Global gender representation in neurointerventional surgery: a natural language processing analysis of journal publications.

Background: Despite increased awareness of diversity and inclusion in neurointerventional surgery, the representation of women in neurointerventional academic publishing has not been systematically quantified. We aimed to evaluate global and temporal gender trends among authors publishing in leading neurointerventional journals using natural language processing (NLP) tools.

Methods: We used the National Center for Biotechnology Information (NCBI) Entrez and Medline APIs (application programming interfaces) to extract metadata from all articles published between 2014 and 2024 in the and , the only two journals dedicated exclusively to the field of neurointerventional surgery. For each publication, we identified the first and last authors, as well as their institutional affiliations. We applied the OpenAI GPT-4 API to infer the country of origin based on last author affiliations and to predict gender using first and last names.

Results: Across 4875 articles published between 2014 and 2024 in two leading neurointerventional surgery journals, female authorship increased modestly over time. In 2014, 10.2% of first authors and 6.1% of last authors were women, compared with 14.2% and 7.6%, respectively, in 2024. A significant upward trend in female first authorship was observed over time using the Cochran-Armitage test for trend (P=0.0002), with the strongest increases in the Americas and Asia. In contrast, no overall increase was seen in female last authorship (P=0.14), with a modest but significant trend observed only in Asia. Geographic disparities were also evident: Austria and Norway had the highest female last authorship rates (62.5% and 40.0%, respectively), whereas the US, despite contributing the most publications, had only 13.8%. The Gender Parity Index (GPI) further highlighted under-representation, with most countries falling well below parity, especially for last authorship.

Conclusions: Analysis using artificial intelligence driven NLP methods suggests persistent gender inequalities in neurointerventional surgery authorship with variation across countries. Although female representation has improved over the past decade, progress is uneven. These findings underscore the need for targeted efforts to promote equity in academic vascular neurosurgery globally.