0945

A prototype of a fully integrated environment for a collaborative work in MR sequence development for a reproducible research
Saulius Archipovas1, Robin Niklas Wilke1, Simon Konstandin1, Jörn Huber1, Daniel Christopher Hoinkiss1, Cristoffer Cordes1, Nora-Josefin Breutigam1, and Matthias Günther1,2

1MR Physics, Fraunhofer MEVIS, Bremen, Germany, 2MR-Imaging and Spectroscopy, Faculty 01 (Physics/Electrical Engineering), University of Bremen, Bremen, Germany

Synopsis

We present a prototype exemplifying conceptual web-based workflow with services and applications which allow MR sequence developers to collaboratively create and work on MR sequences. By automating various error-prone tasks that usually overwhelm new and existing users and focusing on an interplay of web-services and applications, we introduce a novel developer experience. That might help the MR community to establish a sequence development workflow that can greatly benefit from collaborative efforts and make MR sequence developments more transparent and reproducible.

Introduction

In the scientific community, development of MR sequences is typically carried out using state-of-the-art vendor-specific development frameworks. In this way, many novel MR applications could have been implemented, tested and translated into clinical applications. However, the downside of the tools are i) their individual complexity such as requiring advanced programming skills in high-level languages, ii) their inter-vendor incompatibility making it impractical to transfer one sequence development state to another vendor's MR scanner and iii) providing environments that are non adequate for a modern workflow which takes advantage of web-technology and collaborative efforts on single projects as exemplified by e.g. any collaborative project from github.org. Projects such as pulseq-gpi [1] try mitigate to this problem but do not provide solutions for a collaborative workflow. To solve this issue, we introduce a bundle of web-based collaborative and development services and prototype applications that enable not only novices but also experienced developers to develop and maintain MR sequences and reconstruction. As the state of today, it consists of three integrated services that were identified as crucial for sequence development in MR research [2].

Methods

To gather, evaluate and implement functional and non-functional requirements, we had several workshops that were based partly on Design Thinking [3] and its methods, that were applied to suit the complex environment and MR sequence development topic. We also implemented different software prototypes to test the requirements for a streamlined and vendor-agnostic MRI sequence development [4-7].

Results

The working environment consists of three core web-based services: "pm", "sequence" and "recon".

pm

"pm" is a project management service that allows users to create, work and share "sequence" and "recon" projects. By reusing the gitlab api [8] and our backend services, we streamline a convenient workflow for managing changes in MR sequence development. We automate mostly tedious tasks for sequence developers like starting backend services that are needed for running the "sequence" application.

sequence

"sequence" enables users to work on specific MR sequences that are part of a project that the user works on. The development environment holds different plugins [7] that allow users to have different but also inter-connected views and editors of the sequence modules and their properties. This enables users to work more efficiently. At the state of today, we allow users to work on sequences (e.g. extending an EPI-Readout) without breaking modules that other sequence developers are working on. "sequence" itself allows users to re-use different sequences (e.g. EPI, pCASL, Turbo Spin Echo), helpful extensions or reusable modules (e.g. EPI-Readout, Slice Selection). Through the use of the project manager, the service allows for the inclusion of all sequence related developments from other team members.

recon

Along the line of novel MR-image reconstruction techniques that may not be available at all MR scanners, we are confident that providing the option to easily use advanced, novel MR reconstruction methods is desired as well. For this reason, we work on the additional web-based service "recon" allowing the usage of advanced image reconstruction APIs and frameworks such as Gadgetron [9] and BART [10]. To integrate the development of reconstruction in the web-based workflow, we investigate the usage of eclipse che [11] for the reconstruction workstation, see Figure 3.

Discussion

Using the presented services and applications users can create and work on MR sequences. By automating various error-prone tasks that usually overwhelm new and existing users, we introduce a novel developer experience compared to existing frameworks, e.g. Pulseq [1]. By focusing on an efficient interplay of the services, we enable users to create and work on sequences that might help both MR group and on a larger scale, the MR community to establish an efficient sequence development workflow that can greatly benefit of collaborative efforts. In this way, the developments will be much more transparent and reproducible. Decreasing the complexity of the general task of sequence development will also motivate new user groups to this exciting field.

Conclusion

By minimizing the complexity of various MR sequence development tasks, a barrier to enter the development of MR sequences is reduced. Using MR community and our applications and services, the introduced prototypical web-based workflow for developing vendor-agnostic MR sequences enables a user-friendly collaborative and reproducible research.

Acknowledgements

All funding for this study was provided by the internal Attract (600172) funding program of the German Fraunhofer-Gesellschaft.

References

[1.] Ravi KS, Potdar S, Poojar P., Reddy AK, Kroboth S, Nielsen J-F, Venkateson R, Geethanath. Pulseq-Graphical Programming Interface: Open source visual environment for prototyping pulse sequences and integrated magnetic resonance imaging algorithm development. S. Mag. Reson. Imag. 52 (2018): 9-15

[2.] Archipovas S. Nutzerorientierte Gestaltung einer Entwicklungsumgebung von MR-Sequenzen auf Magnetresonanztomographen. In: Endmann A, Fischer H, Krökel M (eds) Mensch und Computer 2015 – Usability Professionals. De Gruyter Oldenbourg, Berlin, pp 311–315 (2015)

[3.] Curedale, Robert. Design Thinking: Process and Methods Manual. Topanga, CA: Design Community College, (2013)

[4.] Honroth T, Cordes C, Archipovas S, Hoinkiss D, Günther M, Porter DA. Platform-independent, rapid prototyping of MR sequences without code compilation. In Proc. Intl. Soc. Mag. Reson. Med. 24:3207 (2016)

[5.] Honroth T, Cordes C, Archipovas S, Günther M, Porter DA. Advanced platform-independent MR prototyping: from EPI to arterial spin labeling without code compilation. In Proc. Intl. Soc. Mag. Reson. Med. 25:3824 (2017)

[6.] Cordes C, Honroth T, Hoinkiss D, Archipovas S, Porter DA, Günther M. Parameter dependency in modular MR sequences using directed graphs. In Proc. Intl. Soc. Mag. Reson. Med. 24:3198 (2016)

[7.] Archipovas S, Honroth T, Cordes C, Günther M, Porter DA. MRI Pulse Sequence Development Using Graphical User Interface Modules. In Proc. Intl. Soc. Mag. Reson. Med. 25:1510 (2017)

[8.] Gitlab API | Gitlab. Retrieved November 6, 2018, from https://docs.gitlab.com/ee/api/

[9.] Hansen MS, Sørensen TS. Gadgetron: An Open Source Framework for Medical Image Reconstruction. Magn Reson Med. 2013 Jun;69(6):1768-76.

[10.] Martin Uecker, Frank Ong, Jonathan I Tamir, Dara Bahri, Patrick Virtue, Joseph Y Cheng, Tao Zhang, and Michael Lustig. Berkeley Advanced Reconstruction Toolbox, Annual Meeting ISMRM, Toronto 2015, In Proc. Intl. Soc. Mag. Reson. Med 23; 2486 (2015)

[11.] Eclipse Che | Eclipse Next-Generation IDE. Retrieved November 6, 2018, from https://www.eclipse.org/che/

Figures

Fig.1: Prototype of the project management tool "pm", that focus on specific tasks needed to start working on MR sequences in "sequence" or "recon" applications

Fig.2: "sequence" is an MRI sequence development tool [7]. The user can work on a particular aspect of an MR sequence by creating own or reusing existing sequence modules built by our team

Fig.3: Gadgetron gadget development workspace in a web-browser. The user can write gadgetron gadgets or directly post-process raw data that was acquired using "sequence" and "driver" [5] application

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)
0945