What attributes to anonymise for a Secondary Capture, Hardcopy ... ?

[guy]

Hello, I finallise my work on dicom PRINT SCU.
I would like if it is possible that you tell me what are the informations to conserve when making a Dicom Secondary Capture, idem for an Hardcopy and Stored Print Object.

For example, I get a Dicom file, and I modify its pixels. When creating the Secondary Capture, do I have to set the informations (like patient's module, study...etc) of the Secondary Capture as there were in the original Dicom. Same thing for Hardcopy and Stored Print.

I Have noticed that DCMTK anonymized all when creating the Hardcopy Grayscale.

Is there a part in the Norme which specifies it ? if yes, can you tell me where ?

thanks

[Thomas Wilkens]

In part 3 of the DICOM standard you will find the IOD (=data structure) definitions of all objects. You can find out from there what information is mandatory and what is user optional.

[guy]

Hello Thomas, I know what you tell me in your post.

I have marked the mandatory and optionnal attributes.
What I want to know is the conservation of the values of the attributes when passing from a dicom file to a secondary capture and then to an harcopy.

Example

Dicom Original

Patient Name = "Doe^John"
Patient Sex = "M"
...


Secondary Capture

Patient Name = ?? ("Doe^John or ^^^^ ?)
Patient Sex = ??


HardCopy

Patient Name = ??
Patient Sex = ??

Stored Print

Patient Name = ??
Patient Sex = ??


I think it's clearer now for you, if you can tell me where find the informations concerning what values to conserve for all the modules (patient, study, series...), it will be great. I notice that DCMTK make anonymous all Secondary Capture and Hardcopies, is it what i have to do ?

[Marco Eichelberg]

You may anonymize the image but you do not have to. Actually there is no rule in DICOM that would define what attributes to keep or to modify when generating a new object (which is what you do here) from an old one. You can keep the patient and study context (Patient ID, Study Instance UID etc.) or modify it if deemed useful. You must modify the series if you change the SOP Class UID, and you must assign a new SOP Instance UID. Our implementation just removes all attributes that would never be sent to the printer anyway. In the DICOM print protocol, the printer never "sees" any specific patient data (unless burned into a bitmap or stored in an annotation box), and it does not have to, because hardcopies are just printed and normally not archived. This data could be communicated to the Print SCP in the Proposed Study Sequence (2130,00A0) and the Original Image Sequence (2130,00C0), both of which are optional and not supported by our implementation.