邻接点算法结合SonoCT技术分析肝硬化声像图纹理的实验研究  

The experimental research on the diagnosis of liver cirrhosis using neighborhood-pixels algorithm combined with SonoCT

作　　者：华兴[1] 李锐[1] 郭艳丽[1] 张萍[1] 郭爱民[1] 

机构地区：[1]第三军医大学西南医院超声科,重庆400038

出　　处：《中华超声影像学杂志》2006年第2期132-134,共3页Chinese Journal of Ultrasonography

基　　金：重庆市科委基础研究与应用课题(2000-8-158)

摘　　要：目的探讨邻接点算法对实验性肝硬化声像图纹理的分析能力和SonoCT技术在提高纹理分析能力中的应用价值。方法以皮下注射四氯化碳建立大鼠肝硬化模型,并将其分为5组:对照组、肝硬化模型6周、8周、10周和12周组,每组16只。在基波条件和SonoCT条件下,以邻接点算法对大鼠肝声像图进行纹理分析并测定其相对不均匀度(HSR),对比分析各组间和各组内不同条件下的HSR测值。结果随着模型时间的增长,各组的HSR逐渐升高,各组间差异有统计学意义(P<0.01);各组内SonoCT条件下获得的HSR较基波条件下更高(P<0.01);各组间两种条件下测得的HSR值差异有统计学意义(P<0.01)。结论邻接点算法测得的HAR值能反映实验性肝硬化声像图纹理的动态变化,SonoCT技术有助于提高邻接点算法的纹理分析能力。Objective To investigate the ability of neighborhood-pixels algorithm in analyzing the echogram of experimental liver cirrhosis and the value of technology of SonoCT in improving the ability of texture analysis. Methods Model of liver cirrhosis was established in 80 rats by subcutaneously injecting CCl4, which were divided into one control group and 4 groups according to modeling time （6,8,10, and 12 weeks respectively）. The texture of hepatic echograms of all the rats was analyzed with neighborhood-pixels algorithm on the conditions of fundamental wave and SonoCT respectively. The relative heterogeneity values （HSR） in different groups and in different conditions were measured and compared. Results HSR in the 5 groups increased with the increase of modeling time, the differences of which were significant （ P 〈0. 01）. In each group, HSR acquired in the condition of SonoCT was higher than that acquired in conventional condition （ P 〈0. 01）. And the difference of HSR in the two conditions varied significantly among groups （ P 〈0. 01 ）. Conclusions HSR measured with neighborhood- pixels algorithm can reflect the dynamic change of the texture of echogram of rats with liver cirrhosis and the technology of SonoCT can enhance its ability of texture analysis.

关 键 词：超声检查 肝硬化 邻接点算法 

分 类 号：R575.2[医药卫生—消化系统]