A new biomechanical model based approach on brain shift compensation

Abstract

We propose a new algorithm for biomechanical model-based brain shift compensation in image guided neurosurgery. It can be used to update preoperative images with intraoperatively acquired information. We derive a model equation with regard to external forces acting on the brain surface during neurosurgery which can be consistently integrated with intraopearatively acquired information, assuming that these forces induce a linear biomechanical response. We treat external forces on the brain boundaries as unknown variables and then estimate them within a framework of inverse finite element analysis. By incorporating additional constraints from prior knowledge, we can solve the derived equations to obtain reasonable estimation results on boundary forces and the entire displacement field. This algorithm is especially beneficial in reducing navigation error of deeper brain structures by updating preoperative images using only exposed surface displacement. In this paper, we describe the derivation of the equations and present examples of two dimensional synthetic data, where the estimated displacement errors are reduced by fifty percent, compared to the standard approach. © Springer-Verlag Berlin Heidelberg 2003.