- To validate (or otherwise) the moderate ability of radiomics to predict outcomes in non small cell lung cancer using radiotherapy CT planning scans.
- To establish a radiomics pipeline which can be applied to other tumour sites and images in future.
- If time permits, to find ways to incorporate dose and structure data from treatment planning, and perhaps additional data from on-treatment imaging, into such a predictive pipeline. In particular the comparison of planned and delivered dose to the heart is of interest.
A cohort of data including outcomes is available from previous patients. It is planned to apply radiomics techniques to the treatment planning scans of this cohort. This will be a replication study of previously published radiomics work, as there a questions about the reproducibility of radiomics work. The radiomics data may be supplemented by dose and structure information from treatment plans and information from cone beam CT scans.
None as yet – work in progress.