2023/02/08 | Research | Artificial Intelligence
Bridging deep learning and clinical radiotherapy
The ARTORG Medical Image Analysis group (MIA) and the Department of Radio-Oncology of the Inselspital have developed an open-source, deep learning framework-independent Python package feasible for processing DICOM RT Structure Sets, in collaboration with Varian a Siemens Healthineers Company. “PyRaDiSe” (Python (Py), radiotherapy (Ra), DICOM-based (Di), auto-segmentation (Se)) goes beyond current 2D reconstruction, potentially supporting acceptance by healthcare professionals.
Web interface of a PyRaDiSe-based auto-segmentation solution deployed on the cloud using Docker containers
(https://doi.org/10.1016/j.cmpb.2023.107374)
Deep learning methodologies for medical imaging in radiotherapy are fast evolving. But auto-segmentation solutions rarely run in clinics due to the lack of open-source frameworks feasible for processing DICOM RT Structure Sets and pixelated contours in 2D reconstruction. The PyRaDiSe package closes the gap between data science and clinical radiotherapy by enabling deep learning segmentation models to be easily transferred into clinical research practice.
PyRaDiSe is open-source, flexible, and provides a framework for building auto-segmentation solutions feasible to operate directly on DICOM data. Its profound DICOM RT Structure Set conversion and processing capabilities also make it suitable for auto-segmentation-related tasks, such as dataset construction for deep learning model training. In a research project for organs-at-risk segmentation in brain tumor patients the package has successfully demonstrated its capabilities.
