DEEP LEARNING-BASED MAMMOGRAPHY ANALYSIS FOR BREAST CANCER: A SYSTEMATIC STUDY

Abstract

Breast cancer remains one of the leading causes of cancer-related mortality among women worldwide, making early and accurate diagnosis essential for improving patient survival. Mammography is considered the gold-standard imaging modality for breast cancer screening; however, traditional interpretation is highly dependent on radiologist expertise and is prone to false-positive and false-negative outcomes. Recent advancements in deep learning have significantly enhanced automated mammographic analysis through architectures such as Convolutional Neural Networks (CNNs), DenseNet, EfficientNet, U-Net, and Vision Transformers (ViTs). This systematic review analyzes 50 peer-reviewed studies published between 2013 and 2024, following the PRISMA 2020 guidelines, to evaluate the effectiveness of deep learning techniques in mammographic breast cancer detection. The review examines datasets, architectures, evaluation metrics, and clinical applicability while identifying major research challenges, including model interpretability, dataset bias, cross-vendor generalization, and privacy constraints. Furthermore, future research directions involving explainable AI, federated learning, and standardized benchmarking frameworks are discussed.