ESTRO 2025 - Abstract Book

S2416

Interdisciplinary – Other

ESTRO 2025

Geographic and gender diversity among stakeholders shaping health-AI guidelines is vital to establishing quality standards that address varied contexts and reduce risks of perpetuating biases. This bibliometric analysis examines disparities that may impact equitable cancer care delivery worldwide. Material/Methods A PRISMA-guided systematic review assessed English-language reporting/appraisal guidelines for AI research in healthcare. MEDLINE, Scopus, EMBASE, and EQUATOR Network were searched. Data were extracted by multiple reviewers and managed through Microsoft Excel. Variables included year of publication, authorship (gender, country, seniority), participating countries, and funding sources. World Bank income classifications (HIC, UMIC, LMIC, LIC) contextualized research contributions. Results Of 35 articles meeting inclusion criteria, 32 (91.4%) were published from 2020-2024 (Figure1). The mean number of authors per article was 11 (SD=7.85). In total, there were 386 individual contributors, and 4 frameworks involved collaborative research-group authorship. First-authorship was balanced across gender (56% vs 44%) , while senior authorship distribution was predominantly male (74% vs 26%). The UK and USA led first-authorship (24.3% each), followed by Canada, the Netherlands and China (10.8% each). For senior authorship, the USA kept its lead at 28.6%, followed by the UK (20.0%), the Netherlands (14.3%) and Canada (11.4%). Of 106 country participations identified, 91.5% were from HICs, 6.6% UMICs, and 1.9% LMICs. Notably, no participation came from LICs. Together, HICs/UMICs accounted for 98.1% of contributions, with the USA accounting for 17%, followed by the UK (14.2%), Canada (11.3%), the Netherlands (9.4%), and Australia (5.7%). When looking at representation per framework, HICs kept at 91.4%, while UMICs and LMICs reached 20.0% and 5.7%, respectively. Notably, there was one (2.9%) article that represented a collaboration spanning HIC, UMIC, and LMIC and another spanning LMICs alone. Authorship of HICs and China was significant, representing 97% of both first and senior authors. 71.4% of frameworks received some form of funding, most often from government institutions (57.1%). The remainder of funding came from academia (31.4%), the private industry and non-profit organizations (22.9% each). Fifteen studies (42.9%) had multiple funding sources (Figure2).