open source imaging informatics
open source imaging informatics
After a long time, the second major version of the OHIF Viewer plugin is officially out of beta.
Major features over 1.0.X:
Distributions for 1.7.4.1/1.7.5.X compatible versions are available here.
(This branch also has versions built around the 1.7.6 distribution, working at the time of writing. But are not yet officially supported as XNAT 1.7.6 isn't finalized.)
Full patch notes are available here.
The plugin has a hard dependency on the ROI plugin by James D'Arcy, which deals with storage and conversion between data types on the backend. At the time of writing the relevant version of the ROI plugin is contained within the distribution.
Once both plugins are installed you may access the viewer from the "View Session" button on a session page to view a session, or the "View Subject" button on the subject page to view all sessions for that subject. E.g.:
You can create contour-based ROIs by clicking on the ROI -> Draw, and giving the ROI a name:
You can then start drawing your ROI with the left mouse button/touch. Drag around your region and back to the origin. Interpolation of contours between slices is on by default, but can be toggled via the Contours sidebar.
You can make corrections to your contour with the Sculpt tool:
The Contours menu lets you add more ROI Contours and switch which one is actively being drawn:
Clicking Export in the Contour menu will bring up a prompt selecting which ROIs to export (all by default), as well as give the collection a name:
After exporting, the ROI Collection will be archived in XNAT as an ROI Collection assessor in the appropriate session:
Another user with read access to the session can click the Import button in the Contours menu and may import the collection to view:
ROI Collections are imported as "locked", to avoid accidentally editing the author's ROI when reviewing it. If you want to edit an ROI Collection, click the Unlock button next to the collection name:
The ROIs will be moved to the "New ROI Contour Collection" and unlocked. You may then edit these as desired and save them back as a new collection.
Overwriting collections is not possible at the time of writing.
You can create mask-based ROIs by clicking on Segment, choosing a brush type and filling in a metadata form:
There are three types of brushes: Manual, Smart CT and Auto:
| Manual | Paints to the canvas based on brush size |
|---|---|
| Smart CT | Paints tissue of a particular type on CT images, gated by Hounsfield units. |
| Auto |
The Smart CT and Auto brushes can be configured using the menu at the bottom of the Segments menu. The desired CT gate can be chosen, or a custom one defined.
Both the Smart CT and Auto brushes have hole-filling and island-removal functionality, which generally means less cleanup is needed compared to just a simple threshold.
You can manage your segments in the Segment menu and export them similarly to Contour-based ROI Collections:
Due to technical limitations you may only import one mask-based segmentation (Which may still contain up to 20 segments) over each scan at a time.
Note
Multiple segmentation layers, as well as better support for segmentations with up to 65535 segments is coming to cornerstone 4.0, and can later be integrated into the XNAT-OHIF viewer.
You may choose which to import in the Segments -> Import menu:
Segmentations don't "lock" on import, as they aren't interactable without explicitly using a brush, unlike Contours. You may edit a mask-based segmentation immediately after import, and simply use Segments -> Export to export it as a different collection.
Clicking on the XNAT Nav button will bring up a sidebar which allows you to navigate to other XNAT views. You will be presented with all projects available to you. The blue buttons at subject level will open a subject-level view, whereas the buttons next to a particular session will open up that session.
If ROI Collections are available for a particular session, this is indicated by a couple of icons and a count of the number of collections.
XNAT is an open source project produced by NRG at the Washington University School of Medicine | NRG Home
Contributions to the XNAT Documentation site are licensed under a Creative Commons Attribution 3.0 Unported License. 
