Dual Model Deep Learning for Alzheimer Prognostication

TL;DR

PROGRESS is a dual-model deep learning framework that predicts Alzheimer's disease progression from a single baseline biomarker, providing calibrated uncertainty and outperforming traditional methods across diverse datasets.

Contribution

It introduces a novel approach that enables accurate, uncertainty-aware prognosis from static baseline data, facilitating early clinical decision-making without extensive longitudinal observations.

Findings

Outperforms Cox, Random Survival Forests, and gradient boosting in survival prediction.

Achieves seven-fold risk stratification differences in patient groups.

Demonstrates robust generalization across diverse clinical centers.

Abstract

Disease modifying therapies for Alzheimer's disease demand precise timing decisions, yet current predictive models require longitudinal observations and provide no uncertainty quantification, rendering them impractical at the critical first visit when treatment decisions must be made. We developed PROGRESS (PRognostic Generalization from REsting Static Signatures), a dual-model deep learning framework that transforms a single baseline cerebrospinal fluid biomarker assessment into actionable prognostic estimates without requiring prior clinical history. The framework addresses two complementary clinical questions: a probabilistic trajectory network predicts individualized cognitive decline with calibrated uncertainty bounds achieving near-nominal coverage, enabling honest prognostic communication; and a deep survival model estimates time to conversion from mild cognitive impairment to dementia. Using data from over 3,000 participants across 43 Alzheimer's Disease Research Centers in the National Alzheimer's Coordinating Center database, PROGRESS substantially outperforms Cox proportional hazards, Random Survival Forests, and gradient boosting methods for survival prediction. Risk stratification identifies patient groups with seven-fold differences in conversion rates, enabling clinically meaningful treatment prioritization. Leave-one-center-out validation demonstrates robust generalizability, with survival discrimination remaining strong across held-out sites despite heterogeneous measurement conditions spanning four decades of assay technologies. By combining superior survival prediction with trustworthy trajectory uncertainty quantification, PROGRESS bridges the gap between biomarker measurement and personalized clinical decision-making.