Orthanc + XNAT architecture

Google Groups archive
1

Hey everybody,

After a half year we managed to build a system where a user uploads a study through Orthanc and the data gets automatically analysed by a list of CADx models (e.g.: lung cancer detection). Orthanc then receives the prediction in DICOM SEG, SR, and GSPS - That all works fine!
We now have the need to edit those predictions and give feedback to the models. We saw the capabilities of XNAT in terms of “fancy” DICOM OHIF viewer with upload, create, store and edit functionalities.

What would be the ideal end goal to move to knowing we’d be embedding this into a hospital PACS (keeping Orthanc’s roadmap in mind)?

  1. Use XNAT instead of Orthanc as a PACS + Viewer + other plugins;

  2. Use XNAT as a research PACS, and Orthanc as an Enterprise PACS;

  3. Use XNAT-OHIF with Orthanc for added create, edit & remove ROI functionalities;

  4. Any other proposed architecture
    Any ideas and recommendations are more then welcome!

2

If I understand correctly, you want to find a viewer that would connect to Orthanc, display the prediction dicom instances and annotate them?

I am not familiar with the XNAT OHIF viewer, but if the viewer can connect to any Dicom SCP , then it will be able to connect to Orthanc without the need for a XNAT server.

Some other viewers also provide annotation abilities (Osimis Pro Viewer, MicroDicom). So I believe either your third solution (or Orthanc + any suitable viewer) would be fine.

Have you already tried to use XNAT-OHIF with Orthanc?