不同方法测量扩散峰度成像和体素内不相干运动模型DWI参数诊断乳腺肿块性病变  

Diffusion kurtosis imaging and intravoxel incoherent motion DWI parameters measured with different methods for breast masses

作　　者：司丽芳[1] 刘小娟[1] 杨开颜[1] 马怡尘 王丽[1] 蒋涛[1] SI Lifang;LIU Xiaojuan;YANG Kaiyan;MA Yichen;WANG Li;JIANG Tao(Department of Radiology, Beijing Chao-yang Hospital,Capital Medical University, Beijing 100020, China)

机构地区：[1]首都医科大学附属北京朝阳医院放射科,北京100020

出　　处：《中国医学影像技术》2019年第5期706-710,共5页Chinese Journal of Medical Imaging Technology

摘　　要：目的探讨不同方法测量扩散峰度成像(DKI)及体素内不相干运动(IVIM)模型DWI参数鉴别诊断乳腺良恶性肿块性病变的价值。方法收集经病理或随访证实的59例乳腺肿块性病变患者(62个病变)。MR检查包括动态增强MRI、IVIM DWI和DKI。分别于动态增强MRI强化最明显处设置ROI,测量其标准扩散系数(ADC_(stand))、慢速扩散系数(ADC_(slow))、平均峰度值(mean kurtosis, MK)和平均扩散系数(mean diffusion, MD);于病灶实性部分最大层面沿病变边缘勾画ROI,测量病灶整体的ADC_(stand)-max、ADC_(slow)-max、ADC_(fast)-max、MK-max、MD-max,比较乳腺良恶性病变间各参数的差异,并绘制ROC曲线,比较AUC。结果 62个病变中,良性36个,恶性26个。良恶性病变间ADC_(stand)、ADC_(stand)-max、ADC_(slow)、ADC_(slow)-max、MK、MK-max、MD、MD-max差异均有统计学意义(P均<0.001)。ROC曲线结果显示ADC_(slow)联合MK的AUC最大(0.915),诊断乳腺良恶性病变的敏感度和特异度分别为88.9%和84.6%。ADC_(stand)与ADC_(stand)-max(Z=1.465,P=0.143)、ADC_(slow)与ADC_(slow)-max(Z=1.013,P=0.311)、MK与MK-max(Z=1.021,P=0.307)、MD与MD-max(Z=1.428,P=0.153)间AUC差异均无统计学意义。结论各DKI和IVIM DWI参数对鉴别乳腺良恶性肿块具有较高诊断价值,不同测量方法之间鉴别诊断效能无明显差异。Objective To investigate the value of diffusion kurtosis imaging(DKI) and intravoxel incoherent motion(IVIM) parameters measured using different methods in differentiating benign and malignant breast mass lesions. Methods Totally 59 patients(62 mass lesions) with breast mass lesions verified by pathologic results or follow up were enrolled. All patients underwent MR scanning, including dynamic contrast-enhanced MRI, IVIM DWI and DKI. ROI were placed at the most enhanced location, and the parameters of standard ADC(ADCstand), slow ADC(ADCslow), mean kurtosis(MK) and mean diffusion(MD) were measured. The whole lesions on the maximum slice were drawn as ROI, and the ADCstand-max, ADCslow-max, ADCfast-max, MK-max, MD-max were obtained. The differences of these parameters between benign and malignant breast lesions were analyzed. The diagnostic performance of these parameters was evaluated by ROC curve, the AUC was compared between the two methods. Results There were 26 malignant lesions and 36 benign lesions. All the parameters(ADCstand, ADCstand-max, ADCslow, ADCslow-max, MK, MK-max, MD, MD-max) were significantly different between malignant and benign lesions(all P<0.001). AUC of ADCslow combined MK was the highest(0.915), with the sensitivity of 88.9% and specificity of 84.6%. The differences of AUC between ADCstand and ADCstand-max(Z=1.465, P=0.143), ADCslow and ADCslow-max(Z=1.013, P=0.311), MK and MK-max(Z=1.021, P=0.307), MD and MD-max(Z=1.428, P=0.153) were not statistically significant. Conclusion For breast mass lesions, all DKI and IVIM DWI parameters are helpful to differentiating malignancy from benign lesions, and these parameters measured with different methods show equal diagnostic efficiency.

关 键 词：乳腺肿瘤 扩散磁共振成像 诊断 鉴别 

分 类 号：R737[医药卫生—肿瘤] R445[医药卫生—临床医学]