SDUM: A Scalable Deep Unrolled Model for Universal MRI Reconstruction

2025-12-18

AI Summary

Clinical MRI protocols exhibit high diversity across anatomical regions, contrast types, sampling patterns, and acceleration factors; however, existing deep learning–based reconstruction methods suffer from poor generalizability and require protocol-specific customization. To address this, we propose the first universal MRI reconstruction framework applicable across all clinical scenarios. Our method introduces a scalable deep-unfolding architecture, revealing a strong logarithmic scaling law between PSNR and parameter count (r = 0.986); novel sampling-aware weighted data consistency (SWDC); stage-wise learned coil sensitivity map estimation (CSME); and universal conditional encoding (UC). Built upon the Restormer backbone with progressive cascaded training, the framework achieves zero-shot cross-protocol generalization without fine-tuning. It achieves state-of-the-art performance on all four tracks of CMRxRecon2025, outperforming PromptMR+ and PC-RNN by 0.55 dB and 1.80 dB, respectively. Ablation studies confirm the significant contribution of each component.

Technology Category

• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Consent frameworks and practices on the web
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Clinical MRI encompasses diverse imaging protocols--spanning anatomical targets (cardiac, brain, knee), contrasts (T1, T2, mapping), sampling patterns (Cartesian, radial, spiral, kt-space), and acceleration factors--yet current deep learning reconstructions are typically protocol-specific, hindering generalization and deployment. We introduce Scalable Deep Unrolled Model (SDUM), a universal framework combining a Restormer-based reconstructor, a learned coil sensitivity map estimator (CSME), sampling-aware weighted data consistency (SWDC), universal conditioning (UC) on cascade index and protocol metadata, and progressive cascade expansion training. SDUM exhibits foundation-model-like scaling behavior: reconstruction quality follows PSNR ${sim}$ log(parameters) with correlation $r{=}0.986$ ($R^2{=}0.973$) up to 18 cascades, demonstrating predictable performance gains with model depth. A single SDUM trained on heterogeneous data achieves state-of-the-art results across all four CMRxRecon2025 challenge tracks--multi-center, multi-disease, 5T, and pediatric--without task-specific fine-tuning, surpassing specialized baselines by up to ${+}1.0$~dB. On CMRxRecon2024, SDUM outperforms the winning method PromptMR+ by ${+}0.55$~dB; on fastMRI brain, it exceeds PC-RNN by ${+}1.8$~dB. Ablations validate each component: SWDC ${+}0.43$~dB over standard DC, per-cascade CSME ${+}0.51$~dB, UC ${+}0.38$~dB. These results establish SDUM as a practical path toward universal, scalable MRI reconstruction.

Problem

• Develops a universal MRI reconstruction model for diverse protocols
• Achieves scalable performance gains with increasing model depth
• Outperforms specialized methods across multiple datasets without fine-tuning

Innovation

• Combines Restormer reconstructor with learned coil sensitivity estimation
• Uses sampling-aware weighted data consistency and universal conditioning
• Trains via progressive cascade expansion for scalable performance