SDUM: A Scalable Deep Unrolled Model for Universal Cardiac MRI Reconstruction

Clinical cardiac MRI spans diverse contrasts, sampling trajectories, accelerations, scanners, and patient populations, yet many deterministic deep-learning reconstruction methods remain protocol specific. We present the Scalable Deep Unrolled Model (SDUM), which integrates a Restormer-based unrolled reconstructor, per-cascade coil sensitivity estimation, sampling aware weighted data consistency, and universal conditioning on cascade index and acquisition metadata. A single SDUM model achieves state-of-the-art performance across all CMRxRecon2025 tracks without task-specific fine-tuning and outperforms PromptMR+ on CMRxRecon2024 by +0.55 dB. Scaling experiments show near-logarithmic gains with depth up to 18 cascades $\left(r=0.986,R^2=0.973\right)$ and continued but diminishing gains from data scaling (32.72 dB at 40% to 33.18 dB at 100%). SDUM also generalizes in a zero-shot setting to unseen in-house chemical exchange saturation transfer (CEST) MRI (43.57 dB PSNR, 0.9769 SSIM). When trained separately on fastMRI brain, SDUM surpasses PC-RNN by +1.8 dB. These results support SDUM as a scalable framework for robust MRI reconstruction across heterogeneous acquisition settings beyond cardiac MRI.

https://github.com/NVIDIA-Medtech/NV-Raw2insights-MRI.

https://huggingface.co/nvidia/NV-Raw2insights-MRI.