Ultra-Fast Cardiovascular Imaging with AI
Published:Dec 25, 2025 12:47
•1 min read
•ArXiv
Analysis
This paper addresses the limitations of current cardiovascular magnetic resonance (CMR) imaging, specifically long scan times and heterogeneity across clinical environments. It introduces a generalist reconstruction foundation model (CardioMM) trained on a large, multimodal CMR k-space database (MMCMR-427K). The significance lies in its potential to accelerate CMR imaging, improve image quality, and broaden its clinical accessibility, ultimately leading to faster diagnosis and treatment of cardiovascular diseases.
Key Takeaways
- •Introduces CardioMM, a generalist reconstruction foundation model for ultra-fast CMR imaging.
- •Presents MMCMR-427K, the largest multimodal CMR k-space database.
- •Demonstrates state-of-the-art performance and strong generalization capabilities.
- •Enables substantial increase in CMR examination throughput without compromising clinical integrity.
Reference
“CardioMM achieves state-of-the-art performance and exhibits strong zero-shot generalization, even at 24x acceleration, preserving key cardiac phenotypes and diagnostic image quality.”