![[2602.20752] OrthoDiffusion: A Generalizable Multi-Task Diffusion Foundation Model for Musculoskeletal MRI Interpretation](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.20752] OrthoDiffusion: A Generalizable Multi-Task Diffusion Foundation Model for Musculoskeletal MRI Interpretation
Summary
OrthoDiffusion is a novel diffusion-based model designed for multi-task interpretation of musculoskeletal MRI scans, improving diagnostic accuracy and efficiency.
Why It Matters
This research addresses the challenges in interpreting complex musculoskeletal MRI scans, which are critical for accurate diagnosis. By leveraging a unified diffusion model, it enhances the ability to identify multiple abnormalities, potentially transforming clinical workflows and patient outcomes.
Key Takeaways
- OrthoDiffusion utilizes three orientation-specific 3D diffusion models for enhanced MRI interpretation.
- The model demonstrates high diagnostic precision even with limited labeled data.
- It achieves robust performance across different clinical settings and MRI field strengths.
- Anatomical representations from knee imaging are transferable to other joints, aiding in multi-disease diagnosis.
- The framework could significantly improve the efficiency of musculoskeletal MRI interpretation in clinical practice.
Computer Science > Computer Vision and Pattern Recognition arXiv:2602.20752 (cs) [Submitted on 24 Feb 2026] Title:OrthoDiffusion: A Generalizable Multi-Task Diffusion Foundation Model for Musculoskeletal MRI Interpretation Authors:Tian Lan, Lei Xu, Zimu Yuan, Shanggui Liu, Jiajun Liu, Jiaxin Liu, Weilai Xiang, Hongyu Yang, Dong Jiang, Jianxin Yin, Dingyu Wang View a PDF of the paper titled OrthoDiffusion: A Generalizable Multi-Task Diffusion Foundation Model for Musculoskeletal MRI Interpretation, by Tian Lan and Lei Xu and Zimu Yuan and Shanggui Liu and Jiajun Liu and Jiaxin Liu and Weilai Xiang and Hongyu Yang and Dong Jiang and Jianxin Yin and Dingyu Wang View PDF HTML (experimental) Abstract:Musculoskeletal disorders represent a significant global health burden and are a leading cause of disability worldwide. While MRI is essential for accurate diagnosis, its interpretation remains exceptionally challenging. Radiologists must identify multiple potential abnormalities within complex anatomical structures across different imaging planes, a process that requires significant expertise and is prone to variability. We developed OrthoDiffusion, a unified diffusion-based foundation model designed for multi-task musculoskeletal MRI interpretation. The framework utilizes three orientation-specific 3D diffusion models, pre-trained in a self-supervised manner on 15,948 unlabeled knee MRI scans, to learn robust anatomical features from sagittal, coronal, and axial views. These view...
