磁共振成像2025，Vol.16Issue(6)：34-41,47,9.DOI:10.12015/issn.1674-8034.2025.06.005
基于神经黑色素成像的纹理分析在帕金森病诊断中的价值研究
Study of the value of texture analysis based on neuromelanin imaging in the diagnosis of Parkinson's disease
摘要
Abstract
Objective:To evaluate the diagnostic value of neuromelanin magnetic resonance imaging(NM-MRI)combined with radiomics for Parkinson's disease(PD)and to explore specific imaging biomarkers.Materials and Methods:Fifty-seven PD patients and thirty-four healthy controls(HC)were included.The substantia nigra of the midbrain was used as the area of interest,the features were acquired using a modified 2D-T1-fast spin echo sequence(2D-T1-FSE).region of interest(ROI)segmentation was performed using 3D-Slicer 4.8.1,and features were selected through least absolute shrinkage and selection operator(LASSO)regression to construct a radiomics score,establishing a combined diagnostic model.Participants were divided into a training set and a test set in a 7∶3 ratio,comparing the diagnostic performance of the radiomics model,traditional clinical diagnostic models,and the combined model,and conducting correlation analyses between high-weighted features and some clinical scales.Results:Ten significant radiomics features(P<0.05)were selected;the area under the curve(AUC)for the combined model,radiomics model,and traditional clinical diagnostic model in the training set were 0.90[95%confidence interval(CI)0.83 to 0.98],0.89(0.81 to 0.97)and 0.70(0.56 to 0.83),the AUCs of the test groups were 0.88(0.76 to 1.00),0.88(0.81 to 1.00)and 0.78(0.60 to 0.96),respectively;the decision curve shows that imaging omics significantly improved clinical evaluation efficiency;in traditional clinical diagnostic models,Head_CNR and Whole_CNR showing significant statistical differences in diagnosing PD(P<0.01);LeastAxisLength,LargeAreaEmphasis,and Head_CNR are negatively correlated with H-Y grading but show no significant correlation with the Unified Parkinson's Disease Rating Scale Part Ⅲ(UPDRS-Ⅲ).Conclusions:The NM-MRI imaging omics combined model can significantly improve the accuracy of PD diagnosis,and the selected imaging omics features such as SurfaceVolumeRatio and Mean have the potential to become PD-specific markers.Meanwhile,Head_CNR and morphological parameters(LeastAxisLength and LargeAreaEmphasis)have clinical application value for the quantitative evaluation of disease severity.关键词
帕金森病/磁共振成像/神经黑色素成像/影像组学/纹理分析Key words
Parkinson's disease/magnetic resonance imaging/neuromelanin imaging/imaging omics/texture analysis分类
医药卫生引用本文复制引用
陈红,柏福运,安鹏,陈月,高平..基于神经黑色素成像的纹理分析在帕金森病诊断中的价值研究[J].磁共振成像,2025,16(6):34-41,47,9.基金项目
Key R&D Project of Hubei Province Science and Technology Plan in 2023(No.2023BCB140) （No.2023BCB140）
Innovative Research Program of Xiangyang No.1 People's Hospital(No.XYY2023SD16). 2023年湖北省科技计划项目重点研发专项(编号:2023BCB140) （No.XYY2023SD16）
湖北省襄阳市第一人民医院科技创新项目(编号:XYY2023SD16) （编号:XYY2023SD16）
