In-situ tumor locating for robotic minimally invasive surgery

Abstract

Manual palpation for verifying the exact location of anomalies in open surgery is not possible through the small incisions of minimally invasive surgery. Robotic palpation is a simple alternative to intra-operatively confirm the tumor location obtained from a CT scan or MRI in order to make a precise cut. This study proposed a novel approach to mimic manual palpation by exerting a constant force to deform the tissue surface with an indenter, much more flexible than using constant displacement of others, and analyzing the corresponding induced surface profile for estimating the location of known abnormalities underneath. Finite element (FE) modeling was used to map the surface indentation for the estimation. The study used porcine liver to be an analog of human liver, where unconfined uniaxial tests were carried out to extract stretch-stress curves from the central region of liver specimens for updating the FE model. The first-order Ogden model was selected as the best hyperelastic model to characterize the liver tissue and indentation experiments were carried out with good agreements achieved. Cancerous computational models were constructed to investigate the mechanical responses of the tumor tissue under indentation. Indentation curves were extracted from the point cloud data of the deformed surfaces of both normal and abnormal tissue. Curve fitting methods were employed to characterize the indentation profile with empirical equations. These equations were used to develop a volume-based method to present how the tumor position affects the surface deformation of the tissue. Indentation experiments were conducted on healthy and "disease" porcine liver specimens with an optical device to obtain the shape of the deformation. Using an optimization algorithm to minimize the differences of the indentation curves between experiments and simulation, the optimal in vivo mechanical parameters of porcine liver were obtained for FE models to generate a set of deformation templates for a variety of tumor locations. These templates can be used in matching the measured data of an indentation for estimation of the tumor location underneath with accuracy.