超声应变弹性成像定量分析与甲状腺结节良恶性的相关性研究

Correlation study of ultrasonic strain elastography quantitative analysis with benign or malignant thyroid lesions

目的 探讨应变弹性成像技术的组织弥散定量分析软件评价甲状腺结节良恶性的可行性.方法 对82例甲状腺患者98个结节[平均（1.6±1.1）cm]进行应变弹性成像检查,获得相应的弹性图,启用组织弥散定量分析软件,得到应变均值、标准偏差、蓝色区域面积、复杂度、峰度、偏度、对比度、均等性、复杂度、一致性、相关性等11个量化特征量,对11个量化特征量均值进行良恶性组间比较,并对有统计学差异的量化指标构建ROC曲线,计算曲线下面积,判断其诊断价值.结果 11个量化特征量中除了相关性差异无统计学意义,其余10个量化特征量良恶性组间差异均有统计学意义（P＜0.05）.将这10个量化特征量绘制ROC曲线,以蓝色区域面积和杂乱度两个特征量诊断价值最高,曲线下面积分别为0.965和0.908,诊断临界值分别为81.96％和0.42,约登指数分别为0.875和0.783,其敏感性、特异性分别为98.4％和89.2％、91.8％和86.5％.结论 应变弹性成像技术的组织弥散定量分析软件有助于甲状腺良恶性结节的诊断,以蓝色区域面积和杂乱度诊断价值最高.

Objective To explore the feasibility for the tissue dispersion quantitative analysis software of ultrasonic strain elastography in diagnosing benign or malignant of thyroid lesions.Methods Eighty-two patients with 98 lesions were examined by ultrasonic strain elastography.There were 11 parameters of elastography imaging obtained by the tissue dispersion quantitative analysis software,including average relative strain value（MEAN）,standard deviation of relative strain value（SD）,area ratio of low-strain region （AREA%）,complexity（COMP）,kurtosis（KURT）,skewness （SKEW）,contrast （CONT）,entropy（ENT）,inverse different moment （IDM）,angular second moment （ASM）,correlation （CORR）.The receiver operating characteristic （ROC） curve was constructed if there were statistically significant differences between benign and malignant lesions,and the areas under the ROC curve were got.Results All parameters except CORR had statistically significant between the groups of benign and malignant thyroid lesions （P 〈0.05）.The AREA% and IDM were the best valuable parameters,the areas under the curve（AUC） of which were 0.965 and 0.908,respectively.Their cut-off point were 81.96% and 0.42,the sensitivity and specificitywere 98.4% and 89.2%,91.8% and 86.5%,respectively.Conclusions The tissue dispersion quantitative analysis software is helpful in the evaluation of benign ormalignant of thyroid lesions,parameters of AREA% and IDM has the highest relationship with pathology and good diagnostic value.