[BUG] Field SourceToSurfaceDistance Missing

[uyousafzai54]

Describe the bug
It seems that while trying to adapt/calibrate a phantom DICOM dataset to import in MatRad, the SourceToSurfaceDistance tag is missing.

I hard coded a value of 50 cm assuming its the distance from the jaw but I just wanted to see if there's any reason why specific meta tags would be missing a value. From the comments in the code it seems that all of this info is defined in the first control point sequence and the only value missing for my dataset is the SourceToSurfaceDistance tag.

Therefore, while trying to import an RT plan as part from the phantom dataset, is there specific info I might need to hard code in to calibrate/add missing data for MatRad to generate its .m file?

I can give specific reproduction steps and/or dataset (anonymized and cleared for sharing) if needed.

[wahln]

It seem the tag is only checked here:

matRad/dicom/matRad_importFieldShapes.m

Line 38 in 3d29b76

'PatientSupportAngle' 'CouchAngle';'SourceToSurfaceDistance' 'SSD'};

SourceToSurfaceDistance (SSD) means the distance from the (virtual) photon source to the patient's surface. It is used to get the correct photon pencil-beam kernel in the SVD algorithm. I am actually not sure why it is necessary here to have the field, as it is commented in this function

matRad/dicom/matRad_importDicomSteeringPhotons.m

Line 83 in 3d29b76

%stf(i).ray.SSD = Fields(ia(i)).SSD;

and it is computed in the dose calculation anyways.

So you should be OK with overriding this value in your DICOM RTPlan / changing the code to not obtain this field at all.

[uyousafzai54]

Sounds good! As an aside, is there any test DICOM dataset for MatRad for simply creating a .mat file that I can use for debugging/understanding the DICOM import process against my dataset?

[wahln]

Unfortunately not really. Particuarly not for RTPlan & RTDose (would be good to have this in a testing pipeline though).

The only thing coming to my mind is the CORT dataset. These are the phantoms/patients within the mat files coming with matRad, and the corresponding DICOMs can be found here: https://academic.oup.com/gigascience/article/3/1/2047-217X-3-37/2682969

[uyousafzai54]

It seem the tag is only checked here:

matRad/dicom/matRad_importFieldShapes.m

Line 38 in 3d29b76

'PatientSupportAngle' 'CouchAngle';'SourceToSurfaceDistance' 'SSD'};

SourceToSurfaceDistance (SSD) means the distance from the (virtual) photon source to the patient's surface. It is used to get the correct photon pencil-beam kernel in the SVD algorithm. I am actually not sure why it is necessary here to have the field, as it is commented in this function

matRad/dicom/matRad_importDicomSteeringPhotons.m

Line 83 in 3d29b76

%stf(i).ray.SSD = Fields(ia(i)).SSD;

and it is computed in the dose calculation anyways.
So you should be OK with overriding this value in your DICOM RTPlan / changing the code to not obtain this field at all.

Also since you aren't sure why the field is necessary as the line is commented out anyways in [matRad/dicom/matRad_importDicomSteeringPhotons.m](https://github.com/e0404/matRad/blob/3d29b765a9a0a02a593890726102a7d9db907af7/dicom/matRad_importDicomSteeringPhotons.m#L83), should I make a PR to address this?

[wahln]

A PR would be nice (of course add your name to the AUTHORS file as well then in the PR, if it is not already there ;-) ).
If you want to walk the extra mile, you could also introduce it as an optional check (i.e., if the field is there it will be read, otherwise not).