剪切波弹性成像结合超声造影诊断乳腺肿块的多因素Logistic回归分析

发布时间：2018-03-20 16:12

本文选题：剪切波弹性成像　切入点：超声造影　出处：《遵义医学院》2015年硕士论文　论文类型：学位论文

【摘要】：目的:利用多因素Logistic回归分析的方法评价剪切波弹性成像、超声造影以及两种诊断方法结合对乳腺肿块的诊断价值。方法:乳腺肿块由一位有5年乳腺超声检查经验的医师根据二维声像图进行BI-RADS(Breast Imaging-Reporting and Data System)分类,当肿块分为3或4类时分别行剪切波弹性成像(Shear Wave Elastography,SWE)及超声造影(Contrast-enhanced Ultrasound,CEUS)检查,并与超声引导下粗针活检(Ultrasound Guided-Core Needle Biopsy,UG-CNB)病理结果或手术病理结果对照。纳入标准:①患者年满18周岁;②肿块未经过任何处理或治疗。排除标准:①患者本人不愿或不能签署知情同意书;②乳腺内有植入物;③孕期或哺乳期妇女;④BI-RADS 1、2或5、6类的肿块;⑤肿块同侧乳腺既往接受过外科手术;⑥常规超声检查未见明显肿块;⑦超声造影剂过敏史;⑧心、肺功能不全或严重高血压等不适宜进行超声造影的患者;⑨完全没有造影剂进入的囊性肿块。SWE采集各定量指标并绘制ROC曲线比较各定量指标的诊断价值;CEUS造影剂选用意大利Bracco公司生产的声诺维Sonovue,经肘正中静脉团注其混悬液4.8ml,观察并记录肿块的造影定性指标并进行卡方检验分析比较各定性指标在良恶性肿块间的差异。采用多因素Logistic回归分析SWE与CEUS中各指标的诊断价值并建立回归方程,运用ROC曲线比较SWE、CEUS及二者结合对乳腺肿块的诊断效能。结果:62个乳腺肿块中良性肿块30个,恶性肿块32个;肿块最大径线4.30~37.40mm,平均(17.02±7.48mm);BI-RADS 3类6个(9.7%),4A类21个(33.9%),4B类26个(41.9%),4C类9个(14.5%)。①单因素分析结果:SWE中,除Emean_avg(肿块弹性平均值的测量均值)与Emean_d(肿块与周围正常脂肪组织弹性平均值的差值)外,其余各弹性指标在乳腺良恶性肿块之间的差异均具有统计学意义(P0.05),其中Emax_d(肿块与周围脂肪组织弹性最大值的差值)曲线下面积最大,为0.882(95%CI=0.798~0.966)。CEUS中,差异具有统计学意义(P0.05)的指标为增强时相、达峰强度、增强后肿块大小和“太阳征”。②多因素分析结果:将SWE中各项指标进行多因素Logistic回归分析,进入Logistic回归方程的是Emax_avg和Emin_d,用上述模型预测乳腺恶性肿块的敏感性为81.3%、特异性为86.7%、准确性为83.9%。将CEUS中各项指标进行多因素Logistic回归分析,进入Logistic回归方程的是增强后肿块大小、校正增强时相和校正达峰强度,用上述模型预测乳腺恶性肿块的敏感性为84.4%、特异性为86.7%、准确性为85.5%。将SWE与CEUS中各项指标进行多因素Logistic回归分析,进入Logistic回归方程的是增强后肿块大小和Emax_avg,用上述模型预测乳腺恶性肿块的敏感性为87.5%、特异性为90.0%、准确性为88.7%。③ROC曲线分析:BI-RADS分类方法、SWE、CEUS以及SWE与CEUS结合绘制的ROC曲线,其曲线下面积分别为0.821、0.915、0.920和0.941。上述结果显示:①剪切波弹性成像、超声造影对乳腺良恶性肿块的诊断效能均高于BI-RADS分类方法;②超声造影对乳腺良恶性肿块的诊断效能与剪切波弹性成像相近,而两种方法结合后的诊断效能高于单一诊断方法。结论:剪切波弹性成像的定量指标与超声造影的定性指标的多因素Logistic回归分析具有良好的可操作性与实用性。剪切波弹性成像与超声造影的诊断方法能较好地预测乳腺肿瘤的良恶性,二者的诊断效能均高于BI-RADS分类方法。而将两种诊断方法结合对乳腺恶性肿瘤的预测敏感性、特异性、准确性以及诊断效能均优于单一诊断方法,并具有安全、快速、可操作性强、重复性高的特点,值得在临床上推广应用。
[Abstract]:Objective: using the multi factor Logistic regression analysis to evaluate the shear wave elastography, ultrasonography and two diagnosis method combined with diagnosis of breast tumors. Methods: breast masses by a 5 years inspection experience of breast ultrasound physicians according to two-dimensional ultrasound (Breast BI-RADS Imaging-Reporting and Data System), when the mass divided into 3 or 4 classes respectively for shear wave elastography (Shear Wave, Elastography, SWE) and contrast-enhanced ultrasound (Contrast-enhanced Ultrasound, CEUS), and ultrasound guided coarse needle biopsy (Ultrasound Guided-Core Needle Biopsy, UG-CNB) pathological results or surgical pathology. Inclusion criteria: patients 18 years of age; II. The mass did not undergo any treatment or treatment. Exclusion criteria: the patients themselves are unwilling or unable to sign the informed consent; the breast implants; the pregnancy or lactation The 1,2 or BI-RADS women; 5,6 class mass; the mass of ipsilateral breast had received surgery; the routine ultrasound examination, no obvious tumor; ultrasound contrast agent allergy history; the heart and lung dysfunction or severe hypertension is not suitable for ultrasound angiography were taken; no diagnostic value of contrast agent enter the cystic mass.SWE acquisition of the quantitative index and ROC curve comparison of the quantitative index; CEUS contrast agent of Italy Bracco company's Sonovue SonoVue, cubital intravenous bolus injection of its suspension 4.8ml, observe and record the mass contrast qualitative indexes and compares the qualitative indexes in the difference between benign and malignant among the masses. The chi square test using multivariate Logistic regression analysis of diagnostic value of each index in SWE and CEUS and the regression equation, using ROC curve SWE, CEUS and the combination of the two of breast lumps The diagnostic efficacy. Results: 62 breast masses in 30 benign lesions, 32 were malignant; the maximal diameter of mass line 4.30~37.40mm, the average (17.02 + 7.48mm); 3 BI-RADS 6 (9.7%), 4A 21 (33.9%), 4B 26 (41.9%), 4C (9 14.5%). Single factor analysis results: SWE, except Emean_avg (mean average elastic mass measurement) and Emean_d (difference between the tumor and the surrounding normal tissue elasticity average), the rest of the elastic index were statistically significant differences between the benign and malignant breast tumors (P0.05), including Emax_d (the largest tumor and surrounding fat tissue elasticity difference) area under the curve of maximum, 0.882 (95%CI=0.798~0.966).CEUS, the difference was statistically significant (P0.05) index for the enhanced phase, peak intensity, enhancement of tumor size and the "sun sign". The results of multivariate analysis: the indexes of SWE in because Logistic regression analysis, the regression equation of the Logistic into Emax_avg and Emin_d, using the model to predict the sensitivity of malignant breast masses was 81.3%, the specificity was 86.7%, accuracy was 83.9%. Logistic multivariate regression analysis of the indexes of CEUS, Logistic entered the regression equation is tumor size after enhancement, enhanced phase correction correction the peak intensity, using the model to predict the sensitivity of malignant breast masses was 84.4%, the specificity was 86.7%, accuracy was 85.5%. Logistic multivariate regression analysis of the indexes of SWE and CEUS, into the Logistic regression equation is enhanced after the tumor size and the Emax_avg, using the model to predict the sensitivity of malignant breast masses was 87.5%, specificity as of 90%, the analysis accuracy of 88.7%. ROC curve: BI-RADS classification, SWE, CEUS, SWE and CEUS combined with ROC curve, the area under the curve 0.821,0.915,0.920 and 0.941. respectively. The results show that the shear wave elastography, effectiveness of contrast-enhanced ultrasonography in diagnosis of benign and malignant breast tumors were higher than that of the BI-RADS classification method; II ultrasound diagnostic efficacy on benign and malignant breast tumors and shear wave elastography are similar, and the diagnostic efficacy of combination of the two methods was higher than single diagnosis method. Conclusion Logistic: multi factor quantitative index and qualitative index of contrast-enhanced ultrasound shear wave elastography in the regression has good maneuverability and practicability of the analysis. The diagnosis method of shear wave elastography and contrast enhanced ultrasound can better predict the benign and malignant breast tumors, two were higher than the diagnostic efficacy of BI-RADS classification method and the two. A diagnosis method combining predictive sensitivity of malignant breast tumor specificity, accuracy of single diagnosis method and diagnosis efficiency are superior, and has safety, It is fast, maneuverable and repeatable. It is worth popularizing in clinical practice.

【学位授予单位】：遵义医学院
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：R445.1;R737.9

【参考文献】