• 首页
  • 期刊简介
  • 编委会
    第六届编委会
  • 投稿须知
    文章要求
    投稿指南
    发表学术论文“五不准”
  • 本刊稿约
  • 期刊公告
  • 联系我们
引用本文:
【打印本页】   【下载PDF全文】     【下载PDF阅读器】  【关闭】
←前一篇|后一篇→ 过刊浏览    高级检索
本文已被:浏览 4130次   下载 0次 本文二维码信息
码上扫一扫!
分享到: 微信 更多
字体:加大+|默认|缩小-
MRI纹理分析在预测肝外胆管癌的胆管外侵犯中的价值
王利敏1, 舒健2, 黄新荞2, 杨春梅2, 郭庆喜2, 苏松2
1.西南医科大学影像医学与核医学专业,四川 泸州 646000;2.西南医科大学附属医院放射科,病理科,肝胆外科,四川 泸州 646000
摘要:
目的 探讨MRI纹理分析在预测肝外胆管癌(ECC)的胆管外侵犯中的价值。方法 回顾性分析65例肝外胆管癌患者的MRI图像,以组织学诊断的胆管癌T分期级别来判定胆管外侵犯情况。使用MaZda软件从MRI图像中提取最佳纹理特征,利用统计学方法对特征进一步筛选,并纳入二元 Logistics 回归建立预测模型,通过预测概率建立受试者工作曲线(ROC)评估预测模型的效能。结果 共筛选出10个有意义的纹理特征,P值均<0.05。预测模型x2=34.8,P<0.001。ROC AUC=0.901,P<0.001。其预测肿瘤T分期≥3期的灵敏度、特异度、准确度分别为86.0%、86.4%、86.2%。结论 基于MRI的纹理分析一定程度上有助于预测肝外胆管癌的胆管外侵犯。
关键词:  肝外胆管癌  纹理分析  T分期
DOI:10.16252/j.cnki.issn1004-0501-2021.01.001
分类号:R445.2
文章编号:1004-0501(2021)01-0001-05
文献标识码:A
基金项目:四川省卫生健康委员会科研课题(编号:19PJ151)
The Value of MRI Texture Analysis in Predicting Extra-Bile Duct Invasion of Extrahepatic Cholangiocarcinoma .
Wang Limin1, Shu Jian2, Huang Xinqiao2, Yang Chunmei2, Guo Qingxi2, Su Song2
1.Southwest Medical University Medical Imaging and Nuclear Medicine,Luzhou,Sichuan 646000;2.Department of Radiology, Affiliated Hospital of Southwest Medical University,Luzhou,Sichuan 646000 ,China.
Abstract:
Objective To investigate the value of MRI texture analysis in predicting extra bile duct invasion of extrahepatic cholangiocarcinoma (ECC). Method MRI images of 65 patients with extrahepatic cholangiocarcinoma were retrospectively reviewed to determine the presence of extrabiliary invasion by histologically diagnosed T-stage grade of cholangiocarcinoma. The best texture features were extracted from MRI images using Mazda software, the features were further screened using statistical methods and incorporated into binary logistic regression to build a prediction model, and the efficacy of the prediction model was evaluated by establishing a receiver operating curve (ROC) with predicted probability. Result A total of 10 meaningful texture features were screened out, and all with P<0.05. The chi-square value of the prediction model was 34.8, P<0.001. ROC AUC=0.901, P<0.001. The sensitivity, specificity, accuracy, positive predictive value, and negative predictive value for predicting tumor stage above 3 stage were 86.0%, 86.4%, 86.2%, 76.0%, and 92.5%, respectively.Conclusion Texture analysis based on MRI is helpful to predict extra-bile duct invasion of extrahepatic cholangiocarcinoma to some extent.
Key words:  extrahepatic cholangiocarcinoma  texture analysis  T staging
您是本站第  88076286  位访问者!蜀ICP备11024768号
版权所有:(C)四川省医疗卫生服务指导中心 2010 CopyRight http://scmj.scyx.org.cn/ all right reserved.
地址:成都市玉林南街2号附3号   电话:028-86136765 邮编:610041
技术支持:北京勤云科技发展有限公司