Radiomics and heart position in thoracic cancer
Research type
Research Study
Full title
Application of radiomics, and incorporation of treatment planning and treatment data including images, to thoracic cancer.
IRAS ID
280961
Contact name
Paul Gordon Roxby
Contact email
Sponsor organisation
Sheffield Teaching Hospitals NHS Foundation Trust
Duration of Study in the UK
1 years, 5 months, 9 days
Research summary
Radiomics concerns medical images and the information we get from them. Usually, little numerical information (other than perhaps the maximum size of tumours) is taken from medical images used in cancer. It is their appearance to a radiologist which matters when diagnosing cancer.
Medical images are also used by oncologists for radiotherapy treatment. The oncologist will create tumour volumes from the images using drawing tools in a software package. This makes a 3D shape which will (once expanded to allow for movement and microscopic tumour extension) receive a high radiation dose.
Radiomics produces several numerical scores from the images by looking at how uniform the shades of grey are within the tumour volume drawn by the oncologist, how bright the region is and what the texture of the tumour is like on the images (e.g. how much fine detail it has in it). It also gives scores to the shape of the tumour (e.g. how round it is, and how spiky).
Several studies have linked some of the scores produced by radiomics with overall survival of patients with lung cancer after radiotherapy. However there is concern that some of these results are due to chance, or are too dependent on the model and settings of the CT scanner used, and cannot be reproduced when similar groups of patients are analysed at a different centre. We will attempt to reproduce several such results to see if this is a real problem.
There is also concern that a higher than expected radiation dose to the heart can result in shorter survival for patients. We will look at this by comparing where the heart is during radiotherapy treatment with where it was expected to be from the pre-treatment CT scan used to plan the treatment.
REC name
N/A
REC reference
N/A