Literature Review:Machine Learning in Cancer Detection

Abstract

Cancer continues to be one of the leading causes of death worldwide, creating an urgent need for accurate, easy-to-use diagnostic tools. Machine learning (ML) has emerged as a transformative force in cancer detection, capitalizing on advances in medical imaging and computational power to improve diagnostic accuracy, efficiency, and scalability. This literature review synthesizes recent (2018-2023) research on the application of ML in cancer detection, focusing on major cancer types, including breast, lung, and skin cancers. We first provide an overview of basic ML techniques such as convolutional neural networks (CNNs) and their adaptation to medical imaging challenges, including limited data and class imbalance. Next, we analyze case studies of various imaging modalities (e.g., mammography, CT scanning, dermoscopy), highlighting state-of-the-art models such as ResNet and U-Net, which achieve radiologist-level performance in tasks such as lung nodule classification and melanoma detection. Despite promising results, key challenges remain: dataset heterogeneity, model interpretability, and integration with clinical workflows. We further discuss future directions such as privacy-preserving joint learning and multimodal fusion of imaging and genomic data. This review highlights the potential of ML to democratize cancer diagnosis while calling for interdisciplinary collaboration to address real-world implementation barriers.