Bridging Data, Semantics, and Clinical Reasoning: A Knowledge Graph Framework for Pediatric Obstructive Sleep Apnea

Background/Objectives: Pediatric obstructive sleep apnea (OSA) is a complex disorder with a variable presentation and often challenging diagnostic testing. The history and physical examination in pediatric OSA frequently differ from those in adults. The treatment options are multifaceted and must be tailored to the individual patient. Artificial intelligence (AI) modalities currently employed in pediatric sleep medicine face several important limitations: modality fragmentation, lack of explainability, and limited semantic integration. Method: Our team proposes a new vision for AI and pediatric sleep medicine. This platform is based on a knowledge graph (KG) framework integrating structured and unstructured data to enable reasoning, personalization, and clinical decision support. Results: This framework represents a conceptual architecture; it has not yet been empirically implemented, and the use cases described herein are illustrative of its intended capabilities. Components of the infrastructure developed for similar applications have been successfully implemented. The quantitative feasibility pilot KG represented 100% multimodal data with >90% semantic completeness. Conclusions: Fully realized and deployed into the clinical space, this pediatric OSA KG system will enhance tertiary care programs and help project tertiary-level pediatric care into underserved regions.