fix: SeriesInstanceUID fix (compatibility with other software) #96
Conversation
| ds.SeriesDescription = getattr(reference_ds, "SeriesDescription", "") | ||
| ds.StudyInstanceUID = reference_ds.StudyInstanceUID | ||
| ds.SeriesInstanceUID = generate_uid() # TODO: find out if random generation is ok | ||
| ds.SeriesInstanceUID = reference_ds.SeriesInstanceUID |
There was a problem hiding this comment.
I believe your solution here is correct. The new images should be part of the previous study.
There was a problem hiding this comment.
@hjmjohnson Thanks for having a look at this! Would you mind reviewing this so it can be merged in a later release?
There was a problem hiding this comment.
This change would violate the DICOM standard, see https://dicom.nema.org/medical/dicom/current/output/chtml/part03/sect_C.8.8.html#sect_C.8.8.1.1
since RTstruct "Modality" tag is "RTSTRUCT", compared to the reference dataset which has "CT" modality, and it is not allowed to mix two modalities in the same series-instance-uid.
@code-lukas if this is the case, could you report this bug to RayStation? They usually reply quickly to bug reports.
fyi @lassoan
|
When will this be merged? |
4 participants
Hi there!
Thank you for providing
rt_utilsto ease the burden of constructing DICOM compliant RTSTRUCTs.We have used your work for research purposes and discovered that generating a random SeriesInstanceUID occasionally breaks functionality in other clinical software (such as RayStation ). So far we have not yet found any problems with the proposed change.
Kind regards