Hi,
Your issues sounds as if it is related to how you are getting the segmentations back into the new XNAT, rather than whether they were actually backed up. You don't specify in your post above whether the segmentations are ones that you created within the viewer or whether you had previously imported them into the original XNAT instance from another source, but I'm guessing the former.
At the moment the correct way of uploading segmentations to XNAT is either via the external app roiuploadassistant (
https://bitbucket.org/icrimaginginformatics/roiuploadassistant/src/master/) or via the XNAT REST API (details of the roi-api are on the Swagger page, accessed via Site Administration -> Miscellaneous), which you would most likely use in conjunction with something like XNATpy.
Hence, what you will need to do is to find and separate out the DICOM-SEG from the image DICOM files in your zip download, then upload the images back to the new instance via your normal route, then treat the DICOM-SEG files as above.
Both the viewer development team at ICR and the XNAT core team at WashU are aware that it would be good to have the upload functionality incorporated directly in XNAT. The situation won't change for the forthcoming XNAT 1.8 release, but beyond this discussions are ongoing as to how best to implement this.
Best wishes,
Simon