合成MRI定量参数图与ADC图直方图特征对乳腺癌Ki-67表达状态的预测研究  

Histogram features of quantitative parameters from synthetic MRI and ADC map in predicting the expression of Ki-67 in breast cancer

作　　者：李方正 李芹[2] 吴莎莎 孙胜君 于海童 陈永升 牛庆亮[2] LI Fangzheng;LI Qin;WU Shasha;SUN Shengjun;YU Haitong;CHEN Yongsheng;NIU Qingliang(School of Medical Imaging,Weifang Medical College,Weifang 261053,China;Center of Medical Imaging,Weifang Traditional Chinese Medicine Hospital,Weifang 261041,China)

机构地区：[1]潍坊医学院医学影像学院,潍坊261053 [2]潍坊市中医院影像中心,潍坊261041

出　　处：《磁共振成像》2022年第7期29-34,67,共7页Chinese Journal of Magnetic Resonance Imaging

基　　金：山东省自然科学基金(编号:ZR202103060229)。

摘　　要：目的评估合成MRI定量参数图直方图特征及表观扩散系数(apparent diffusion coefficient,ADC)图直方图特征预测乳腺癌Ki-67表达状态的价值。材料与方法回顾性分析2019年12月至2021年3月经潍坊市中医院病理证实为乳腺癌的146名患者的临床及影像学资料,所有患者活检或手术前均行MRI常规序列成像、动态对比增强MRI(dynamic contrast-enhanced MRI,DCE-MRI)及合成MRI序列扫描成像。使用PyRadiomics软件提取合成MRI定量参数T1、T2、质子密度(proton density,PD)值及ADC值这四个数值的直方图特征。根据Ki-67表达情况将乳腺癌患者分为高表达组(≥30%)与低表达组(<30%)。采用卡方检验、独立样本t检验及Mann-Whitney U检验比较两组患者的临床、病理特征及合成MRI定量参数图(T1-mapping、T2-mapping、PD-mapping)、ADC图的直方图特征差异。单因素及多因素logistic回归分析乳腺癌Ki-67表达状态与MRI定量参数图直方图特征之间的关系,并绘制受试者工作特征(receiver operating characteristic,ROC)曲线,计算曲线下面积(area under the curve,AUC),比较各直方图特征对Ki-67表达状态的预测效能。结果单因素logistic分析结果显示Ki-67高、低表达组间,ADC图直方图特征、临床及病理特征差异无统计学意义(年龄,P=0.13;最大径,P=0.09;形态,P=0.94;边缘,P=0.23;强化方式,P=0.13;纤维腺体类型,P=0.26),T1-平均值、T1-10^(th)位数(T1百分之十位数)、T2-平均值、T2-10th位数(T2百分之十位数)、PD-熵值和PD-峰度的差异均有统计学意义(P<0.01)。多因素logistic分析结果显示T1-10^(th)位数及T2-10^(th)位数是Ki-67表达状态的独立预测因素,两者联合构建模型预测乳腺癌Ki-67表达状态的AUC为0.809,敏感度为64.8%,特异度为87.5%,准确度为72.8%。结论合成MRI定量参数有助于预测乳腺癌Ki-67的表达状态,为术前无创性评估乳腺癌的肿瘤增殖情况提供了一种有效的辅助诊断方法。Objective:To evaluate the value of the histogram features of quantitative parameters from synthetic MRI and apparent diffusion coefficient(ADC)in predicting the expression of Ki-67 in breast cancer.Materials and Methods:The clinical and imaging data of 146 patients with breast cancer confirmed by pathology in our hospital from December 2019 to March 2021 were retrospectively analyzed.All patients underwent MRI routine sequence imaging,dynamic contrast-enhanced MRI(DCE-MRI)and synthetic MRI sequence scan imaging before biopsy or surgery.The histogram features of the quantitative parameters T1、T2、proton density(PD)of synthetic MRI and ADC values were extracted by PyRadiomics software.According to the expression of Ki-67,breast cancer patients were divided into high expression group(≥30%)and low expression group(<30%).Theχ2 test,independent sample t-test or Mann-Whitney U test were used to compare the differences of clinical and pathological characteristics,the histogram features of synthetic MRI quantitative parameter maps(T1-mapping,T2-mapping,PD-mapping)and ADC map between the two groups.Logistic regression analysis was used to analyze the relationship between the expression status of Ki-67 in breast cancer and quantitative parameters of MRI,and drawn the receiver operating characteristic(ROC)curve,calculated the area under curve(AUC)to compare the predictive efficacy of each histogram feature in predicting the expression status of Ki-67.Results:Univariate logistic analysis showed that there were no significant differences in the histogram characteristics of ADC map,clinical and pathological characteristics between the high and low expression groups of Ki-67(age,P=0.13;maximum diameter,P=0.09;shape,P=0.94;border,P=0.23;reinforcement mode,P=0.13;fibrous gland type,P=0.26).There was statistically significant difference in T1-mean,T1-10^(th) percentile,T2-mean,T2-10th percentile,PD-entropy,and PD-kurtosis between the high and low expression groups Ki-67 in breast cancer(P<0.01).Multivariate logistic analysis

关 键 词：乳腺癌 磁共振成像 定量 合成磁共振 细胞增殖指数 

分 类 号：R445.2[医药卫生—影像医学与核医学] R737.9[医药卫生—诊断学]