超声和X线检查对乳腺病变的诊断效能分析

Diagnostic efficacy analysis of ultrasonography and mammography for breast lesions

目的 探讨超声和X线检查对乳腺病变的诊断效能。方法 回顾性分析2007—2014年在北京市顺义区妇幼保健院乳腺中心就诊的1 083例患者乳腺病变的超声和X线影像诊断资料。根据年龄将患者分为〈40岁、40~49岁、50~59岁和≥60岁4个组,根据腺体密度将患者分为Ⅰ~Ⅳ度4个组。以病理诊断为金标准,超声和X线BI-RADS分级4~5级为阳性结果,采用受试者工作特征（ROC）曲线对两者诊断效能进行评价。采用Kappa检验评价超声和X线检查单独及联合使用的一致性。结果 1 083例患者均有明确病理诊断,包括乳腺癌386例和良性疾病697例。在386例乳腺癌患者中,超声单独诊断乳腺癌的敏感度和特异度分别为93.0%（359/386）和87.8%（612/697）,X线单独诊断乳腺癌的敏感度和特异度分别为76.2%（294/386）和92.4%（644/697）,两者联合诊断乳腺癌的敏感度和特异度分别为95.1%（367/386）和82.6%（576/697）。以病理诊断为金标准,超声和X线检查与金标准的一致性检验Kappa值分别为0.782和0.701,联合诊断的Kappa值为0.734（P均〈0.001）。乳腺超声和X线诊断的ROC曲线下面积分别为0.904（95%CI：0.885~0.921,P〈0.001）和0.843（95%CI：0.820~0.864,P〈0.001）,超声的诊断价值优于X线（Z=5.930,P〈0.001）。对于4个年龄组研究对象,超声的诊断价值均优于X线（Z=2.549、5.332、3.584和2.031,P均〈0.050）。超声和X线对于乳腺腺体分型为Ⅱ度者诊断价值相当（Z=0.178,P=0.859）,而超声对腺体分型为Ⅰ度、Ⅲ度和Ⅳ度者的诊断价值均高于X线（Z=2.129、5.847、3.437,P均〈0.050）。结论 超声和X线对乳腺病变均有较好的诊断价值,超声的诊断价值略优于X线。建议临床实践中对于致密型乳腺,在X线检查的基础上结合超声检查,以提高乳腺病变检出率。

Objective To investigate the diagnostic performance of uhrasonography （US） and mammography （MAM） for breast lesions. Methods The US and MAM imaging data of 1 083 patients with breast diseases in Shunyi Health Care Hospital for Women and Children in 2007-2014 were retrospectively analyzed. The patients were divided into 4 groups ： 〈40 years, 40-49 years, 50-59 years, 〉160 years, or the patients were divided into 4 groups of Ⅰ-Ⅳ based on breast density. Pathological diagnosis of the lesions was set as the gold standard, and the BI-RADS grades 4 and 5 of US and MAM were regarded as positive results. ROC curve analysis was used to evaluate the diagnosis performances of US and MAM. Kappa test was used to evaluate the consistency of US, MAM and combined diagnosis of both. Results All patients had definite pathological diagnosis, including 386 cases of breast cancer and 697 cases of benign diseases. Among 386 cases of breast cancer, the sensitivity and specificity of US alone for diagnosis of breast cancer were 93.0% （359/386） and 87.8% （612/697）, and the sensitivity and specificity of MAM alone were 76.2% （294/386） and 92.4% （644/697）, and the sensitivity and specificity of combined diagnosis were 95. 1% （367/386） and 82. 6% （576/697）. With pathological diagnosis as the gold standard, the Kappa values of US and MAM in consistency test were 0. 782 and 0. 701 respectively, and the Kappa value of combined diagnosiswas 0. 734 （ all P〈0. 001 ）. The areas under ROC curves of US and MAM for diagnosis of breast lesions were 0. 904 （95% CI： 0. 885-0. 921, P〈0. 001 ） and 0. 843 （95% CI： 0. 820-0. 864, P〈0. 001 ）, respectively. US was significantly superior to MAM （Z = 5. 930, P〈0. 001 ）. In 4 age groups, US had a better diagnostic value compared with MAM （Z=2.549, 5.332, 3.584, 2.031, all P〈0.050）. US and MAM had an equivalent diagnostic value for lesions of grade Ⅱ breast density （ Z = 0. 178, P = 0. 859）, while US had an significantly better diagnostic value for lesions of grades Ⅰ , Ⅲ and Ⅳ breast density （ Z = 2. 129, 5. 847, 3.437, all P〈0. 050）. Conclusions Both US and MAM show a good diagnostic value for breast lesions, while US is more effective than MAM. MAM combined with US is recommended for women with dense breast in clinical practice, in order to improve the detection rate of breast lesions.