Cross-Modal Knowledge Distillation for PET-Free Amyloid-Beta Detection from MRI

2026-04-14

AI Summary

This study addresses the limitations of costly and invasive amyloid-β (Aβ) positron emission tomography (PET) imaging in early Alzheimer’s disease screening by proposing a purely MRI-based prediction method that requires neither clinical covariates nor PET images during inference. The approach employs a PET-guided knowledge distillation framework, integrating a BiomedCLIP teacher model, cross-modal attention mechanisms, and Centiloid-aware online negative sampling within a triplet contrastive learning scheme to jointly train the MRI student model at both feature and logit levels. Evaluated on the OASIS-3 and ADNI datasets, the method achieves AUCs of 0.74 and 0.68, respectively, and saliency visualizations confirm its focus on Aβ-associated cortical regions, offering a clinically viable pathway for Aβ screening in populations without access to PET imaging.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Computer Vision: Medical and Biological Imaging

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: User modeling for targeted and personalized online advertising

Abstract

Detecting amyloid-$β$ (A$β$) positivity is crucial for early diagnosis of Alzheimer's disease but typically requires PET imaging, which is costly, invasive, and not widely accessible, limiting its use for population-level screening. We address this gap by proposing a PET-guided knowledge distillation framework that enables A$β$ prediction from MRI alone, without requiring non-imaging clinical covariates or PET at inference. Our approach employs a BiomedCLIP-based teacher model that learns PET-MRI alignment via cross-modal attention and triplet contrastive learning with PET-informed (Centiloid-aware) online negative sampling. An MRI-only student then mimics the teacher via feature-level and logit-level distillation. Evaluated across four MRI contrasts (T1w, T2w, FLAIR, T2*) and two independent datasets, our approach demonstrates effective knowledge transfer (best AUC: 0.74 on OASIS-3, 0.68 on ADNI) while maintaining interpretability and eliminating the need for clinical variables. Saliency analysis confirms that predictions focus on anatomically relevant cortical regions, supporting the clinical viability of PET-free A$β$ screening. Code is available at https://github.com/FrancescoChiumento/pet-guided-mri-amyloid-detection.

Problem

• Amyloid-Beta
• Alzheimer's disease
• PET-free detection
• MRI
• Early diagnosis

Innovation

• cross-modal knowledge distillation
• PET-free amyloid-beta detection
• BiomedCLIP
• triplet contrastive learning
• MRI-based Alzheimer's screening