Posters

OBJECTIVE: Precise morphological and functional knowledge about the mitral valve (MV) is a prerequisite especially in patients undergoing MV repair. We propose a novel automatic system for patient-specific four-dimensional (4D) modeling and quantification of mitral valves, which operates on real-time three-dimensional (3D) echocardiography data.
METHODS: Patients underwent MV repair due to severe mitral valve regurgitation. Real-time 3D transesophageal echocardiographic examination of the MV was performed pre- and postoperatively. Data was collected and included into the MV model (Siemens Corporate Research, Princeton, NJ, USA). Mitral valve modeling includes three steps: Step 1, MV localization in the 3D volume; Step 2, detection of anatomic landmarks (trigones, commissures, papillary muscles and leaflet tips); Step 3, detection of shape and dynamics (annulus, anterior and posterior leaflets). The modeling was constructed automatically and the total computation time of Step 1 - 3 was taken less than 2 minutes.
RESULTS: A large variety of morphological and functional measurements as well as MV images are derived from the model over the complete cardiac cycle: mitral annular area and circumference, annular diameters, leaflet lengths and areas, annular height, commissural width, annular height-to-commissural width ratio, tenting height, and others, respectively. 4D visualization of the complex MV dynamic is applicable.
CONCLUSIONS: We herein present a brand-new patient-specific MV model, which allows quantification of MV functional and morphological parameters. The model presents another step towards a full patient-specific simulation of MV repair.