Radiomics in Head and Neck Cancer: A Web of Science-Based Bibliometric and Visualized Analysis

This study is the first to employ bibliometric analysis methods to conduct a visual analysis of the literature on the application of radiomics in the field of head and neck cancer from 2014 to mid-2025. Based on 428 documents obtained from the Web of Science database, the distribution of national (regional) collaborative networks, institutions, journals and authors' contributions, as well as the evolution of keywords were visualized and analyzed using VOSviewer and CiteSpace software. 428 documents were included in the study, and the overall trend of the number of publications showed a rapid increase after 2017, and gradually stabilized after 2021. China (118 articles) and the United States (108 articles) were at the core of the field in terms of the number of publications. However, in terms of international cooperation, China lacked compared with Europe and the United States. Maastricht University (Maastricht University, Netherlands) was the institution with the most publications; Cancers (35), Scientific Reports (32) were the journals with high impact in the field; Forghani Reza (12) was the author with the most publications. In the last 10 years, the keywords "radiomics", "head and neck cancer", and "machine learning" have appeared more frequently. The research focus has transitioned from tumor heterogeneity characterization to local tumor control optimization, demonstrating tangible progress toward clinical implementation. To realize the full potential of these advances, two critical requirements must be addressed. First, extensive validation through multicenter prospective trials with large sample sizes is essential. Second, comprehensive standardization of radiomics protocols across all stages including image acquisition, feature extraction, and analytical processing must be established to ensure reproducibility and facilitate clinical adoption.