Skip to content

UPDATE: We are delighted to announce that EPSRC have funded our CDT for five more years under the new name EPSRC Centre for Doctoral Training in Intelligent Integrated Imaging in Healthcare (i4health).

Visit our new website at:

08 Sep

Computational modelling for non-invasive cancer microstructure estimates with MRI

Current Projects

In spite of significant advances in diagnosis and treatment, cancer remains a major killer. Diagnosis is often inaccurate and subsequent treatment ineffective for far too many patients. Cancer diagnosis still relies on traditional histology for specific morphological information to determine cancer types and inform treatment. However, histology is an uncomfortable procedure where tissue is extracted with a biopsy needle that can have side-effects. Furthermore, because it targets only a small area, it can often miss the most important tumour regions and so require repetition. This project develops computational models to support non-invasive imaging techniques capable of estimating the same cellular characteristics as histology, such as cell size and density. Besides avoiding biopsies, these methods allow sampling of the whole organ avoiding false negatives from poor targeting. Specifically, the project will use diffusion Magnetic Resonance Imaging to reveal non-invasive cancer biomarkers. The models will be based on the VERDICT-MRI framework [1,2] that allows probing of tumour microstructure non-invasively. Successful application to prostate cancer has already been achieved [2]; the current project will use this framework for other tumours. Integration of the proposed method into clinical practice will lead to major patient benefits by providing more appropriate, targeted therapy while reducing the need for biopsies. Furthermore, precise identification of patients with good prognosis based on their microstructural imaging features may permit organ-preserving rather than radical surgery, considerably reducing patient side-effects while reducing healthcare costs. References:1)Panagiotaki, et al Cancer Research 74 (7), 1902-1912, 2014 2)Panagiotaki, et al Investigative Radiology 50 (4), 218-227, 2015

Cancer Imaging

Image Acquisition

Key Partners