Detection of apnea and hypopnea events using a wireless Abdomen-Worn sensor with SpO2 integration

Abstract

Purpose Polysomnography is the gold standard for diagnosing sleep apnea (SA), but it is costly and not widely available. Home sleep apnea testing (HSAT) offers a more accessible, lower-cost alternative. Type III HSAT typically uses nasal pressure sensors, respiratory inductance plethysmography (RIP) belts, and pulse oximetry; however, nasal sensors and RIP belts may be uncomfortable or unsuitable for some patients. This study assessed whether apnea and hypopnea events can be detected using a wireless abdomen-worn sensor (Soomirang) combined with SpO(2) monitoring, eliminating the need for nasal pressure sensors or RIP belts. Methods Data from 37 participants were collected using the Soomirang device and a typical type III HSAT (AL). Two models were developed for apnea and hypopnea event detection: SoomOxy, combining abdominal and body movement data with SpO(2), and Soom, using abdominal and body movement data alone. Their performance was evaluated against AL. Results SoomOxy demonstrated strong agreement with AL, achieving an area under the curve of 0.9447 for apnea and 0.8702 for hypopnea detection, a predicted apnea hypopnea index correlation of 0.96, and an average accuracy of 0.8286 across all severity categories. The SoomOxy outperformed Soom model in detecting hypopnea events. Conclusion A wireless abdomen-worn sensor combined with SpO(2) monitoring can accurately detect and classify apnea and hypopnea events without nasal pressure and RIP belts, offering a practical and more comfortable alternative to conventional HSAT setups.