Spatial amyloid-informed multimodal brain age as an early marker of Alzheimer's-related vulnerability and risk stratification.

Higher BAG was associated with greater odds of cognitive impairment across the AD continuum, with stronger effects in amyloid-positive individuals.

• The cohort included 990 community-dwelling adults spanning normal cognition, subjective cognitive decline (SCD), mild cognitive impairment (MCI), and dementia.
• Participants were recruited from the Chinese Preclinical Alzheimer's Disease Study (CPAS) from community settings and memory clinics.
• Effects of BAG on cognitive status classification were stronger in Aβ-positive individuals compared to the overall cohort.
• Cross-sectional analysis using integrated machine-learning models was used to derive BAG estimates.

BAG explained more cognitive variance than global amyloid-β burden and was linked to multidomain cognitive deficits.

• Regional Aβ-PET and structural MRI were both used to inform BAG estimation in the multimodal model.
• The multimodal BAG model outperformed global Aβ burden alone in explaining variance in cognitive test performance.
• Cognitive associations spanned multiple domains, indicating multidomain cognitive deficits associated with elevated BAG.

Elevated BAG corresponded to adverse plasma biomarker profiles indicative of early Alzheimer's-related pathology.

• Higher BAG was associated with higher p-tau217, p-tau181, neurofilament light (NfL), and glial fibrillary acidic protein (GFAP) levels.
• Higher BAG was also associated with lower Aβ42/40 ratio in plasma.
• These associations indicate that elevated BAG corresponds to early biomarker alterations across the AD continuum.
• Plasma biomarkers assessed included p-tau217, p-tau181, NfL, GFAP, and Aβ42/40.

Elevated BAG was associated with reduced hippocampus-default mode network (DMN) functional connectivity.

• Resting-state fMRI was used to assess hippocampus-DMN connectivity.
• Higher BAG corresponded to reduced connectivity between the hippocampus and the default mode network.
• This finding links accelerated brain aging as captured by the multimodal BAG to functional network disruption relevant to AD.

Incorporating regional Aβ-PET data alongside structural MRI improved the sensitivity of the BAG model to early AD processes compared to MRI-only models.

• MRI-only brain age models were noted to insufficiently reflect Alzheimer's disease pathology.
• The multimodal model integrated regional Aβ-PET with structural MRI for BAG estimation.
• The amyloid-informed BAG captured convergent AD-related pathology beyond what amyloid burden alone provided.
• The model was developed and tested in a sample of 990 community-dwelling adults from the CPAS cohort.