Researchers: Nathan Sjoquist, Graham Treece
This project involves the development of novel techniques and algorithms to better detect and visualize metastatic cancer in bone within computed tomography (CT) data sets. Specifically, using features of symmetric and asymmetric bone regions to detect cancer.
To achieve this, we will need to automate accurate 3D segmentation of bone regions within CT data, following which corresponding left and right sides will be examined by mirroring one to match the other, using locally adaptive registration. Once the bone is matched, the relative bone densities and asymmetry will be visualised, using novel methods which are easily understood by the clinician. Local differences in bone may be the result of osteolytic or osteoblastic metastases due to cancer, both of which can change the bone material and structure.