乳腺癌前哨淋巴结转移的超声及病理危险因素分析

目的:联合临床病理及超声图像特征，采用二元logistic回归，分析乳腺癌前哨淋巴结(SLN)转移的临床病理及超声危险因素。方法:回顾性分析404例接受乳腺癌手术治疗的患者的临床病理及超声资料，以病理诊断结果为金标准，分析SLN转移与病理和超声特征的相关性，对单因素和多因素分析均有统计学意义的特征建立logistic回归模型，并绘制受试者工作特征(ROC)曲线，评估模型效能。结果:单因素分析与多因素分析显示，肿瘤最大径、内部回声、边界、边缘、血流分级五种超声特征及肿瘤增殖指数Ki-67、病理类型两种临床病理特征是乳腺癌SLNM的独立危险因素。ROC曲线下面积(AUC)为0.799,灵敏度79.7%,特异度67.5%。结论:乳腺癌SLNM的独立危险因素分别是超声特征中的肿瘤最大径≥20 mm、内部回声不均匀、边界不清、边缘不光整、血流信号丰富(Ⅱ-Ⅲ级)和肿瘤增殖指数Ki-67>14%和病理类型。

Risk factors analysis of sentinel lymph node metastasis in ultrasonographic and clinicopathologic with breast cancer

Objective: Combined with clinicopathologic and ultrasonic features, binary logistic regression was used to analyze the clinicopathologic and ultrasonic risk factors of sentinel lymph node (SLN) metastasis in breast cancer. Methods: The clinicopathological and ultrasonographic data of 404 patients who received breast cancer surgery in our hospital were retrospectively analyzed, and the pathological diagnosis was used as the gold standard to analyze the correlation between SLN metastasis and pathological and ultrasonographic characteristics. Logistic regression models were established for the features with statistical significance in both univariate and multivariate analysis, and receiver operating characteristic (ROC) curves were drawn to evaluate the model performance. Results: Univariate analysis and multivariate analysis showed that five ultrasonographic features, tumor maximum diameter, internal echo, boundary, margin, and blood flow grade, and two clinicopathological features, tumor proliferation index Ki-67 and pathological type, were independent risk factors for breast cancer SLNM. The area under the ROC curve (AUC) was 0.799, with a sensitivity of 79.7 % and specificity of 67.5 %. Conclusion: The independent risk factors for SLNM in breast cancer are ultrasound characteristic tumor maximum diameter ≥20mm, internal echo heterogeneity, unclear boundary, irregular edge, abundant blood flow signal (grade Ⅱ-Ⅲ) , tumor proliferation index KI-67 > 14 %, and the pathological type.