MODELING OF INTRAOPERATIVE BRAIN SHIFT FOR REGISTRATION

摘要

We present preliminary results of our research on intraoperative brain shift modeling. This is a part of our wider research interest on intraoperative brain MR registration. We try to increase the accuracy of preoperative to intraoperative MR image registration. We are particularly interested in intraoperative images obtained after a tumour resection had been performed. The soft brain tissue deforms as the body of the tumour is being removed. The topic of this paper is modeling of gravity-induced brain shift appearing due to tumour resection, in order to guide a preoperative to intraoperative brain registration algorithm. The idea is to segment the brain from the intraoperative images and to construct an elastic model of the brain based on segmentation results. Elastic parameters of the brain model are obtained from in-vivo experiments. In order to model the effect of gravity force we apply the force equal to gravitation but with opposite direction. Such model of the brain without tumour body is then deformed using Finite Elements Method (FEM). In this research we used National Library of Medicine Insight Segmentation and Registration Toolkit (ITK).