NA-MIC Project Weeks

Investigation of Kaapana and XNAT as platforms for data management and collaborative research

Key Investigators

Nadya Shusharina (BWH & MGH)
Klaus Kades (DKFZ)
Hanno Gao (DKFZ)
Randy Gollub (MGH)
Andrey Fedorov (BWH)

Project Description

We have installed the platforms on the Googe Cloud Virtual Machines. To investigate the functionalities, we populated the platforms with publicly available datasets.

Objective

Ultimately, our goal is to have a platform that can be used to support data management needs for internal research activities.

Archival of DICOM images
Flexible and extensible interface to explore archived data
In-browser visualization of images and annotations
In-browser segmentation of images
Integration of analysis tools and their application to the data available in the platform

Approach and Plan

Continue working on the evaluation of the platforms
Connect with any of the groups that worked on either Kaapana or XNAT to share experience
Meet with Kaapana maintainers to debug outstanding issues.
Work on resolving issues in the GCP installation.
Clarified status of XNAT OHIF plugin in the OHIF session (see minutes)

Progress and Next Steps

Current unresolved issues

XNAT:

not clear how to extend the search of DICOM metadata to include arbitrary attributes
not clear if/how to integrate analysis tools
workflow to access existing SEG/RTSTRUCT from XNAT-OHIF plugin is unclear (related post here)
not clear if integration of desktop annotation tools like Slicer is possible

Kaapana:

fixed MITK segmentation flow
need to build platform from source - deployment from prebuilt solution has limitations
need to work on integration of Slicer the same way as MITK - should be possible, can follow example, just need time
overall, no blocking issues identified, just need time to debug/implement