1	High Fidelity Simulation of Interventional Neuroradiology Procedures Stephane COTIN, Vincent LUBOZ, Julien LENOIR, Christian DURIEZ, Xunlei WU, Paul NEUMANN, Vincent PEGORARO, Dr. James RABINOV, Dr. Steven DAWSON The Simulation Group, CIMIT, MGH, Harvard Medical School http://www.medicalsim.org/
2	Introduction Catheterization is the main therapy for ischemic stroke, which is the 3^rd leading cause of death in US. Due to its complexity and training requirements, interventional radiology is an ideal platform to introduce computer-based training. The US FDA mandated proficiency training including simulation for carotid stenting in 2004.
3	Previous work Virtual Reality systems on IR have been developed over the past decade. We propose a real-time high fidelity simulator for IR training and future procedure planning with new approaches in rendering, physics-based modeling, and unified anatomical representations.
4	Components Vascular segmentation Catheter behavior with collision response Fluid flow computation Contrast agent propagation Fluoroscopic rendering Device tracking
5	Semi-automatic Vascular Segmentation
6	Vascular Surface Model
7	Catheter Behavior 3D beam theory with collision response can be used to represent non-linear catheter and guidewire behavior.
8	Blood Flow Computation Approximate flow with curvilinear model based on average radii, circular cross sections and boundary conditions. Image shows vascular surface color coded based on computed blood pressure values superimposed over a synthetic fluoroscopic image. Investigating high order flow algorithms.
9	Synthetic Angiography Simulate contrast agent (CA) propagation with an advection equation on CA concentration within the blood flow stream.
10	3D Fluoroscopic Rendering Interactive fluoroscopic rendering which replicates the X-ray process with any CT data set by utilizing hardware accelerated volume rendering techniques.
11	Catheter/Guidewire Tracking Xitact’s VSP device (www.xitact.com) Simultaneous tracking and force feedback on up to 3 coaxial instruments Integrates devices for contrast injection, angioplasty and stent deployment C-arm and patient table control console, foot pedals and X-ray controls
12	Results Implemented on P4 3.0GHz desktop with NVIDIA GeForce FX5900 GPU. Render 256 CT slices at 18 frames/second. This frame rate includes fluoroscopic rendering, collision detection and response, catheter/guidewire deformation, vascular flow, and CA propagation.
13	Results
14	Live Demonstration
15	Conclusion and Future Work A set of simulation components have been developed and integrated into a real-time training system for the treatment of stroke. Focus on high fidelity visual feedback and physically accurate modeling. Cost-effective and compact design leads to cross-specialty training and increases accessibility to institutions and hospitals. A work in progress. In the future, Simulate balloons and carotid stents with deformable models. System validation study. Incorporate an educational curriculum to unleash the potential of computer-based training and procedure planning.
16	The Simulation Group Members of The SIM Group at CIMIT have been actively involved in the development of medical simulation since the mid-1990’s. Formed in February, 2001 the multi-disciplinary research team was organized under the leadership of Dr. Steve Dawson to investigate how technology can improve medical education and increase patient safety. With expertise in all areas of medical simulation, CIMIT Simulation is working to make both surgery and medical practice safer. http://www.medicalsim.org/
17	Acknowledgements Funding support from TATRC (DAMD 17-02-2-0006) and 2005 CIMIT Application Development Award. Special thanks to Dr. Karl Krissian from Surgical Planning laboratory (SPL) for assisting on the vascular segmentation.