Choroid plexus volume as a novel candidate neuroimaging marker of the Alzheimer's continuum

Abstract

Background: Enlarged choroid plexus (ChP) volume has been reported in patients with Alzheimer's disease (AD) and inversely correlated with cognitive performance. However, its clinical diagnostic and predictive value, and mechanisms by which ChP impacts the AD continuum remain unclear.

Methods: This prospective cohort study enrolled 607 participants [healthy control (HC): 110, mild cognitive impairment (MCI): 269, AD dementia: 228] from the Chinese Imaging, Biomarkers, and Lifestyle study between January 1, 2021, and December 31, 2022. Of the 497 patients on the AD continuum, 138 underwent lumbar puncture for cerebrospinal fluid (CSF) hallmark testing. The relationships between ChP volume and CSF pathological hallmarks (Aβ₄₂, Aβ₄₀, Aβ_42/40, tTau, and pTau₁₈₁), neuropsychological tests [Mini-Mental State Examination (MMSE), Montreal Cognitive Assessment (MoCA), Neuropsychiatric Inventory (NPI), and Activities of Daily Living (ADL) scores], and multimodal neuroimaging measures [gray matter volume, cortical thickness, and corrected cerebral blood flow (cCBF)] were analyzed using partial Spearman's correlation. The mediating effects of four neuroimaging measures [ChP volume, hippocampal volume, lateral ventricular volume (LVV), and entorhinal cortical thickness (ECT)] on the relationship between CSF hallmarks and neuropsychological tests were examined. The ability of the four neuroimaging measures to identify cerebral Aβ₄₂ changes or differentiate among patients with AD dementia, MCI and HCs was determined using receiver operating characteristic analysis, and their associations with neuropsychological test scores at baseline were evaluated by linear regression. Longitudinal associations between the rate of change in the four neuroimaging measures and neuropsychological tests scores were evaluated on the AD continuum using generalized linear mixed-effects models.

Results: The participants' mean age was 65.99 ± 8.79 years. Patients with AD dementia exhibited the largest baseline ChP volume than the other groups (P < 0.05). ChP volume enlargement correlated with decreased Aβ₄₂ and Aβ₄₀ levels; lower MMSE and MoCA and higher NPI and ADL scores; and lower volume, cortical thickness, and cCBF in other cognition-related regions (all P < 0.05). ChP volume mediated the association of Aβ₄₂ and Aβ₄₀ levels with MMSE scores (19.08% and 36.57%), and Aβ₄₂ levels mediated the association of ChP volume and MMSE or MoCA scores (39.49% and 34.36%). ChP volume alone better identified cerebral Aβ₄₂ changes than LVV alone (AUC = 0.81 vs. 0.67, P = 0.04) and EC thickness alone (AUC = 0.81 vs.0.63, P = 0.01) and better differentiated patients with MCI from HCs than hippocampal volume alone (AUC = 0.85 vs. 0.81, P = 0.01), and LVV alone (AUC = 0.85 vs.0.82, P = 0.03). Combined ChP and hippocampal volumes significantly increased the ability to differentiate cerebral Aβ₄₂ changes and patients among AD dementia, MCI, and HCs groups compared with hippocampal volume alone (all P < 0.05). After correcting for age, sex, years of education, APOE ε4 status, eTIV, and hippocampal volume, ChP volume was associated with MMSE, MoCA, NPI, and ADL score at baseline, and rapid ChP volume enlargement was associated with faster deterioration in NPI scores with an average follow-up of 10.03 ± 4.45 months (all P < 0.05).

Conclusions: ChP volume may be a novel neuroimaging marker associated with neurodegenerative changes and clinical AD manifestations. It could better detect the early stages of the AD and predict prognosis, and significantly enhance the differential diagnostic ability of hippocampus on the AD continuum.

Keywords: Alzheimer’s disease; Amyloid-beta; Choroid plexus; Longitudinal studies; Mediation analysis; ROC curve.

Fig. 1

Comparisons among three representative 3D-T1 weighted images of the ChP volume in the three groups. The ChP volume (red) is the largest in patients with AD dementia, followed by those with MCI and HCs. All participants are 72-year-old men. The MMSE and MoCA scores of the HC are 30 and 28, respectively (A). The MoCA score is 23 in patients with MCI with 10 years of education (B), whereas the MoCA score is 17 in patients with AD with 13 years of education (C)

Fig. 2

Heatmap of the baseline correlation between ChP volume (and ChP/eTIV) and CSF biomarkers, neuropsychological tests, and multimodal neuroimaging in the Alzheimer’s continuum. Partial Spearman’s correlation is performed after adjusting for age, sex, years of education, APOE ε4 status, and eTIV. -v, volume of the brain region; -t, cortical thickness of the brain region; -c, corrected cerebral blood flow (cCBF) of the brain region. The number of CSF analyses is 138, the number of regional cCBF is 288, and the number of neuropsychological assessments and brain structural measures is 497. The asterisks represent significant results: *P < 0.05; **P < 0.01; ***P < 0.001

Fig. 3

Simple mediating effects of ChP volume, hippocampal volume, LVV, and ECT alone on the association of CSF hallmarks and neuropsychological tests

Fig. 4

Receiver operating characteristic curves of four neuroimaging measures for discriminating the cerebral Aβ₄₂ change and distinguishing different stages of the AD continuum. (A) The AUC of ChP volume alone is higher than that of LVV (P = 0.038) and ECT (P = 0.008) alone, and that of combined ChP and hippocampal volume is higher than that of hippocampal volume alone (P = 0.033). However, the combination of four neuroimaging measures (P = 0.443) or that of ChP and hippocampal volume (P = 0.735) does not show significant advantage over ChP volume alone; (B) The AUC of ChP volume alone is higher than those of hippocampal volume (P = 0.009) and LVV (P = 0.031) alone. The AUC of combined ChP and hippocampal volumes is higher than that of hippocampal volume alone (P = 0.009) but not higher than that of ChP volume alone (P = 0.654); (C) The AUC of combined ChP and hippocampal volumes is higher than those of hippocampal volume alone (P = 0.008) and ChP volume alone(P < 0.001); (D) The AUC of combined ChP and hippocampal volumes is higher than those of hippocampal volume alone (P = 0.004) and ChP volume alone (P = 0.001)