Release Notes

Important Note

Version 3.0 of the Viewer plugin requires XNAT 1.8 in order to run correctly. See XNAT OHIF Plugin Compatibility Matrix



  • Fixed JSON metadata not being updated when new data is added to existing session. This was causing viewer to only display part of the session.
  • Fixed scan date issue.
  • Added milliseconds to the ROI assessor labels.


Version 3.0.0 of the XNAT OHIF Viewer focuses on defining regions-of-interest, with the "ROI Collection" datatype now a part of core XNAT, which removes the dependency on having a separate datatype plugin. There are new mask creation tools and an integration of NVIDIA Clara's AI-Assisted Annotation, which seamlessly applies machine-learning models to auto-segment ROIs that are both editable within the viewer and directly save-able within XNAT. There is support for DICOM "fractional segmentations", allowing the representation of volumetric masks that are probabilistic (e.g., allowing the representation of the degree of certainty of machine learning models). ROI creation on 2-D image modalities is also now available. There are additional enhancements to support for contour datatypes

We track core upgrades to the architecture of the main OHIF project, in particular the transition to a more modern React JavaScript framework to improve viewer performance in XNAT and allow the inclusion of major new functionality, such as a 3D-MPR mode for the first time. 

New & Improved


  • Fix event-triggered JSON metadata generation for session uploads.
  • JSON metadata format update to support OHIF 4.2.7 requirements.