摘要
目的 探讨合成MRI(synthetic MRI, syMRI)技术在鉴别乳腺环形强化病变良恶性中的价值。材料与方法 收集MRI检查显示乳腺环形强化病例52个,共54个病灶,MRI检查包括平扫、增强syMRI MAGiC (magnetic resonance image complication)序列、扩散加权成像、动态对比增强MRI,MRI检查后1周内行手术或穿刺并取得病理结果。经syMRI后处理软件测量肿块增强前的T1(T1pre)、增强前的T2(T2pre)、增强前的质子密度(proton density, PD)值(PDpre)以及增强后的T1(T1post)、增强后的T2(T2post)、增强后的PD值(PDpost),计算增强前后的T1相对变化率(ΔT1%)、T2相对变化率(ΔT2%)、 PD值相对变化率(ΔPD%),测量乳腺环形强化病变环壁以及环内的表观扩散系数(apparentdiffusion coefficient, ADC)值。将病例根据病理结果分为良性组和恶性组,采用独立样本t检验和Wilcoxon秩和检验,分析上述syMRI定量值、ADC值在良恶性组别间的差异(P<0.05),采用χ2/Fisher精确检验分析良恶性组别间对比增强成像上的影像特征[形态、边缘、时间-信号强度曲线(time signal intensity curve, TIC)]的差异,将差异有统计学意义的参数分别使用logistic回归建立单因素及多因素预测模型,绘制ROC曲线并进行DeLong检验评价预测模型的诊断效能。结果 54个病灶中,恶性42个、良性12个。良性组T2pre [101.17 (28.42)]高于恶性组T2pre [88.67 (21.75)],良性组PDpost(101.90±14.60)大于恶性组PDpost(84.90±20.00),良性组与恶性组之间TIC类型(P<0.001)、环壁ADC值(P<0.001)的差异具有统计学意义(P<0.05),PDpost的ROC曲线下面积(area under the curve, AUC)为0.763,高于T2pre(AUC=0.758),低于TIC类型(AUC=0.883)、环壁ADC值(AUC=0.901)。上述参数单独应用时,环壁ADC值的诊断效能最高,参数联合应用时,四者联合的多因素模型诊断效能最高(AUC=0.998),TIC类型、T2pre、PDpost作为单因素模型,其诊断效能与四者联合的多因素模型诊断效能的差异具有�
Objective:To explore the application value of synthetic MRI(syMRI)in the differential diagnosis of benign and malignant breast lesions with peripheral rim enhancement.Materials and Methods:A total of 52 patients(54 breast lesions with peripheral rim enhancement)which underwent MRI examination were included in the study.All subjects underwent plain scan and enhanced scan of syMRI,diffusion weighted imaging scan and dynamic contrast-enhanced imaging MRI(DCE-MRI)scan,and then all of them received pathology results by surgery or biopsy within one week after MRI examination.The T1,T2,PD values of breast lesions pre-contrast and post-contrast injection(T1pre,T2pre,and PDpre;T1post,T2post,and PDpost)were estimated from syMRI data by using syMRI post-processing software.The relative change rate was calculated(ΔT1%,ΔT2%,ΔPD%).The apparent diffusion coefficient(ADC)values of the ring wall and inside the ring were measured.According to the result of pathology,all lesions were divided into benign and malignant groups.Independent sample t-test and Wilcoxon rank-sum test were used to analyze whether the data were significantly different(P<0.05).Theχ2 test or Fisher exact test were used to analyze the image features of DCE-MRI[morphology,edge and time-signal intensity curve(TIC)].Logistic regression was used to establish single-factor and multi-factor prediction models for statistically significant parameters,and the ROC curve of the subjects was drawn and the diagnosis efficiency of the prediction model was evaluated by DeLong test.Results:In 54 lesions,malignant 42 cases,12 cases were benign.The T2pre of benign groups[101.17(28.42)]higher than malignant group[88.67(21.75)],the PDpost of benign groups(101.90±14.60)higher than malignant group(84.90±20.00).There were significant differences in TIC(P<0.001)and the ADC values of the ring wall(P<0.001)between benign and malignant groups(P<0.05).The area under the curve(AUC)of PDpost was 0.763 and higher than T2pre(AUC=0.758),lower than TIC(AUC=0.883)and the ADC values of the
作者
童秋云
余雪燕
李迅
蒋宇
戴文海
周智鹏
TONG Qiuyun;YU Xueyan;LI Xun;JIANG Yu;DAI Wenhai;ZHOU Zhipeng(Department of Medical Imaging,Affiliated Hospital of Guilin Medical University,Guilin 541001,China)
出处
《磁共振成像》
CAS
CSCD
北大核心
2023年第10期78-83,共6页
Chinese Journal of Magnetic Resonance Imaging
基金
广西自筹经费科研课题项目(编号:Z20201476)。
关键词
乳腺
环形强化
鉴别诊断
合成磁共振成像
磁共振成像
breast
peripheral rim enhancement
differential diagnosis
synthetic magnetic resonance imaging
magnetic resonance imaging
