Incorrect character encoding in Dicom RDSR

[Rob Ireland]

I am posting in case this is helpful to other users.

Is there anything in Orthanc / OpenREM that could cause incorrect character encoding or are the files in the Orthanc storage exactly as query retrieved from PACS?

Installation: OpenREM Version 1.0.0b3 with Orthanc 26.1.0 running on Windows Server 2022

Issue started: After GE CT Revolution EVO software update Qilin 2.0 EVO (version 24BW11.73).

Issue: Acquisition protocol names that include a % sign are wrongly displayed within OpenREM (correctly displayed before GE update). Replacement of % is not consistent although some do repeat. NAN occurs for some.

Example: ‘1.2 Helical Head 50-0X1.1ACA800000001P+1020SIR V’ or ‘1.2 Helical Head 50-NANSIR V’ instead of ‘1.2 Helical Head 50% ASIR V’.

Investigation: On PACS and on the GE scanner, the protocol name is displayed correctly. By suspending the automatic deletion of Dicom RDSR in Orthanc, I was able to locate a GE CT RDSR and confirm that the acquisition protocol name is wrongly displayed. Does this imply the original file on PACS was corrupt or is it possible the corruption has been caused by Orthanc / OpenREM?

Solution: Deletion of the % sign from all protocol names solved the problem. I do not know if other characters can be wrongly encoded within these RDSR.

Thank you

Rob Ireland

Royal Cornwall Hospital, Truro, UK