See discussion here: https://github.com/MGHComputationalPathology/dicomweb-client/issues/41 Looks like we should be following this older version of the standard where it's stated clearly: http://dicom.nema.org/medical/dicom/2018e/output/chtml/part18/sect_6.5.3.2.html