基于三维密集连接卷积网络鉴别高级别胶质瘤与单发脑转移瘤  

Differentiation Between High-Grade Glioma and Single Brain Metastases Based on Three-Dimensional DenseNet

作　　者：张斌 黄陈翠 薛彩强 李昇霖 周俊林[1] ZHANG Bin;HUANG Chencui;XUE Caiqiang;LI Shenglin;ZHOU Junlin(Department of Radiology,Lanzhou University Second Hospital,Second Clinical School,Lanzhou University,Key Laboratory of Medical Imaging of Gansu Province,Gansu International Scientific and Technological Cooperation Base of Medical Imaging Artificial Intelligence,Lanzhou 730030,China)

机构地区：[1]兰州大学第二医院放射科,兰州大学第二临床医学院,甘肃省医学影像重点实验室,医学影像人工智能甘肃省国际科技合作基地,甘肃兰州730030 [2]北京深睿博联科技有限责任公司,北京100080

出　　处：《中国医学影像学杂志》2024年第2期119-124,共6页Chinese Journal of Medical Imaging

基　　金：国家自然科学基金面上项目(82071872,82371914);甘肃省科技计划项目(21YF5FA123)。

摘　　要：目的探讨三维密集连接卷积网络(3D-DenseNet)通过MRI鉴别诊断高级别胶质瘤(HGGs)与单发脑转移瘤(BMs)的价值,并比较不同序列建立的模型诊断性能。资料与方法回顾性收集兰州大学第二医院2016年6月—2021年6月经手术病理证实的230例HGGs和111例BMs的T2WI及T1WI对比增强(T1C)影像资料,预先勾画出三维模型下的感兴趣区体积作为输入数据,以7∶3随机分为训练集254例和验证集87例,基于3D-DenseNet分别构建T2WI、T1C及两种序列融合的预测模型(T2-net、T1C-net和TS-net),通过受试者工作特征曲线评价各模型的预测效能并进行比较。结果T1C-net、T2-net和TS-net在训练集和验证集的曲线下面积(AUC)分别为0.852、0.853,0.802、0.721,0.856、0.745。T1C-net在验证集的AUC及准确度高于T2-net和TS-net,TS-net在验证集的AUC及准确度高于T2-net,T1C-net与T2-net在验证集的AUC差异有统计学意义(P<0.05),而TS-net与T2-net、T1C-net与TS-net的AUC差异无统计学意义(P>0.05)。基于3D-DenseNet的T1C-net模型性能最优,验证集的准确度为80.5%,敏感度为90.9%,特异度为62.5%。结论基于MRI常规序列的3D-DenseNet模型具有较好的诊断效能,在鉴别HGGs和BMs时T1C-net序列所建模型性能更好,深度学习模型可成为鉴别HGGs和BMs并指导临床制定精准化治疗方案的潜在工具。Purpose To explore the value of three-dimensions densely connected convolutional networks(3D-DenseNet)in the differential diagnosis of high-grade gliomas(HGGs)and single brain metastases(BMs)via MRI,and to compare the diagnostic performance of models built with different sequences.Materials and Methods T2WI and T1WI contra-enhanced(T1C)imaging data of 230 cases of HGGs and 111 cases of BMs confirmed by surgical pathology in Lanzhou University Second Hospital from June 2016 to June 2021 were retrospectively collected,and the volume of interest under the 3D model was delineated in advance as the input data.All data were randomly divided into a training set(n=254)and a validation set(n=87)in a ratio of 7∶3.Based on the 3D-DenseNet,T2WI,T1C and two sequence fusion prediction models(T2-net,T1C-net and TS-net)were constructed respectively.The predictive efficiency of each model was evaluated and compared by the receiver operating characteristic curve,and the predictive performance of models built with different sequences were compared.Results The area under curve(AUC)of T1C-net,T2-net and TS-net in the training and validation sets were 0.852,0.853,0.802,0.721,0.856 and 0.745,respectively.The AUC and accuracy of the validation set of T1C-net were significantly higher than those of T2-net and TS-net,respectively,and the AUC and accuracy of the validation set of TS-net were significantly higher than those of T2-net.There was a significant difference between T1C-net and T2-net models(P<0.05),while there were no statistical differences between the models of TS-net and T2-net,T1C-net and TS-net(P>0.05).The T1C-net model based on 3D-DenseNet had the best performance,the accuracy of the validation set was 80.5%,the sensitivity was 90.9%,the specificity was 62.5%.Conclusion The 3D-DenseNet model based on MRI conventional sequence has better diagnostic performance,and the model built by T1C-net sequence has better performance in differentiating HGGs and BMs.Deep learning models can be a potential tool to identify HGGs and BMs

关 键 词：高级别胶质瘤 脑转移瘤 磁共振成像 深度学习 诊断 鉴别 

分 类 号：R739.41[医药卫生—肿瘤] R445.2[医药卫生—临床医学]