PaLM Conf Minutes 2018-July-11
|Riki Merrick||Alesandro Sulis|
|Raj Dash||Francesca Frexia|
|David Beckman||Nicholas Jones|
|David Clunie||Dan Rutz|
|Kevin Schap||Mario Vallace|
|Mary Kennedy||Nick Hasselhorst|
Riki reviewed the agenda. SET update: Alessandro is building the z structure messages and will share the proposed structures and issues for each event. His goal is the end of this week or prior to the next call (8/8/2018). We will use the second hour for his updates. See the follow up on SET related items on the F2F summary slides #3/5.
Riki is finishing the LCC. The updated LCC publication date is July 15.th Riki will follow up with Mary Jungers on TF publication TMA: The changes have not been made yet; it will be published for trial implementation after 8/8 SNOMED CT follow up: Riki will send a note to with Chris Carr requesting status. Digital pathology: Raj reviewed the work from the F2F.
- Acquisition is top priority = looking at workflow and how to split between HL7 and DICOM transactions
- There MUST be an HL7 option instead of RAD-49 = Instance availability notification
- Image archive to support both, but in the lab side mostly using HL7 only
- Order filler to Acquisition manger transactions:
o Current HL7 interfaces don’t have receive messages at this time – DICOM transaction can tell you , if they worked and include meta data about the image
- Issues that are standard independent:
o Both DICOM and HL7 support acknowledgements for receipt and application level
- Radiology just defined new ORU message in encounter-based workflow profile: https://ihe.net/uploadedFiles/Documents/Radiology/IHE_RAD_Suppl_EBIW_Rev1.0_PC_2018-02-21.pdf
- Broker use is attractive for the architecture = acquisition manager actor has value to help with install base (Is this query or broadcast?)
- Scanners do not want push notification of all possible s=order vs query for order, when barcode has been scanned; use of queries in HL7 world has been elusive. This is where LAW comes in – defines both push and query.
- Protocol assignment based on accession number is how a lot of the clinical chemistry analyzers work. No issue to have IHE create both options and see what comes through in the market
- There are three models of Acquisition:
o PUSH = explicit order o QUERY o Image identification reconciliation
- Real time notification ahead of image being available
- EHR may not want individual slide level detail ahead of time – when you need to pass images around outside the report is different
- Hierarchical ID = study instance ID that is the highest level, that then have unique IDs for each slide – proxy is accession number in RAD, probably the same in AP
- AP comes in asynchronously from the slide perspective compared to a single radiology process in a larger study (e.g. a CT scan is not missing a slide in the set of CTs)
- Order filler here is AP LIS – that needs to know that ALL slides have been produced.
- Need to be able to identify when case is incomplete – how do we translate that to digital pathology?
- Order filler vs acquisition manager – protocol defines what a complete study needs – that can be specified between order filler and acquisition manager
- There is still a need for human interactions in the AP workflow – so you don’t have to standardize EVERYTHING; need to have meta-data about what is available – but to decide if that is everything you need for a case should be human decision for now (or defined by proprietary protocol)
- Keep transaction for now at the individual image level for starter. Include explanation of how some of this will work in addition to the definition of testable profiles
- Need to include thoughts on how systems can break / error handling = that can include normative rules
- Look at RAD-14 and 16 for whole object transactions
- There are new TCCP transactions DICOM WG26 has been working with – web image access uses these RESTful transactions (Radiology – may be reusable but adjust for AP microscopy – part 18 in DICOM
The call ended at 10:10am Central. 1