Tamura纹理分析应用于HIFU监控的离体研究  

Tamura Textural Analysis over Ultrasonography Monitoring of High Intensity Focused Ultrasound on Tissue Ablation in Vitro

作　　者：冯玉洁[1] 邹建中[1] 王华[1] 李雁[1] 欧阳挺[1] 

机构地区：[1]重庆医科大学生物医学工程系超声医学工程重庆市市级重点实验室,重庆400016

出　　处：《中国医科大学学报》2010年第8期610-612,共3页Journal of China Medical University

基　　金：国家自然科学基金资助项目(30970767)

摘　　要：目的探讨在高强度聚焦超声(HIFU)辐照过程中,超声图像与HIFU辐照产生凝固性坏死的关系,以提高监控超声对HIFU凝固性坏死的判断灵敏度。方法在相同声强、辐照时间、辐照深度情况下,HIFU定点辐照离体牛肝,观测靶区辐照前及辐照结束后即刻、1min、5min的声像图变化和灰度值变化,并对靶区声像图进行小波变换,对经过小波变换后得到的9幅图进行tamura纹理分析,再用支撑矢量机(SVM)筛选参数并得到决策超平面。结果超声图像基于小波变换的tamura纹理分析可以评价HIFU凝固性坏死情况,其灵敏度高于用灰度评价(χ2=15.211,P<0.05)。结论声像图基于小波变换的tamura斑纹分析判断HIFU辐照的靶组织有无凝固性坏死优于灰度评价。Objective To improve the determination sensitivity of coagulation necrosis caused by high intensity focused ultrasound（HIFU） exposure.Methods The bovine liver in vitro was dot-exposed to HIFU with constant setting of intensity,duration series and depth.The ultrasonographys generated at seven time points were inspected,i.e.prior to the exposure,instantaneous after the exposure,1 minute and 5minute after the exposure.The gray values of the graphs were measured,and the tamura texture features of the 9 graphs which obtained from target area wavelet transforming was computed.Based on Support Vector Machine,our method automatically selects the parameters for the determination function.Results Tamura textural analysis based on wavelet transforming can be used to predict HIFU coagulation necrosis,and its sensitivity is higher than the gray-scale method（Χ^2=15.211,P〈0.05）.Conclusion Tamura textural analysis based on wavelet transforming over ultrasonographic data can be used to predict coagulation necrosis of target tissues exposed with HIFU.According to experiments,our method is better than traditional gray-scale evaluation method.

关 键 词：高强度聚焦超声 小波变换 tamura理纹分析 支撑矢量机 

分 类 号：R445.1[医药卫生—影像医学与核医学]