Image guidance in cardiac electrophysiology

Abstract

Cardiac arrhythmias are characterized by a disruption or abnormal conduction of electrical signals within the heart. Treatment of arrhythmias has dramatically evolved over the past half-century, and today, minimally-invasive catheter-based therapy is the preferred method of eliminating arrhythmias. Using an electroanatomical (EA) mapping system, which precisely tracks the position of catheters inside the patient's body, it is possible to construct three-dimensional maps of the ventricular and atrial chambers of the heart. Each point of these maps is annotated based on bioelectrical signals recorded from the electrodes located at the tip of the catheter. These maps are then used to guide catheter ablation within the heart. However, the electroanatomical mapping procedure results in relatively sparse sampling of the heart and a significant amount of time and skill are require to generate these maps. In this thesis, we present our software system for the integration of pre-operative, patient-specific magnetic resonance (MR) or computed tomography (CT) imaging data with real-time electroanatomical mapping (EAM) information.

(cont.) Following registration between the EAM and imaging data, the system allows for real-time catheter navigation within patient-specific anatomy. We then evaluate candidate registration strategies to rapidly and accurately align the pre-operative imaging data with the intra-operative mapping data using simulated electroanatomical mapping data using the great cardiac vessels including the aorta, superior vena cava, and coronary sinus. Based on these in vitro results, we focus on a registration strategy which is constrained by the ascending and descending aorta. In vivo prospective evaluation of the resulting image integration was then performed (n>200) in both experimental and clinical electrophysiology procedure. To compensate for residual error following registration or patient movement during a procedure, we present and evaluate warping strategies for deforming the pre-operative imaging data into agreement with the intra-operative mapping information.