Radiological trends in convolutional neural networks for breast cancer diagnosis

Breast cancer remains a leading cause of cancer-related mortality worldwide, making early and accurate diagnosis crucial for patient outcomes. Future diagnosis may employ convolutional neural networks (CNNs), which have established themselves as powerful multi-layered artificial intelligence (AI) tools for computer vision tasks, with growing applications in breast cancer detection, diagnosis and classification. To provide insight into this field’s intellectual, social, and conceptual knowledge structures, we conducted a bibliometric review of its 100 most-cited articles. The review looked at articles from January 1, 1995 to August 23, 2024. Our network analyses encourage increased inter-country collaboration. Thematic mapping highlights the increasing role of CNNs as foundational components in present and future AI applications. Multiple correspondence analyses track progress in diagnostic accuracy, system performance, and advanced classification techniques. Study design analyses suggest a need for future CNN research to be benchmarked against human readers and foster closer collaboration between technical and clinical researchers. In this bibliometric analysis, we summarise key contributions, examine emerging research trends, and provide an overview of the evolving landscape of CNN applications in breast cancer diagnostics. • Considerable interest in convolutional neural networks for breast cancer diagnosis • China and USA are major contributors, but further global collaboration encouraged • More research needs to be benchmarked against human readers • More metrics needed when evaluating predictive models trained on imbalanced datasets • Models could integrate multiple modalities to enhance diagnostic precision