Applications of artificial intelligence in endobronchial ultra-sound for lung cancer diagnosis and staging: a scoping review

This scoping review evaluated the use of artificial intelligence (AI) in endobronchial ultrasound (EBUS) for lung cancer diagnosis and staging. A total of 26 studies published between 2015 and 2026 were included, with most employing deep learning–based models. The main clinical application was diagnostic classification of malignancy, followed by segmentation, cytological analysis, anatomical navigation, and multimodal predictive models. Most studies used EBUS-derived images or videos from both Convex Probe-EBUS and Radial Probe-EBUS procedures. Although findings highlight the growing potential of AI to improve diagnostic accuracy and image analysis, current evidence remains limited by retrospective designs, small sample sizes, and single-center studies.