Predicting Amyloid-β Positivity In Alzheimer's Disease Neuroimaging Initiative (ADNI) Cohort: Implications For Clinical Pre-screening

Anti-amyloid therapies have increased the need for efficient identification of amyloid-β-positive (Aβ+) individuals.

Amyloid PET and cerebrospinal fluid biomarkers are costly and not universally accessible.

Clinical Amyloid Positivity Score (CAPS) predicts Aβ positivity using clinical and imaging variables.

This study evaluates the applicability of CAPS within the larger and more demographically diverse Alzheimer’s Disease Neuroimaging Initiative (ADNI) cohort.

Objective: To evaluate clinical, cognitive, neuropsychiatric, and imaging predictors of amyloid positivity and assess the performance of the CAPS in a large independent cohort.

Univariate Analysis

Compared with the Aβ− group, participants in the Aβ+ group had:
Lower baseline MMSE scores
Greater annual MMSE decline
Higher NPI-Q scores
Reduced hippocampal & entorhinal volumes
Gender and education did not differ significantly between groups.

Multivariate Regression Analysis
NPI-Q, cognitive decline (MMSE), & hippocampal atrophy independently predicted amyloid positivity.

After multivariate adjustments, hippocampal volume remained significant whereas entorhinal volume did not. Hippocampal atrophy may be a stronger imaging marker of amyloid positivity.

The modified CAPS demonstrated higher sensitivity (76.9% vs. 74.0%), slightly lower specificity (50.0% vs. 51.5%), and a greater AUC (0.682 vs 0.653).

In ADNI, amyloid positivity is most strongly associated with neuropsychiatric symptoms, cognitive decline, and hippocampal atrophy. Cohort-specific adaptations improved predictive accuracy of moderate CAPS performance in ADNI. Accessible clinical pre-screening tools for amyloid positivity should continue to be developed and validated.