Hi everyone, just wondering if there’s an accepted normal regarding image overwriting in PACS.
We have modalities where corrections sent to PACS with the same SOP instance UID overwrite the original image (the UID is identical, but checksum is different)
I think it’s medicolegaly irresponsible to allow an image overwrite automatically, but I’m not seeing anything in my quick look in IHE and Dicom standards to back me up.
Does anyone have any concrete guidance on this please?
Thanks
Stu
The IHE profiles and their use of DICOM (and the expectations put on the actors, the Modality and the PACS in this case) are quite clear. This does not mean that your Modality and PACS can do this, or that it is tricky for the suppliers to set up so corners may have been cut.
The correct way for the Modality to issue the update is to generate a new object (new SUID) and then the IHE IOCM (Image Object Context Management) profile to update the PACS records with a new object in a managed way. It appears as present in many Modality and PACS conformance statements but is not always implemented. Using this profile would allow the Modality to inform the PACS that “this Study has been updated”.
The PACS has choices on how to implement the behaviour of corrections or updates, but it should deprecate old images and display updated ones. IOCM can also be used for a system add additional ROIs or annotation to images and have them transferred between systems as updates, not overwrite.
Without this controlled IOCM capability implementors are left with the difficult choice of how to act on the updated images just sent in by a Modality with the same SUID (if the object is different it really should have a new SUID, keeping the old one breaks the DICOM standard rules) - and usually that is to “overwrite”. You are correct in thinking this convention is less than perfect - it removes an audit trail from the import activity and it eaves an “integration behaviour by convention” that is uncontrolled by the systems an may lead to future safety issues if the modality is replaced or if a PACS software upgrade does not correctly replicated the expected behaviour for that modality.
IOCM was drafted in 2011 and finalised in 2014. This standard has been in place for more than 10 years and my view is that Modality and PACS suppliers should be pushed to use it.
Thanks Robin, I appreciate the detailed answer, I’ll see if I can make sense of the IOCM documentation now!
Cheers
Stu