Neural risks, protection, and monitoring in endoscopic sinus surgery: anatomy, imaging, outcomes, and emerging AI tools

Endoscopic sinus surgery traverses a narrow corridor bordering the orbit and the anterior skull base, where rare errors can cause vision loss, cerebrospinal fluid leak, or vascular injury. This review summarizes neural risk, protection, and monitoring in endoscopic sinus surgery and outlines the role of artificial intelligence. Risk stratification uses high-resolution computed tomography and selective magnetic resonance imaging to link anatomic variants such as Keros depth, Onodi cells, Haller cells, and dehiscence of the carotid or optic canal to disciplined intraoperative behavior. Preoperative planning with image-guided surgery converts imaging into coordinated team action. Intraoperatively, meticulous visualization, conservative instrument choreography, and scenario specific safeguards support neural protection. Visual evoked potentials for intraoperative neuromonitoring are selectively useful near the optic apparatus. Emerging artificial intelligence tools including computer vision landmarking, registration of operative video to computed tomography with augmented reality overlays, and proximity alerts can provide cognitive support when integrated with low latency, confidence gating, and fail-safe defaults. A complication-centered checklist that links predictors to prevention and first line management may help teams reduce already rare neural complications while preserving vision, olfaction, and quality of life.