| 研究概要 |
The principal investigator developed four image analysis approaches including computational anatomy models for treatment planning in computer-assisted particle therapy. The inadequate outcome of particle therapy could be caused by several uncertainties such as inter-observer variation of target (i.e., cancerous lesion) delineations, inter-fractional position and shape variations of the targets, and patient setup errors, and so on. Furthermore, sharp dose distributions could greatly deteriorate due to the setup errors and/or organ motions depending on particle beam direction. To mitigate these issues, we have developed mainly four image analysis approaches, i.e., (1) an automated extraction of tumors using a machine learning classifier with knowledge of radiation oncologists, (2) a framework for construction of statistical clinical target volume models with inter-observer delineation and inter-fractional variations in radiation treatment planning, (3) a quantitative evaluation approach of the robustness of beam directions against patient setup errors in the particle therapy, and (4) a similar case based treatment planning system. These technologies can be organized for construction of computer-assisted particle therapy systems, which can support radiation therapy practitioners.
|
