摘要
目的:探讨胰腺癌能谱CT图像纹理分析在淋巴结转移评估中的可行性及其影响因素。方法:选取本院行术前CT能谱成像检查且病理证实为胰腺导管腺癌的155例患者。CT增强扫描采用能谱扫描(GSI)模式,期相取动脉晚期(AP)、门脉期(PP)采用标准算法重建图像,层厚分别为1.25mm和5mm。根据淋巴结转移情况,将研究对象分为两组,应用Ma Zda软件通过手动勾画感兴趣区(ROI)的方式提取病变的纹理特征,通过特征选择方法选出最具判断胰腺癌是否淋巴转移的纹理特征。方法包括Fisher系数(Fisher)、分类错误概率联合平均相关系数(POE+ACC)、互信息(MI)及上述3种方法的联合法(FPM)。用不同分类统计方法判断胰腺癌是否存在淋巴转移,结果以判错率形式表示。结果:特征选择方法中,3种方法联合法(FPM)选择的纹理特征判断胰腺癌是否淋巴结转移的判错率最低,为20.65%(32/155);分类统计方法中,非线性判别分析(NDA)区分胰腺癌淋巴结转移的判错率低于原始数据分析(RDA)、主成分分析(PCA)、线性判别分析(LDA)。能谱CT图像各序列中,动脉晚期5mm层厚(5mmAP)图像的纹理分析判错率最低,但各序列之间的差异无统计学意义。结论:胰腺癌能谱CT图像纹理分析可作为术前评估是否发生淋巴结转移的辅助工具。
Purpose: To investigate the feasibility of texture analysis in prediction nodal involvement of pancreatic ductal adenocarcinoma with spectral CT images. Methods: One hundred and fifty-five patients with pathologically confirmed pancreatic ductal adenocarcinoma admitted to Ruijin Hospital were enrolled retrospectively. Gemstone spectral imaging(GSI) was performed in late arterial phase(AP) and portal venous phase(PP). Images were reconstructed with standard algorithm and slice thickness of 1.25 mm and 5 mm. Study subjects were divided into two groups according to lymph node metastatic state. Texture features were extracted from manually drawn ROIs by using MaZda software. Feature selection methods including Fisher coefficient(Fisher), classification error probability combined with average correlation coefficients(POE+ACC), mutual information(MI), and combination of above three methods(FPM) were used. With the selected features, four classification algorithms were used to predict the categories of the study cases, which include raw data analysis(RDA), principal component analysis(PCA), linear discriminant analysis(LDA), and nonlinear discriminant analysis(NDA). The results were presented by misclassification rate.Results: For the feature selection methods, FPM was with the lowest misclassification rate. Among the classification algorithms, the misclassification rate of NDA was lower than those of RDA, PCA, and LDA. Among the four CT image series, the images of late arterial phase(AP) with 5 mm slice thickness obtained the lowest misclassification rate, but there was no significant difference among the different series. Conclusions: Texture analysis of spectral CT images can be used as an assistant tool for predicting nodal involvement of pancreatic ductal adenocarcinoma.
作者
房炜桓
李旭东
张静
潘自来
陈克敏
林晓珠
FANG Wei-huan;LI Xu-dong;ZHANG Jing;PAN Zi-lai;CHEN Ke-min;LIN Xiao-zhu
出处
《中国医学计算机成像杂志》
CSCD
北大核心
2019年第2期141-145,共5页
Chinese Computed Medical Imaging
基金
国家自然科学基金No.81201145~~
关键词
胰腺导管腺癌
淋巴结转移
能谱CT成像
纹理分析
特征选择
Pancreatic ductal adenocarcinoma
Lymph node metastasis
CT spectral imaging
Texture analysis
Feature selection
