Statistical atlas based registration and planning for ablating bone tumors in minimally invasive interventions

Bone tumor ablation has been a viable treatment in a minimally invasive way compared with surgical resections. In this paper, two key challenges in the computer-assisted bone tumor ablation have been addressed: 1) establishing the spatial transformation of patient's tumor with respect to a global map of the patient using a minimum number of intra-operative images and 2) optimal treatment planning for large tumors. Statistical atlas is employed to construct the global reference map. The atlas is deformably registered to a pair of intra-operative fluoroscopy images, constructing a patient-specific model, in order to reduce the radiation exposure to the sensitive patients such as pregnant and infants. The optimal treatment planning system incorporates clinical constraints on ablations and trajectories using a multiple objective optimization, which obtains optimal trajectory planning and ablation coverage using integer programming. The proposed system is presented and validated by experiments.