Our scanner is generating Whole Slide DICOM Images.
We use Photometric Interpretation = MONOCHROME2

For some reasons, we cannot use 0022,0055 (Illumination Wave Length.
As per DICOM standard, then we need to use 'llumination Color Code Sequence Attribute'

How to use this sequence to convey our color information.
Lets say, we have a channel with red color. How to assign this information to ''Illumination Color Code Sequence Item' ?

We use SLIM viewer for testing our images. When we view these images, it is all black and while image.

Hi Anil,

> Our scanner is generating Whole Slide DICOM Images.
> We use Photometric Interpretation = MONOCHROME2
>
> For some reasons, we cannot use 0022,0055 (Illumination Wave Length.
> As per DICOM standard, then we need to use 'llumination Color Code Sequence Attribute'
>
> How to use this sequence to convey our color information.
> Lets say, we have a channel with red color. How to assign this information to ''Illumination Color Code Sequence Item' ?

according to DICOM PS3.3 Section A.32.8.3.1.1, a code from the Context Group 8122 "Microscopy Illuminator and Sensor Color" [1] may be used for Illumination Color Code Sequence (0048,0108), e.g. (371240000, SCT, "Red").

[1] https://dicom.nema.org/medical/dicom/current/output/chtml/part16/sect_CID_8122.html

> We use SLIM viewer for testing our images. When we view these images, it is all black and while image.

I haven't used viewer yet, but for monochrome WSI that should be displayed in color I would suggest to explicitly provide a color palette in the Palette Color Lookup Table Sequence (0048,0120) [2] of the respective optical path.

[2] https://dicom.nema.org/medical/dicom/current/output/chtml/part03/sect_C..8.12.5.html#para_45f5394e-e16a-4e4e-965d-58f874a5e7cb

Regards,
Jörg

Follow-up: Actually, I already used this viewer, but didn't know that the name was "Slim" :-)

Here's are the official links:
- Project page: https://github.com/ImagingDataCommons/slim
- Online demo: https://imagingdatacommons.github.io/slim/

On Friday, August 25, 2023 at 1:24:30 PM UTC+5:30, Jörg Riesmeier wrote:
> Hi Anil,
> > Our scanner is generating Whole Slide DICOM Images.
> > We use Photometric Interpretation = MONOCHROME2
> >
> > For some reasons, we cannot use 0022,0055 (Illumination Wave Length.
> > As per DICOM standard, then we need to use 'llumination Color Code Sequence Attribute'
> >
> > How to use this sequence to convey our color information.
> > Lets say, we have a channel with red color. How to assign this information to ''Illumination Color Code Sequence Item' ?
> according to DICOM PS3.3 Section A.32.8.3.1.1, a code from the Context Group 8122 "Microscopy Illuminator and Sensor Color" [1] may be used for Illumination Color Code Sequence (0048,0108), e.g. (371240000, SCT, "Red").
>
> [1] https://dicom.nema.org/medical/dicom/current/output/chtml/part16/sect_CID_8122.html
> > We use SLIM viewer for testing our images. When we view these images, it is all black and while image.
> I haven't used viewer yet, but for monochrome WSI that should be displayed in color I would suggest to explicitly provide a color palette in the Palette Color Lookup Table Sequence (0048,0120) [2] of the respective optical path.
>
> [2] https://dicom.nema.org/medical/dicom/current/output/chtml/part03/sect_C.8.12.5.html#para_45f5394e-e16a-4e4e-965d-58f874a5e7cb
>
> Regards,
> Jörg

Thanks Jorg. Will do as suggested.

Regards,
Anil TA

Hi Anil

It is important to distinguish what the channel is, how it was acquired, and how it is recommended that it be displayed.

Slim might not pay any attention to a Palette Color Lookup Table encoded in the image, since I don't think that has been encountered in practical images (from scanner vendors or converted) though it is possible some of the test fluorescence images on the NEMA ftp server have them [1] (haven't checked recently); nor has labeling individual images acquired with a single color in Illumination Color Code Sequence.

If they don't work and you want these features to be considered for addition to Slim, add an issue on the github site, and provide sample images to test with along with a screenshot of expected results.

What does your "red" channel represent in semantic terms (i.e., regardless of the color that might be recommended for display, what does it represent, e.g., what antibody or target or whatever)?

If you really did illuminate the specimen with "red" light, as opposed to just wanting it to look red when displayed, then that's what Illumination Color Code Sequence is actually for (and that does not mean it has to be displayed as red).

Ideally, the "presentation" is really not something that would be in the image (as opposed to, say, a separate Advanced Blending Presentation State, which is what Slim uses for this esp. when there are multiple channels to be superimposed [2]), even though Palette Color Lookup Table is included in the IOD for the purpose of providing a default.

> On Friday, August 25, 2023 at 1:24:30 PM UTC+5:30, Jörg Riesmeier wrote:
> > I haven't used viewer yet, but for monochrome WSI that should be displayed in color I would suggest to explicitly provide a color palette in the Palette Color Lookup Table Sequence (0048,0120) [2] of the respective optical path.

David

1. ftp://medical.nema.org/medical/dicom/DataSets/WG26/
2. https://viewer.imaging.datacommons.cancer.gov/slim/studies/2.25.219049900737444567890964056686598614067/series/1.3.6.1.4.1.5962.99.1.2343322182.1764456793.1655905763910.4.0?state=1.2.826.0.1.3680043.10.511.3.21962974493472368945261909485101759

Hi David,

thank you for the additional, clarifying information to the OP.

> Slim might not pay any attention to a Palette Color Lookup Table encoded in the image, since I don't think that has been encountered in practical images (from scanner vendors or converted) though it is possible some of the test fluorescence images on the NEMA ftp server have them [1] (haven't checked recently); nor has labeling individual images acquired with a single color in Illumination Color Code Sequence.

I just created the Docker image of Slim and started it locally. After uploading a DICOM WSI fluorescence image with multiple optical paths and Palette Color Lookup Tables, I get a rather dark image on the screen, but the following text in the "Display Settings" of the individual optical paths seems to indicate that the Palette Color Lookup Table is actually used: "Custom pseudo-coloring is disabled because pixels are colorized via a provided palette color lookup table."

But if there is no Palette Color Lookup Table, how does the Slim viewer do the "pseudo-coloring"?

In any case, I'll play around a bit with this tool as time permits and further investigate the above findings...

Regards,
Jörg