Enabling Neurodesk Workflows on the Scanner Console
Poster No:
1516
Submission Type:
Abstract Submission
Authors:
Marshall Xu1, Steffen Bollmann2, Stefan Sommer3, Daniel Nanz4, Pim Pullens5, Kerrin Pine6, Kelvin Chow7, Rainer Schneider8, Tobias Wuerfl8, Markus Barth9, Daniel Güllmar10
Institutions:
1The University of Queesland, Brisbane, Queensland, 2The University of Queensland, Brisbane, Queensland, 3Advanced Clinical Imaging Technology, Siemens Healthineers International AG, Zurich, Switzerland, 4Swiss Center for Musculoskeletal Imaging (SCMI) Balgrist Campus, Zurich, Switzerland, 5Ghent University Hospital, Ghent, Belgium, 6Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany, 7Cardiovascular MR R&D, Siemens Healthcare Ltd., Calgary, Alberta, 8Magnetic Resonance Imaging, Siemens Healthineers AG, Erlangen, Germany, 9The University of Queensland, Brisbane, Australia, 10Institute for Diagnostic and Interventional Radiology, Jena, Germany
First Author:
Co-Author(s):
Stefan Sommer
Advanced Clinical Imaging Technology, Siemens Healthineers International AG
Zurich, Switzerland
Advanced Clinical Imaging Technology, Siemens Healthineers International AG
Zurich, Switzerland
Introduction:
Neurodesk (Renton et al., 2024) is a powerful tool for reproducible neuroimaging , offering software containers for image analysis accessible via a browser or command-line interface. Traditional workflows require exporting and re-importing DICOM data, a tedious, error-prone process lacking immediate feedback essential for quality assurance. While scanner vendors provide internal pipelines like Siemens ICE, these are complex, vendor-specific, and incompatible with common neuroimaging tools. With open formats like ISMRMRD (MRD) (Inati et al., 2017) and frameworks like Gadgetron (Hansen & Sørensen, 2013), imaging data can now integrate directly with software containers. We extended Neurodesk to support Open Recon (Wallace et al., 2024) containers, enabling seamless image processing directly on scanners, improving workflow flexibility and usability for developers and users.
Methods:
The Open Recon framework requires software containers built with a specific Docker version, specific JSON configuration files and MRD server software that can receive data from the reconstruction system. This container by default runs on the scanner hardware itself and doesn't require any external setup. We expanded Neurodesk's build scripts to integrate the open-source Python MRD server (https://github.com/kspaceKelvin/python-ismrmrd-server) using Neurodocker syntax, allowing easy addition to any Python-based container. You can see an example setup in Neurodesk's repository here: https://github.com/NeuroDesk/neurocontainers/blob/master/recipes/vesselboost/build.sh and in Figure 1 (a)(b).
Additionally, we developed an automated Open Recon build workflow on GitHub, leveraging GitHub Actions to generate and test Open Recon files in the required formats. This automated workflow streamlines the build process, ensuring each build is consistent and ready for deployment on the scanner. For an example of an Open Recon build workflow, see https://github.com/NeuroDesk/openrecon/tree/main/recipes/vesselboost
We documented the necessary workflow in a community editable website and are inviting the community to contribute their image processing workflows: https://www.neurodesk.org/docs/getting-started/neurocontainers/openrecon/
Additionally, we developed an automated Open Recon build workflow on GitHub, leveraging GitHub Actions to generate and test Open Recon files in the required formats. This automated workflow streamlines the build process, ensuring each build is consistent and ready for deployment on the scanner. For an example of an Open Recon build workflow, see https://github.com/NeuroDesk/openrecon/tree/main/recipes/vesselboost
We documented the necessary workflow in a community editable website and are inviting the community to contribute their image processing workflows: https://www.neurodesk.org/docs/getting-started/neurocontainers/openrecon/
·Figure 1 – (a) Overview of the workflow of building Neurocontainer for OpenRecon. (b) Subprocess of building an application using VesselBoost as an example. (c) Overview of the proposed build system
Results:
We developed and tested several containers for Open Recon on two sites with a MAGNETOM Terra.X, XA60, building on existing Neurodesk containers. The first example is a synthetic CT application using a U-Net model on T1-weighted SPACE sequences with GPU acceleration (Figure 2a). The second example is based on the recently released VesselBoost (Xu et al., 2024) toolkit, which segments Magnetic Resonance Angiography (MRA) data (Figure 2b). We also developed a brain extraction container, which uses the well-known BrainExtractionTool (BET) (Smith, 2002) from the FSL toolbox (Figure 2c).
Further Neurodesk-based OpenRecon containers include Synthstrip (Hoopes et al., 2022), which removes the skull from brain images (like BET) but can also efficiently deface a participant for anonymization purposes using Quickshear (Schimke & Hale, 2011). Additional containers include an AFI B1 mapping tool, with many more tools expected to be contributed by the community shortly.
Further Neurodesk-based OpenRecon containers include Synthstrip (Hoopes et al., 2022), which removes the skull from brain images (like BET) but can also efficiently deface a participant for anonymization purposes using Quickshear (Schimke & Hale, 2011). Additional containers include an AFI B1 mapping tool, with many more tools expected to be contributed by the community shortly.
·Figure 2 – (a) Synthetic CT contrast running on scanner’s GPU (b) VesselBoost segmentation pipeline applied on a ToF (c) FSL BET applied directly on an MPRAGE
Conclusions:
Simplifying the container-building process for Open Recon makes building such containers accessible to a broader range of contributors. This work allows anyone to develop and test imaging pipelines within the Neurodesk framework and seamlessly deploy them on the scanner console via Open Recon (see Figure 1c).
Modeling and Analysis Methods:
Methods Development 1
Neuroinformatics and Data Sharing:
Workflows 2
Keywords:
Data analysis
Development
Open-Source Software
Workflows
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
Hoopes, A., Mora, J. S., Dalca, A. V., Fischl, B., & Hoffmann, M. (2022). SynthStrip: Skull-Stripping for Any Brain Image. NeuroImage, 260, 119474. https://doi.org/10.1016/j.neuroimage.2022.119474
Inati, S. J., Naegele, J. D., Zwart, N. R., Roopchansingh, V., Lizak, M. J., Hansen, D. C., Liu, C., Atkinson, D., Kellman, P., Kozerke, S., Xue, H., Campbell‐Washburn, A. E., Sørensen, T. S., & Hansen, M. S. (2017). ISMRM Raw data format: A proposed standard for MRI raw datasets. Magnetic Resonance in Medicine, 77(1), 411–421. https://doi.org/10.1002/mrm.26089
Renton, A. I., Dao, T. T., Johnstone, T., Civier, O., Sullivan, R. P., White, D. J., Lyons, P., Slade, B. M., Abbott, D. F., Amos, T. J., Bollmann, S., Botting, A., Campbell, M. E. J., Chang, J., Close, T. G., Dörig, M., Eckstein, K., Egan, G. F., Evas, S., … Bollmann, S. (2024). Neurodesk: An accessible, flexible and portable data analysis environment for reproducible neuroimaging. Nature Methods, 21(5), 804–808. https://doi.org/10.1038/s41592-023-02145-x
Schimke, N., & Hale, J. (2011). Quickshear defacing for neuroimages. Proceedings of the 2nd USENIX Conference on Health Security and Privacy, 11–11.
Smith, S. M. (2002). Fast robust automated brain extraction. Human Brain Mapping, 17(3), 143–155. https://doi.org/10.1002/hbm.10062
Wallace, T., Morales, M., Assana, S., Demirel, O., Chow, K., & Nezafat, R. (2024). Accelerating Cardiac MRI with Inline Generative AI and Open Recon. MAGNETOM Flash, 87.
Xu, M., Ribeiro, F. L., Barth, M., Bernier, M., Bollmann, S., Chatterjee, S., Cognolato, F., Gulban, O. F., Itkyal, V., Liu, S., Mattern, H., Polimeni, J. R., Shaw, T. B., Speck, O., & Bollmann, S. (2024). VesselBoost: A Python Toolbox for Small Blood Vessel Segmentation in Human Magnetic Resonance Angiography Data. Aperture Neuro, 4. https://doi.org/10.52294/001c.123217