Breast Density: Developing Imaging Biomarkers for Breast Cancer

Mammography is widely used for screening the asymptomatic population for early signs of cancer, and the advent of digital imaging has opened the door to the development of automated techniques, both for detecting cancer and identifying women at increased risk.

We have shown that mammographic density, which describes the quantity of radiodense and fatty tissues in a woman’s breasts, does not fully encapsulate the information in mammograms associated with breast cancer risk, and have developed approaches using Artificial Intelligence which are predictive of the development of cancer and the efficacy of mammography as a screening tool. Our ongoing work involves extending the approach to mammograms which may be excluded by other methods and to processed images from a range of manufacturers.

We have also used AI to develop a method for measuring breast density in mammograms taken at a tenth of the usual x-ray dose. These cannot be examined by clinical experts because of poor image quality, but we have demonstrated that we can make reliable breast density measurements from them. This approach is being used in a clinical trial which is investigating predicting the risk of future breast cancer in young women. The method could also be used for stratification in risk-adapted screening.

Our methods have been used in clinical trials to predict risk and assess the potential impact of preventive interventions and treatment on future breast cancer risk. We are also investigating the relationship of imaging biomarkers with other risk factors for breast cancer, and methods for avoiding bias in AI.

Lay summary: We are developing methods that can predict the development of cancer, tell us accurately whether treatment to reduce risk is working, and give us an indication of the most appropriate methods of imaging an individual woman’s breasts. 

Evaluation of New Technologies

We have undertaken systematic evaluations of technologies including Digital Breast Tomosynthesis (DBT), Contrast Enhanced Mammography (CEM), Electrical Impedance (EI), Breast Density measures and Computer Aided Detection (CAD), and further evaluations are in progress.  We assess not only the stand-alone performance of technologies, but their impact on the clinical experts using them. A particular interest at present is Explainable AI. 

Lay summary: All new technologies require systematic evaluation to determine the circumstances in which they can potentially assist radiologists, and those in which performance may be degraded by using them. We are experienced in making these assessments in a way which is fair and robust.

Other Areas of Research

Understanding Reader Performance: We are interested in the ways in which expert clinical readers assess images, particularly with respect to breast density. We are conducting experiments to assess reader agreement and to identify potential reasons why readers disagree and investigating whether we can detect these cases automatically using AI.

Detection of Mammographic Abnormalities: We have investigated the use of Bayesian statistics to improve specificity in the detection of microcalcifications, which are one of the earliest signs of cancer, and found that combining different cues was effective in improving detection. The detection of asymmetry is is more difficult, as the breasts are variable in appearance and differ naturally. The technique we have developed is based on identifying anatomically similar regions using the transportation algorithm to measure similarity.  We have also investigated the detection of spiculated masses and distortion; one of the key ideas behind this approach was model-based classification of linear structures in the digitised images. More recently we have started to develop menthods for predicting subtypes of breast cancer using Artificial Intelligence.

Lesion Modeling:  We have developed a method for generating realistic synthetic masses by statistically modelling spiculated breast lesions, many of which are indistinguishable from real lesions by consultant breast radiologists. We have also looked at the relationship of the lesions to normal breast tissue. Our longer term aim is to consider how these might be used in personalised training.

DME Lead for Intercalated Degrees

DME Associate Lead for Personal Excellence Pathway (APEP Lead)

Prof Gareth Evans

Prof Tony Howell

Dr Sacha Howell

Dr Elaine Harkness

Prof Bob Nishikawa (University of Pittsburgh, USA)

Prof Juhun Lee (University of Pittsburgh, USA)

Prof Anne Martel (Sunnybrook Research Institute, Canada)