Scientific Mapping of Deep Learning Research in Breast Cancer

Deep learning is reshaping breast cancer research across imaging, pathology, and outcome prediction; however, the structure and recent evolution of this literature have not been comprehensively mapped. This bibliometric review analysed Web of Science Core Collection records published from 2022 to 2026 using CiteSpace 6.4.R1. The keyword network comprised 265 nodes and 1,069 links and resolved into 11 thematic clusters, with mammography occupying a central position. Emerging topics included antibody-drug conjugates, large language models, and prediction of treatment response. The temporal analyses indicate a gradual shift from technical performance evaluation toward multimodal modelling and precision oncology. Future progress will depend on externally validated systems that perform reliably across institutions and patient populations.