Issue with special characters when scans where uploaded with SpecificCharacterSet dicom tag value as ISO_IR 13

General
1

Greeting everyone,

We are using orthancteam/orthanc:25.2.0 docker image. When we uploaded a scan which has dicom tag SpecificCharacterSet value as ISO_IR 13, we noticed that few of the tags were having the delimiter value as ¥

Ex:

  • 0020,0032 (ImagePositionPatient): -20.717017102556¥-99.021923065871¥144.95523497564

  • 0020,0037 (ImageOrientationPatient): 1.37e-13¥1¥4.895e-12¥-.0279215878283¥4.897e-12¥-.999610116462

Because of this the scans were not opening in ohif viewer.

However what we noticed in jodogne/orthanc-plugins:1.12.3 docker image was that the same tags were appearing like the following

  • 0020,0032 (ImagePositionPatient): -9.6000003814697-130\150

  • 0020,0037 (ImageOrientationPatient): 0\1\0\0\0-1

\ is being used as delimiter

Kindly help us resolve the issue in orthancteam/orthanc:25.2.0 docker image

2

Hi @adhithya

Could you share a sample file for us to reproduce on our side ?

Thanks,

Alain

3

Hi @alainmazy

Please find the anonymized scan. I have used the anonymize option in Orthanc UI.
MR000005.dcm (86.5 KB)
MR000004.dcm (85.7 KB)
MR000003.dcm (84.7 KB)
MR000002.dcm (84.1 KB)
MR000001.dcm (84.3 KB)

Have attached 5 instances of a series

4

Hi @alainmazy Was the attached files helpful in observing the issue

5

Hi @adhithya

Thanks for the sample files.
I just performed a quick test and here are the results:

  • Windows installers: OK
  • jodogne/orthanc:1.12.8: OK
  • orthancteam/orthanc:25.6.4: KO

So, this means that this is related to the way the orthancteam/orthanc images are compiled. I’ll have a deeper look when I find some time but this might take some time …

Best,

Alain.

6

Thank you @alainmazy . Will wait for your analysis and possible solution

7

Hi @alainmazy,

May I kindly know if there is any solution for this problem

8

Hi,

No progress on that topic so far.

9

Hello,

I have just committed a fix that should solve your issue.

To make a long story short: In the ISO_IR 13 encoding, backslash characters (i.e., \, which is encoded as the hexademical value 0x5C in the ASCII character set) are interpreted as the Yen character (i.e., ¥). But, the DICOM standard specifies that, for ISO_IR 13, \ should only be interpreted as ¥ if the value representation of the DICOM element is UT, ST, or LT. This is precisely what does the fix above. Check out the DICOM specification for more information.

Best Regards,
Sébastien-

10

Thank you so much. We are using orthancteam/orthanc docker images. Currently I see that the latest update was 22 days ago. May I please know when this fix will be available in the docker images.

11

The fix will be available when we release Orthanc 1.12.9, likely within 2 weeks.

However, it is already available in the mainline binaries