Süreyya Saridas DEMİR
Perinatal Journal - 2026;34(2):592-598
To map the global bibliometric landscape of Artificial Intelligence (AI) applications in fetal ultrasonography indexed in Web of Science (WoS), Scopus, and PubMed between 2015 and 2025. A systematic bibliometric analysis was performed using WoS Core Collection, Scopus, and PubMed. Eligible publications were English-language original articles and reviews on AI in fetal ultrasonography published from 2015 to 2025. Bradford's, Lotka's, and Price's laws were applied; VOSviewer (v1.6.20) was used for co-authorship and keyword co-occurrence network analysis. A total of 381 publications from 44 countries were retained. Annual output rose from 2 articles in 2015 to 103 in 2025 (annual growth rate 48.3%). Total citations reached 6,516 (mean 17.1/article; h-index 40). China (27.8%), USA (14.7%), and England (13.9%) led output. Bradford Zone 1 comprised 11 core journals. Of 1,548 unique authors, 78.9% published a single article. VOSviewer identified five thematic clusters, with transformer architectures emerging as the most recent paradigm. AI research in fetal ultrasonography expanded 51.5-fold over the past decade. Deep learning is the dominant methodology; future research should prioritise federated learning, multi-centre validation, and AI tools for low-resource settings.
