Enhancing Accuracy in Breast Density Assessment using Deep Learning

Kvak, D., Biroš, M., Hrubý, R., Dandár, J., Janů, E., & Atakhanova, A. (2024). Human-LevelComputer-Aided Approach for BI-RADS Breast Density Classification: Multi-Reader, Multi-Centric Study. European Congress of Radiology 2024, Vienna.

Abstract:

Breast cancer is the most common cancer type in women and is the main cause of cancer-related death in women population in the world. The diagnosis of abnormalities in mammography studies is considered a difficult task because of variable sizes and shapes of lesions, difficult differentiation of lesion borders and some extremely small lesions. A pivotal factor affecting the detection of breast cancer is the density of the breast tissue, which can reduce visibility of tumors and is independently associated with an increased risk of breast cancer. The Breast Imaging Reporting and Data System (BI-RADS) created by the American College of Radiology offers a standardized classification system for breast density assessment. However, this system has faced challenges due to the substantial interobserver variability among radiologists, introducing uncertainty and inconsistency in evaluation.