Systematic Review of Contemporary Breast Cancer Detection Techniques Using Machine Learning

Breast cancer, the main cause of mortality and disability in women globally, requires novel techniques to early detection, precise diagnosis, and successful treatment. In recent years, the combination of feature selection methodologies with CNN (convolutional neural network) techniques has emerged as a possible strategy to tackling the intricate nature of breast cancer analysis. This systematic review synthesizes the research at the intersection of feature selection and CNN techniques in breast cancer analysis, aiming to provide insights into the synergistic potential of these methodologies in advancing precision oncology. Through a rigorous search and selection process, articles extracted from peer reviewed journals and proceedings of A grade conferences. The review elucidates the diverse modalities of feature selection techniques employed in conjunction with CNN architectures, encompassing filter-based methods, wrapper approaches, embedded feature selection, and hybrid strategies. Presented work showcased the literature, using important parameters like classification of histopathological images, molecular subtype prediction, and treatment response assessment and unveils the transformative potential of feature selection and CNN techniques in augmenting the accuracy, interpretability, and generalizability of breast cancer analysis. Moreover, scrutinize the methodological rigor and translational relevance of the reviewed studies, delineating key challenges, limitations, and avenues for future research.