Multiple modality LUT

[mail2...@gmail.com]

I have 2 doubts.
1. I got an image and a PR generated by Philips modality. Both image and PR are having linear modality LUT, but the values are different.
Is it ok to have different values for the modality LUT in PR and image?

2. In the application I am working, there is an option to switch PRs. So if there are multiple PRs referring the same image, we can switch between them.
Is there a chance that modality LUTs in multiple PRs referring same image have different values?

[Marco Eichelberg]

Dear Anonymous,

> 1. I got an image and a PR generated by Philips modality. Both image and PR are having linear modality LUT, but the values are different.
> Is it ok to have different values for the modality LUT in PR and image?

Yes. The values in the PR override the values in the image.

> 2. In the application I am working, there is an option to switch PRs. So if there are multiple PRs referring the same image, we can switch between them.
> Is there a chance that modality LUTs in multiple PRs referring same image have different values?

Yes. I cannot think of any *useful* use case for this, but
certainly it' permitted by the standard, so technically it's possible,
and with PRs generated by different devices it might also
practically happen.

Best regards,
Marco Eichelberg
OFFIS

[mail2...@gmail.com]

Hi Marco Eichelberg,

Modality LUT is used to transform the pixel values to DICOM values. That means modality LUT is a mapping betweeen the pixel values and the final DICOM values.
The output range(DICOM values) should remain the same and the input values(pixel values) are also same, then how come the mapping functions\LUT can change?

I was planning to do an optimization in my application for applying non-linear modality LUTs. Right now I will have to store the original pixel data and also the buffer on which the non-linear modality LUT is applied. The original pixel data is stored so that whenever user switches from one PR to another, the modlaity LUT can be reapplied. The problem with this approach is that memory usage is doubled. So I wanted to apply the modality LUT only once and store only that buffer. But from your explanantion I think it will not be possible, since there is a chance that different PRs may contain different modality LUTs.

Regards,
R Manu

[Wim Corbijn]

Dear Manu,

If this is for MR data then the values should be same if they are
present in both objects.
Only when you remove the information from the Image objects you will
see a difference.
This removal is an option for the transfer (configured for the node)
to help systems that can't handle the presence for MR objects.

But then you will only have no Modality LUT in the image objects and
it will only be present in the PR object.

Regards,
Wim

[mail2...@gmail.com]

Hi Wim,

The images belongs to MR modality and the modality LUT tag is present in the image as well as in the PR. If modality LUT was not present in the image, then it was understandable, but in this case its present in both(image and PR) and are having different values.

Is it mandatory that the modality LUT values in image(if present) and PR should be same for some modalities?

Regards,
R Manu

[Wim Corbijn]

Hi Manu,

This would be a discussion on its own as the Modality LUT is not part
of the Classic MR object definition.
For the Enhanced object definition it is.
Consistency between the two objects is not defined in the standard.
For us we have internally no reason to change them and as such they
should be the same.
Normally PR would overrule everything in the image, but for this one I
would tend to stay with the image values.
In the meantime interested in seeing some samples to see if I
understand the difference.

Regards,
Wim