Potential and challenges of generative adversarial networks for super-resolution in 4D Flow MRI

2025-08-20

AI Summary

Low spatial resolution and noise in 4D Flow MRI impede accurate reconstruction of near-wall blood velocity, limiting clinical utility. To address this, we introduce generative adversarial networks (GANs) for the first time to super-resolution reconstruction of 4D Flow MRI data. We systematically evaluate vanilla, relativistic, and Wasserstein GAN architectures and propose a medical flow-field–specific adversarial learning strategy. Training and validation leverage synthetic data generated from patient-specific, in-silico cerebral vascular models and realistic MR reconstruction pipelines. Experimental results demonstrate that the Wasserstein GAN significantly improves training stability and near-wall velocity recovery accuracy—reducing vector-normalized RMSE to 6.9% (a 0.3-percentage-point improvement over non-adversarial baselines). It further exhibits superior robustness under low signal-to-noise ratio conditions and in anatomically complex vascular regions. This work establishes a new paradigm for high-fidelity hemodynamic analysis from clinical 4D Flow MRI.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Machine Learning: Adversarial Learning & Robustness
• Multiagent Systems: Adversarial Agents

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Web data generation and simulation
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

4D Flow Magnetic Resonance Imaging (4D Flow MRI) enables non-invasive quantification of blood flow and hemodynamic parameters. However, its clinical application is limited by low spatial resolution and noise, particularly affecting near-wall velocity measurements. Machine learning-based super-resolution has shown promise in addressing these limitations, but challenges remain, not least in recovering near-wall velocities. Generative adversarial networks (GANs) offer a compelling solution, having demonstrated strong capabilities in restoring sharp boundaries in non-medical super-resolution tasks. Yet, their application in 4D Flow MRI remains unexplored, with implementation challenged by known issues such as training instability and non-convergence. In this study, we investigate GAN-based super-resolution in 4D Flow MRI. Training and validation were conducted using patient-specific cerebrovascular in-silico models, converted into synthetic images via an MR-true reconstruction pipeline. A dedicated GAN architecture was implemented and evaluated across three adversarial loss functions: Vanilla, Relativistic, and Wasserstein. Our results demonstrate that the proposed GAN improved near-wall velocity recovery compared to a non-adversarial reference (vNRMSE: 6.9% vs. 9.6%); however, that implementation specifics are critical for stable network training. While Vanilla and Relativistic GANs proved unstable compared to generator-only training (vNRMSE: 8.1% and 7.8% vs. 7.2%), a Wasserstein GAN demonstrated optimal stability and incremental improvement (vNRMSE: 6.9% vs. 7.2%). The Wasserstein GAN further outperformed the generator-only baseline at low SNR (vNRMSE: 8.7% vs. 10.7%). These findings highlight the potential of GAN-based super-resolution in enhancing 4D Flow MRI, particularly in challenging cerebrovascular regions, while emphasizing the need for careful selection of adversarial strategies.

Problem

• Improving low spatial resolution in 4D Flow MRI
• Recovering accurate near-wall blood velocity measurements
• Addressing GAN training instability for medical super-resolution

Innovation

• GAN-based super-resolution for 4D Flow MRI
• Wasserstein GAN for stable training
• Improved near-wall velocity recovery