Open Source Acquisition & Reconstruction Software
Stefan Kroboth1

1Department of Radiology, Medical Physics, Freiburg, Germany

Synopsis

This educational lecture will introduce the audience to open source image acquisition and image reconstruction tools which can be used to set up an open source pipeline for rapid prototyping of MR imaging and reconstruction methods.

Target Audience

Researchers interested in setting up a fully or partially open source prototyping pipeline from image acquisition to image reconstruction.

Outcome/Objectives

The audience will receive an overview over available tools and how they fit into the imaging pipeline. Furthermore the attendees of this course will learn how to implement a sequence in Pulseq.

Purpose

In recent years a large community around open source projects related to magnetic resonance imaging has evolved.The recently launched Open Source Imaging Initiative (OSI2 [1]), which aims at channelizing these efforts, is a good indicator for this trend and showcases the relevance of open source in MRI. Not only do open source projects offer an alternative to (by definition less transparent) closed source industrial development. They also provide flexiblity where restrictions of vendors may prohibit rapid prototyping, or they provide a vendor-agnostic way of solving research problems. All of this simplifies the exchange of data and methods and improves the repeatability of research.

While image reconstruction and post-processing software tools have made their open-source debut many years ago (see [2,3,4,5,6,7,8,9] for a small selection), the interfaces to the MRI machines have long been dominated by the vendor-specific tools.

This often complicates research which needs to interface with the MRI scanner in some way. For instance, testing a novel reconstruction algorithm may require direct access to the data export of the scanner to enable rapid prototyping. This gap has been bridged by tools such as Gadgetron ([10]) and Yarra ([11]).

Another important interface of the user to the scanner is the pulse sequence. Unfortunately, sequences developed for scanners of one vendor can not be run on scanners of other vendors. Sometimes even sharing sequences between different scanners of one vendor is problematic. Open source alternatives such as ODIN ([12]), SequenceTree ([13]), TOPPE ([14]) and Pulseq ([15,16]) provide abstractions over the hardware to mitigate some or all of these problems.

The myriad of available tools allows researchers to build a pipeline to their liking; however, it may be difficult to obtain an overview. This talk aims to provide such an overview over many available and actively developed open source tools and will show where they fit into a fully open source software pipeline.

In this talk we will set up a "virtual open source pipeline" on a "virtual scanner" to enable rapid prototyping. We will start off with a simple gradient echo sequence and together develop it into a spoiled gradient echo sequence. Since the presenter is the maintainer of Pulseq, a large portion of this talk will focus on rapid sequence development with Pulseq.

Conclusion

An established open source prototyping pipeline would allow researchers to quickly develop and test new methods. Ideally, such a pipeline would provide an abstraction from the scanner which would enable an easy transfer of methods to scanners of other vendors. This would substantially decrease the effort necessary for multi-center studies and would improve the reproducability of research.

Acknowledgements

The author thanks Maxim Zaitsev and Kelvin Layton for their help.

References

[1] Felix Arndt et.al., "Open Source Imaging Initiative (OSI²) – Update and Roadmap", ISMRM 2017

[2] X. L. Wu et al., "Impatient MRI: Illinois Massively Parallel Acceleration Toolkit for image reconstruction with enhanced throughput in MRI," 2011 IEEE International Symposium on Biomedical Imaging: From Nano to Macro, Chicago, IL, 2011, pp. 69-72.

[3] M. Freiberger, F. Knoll, K. Bredies, H. Scharfetter and R. Stollberger, "The Agile Library for Biomedical Image Reconstruction Using GPU Acceleration," in Computing in Science & Engineering, vol. 15, no. 1, pp. 34-44, Jan.-Feb. 2013.

[4] Martin Uecker, Frank Ong, Jonathan I Tamir, Dara Bahri, Patrick Virtue, Joseph Y Cheng, Tao Zhang, and Michael Lustig, "Berkeley Advanced Reconstruction Toolbox", ISMRM 2015

[5] Knoll, F.; Schwarzl, A,; Diwoky, C.; Sodickson DK., "gpuNUFFT - An Open-Source GPU Library for 3D Gridding with Direct Matlab Interface", ISMRM 2014.

[6] SPM (Statistical Parameter Mapping): http://www.fil.ion.ucl.ac.uk/spm/

[7] FSL (FMRIB Software Library): https://fsl.fmrib.ox.ac.uk/fsl/fslwiki

[8] AFNI (Analysis of Functional NeuroImages): https://afni.nimh.nih.gov/

[9] FreeSurfer: https://surfer.nmr.mgh.harvard.edu/

[10] Hansen, M. S. and Sørensen, T. S. (2013), Gadgetron: An open source framework for medical image reconstruction. Magn Reson Med, 69: 1768-1776.

[11] Yarra: https://yarra.rocks/doc/

[12] Jochimsen TH, von Mengershausen M. ODIN—Object-oriented Development Interface for NMR. J. Magn. Reson. 2004;170:67–78.

[13] Magland JF, Li C, Langham MC, Wehrli FW. Pulse sequence programming in a dynamic visual environment: SequenceTree. Magn. Reson. Med. 2016;75:257–265.

[14] Nielsen, J. and Noll, D. C. (2018), TOPPE: A framework for rapid prototyping of MR pulse sequences. Magn. Reson. Med., 79: 3128-3134.

[15] Layton, K. J., Kroboth, S. , Jia, F. , Littin, S. , Yu, H. , Leupold, J. , Nielsen, J. , Stöcker, T. and Zaitsev, M. (2017), Pulseq: A rapid and hardware‐independent pulse sequence prototyping framework. Magn. Reson. Med., 77: 1544-1552.

[16] R. Keerthi Sravan, Sneha Potdar, Pavan Poojar, Ashok Kumar Reddy, Stefan Kroboth, Jon-Fredrik Nielsen, Maxim Zaitsev, Ramesh Venkatesan, Sairam Geethanath, "Pulseq-Graphical Programming Interface: Open source visual environment for prototyping pulse sequences and integrated magnetic resonance imaging algorithm development." Magnetic Resonance Imaging (preprint)


Proc. Intl. Soc. Mag. Reson. Med. 26 (2018)