Needle Tumor Puncture Detection Using a Force and Position Sensor

Preforming a tumor ablation procedure involves precisely positioning an ablation probe in a tumor. Limited sensing accuracy and variations in tissue properties require that surgeons often identify when they have correctly positioned the probe through their sense of touch. In this paper, supervised learning is used to detect the point at which a needle has entered a tumor using a combination of force and position data. A classifier is developed that uses force data to detect a puncture event, and a Bayesian network that incorporates a position estimate to distinguish between puncture events that may occur because of tissue variations, and puncture events that indicate the target has been reached.