利用X射线衍射增强成像技术诊断肝纤维化

Investigation on diagnosing hepatic fibrosis with X-ray diffraction-enhanced imaging

目的:利用X射线衍射增强成像(diffraction enhance imaging,DEI)的空间高分辨率和高相位衬度特征,对小鼠肝纤维化样本进行成像,观察纤维化病变在DEI图像中的表象特征,并对正常肝组织及肝纤维化图像进行纹理特征分析,探讨量化诊断肝纤维化病变的有效性。方法:实验在北京同步辐射装置(Beijing Synchrotron Radiation Facility,BSRF)4W1A光束线形貌学实验站完成,样品为四氯化碳(carbon tetrachloride,CCl_4)诱导的ICR小鼠肝纤维化模型及正常肝脏组织标本。对样品进行DEI成像后,对肝脏组织进行马松(Masson)染色,将DEI图像与病理结果进行比对,分析得到不同程度纤维化病变的影像学特征。基于灰度共生矩阵(gray level co-occurrence matrix,GLCM),计算DEI图像感兴趣区域(region of interest,ROI)的九种纹理特征参数,分析纹理特征参数与纤维化病变程度的相关性。结果:正常及不同程度的肝纤维化组织在DEI图像上差异明显,能量、熵、惯性矩、逆差距、差的熵以及差的均值六种纹理特征值能够明显区分不同程度肝纤维化(P<0.05),其中,熵、惯性矩、差的熵、差的均值随着肝纤维化程度的加重而增大,而能量、逆差距随着肝纤维化程度的加重而减小,熵的变化程度最为明显。结论:正常及不同程度的肝纤维化模型在DEI图像上差异明显,基于GLCM提取的图像纹理特征参数能够对不同程度的肝纤维化进行区分。DEI成像为肝纤维化的诊断提供了一种潜在的影像学途径。

Objective： X-ray diffraction-enhanced imaging （DEI） method was used to image hepatic fibrous samples,and the texture measurements based on DEI images were calculated and analyzed for investigating the feasibility of diagnosing fibrosis quantitatively. Methods： Hepatic fibrosis of different severity in ICR mice was induced by injecting with carbon tetrachloride （CCl4） olive oil solution for 21 d,42 d and 84 d,respectively. The excised liver tissues were then imaged at the 4 W1 A Topography Imaging Station of Beijing Synchrotron Radiation Facility （BSRF）. After imaging,the liver tissues were sectioned and stained for Masson trichrome using standard histological techniques. The imaging features of hepatic fibrosis of different severity were analyzed by comparing DEI images with histological examinations. The nine basic texture parameters on regions of interest （ROIs） of DEI images were extracted based on the gray level co-occurrence matrix （GLCM）,and then evaluated for their feasibility to distinguish stages of fibrosis. Results： The results showed that great differences could be observed in the DEI images between the normal and diseased mice in different stages of liver fibrosis. It was demonstrated that there were six texture measurements （entropy,inertia,difference entropy,difference average increase,energy and inverse difference moment） that could distinguish between the normal and diseased mice in different stages of liver fibrosis （P〈 0. 05）. The texture parameters of entropy,inertia,difference entropy,difference average increased with the development of fibrosis,while the texture parameters of energy and inverse difference moment decreased with the development of fibrosis. The variation in entropy of different severity of fibrosis was most obvious among the six texture features. Conclusion： The degree of hepatic fibrosis can be discriminated by morphological features in DEI images. Moreover,the degree of hepatic fibrosis can be graded quantitatively by the texture features. These results suggest that DEI can be a promising noninvasive technique to diagnose and distinguish hepatic fibrosis of different degrees.