Prior knowledge-based multi-objective optimization method for intensity-modulated radiotherapy planning of head and neck cancer(PDF)
《中国医学物理学杂志》[ISSN:1005-202X/CN:44-1351/R]
- Issue:
- 2021年第3期
- Page:
- 265-276
- Research Field:
- 医学放射物理
- Publishing date:
Info
- Title:
- Prior knowledge-based multi-objective optimization method for intensity-modulated radiotherapy planning of head and neck cancer
- Author(s):
- LIU Jingguang1; 2; LIU Guocai1
- 1. School of Electrical and Information Engineering, Hunan University, Changsha 410082, China 2. School of Computational Science and Electronics, Hunan Institute of Engineering, Xiangtan 411104, China
- Keywords:
- Keywords: head and neck cancer intensity modulated radiotherapy multi-criteria optimization Pareto frontier approximation
- PACS:
- R318；R811.1
- DOI:
- DOI:10.3969/j.issn.1005-202X.2021.03.001
- Abstract:
- Abstract: Intensity-modulated radiotherapy (IMRT) planning of head and neck cancer is a complex multi-criteria optimization with more than 20 criteria and thousands of decision variables. So far, there is no effective method that can determine the accurate Pareto optimal solution set. Herein an effective approximate solution method is proposed. Based on the prior knowledge of the dose distribution of 71 patients with head and neck cancer, R-mode clustering analysis is conducted on the linear EUD of all organs-at-risk according to correlation coefficients, and all organs-at-risk are classified into 3 categories. Secondly, the mean value of linear EUD of organs-at-risk of the same category is regarded as the criteria, and then is combined with the constraints of tumor target to construct a multi-criteria optimization model with a lower dimension. Finally, the enhancement sandwich algorithm based on corner solutions is used to approximate the low-dimensional and partial Pareto frontier, and an IMRT plan is selected through the visual navigation method quickly. Compared with the traditional multi-criteria optimization method, this method can significantly reduce the number of criteria and remarkably simplify Pareto frontier. Experiment results show that the number of criteria for the multi-criteria optimization model of IMRT plan for head and neck cancer is reduced from 20-21 to 6-7, and that the approximate Pareto frontier can be determined effectively, an IMRT plan can be derived with the same quality as the clinical plan, with some indexes even better.
Last Update: 2021-03-30