Tailored software for post-radiotherapy lung radiomics analysis

Abstract Introduction Radiotherapy is a crucial method of treating lung cancer. However, potential toxicity such as radiation-induced lung injury (RILI) must be considered. In order to quantitatively describe different forms of RILI, a next generation of a previously discussed system has been developed to analyse the subsequent series of CT scans performed by patients during radiation therapy and post-RT follow-up. Materials and methods A 3D CT scan registration module has been developed as part of the system. The analysis management module has been expanded compared to previous versions of the system. The environment has been prepared and the system has been deployed. In order to determine the effectiveness of the system, 1598 analyses, based on automated image pre-processing (registration and segmentation) and first-order delta-radiomics features, have been performed on a data set from 50 Patients. Results The efficiency of the execution of the analysis was evaluated in terms of time, including a breakdown into individual stages of the pipeline. The results of the analyses were visualised and exported to a CSV file, then evaluated from a clinical perspective. Conclusions The system allows, thanks to the use of a 3D registration algorithm, to analyse changes more accurately than before, and thanks to the use of a more flexible architecture, to introduce subsequent registration and calculation algorithms more easily. The current main research and development areas of the system include the extension of the analytical capabilities of the system in the field of radiomics and the expansion of the number of image analysis algorithms used within the system.