The challenge
High breast density makes screening difficult
Breast cancer is the most common cancer affecting women in Australia. According to the Australian Institute of Health and Welfare one in eight women will be diagnosed with breast cancer in their lifetime. With early detection, a greater range of treatment options are available and chances of survival increase.
Regular population-level screening, especially for women aged 50 – 74 years, can detect breast even cancer before symptoms appear. A specialised X-ray image called a mammogram is routinely used in the screening process.
Women with high breast tissue density are at higher risk of breast cancer. However, it is more difficult to detect cancers using mammograms, and larger tumours are found at diagnosis in women with high breast density.
To help detect cancers earlier, Australian researchers and clinical services need fast, automated and practical tools to reliably measure breast density from mammograms.
Our response
Software to automatically assess breast density
Working with the University of Melbourne we developed AutoDensity, a fully automated breast density measurement software package to help identify women at higher risk of breast cancer and women who might benefit from alternative screening strategies.
AutoDensity is capable of analysing digital and digitised mammogram images from any mammography machine in use across Australia. It automatically identifies the breast area in an image and demarcates and measures dense regions inside the breast. Dense tissue appears as relatively bright components of the breast image. By applying a range of image processing algorithms, the software finds an optimal threshold between darker and whiter regions in the breast.
AutoDensity is fast and simple to use, is fully automated, and provides a friendly and intuitive user experience. It can be easily installed on Windows, Linux, or Mac computers, and does not require a trained radiographer to operate.
Our software works across image formats, mammogram machines brands and image formats. This makes it possible to assess breast density in a standardised way from different clinics, analyse older mammogram films that have been digitised, track individual woman over time, and study the role of breast density in long-term breast cancer risk.
The results
Improved screening for breast cancer risk
We have validated the effectiveness of AutoDensity to assess breast cancer risk and screening clinical outcomes on a large pool digitised film mammograms from more than 6,000 Australian women. Our study found that AutoDensity's automated measurements of breast density from mammograms predict breast cancer risk and tumour size at diagnosis similarly to a labour-intensive method commonly used in research settings.
AutoDensity is currently undergoing evaluations with clinical and population health research groups in Australia and internationally. Together we are working to create a powerful tool to improve breast cancer outcomes for Australian women.