An Efficient Registration Algorithm for Advanced Fusion of 2D/3D Angiographic Data

Abstract

Computed tomography angiography (CTA) is often used for pre-interventional diagnosis and planning, whereas nowadays, mostly 2D angiograms are acquired for intra-interventional catheter guidance. Spatial information from pre-interventional scans is not transferred to the intervention yet since existing 2D-3D registration methods either require a good initial manual alignment or have rather long runtime and thus lack clinical usefulness. We propose a fast and automatic method for 2D-3D registration and evaluate methods for intra-interventional visualization and navigation. Moreover, we introduce an easy clinical work- flow for transferring a planned roadmap from pre-interventional 3D to intra-interventional 2D using the registration. We demonstrate the good quality of fit and the fast runtime of the algorithm on one phantom and three patient data sets.