MRN: Harnessing 2D Vision Foundation Models for Diagnosing Parkinson's Disease with Limited 3D MR Data

2025-09-22

AI Summary

Sparse 3D magnetic resonance quantitative susceptibility mapping (QSM) and neuromelanin-sensitive MRI (NM-MRI) data, coupled with inconsistent voxel spacing and modality mismatches, hinder robust model training and cross-modal transfer for Parkinson’s disease (PD) diagnosis. Method: We propose a lightweight, 2D vision foundation model–based framework for automatic PD diagnosis. First, multi-region-of-interest (ROI) axial 2D slices are extracted and encoded—bypassing the modality adaptation bottleneck of 3D pretraining. Second, a multi-branch network fuses ROI features, augmented by an auxiliary segmentation head to enforce anatomical consistency. Third, a multi-ROI supervised contrastive learning strategy is introduced to enhance discriminative capability under extreme data scarcity. Contribution/Results: Evaluated on the MICCAI 2025 PDCADxFoundation Challenge, our method achieves 86.0% accuracy—ranking first and outperforming the runner-up by 5.5%. This demonstrates the efficacy and generalizability of adapting 2D foundation models to sparse, heterogeneous 3D neuroimaging diagnostics.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

The automatic diagnosis of Parkinson's disease is in high clinical demand due to its prevalence and the importance of targeted treatment. Current clinical practice often relies on diagnostic biomarkers in QSM and NM-MRI images. However, the lack of large, high-quality datasets makes training diagnostic models from scratch prone to overfitting. Adapting pre-trained 3D medical models is also challenging, as the diversity of medical imaging leads to mismatches in voxel spacing and modality between pre-training and fine-tuning data. In this paper, we address these challenges by leveraging 2D vision foundation models (VFMs). Specifically, we crop multiple key ROIs from NM and QSM images, process each ROI through separate branches to compress the ROI into a token, and then combine these tokens into a unified patient representation for classification. Within each branch, we use 2D VFMs to encode axial slices of the 3D ROI volume and fuse them into the ROI token, guided by an auxiliary segmentation head that steers the feature extraction toward specific brain nuclei. Additionally, we introduce multi-ROI supervised contrastive learning, which improves diagnostic performance by pulling together representations of patients from the same class while pushing away those from different classes. Our approach achieved first place in the MICCAI 2025 PDCADxFoundation challenge, with an accuracy of 86.0% trained on a dataset of only 300 labeled QSM and NM-MRI scans, outperforming the second-place method by 5.5%.These results highlight the potential of 2D VFMs for clinical analysis of 3D MR images.

Problem

• Diagnosing Parkinson's disease using limited 3D MR data
• Overcoming data scarcity and modality mismatches in medical imaging
• Leveraging 2D vision models for 3D medical image analysis

Innovation

• Leveraging 2D vision foundation models for 3D MR image analysis
• Using auxiliary segmentation to guide feature extraction
• Applying multi-ROI supervised contrastive learning for classification