Parcel2Progression: An Anatomy-aware Longitudinal Framework for Alzheimer's Disease Diagnosis

2026-08-09

AI Summary

This work addresses the challenge of simultaneously capturing high-resolution spatial details and efficiently modeling variable-length longitudinal brain scans for early Alzheimer’s disease (AD) prediction. The authors propose an anatomy-aware longitudinal Transformer framework that leverages an atlas-guided regional segmentation encoder to convert 3D MRI into anatomically semantic tokens, further integrating patient age information to enable linear-complexity modeling of arbitrary-length 4D structural MRI sequences. By circumventing the quadratic complexity bottleneck of conventional 4D Vision Transformers, the method achieves significant performance gains on ADNI, AIBL, and MIRIAD datasets, improving balanced accuracy by 7% for MCI-to-AD conversion prediction and by 5% for AD/CN classification. It also uncovers clinically consistent atrophy patterns and demonstrates generalizability to other neurodegenerative conditions such as frontotemporal dementia.

Abstract

Alzheimer's disease (AD) progression is a longitudinal process with subtle pathological cues in the early stages. Yet, computational constraints have limited most neuroimaging models to either compromise spatial information or limit the number of longitudinal scans. We aim to overcome this bottleneck and fully leverage high-resolution, variable-length T1w structural MRI (4D sMRI) scan sequences. We introduce Parcel2Progression (P2P), a Longitudinal Transformer Framework which tackles this challenge using an Atlas-guided Parcel Encoder that tokenizes 3D scans into a set of richer anatomically grounded representations. A Longitudinal Transformer then integrates irregular, arbitrary-length longitudinal visits with patient age. This synergy delivers two key advantages: (1) parcel-specific interpretability, and (2) computational tractability for long-term analysis, which scales linearly with the number of scans compared to a naive quadratic 4D ViT cost. P2P outperforms prior works and baselines in both MCI (Mild Cognitive Impairment) to AD conversion prediction and AD vs. CN (Cognitively Normal) classification tasks across ADNI, AIBL, and MIRIAD datasets. Leveraging longitudinal scans boosts performance over single-scan baselines by up to 5% and 7% in balanced accuracy for AD classification and MCI conversion prediction tasks, respectively. Interpretability analysis using parcel saliencies and attention rollouts reveals clinically consistent atrophy patterns in AD and MCI subjects. We also demonstrate the frameworks' reliability in anomaly detection using a synthetic dataset, and test the model's generalizability for other neurodegenerative diseases like Frontotemporal Dementia.

Problem

• Alzheimer's disease
• longitudinal analysis
• structural MRI
• computational constraints
• early diagnosis

Innovation

• Longitudinal Transformer
• Atlas-guided Parcel Encoder
• 4D sMRI
• Computational Tractability
• Anatomy-aware Representation