基于弥散加权成像影像组学在儿童颅内肿瘤分级诊断中的应用价值  

Application of Diffusion Weighted Imaging Imagomics in the Grading Diagnosis of Intracranial Tumors in Children

作　　者：马秋红 易婷 刘雨晴 金科 MA Qiuhong;YI Ting;LIU Yuqing(Department of Radiology,The Affiliated Children's Hospital of Xiangya School of Medicine,Central South University(Hunan children's hospital),Changsha Hunan 410007)

机构地区：[1]中南大学湘雅医学院附属儿童医院(湖南省儿童医院),湖南长沙410007

出　　处：《医学临床研究》2025年第1期37-40,共4页Journal of Clinical Research

基　　金：湖南省卫生健康委2022年度科研计划项目(202209012935)。

摘　　要：【目的】探讨基于弥散加权成像(DWI)影像组学在儿童低级别颅内肿瘤(LGG)和高级别颅内肿瘤(HGG)鉴别诊断中的应用价值。【方法】回顾性分析2015年1月至2021年4月在本院行手术治疗的92例颅内肿瘤患儿,根据患儿肿瘤分级情况将其分为低级别肿瘤组(LGG组,Ⅰ~Ⅱ级,n=53)和高级别肿瘤组(HGG组,Ⅲ~Ⅳ级,n=39)。两组均行DWI检查,观察其影像特征,使用MRMicro GL软件提取病灶特征,通过ITK-SNAP软件进行影像组学特征提取和模型评估,筛选出具有诊断价值且与肿瘤级别显著相关的影像组学特征,采用受试者工作特征(ROC)曲线评估表现弥散系数(ADC)、DWI单独诊断及ADC+DWI诊断颅内肿瘤分级的价值。【结果】从颅内肿瘤ADC、DWI图像中各选择108个影像组学特征进行降维筛选及模型评估,ADC模型曲线下面积(AUC)为0.7865,诊断敏感性、特异性分别为1.0000、0.5000;DWI模型AUC为0.7135,诊断敏感性、特异性分别为0.4375、1.0000,相关系数一致性一般;ADC+DWI模型的AUC为0.9479,相关系数一致性好,诊断效能最佳(P<0.05)。【结论】ADC、DWI及ADC+DWI模型可以鉴别儿童颅内肿瘤级别,ADC+DWI模型对高低级别肿瘤的鉴别效能较高,可以提高模型诊断性能,有助于指导临床医生手术及治疗方案的制定。【Objective】To explore the value of diffusion weighted imaging(DWI)based imaging-omics in the differential diagnosis of low-grade gliomas(LGG)and high-grade gliomas(HGG)in children.【Methods】A total of 92 children with intracranial tumors who underwent surgical treatment in our hospital from January 2015 to April 2021 were retrospectively analyzed and divided into low-grade gliomas group(LGG group)(gradeⅠ-Ⅱ,n=53)and high-grade gliomas group(HGG group)(gradeⅢ-Ⅳ,n=39)according to tumor grade.DWI examination was performed on both groups to observe their image features.MRMicro GL software was used to extract lesion features,and ITK-SNAP software was used to extract and evaluate the image omics features,which had diagnostic value and were significantly correlated with tumor grade.Receiver operating characteristic(ROC)curve was used to evaluate the value of ADC,DWI alone and ADC+DWI in the diagnosis of intracranial tumors.【Results】108 image features were selected from ADC and DWI images for dimensionality reduction screening and model evaluation.The ADC model area under the curve(AUC)was 0.7865,and the diagnostic sensitivity and specificity were 1.0000 and 0.5000,respectively.The AUC of DWI model was 0.7135,the diagnostic sensitivity and specificity were 0.4375 and 1.0000,respectively,and the correlation coefficients were generally consistent.The AUC of ADC+DWI model was 0.9479,the correlation coefficient was consistent well,and the diagnostic efficiency was the best(P<0.05).【Conclusion】ADC,DWI and ADC+DWI models can identify intracranial tumors in children,and ADC+DWI models have high efficiency in differentiating high and low grade gliomas,which can improve the diagnostic performance of the models and help to guide clinicians to make surgery and treatment plans.

关 键 词：脑肿瘤 磁共振成像 儿童 诊断 

分 类 号：R739.41[医药卫生—肿瘤]