Why did we fund this fellowship?
This award aims to reduce the number of animals needed in longitudinal CT imaging studies by developing a standardised radiomics workflow for preclinical imaging analysis.
CT imaging is a non-invasive method used to monitor disease progression and response to drugs in animal studies and in the clinic. For the former, the technique enables longitudinal studies to be performed reducing the number of animals needed in a study overall. However, in cancer studies the genetic properties of the tumour also need to be validated, for example if the tumour is homogenous or heterogeneous, which requires invasive biopsies or killing animals at various timepoints for tissue samples, increasing the number of animals required. Radiomics is an advancing field in medical imaging to enhance data from existing imaging techniques. Using computer modelling, disease characteristics not able to be discerned by eye can be determined, for example by analysing spatial distribution of signal intensities, which enables clinicians to make more informed decisions. For animal research, radiomics has the potential to avoid the need to use animals for tissue phenotyping.
In this Fellowship, Dr Kathryn Brown will develop computational models to extract radiomics data from small animal CT images of lung cancer. She will use the models to establish relationships between tissue characteristics, phenotype and imaging features to maximise the amount of data that can be gleaned from imaging studies. Kathryn will investigate predictive biological features and radiotherapy responses of tumours in multiple mouse strains, reducing the number of animals needed for longitudinal studies. She will develop skills in phantom imaging, bioluminescence imaging and statistical methods.
Imaging is an essential tool in animal studies that allows investigators to look inside the bodies of mice to assess the size, shape and location of different tissues. Imaging can be used to monitor how different disease progress and to show how different tissues such as tumours respond to experimental treatments. In humans, new techniques are being developed to analyse routinely collected computed tomography (CT) scans. These images are then analysed by a computer to extract hundreds of features, termed 'radiomics features', which have the potential to uncover different disease characteristics that cannot be detected by the naked eye. This technique can complement traditional diagnosis methods and may help clinicians to make more informed personalised treatment choices and could replace the need for traditional invasive biopsies.
In this project, we will develop a similar approach for radiomics in mouse tissues. We will adapt existing methods used in the patients, to analyse CT scans from different mouse tissues and tumours, and setup a standard procedure for this process. Using this method, we will determine the relationship between the imaging features of different mouse tissues with the particular biological characteristics to develop a new way to monitor disease progression and response to treatment. This approach will deliver extra data from CT scans on the features of tissues, reducing the need for large animal numbers and invasive procedures.