Artificial Intelligence and Deep Learning Architectures for Automated Bone Tumor Detection and Classification: A Comprehensive Survey

Bone malignancies represent critical medical conditions requiring rapid and precise diagnostic intervention. Traditional diagnostic methodologies, heavily dependent on manual radiological interpretation, face significant limitations including inter-observer variability, time consumption, and delayed detection rates. Recent advances in Artificial Intelligence (AI) and Deep Learning (DL) technologies have demonstrated transformative potential in automating tumor detection workflows. This comprehensive survey examines contemporary research leveraging sophisticated AI frameworks including Convolutional Neural Networks (CNN), transfer learning paradigms, and optimization-driven hybrid architectures for medical imaging applications in osseous tumor diagnosis. Through systematic analysis of four representative studies, we evaluate performance metrics, diagnostic reliability, and clinical applicability of AI-driven approaches across multiple imaging modalities. Our findings demonstrate that deep learning methodologies consistently surpass conventional manual interpretation techniques, achieving accuracies of 84-95% with superior sensitivity and specificity. The paper culminates with strategic recommendations for advancing model interpretability, expanding dataset heterogeneity, and facilitating real-world clinical integration.