Miao Lin, Qingze Zeng, Siyan Zhong, Shaoping Nie, R. Zhang, Hui Hong, Lingyun Liu, Shuyue Wang, Kaicheng Li, Xiao Luo, Yanxing Chen, Peiyu Huang, for the Alzheimer's Disease Neuroimaging Initiative (ADNI)
BACKGROUND AND OBJECTIVES: The choroid plexus (ChP) plays a crucial role in maintaining brain health. Alzheimer disease (AD) pathologies may damage the ChP and accelerate neurodegeneration. Previous imaging studies have found overall increased ChP volume in patients with AD. In this study, we aim to investigate AD-related ChP morphological features and their associations with cognitive decline. METHODS: This retrospective study used data from participants without dementia enrolled in the Alzheimer's Disease Neuroimaging Initiative, a large multicenter project. We included participants with complete 3D T1 and 3D fluid-attenuated inversion recovery MRI, β-amyloid (Aβ) and tau PET, and cognitive assessments. Follow-up cognitive data were collected when available. We developed a deep learning-based segmentation method for accurate ChP segmentation and analyzed the high-order morphological features of the ChP. First, we selected AD-related ChP features using partial correlation analyses. Then, partial correlation analyses and chain mediation analyses were performed to investigate the associations between the ChP features and cognitive scores. Finally, linear mixed models were used to investigate whether the ChP features could predict cognitive decline over time. RESULTS: = 0.022). DISCUSSION: Our approach captured the complex morphological alterations of the ChP in AD. These findings provide valuable knowledge that AD pathologies may disrupt ChP structure, contributing to clinical progression. The pathologic substrates underlying these ChP morphological features remain to be elucidated, necessitating further investigation through imaging-histopathological comparison studies.