[1]王翰林,刘嘉城,王清莹,等.深度强化学习在直肠癌IMRT自动计划的应用[J].中国医学物理学杂志,2022,39(1):1-8.[doi:DOI:10.3969/j.issn.1005-202X.2022.01.001]
　WANG Hanlin,LIU Jiacheng,WANG Qingying,et al.Application of deep reinforcement learning in automatic IMRT planning for rectal cancer[J].Chinese Journal of Medical Physics,2022,39(1):1-8.[doi:DOI:10.3969/j.issn.1005-202X.2022.01.001]
点击复制

深度强化学习在直肠癌IMRT自动计划的应用()
分享到：

《中国医学物理学杂志》[ISSN:1005-202X/CN:44-1351/R]

卷:
39卷
期数:
2022年第1期
页码:
1-8
栏目:
医学放射物理
出版日期:
2022-01-17

文章信息/Info

Title:
Application of deep reinforcement learning in automatic IMRT planning for rectal cancer
文章编号:
1005-202X（2022）01-0001-08
作者:
王翰林; 刘嘉城; 王清莹; 岳海振; 杜乙; 张艺宝; 王若曦; 吴昊
北京大学肿瘤医院暨北京市肿瘤防治研究所放疗科/恶性肿瘤发病机制及转化研究教育部重点实验室， 北京 100142
Author(s):
WANG Hanlin;  LIU Jiacheng;  WANG Qingying;  YUE Haizhen;  DU Yi;  ZHANG Yibao;  WANG Ruoxi;  WU Hao
Key Laboratory of Carcinogenesis and Translational Research (Ministry of Education)/Department of Radiotherapy, Peking University Cancer Hospital & Institute, Beijing 100142, China
关键词:
直肠癌; 自动优化; 深度强化学习; 脚本应用程序接口; 优化调整决策网络
Keywords:
Keywords: rectum cancer automatic optimization deep reinforcement learning Eclipse scripting application programming interface optimization adjustment policy network
分类号:
R318；R811.1
DOI:
DOI:10.3969/j.issn.1005-202X.2022.01.001
文献标志码:
A
摘要:
目的：对于调强放疗（IMRT）计划，优化过程较为耗时，且计划的质量取决于计划人员的经验和时间，本文探讨并实现一种无监督IMRT自动优化的方案，使其能够模拟人工操作方式进行治疗计划优化。方法：本研究基于深度强化学习框架，提出一种优化调整决策网络（OAPN）自动化计划优化的方法。利用Varian Eclipse 15.6 TPS的脚本应用程序接口（ESAPI）实现OAPN与TPS之间的交互，以剂量体积直方图作为信息输入，通过强化学习的训练方式，OAPN学习TPS中目标参数的调整策略，从而逐步改善并获得较高质量的计划。实验从临床数据库中选取18例既往已完成治疗的直肠癌病例，其中5例计划案例用于OAPN网络训练，其余13例计划案例用于评估训练后OAPN的可行性与有效性，引入第三方计划评分工具来衡量计划质量。结果：用于测试的13例直肠癌计划，使用统一的初始优化目标参数（OOPs）所获得的平均得分为（45.53±4.58）分（计划得分上限值为110），经过OAPN对OOPs调整后计划所获得的平均得分为（88.67±6.74）分。结论：OAPN借助ESAPI实现与TPS之间数据交互，通过深度强化学习的方式形成行为价值策略，经过训练后的OAPN可以对目标参数进行高效率的调整，同时获得较高质量计划。
Abstract:
Abstract: Objective The optimization of intensity-modulated radiotherapy planning is often time-consuming, and the plan quality depends on the experience of the planner and the available planning time. An unsupervised automatic intensity-modulated radiotherapy optimization procedure is discussed and implemented to simulate the human operation during the whole optimization process. Methods Based on the framework of deep reinforcement learning (DRL), an optimization adjustment policy network (OAPN) was proposed to automate the process of treatment planning optimization. The scripting application programming interface (ESAPI) of Varian Eclipse 15.6 TPS was used to realize the interaction between OAPN and TPS. Taking dose-volume histogram as the information input, OAPN learned the adjustment strategy of objective parameters in TPS by the training mode of reinforcement learning, so as to gradually improve and obtain high-quality plans. A total of 18 cases of rectum cancer which had completed treatment were selected from the clinical database. Five of the cases were used for OAPN training, and the remaining 13 for evaluating the feasibility and effectiveness of OAPN after training. Finally, a third-party scoring tool was used to evaluate plan quality. Results The average score of 13 tested plans using uniform initial optimization objective parameters (OOPs) was 45.53±4.58 (the upper limit value was 110). After adjusting OOPs by OAPN, the average plan score was 88.67±6.74. Conclusion OAPN can realize the data interaction with TPS through ESAPI, and form an action-value strategy through DRL. After training, OAPN can efficiently adjust OOPs and obtain a high-quality plan.

相似文献/References:

[1]任信信,戴建荣,张 岳,等.基于专家库的γ射线立体定向放疗计划优化方法[J].中国医学物理学杂志,2015,32(03):301.[doi:10.3969/j.issn.1005-202X.2015.03.001]
[2]张武哲,陈志坚,彭 逊,等.不同角度固定野动态调强在直肠癌术前放疗中的剂量学比较[J].中国医学物理学杂志,2014,31(02):4735.[doi:10.3969/j.issn.1005-202X.2014.02.004]
[3]李忠伟,张九堂,吴智理,等.R620-ACF型腹部碳纤维托架对直肠癌调强放射治疗剂量的影响[J].中国医学物理学杂志,2014,31(05):5148.[doi:10.3969/j.issn.1005-202X.2014.05.012]
[4]吴虹,翟福山,刘明,等.应用CBCT研究膀胱体积大小对直肠癌术后放疗剂量学的影响[J].中国医学物理学杂志,2013,30(04):4253.[doi:10.3969/j.issn.1005-202X.2013.04.010]
[5]张艺宝,吴昊,李莎,等.临床前验证与几何对比分析基于图谱库的危及器官自动勾画[J].中国医学物理学杂志,2015,32(06):761.[doi:doi:10.3969/j.issn.1005-202X.2015.06.001]
　[J].Chinese Journal of Medical Physics,2015,32(1):761.[doi:doi:10.3969/j.issn.1005-202X.2015.06.001]
[6]王沛沛,刘哲铭,李彩虹,等.容积旋转调强与固定野动态调强在直肠癌新辅助放疗中的剂量学比较[J].中国医学物理学杂志,2015,32(06):901.[doi:doi:10.3969/j.issn.1005-202X.2015.06.030]
　[J].Chinese Journal of Medical Physics,2015,32(1):901.[doi:doi:10.3969/j.issn.1005-202X.2015.06.030]
[7]姚文良,许波,戚金凤,等.直肠癌术前不同照射技术剂量学比较[J].中国医学物理学杂志,2016,33(3):317.[doi:DOI:10.3969/j.issn.1005-202X.2016.03.021]
　[J].Chinese Journal of Medical Physics,2016,33(1):317.[doi:DOI:10.3969/j.issn.1005-202X.2016.03.021]
[8]黎旦,宾石珍,程品晶,等.非晶硅电子射野影像系统与Arc CHECK在直肠癌容积旋转调强剂量验证中的应用[J].中国医学物理学杂志,2016,33(7):674.[doi:10.3969/j.issn.1005-202X.2016.07.006]
　[J].Chinese Journal of Medical Physics,2016,33(1):674.[doi:10.3969/j.issn.1005-202X.2016.07.006]
[9]王贝,全红,邱杰,等. 容积旋转调强技术用于直肠癌时最佳机架角度设置研究[J].中国医学物理学杂志,2016,33(9):898.[doi:10.3969/j.issn.1005-202X.2016.09.006]
　[J].Chinese Journal of Medical Physics,2016,33(1):898.[doi:10.3969/j.issn.1005-202X.2016.09.006]
[10]孙永健,曹洋森,代智涛,等.Monaco设计直肠癌固定机架角调强和容积旋转调强治疗的剂量[J].中国医学物理学杂志,2017,34(2):161.[doi:10.3969/j.issn.1005-202X.2017.02.010]
　Dosimetric comparison of seven- field intensity modulated radiotherapy and volumetric modulated arc therapy for rectal cancer designed with Monaco[J].Chinese Journal of Medical Physics,2017,34(1):161.[doi:10.3969/j.issn.1005-202X.2017.02.010]

备注/Memo

备注/Memo:
【收稿日期】2021-07-10 【基金项目】国家重点研发计划（2019YFF01014405）；北京市医管局培育计划（PX2019042）；北京市自然科学基金（1202009）；国家自然科学基金（12005007）；中央高校基本科研业务费/北京大学临床医学+X青年专项（PKU2020LCXQ019） 【作者简介】王翰林，硕士，研究方向：放射治疗物理学，E-mail: Wanghanlins@163.com 【通信作者】吴昊，高级工程师，研究方向：医学物理，E-mail: hao.wu@bjcancer.org
更新日期/Last Update: 2022-01-17