FLAME Scorecard: Alzheimer's Disease Risk Prediction Model

A machine learning model for predicting Alzheimer's disease risk developed using data from 666 participants from the Alzheimer's Disease Neuroimaging Initiative (ADNI). The FLAME scorecard was externally validated on 4,876 participants from the National Alzheimer's Coordinating Center (NACC) and integrates cognitive measures, daily functioning, and demographics. The research was authored by Yumiko Wiranto and published on figshare in April 2026.

Use Cases

• Predict Alzheimer's disease development risk based on cognitive test scores and demographic data.
• Validate clinical risk models using external datasets like the National Alzheimer's Coordinating Center.
• Implement interpretable scoring systems into primary care workflows for early dementia screening.
• Analyze the predictive power of specific cognitive measures like Category Animal, Trail Making Test B, and Logical Memory Delayed.
• Study the relationship between daily functioning scores (FAQ) and Alzheimer's disease progression.

Strengths

• Model trained on 666 participants from the Alzheimer's Disease Neuroimaging Initiative.
• External validation performed on a large cohort of 4,876 participants from the National Alzheimer's Coordinating Center.
• Strong predictive performance reported with AUC values between 0.868 and 0.892 in the development set.
• Scorecard uses a limited set of 4 interpretable measures, making it suitable for clinical settings.

Limitations

• Row count for the underlying dataset is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data may reflect temporal and source bias inherent to the specific research cohorts used.

Provenance

Source

Alzheimer's Disease Neuroimaging Initiative (ADNI) and National Alzheimer's Coordinating Center (NACC)

Collection Method

Data used to train and validate a machine learning risk score model via the FasterRisk algorithm.

Freshness

Last updated 2026-04-10 06:03:27; freshness should be verified.