Advances of Deep learning in Breast Cancer Modeling

Deep learning (DL) has recently gained popularity in forecasting, detecting, categorizing, and diagnosing for breast cancer with promising results. Developing a review paper to assess the efficacy of DL methods in this context is essential. We've established a standardized database initially containing fundamental publications for methodical reviews. The primary objective of this review is to systematically present the current state-of-the-art, using an updated PRISMA guidelines to better review and evaluate the DL's effectiveness in breast cancer applications. The research follows three main stages: data collection, data analysis, and summarization of initial outcomes. The results highlight accuracy as the prevailing and comprehensive metric used in evaluating DL tools across varied breast cancer applications. Convolutional neural networks (CNNs) have found to have widespread utility, notably surpassing other DL methods. In contrast, collaborative teamwork and employing advanced DL techniques yield optimal performance.