Artificial Intelligence in Radiology: Global Research Trends and Insights (2000-2025)

Abstract

Background and Objectives: Artificial Intelligence (AI) applications in radiology are crucial for assisting radiologists in detecting abnormal findings in imaging examinations and reaching a diagnosis. Hence, this study conducted a bibliometric analysis to uncover global research trends on AI applications in radiology,

Methodology: An electronic search of the Scopus database was conducted on May 02, 2025, using specific keywords to retrieve documents on AI applications in radiology. The search specifically targeted documents published over 26 years, from January 2000 to May 2025. The collected data were downloaded as a plain text file and analyzed using RStudio 2024.12.1, Bibliometrix (biblioshiny), and the Visualization of Similarities (VOS) viewer software (version 1.6.20). The "Article" type documents published in English were included.

Results: 12,139 research documents on AI applications in radiology were published by global researchers, with a peak publication count in 2024. The University of California, United States, is the leading contributor with 788 documents. Saba L. and Suri JS had 37 publications. China was the most productive country with 3,443 research documents, while the United States published 3,145 documents but showed the highest citation count (n=98,928). The strongest collaboration was found between China and the United States, with 415 research documents.

Conclusion: The publication of AI applications in radiology has improved extensively since 2018 and is expected to peak in 2024. Global researchers can further progress their affinity for AI in radiology by producing additional high-quality research documents in the future. Further, expanding international research collaboration networks across various countries is warranted.