Radiology meets AI: Evolving trends in national vs. international conferences (2019–2024)

Abstract

This study examines the prevalence and evolution of artificial intelligence (AI)-related topics in radiology conferences from 2019 to 2024, comparing national and international congresses to trends in the literature. The AI-related sessions of national radiology congresses, including TURKRAD, RÖKO, JRC, KCR, RANZCR, and BIR, were analyzed alongside international conferences such as RSNA, ECR, and EuSoMII. Additionally, interventional radiology congresses (CIRSE, PAIRS, and TSIR) were included. The programs of congresses over the last six years were systematically reviewed by two experienced radiologists. Sessions were categorized based on type (lectures and workshops) and keywords. Concurrently, AI-related radiology research trends were analyzed using Web of Science (WOS) data, focusing on correlations between conference discussions and scientific publications. The analysis revealed a significant increase in AI-related sessions in radiology congresses over time, with RSNA leading in both the number and diversity of AI topics. Among national congresses, JRC hosted the most AI sessions, while interventional radiology meetings showed comparatively fewer AI discussions. Workshops were more prevalent in international conferences, likely due to industry support. Keyword analysis indicated strong overlaps between conference themes and WOS publications, with topics such as neuroradiology and breast cancer frequently appearing. However, emerging AI subfields, such as large language models, showed weaker, likely due to their rapid development and ease of publication. AI’s integration into radiology is expanding rapidly, with conferences serving as a barometer for trends. Understanding the alignment between congress and literature can help guide future research priorities, industry investments, and educational initiatives in radiology. • AI sessions at radiology conferences rose sharply from 2019–2024, led by RSNA and ECR in volume and topic variety. • Conference AI topics closely mirror scientific literature, especially in neuro, breast imaging, and radiomics. • Structured training, ethics, and regulation in AI remain underrepresented, limiting broader clinical adoption.

Ähnliche Arbeiten