Stream: implementers
Topic: ImagingStudy vs IHE WIA
Øyvind Aassve (Oct 16 2020 at 07:21):
Hi, we are working with a purchase of a multimedia archive supporting DICOM. In addition to ImagingStudy the IHE-profile Web Image Access (WIA) is mentioned. Can anyone briefly describe the relationship between the two with regards to adaption to use cases?
René Spronk (Oct 16 2020 at 08:17):
ImagingStudy is the FHIR equivalent of a KOS-object, and WIA is effectively MHD for imaging, where (speaking from memory) the content.attachment of the DocumentReference is an ImagingStudy.
As I recall WIA was affectively abandoned in favor of the use of DICOMWeb based URLs in combination with DocumentReference (DocRef.content.attachment is the DICOMWeb URL to an imaging object).
Bjørn Ravnestad (Oct 16 2020 at 11:28):
WIA replaced MHD-I,which was deprecated in 2017. WIA prescribes the use of restful services (DICOMweb) but is agnostic about the infrastructure behind the Document Responder actor. The Document Responder thus can act as a facade to underlying imaging services. "WIA enables retrieval of imaging studies shared within an enterprise and across enterprises using
RESTful services. WIA can be used with different image sharing infrastructures, including but
not limited to XDS / XDS-I and DICOM / DICOMweb."
John Moehrke (Oct 16 2020 at 14:03):
@Elliot Silver ?
Elliot Silver (Oct 16 2020 at 22:53):
René Spronk said:
ImagingStudy is the FHIR equivalent of a KOS-object
Actually, it is closer to a C-FIND result. The now-deprecated ImagingManifest is closer to a KOS object.
@Bjørn Ravnestad is correct about WIA and MHD-I. Just as MHD can be seen as "XDS on FHIR," MHD-I could be seen as "XDS-I on FHIR." MHD-I used ImagingManifest as the FHIR representation of the KOS.
(All of the following is going off memories of stuff I haven't looked at recently, so pardon me if I get it wrong.)
However, MHD-I inherited many of the the shortcomings of XDS-I, both of which trace back to the limited information present in the KOS. That limited information makes both XDS-I and MHD-I suitable for study transfer, but less useful for on-demand viewing. For example, (going off memory) I don't believe image number, or flagging is present in either, which makes it hard to know which images to retrieve first.
On the other hand, WIA avoids the KOS and sticks to QIDO and WADO (or at least allows a facade in front of the KOS), and is better positioned for interactive use.
Back to @Øyvind Aassve's original question... I'd say Imaging Study is good for non-imaging-centric systems that need to be aware of the existence of a DICOM study, and a few of the key parameters of the study. As long as you know how to create a WADO-RS url, you can also display or retrieve the images. On the other hand, you WIA assumes a higher level of DICOM awareness, since it uses DICOM query, etc. and would be more appropriate for systems willing to "step into" the DICOM world in exchange for deeper information and capabilities.
Brendan Keeler (Oct 17 2020 at 06:30):
All the acronyms
Øyvind Aassve (Oct 17 2020 at 07:25):
Thanks, a lot of good information here about the imaging standards acronym-puzzle :). @Elliot Silver - to what extent would you say that EHRs are or should be willing to "step into" the DICOM world?
Elliot Silver (Oct 17 2020 at 08:31):
@Brendan Keeler : KOS - a DICOM Key Object Selection SOP class; WIA - IHE’s Web Image Access profile; MHD - IHE’s Mobile Access to Health Documents profile; MHD-I - MHD for Imaging; XDS - IHE’s Cross-Enterprise Document Sharing profile; XDS-I - XDS for imaging; C-FIND - traditional DICOM query; QIDO - RESTful DICOM query; WADO - RESTful DICOM retrieve. That should be most of them, except the ones I introduced in this response.
Elliot Silver (Oct 17 2020 at 08:42):
@Øyvind Aassve that is perhaps a better question for someone on the EMR side of the aisle. I think image viewing within the EMR will increase, but image archive will remain separate. New uses of DICOM outside of radiology, such as in pathology or dermatology, as well as point-of-care ultrasound and other encounter-based imaging will drive the need for more systems that deal with both worlds. I suspect some AI and CDS systems will be heavily EMR-focused but will want to use high-level imaging info.
Brendan Keeler (Oct 18 2020 at 05:05):
Oh, sorry, I'm familiar with all of them. It just reminds me how the imaging space is still very all-over-the-place with standards and formats, like this blog from a few years ago talks about - http://dclunie.blogspot.com/2016/03/how-many-medical-image-exchange.html?m=1
John Silva (Oct 18 2020 at 17:24):
Is this because of competition in the "imaging space" or just a growing maturity and recognition of the need to make "imaging systems" more interoperable with the "rest of the HealthCare IT ecosystem"?
Brendan Keeler (Oct 19 2020 at 13:59):
Some countries have converged on a standard. The Netherlands uses XDS-I in a widespread manner
Last updated: Apr 12 2022 at 19:14 UTC
