Validation of novel automatic ultra-wideband radar for sleep apnea detection

摘要

Background: To validate the accuracy of ultra-wideband (UWB) wireless radar for the screening diagnosis of sleep apnea. Methods: One hundred and seventy-six qualified participants were successfully recruited. Apnea-hypopnea index (AHI) results from polysomnography (PSG) were reviewed by physicians, while the radar device automatically calculated AHI values with an embedded chip. All results were statistically analyzed. Results: A UWB radar-based AHI algorithm was successfully developed according to respiratory movement and body motion signals. Of all 176 participants, 63 exhibited normal results (AHI 5/hr) and the remaining 113 were diagnosed with obstructive sleep apnea. Significant correlation was detected between radar AHI and PSG AHI (Intraclass correlation coefficient 0.98, P0.001). Receiver operating characteristic curve (ROC) analysis revealed high sensitivity and specificity. High concordance in participants with varying gender, age, BMI, and PSG AHI was reached. Conclusions: The UWB radar may be a portable, convenient, and reliable device for obstructive sleep apnea screening.