Aplicações e desafios da inteligência artificial na radiologia: revisão narrativa da literatura

Abstract

Artificial Intelligence (AI) has become an essential tool in Radiology, standing out for its ability to optimize workflows, increase diagnostic accuracy, and support clinical decision-making. This study consists of a Narrative Literature Review, conducted through structured searches in the CAPES Periodicals Portal, applying terms related to AI and radiological diagnosis. After filtering for open access, publication years, and peer review, 20 articles were selected for qualitative analysis. The results indicate that AI performs well in various clinical applications, such as the detection of ophthalmological pathologies and cardiopulmonary analysis, achieving sensitivity and specificity comparable to those of radiologists. However, the literature highlights significant challenges, including educational gaps, infrastructure limitations, algorithmic bias, and ethical issues related to data protection. Human-AI integration proved to be the most efficient model, combining computational speed with clinical judgment. In conclusion, AI has the potential to transform radiological practice, but its safe implementation depends on continuous professional training, adequate regulations, and rigorous human supervision, ensuring more accurate, ethical, and reliable diagnoses.