FSE 5: Localising breast cancer risk in mammograms

Faculty: Faculty of Science and Engineering

Supervisors: Dr Faraz Janan; Dr Louise Wilkinson (external); Dr Silvia Cirstea

Location: Cambridge

The interview for this project is expected to take place on Tuesday 25 April.

This project aims to develop a breast cancer diagnosis support system by localising the future risk of developing cancer, in particular associated with mammographic focal densities.

By analysing time series data with new mammographic image analysis framework, it aims to propose a CAD system that can flag women likely to develop breast cancer in near future.

The system should also be able to suggest among the bilateral breast, as well as the location/quadrant where the risk of developing cancer is higher.

Applications are invited for a PhD studentship to work on a project using a range of advanced AI and image analysis techniques to study mammographic images in the context of breast cancer risk.

The project will require a full-time research commitment and will be based in School of Computing and Information Science on ARU's Cambridge campus.

It will combine mammographic density quantification with an AI based classification and pattern recognition framework (desirably deep learning methods).

It would evaluate the methods developed on mammograms acquired from the Optimam that have negative priors, CC and MLO views available for both breasts and depict biopsy-proven cancers, as well as normal cases. This would help us to assess the effectiveness of a CAD system and its suitability in a clinical set up.

The three-year project will be carried out in close collaboration with scientists and breast radiologists at Oxford. You'll be required to apply, develop and program algorithms in the area of computer vision and machine learning, while applied to mammographic images -including but not limited to x-ray and digital breast tomosynthesis (DBT).

You'll learn mammographic image analysis, clinical aspects of breast cancer detection and diagnostics, and deep learning methods applied to medical imaging.

You'll have a strong academic track record, with a 2:1 or higher degree in computer science, mathematics, biomedical engineering, electrical (computer) engineering, or equivalent if your degree was awarded by a university outside of the UK. You'll be able to demonstrate excellent performance in a relevant postgraduate degree.

You'll be expected expected to demonstrate expertise of coding in Matlab and Python, with good knowledge of image processing techniques. A prior working experience of deep learning methods is desirable. You must be willing to work in close collaboration with clinical radiologists.

If you would like to discuss this research project prior to application, please contact faraz.janan@aru.ac.uk

Apply online by 19 March 2023

Funding notes

This successful applicant for this project will receive a Vice Chancellor’s PhD Scholarship which covers Home tuition fees and provides a UKRI equivalent minimum annual stipend for three years. For 2022/3 this was £17,688 per year. The award is subject to the successful candidate meeting the scholarship terms and conditions. Please note that the University asserts the right to claim any intellectual property generated by research it funds.

Download the full terms and conditions.