期刊文献+

基于肺部PET/CT图像不同纹理特征的K最近邻分类器 预览 被引量:1

K-nearest neighbor classifier based on different texture features of pulmonary nodules from PET/CT images analysis
在线阅读 下载PDF
收藏 分享 导出
摘要 目的对PET/CT图像高维纹理参数进行降维,基于不同纹理参数建立肺结节良恶性的K最近邻(K-nearest neighbor,KNN)分类器,探究最佳建模方法,提高分类的准确率。方法采用回顾性研究的方式,收集52例首都医科大学宣武医院核医学科肺结节患者的PET/CT图像,对图像的感兴趣区域基于Contourlet变换提取灰度共生矩阵的纹理参数。对肺结节PET/CT图像的纹理参数首先采用单因素分析的方法,根据ROC曲线下面积筛选纹理参数,再对其进行主成分分析提取主要成分。基于主成分、根据ROC曲线筛选的纹理及原始纹理分别采用K最近邻分类算法建立肺结节良恶性的分类器,通过正确率、灵敏度、特异度、阳性预测值(positive predictive value,PPV)、阴性预测值(negative predictive value,NPV)、ROC曲线下面积(area under curve,AUC)这些指标评价分类效果。结果 PET/CT图像共提取1344个原始纹理参数,经单因素分析后筛选出89个纹理参数,对筛选后的纹理共提取11个主成分。基于主成分、筛选纹理、原始纹理的分类模型正确率分别为0.614、0.579、0.263;AUC分别为0.645、0.610、0.515。结论在主成分纹理、单因素分析筛选的纹理、原始纹理中,基于主成分纹理建立K最近邻分类器的效果最好。 Objective To reduce the dimension of the high-dimensional texture parameters of PET/CT images and to improve the accuracy of classification by building the K-nearest neighbor( KNN) classifier based on different texture features. Methods The study retrospectively collected 52 cases with pulmonary nodules who underwent18~F-FDG PET/CT from department of Nuclear Medicine,Xuanwu Hospital Capital Medical University.Co-occurrence matrix texture features were extracted from the contourlet transformed PET/CT images. Univariate analysis was applied first to reduce dimensionality of texture features according to c value before principle components analysis. Principal components of texturefeatures from selected texture features were extracted by PCA. We built the KNN classifier for original textures,selected textures and principle components respectively to distinguish benign and malignant nodules,comparing the efficacy of models based on the evaluation indices such as accuracy,sensitivity,specificity and AUC.Results 1344 original texture features were extracted from the region of interest of PET/CT images,from which89 texture features were selected. Eleven principal components were extracted through the PCA procedure. The accuracy of KNN classifiers based on principal components,selected textures and original textures are 0. 614,0. 579 and 0. 263 with AUC of 0. 645,0. 610,0. 515 respectively. Conclusions The KNN classifier based on the texture of principal components is the best one among the classifiers based on original texture features,the selected texture features through univariate analysis and the texture of principal components.
作者 马圆 田思佳 冯巍 梁志刚 崔春蕾 郭秀花 MA Yuan1,2,TIAN Sijia1,2,FENG Wei1,2,LIANG Zhigang3,CUI Chunlei3, GU0 Xiuhua1,2 (1 Department of Epidemiology and Biostatistics, School of Public Health,Capital Medical University,Beijing 100069;2 Beijing Key laboratory of Epidemiology,Capital Medical University,Beijing 100069;3 Department of Nuclear Medicine,Xuanwu Hospital Capital Medical University,Beijing 100053)
出处 《北京生物医学工程》 2018年第1期57-61,85共6页 Beijing Biomedical Engineering
基金 国家自然科学基金(81530087)资助
关键词 K-最近邻分类器 肺癌 纹理特征 PET/CT K-nearest neighborr classifier lung cancer texture features PET/CT
作者简介 通信作者:郭秀花,教授.E-mail:guoxiuh@ccmu.edu.cn
  • 相关文献

参考文献3

二级参考文献55

  • 1韩玉成,郎志谨,张连君,初建国,王绍武,曲永业,杜长春,.高分辨率CT对周围型小肺癌的诊断价值[J].中华放射学杂志,1994,28(11):737-740. 被引量:100
  • 2穆殿斌,王绍平,杨文锋,付政,陈绪霞,孙晓蓉,于金明.食管癌组织中葡萄糖转运蛋白1表达和Ki-67抗原标记指数与PET/CT显示的^18F-FDG摄取水平相关[J].中华肿瘤杂志,2007,29(1):30-33. 被引量:18
  • 3袁双虎,于金明,于甬华,付政,郭洪波,刘同海,杨新华,杨国仁,李文武.FDG PET/CT与PET对食管癌淋巴结转移的诊断价值比较[J].中华肿瘤杂志,2007,29(3):221-224. 被引量:11
  • 4张祖进,李辉,郭召平,刘大鹏,张红远,张宗礼.PET/CT的技术性能及临床应用[J].医疗卫生装备,2007,28(10):59-60. 被引量:15
  • 5Basu S,Kwee TC, Gatenby R, et al. Evolving role of molecularimaging with PET in detecting and characterizing heterogeneity ofcancer tissue at the primary and metastatic sites, a plausible ex-planation for failed attempts to cure malignant disorders. Eur JNucl Med Mol Imaging, 2011,38(6):987-991. 被引量:1
  • 6van Velden FH, Cheebsumon P, Yaqub M, et al. Evaluation of acumulative SUV-volume histogram method for parameterizingheterogeneous intratumoural FDG uptake in non-small cell lungcancer PET studies. Eur J Nucl Med Mol Imaging, 2011,38(9):1636-1647. 被引量:1
  • 7El Naqa I,Grigsby P, Apte A, et al. Exploring feature-based ap-proaches in PET images for predicting cancer treatment out-comes. Pattern Recognit, 2009,42(6) :1162-1171. 被引量:1
  • 8Tixier F, Le Rest CC, Hatt M, et al. Intratumor heterogeneitycharacterized by textural features on baseline 18 F-FDG PET ima-ges predicts response to concomitant radiochemotherapy in esoph-ageal Cancer. J Nucl Med, 2011, 52(3) : 369-378. 被引量:1
  • 9Tomasi G, Turkheimer F, Aboagye E. Importance of quantifi-cation for the analysis of PET data in oncology: Review of cur-rent methods and trends for the future. Mol Imaging Biol, 2012,14(2) :131-146. 被引量:1
  • 10Dehdashti F,Mortimer JE,Trinkaus K,et al.PET-based estradiol challenge as a predictive biomarker of response to endocrine therapy in women with estrogen-receptor-positive breast cancer.Breast Cancer Res Treat,2009,113(3):509-517. 被引量:1

共引文献13

同被引文献8

引证文献1

投稿分析

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部 意见反馈