基于表观扩散系数直方图分析预测多形性胶质母细胞瘤MGMT基因启动子甲基化状态  

ADC-based histogram analysis for prediction of MGMT gene promoter methylation status in glioblastoma multiform

作　　者：陈鑫[1] 张国琴 阮秀杭 魏新华[1] CHEN Xin;ZHANG Guoqin;RUAN Xiuhang;WEI Xinhua(Department of Radiology,Guangzhou First People’s Hospital,Guangzhou 510180,China)

机构地区：[1]广州市第一人民医院放射科,广东广州510180

出　　处：《实用放射学杂志》2020年第6期965-968,共4页Journal of Practical Radiology

基　　金：国家自然科学基金项目(81601469);广州市卫生健康科技项目(20191A011002);广州市科技计划项目(201804010032)。

摘　　要：目的探讨ADC直方图分析预测多形性胶质母细胞瘤(GBM)MGMT基因启动子甲基化状态的价值。方法回顾性收集经病理证实的119例GBM患者的术前MR图像和临床资料,其中61例患者为MGMT启动子甲基化,58例为非甲基化。所有患者术前均行MR DWI。使用Image J软件在肿瘤最大层面的ADC图上,分别手动勾画增强后T1 WI对应肿瘤强化部分和液体衰减反转恢复(FLAIR)序列对应的瘤周水肿部分ROI,计算相应ROI的直方图参数,包括平均数、标准差(SD)、模式、最小值、最大值、峰度和偏度等;采用Mann-Whitney U检验比较甲基化组与非甲基化组的ADC直方图参数的差异,对有统计学差异的参数,使用ROC曲线评价其预测肿瘤MGMT启动子甲基化状态的效能。结果MGMT启动子甲基化组的肿瘤强化部分的SD和最大值较非甲基化组小,且均有统计学意义(P<0.001)。其中SD鉴别MGMT启动子甲基化状态的AUC为0.777(95%CI:0.693~0.860),以17.003为临界值时,敏感度为0.737,特异度为0.717。最大值鉴别二者的AUC为0.703(95%CI:0.608~0.798),当临界值为157.5时,其敏感度和特异度分别为0.754和0.617。结论基于ADC直方图分析有助于预测GBM的MGMT基因启动子甲基化状态,其中肿瘤实质部分对应的ADC直方图参数SD诊断效能最佳。Objective To investigate the value of ADC-based histogram analysis in predicting the status of MGMT gene promoter methylation in glioblastoma multiform(GBM)patients.Methods Pre-surgical MR images and clinical data of 119 patients with pathologically confirmed GBM were retrospectively reviewed.Of 119 patients,61 patients were with MGMT promoter methylation and 58 patients with unmethylation.All patients underwent pre-operative MR DWI.The ROI of enhanced portion of tumor(referring to post-contrast enhanced T1 WI)and peritumoral edema(referring to FLAIR)were manually drawn by using Image J software on the largest tumor cross-section of ADC map.Then the histogram metrics,including mean,standard deviation(SD),model,minimum,maximum,kurtosis and skewness were calculated accordingly.Mann-Whitney U test was used to compare the difference of ADC histogram metrics between methylation and unmethylation groups.The performance of the metrics with statistical difference between two groups in predicting MGMT promoter methylation status were evaluated using ROC curve.Results SD and maximum value of enhanced portion of tumor with MGMT promoter methylation were statistically lower than those of unmethylation(P<0.001).The AUC of SD in predicting the status of MGMT promoter methylation was 0.777(95%CI:0.693-0.860).And the sensitivity and specificity were 0.737 and 0.717 when the cutoff value was 17.003.The AUC of maximum was 0.703(95%CI:0.608-0.798),and the sensitivity and specificity were 0.754 and 0.617 when the cutoff value was 157.5,respectively.Conclusion ADC-based histogram analysis is helpful in predicting MGMT gene promoter methylation status of GBM.And SD of enhanced portion of tumor has better performance.

关 键 词：表观扩散系数 直方图分析 胶质母细胞瘤 MGMT基因启动子 

分 类 号：R445.2[医药卫生—影像医学与核医学] R739.41[医药卫生—诊断学]