Click here to

Session: Multi-Disciplinary: Research Tools and Databases [Return to Session]

PIXNAT: An XNAT Imaging Archive for Collaborative Radiotherapy Research

E LoCastro*, A Apte, A Iyer, J Deasy, Department of Medical Physics, Memorial Sloan-Kettering Cancer Center, New York, NY


TU-IePD-TRACK 4-5 (Tuesday, 7/27/2021) 12:30 PM - 1:00 PM [Eastern Time (GMT-4)]

Purpose: We introduce our research imaging archive, the “Predictive Informatics XNAT” (“PIXNAT”). The platform is built on the well-known XNAT platform, which we work to enable with the CERR computational platform, enhanced with custom pipelines to assist with conversion between DICOM and analysis-friendly data formats. PIXNAT-housed data is secure and accessible to external researchers for research and processing.

Methods: The PIXNAT server is a RedHat Enterprise Linux instance running XNAT v1.7.5. Pipeline processes securely run on institutional HPC cluster with multi-GPU for rapid deep network processing. PIXNAT is approved by institutional information security for housing and sharing research data. All imaging data is fully anonymized using CTP and according to 18-point institutional guidelines. PIXNAT is extended with OHIF-XNAT and ROI datatype plugins. CERR computational routines are implemented as PIXNAT pipelines to provide access to a variety of image analysis tools. Deep learning segmentation pipelines from CERR work directly from PIXNAT imaging data and produce RTSTRUCT-based segmentations to be imported and overlaid in the viewer. CERR texture map analysis is also implemented on the system, the resulting maps are converted to DICOM and imported into the session list for review with the integrated OHIF viewer. We have enabled a launch pipeline for JupyterLab notebooks which can be pre-loaded with images from PIXNAT to interactively run Python and CERR routines and visualize in-line results.

Results: PIXNAT currently houses 24 projects, 2243 subjects, and 3067 imaging sessions. It has been used to collaborate with 5 different institutions.

Conclusion: We have demonstrated a close integration between CERR and XNAT, resulting in a useful inter-institutional collaboration paradigm. We have furnished PIXNAT as a destination for medical physics research, open source image processing and data sharing.

Funding Support, Disclosures, and Conflict of Interest: Research was supported by the following grants: (1) P30 CA008748/CA/NCI NIH HHS/United States (2) R01 CA198121/CA/NCI NIH HHS/United States



    Not Applicable / None Entered.


    Not Applicable / None Entered.

    Contact Email