Integrated algorithm combining plasma biomarkers and cognitive assessments accurately predicts brain β-amyloid...

Area of Science:

• Biomarkers and Computational Modeling in Neuroscience
• Alzheimer's Disease Research
• Neurodegenerative Disease Diagnostics

Background:

• Accurate prediction of cerebral amyloidosis is crucial for Alzheimer's disease (AD) diagnosis and treatment.
• Easily available indicators are needed for non-invasive assessment of amyloid pathology.

Purpose of the Study:

• To develop and validate data-driven computational models for predicting brain β-amyloid (Aβ) pathology.
• To assess the efficacy of plasma biomarkers, APOE genotypes, and demographics in predicting Aβ positivity.

Main Methods:

• Examined plasma biomarkers (Aβ42, Aβ40, T-tau, P-tau181, NfL), APOE genotypes, cognitive scores, and demographics in a Chinese cohort (N=609).
• Developed integrated computational models to predict brain Aβ pathology.
• Validated models using the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort.

Main Results:

• Computational models accurately predicted brain Aβ positivity (AUC=0.94), validated in the ADNI cohort.
• Models showed highest prediction power in mild cognitive impairment (MCI) participants (AUC=0.97).
• A model using only plasma biomarkers predicted Aβ positivity in amnestic MCI (aMCI) patients with AUC=0.89.

Conclusions:

• Innovative integrated models offer a non-invasive and cost-effective method for assessing Aβ pathology.
• These models can facilitate AD drug development, early screening, clinical diagnosis, and prognosis evaluation.