Purpose: To develop a highly efficient means to automatically determine optimal beam isocenter locations and beam sizes in the planning for Leksell Perfexion Gamma knife based SRS treatments. Methods: 2D contours of the tumor were exported from the Gamma knife planning system and re‐ digitized as a 3D volume of voxels. A fast morphological method was used to generate the skeleton of the 3D volume. A group of initial sampling points were generated inside the target volume based on the topology and geometry of the skeleton. A fast 3D Voronoi diagram algorithm was developed, using these sampling points as the starting location of diagrams, to iteratively divide the target volume into anisotropic 3D diagrams with the size in approximation to the Leksell Perfexion Gamma Knife beams. The centers of Voronoi diagrams were used as the location of isocenters and the sizes of diagrams determine the size for each individual beam to achieve optimal conformity of the dose distribution and the target volume. Different weightings were assigned to different beam size to encourage larger beam size and shorter overall treatment time. Results: The proposed method was tested by generating virtual beams for numerical phantoms and tumors of previous SRS patients (treated using Radionics xKnife system). For most target volumes (20 out of 23) that require multiple isocenters, with the proposed approach one was able to determine the size and location of isocenters that covers >95% of the target volume with the 50% isodose surface. The overall conformity index was less than 1.33. The average time of the procedure was about 30 seconds. Conclusions: The proposed method can be used to efficiently determine the isocenters and the beam sizes for a Gamma knife unit to enhance the efficiency of the treatment planning process for SRS patients with irregular shape tumors.
All Science Journal Classification (ASJC) codes
- Radiology Nuclear Medicine and imaging