Artificial Intelligence in the evaluation of pulmonary nodules: performance, workflow integration, and lessons for the Brazilian Unified Health System (SUS)

This study analyzes the advances, challenges, and clinical applications of artificial intelligence (AI) for the early detection of lung cancer through imaging, as a support tool for radiologists. Based on a narrative literature review conducted in databases such as PubMed, Scopus, Web of Science, and Google Scholar, AI was found to have progressed significantly in the field, assisting in the detection of pulmonary nodules. Deep learning models, such as convolutional neural networks (CNNs), have demonstrated the ability to increase diagnostic sensitivity, reduce interpretation time, minimize reading errors, and support more accurate clinical decisions compared to isolated reading, albeit at the cost of a higher false-positive rate. Evidence highlights the importance of AI in contexts with a shortage of specialized professionals, such as public health systems often marked by resource limitations and structural neglect, like the Brazilian Unified Health System (SUS). Despite these advances, challenges remain, including the need for large-scale clinical validation, integration into hospital systems, ethical and legal regulation, and greater acceptance by professionals. AI thus presents itself as a tool to support radiologists by increasing sensitivity and reducing reading time, but its incorporation depends on overcoming technical, operational, and ethical barriers, as well as validation in different healthcare contexts.