乳腺癌前哨淋巴结转移多因素Logistic回归分析  

MULTIVARIATE LOGISTIC REGRESSION ANALYSIS OF BREAST CANCER SENTINEL LYMPH NODE METASTASIS

作　　者：康淑娟[1] 胡大为[2] 马立辉[2] 王助新[2] 孟庆来[2] 张宏旭[2] 

机构地区：[1]承德医学院,河北承德067000 [2]承德医学院附属医院乳腺外科

出　　处：《承德医学院学报》2016年第3期204-207,共4页Journal of Chengde Medical University

摘　　要：目的:探讨乳腺癌前哨淋巴结(SLN)转移的影响因素,并建立多因素Logistic回归模型。方法:分析167例早期乳腺癌患者的年龄、肿瘤位置、肿瘤直径、病理分类、分子亚型等与SLN转移的关系,取有统计学意义变量进行多因素Logistic回归分析。结果:肿瘤直径、病理分类、分子亚型与SLN转移有关。影响SLN转移的风险比值分别为:HER2过表达(OR=22.491)、Luminal B(HER2+()OR=6.304)、Luminal B(HER2-)(OR=5.401)、三阴型(OR=4.127)、肿瘤直径(OR=2.825)、病理分类(OR=15.240)。多因素Logistic回归预测模型为:P=1/1+ExpΣ(5.319-1.038X1-2.724X2-1.841X3-1.687X4-3.113X5-1.418X6),该模型预测乳腺癌SLN转移的准确度为79.0%、灵敏度为61.5%、特异度为87.0%。结论:肿瘤直径、病理分类、分子亚型是SLN转移的独立影响因素,所建立的多因素Logistic回归模型能较好地预测乳腺癌SLN转移的概率。Objective： To investigate the infl uence factors of breast cancer sentinel lymph node（SLN） metastasis and establish Logistic regression model. Methods： The age, tumour location, tumour diameter, pathological type and molecular subtype of 167 early breast cancer patients were analyzed and performed multivariate logistic regression analysis. Results： Tumour diameter, pathological type and molecular subtype were related to SLN metastasis. The risk ratio of influencing factors of SLN metastasis were respectively HER2 over express type（OR=22.491）, Luminal B（HER2＋） type（OR=6.304）, Luminal B type（HER2-）（OR=5.401）, triple negative type（OR=4.127）, tumor diameter（OR=2.825）, pathological type（OR=15.240）. Multivariate logistic regression prediction model was P=1/1＋ExpΣ（5.319-1.038X1-2.724X2-1.841X3-1.687X4-3.113X5-1.418X6）, whose accuracy, sensitivity and specifi city that predict breast cancer SLN metastasis were respectively 79.0%, 61.5% and 87.0%. Conclusions： Tumor diameter, pathological type and molecular subtype are independent influencing factors of SLN metastasis. Multivariate logistic regression prediction model in this study can predict the probability of breast cancer SLN metastasis better.

关 键 词：乳腺癌 前哨淋巴结转移 回归分析 

分 类 号：R737.9[医药卫生—肿瘤]