结构网络拓扑属性改变对早期帕金森病伴轻度认知障碍的预测价值

Predictive value of alterations of brain structural network topology in early-stage Parkinson’s disease with mild cognitive impairment

目的利用弥散张量成像(diffusion tensor imaging, DTI)探讨脑白质网络拓扑属性改变对早期帕金森病(Parkinson′s disease,PD)患者发生轻度认知障碍(mild cognitive impairment,MCI)的潜在预测价值。材料与方法 从Parkinson′s Progression Markers Initiative (PPMI)数据库中纳入83例在基线时认知正常的PD患者,且均行4年的随访;4年后,有26例转化为PD-MCI,其余57例仍保持正常认知(PD patients with normal cognition,PD-NC)。采用图论分析研究PD-MCI患者结构网络全局和局部拓扑属性的改变;采用受试者工作特征曲线和向前逐步逻辑回归分析评估网络拓扑属性和认知量表对PD-MCI的预测价值。结果 PD-MCI患者在第4年随访时网络属性较PD-NC患者发生广泛改变。PD-MCI患者4年随访后的全局效率、局部效率较基线时显著降低,同时伴有特征路径长度的增加(P<0.05);且在前额叶、颞叶、枕叶、顶叶及纹状体-边缘系统多个节点的中心性降低(P<0.05)。此外,右侧眶内额上回的节点中心性的降低以及蒙特利尔认知评估量表、数字字母排序测试是PD-MCI的重要预测指标(P<0.01)。结论 右侧眶内额上回的局部网络属性的改变可作为PD-MCI影像学预测标志物,且白质结构网络拓扑属性对早期PD患者认知功能下降具有重要的预测价值。

Objective: Useing diffusion tensor imaging(DTI) to explore the potential predictive value of changes of white matter(WM) structural network topological properties on mild cognitive impairment in early-stage Parkinson’s disease(PD). Materials and Methods: Eighty-three PD patients with normal cognition at baseline were included from the Parkinson’s Progression Markers Initiative(PPMI) database, and all completed a 4-year follow-up. Among the 83 PD patients, 26 developed mild cognitive impairment(PD-MCI) and 57 retained normal cognition(PD-NC). Graph theory was utilized to evaluate the structural WM networks alterations in PD-MCI, and receiver operating characteristic analysis followed by stepwise logistic regression were performed to assess the predictive performance of network topology properties and cognitive measures. Results: The patients with PD-MCI showed longitudinal decreased global efficiency and local efficiency, increased characteristic path length(P<0.05). Locally, patients with PD-MCI exhibited longitudinal reduced nodal centralities, mainly in the frontal, temporal, occipital, parietal and striatal-limbic system regions over time(P<0.05). Moreover, the longitudinal decline in the degree centrality and nodal efficiency of the right medial orbital superior frontal gyrus, and patient Montreal Cognitive Assessment and Letter–Number Sequencing scores predicted the development of cognitive impairment in early-stage PD(P<0.01). Conclusions: The current study indicates that local network properties in the right medial orbital superior frontal gyrus can predict the onset of cognitive impairment in PD, and highlighting the value of network topology properties as sensitive biomarkers of cognitive decline in early-stage PD patients.