IMAGE SYNTHESIS OF DEFORMED TISSUE: APPLICATION TO ULTRASOUND FOR PROSTATE BRACHYTHERAPY

Training simulators are required to mimic the med-ical image modalities to visually immerse the user in the virtual reality. Medical simulations often discretize the tissue to model its deformation, which is very common during medical interventions. Modeling techniques, such as the Finite Element Method, usually employ meshes with much coarser elements than the resolu-tion of typical medical images. This paper presents an image synthesis procedure inside a deformed volume, given the node displacements of its discretization. In this method, directional imaging artifacts are neglected which is an assumption that most volume reconstruc-tion methods also employ. The presented technique is applicable to CT, MRI, or ultrasound volume data. A specific application to transrectal ultrasound (TRUS) is given, since the visual plausibility of speckle patterns on ultrasound images is a concern especially when images are interpolated from two or more initial data planes. Our implementation synthesizes B-mode TRUS images during needle insertion simulation of prostate brachy-therapy. The synthesized images do not contain sig-nificant interpolation artifacts. Achieving TRUS image generation in real-time is our focus of future research