Catheter navigation based on probabilistic fusion of electromagnetic tracking and physically-based simulation

Minimally invasive endovascular procedures including robotically assisted intervention require effective intraoperative guidance. This is mainly achieved through intraoperative imaging such as fluoroscopy. Concerns over excessive x-ray radiation and nephrotoxicity due to repeated injection of contrast agents have motivated the development of effective catheter navigation schemes based on limited imaging data. This paper presents a catheter navigation technique based on probabilistic fusion of in situ real-time electromagnetic tracking with physically-based simulation of the mechanical characteristics of the catheter. A catheter with multiple electromagnetic sensors placed along its length has been developed. The sensor data and the catheter insertion-length are used as the boundary condition for determining the shape and position of the catheter within the vasculature. A probabilistic framework based on a Kalman Filter is used to combine the information from the catheter motion algorithm and the electromagnetic tracking data. This provides continuous visualization of the catheter within the lumen without the need of continuous fluoroscopy and contrast injection. The proposed approach has been validated with detailed in vitro experiments demonstrating the potential clinical application of the technique.