RSA‐KG: A Graph‐Based Rag Enhanced AI Knowledge Graph for Recurrent Spontaneous Abortions Diagnosis and Clinical Decision Support
Yibo He, Yue Chai, Yonglin Liu, Jiajia Chi, Yiran Fei, Beihui He, Ting Zhang, Shiliang Chen, Ying Yu, Zhezhong Zhang, Jinwei Li, Yun Sun, Shiyuan Tong, Guoyin Kai
Abstract
ABSTRACT Recurrent spontaneous abortion (RSA), affecting 1%–5% of reproductive‐aged women, presents diagnostic challenges due to its complex multifactorial causes. Conventional guidelines are inadequate for idiopathic cases and emerging biomarkers, whereas artificial intelligence (AI) models struggle to integrate multimodal data. To address these issues, we developed RSA‐KG, a graph‐based AI knowledge graph that synthesizes multimodal clinical data and adapts to evolving guidelines. RSA‐KG integrates 5 international RSA guidelines, utilizing natural language processing (NLP) and multimodal models for data processing. Evaluation of RSA‐KG showed that LLMs enhanced by RSA‐KG outperformed naive retrieval‐augmented generation (RAG) and raw models in diagnostic accuracy. Reproductive specialists also rated the output of the RSA‐KG system more favorably than raw models and medical large language models (LLM). RSA‐KG represents a novel approach to RSA management, overcoming limitations of traditional AI by modeling systemic interactions and integrating real‐time evidence. Further validation through multicenter trials is required for broader clinical adoption.