体素内不相干运动磁共振成像在乳腺病变的诊断及预后评估中的价值

发布时间：2018-06-01 02:20

本文选题：扩散加权成像 + 体素内不相干运动成像　；参考：《天津医科大学》2017年硕士论文

【摘要】：目的:比较体素内不相干运动(IVIM)磁共振成像双指数模型、拉伸指数模型与传统扩散加权成像(DWI)单指数模型各相关参数在乳腺良、恶性病变的鉴别诊断中的价值,并由ROC曲线确定各参数诊断良、恶性病变的最佳阈值,并进一步分析IVIM各参数与乳腺癌预后因子及分子分型之间的相关性。方法:回顾性分析257例于天津医科大学肿瘤医院行乳腺MRI检查并经病理证实的乳腺病变患者(共276个病灶,其中包括197个恶性病变,79个良性病变),同时选取正常的乳腺组织作为对照组。所有患者均行MRI常规检查及IVIM(多b值DWI)检查。由单指数模型、双指数模型及拉伸指数模型分别获得并分析各个病变的表观扩散系数(ADC),慢速表观扩散系数(slow ADC),快速表观扩散系数(fast ADC),灌注分数(f),扩散分布指数(DDC)及扩散异质性指数(α)。各参数在正常乳腺对照组、乳腺良性病变组及恶性病变组中的差异采用K个独立样本检验进行分析,P0.05认为差异有统计学意义;进一步行两两比较(恶性组VS良性组,恶性组VS正常对照组,良性组VS正常对照组)采用Mann-Whitney U检验,P0.02认为差异有统计学意义。采用受试者工作特征(ROC)曲线评价各参数对乳腺良、恶性病变的诊断效能并确定诊断恶性病变的最佳阈值及其在该阈值下各参数诊断的敏感性、特异性及准确性。分析187个乳腺癌的病理类型及免疫组化指标,将10例导管内癌及151例浸润性导管癌纳入第二部分的研究,采用非参数U检验及Kruskall-Wallis检验进一步分析IVIM各参数在乳腺癌分子分型、预后因子之间的差异。结果:由常规DWI得到的ADC值和由IVIM模型得到的slow ADC、fast ADC、f、DDC及α在正常乳腺组织、乳腺良性病变及恶性病变中差异均有统计学意义(P0.001),进一步两两比较,乳腺恶性病变与良性病变及正常腺体组织中ADC、slow ADC、fast ADC、f、DDC和α差异均有统计学意义(P0.001),且ADC、slow ADC、f、DDC和α在正常乳腺组织、乳腺良性病变、恶性病变中依次减低,fast ADC值在正常乳腺组织、乳腺良性病变、恶性病变中依次升高;ADC、slow ADC、DDC和α在乳腺正常腺体组织和乳腺良性病变之间差异有统计学意义(P0.001)。由ROC曲线得到ADC、slow ADC、fast ADC、f、DDC和α诊断乳腺良、恶性病变的最佳诊断阈值分别为1.105×10-3mm2/s、0.883×10-3mm2/s、4.750×10-3mm2/s、0.215、1.025×10-3mm2/s和0.842,诊断的敏感性分别为87.2%、90.3%、68.5%、64.5%、81.6%、76.0%,特异性分别为74.4%、74.4%、54.4%、75.9%、79.5%、67.9%,ROC曲线下面积(AUC)分别为0.865、0.861、0.611、0.742、0.850和0.735。与传统的DWI比较,slow ADC值的敏感性(90.3%)较ADC值(87.2%)高,DDC的特异性(79.5%)较ADC值(74.4%)高,但差异均无统计学意义。双指数模型中slow ADC值与fast ADC值联合诊断的曲线下面积为0.882;拉伸指数模型DDC与α联合诊断的曲线下面积为0.853,差异均无统计学意义。Ki-67低表达组乳腺癌slow ADC值高于Ki-67高表达组(P值为0.045);Her-2过表达型乳腺癌较其他类型乳腺癌患者的slow ADC值更高(P值为0.014);乳腺浸润性导管癌较导管内癌slow ADC、f、DDC及ADC值更低(P值分别为0.013,0.028,0.01,0.003)。结论:传统DWI和IVIM磁共振成像对乳腺良、恶性病变的诊断均具有较高价值,IVIM双指数模型中slow ADC值与fast ADC值联合诊断效能较高。IVIM相关参数对预测乳腺癌的Ki-67指数、分子分型及病理浸润与否具有潜在的价值。
[Abstract]:Objective: To compare the value of the two exponential model of the incoherent motion (IVIM) magnetic resonance imaging, the parameters of the tensile index model and the traditional diffusion-weighted imaging (DWI) single index model in the differential diagnosis of benign and malignant breast lesions, and to determine the best threshold for the diagnosis of benign and malignant lesions by the ROC curve, and to further analyze the IVIM The correlation between the parameters and the prognostic factors and molecular typing of breast cancer. Methods: a retrospective analysis of 257 cases of breast lesions in the Cancer Hospital of Medical University Of Tianjin (276 lesions, including 197 malignant lesions, 79 benign diseases), and normal breast tissue as control, were performed in 257 cases of the breast cancer hospital in Medical University Of Tianjin. MRI routine examination and IVIM (multi b value DWI) examination were performed in all patients. The apparent diffusion coefficient (ADC), slow apparent diffusion coefficient (slow ADC), rapid apparent diffusion coefficient (fast ADC), perfusion fraction (f), diffusion distribution index (DDC) and diffusion difference were obtained from single index model, double exponential model and tensile index model respectively. Qualitative index (alpha). The difference of the parameters in the normal breast control group, the benign breast lesion group and the malignant lesion group was analyzed by K independent sample test, and P0.05 thought the difference was statistically significant. Further 22 comparison (malignant group VS benign group, malignant group VS normal control group, benign group VS normal control group) adopted Mann-Whitney U Test, P0.02 thought the difference was statistically significant. Using the ROC curve to evaluate the diagnostic effectiveness of the parameters for benign and malignant breast lesions and to determine the best threshold for the diagnosis of malignant lesions and the sensitivity, specificity and accuracy of the diagnosis of all parameters under this threshold. The pathological types and immune groups of 187 breast cancers were analyzed. 10 cases of intraductal carcinoma and 151 cases of invasive ductal carcinoma were included in the study. The non parametric U test and Kruskall-Wallis test were used to further analyze the differences between the IVIM parameters in the molecular classification of breast cancer and the prognostic factors. Results: the ADC values obtained from the conventional DWI and the slow ADC, fast ADC, F, DDC and alpha obtained from the IVIM model were obtained. There were significant differences in normal breast tissue, benign breast lesions and malignant lesions (P0.001). Further 22, the differences of ADC, slow ADC, fast ADC, F, DDC and alpha in benign lesions and normal gland tissues were statistically significant (P0.001), and ADC, slow ADC, and alpha in normal mammary tissues and mammary glands. Fast ADC values in normal breast tissue, benign breast lesions and malignant lesions were increased in turn, and ADC, slow ADC, DDC and alpha were statistically significant (P0.001) between normal mammary gland tissue and benign breast lesions (P0.001). The best diagnostic thresholds were 1.105 x 10-3mm2/s, 0.883 x 10-3mm2/s, 4.750 x 10-3mm2/s, 0.215,1.025 x 10-3mm2/s and 0.842, and the diagnostic sensitivity was 87.2%, 90.3%, 68.5%, 64.5%, 81.6%, 76%, respectively. The specificity of the specificity was 0.865,0.861,0.611,0.742,0.850 and the area under the ROC curve (AUC), respectively. Compared with the traditional DWI, the sensitivity of the slow ADC value (90.3%) was higher than that of the ADC value (87.2%), and the specificity of the DDC (79.5%) was higher than the ADC value (74.4%), but the difference was not statistically significant. The area under the curve under the joint diagnosis of slow ADC value and fast ADC was 0.882 in the double exponential model, and the area under the curve under the joint diagnosis of the tensile exponential model DDC and alpha was 0.853, The difference was not statistically significant in.Ki-67 low expression group, the value of slow ADC in breast cancer was higher than that in Ki-67 high expression group (P value was 0.045); Her-2 overexpressed breast cancer was higher on slow ADC than other types of breast cancer patients (P value was 0.014), and invasive ductal carcinoma of the breast was higher than that of ductal carcinoma slow ADC. Conclusion: traditional DWI and IVIM MRI are of high value for the diagnosis of benign and malignant breast lesions. The combination of slow ADC value and fast ADC value in IVIM double index model is of potential value for predicting Ki-67 index, molecular typing and pathological infiltration of breast cancer.
【学位授予单位】：天津医科大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R445.2;R737.9

【参考文献】