3D coronary reconstruction from calibrated motion-compensated 2D projections based on semi-automated feature point detection

Abstract

3D rotational coronary angiography (3DRCA) is one of the application areas of 3D rotational X-Ray imaging. In this application a sequence of projection images is acquired when the C-arm is rotated around the patient. Since the heart is a moving object, only projections can be used which correspond to the same phase of the cardiac cycle. This significantly limits the number of projections available for reconstruction causing streaking artefacts in the reconstructed image due to angular undersampling. The involvement of additional projections in the reconstruction procedure from different viewing angles would increase the quality of the volume data. Each successive acquired projection is slightly different compared with the previous one due to two reasons: First, there is a motion to the deformation of the heart, second there is an induced deformation owing to the change in the projection angle. The purpose of this work is to determine the motion owing to the heart deformation, so as to compensate for this motion in projection images in a different heart phase. Hereto we propose to use concepts from coronary modeling in combination of conventional reconstruction procedures. The proposed method facilitates the use of additional projections in the reconstruction. Motion-compensated reconstructed volume data are presented for coronary arteries in an animal (pig) model.