单指数及DKI模型在乳腺良恶性病变鉴别诊断中的价值

发布时间：2018-02-21 10:20

本文关键词： 乳腺肿瘤单指数模型 DKI模型　出处：《临床放射学杂志》2017年02期 　论文类型：期刊论文

【摘要】：目的探讨扩散加权成像(DWI)单指数模型及扩散峰度成像(DKI)模型各参数在乳腺良恶性病变鉴别诊断中的价值。方法回顾性分析经病理穿刺或手术证实的40例乳腺病变患者,共有46个病灶,其中良性病灶27个,恶性病灶19个。40例均行常规T_1WI、T_2WI、DWI、DKI及动态增强(DCE-MRI)扫描,通过Advantage Windows4.5后处理工作站获得由DWI扫描所得参数表观扩散系数(ADC)及由DKI扫描所得参数平均扩散系数(MD)、平均扩散峰度(MK)值。采用Mann-Whitney U秩和检验比较各参数在良性及恶性病变中的统计学差异,采用ROC曲线评价各参数的诊断效能,采用Spearman秩相关分析ADC值与MD值、MK值的相关性。结果恶性病变和良性病变的ADC值、MD值、MK值差异均有统计学意义。恶性病变ADC值、MD值明显低于良性病变(P值均0.05),恶性病变MK值明显高于良性病变(P0.05)。ADC值、MD值及MK值3个参数的ROC曲线下面积分别为0.857、0.876、0.896。以最大约登指数确定各参数最佳诊断阈值,分别以1.355×10~(-3)mm~2/s、1.215×10~(-3)mm~2/s、0.4925为阈值,判断乳腺良恶性病变的敏感性分别为85.2%、88.9%、89.5%,特异性分别为75.2%、76.8%、77.8%。ADC值与MD值呈正相关性,相关系数r=0.704,P0.05;与MK值呈负相关,相关系数r=-0.495,P0.05。结论乳腺DWI单指数模型及DKI模型能够对乳腺良恶性病变做出诊断,但DKI模型具有更高的敏感性及特异性,能够反映组织微环境的变化。
[Abstract]:Objective to evaluate the value of DWI single index model and DKI model in differential diagnosis of benign and malignant breast lesions. Methods 40 cases of breast lesions proved by pathological puncture or surgery were retrospectively analyzed. There were 46 lesions, including 27 benign lesions and 19 malignant lesions. All the patients underwent routine T1WII-T2WII-DKI and dynamic enhanced DCE-MRI scans. Advantage Windows4.5 postprocessing workstation was used to obtain the values of parameter apparent diffusion coefficient obtained from DWI scan and average diffusion coefficient from DKI scan. Mann-Whitney U rank sum test was used to compare the parameters in benign and evil. Statistical differences in sexual lesions, The ROC curve is used to evaluate the diagnostic efficiency of each parameter. Spearman rank correlation analysis was used to analyze the correlation between ADC value and MD value and MK value. Results there were significant differences in ADC value and MK value between malignant lesions and benign lesions. The ADC value of malignant lesions was significantly lower than that of benign lesions (P < 0.05), and that of malignant lesions was lower than that of benign lesions (P = 0.05). The area under the ROC curve of the MK value of the lesion was significantly higher than that of the benign lesion (P 0.05, ADC value, MD value and MK value), and the area under the ROC curve was 0.896. The best diagnostic threshold of each parameter was determined by the maximum Jordan index. Taking 1.355 脳 10 ~ (-1) ~ (-3) mm ~ (-1) / s ~ (-1) 215 脳 10 ~ (-1) ~ (-3) mm ~ (-1) ~ 0.4925 as the threshold value, the sensitivity of judging benign and malignant breast lesions was 85.2 ~ 88.9% and the specificity was 75.2 ~ 76.8mm ~ (-1) respectively. The correlation coefficient was 0.704 (P _ (0.05)), and the correlation coefficient was negative correlation with MK value, and the correlation coefficient was 0.704 (P ~ (0.05)), and the correlation coefficient was 0.704 (P ~ (0.05)), which was negatively correlated with MK value. Conclusion DWI single index model and DKI model can be used to diagnose benign and malignant breast lesions, but DKI model is more sensitive and specific and can reflect the changes of tissue microenvironment.
【作者单位】：郑州大学第一附属医院MRI室;
【分类号】：R445.2;R737.9

【相似文献】