基于影像组学的非小细胞肺癌淋巴结转移预测

王超; 刘侠; 董迪; 臧丽亚; 刘再毅; 梁长虹; 田捷

doi:10.16383/j.aas.c160794

基于影像组学的非小细胞肺癌淋巴结转移预测

doi: 10.16383/j.aas.c160794

王超^1,,
刘侠^1,,
董迪^2,,
臧丽亚^2,,
刘再毅^3,,
梁长虹^3,,
田捷^2, ,

1.
哈尔滨理工大学自动化学院哈尔滨 150080
2.
中国科学院自动化研究所北京 100190
3.
广东省人民医院广州 510080

基金项目:

国家重点研发项目 2017YFC1308701

黑龙江省然科学基金 12541105

国家自然科学基金 61231004

国家自然科学基金 81227901

国家自然科学基金 81501616

国家重点研发项目 2016YFC0103803

国家自然科学基金 81771924

北京市科学技术委员会Commission Z161100002616022

国家重点研发项目 2017YFA0205200

国家自然科学基金 81301346

国家自然科学基金 61672197

国家重点研发项目 2017YFC1308700

国家自然科学基金 81527805

国家重点研发项目 2017YFC1309100

北京市科学技术委员会Commission Z171100000117023

国家自然科学基金 81671851

黑龙江省然科学基金 F201311

详细信息

作者简介:
王超  中国科学院自动化研究所和哈尔滨理工大学自动化学院联合培养硕士研究生.主要研究方向为医学图像处理和模式识别.E-mail:wangchao2015@ia.ac.cn

刘侠  哈尔滨理工大学自动化学院教授.2006年获得哈尔滨工程大学自动化学院博士学位.主要研究方向为模式识别与智能系统.E-mail:liuxia@hrbust.edu.cn

董迪  中国科学院自动化研究所副研究员.2013年获得中国科学院自动化研究所博士学位.主要研究方向为影像组学.E-mail:di.dong@ia.ac.cn

臧丽亚：臧亚丽  中国科学院自动化研究所副研究员.2013年获得中国科学院自动化研究所博士学位.主要研究方向为影像组学.E-mail:yali.zang@ia.ac.cn

刘再毅  广东省人民医院主任医师.2004年获得四川大学博士学位.主要研究方向为腹部影像诊断和分子影像学.E-mail:zyliu@163.com

梁长虹  广东省人民医院影像医学部主任.1998年获得广东省心血管病研究所博士学位.主要研究方向为腹部影像诊断和分子影像学.E-mail:cjr.lchh@vip.163.com

通讯作者:
田捷中国科学院自动化研究所研究员.1992年获得中国科学院自动化研究所博士学位.主要研究方向为模式识别, 医学图像处理和分子影像.本文通信作者.E-mail:tian@ieee.org

计量
- 文章访问数: 1875
- HTML全文浏览量: 395
- PDF下载量: 395
- 被引次数: 0
出版历程
- 收稿日期: 2016-11-30
- 录用日期: 2017-08-17
- 刊出日期: 2019-06-20

Radiomics Based Lymph Node Metastasis Prediction in Non-small-cell Lung Cancer

WANG Chao^1
,,
LIU Xia^1
,,
DONG Di^2
,,
ZANG Li-Ya^2
,,
LIU Zai-Yi^3
,,
LIANG Chang-Hong^3
,,
TIAN Jie^{2
, ,}

1.
School of Automation, Harbin University of Science and Technology, Harbin 150080
2.
Institute of Automation, Chinese Academy of Sciences, Beijing 100190
3.
Guangdong General Hospital, Guangzhou 510080

Funds:

National Key Research and Development Program of China 2017YFC1308701

Natural Science Foundation of Heilongjiang Province 12541105

National Natural Science Foundation of China 61231004

National Natural Science Foundation of China 81227901

National Natural Science Foundation of China 81501616

National Key Research and Development Program of China 2016YFC0103803

National Natural Science Foundation of China 81771924

Beijing Municipal Science and Technology Commission Z161100002616022

National Key Research and Development Program of China 2017YFA0205200

National Natural Science Foundation of China 81301346

National Natural Science Foundation of China 61672197

National Key Research and Development Program of China 2017YFC1308700

National Natural Science Foundation of China 81527805

National Key Research and Development Program of China 2017YFC1309100

Beijing Municipal Science and Technology Commission Z171100000117023

National Natural Science Foundation of China 81671851

Natural Science Foundation of Heilongjiang Province F201311

More Information

Author Bio:
Master student at Institute of Automation, Chinese Academy of Sciences and School of Automation, Harbin University of Science and Technology. His research interest covers imaging processing and pattern recognition

Professor at the School of Automation, Harbin University of Science and Technology. He received his Ph. D. degree from Harbin Engineering University in 2006. His research interest covers pattern recognition and intelligent systems

Associate professor at Institute of Automation, Chinese Academy of Sciences. He received his Ph. D. degree from Chinese Academy of Sciences in 2013. His research interest is radiomics

Associate professor at Institute of Automation, Chinese Academy of Sciences. She received her Ph. D. degree from Chinese Academy of Sciences in 2013. Her research interest is radiomics

Director of Guangdong General Hospital. He received his M. D. degree from SiChuan University in 2004. His research interest covers abdominal imaging diagnosis and molecular imaging

Director of Imaging Medicine Department of Guangdong General Hospital. He received his M. D. degree from Guangdong Cardiovascular Disease Research Institute in 1998. His research interest covers abdominal imaging diagnosis and molecular imaging

Corresponding author: TIAN Jie Professor at Institute of Automation, Chinese Academy of Sciences. He received his Ph. D. degree from Institute of Automation, Chinese Academy of Sciences in 1992. His research interest covers pattern recognition, medical image processing and molecular imaging. Corresponding author of this paper

摘要

摘要: 在非小细胞肺癌的临床诊疗中，淋巴结是否转移对于医生制定手术方案有重要指导意义.但是目前临床上缺乏能够安全准确地预测非小细胞肺癌淋巴结转移的方法.本文应用影像组学方法对肺部CT影像进行定量分析来实现对非小细胞肺癌淋巴结是否转移的预测.从广东省人民医院收集了564例满足分析要求的非小细胞肺癌病例数据，并从每例CT影像中提取了386个定量影像特征，包括肿瘤的三维形状特征，表面纹理特征，Gabor特征以及小波特征：然后利用Lasso logistic regression（LLR）来构造非小细胞肺癌淋巴结转移的影像组学标签（Rad-score），并结合临床信息进行多元分析，构造了诺模图个性化预测模型.其中，LLR淋巴结转移预测模型性能在训练集上AUC为0.710，测试集AUC为0.712：在个性化诺模图上，用所有数据进行预测，得到C-index为0.724（95% CI：0.678~0.770），在一致性曲线上表现较佳，可为临床决策提供有价值的信息.
- 影像组学 /
- 淋巴结转移 /
- 套索逻辑斯特回归 /
- 诺模图 /
- 一致性曲线
Abstract: In the clinical diagnosis and treatment of non-small cell lung cancer, it makes great sense for guiding doctors to make operation plans according to lymph node metastasis status. However, a current weak point in clinical is the lack of a method that can be used to predict lymph node metastasis for non-small cell lung cancer safely and accurately. In this article, radiomic method was applied to the lung CT images to achieve quantitative analysis for the prediction of lymph node metastasis for non-small cell lung cancer. We collected 564 non-small cell lung cancer cases that could satisfy the data recruitment from Guangdong General Hospital, from which 386 quantitative radiomic features were extracted each, including the tumor's three-dimensional shape features, texture features, Gabor features and wavelet features. Then, Lasso logistic regression (LLR) was used to construct the radiomic signature (Rad-score) for the lymph node metastasis of NSCLC. With multivariate analysis of clinical information, the customized prediction nomogram model was built. The performance of the LLR model was shown to have an AUC of 0.710 in the training set and 0.712 in the validation set. Our nomogram model had a C-index of 0.724 (95% CI:0.678~0.770) and performed well on the consistency, providing valuable information for clinical decision-making.
- Radiomics /
- lymph node metastasis /
- Lasso logistic regression /
- nomogram /
- calibration curve
注释:

1) 本文责任编委朱朝

HTML全文

图 1 数据筛选流程图

Fig. 1 Data filtering flow chart

下载: 全尺寸图片幻灯片

图 2 三维病灶的分割

Fig. 2 3D tumor segmentation

下载: 全尺寸图片幻灯片

图 3 淋巴结转移预测模型构造图

Fig. 3 Structure of lymph node metastasis prediction model

下载: 全尺寸图片幻灯片

图 4 $\lambda $与变量数目对应走势

Fig. 4 The trend of the parameters and the number of variables

下载: 全尺寸图片幻灯片

图 5 系数随$\lambda $参数变化图

Fig. 5 The coefficient changes with the parameters

下载: 全尺寸图片幻灯片

图 6 测试集ROC曲线

Fig. 6 ROC curve of test set

下载: 全尺寸图片幻灯片

图 7 验证诺模图

Fig. 7 Verifies the nomogram

下载: 全尺寸图片幻灯片

图 8 一致性曲线

Fig. 8 Consistency curves

下载: 全尺寸图片幻灯片

表 1 训练集和测试集病人的基本情况

Table 1 Basic information of patients in the training set and test set

基本项	训练集($N=400$)	$P$值	测试集($N=164$)	$P$值
性别	男	144 (36 %)	0.896	78 (47.6 %)	0.585
性别	女	256 (64 %)		86 (52.4 %)
吸烟	是	126 (31.5 %)	0.030*	45 (27.4 %)	0.081
吸烟	否	274 (68.5 %)		119 (72.6 %)
EGFR	缺失	36 (9 %)		4 (2.4 %)
	突变	138 (34.5 %)	$ < $0.001*	67 (40.9 %)	0.112
	正常	226 (56.5 %)		93 (56.7 %)

下载: 导出CSV

表 2 Lasso选择得到的参数

Table 2 Parameters selected by Lasso

Lasso选择的参数	含义	数值	$P$值
$V0$	截距项	2.079115
$V179$	横向小波分解90度共生矩阵Contrast特征(Contrast_2_90)	0.0000087	< 0.001***
$V230$	横向小波分解90度共生矩阵Entropy特征(Entropy_3_180)	$-$3.573315	< 0.001***
$V591$	表面积与体积的比例(Surface to volume ratio)	$-$1.411426	< 0.001***

下载: 导出CSV

表 3 不同方法对比结果

Table 3 Comparison results of different methods

方法	训练集(AUC)	测试集(AUC)	召回率
LLR	0.710	0.712	0.75
SVM	0.698	0.654	0.75
NB	0.718	0.681	0.74

下载: 导出CSV

参考文献(16)

[1]	Jemal A, Bray F, Center M M, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer Journal for Clinicians, 2011, 61(2):69-90 doi: 10.3322/caac.v61:2
[2]	Howlader N, Noone A M, Krapcho M, Neyman N, Aminou R, Waldron W, et al. SEER cancer statistics review, 1975-2008. Bethesda, MD:National Cancer Institute[Online], available:http://seer.cancer.gov/csr/1975_2008/, August 1, 2015
[3]	Glasgow S C, Bleier J I S, Burgart L J, Finne C O, Lowry A C. Meta-analysis of histopathological features of primary colorectal cancers that predict lymph node metastases. Journal of Gastrointestinal Surgery, 2012, 16(5):1019-1028 doi: 10.1007/s11605-012-1827-4
[4]	Smith A J, Driman D K, Spithoff K, Hunter A, McLeod R S, Simunovic M, et al. Guideline for optimization of colorectal cancer surgery and pathology. Journal of Surgical Oncology, 2010, 101(1):5-12 doi: 10.1002/(ISSN)1096-9098
[5]	Toiyama Y, Inoue Y, Shimura T, Fujikawa H, Saigusa S, Hiro J, et al. Serum angiopoietin-like protein 2 improves preoperative detection of lymph node metastasis in colorectal cancer. Anticancer Research, 2015, 35(5):2849-2856 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=257db21275433b4ca2fe9da0f66d802c
[6]	Song J, Liu Z, Zhong W, Ma Z, Dong D, Liang C, et al. Nonsmall cell lung cancer:quantitative phenotypic analysis of CT images as a potential marker of prognosis. Scientific Reports, 2016, 6(1):Article No. 38282 doi: 10.1038/srep38282
[7]	Gillies R J, Kinahan P E, Hricak H. Radiomics:images are more than pictures, they are data. Radiology, 2016, 278(2):563-577 doi: 10.1148/radiol.2015151169
[8]	Kuo M D, Gollub J, Sirlin C B, Ooi C, Chen X. Radiogenomic analysis to identify imaging phenotypes associated with drug response gene expression programs in hepatocellular carcinoma. Journal of Vascular and Interventional Radiology, 2007, 18(7):821-830 doi: 10.1016/j.jvir.2007.04.031
[9]	Lambin P, Rios-Velazquez E, Leijenaar R, Carvalho S, van Stiphout R G, Granton P, et al. Radiomics:extracting more information from medical images using advanced feature analysis. European Journal of Cancer, 2012, 48(4):441-446 doi: 10.1016/j.ejca.2011.11.036
[10]	Sauerbrei W, Royston P, Binder H. Selection of important variables and determination of functional form for continuous predictors in multivariable model building. Statistics in Medicine, 2007, 26(30):5512-5528 doi: 10.1002/(ISSN)1097-0258
[11]	Song J D, Yang C Y, Fan L, Wang K, Yang F, Liu S Y, et al. Lung lesion extraction using a toboggan based growing automatic segmentation approach. IEEE Transactions on Medical Imaging, 2015, 35(1):337-353 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=25538c05c98e59d5f264491e0cd769c9
[12]	Aerts H J W L, Velazquez E R, Leijenaar R T H, Parmar C, Grossmann P, Carvalho S, et al. Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach. Nature Communications, 2014, 5:Article No. 4006 doi: 10.1038/ncomms5006
[13]	Vickers A J, Cronin A M, Elkin E B, Gonen M. Extensions to decision curve analysis, a novel method for evaluating diagnostic tests, prediction models and molecular markers. BMC Medical Informatics and Decision Making, 2008, 8:Article No. 53 doi: 10.1186/1472-6947-8-53
[14]	Zhang L J, Yang T B, Jin R, Zhou Z H. Stochastic proximal gradient descent for nuclear norm regularization. arXiv:1511.01664v2, January 1, 2015
[15]	Huang Y Q, Liang C H, He L, Tian J, Liang C S, Chen X, et al. Development and validation of a radiomics nomogram for preoperative prediction of lymph node metastasis in colorectal cancer. Journal of Clinical Oncology, 2016, 34(18):2157-2164 doi: 10.1200/JCO.2015.65.9128
[16]	Liang W H, Zhang L, Jiang G N, Wang Q, Liu L X, Liu D R, et al. Development and validation of a nomogram for predicting survival in patients with resected non-small-cell lung cancer. Journal of Clinical Oncology, 2015, 33(8):861-869 doi: 10.1200/JCO.2014.56.6661