Computer technology is key to detecting early signs of breast cancer

Software developed at the Schulich School of Engineering in collaboration with the Alberta Breast Cancer Screening Program has been shown to flag subtle signs of breast cancer sooner than ever before.

Raj Rangayyan, Shantanu Banik and Leo Desautels combined advanced computer technology with mathematical algorithms to process images of mammograms and digitally analyze them. They are the first experts to focus on architectural distortion for early detection of breast cancer using mammograms taken before a clinical diagnosis.

The system focuses on detecting architectural distortion, which refers to specific patterns in breast tissue that are often missed during routine screening. These patterns can be precursors to tumours and they may appear more than a year before more obvious signs of breast cancer such as lumps in breast tissue.

"This study demonstrates that it is possible to develop sophisticated computational methods to detect early signs of breast cancer," explains Raj Rangayyan, professor in electrical and computer engineering at the Schulich School of Engineering and a collaborator with the Department of Radiology at the University of Calgary's Faculty of Medicine. "Until now, computer-aided diagnosis systems have focused mostly on finding lumps and calcification in breast tissue."

Because architectural distortion usually appears before tumours, detecting it could lead to earlier diagnosis of breast cancer and improve the survival rate.

Before this technology can become a regular part of breast cancer screening, experiments with larger sample sizes are needed. The number of false alarms needs to be reduced by refining algorithms and mathematical characterization techniques. The new method also must be tested on the latest version of digital images that are currently used in mammography. As computer technology advances and image quality improves, research results are expected to be even better in the future.

This research was recognized as making a significant contribution to cancer research by Canada's Institute for Cancer Research, which awarded a Publication Prize to the research team.