Modify dose output .nii.gz file name

[TmnGitHub]

Hi,

Could some please guide me on how to modify the dose file format to add RTPlanLabel. Below code is working but {RTPlanLabel} dose not show up in file name.

Thanks!
output_dose_name_format='{parent_sorting_data}_{study_uid_index}_{DoseSummationType}_{RTPlanLabel}' process_dicom_directory( dicom_directory = input_dicom_dir, output_directory = output_dir, output_dose_name_format=output_dose_name_format)

[TmnGitHub]

Any suggestion @pchlap @rnfinnegan Thanks!

[pchlap]

Hi @TmnGitHub, hard to say exactly why this isn't working.

Check the contents of all_naming_fields after it is populated here:

platipy/platipy/dicom/io/crawl.py

Line 1017 in b1168be

all_naming_fields = [

(You can use a print statement or put a debugging break point).

If RTPlanLabel is in that list as expected, then next check that it is being loaded into the naming_info_dict in this loop:

platipy/platipy/dicom/io/crawl.py

Line 1027 in b1168be

for dicom_field in dicom_header_tags:

Let us know what you find.

[TmnGitHub]

When I change that parameter it dose not work. I have to give the exact parameter or there is this error.
output_dose_name_format="{parent_sorting_data}{study_uid_index}{RTPlanLabel}"

There are other errors:

` 1011 all_naming_fields = [
1012 i[i.find("{") + 1 :]
1013 for i in output_structure_name_format.split("}")
1014 if len(i) > 0
1015 ]
1016 elif dicom_type == "DOSES":
1017 all_naming_fields = [
1018 i[i.find("{") + 1 :]
-> 1019 for i in output_dose_name_format.split("}")
1020 if len(i) > 0
1021 ]
1023 # Now exclude those that aren't derived from the DICOM header
1024 dicom_header_tags = [i for i in all_naming_fields if i not in special_name_fields]

AttributeError: 'tuple' object has no attribute 'split'`

[rnfinnegan]

Thanks for your question!

RTPlanLabel is a DICOM tag found in the RTPLAN file, not the RTDOSE file.

It's probably going to be quite challenging to create a simple solution, because to get that information you would have to read in the RTPLAN file corresponding to the dose grid you are interested in to obtain that information, which isn't currently done with this code.

Here are some steps to go about implementing this:

1. Read in the DICOM RTDOSE file: ds = pydicom.dcmread("rtdose.dcm")
2. Get the referenced plan UID: ref_uid = ds["ReferencedRTPlanSequence"][0]["ReferencedSOPInstanceUID"] - note that this assumes the dose is from a single plan (i.e., not a plan sum)
3. Match the correct RTPLAN file, for example if you had multiple plan files you'd need to iterate through each of them and find the DICOM file where the SOPInstanceUID is the same as the ref_uid defined above
4. Read in the correct RTPLAN file and get the label: ds_plan_match = pydicom.dcmread("matching_rtplan.dcm"); plan_label = ds_plan_match["RTPlanLabel"]
5. Create the SimpleITK image for the dose grid: dose_grid = sitk.ReadImage("rtdose.dcm") * ds["DoseGridScaling"]
6. Save dose grid with desired filename: sitk.WriteImage(dose_grid, f"DOSE_{plan_label}.nii.gz")

Hope this helps 🚀

[pchlap]

@TmnGitHub, I'll also point out that we have developed an additional library, PyDicer, which handles converting large messy Radiotherapy DICOM datasets to Nifti in a standardized folder structure. Might be worth checking out to see if it would better suit your usecase:
Code: https://github.com/AustralianCancerDataNetwork/pydicer
Docs: https://australiancancerdatanetwork.github.io/pydicer/
Paper: https://www.sciencedirect.com/science/article/pii/S2352711024003807?via%3Dihub