MRI定量参数联合血清HE4构建的logistic回归模型对子宫内膜样腺癌病理分级的预测价值  

The predictive value of a logistic regression model constructed by combining MRI quantitative parameters with serum HE4 for pathological grading of endometrioid adenocarcinoma

作　　者：岳晓宁 贺小玉 张亚婷 王成伟[1] YUE Xiao-ning;HE Xiao-yu;ZHANG Ya-ting(Department of CT and MRI,Xinjiang 832000,China)

机构地区：[1]石河子大学医学院第一附属医院CT/MRI室,新疆石河子832000

出　　处：《放射学实践》2023年第8期1026-1032,共7页Radiologic Practice

基　　金：石河子大学医学院第一附属医院青年基金项目(QN202114)。

摘　　要：目的:应用logistic回归分析评估动态增强MRI(DCE-MRI)半定量参数、表观扩散系数(ADC)值联合血清人附睾蛋白4(HE4)水平对子宫内膜样腺癌(EAC)病理分级的诊断价值。方法:回顾性搜集91例EAC患者[年龄32~76岁,平均(53.4±9.3)岁]的临床和影像资料。所有患者术前行3.0T盆腔MRI检查(包括T_(1)WI、T 2WI、DWI、DCE)及血清HE4检测。对DCE-MRI及DWI图像分别进行后处理,测量病灶的正性增强积分(PEI)、最大上升斜率(MSI)、达峰时间(TTP)、最大下降斜率(MSD)和ADC值。根据肿瘤分化程度,将EAC分成高级别组(G3)和低级别组(G1~G2),比较两组间各项MRI后处理参数及血清HE4水平,将差异有统计学意义的参数纳入二元logistic回归分析并建立联合诊断模型,并应用ROC曲线分析各单一有效参数及联合诊断模型对EAC病理分级的诊断效能。结果:ADC、PEI、MSI值及血清HE4水平在高级别组与低级别组EAC之间的差异具有统计学意义(P<0.05)。ADC、PEI、MSI值及血清HE4水平预测高级别EAC的ROC曲线下面积(AUC)分别为0.858、0.776、0.720和0.781;这4项指标建立的联合诊断模型预测高级别EAC的AUC为0.954,敏感度为92.86%,特异度为90.91%,且其诊断效能优于PEI、MSI值及血清HE4水平(Z=3.874、3.908、2.768,均P<0.05)。结论:DCE-MRI半定量参数、ADC及血清HE4水平可以用来术前预测高级别EAC,MRI多参数联合血清HE4水平建立的联合模型可进一步提高预测效能。Objective:To investigate the value of dynamic enhanced magnetic resonance imaging(DCE-MRI)semi-quantitative parameters,apparent diffusion coefficient(ADC)value combined with human epididymis protein 4(HE4)level in predicting the pathological grade of endometrioid adenocarcinoma(EAC)by logistic regression.Methods:The clinical and MR imaging data of 91 patients with EAC[mean:(53.4±9.3)years,range:32~76 years)]were retrospectively collected.All patients underwent pelvic MRI scanning(including T_(1)WI,T_(2)WI,DWI and DCE-MRI)at a 3.0T MR scanner and serum HE4 detection before surgery.The positive enhancement integral(PEI),maximum slope of increase(MSI),time to peak(TTP),maximum slope of decrease(MSD)and ADC values were obtained by post-processing of DCE and DWI images respectively.EAC was divided into high-grade group(G3)and low-grade group(G1~G2)according to the degree of differentiation.By comparing MRI postprocessing parameters and serum HE4 levels between the two groups,parameters with statistically signi-ficant differences were included in binary logistic regression analysis and a combined diagnosis model was established,and receiver operating characteristic(ROC)curve was used to analyze the diagnostic efficacy of single effective parameter and the combined diagnosis model for EAC pathological grading.Results:There were significant differences in ADC values,PEI,MSI and HE4 between high-grade and low-grade(P<0.05).The area under the curve(AUC)of ADC,PEI,MSI and HE4 for the diagnosis of high-grade EAC were 0.858,0.776,0.720 and 0.781,respectively.The combined diagnostic model established by these four indicators predicted that the AUC of high-grade EAC was 0.954,the sensitivity was 92.86%,and the specificity was 90.91%,and its diagnostic efficiency was better than PEI,MSI value and serum HE4(Z=3.874,3.908 and 2.768,all P<0.05).Conclusion:DCE-MRI semi-quantitative parameters,ADC value and serum HE4 levels can be used for preoperative prediction of high-grade EAC,and the combined model established by MRI multi-para

关 键 词：子宫内膜样腺癌 肿瘤分级 磁共振动态增强成像 扩散加权成像 人附睾蛋白4 诊断效能 

分 类 号：R445.2[医药卫生—影像医学与核医学] R737.33[医药卫生—诊断学]