Deep learning for breast cancer diagnosis from histopathological images: classification and gene expression: review

Histopathology, the microscopic analysis of tissue to study the symptoms of the disease, is used to diagnose breast cancer. Breast cancer is specifically examined using tissue examination. The recent progress in deep learning has reinforced the potential of histopathological analysis by automating diagnostic processes. This review focuses on the integration of deep learning methods into histopathological analyses of breast cancer to categorise different types of cancer and predicting gene expression. We discuss the challenges associated with the use of deep learning models, such as the use of CNN variants to classify histopathological images into benign or malignant categories and specifically to identify different subtypes. We also observe the use of hybrid methods, as well as other approaches such as GNN, vision transformer etc. In addition, we investigate the capacity of deep learning to contribute to the interpretation of gene expression data that facilitates the prediction of breast cancer genes from the whole slide images, with the purpose of supporting personalised medicine. The majority of the selected studies are based on publicly available datasets and use techniques such as noise removal, image normalization, etc. Finally, our work looks to the future of deep learning in histopathology, exploring its role in therapeutic decision-making, predicting treatment outcomes and integrating histopathological and genetic data to increase results. The review can help researchers discover which deep learning techniques are most effective for a specific dataset and which features are significant for breast cancer detection.