Event classification in percutaneous treatments based on needle insertion force pattern analysis

Abstract

Percutaneous treatments are becoming a common minimally invasive treatment for cancer. Surgeons must introduce a needle into a cancerous area to perform a biopsy or kill the cancer, but this is a complex procedure that often misses the target. Novel robotic devices to assist in percutaneous treatments are being developed. Most systems use preoperative FEM simulation to calculate a needle trajectory that will hit the target, or US image needle guidance. However, neither of them are fully satisfactory. Ideally, real-time simulators should be used for intraoperative robot control, but, they lack accuracy. We propose a new method to provide information about the current situation during a needle insertion using needle insertion force pattern recognition. This information can be used as feedback for simulators or robot control in order to increase their accuracy.