摘要
目的探讨乳管镜图像特征多元参数在伴乳头溢液乳腺癌中的诊断价值。方法回顾性分析1159例病理资料完整的伴乳头溢液乳腺病患者的乳管镜资料,通过χ2检验筛选出乳管镜下诊断乳腺癌的相关图像特征,用Logistic回归分析评价它们在伴乳头溢液乳腺癌中的诊断价值。结果乳管镜诊断相关的图像特征包括病变的位置、管腔改变、管壁弹性、管壁表面形态、乳管末稍出血、病变在乳管内的位置、病变形状、病变颜色、病变表面形状、病变数量及病变表面出血等(P<0.05),行Logistic回归分析筛选出各乳管镜图像特征的OR值分别为:OR管壁表面形态=3.05,OR乳管末梢出血=2.22,OR病变数量=2.09,OR病变颜色=1.53,OR病变表面形状=1.53,OR病变形状=1.41,OR管壁弹性=0.43。根据OR值对各乳管镜图像特征在伴乳头溢液乳腺癌中的诊断价值进行排序:管壁表面形态>乳管末梢出血>病变数量>病变颜色=病变表面形状>病变形状>管壁弹性。结论乳管镜图像特征对伴乳头溢液乳腺癌的诊断有重要价值。
Objective To explore the diagnostic value of mammary ductoscopic appearances in breast cancer with nipple discharge by logistic regression model. Methods A total of 1159 patients that received mammary ductoscopy with complete pathological results were analyzed retrospectively. We selected several diagnostic parameters for breast cancer with nipple discharge and assessed their diagnostic value by logistic stepwise regression model. Results The mammary ductoscopic appearances for breast cancer with nipple discharge included lesion orientation, duct cavity, elasticity of duct wall, surface of duct wall, bleeding of duct end, lesion location, lesion shape,lesion color, lesion surface profile, lesion amount and lesion bleeding (P 〈 0.05). The result of logistic regression model showed that surface of duct wall (OR = 3.05 ), duct end bleeding( OR = 2.22), lesion amount ( OR = 2.09), lesion color ( OR = 1.53 ) ,lesion surface profile ( OR = 1.53),lesion shape ( OR = 1.41 ), elasticity of duct wall(OR = 0. 43) were valuable for the diagnosis of breast cancer with nipple discharge. Accordingly, the rank of diagnostic value of those parameters was as follows : surface of duct wall 〉 duct end bleeding 〉 lesion amount 〉 lesion color = lesion surface profile 〉 lesion shape 〉 elasticity of duct wall. Conclusions Mammary ductoscopic appearances are helpful to diagnose breast cancer with nipple discharge.
出处
《中华乳腺病杂志(电子版)》
CAS
2011年第4期29-32,共4页
Chinese Journal of Breast Disease(Electronic Edition)
关键词
乳管内视镜
回归分析
乳头溢液
乳腺肿瘤
mammary ductoscopy
regression analysis
nipple discharge
breast neoplasms