@article{article_1960958, title={Artificial Intelligence–Based Topic Modeling of Cadaveric Anatomy Literature: A 25-Year BERTopic Analysis (2000–2025)}, journal={Journal of Contemporary Medicine}, volume={16}, pages={179–184}, year={2026}, url={https://izlik.org/JA54PS37PU}, author={Keskin, Arif and Aygün, Tayfun and Palancı, Özgür}, keywords={Kadavra, Diseksiyon, Anatomi, Doğal Dil İşleme, Bibliyometri}, abstract={Aim: This study aimed to identify research topics in the cadaveric anatomy literature between 2000 and 2025, reveal semantic relationships between these topics, and analyze temporal trends. Methods: A systematic search of the Scopus database yielded 2,465 articles published between 2000 and 2025; after preprocessing, 2,455 were included. The Bidirectional Encoder Representations from Transformers Topic Modeling algorithm — integrating Sentence-BERT, Uniform Manifold Approximation and Projection, and Hierarchical Density-Based Spatial Clustering of Applications with Noise — was applied to article titles and abstracts. Topic coherence was assessed using the C_v score, and temporal trends were evaluated with the Mann-Kendall trend test. Results: A total of 62 topics were identified (mean C_v=0.664). Of all articles, 29.8% were classified as outliers. Eighteen topics showed a statistically significant increasing trend; no topic showed a decrease. The strongest increase was observed in sciatic nerve and related clinical anatomy topics. Education, endoscopic surgery, and vascular anatomy emerged as prominent thematic areas. The United States accounted for the largest share of publications (34.4%), with contributions from 60 countries overall. Conclusion: To our knowledge, this is the first transformer-based topic modeling study to map the cadaveric anatomy literature. The findings demonstrate significant growth in education-focused research alongside sustained expansion across surgical and clinical anatomy topics, confirming that cadaver studies maintain their central role in modern medical training and research.}, number={4}