Room: Exhibit Hall
Purpose: For modern radiotherapy setups like adaptive radiotherapy it is often not possible to use real patient or phantom data to test crucial details of a treatment planning system or the delivery chain. The presented toolkit offers different options to test, generate and manipulate DICOM image,-plan and -structure data.
Methods: Based on IDL (Interactive Data Language, Harris Geospatial solutions) and Python scripts the toolkit was developed. Originally designed to test elastic registrations with manipulated images the functionality has been enlarged. DICOM image data sets can be rotated, shifted, resized, deformed and the contrast can be changed by simple equations. The resulting deformation vector fields are therefore well known. The re-deformed data set from the Treatment Planning System TPS (Raystation, Raysearch) and the vector field used there can be extracted and used to check the accuracy of the registration. The toolkit also allows working on DICOM RT plans where MU and field positions can be manipulated or EPID MLC sequences can be inserted to be used for recalculation in the TPS or for testing the QA system at the treatment device. For manipulated DICOM structure sets a conformity check was introduced to guaranty the import into the TPS.
Results: The toolbox has been proven to be helpful to test the adaptive workflow setup in our institution. Elastic registrations, EPID based recalculations of treatment plans and tests of the QA chain have been made possible. Modules can be combined and extended to specific needs.
Conclusion: The necessity of generating manipulated DICOM data to test an adaptive workflow and the related QA system lead to the presented toolbox. Although some of the topics can already be done by commercial software tools, the self-developed toolbox is much more flexible and can be adapted to special needs for example statistical evaluations.