SU‐E‐T‐251: 3D Collision Viewer for Proton Therapy Treatment Planning

• Published in: Medical Physics 2013-06, Vol.40 (6), p.262-262
• Main Authors: Hojo, Y, Gillin, M, Suzuki, K, Wu, R, Zhu, X
• Format: Article
• Language: English
• Subjects: Computer software; Medical treatment planning; Proton therapy; Protons; Therapeutics
• ISSN: 0094-2405; 2473-4209
• DOI: 10.1118/1.4814686

Purpose: In proton treatment, in order to reduce beam penumbra, it is desirable to place the treatment snout as close as possible to the patient skin. However, collision between treatment couch/patient and snout is not predicted by current treatment planning system. We developed and validated an interactive 3D Collision Viewer for patient clearance check during treatment planning. Methods: The 3D Collision Viewer calculates the shortest distance between two major groups of objects, couch group and nozzle group. The couch group includes a couch and a patient and the nozzle group includes a nozzle, a snout and extensions for snout. The software displays the object coordinates and the shortest distance in the 3D Viewer. The tool was developed with a video game development framework to realize 3D interactive performance. Surface coordinates of equipment including the patient were defined relative to the iso‐center and created by commonly used CAD software. The equipment coordinates and the shortest distance are interactively updated on any change of equipment setting by the user. The video game framework takes care of equipment surface coordinate motion according to each equipment setting. The shortest distance was obtained by comparing distances between all combination of nozzle meshes and couch meshes. To confirm the numerical accuracy of the shortest distance calculation, the distance between nozzle and 2 cm diameter sphere was compared with the expected distance. Results: Numerical accuracy of the distance calculation is better than about 1/40 of mesh size of the sphere. The accuracy mainly depends on the surface data preparation. Average calculation time for a sample configuration, which had 156 vertices and 94 meshes for the nozzle and 2707 vertices and 3758 meshes for couch, was less than 0.6 seconds. Conclusion: 3D Collision Viewer is capable of detecting potential collision during treatment planning.