OpenMRF: A Modular, Vendor-Neutral Open-Source Framework for Reproducible Magnetic Resonance Fingerprinting using...

Area of Science:

• Quantitative MRI
• Biomedical Engineering
• Medical Imaging Physics

Background:

• Magnetic Resonance Fingerprinting (MRF) adoption is hindered by fragmented tools and lack of unified frameworks.
• Reproducibility and consistency challenges limit widespread MRF implementation and advancement.

Purpose of the Study:

• Introduce OpenMRF, a comprehensive Pulseq-based solution to overcome MRF adoption barriers.
• Enable consistent, reproducible, and transferable MRF research across vendors, sites, and field strengths.

Main Methods:

• OpenMRF integrates modular Pulseq sequence design, Bloch-simulation dictionary creation, and iterative low-rank subspace reconstruction.
• Evaluated using digital phantoms, multi-site phantom studies (0.55 T, 1.5 T, 3 T), and in vivo scans (liver, myocardium, brain).

Main Results:

• Simulations showed high mapping accuracy (T1: 0.03+/-0.32%, T2: 0.12+/-1.94%).
• Multi-site phantom study yielded relaxation times consistent with reference values across field strengths.
• In vivo acquisitions demonstrated high-quality parameter maps on various platforms.

Conclusions:

• OpenMRF offers a robust, open-source, end-to-end Pulseq solution for MRF.
• Facilitates reproducible sequence implementation, accurate dictionary simulation, and advanced reconstruction.
• Aims to accelerate reproducible and harmonized quantitative MRI research through a unified platform.