0460 Detecting Respiratory Events By Respiratory Effort Derived From 3D Time-of-Flight Camera And SpO2. (12th April 2019)

• Record Type: Journal Article
• Authors: Coronel, Carmina; Wiesmeyr, Christoph; Garn, Heinrich; Kohn, Bernhard; Wimmer, Markus; Mandl, Magdalena; Glos, Martin; Penzel, Thomas; Kloesch, Gerhard; Stefanic-Kejik, Andrijana; Boeck, Marion; Kaniusas, Eugenijus; Seidel, Stefan
• Is Part Of: Sleep. Volume 42(2019)Supplement 1
• Page Start: A185
• Page End: A185
• Publication Date: 2019-04-12
• Subjects: Sleep -- Physiological aspects -- Periodicals; Sleep disorders -- Periodicals; Sommeil -- Aspect physiologique -- Périodiques; Sommeil, Troubles du -- Périodiques; Sleep disorders; Sleep -- Physiological aspects; Sleep -- physiological aspects; Sleep Wake Disorders; Psychophysiology; Electronic journals; Periodicals; 616.8498
• DOI: 10.1093/sleep/zsz067.459
• Languages: English
• ISSNs: 0161-8105

Abstract

Abstract: Introduction: To calculate the Apnea-Hypopnea Index (AHI), the detection of apneas and hypopneas is performed using multiple sensors of a polysomnography (PSG) such as nasal airflow and respiratory inductance plethysmography (RIP). This setup is uncomfortable and labor intensive. Therefore, we used a contactless 3D time-of-flight (TOF) camera to monitor the respiratory effort. Using this respiratory effort and SpO2, we developed an algorithm to detect apnea and hypopnea events without the need of airflow sensors and RIP belts. Methods: 3D Video-PSG was performed for 53 patients with suspected sleep apnea syndrome at Advanced Sleep Research GmbH, Berlin and Johannes Kepler University Clinic, Linz, as approved by the relevant ethical committees. The respiratory effort signal (Effort3D) was derived from the camera's depth information over the upper body region. An Effort3D-SpO2 algorithm for detecting apnea and hypopnea events was developed wherein an event is a segment of at least 10 seconds with a substantial decrease in Effort3D and an associated 4% desaturation in SpO2 . The 3DSpO2 -AHI was calculated from the number of detected events and the total sleep time (TST), obtained from the PSG's hypnogram. The intraclass correlation coefficient (ICC) with its 95% confidence interval (CI) was used to measure reliability with manually scored m-AHI according to the American Academy of Sleep Medicine (AASM) scoring manual and automatically calculated PSG-AHI from the PSG. Results: The ICC for 3DSpO2 -AHI versus m-AHI is 0.97 (CI: 0.95-0.98) and the median absolute difference is 4. On the other hand, the ICC for 3DSpO2 -AHI and PSG-AHI is 0.91 (0.85-0.95) and the median absolute difference is 4. Between m-AHI and PSG-AHI, the ICC is 0.93 (0.88 - 0.96), and the median absolute difference is 4. Conclusion: The respiratory effort derived from 3D TOF camera together with SpO2 is a promising option in detecting respiratory events as it has shown excellent reliability scores, comparable to automated-PSG and manual scoring. The contactless 3D TOF camera and SpO2 is a more comfortable alternative for overnight recordings. Support (If Any): Austrian Research Promotion Agency (FFG), project ID 859622.