Open Access
Ultrasound Int Open 2017; 03(01): E34-E42
DOI: 10.1055/s-0042-124503
Technical Development
© Georg Thieme Verlag KG Stuttgart · New York

Ultrasonographic Detection of Airway Obstruction in a Model of Obstructive Sleep Apnea

Amal Isaiah
1   Otorhinolaryngology – Head and Neck Surgery, University of Maryland School of Medicine, Baltimore, United States
,
Reuben Mezrich
2   Radiology, University of Maryland School of Medicine, Baltimore, United States
,
Jeffrey Wolf
3   Otolaryngology – Head and Neck Surgery, University of Maryland School of medicine, Baltimore, United States
› Institutsangaben
Weitere Informationen

Publikationsverlauf

received 19. Juli 2016
revised 06. Dezember 2016

accepted 08. Dezember 2016

Publikationsdatum:
22. März 2017 (online)

Preview

Abstract

Purpose Obstructive sleep apnea (OSA) is a common clinical disorder characterized by repetitive airway obstruction during sleep. The gold standard for diagnosis of OSA, polysomnogram (PSG), cannot anatomically localize obstruction. Precise identification of obstruction has potential to improve outcomes following surgery. Current diagnostic modalities that provide this information require anesthesia, involve ionizing radiation or disrupt sleep. To mitigate these problems, we conceived that ultrasound (US) technology may be adapted (i) to detect, quantify and localize airway obstruction and (ii) for translational application to home-based testing for OSA.

Materials and Methods Segmental airway collapse was induced in 4 fresh cadavers by application of negative pressure. Following visualization of airway obstruction, a rotary US probe was used to acquire transcervical images of the airway before and after induction of obstruction. These images (n=800) were analyzed offline using image processing algorithms.

Results Our results show that the non-obstructed airway consistently demonstrated the presence of a US air-tissue interface. Importantly, automated detection of the air-tissue interface strongly correlated with manual measurements. The algorithm correctly detected an air-tissue interface in 90% of the US images while incorrectly detecting it in 20% (area under the curve=0.91).

Conclusion The non-invasive detection of airway obstruction using US represents a major step in expanding OSA diagnostics beyond PSG. The preliminary data obtained from our model could spur further research in non-invasive localization of obstruction. US offers the benefit of precise localization of the site of obstruction, with potential for improving outcomes in surgical management