《放射性肺损伤与三维剂量分布的相关性研究》一文的相关的文档和代码等资料
如果本资料库代码或文章对您有帮助，请引用文章(http://zgyxwlxzz.paperopen.com/oa/DArticle.aspx?type=view&id=202008042#)：
[1] 范晓鸿,董可,杨银,张建平.放射性肺损伤与三维剂量分布的相关性研究[J].**医学物理学杂志,2021,38(6):692-698.[doi:DOI:10.3969/j.issn.1005-202X.2021.06.007]
FAN Xiaohong,DONG Ke,YANG Yin,et al.Correlation between radiation-induced lung injury and 3D dose distribution[J].Chinese Journal of Medical Physics,2021,38(6):692-698.[doi:DOI:10.3969/j.issn.1005-202X.2021.06.007]
放射性肺损伤与三维剂量分布的相关性研究
范晓鸿1，董 可2，杨 银1，张建平1
（1.湘潭大学数学与计算科学学院，湖南 湘潭 411105；2. 湘潭市中心医院放疗科，湖南 湘潭 411101）
【摘要】目的：应用影像组学开展肺癌患者放射性肺损伤与放疗计划剂量分布相关性分析的研究，筛选出与放疗CT影像改变相关联的特征。方法：选取放疗后参与随访的肺癌患者18例，通过计划和诊断CT之间的配准变形场得到相应诊断CT的剂量分布；将对侧肺与同侧肺及各设定剂量区间内的肺实质分割为相应感兴趣区并提取其影像组学特征，计算治疗前后特征改变量，比较特征改变量在不同剂量区间、时间节点下的差异并筛选特征。结果：（1）筛选出28个特征的筛查剂量响应曲线在放疗第2周与放疗第4、6及10周有较明显差异，而放疗第4周与第6、10周整体差异较小；（2）筛选出27个特征的肺损伤剂量响应曲线在低剂量区（020 Gy）或高剂量区（4565 Gy）响应较强烈，且随放疗后时间越久特征改变量越大；（3）筛选出15个特征在对侧肺与同侧肺的特征改变量有较明显分层现象。结论：通过分析肺癌患者放疗前后影像组学特征改变量，证实了放射性肺损伤与三维剂量分布具有显著相关性。追踪这些特征的动态变化，对评估放射性肺损伤随时间及位置的发展规律有潜在优势．
关键词：肺癌；剂量分布；影像组学；特征筛选；放射性肺损伤
The study of correlation between radiation-induced lung injury and 3D dose distribution
FAN Xiaohong1，DONG Ke2，YANG Yin1，ZHANG Jianping1
(1. School of Mathematics and Computational Science, Xiangtan University, Xiangtan, 411105, China;
2. Department of Radiotherapy, Xiangtan Central Hospital, Xiangtan, 411101, China)
Abstract：Objective: To analyze the correlation between radiation-induced lung injury and 3D dose distribution in lung cancer patients by using radiomics technique, and identify the features related to CT image texture change. Methods: A total of 18 lung cancer patients who were followed up after radiotherapy were selected in this study. The dose distribution of diagnosed CT was obtained from the planned CT dose distribution’s deformation by using the registration deformation field between planned CT and diagnosed CT. The ipsilateral and contralateral lungs and the lung parenchyma in each dose interval were considered as the region of interest (ROI), the radiomic features were extracted from ROI. The features’ changes between before radiation and after treatment were compared at different dose interval and time, the distinct difference features were selected. Results:（1）For dose-response curve of screening, 28 features who were significantly different between the 2 week of radiotherapy and the 4, 6 and 10 weeks of radiotherapy were selected , while the overall differences between the 4 week of radiotherapy and the 6 and 10 weeks of radiotherapy were relatively small.（2）For dose-response curve of lung injury, 27 features who were stronger responded in the low dose area (0-20Gy) or the high dose area (45-65Gy) were selected, and the features’ changes increased with the time after radiotherapy.（3）For the ipsilateral and contralateral lungs, 15 features who were obviously stratified were selected. Conclusions: By analyzing the features’ changes of lung cancer patients between before radiation and after treatment, the significant correlation between radiation lung injury and three-dimensional dose distribution was confirmed. Monitoring the dynamic changes of these features has potential advantages in assessing the pattern of radioactive lung injury over time.
Key words：Dose distribution; Radiomics; Feature selection; Radiation-induced lung injury; Lung cancer
图1 影像组学分析流程
Fig.1 Flowchart of radiomics analysi
The code is provided to support reproducible research. If the code is giving syntax error in your particular python configuration or some files are missing then you may open an issue or directly email me at fanxiaohong@smail.xtu.edu.cn