Machine learning prediction of tau‐PET in Alzheimer's disease using plasma, MRI, and clinical data

Key Points

• This research aims to determine how accessible variables can estimate tau tangle load in Alzheimer's disease.
• Utilized plasma phosphorylated tau-217 and MRI data as predictors for tau-PET.
• Developed machine learning models to analyze data from various Alzheimer’s disease cohorts.
• Assessed model generalizability across different cohorts.
• High predictive accuracy for tau-PET observed using plasma p-tau217 and MRI data.
• Machine learning models demonstrated strong generalizability across various Alzheimer's disease cohorts.