Abstract
Objectives: To investigate the computer graphics (CG) reconstruction methods as visualization tools for cerebrovascular surgery simulations and their possible clinical applications.
Methods: The following CG technologies were used to improve three-dimensional fusion images reconstructed from medical images: particle-based fluid simulation enabling blood flow analysis; rigging and mesh editing for virtual tissue deformation and fusion with CG of normal anatomy, machine learning technology for automatic tissue segmentation, mixed reality technology enabling fusion with physical space, and rendering with texturing and shading. Problems and contributions of the proposed CG methods in preoperative surgical simulation were assessed in 299 cases of cerebrovascular surgery.
Results: Blood flow analysis enabled using the proposed CG method contributed to the differentiation of abnormal vessels, arteries, and veins in cerebral arteriovenous malformations. Mixed reality technology has contributed to the identification of anatomical features within a complex surgical field. Virtual tissue deformation and fusion with CG of normal anatomy offered a more flexible preoperative assessment than 3D fusion images constructed from medical images alone. The modified rendering improved the visibility. On the other hand, neurosurgeons thought that the image processing required for the proposed CG method was complicated and challenging. Image processing is also time-consuming.
Conclusion: The proposed methods have the potential to be more useful than conventional medical images in preoperative assessment for cerebrovascular surgery, but technical issues remain to be addressed.
