Overiew

The driving vision for this NSF ITR research is the development of a practical grid architecture that will allow access to multi-teraflop high performance computing architectures and associated computational science algorithms, directly from a surgery room, for real-time computer assisted surgery. Such access is essential for solving the next generation of advanced biomechanical models within the time constraint of image guided neurosurgery (IGNS). The key surgical challenge neurosurgeons have during tumor resection is the intra-operative shape deformation of the brain. Preoperative investigations can not account for this type of deformation, which occurs as a result of the actual surgical process. It is therefore of great importance to be able to quantify and correct for these deformations while a surgery is in progress, by updating pre-operative images in a way that enables surgeons to react to changing conditions. To have practical applicability, nonrigid biomechanical simulation and alignment of pre- and intra-operative images need to be performed within time constraints of a few minutes.

Major effort is being devoted towards demonstrating feasibility of a) acquiring data intraoperatively at the Brigham and Women’s Hospital (BWH), b) processing that data for the finite element model, c) transferring the data to the San Diego Supercomputer Center (SDSC), d) solving a large advanced finite element model, e) transferring the results back to BWH, and f) displaying new intraoperative images to neurosurgeons during the surgery. All tasks need to be performed within the time constraints of neurosurgery, such that this whole process is of practical and reliable use to surgeons