应用Logistic回归模型评价超声弹性成像在乳腺结节鉴别诊断中的应用价值

Logistic Regression Model to Evaluate the Value of Ultrasound Elastography in the Differential Diagnosis of Breast Nodules

目的运用Logistic回归模型筛选出能够鉴别乳腺结节良恶性的超声声像图特征,并建立以声像图特征为自变量的Logistic回归模型,评价常规超声及超声弹性成像在乳腺结节良恶性鉴别诊断中的应用价值。方法收集首都医科大学宣武医院2014年1月~2015年12月136例女性乳腺结节手术患者的临床资料,对患者乳腺结节的二维、彩色多普勒及超声弹性成像检查图像进行回顾性分析,以病理结果为金标准,建立回归模型。比较进入方程中的变量的优势比(OR),评价各变量尤其是弹性成像的鉴别诊断效能。结果利用Logistic回归分析评价各个超声特征,筛选出毛刺征、微钙化、血流分级及弹性评分4个在乳腺结节良恶性鉴别诊断中有统计学意义的特征变量,其中弹性评分的OR值高于其他自变量。结论应用Logistic回归模型筛选出对乳腺结节病理性质有鉴别诊断意义的特征变量,超声弹性成像较其他超声特征更有优势,联合应用超声弹性成像及二维声像图特征对于乳腺结节的确诊具有重要临床意义。

Objective To select sonogram features for the differential diagnosis of benign and malignant breast nodules by Logistics regression analysis,and to contribute the Logistic regression model of sonogram features as independent variable and evaluate the value of conventional uhrasonography and uhrasound elastography （UE） in the differential diagnosis of benign and malignant breast nodules. Methods In Xuanwu Hospital from January 2014 to December 2015, the clinical data of 136 cases of breast nodules surgery patients, patients with breast nodules 2D ultrasound, color Doppler and ultrasound elastography （UE） were retrospectively analyzed and pathology results were as the gold standard. The odds ratio of variables in the equation were compared to assess various variables, especially the effi- cacy of elastography in the diagnosis. Results Four statistically significant features were finally entering the Logistic stepwise regression model,including spicule sign, mierocalcifications, type of blood and elasticity score. And the odds ratio of the elasticity score was higher than other features. Conclusion The analysis of Logistic regression can select the valuable variables for the diagnosis of pathological na- ture of breast nodules. UE has much more dominances than other features. The combinated application of UE and 2D ultrasonic features plays a great clinical role in the final diagnosis of breast nodules.