Radiotherapy integration after transoral robotic surgery in head and neck cancer: a bibliometric and knowledge-mapping analysis of adjuvant treatment, de-escalation, functional preservation, and survival outcomes

Abstract

Background Transoral robotic surgery (TORS) is increasingly considered within multimodality treatment pathways for selected head and neck cancers, in which postoperative pathology may influence radiotherapy (RT), chemoradiotherapy (CRT), treatment de-intensification, and functional preservation. This study characterized the development, knowledge structure, and thematic evolution of research on RT integration after TORS. Methods Publications from 1 January 2010 to 1 August 2026 were retrieved from the Web of Science Core Collection, Scopus, and PubMed. After deduplication, exclusion of records lacking predefined fields required for downstream analyses, and manual topical screening, 215 publications were included. Bibliometric and knowledge-mapping analyses assessed publication and citation trends, contributors and collaboration, co-citation, bibliographic coupling, keyword and thematic evolution, and exploratory logistic life-cycle and latent Dirichlet allocation (LDA) models. Results The 215 publications were distributed across 54 sources, with an annual growth rate of 14.72%; 60.0% were published during 2020–2026. International co-authorship accounted for 13.49%. The United States was the leading corresponding-author country, accounting for 96 publications (44.7% of the total corpus). Citation and co-citation structures linked foundational TORS studies with evidence on HPV-associated prognosis and postoperative RT/CRT. Keyword and thematic analyses showed persistent prominence of TORS and oropharyngeal squamous cell carcinoma, alongside increasing representation of adjuvant treatment, HPV-associated disease, swallowing, and functional outcomes. Bibliographic coupling and factorial analysis indicated substantial thematic overlap rather than clearly separated research domains. Exploratory logistic modeling showed only moderate fit (R² = 0.778) and did not fully capture the empirical publication peak, while the five-topic LDA solution had limited bootstrap stability (0.3356). Conclusion The literature has shifted from predominantly procedure-centered investigation toward greater attention to risk-adapted postoperative management. Current research increasingly integrates treatment selection, RT dose and target-volume modification, and functional outcomes. Prospective multicenter studies with standardized postoperative risk definitions and long-term oncologic and functional assessment are needed to define the appropriate limits of treatment de-intensification.