A porcine liver model for validation of registration accuracy in image-guided surgery

Abstract

Correct registration between pre-operative high-resolution images and intra-operative data of lower detail is a fundamental requirement in image-guided liver surgery. We propose a multi modality liver model for measuring the accuracy of such registration methods. A freshly explanted porcine liver is artificially perfused by a peristaltic pump and liver motion is simulated by means of inflatable objects positioned around the liver. Co-registered ultrasound and CT data sets are acquired in different deformation scenarios and allow comparing registration outcomes with a CT data set serving as ground truth. The pre-sent work describes the experimental setup and summarizes the results from ultrasound and CT imaging.