[Referring to http://dicom.nema.org/medical/dicom/2016e/output/chtml/part18/sect_6.5.html]
When we are talking about WADO-RS, NEMA mentions that:
Every request (we'll leave out /metadata & /rendered requests for now) can have accept-type of three types:
1. multipart/related; type="application/dicom" [dcm-parameters] ------- (DICOM File format as mentioned in PS3.10)
2. multipart/related; type="application/octet-stream" [dcm-parameters] ------- (Bulk data)
3. multipart/related; type="{media-type}" [dcm-parameters] ------- (Bulk data)
For all these accept types, response is created as multipart with each part corresponding to a particular Instance. Now I understand the first case (application/dicom) in which we'll have fill each response part with each SOP Instance's .dcm counterpart. (for e.g., if the WADO RS is for a Study, then the multipart response will have one part for each SOP Instance's Dicom File Stream) But when it comes to the bulk data I have few questions:
- What exactly is bulk-data in WADO-RS standard? Is it only the 7FE00010 tag, or is it all the binary tags of an SOP Instance combined into one single binary data?
- If it is just 7FE00010, then there will be one http response part each for every SOP Instance. Then how will the WADO-RS client come to know which bulk data is of which SOP Instance?
Information about this is limited on the internet. Hence asking here. If any one has any article about this, that's welcome too.
Ps: I am new to DICOM/DICOMWeb
What I've generally used is to look at what's included (or should it be excluded?) in the Composite Instance Retrieve Without Bulk Data service class: