BART Online Open-Source Sequence Toolbox for Computational MRI

Area of Science:

• Medical Imaging
• Computational MRI
• Open-Source Software

Background:

• Advanced computational MRI techniques require joint design of acquisition and reconstruction for reproducibility.
• Clinical integration of these methods faces challenges in long-term reproducibility and maintenance.

Purpose of the Study:

• To provide a fully integrated open-source framework for advanced computational MRI.
• To ensure reproducibility and maintainability of MRI acquisition and reconstruction techniques.
• To enable seamless integration with clinical MRI scanners.

Main Methods:

• Developed a software framework for pulse sequence development within the BART toolbox.
• Created a vendor-specific driver sequence for clinical MRI scanner integration with online parameter adjustment.
• Utilized the Pulseq format for offline sequence reproducibility.
• Implemented quantitative MRI methods (T1, R2*, B0 mapping) with radial FLASH and model-based reconstruction.

Main Results:

• Successfully implemented quantitative MRI methods within the BART framework.
• Enabled online adaptation of acquisition parameters and Field of View (FOV) on a clinical MRI system.
• Demonstrated agreement in quantitative parameter maps between online and offline Pulseq acquisitions.

Conclusions:

• The developed framework enables end-to-end reproducibility of advanced computational MRI methods.
• This work facilitates the use and maintenance of complex MRI techniques in clinical settings.
• The open-source approach promotes wider adoption and further development of computational MRI.