SOMNAR: A platform to localize airway obstruction in sleep apnea

Bairitone Health Inc., Houston TX

Investigators

Abstract

Abstract - An estimated 24 million undiagnosed or untreated people with OSA (obstructive sleep apnea) contribute to a $150B economic burden in the US. Most people with OSA are prescribed continuous positive airway pressure (CPAP) as a first line treatment, but CPAP has a 50% failure rate. Highly effective treatments that target the anatomic causes of OSA are available but are only effective when the anatomical source of apnea is appropriately identified. Currently there is no accessible diagnostic that can query airway anatomy during natural sleep. Increasingly, sleep surgeons use Drug-Induced Sleep Endoscopy (DISE) to investigate the upper airway anatomy for OSA surgical planning. This diagnostic is subjective, a drain on surgeon time and operating room resources, inconvenient for patients, and cannot be performed during natural sleep. Several methods have been investigated to identify obstruction sites during natural sleep, such as using MRI to image the airway of sleeping patients or introducing a thin tube with multiple pressure sensors into the upper airway. However, these methods are cost prohibitive, impractical for clinical use, or not tolerated by every patient. Bairitone Health proposes to develop a platform for the anatomic diagnosis of obstructive sleep apnea (OSA). The technology detects tissue-borne sounds, recognizes their association with normal breathing or airway obstruction, and identifies the location in the upper airway they emanate from. The overall goal of this Phase I project is to demonstrate that Bairitone Health’s approach for localization of airway obstructions is technically feasible and can be developed into a clinical application. To accomplish this goal, Bairitone Health has four specific aims: 1) develop and train an algorithm to recognize post-apnea snore events indicative of airway reopening, 2) develop speech-based calibration and airway-obstruction localization methods, 3) optimize the localization method on a 3D-printed human head model by generating sounds at specific locations within the model's airway, and 4) develop an updated prototype of the sound recording device and sensor configuration. If successful, the platform will allow the diagnosis of OSA and determination of its anatomical source in one single procedure, streamlining the process and cutting costs. Completing these aims will validate the technical feasibility of the platform and usher the technology towards commercial application. Moving the anatomy assessment from the operating room to the home will result in higher OSA diagnostic capacity. In the future, Bairitone Health plans to integrate traditional diagnostics collected in current home sleep tests into the home-based anatomy assessment. With the anatomy information the diagnostic platform can provide, this product will become the first-choice testing tool for obstructive sleep apnea diagnosis and personalized treatment planning.