Stream: Dicom-sr
Topic: Open Points on DICOM SR to FHIR Mapping
Diana_Ovelgoenne (Dec 07 2021 at 07:41):
Hi @Jonathan Whitby @Christopher Lindop ,
While doing the exercise of mapping DICOM SRs to FHIR based on the definitions in Confluence, I see the following open points:
a) Which wrapper shall be used, is it a Composition or a Diagnostic Report? If I refer to SR Objects created by an AI algorithm, definitely it isn't a Diagnostic Report. If I query for the Diagnostic Reports for this patient, I don't want to see data created by a machine, but the legal document signed by a radiologist.
b) Another point is the code of the wrapper, which code to be used? DICOM TID1500? If so, how to encode this? TID_1500? TID1500? Or is it a LOINC code as it is no longer a DICOM artifact but a FHIR generated one? If so, when querying LOINC there are the following options, which one is the correct one?
18782-3 Radiology Study observation (narrative) Study observation
19005-8 Radiology Imaging study [Impression] (narrative) Imaging study
68604-8 Radiology Diagnostic study note Diagnostic study note
or even
18834-2 Radiology Comparison study (narrative) Comparison.study
c) Still the mapping from the 3 flags to status, is then the preliminary flag mapped to the observation status and verification flag to the wrapper status flag? Are the 3 flags really needed?
d) Which are the equivalent codes for the different flags, some are straight forward like preliminary; unverified for example which value is the best? An extension to the valueSet cannot be considered as an option.
I hope these points get addressed soon, as the rest of the mapping documented works quite fine.
Last updated: Apr 12 2022 at 19:14 UTC
