An Explainable, Knowledge-Driven System for Difficult Airway Risk Stratification and Planning

Anticipating a difficult airway during surgery is critical for patient safety, yet recognition and escalation can be inconsistent in practice. We present an ontology-based clinical decision support system that standardizes pre-anesthesia difficult-airway assessment and proposes evidence-based plans. The ontology models patient descriptors, predictive scores, and airway techniques. The system computes the ARNE score from recorded findings, applies an expert-curated mapping to provide appropriate techniques recommendations. The pipeline was evaluated on 20 simulated cases covering typical and edge presentations with 80% agreement on priority intubation technique. The system executed end-to-end for all cases, producing consistent, reasoner-validated outputs. In conclusion, these findings indicate that transparent, institution-aware decision support for airway planning is feasible and aligned with clinical realities.