基于DWI和DCE-MRI的影像特征对四种不同分子亚型乳腺癌的诊断价值探讨  

Diagnostic value of DWI and dynamic contrast-enhanced MRI in patients with four different types of breast cancer

作　　者：徐丽娜[1] 唐竹晓[1] 李双标[1] 李莉[1] 徐文丽 李瑞南 XU Lina;TANG Zhuxiao;LI Shuangbiao;LI Li;XU Wenli;LI Ruinan(Department of Imaging,Cangzhou Combination of Traditional Chinese and Western Medicine of Hebei Province, Hebei Cangzhou 061001,China)

机构地区：[1]河北省沧州中西医结合医院影像科,河北沧州061001

出　　处：《现代肿瘤医学》2021年第1期116-120,共5页Journal of Modern Oncology

基　　金：沧州市科技局项目基金(编号:162302052)。

摘　　要：目的:基于弥散加权成像(diffusion weighted imaging,DWI)和动态增强MRI(contrast-enhancement magnetic resonance imaging,DCE-MRI)的影像特征对Luminal A型、Luminal B型、人类表皮生长因子受体(human epidermal growth factor receptor-2,HER-2)过表达型和三阴型(triple negative,TN)四种不同分子亚型乳腺癌的诊断价值探讨。方法:选取2016年9月至2018年9月我院收治确诊为乳腺癌的患者98例,所有患者行DWI及DCE-MRI扫描,后行乳腺癌手术行病理分子分型鉴定。采用Mann-Whitney U检验和单因素logistic回归法进行各变量单因素分析,采用多因素logistic回归法对单因素分析中有统计学意义的变量进一步分析并建模,使用受试者工作曲线(receiver operating characteristic curve,ROC)评估模型的诊断效能,采用Hosmer-Lemeshow检验对模型的拟合优度进行检验。结果:患者分子病理分型Luminal A型32例,Luminal B型45例,HER-2过表达型10例,TN型11例。ADC、IER值在鉴别乳腺癌分子分型时AUC<0.07,诊断效能较低;影像学特征鉴别分子分型时,显示DCEbior3.13correlation鉴别Luminal A型乳腺癌AUC=0.732;ADCrbio1.11sumvariance鉴别Luminal B型乳腺癌AUC=0.722;鉴别HER-2过表达型乳腺癌ADCLG2.5min AUC=0.747;鉴别TN时5个影像学特征AUC均> 0.70。鉴别Luminal A和非Luminal A型的模型为0.005×ADCrbio1.11sumvariance+0.032×DCEbior3.13correlation-0.273;鉴别Luminal B和非Luminal B型的模型为-0.008×ADCrbio1.11sumvariance-0.003×DCErbio3.13variance+3.204;鉴别TN和非TN的模型为-0.163×DCELG2.5autocorrelation+8.904,各个鉴别诊断的AUC分别为0.787 6、0.744和0.773。经Hosmer-Lemeshow检验各模型P值均> 0.05,各个模型预测值与观测值之间差异无统计学意义,模型拟合效果较好。结论:MRI影像学指标ADC、IER值鉴别乳腺癌分子亚型价值有限,DWI和DCE-MRI相关影像学特征鉴别乳腺癌分子亚型有一定价值。Objective:To explore the diagnostic value of DWI and dynamic contrast-enhanced MRI in patients with four different types of breast cancer.Methods:98 patients diagnosed as breast cancer in our hospital from september 2016 to september 2018 were selected,and all were scanned with DWI and DCE-MRI,and then underwent breast cancer surgery for pathological molecular typing.Mann Whitney U test and univariate logistic regression were used to analyze each variable and univariate.Multivariate logistic regression was used to further analyze and model the variables with statistical significance in univariate analysis.Receiver operating characteristic curve was used to evaluate the diagnostic efficiency of the model.Hosmer lemeshow test was used to test the goodness of the model.Results:There were 32 cases of luminal A,45 cases of luminal B,10 cases of HER-2 overexpression and 11 cases of TN.The AUC of ADC and IER were lower than 0.07 and were limited in differentiating molecular classification of breast cancer.The AUC of DCE_bior3.1_3_correlation was 0.732 when diagnosing luminal A breast cancer.The AUC of ADC_rbio1.1_1_sum_variance was 0.722 when diagnosing luminal B breast cancer.The AUC of ADC_L_G_2.5_min was 0.747 when diagnosing HER-2 overexpressing breast cancer.The AUC of five imaging features in differentiating LN breast cancer were all over 0.07.The model for identification of luminal A and non luminal A was 0.005×ADC_rbio1.1_1_sum_variance+0.032×DCE_bior3.1_3_correlation-0.273.The model for identification of luminal B and non luminal B was-0.008×ADC_rbio1.1_1_sum_variance-0.003×DCE_rbio3.1_3_variance+3.204.The model for identification of TN and non-TN was-0.163×DCE_L_G_2.5_autocorrelation+8.904.And the AUC of the three model was 0.7876,0.744 and 0.773.P value of each model was more than 0.05 by Hosmer lemeshow test.There was no statistical significance between the predicted value and the observed value of each model,and the model fitting effect was good.Conclusion:ADC and IER are limited in differentiating bre

关 键 词：乳腺癌 磁共振 分子亚型 诊断 

分 类 号：R737.9[医药卫生—肿瘤]