北京生物医学工程

多层线性模型在肺癌患者肿瘤进展中的应用

Application of hierarchical linear model in thedevelopment of tumor in lung cancer

作者：孙延辉方丽英王普

单位：北京工业大学电子信息与控制工程学院(北京100124)

分类号：R318.04

出版年·卷·期（页码）：2017·36·1（77-81）

摘要：

目的通过分析肺癌患者以随访方式收集的一组纵向数据，研究肺癌患者肿瘤大小随时间变化的趋势，同时分析这种变化趋势的个体症状之间的差异性。方法本文采用多层线性模型，对肿瘤发展的纵向数据进行了深入研究。首先对213例非小细胞肺癌患者进行了2年的随访记录，记录患者性别、是否抽烟、症状（咳嗽、痰中带血、胸痛、发热、神疲乏力）和肿瘤大小数据，然后通过建立多层线性模型分析肿瘤大小与患者性别、是否抽烟及患者所表现出症状之间的关系。结果通过对肺癌患者数据建立多层线性模型并进行数据分析，发现患者性别、是否抽烟与肿瘤的大小有密切关系，且不同个体特征对肿瘤大小的影响存在差异。结论多层线性模型可以分析肺癌患者肿瘤发展的纵向数据，得到肿瘤大小的发展趋势，并可以进一步分析这种趋势在个体之间的差异性。

Objective By analyzing a set of longitudinal data collected from patients with lung cancer during follow-up, the trend of tumor size in patients with lung cancer was studied, and the differences between individual symptoms of this trend were analyzed. Methods In this paper, hierarchical linear model (HLM) was used to study the longitudinal data of tumor development. Firstly, we tracked 213 cases of non-small cell lung cancer patients for up to two years and recorded patient gender, the habit of smoking and some other related symptoms (cough, bloody sputum, chest pain, fever, god fatigue) and tumor size data. Then, through the establishment of hierarchical linear model, we analyzed the relationships between tumor size and gender, smoking, symptoms of the patients. Results By establishing HLM and analyzing the longitudinal data of cancer, we found that close relationships existed between the tumor size and the gender, the habit of smoking. Additionally, the effects of different individual characteristics to their tumor size were significantly different. Conclusions HLM can help to dig the longitudinal data of tumor development and obtain the development trend of tumor size, which can be further analyzed.

参考文献：

［1］Liang KY, Zeger ST.Longitudinal data analysis using generalized linearmodels［J］. Biometrics,1986, 73(1): 13.

［2］Diggle P. Analysis of Longitudinal Data［M］. 2nd ed. New York: Oxford University Press,2002: 379.

［3］Tsiatis AA, Davidian M. Joint modeling of longitudinal and time-to-event data: an overview［J］. Statistica Sinica, 2004, 14(3): 809-834.

［4］Plumper T, Troeger VE, Manow P. Panel data analysis in comparative politics: linking method to theory［J］. European Journal of Political Research,2005, 44(2): 327-354.

［5］毛新娜, 王秀旺. 纵向数据常用参数分析方法比较［J］. 新乡学院学报:自然科学版, 2009, (2): 10-14.

Mao Xinna, Wang Xiuwang. Comparison among the common parametric methods in studying longitudinal data ［J］. Xinxiang University: Natural Science Edition,2009,(2): 10-14.

［6］刘红云，孟庆茂. 纵向数据分析方法［J］. 心理科学进展,2003,11(5): 586-592.

Liu Hongyun, Meng Qingmao. A review on longitudinal data analysis method and it’s development［J］. Advances in Psychological Science, 2003, 11(5):586-592.

［7］Raudenbush SW, Bryk AS.Hierarchical Linear Models: Applications and Data Analysis Methods［M］. New York: Sage Publications, 2002.

［8］盖笑松，张向葵. 多层线性模型在纵向研究中的运用［J］. 心理科学,2005,28(2): 429-431.

Gai Xiaosong, Zhang Xiangkui. The application of multilevel model in longitudinal research［J］. Psychological Science, 2005,28(2): 429-431.

［9］王济川，谢海义，姜宝法. 多层统计分析模型：方法与应用［M］.北京: 高等教育出版社, 2008: 81-127.

Wang Jichuan,Xie Haiyi,Jiang Baofa. Multilevel Statistical Model Analysis: Methods and Applications ［M］. Beijing: Higher Education Press, 2008: 81-127.

［10］Teo T. Basic concepts of Hierarchical Linear Modelling (HLM) for educational technology research［J］. British Journal of Educational Technology, 2011, 42(3): E51-E53.

［11］索文莉. 线性混合模型在艾滋病疗效预测中的应用［D］. 天津:天津工业大学, 2008.

Suo Wenli.Linear mixed model to predict the effectiveness of AIDS in the application ［D］.Tianjin:Tianjin University of Technology, 2008.

［12］游达明, 杨晓辉, 杨立,等. 基于多层线性模型的就业影响因素研究［J］. 统计与决策, 2011,(3):41-44.

［13］鲁庆, 穆志纯. 多层线性模型在碳钢土壤腐蚀规律中的应用研究［J］. 北京科技大学学报, 2013,35(11): 1458-1464.

Lu Qing, Mu Zhichun. Application study of hierarchical linear models in carbon steel corrosion in soils ［J］. Beijing University of Science and Technology, 2013, 35(11):1458-1464.

［14］叶青青, 刘艳芳, 刘耀林,等. 基于多层线性模型的湖北省县域建设用地集约利用影响因素研究［J］. 中国土地科学, 2014,8(8):33-39.

Ye Qingqing, Liu Yanfang, Liu Yaolin, et al. Study on the impact factors of construction land use intensity at county level in hubei province based on hierarchical linear model［J］.China Land Science, 2014 ,8(8): 33-39.

服务与反馈：

【文章下载】【加入收藏】

提示：您还未登录，请登录！点此登录