目的评价MRI动态增强鉴别乳腺 良恶性肿瘤和病理分级的作用。方法 回 顾性分析我院2012年1月1日至2014年12 月31日期间经病理证实行影像学动态增强 MRI检查的患者85例，选择第1个动态序列 最大轴面肿瘤，选择3个明显强化和3个强 化最弱的区域勾画兴趣区域(ROI)计算ROI 的各参数。以病理检查结果为金标准， 采用受试者曲线下面积(ROC)评估初始斜 率、流入\流出斜率梯度、曲线形态、环 形强化、不均匀强化程度来鉴别乳腺肿瘤 的良恶性。45例浸润性导管癌患者不同病 理分级各参数的比较采用χ2 或KruskalWallis秩和法分析各动态参数等非参数 资料，P＜0.05为差异有统计学意义。结 果 所有参数中热点初始斜率ROC曲线下面 积(AUC)最大(0.675)，其敏感度和特异度 分别为68.6%和67.6%，对鉴别乳腺良性和 恶性肿瘤的意义较大，但与其他参数差异 未见统计学意义(P值均>0.05)。45例浸 润性导管癌患者中，I、Ⅱ、Ⅲ级分别为 13、17、15例，冷点初始斜率中位数为 6.2×10-3、5.5×10-3、1.6×10-3，流入斜 率梯度中位数为6.7、9.9、16.8，差异均 有统计学意义(P＜0.05)；强化特征内部 均匀性，I级表现为均匀、较不均匀、极 不均匀分别为3、7、3例，Ⅱ级为1、11、 5例，Ⅲ级为0、5、10例，差异有统计显 著性意义(P＜0.05)。结论 MRI动态增强 扫描利用相应参数特征有助于鉴别乳腺肿 瘤的良恶性和病理分级。

Objective To evaluate clinical value of dynamic contrast enhanced MRI (DCE-MRI) to identify benign and malignant breast tumor and predicting histological grades for breast cancer patients. Methods DCE-MRI from 86 patients who admitted from Jan. 1, 2012 to Dec. 31,2014, in our hospital, were retrospectively analyzed. in the first dynamic phase, the strengthened and weakest 3 areas (hot spot, cold spot)in the maximum axial tumors were selected as ROIs. Then the parameters of the ROI were calculated and compared. The ROC curve analysis was applied to test the diagnostic power of DCE-MRI parameters (initial slope, inflow/washout slope ratio, fast inflow/ washout, ring enhancement, internal homogeneity) for differentiating benign breast tumors from malignant ones. χ2 or Kruskal-Wallis test were used to analyze categorical data. Results the area under ROC curve (AUC) of hot spot initial slope is 0.675, biggest among all MRI parameters (P> 0.05). its sensitivity and specificity were 68.6% and 67.6% respectively. The cold initial slope, inflow slope ratio and internal homogeneity showed a significant correlation with histologic grade (P <0.05). the median and quartile of cold initial slope in pathologic I, Ⅱand Ⅲ were 6.2×10-3,5.5×10-3 and 1.6×10-3; inflow slope ratio: 6.7,9.9 and 16.8 ; pathologic I, Ⅱand Ⅲ showed 3,7 and 3 patients with homogeneity enhancement; intermediate enhancement: 1,11 and 5 patients; heterogeneity enhancement: 0,5 and 10 patients. Conclusion The DCE- MRI is a potential tool in helping differentiate malignant tumors from benign ones, and predict the histologic grade in breast cancers.