亚实性肺结节CT阈值分割:实性成分识别与定量

发布时间：2018-03-13 20:36

本文选题：亚实性结节　切入点：腺癌　出处：《中国肺癌杂志》2017年05期 　论文类型：期刊论文

【摘要】：背景与目的肺内亚实性结节(subsolid nodule,SSN)实性成分的识别与定量对SSN的鉴别诊断,病理预测和预后评估具有重要价值,但目前缺乏公认且客观的标准。本研究旨在应用计算机断层扫描(computed tomography,CT)阈值分割方法确定用于SSN实性成分识别与定量的最佳阈值。方法回顾性分析102例SSN的CT图像。由观察者确定SSN内是否存在实性成分并借助软件确定实性成分的体积,以此作为参照标准。应用CT阈值分割方法对所有SSN进行分析,测量其内不同CT阈值对应的体素体积,假定该体积即为实性成分体积,计算实性成分与结节整体的体积比率。所得结果与相应参照标准进行比较,应用受试者工作特征曲线及Wilcoxon检验筛选最佳阈值。结果阈值设为-250 HU时,识别SSN内实性成分的诊断价值最大(曲线下面积为0.982),以体积比率1.10%作为确认实性成分存在的界限值,其诊断效能最高。阈值设为-300 HU时诊断效能次之,对应曲线下面积和体积比率界限值分别为0.977、6.14%。阈值设为-250 HU、-300 HU时所得实性成分中位体积分别为202.7 mm~3(598.2 mm~3)、247.1 mm~3(696.0 mm~3),与参照标准199.5 mm~3(743.1mm~3)间无显著差异(P=0.125,1,0.061,3)。结论 CT阈值分割方法可对SSN实性成分进行准确识别与定量,阈值可设为-250 HU或-300 HU。
[Abstract]:Background & objective the identification and quantification of solid components of subsolid nodules of the lung have important value in differential diagnosis, pathological prediction and prognosis evaluation of SSN. The purpose of this study is to determine the best threshold for identifying and quantifying real components of SSN by using computed tomography threshold segmentation method. Methods the CT images of 102 cases of SSN were retrospectively analyzed. . the observer determines whether there is a real component in the SSN and the volume of the real component is determined by software, Using CT threshold segmentation method to analyze all SSN and measure the volume of voxels corresponding to different CT thresholds, the volume is assumed to be the volume of real components. The volume ratio of solid components to the whole nodule was calculated. The results were compared with the corresponding reference criteria. The optimal threshold was screened by using the operating characteristic curve and Wilcoxon test. When the threshold value was -250HU, the optimal threshold was obtained. The diagnostic value of identifying the real components in SSN is the greatest (the area under the curve is 0.982g). The diagnostic effectiveness is the highest when the volume ratio is 1.10%, and the diagnostic efficacy is the second when the threshold is -300HU. The threshold values of area and volume ratio under the corresponding curve are 0.977 卤6.14, respectively. When the threshold value is -250hu -300HU, the median volume of solid component is 202.7 mm~3(598.2 / mm, respectively, and there is no significant difference between the threshold value of real component and the reference standard (199.5 mm~3(598.2 / 3). Conclusion CT threshold segmentation method can be used for CT segmentation. Accurate identification and quantification of SSN real components, The threshold can be set to -250 Hu or -300 Hu.
【作者单位】：天津医科大学总医院医学影像科;
【基金】：国家自然科学基金面上项目(No.81171345) 中央补助地方公共卫生专项资金肺癌早诊早治项目高等学校博士学科点专项科研基金(No.20121202110005)资助~~
【分类号】：R730.44;R734.2

【相似文献】