부산대학교 도서관

In neurosurgeries, even a small mistake may pose a serious risk to a patient. Hence, a neuronavigation has been used in neurosurgeries to find accurate spatial position of lesion. However, the neuronavigation localizes the lesion based on preoperative medical images, therefore the brain deformation during a surgery, which is referred to as brain shift, has been a serious problem in the neuronavigation. Brain shift occurs in craniotomy and it is caused by gravity, cerebrospinal fluid (CSF) leakage, and so on. In order to present accurate information on a neuronavigation during surgery, a simple brain shift estimation method is proposed to predict and reproduce intraoperative brain deformation. The proposed method is validated by using a Finite Element (FE) brain model generated from medical images of a real patient. A brain shift simulation using a patient brain model is conducted. The comparison of the simulation result and the intraoperative medical images of the patient show that the simulation can reproduce the brain shift accurately. Most earlier studies have verified their brain shift estimation methods in 2D coordinate by using the 2D medical images. In this study, the accuracy was verified in x-y-z coordinate by using the 3D models.