Computer aided detection of masses in digital breast tomosynthesis: A review

Mammography is currently the most effective early detection tool for breast cancer screening. To provide a reliable and efficient second-reader to aid radiologists, recent research has been developing computer-aided detection (CAD) tools for mammography. Although these tools have shown promise in identifying calcifications, detecting masses has proven relatively more difficult primarily due to the presence of dense overlying tissue in a mammogram. Digital Breast Tomosynthesis (DBT) is a new three-dimensional breast imaging technique that has overcome the tissue superimposition problem for better lesion detection. It is possible that CAD when implemented with DBT performs better compared to mammogram images because of better visibility of masses. This paper presents a review on the existing CAD algorithms for the detection of masses in DBT. The performance of all these CAD tools in terms of sensitivity is also reviewed. Further, implementation of the CAD algorithm for DBT in hardware such as digital signal processor is proposed in this paper.