磁共振动态对比增强及弥散加权成像预测乳腺癌前哨淋巴结转移的价值

目的回顾性分析磁共振动态对比增强成像（DCE-MRI）及扩散加权成像(DWI)预测乳腺癌前哨淋巴结（SLN）转移的价值。方法收集我院乳腺癌患者216例，包括SLN转移39例，无SLN转移177例。比较转移组、未转移组间影像特征及临床病理特征的差异，通过多元Logistic回归分析SLN转移的危险性因素，利用受试者工作曲线分析比较影像和临床病理特征对SLN的预测效能。结果两组患者影像、临床病理特征比较，患者年龄、病灶大小、内部强化方式、瘤周/肿瘤表观扩散系数(ADC)比值具有统计学意义（P=0.003、0.001、0.020、0.001）。经Logistic回归分析，肿瘤大小、边缘增强和不均匀强化、瘤周/肿瘤ADC值比与SLN转移独立相关（OR=1.15, P < 0.001；OR=15.54, P=0.007；OR=5.24, P=0.006；OR=70.62, P < 0.001）；临床病理资料中年龄、CD31与SLN转移独立相关（OR=0.96, P=0.012；OR=2.90, P=0.049）。MRI特征预测效能显著高于临床病理变量（P=0.048）。结论术前DCE-MRI病灶大小、内部增强模式和瘤周/肿瘤ADC比可能有助于预测乳腺癌患者的SLN转移。 

Value of DCE-MRI and DWI in predicting sentinel lymph node metastasis of breast cancer

ObjectiveTo retrospectively analyze the value of magnetic resonance dynamic contrast enhancement (DCE-MRI) and diffusion-weighted imaging (DWI) in predicting sentinel lymph node (SLN) metastasis in breast cancer.MethodsA total of 216 breast cancer patients in our hospital were collected, including 39 cases with SLN metastasis and 177 cases without SLN metastasis. The difference of image features and clinicopathological features were compared between metastatic and nonmetastatic groups. The risk factors of SLN metastasis were analyzed by multiple logistic regression. And by receiver operating characteristic (ROC), the predictive power of imaging and clinicopathological features were compared on SLN.ResultsThe imaging and clinicopathological characteristics of the two groups were statistically significant. The age, lesion size, internal enhancement method, and the ratio of peritumoral/tumor apparent diffusion coefficient (apparent diffusion coefficient, ADC) were statistically significant (P=0.003, 0.001, 0.020, 0.001). Logistic regression analysis showed that tumor size, edge enhancement and uneven enhancement, peritumoral/tumor ADC value ratio were independently related to SLN metastasis (OR= 1.15, P < 0.001; OR=15.54, P=0.007; OR=5.24, P=0.006; OR=70.62, P < 0.001); age, CD31 and SLN metastasis were independently related in the clinicopathological data (OR=0.96, P=0.012; OR=2.90, P=0.049). The predictive power of MRI features was significantly higher than clinicopathological variables (P=0.048).ConclusionThe preoperative DCE-MRI lesion size, internal enhancement pattern, and peritumoral/tumor ADC ratio may help predict SLN metastasis in breast cancer patients.