Artificial intelligence for automated lung nodule detection in chest computed tomography scans

Automated detection of pulmonary nodules on computed tomography has evolved from rule-based image processing algorithms to sophisticated artificial intelligence (AI) systems capable of nodule localisation, volumetric measurement and malignancy risk stratification. Four landmark papers published between 1994 and 1999 established the computational foundations of this field, introducing grey-level thresholding, morphological feature extraction, neural networks and three-dimensional analysis as core detection principles. We provide an overview of these foundational studies and trace the subsequent evolution of AI-based nodule detection, from early computer-aided detection systems to modern deep learning frameworks evaluated in the context of lung cancer screening programmes.