(MRC DTP) Quantifying change in mammographic appearance using Artificial Intelligence

Breast cancer is the most common cancer among women, with over 54,000 new cases diagnosed each year in the UK. The NHS screening programme offers mammography every three years with the aim of detecting cancers at an early stage, however breast cancer risk varies between individuals, and some women have dense breasts in which mammography is less effective. Measurement of breast density is thus important for personalising breast screening. Women at high risk of developing the disease and those with dense breasts may benefit from more frequent screening and alternative or supplemental screening with different imaging techniques. Change in breast density is also a useful indication of whether a preventive intervention is succeeding in lowering an individual’s breast cancer risk [Cuzick 2011].
We have shown that expert radiologists are better than many computer-based algorithms at assessing whether a mammographic density pattern corresponds to high risk of cancer [Astley 2018]. We have also shown that this can be replicated by training a neural network on radiologists’ judgements, with the AI-based approach outperforming more conventional breast density methods [Ionescu 2019]. In this project we will focus on learning how breast appearance changes over time, using machine learning to determine whether we can distinguish between normal change and change due to developing abnormality. We will also apply the methods developed to assess change in response to preventive interventions. The work involves a number of technical challenges, the solutions of which are likely to be useful for a wide range of applications.
The student will work at the intersection of advanced image analysis, radiology, oncology and epidemiology. They will benefit from a multidisciplinary supervisory team, with expertise in breast radiology, oncology and cancer prevention, deep learning and imaging science. They will be actively involved in national and international collaborations involving density assessment, and will be encouraged to seek further opportunities to exploit and evaluate their methodology in a range of settings.

https://www.research.manchester.ac.uk/portal/sue.astley.html

Entry Requirements:
Applications are invited from UK/EU nationals only. Applicants must have obtained, or be about to obtain, at least an upper second class honours degree (or equivalent) in a relevant subject.