Closed-Loop Control Algorithms for Adaptive Eustachian Tube Treatment

About This Grant

Project Summary Eustachian tube dysfunction (ETD) affects millions of children annually, causing otitis media that results in 30 million pediatric visits in the US. Current clinical practice leaves pediatricians with no treatment options when middle ear effusion is diagnosed, forcing months of observation during critical developmental periods or eventual surgical intervention. This project addresses this significant treatment gap by developing closed-loop control algorithms for the first point-of-care ETD treatment device using masseter reflex stimulation. The tensor veli palatini muscle opens the Eustachian tube through trigeminal nerve innervation. Masseter reflex stimulation provides a non-invasive pathway to trigger this mechanism, but clinical implementation faces a critical engineering challenge. Pediatric effusions vary dramatically in viscosity from thin serous fluids (1-2 centipoise) to thick glue ear consistency (500-2000+ centipoise), creating vastly different drainage requirements that cannot be determined visually during examination. This demands intelligent control systems that automatically discover optimal stimulation parameters without subjecting children to prolonged treatment attempts. This research will develop safe Bayesian optimization algorithms using computational fluid dynamics modeling to create adaptive therapeutic systems. Our approach employs finite element analysis to characterize stimulation-response relationships across the complete clinical viscosity spectrum, then implements Gaussian process surrogate models to efficiently navigate the three-dimensional parameter space of stimulation intensity, duration, and frequency. The closed-loop system uses real-time tympanometry feedback to iteratively optimize treatment protocols, converging on optimal parameters within 3 treatment attempts while maintaining established safety bounds for pediatric neuromuscular stimulation. The innovation represents the first closed-loop medical device for routine pediatric conditions requiring sophisticated multi-dimensional optimization algorithms that adapt to extreme biological variability. Success will deliver the computational framework essential for clinical translation of masseter reflex stimulation, enabling the first point-of-care treatment that restores hearing when effusion is diagnosed. This paradigm shift from passive observation to active intervention addresses the most common reason children seek medical care, potentially transforming pediatric otolaryngology by providing immediate therapeutic options during critical developmental periods when hearing loss has lasting developmental consequences.