Breast cancer (BC) is the most common cancer in women globally. Improving survival rates and treatment results requires early and precise identification. This research study analyzes the use of deep learning methods on tomosynthesis images to enhance the efficiency and diagnostic accuracy of breast cancer screening. This study is based on a comprehensive review of 25 research publications published between 2018 and 2024. The study used different types of deep learning (DL) techniques, including convolutional neural networks (CNNs), Deep CNNs, RNN, and faster region-based convolutional neural network (Faster-RCNN). The study's results showed that breast cancer detection and classification accuracy has significantly improved over time, with some studies reporting accuracy levels of classification, detection, and segmentation as 95%, 96%, and 99%, respectively. The paper analyzes current breast cancer detection, classification, and segmentation advances. However, more research is necessary to support these techniques in medical practices and increase their efficiency.