# Real-World Multimodal Machine Learning for Risk Enrichment Across the Alzheimer’s Disease Spectrum

## TL;DR

This study uses machine learning with clinical and brain data to better identify Alzheimer's disease risk in patients with mild cognitive impairment.

## Contribution

The novel use of multimodal machine learning with real-world clinical data and neuroimaging for biologically informed risk enrichment in Alzheimer's disease.

## Key findings

- Adding more data types improved model performance, with FDG-PET showing the biggest impact.
- Low metabolism in posterior default mode network regions was the strongest predictor of Alzheimer's risk.
- AD-like scores in MCI patients showed a continuous pattern, supporting biological enrichment.

## Abstract

Background and Objectives: Mild cognitive impairment (MCI) is heterogeneous within the Alzheimer’s disease (AD) continuum, and categorical labels may not reflect biological variability. We evaluated whether multimodal machine learning using routine clinical data and neuroimaging could support biologically informed enrichment across MCI and AD in a real-world memory clinic cohort. Methods: We analyzed 474 patients (1547 visits) with clinical and cognitive measures, laboratory parameters, MRI regional volumes, and FDG-PET regional uptake. Elastic Net and gradient boosting models were trained using nested cross-validation with strict patient-level separation. Results: Model discrimination improved as additional data modalities were added, and FDG-PET contributed the largest performance improvement. Hypometabolism in posterior default mode network regions consistently emerged as the most influential predictor. In the MCI subgroup, AD-like scores showed a continuous distribution consistent with biological enrichment. Conclusions: Multimodal models may provide an interpretable enrichment framework in heterogeneous memory clinic populations.