GENRE-CMR: Generalizable Deep Learning for Diverse Multi-Domain Cardiac MRI Reconstruction

2025-08-28

AI Summary

To address the weak generalization across diverse acquisition settings and the trade-off between image quality and scanning efficiency in accelerated cardiovascular magnetic resonance (CMR) reconstruction, this paper proposes a residual deep-unfolding network. The method innovatively incorporates an Edge-Aware Region (EAR) loss and a Statistical Distribution Alignment (SDA) loss within a generative adversarial framework to jointly optimize structural fidelity and distribution consistency. Leveraging iterative unfolding, cascaded convolutional subnetworks, and residual connections, the architecture enhances representational capacity and robustness. Evaluated on unseen acquisition scenarios, the model achieves 0.9552 SSIM and 38.90 dB PSNR—outperforming state-of-the-art methods—and demonstrates superior generalization capability and high-fidelity reconstruction accuracy.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Multi-modal Vision
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

Accelerated Cardiovascular Magnetic Resonance (CMR) image reconstruction remains a critical challenge due to the trade-off between scan time and image quality, particularly when generalizing across diverse acquisition settings. We propose GENRE-CMR, a generative adversarial network (GAN)-based architecture employing a residual deep unrolled reconstruction framework to enhance reconstruction fidelity and generalization. The architecture unrolls iterative optimization into a cascade of convolutional subnetworks, enriched with residual connections to enable progressive feature propagation from shallow to deeper stages. To further improve performance, we integrate two loss functions: (1) an Edge-Aware Region (EAR) loss, which guides the network to focus on structurally informative regions and helps prevent common reconstruction blurriness; and (2) a Statistical Distribution Alignment (SDA) loss, which regularizes the feature space across diverse data distributions via a symmetric KL divergence formulation. Extensive experiments confirm that GENRE-CMR surpasses state-of-the-art methods on training and unseen data, achieving 0.9552 SSIM and 38.90 dB PSNR on unseen distributions across various acceleration factors and sampling trajectories. Ablation studies confirm the contribution of each proposed component to reconstruction quality and generalization. Our framework presents a unified and robust solution for high-quality CMR reconstruction, paving the way for clinically adaptable deployment across heterogeneous acquisition protocols.

Problem

• Accelerated cardiac MRI reconstruction faces scan time versus quality trade-off
• Generalizing across diverse acquisition settings remains a critical challenge
• Existing methods struggle with reconstruction blurriness across different protocols

Innovation

• GAN-based unrolled residual network for MRI reconstruction
• Edge-aware and statistical loss functions enhance quality
• Improves generalization across diverse acquisition protocols