摘要
目的基于磁共振扩散张量成像(Diffusion Tensor Imaging,DTI)数据,构建区分帕金森(Parkinson’s Disease,PD)患者和健康对照组的机器学习模型,探索其脑影像生物学标志物。方法收集我院2015年6月至2019年12月PD患者289例,正常对照志愿者131例。按照7:3的比例分为训练集和验证集。基于DTI数据构建包含各向异性分数、平均扩散率、轴向弥散系数、径向弥散系数值的向量。使用机器学习方法对该向量降维,构建分类模型并进行模型评估。结果构建的五种机器学习分类模型中,使用12个DTI脑区参数的SVMLinear模型具有最优的分类性能。模型评估结果显示,训练集中AUC为0.897,敏感度为83.3%,特异度为89.0%;验证集中AUC为0.878,敏感度为79.3%,特异度为88.4%。结论基于DTI数据构建的机器学习分类模型能有效区分PD患者和健康对照者,胼胝体、扣带回、穹窿等脑区的DTI参数有作为PD的影像学标志物的潜力。
Objective To establish a machine learning model that can distinguish Parkinson’s disease(PD)patients from healthy controls based on the data of MR diffusion tensor imaging(DTI),and to explore the brain imaging and biological markers of PD patients.Methods A total of 289 patients with PD in our hospital from June 2015 to December 2019 were collected as PD group,and 131 healthy controls were recruited as control group.All subjects were divided into training set and validation set according to the ratio of 7:3.Vectors which included parameters such as fractional anisotropy,mean diffusivity,axial and radial diffusion coefficient was constructed based on the data of DTI.The dimension of vectors was reduced based on machine learning methods to build a classification model,and the model evaluation was performed by receiver operating characteristic.Results Among the five machine learning classification models constructed,the SVMLinear model using twelve parameters of DTI for brain region had the best classification performance.The results of model evaluation showed that the area under curve(AUC)in the training set was 0.897,the sensitivity was 83.3%,and the specificity was 89.0%.The AUC in the validation set was 0.878,the sensitivity was 79.3%,and the specificity was 88.4%.Conclusion The machine learning classification model based on the data of DTI can effectively distinguish PD patients from healthy controls.The DTI parameters of corpus callosum,cingulate gyrus,fornix and other brain regions have the potential to be used as imaging markers of PD.
作者
李静
范文亮
雷子乔
余建明
LI Jing;FAN Wenliang;LEI Ziqiao;YU Jianming(Department of Radiology,Union Hospital,Tongji Medical College,Huazhong University of Science and Technology,Wuhan Hubei 430022,China;Hubei Province Key Laboratory of Molecular Imaging,Wuhan Hubei 430022,China)
出处
《中国医疗设备》
2021年第10期32-35,共4页
China Medical Devices
基金
国家自然科学基金(81701673)
湖北省自然科学基金(2017CFB796)。
关键词
弥散张量成像
机器学习
帕金森
磁共振扩散张量成像
diffusion tensor imaging
machine learning
Parkinson’s disease
MR diffusion tensor imaging
