摘要
目的:探讨乳腺导管原位癌(DCIS)的MRI特征对其内存在微浸润癌成分的预测诊断价值。方法:回顾分析我院2018年9月—2020年3月经术后病理确诊且有完整MRI资料的单纯DCIS和导管原位癌伴微浸润(DCIS-IC)病例,进行分组研究。分析每个病例术前的MRI图像,提取包括乳腺纤维腺体类型,背景实质强化类型,病灶的类型和最长径,是否为多中心病灶,乳头复合体受累及病灶弥散是否受限等MRI特征。采用Pearson相关分析MRI测量的病灶最大直径与手术后病理测量的病灶最大直径的相关性。通过单因素分析比较两组间各种MRI特征的差异,再将单因素分析后有统计学意义的MRI特征进行多因素Logistic回归分析,采用ROC曲线评价回归模型对DCIS-IC的预测诊断价值。结果:最终纳入62例进行研究,其中35例为单纯DCIS,27例(43.5%)为DCIS-IC。MRI测量的病灶最大直径与手术后病理测量的病灶最大直径间具有较好的正相关性(r=0.750,P=0.000)。单因素分析结果显示,病灶同侧乳腺实质背景明显强化(F=5.034,P=0.025),存在同侧乳头复合体受累(F=10.070,P=0.002),较大的病灶(t=3.061,P=0.003),弥散明显受限的病灶(F=6.042,P=0.014),不均质和簇样环形强化的非肿块样强化病灶(F=4.560,P=0.049)是对DCIS-IC预测诊断有价值的MRI特征;多因素回归分析结果显示,对DCIS-IC预测诊断最有价值的特征是同侧乳头复合体受累(OR=5.168,P=0.006)和弥散明显受限的病灶(OR=3.254,P=0.048),回归模型对DCIS-IC预测诊断的ROC曲线下面积为0.761。结论:乳腺DCIS术前MRI特征对乳腺DCIS-IC的预测诊断有一定价值。
Objective:To study the value of preoperative breast MRI features in predicting ductal carcinoma in situ(DCIS)with invasive components(DCIS-IC).Methods:Patients with DCIS and DCIS-IC proved by pathology after surgery in our hospital from September 2018 to March 2020 were retrospectively analyzed.The MRI features such as the fibroglandular tissue(FGT)type,background parenchymal enhancement(BPE),the type and maximum diameter of lesion,whether or not to be mul-ticentric lesion and nipple-areolar complex(NAC)invasion were analyzed.Pearson correlation analysis was used to analyze the correlation between the maximum diameter of lesions measured by MRI and the maximum diameter of lesions measured by pathology after operation.The differences of various MRI features between the two groups were compared by univariate analy-sis,and then the statistically significant MRI features after univariate analysis were analyzed by multivariate logistic regression.ROC curve was used to evaluate the predictive diagnostic value of regression model for DCIS-IC components.Results:Sixtytwo cases with 35 pure DCIS and 27(43.5%)DCIS-IC were recruited in the study.The maximum diameter measured by MRI was positively correlated with the maximum diameter measured by pathology(r=0.750,P=0.000).Univariate analysis revealed that the obviously ipsilateral background parenchymal enhancement(F=5.034,P=0.025),NAC invasion(F=10.070,P=0.002),larger tumor diameter(t=3.061,P=0.003),obvious limited DWI(F=6.042,P=0.014)and heterogenous or rim enhancement pattern of NMLE(F=4.560,P=0.049)were valuable MRI features for the prediction and diagnosis of DCIS-IC.Multivariate regression analysis showed that the NAC invasion(OR=5.168,P=0.006)and obvious limited DWI(OR=3.254,P=0.048)were the most valuable features for the prediction and diagnosis of DCIS-IC,and the AUC of the model was 0.761.Conclusion:Preoperative MRI features of DCIS are of certain value in the prediction and diagnosis of DCIS with invasive components.
作者
罗红兵
王春华
刘圆圆
任静
周鹏
LUO Hong-bing;WANG Chun-hua;LIU Yuan-yuan;REN Jing;ZHOU Peng(Department of Radiology,Sichuan Cancer Hospital&Institute,School of Medicine,University of Electronic Science and Technology of China,Chengdu 610041,China)
出处
《中国临床医学影像杂志》
CAS
CSCD
2021年第11期789-793,798,共6页
Journal of China Clinic Medical Imaging
基金
四川省科技厅的重点研发项目(2021YFS0225)
成都市科技局技术创新研发项目(2020YF0500123)。
关键词
癌
导管
乳腺
磁共振成像
Carcinoma
Ductal
Breast
Magnetic Resonance Imaging
