Studies Database
Li Y, Brodlie K and Phillips N. Web-based VR training simulator for percutaneous rhizotomy. Study in Health Technology and Informatics. 2000;70:175-81
PMID: 10977535Abstract
Virtual Reality offers great potential for surgical training--yet is typically limited by the dedicated and expensive equipment required. Web-based VRhas the potential to offer a much cheaper alternative, in which simulations of fundamental techniques are downloaded from a server to run within a web browser. The equipment requirement is modest--an Internet-connected PC or small workstation--and the simulation can be accessed worldwide. In a collaboration between computer scientists and neurosurgeons, we have studied the use of web-based VR to train neurosurgeons in PercutaneousRhizotomy--a treatment for the intractable facial pain which occurs in trigeminal neuralgia. This involves the insertion of a needle so as to puncture the foramen ovale, and lesion the nerve. Our simulation uses VRML to provide a 3D visualization environment, but the work immediately exposes a key limitation of VRML for surgical simulation. VRML does not support collision detection between objects--only between viewpoint and object. Thus collision between needle and skull cannot be detected and fed back to the trainee. We have developed a novel solution in which the trainingsimulation has linked views: a normal view, plus a view as seen from the tip of the needle. Collision detection is captured in the needle view, and fed back to the viewer. A happy consequence of this approach has been the chance to aid the trainee with this additional view from needle tip, which helps locate the foramen ovale. The technology to achieve this is Java software communicating with the VRML worlds through the External Authoring Interface (EAI). The training simulator is available on the Web, with accompanying tutorial on its use. A major advantage of web-based VR is that the techniques generalize to a whole range of surgical simulations. Thus we have been able to use exactly the same approach as described above for neurosurgery, to develop a shoulder arthroscopy simulator--where again collision detection, and the view from the scope, are fundamental.
Author's contacts: ying@scs.leeds.ac.uk