This technique is currently being developed in both Ovarian and Renal cancer.
Tissue-specific and interpretable sub-segmentation of whole tumour burden on CT images by unsupervised fuzzy clustering, Rundo, L. et al., Computers in Biology and Medicine, 2020
Summary: By using novel automated segmentation methods, novel radiomics extraction and data integration we are working on predicting response to chemotherapy in ovarian cancer. We have started to optimize automated segmentation of Computer Tomography images building on our work using a U-Net version. By using a novel two-stage framework for tissue specific solid tumour segmentation based on unsupervised fuzzy clustering, we can take the prior domain knowledge of the typical sub-regions densities into account and are able to deliver interpretable results for the clinicians while keeping them optimised for radiomics purposes.
Further information can be found on the Radiogenomics and Quantitative Imaging Group page.
Further information on the All-in-one Cancer Imaging project funded by the Wellcome Trust Innovator Award.
The Mark Foundation Institute for Integrated Cancer Medicine (MFICM) at the University of Cambridge aims to revolutionise cancer care by affecting patients along their treatment pathway.