Sequential reconstruction of vessel skeletons from X-ray coronary angiographic sequences.

X-ray coronary angiography (CAG) is one of widely used imaging modalities for diagnosis and interventional treatment of cardiovascular diseases. Dynamic CAG sequences acquired from several viewpoints record coronary arterial morphological information as well as dynamic performances. The aim of this work is to propose a semi-automatic method for sequentially reconstructing coronary arterial skeletons from a pair of CAG sequences covering one or several cardiac cycles acquired from different views based on snake model. The snake curve deforms directly in 3D through minimizing a predefined energy function and ultimately stops at the global optimum with the minimal energy, which is the desired 3D vessel skeleton. The energy function combines intrinsic properties of the curve and acquired image data with a priori knowledge of coronary arterial morphology and dynamics. Consequently, 2D extraction, 3D sequential reconstruction and tracking of coronary arterial skeletons are synchronously implemented. The main advantage of this method is that matching between a pair of angiographic projections in point-by-point manner is avoided and the reproducibility and accuracy are improved. Results are given for clinical image data of patients in order to validate the proposed method.