dcmrtstruct2nii

DICOM RT-Struct to nii-mask. This is a naïve approach to rasterizing rt-struct to masks in nii format. If there's holes in your RT-Struct then this approach will most likely not work. The RT-Struct needs to be within the bounds of the slices of the original DICOM. Rasterization is done on a slice-by-slice basis, interpolation between slices is currently unsupported.

Rasterization

Currently the RT-Struct is treated as a polygon and is rasterized trough this function on a slice-by-slice basis. We'll hapilly add other rasterization algorithms (e.g. ones that can deal with holes in the polygon) if you point us to some code to do that or if you send in a pull request.

Interpolation

Interpolation of the mask between slices is currently unsupported. Send us an algorithm or a pull requests and we'll happly add it.

Input file format

The DICOM and RT-Struct inputs need to be unzipped in a directory. Currently this is the only way to read the input files.

CLI Tool

# install using pip and show tool help
pip install dcmrtstruct2nii
dcmrtstruct2nii --help

# list structures in DICOM RT Struct
dcmrtstruct2nii list -r /path/to/rtstruct/file

# convert help output
dcmrtstruct2nii convert --help

# convert DICOM RT Structs to .nii.gz masks
dcmrtstruct2nii convert -r /path/to/rtstruct/file.dcm -d /path/to/original/extracted/dicom -o /output/path

Python API

# install using pip and show tool help
pip install dcmrtstruct2nii

# lets test it
from dcmrtstruct2nii import dcmrtstruct2nii, list_rt_structs

print(list_rt_structs('/path/to/dicom/rtstruct/file.dcm'))

dcmrtstruct2nii('/path/to/dicom/rtstruct/file.dcm', '/path/to/original/extracted/dicom/files', '/output/path')

License and academic use

The program is licensed Apache license 2.0.

For academic use, use a presistent copy from .

Please cite:

Phil, T. (2020). Sikerdebaard/dcmrtstruct2nii: v1.0.19 (v1.0.19) [Computer software]. Zenodo. https://doi.org/10.5281/ZENODO.4037865