扩散峰度成像直方图联合动态对比增强MRI对乳腺良恶性病变的诊断价值  

The value of histogram analysis of diffusion kurtosis imaging combined with dynamic contrast-enhanced MRI in differential diagnosis of benign and malignant breast lesions

作　　者：牟方胜 蔡小丽 温云 李建蓉 MOU Fangsheng;CAI Xiaoli;WEN Yun;LI Jianrong(Department of Radiology,Three Gorges Hospital Affiliated to Chongqing University,Chongqing 404000,China)

机构地区：[1]重庆大学附属三峡医院放射科,重庆404000

出　　处：《实用放射学杂志》2023年第10期1611-1616,共6页Journal of Practical Radiology

摘　　要：目的评估扩散峰度成像(DKI)直方图联合动态增强磁共振成像(DCE-MRI)定量参数对乳腺良恶性病变的诊断价值。方法选取术前行多模态MRI检查并获得明确病理诊断的155例乳腺病变患者(共160个病灶)。利用第三方后处理软件得到DKI模型的平均峰度(MK)、平均扩散率(MD)直方图参数及DCE-MRI的容量转移常数(K^(trans))、速率常数(K_(ep))及血管外细胞外间隙容积比(V_(e))。采用两独立样本t检验或Mann-Whitney U检验比较乳腺良恶性病灶的K^(trans)、K_(ep)、V_(e)、MK及MD直方图参数的差异。根据病理结果,绘制受试者工作特征(ROC)曲线评价有统计学意义的参数独立和联合诊断乳腺良恶性病变的价值,Z检验评价联合诊断曲线下面积(AUC)的差异。单变量logisti回归分析(P<0.001)的变量纳入多变量logisti回归分析独立影响因素。结果乳腺恶性病变118个,良性病变42个。MD直方图[第10百分位数(P_(10th))、第90百分位数(P_(90th))、熵、峰度、均值、中位数、最小值、偏度]、MK直方图(P_(10th)、P_(90th)、熵、峰度、最大值、均值、中位数、最小值)、K^(trans)、K_(ep)、V_(e)鉴别乳腺良恶性病变单因素分析有统计学意义(P<0.05)。AUC最大为K_(ep)(0.782),灵敏度最高为MK-P_(90th)(0.898),特异度最高为MK-P_(90th)和MK-熵(0.786),准确度最高为MD-熵(0.80)。DCE-MRI参数联合及DKI直方图联合DCE-MRI诊断的AUC、灵敏度、特异度、准确度分别为0.775(0.935)、0.712(0.949)、0.738(0.833)、0.781(0.906)。DKI直方图联合DCE-MRI的AUC显著大于DCE-MRI、MD、MK各参数联合(P<0.05)。MD-熵[比值比(OR)=4.75]和MD-偏度(OR=3.67)为乳腺恶性病变的独立危险因数。结论DKI联合DCE-MRI诊断乳腺良恶性病变的价值显著大于单一参数。MD-熵和MD-偏度为乳腺恶性病变的独立危险因数。Objective To assess the value of histogram analysis of diffusion kurtosis imaging(DKI)combined with dynamic contrast-enhanced magnetic resonance imaging(DCE-MRI)in differential diagnosing benign and malignant breast lesions.Methods 155 patients with 160 breast lesions confirmed by histopathology underwent preoperative multimodal MRI examination.The histogram of mean kurtosis(MK)and mean diffusivity(MD)of the DKI model,and the volume transfer constant(K^(trans)),rate transfer constant(K_(ep))and volume of extravascular extracellular space per unit volume of tissue(V_(e))were performed using other post processing software.The K^(trans),K_(ep),V_(e),MK and MD histogram parameters of the benign and malignant breast lesions were compared by Student’s t test or the Mann-Whitney U test.According to the pathological results,receiver operating characteristic(ROC)curve was drawn to evaluate the value of the independent parameter with statistical significance and combination in diagnosis of malignant from benign breast lesions.The Z test was used to evaluate the area under the curve(AUC)of combined diagnosis.The variables of univariable logistic regression analysis(P<0.001)were included in the multivariate logistic regression analysis to analyze independent influencing factors.Results There were 118 malignant breast lesions and 42 benign lesions.MD histogram parameters[10th percentile(P_(10th)),90th percentile(P_(90th)),entropy,kurtosis,mean,median,minimum,skewness],MK histogram parameters(P_(10th),P_(90th),entropy,kurtosis,maximum,mean,median,minimum),Ktrans,Kep,and Ve were statistically significant in differentiating benign and malignant breast lesions by using single factor analysis(P<0.05).The maximum AUC was K_(ep)(0.782);the highest sensitivity was MK-P_(90th)(0.898);the highest specificity were MK-P_(90th) and MK-entropy(0.786);the highest accuracy was MD-entropy(0.80).The AUC,sensitivity,specificity and accuracy of DCE-MRI parameters and DKI histogram combined with DCE-MRI were 0.775(0.935),0.712(0.949),0.738(0.83

关 键 词：乳腺病变 扩散峰度成像 直方图 动态对比增强磁共振成像 

分 类 号：R655.8[医药卫生—外科学] R445.2[医药卫生—临床医学]