The relationship between clinical severity of obstructive sleep apnea based on polysomnography and drug-induced sleep endoscopy with 3D, 2D, linear, and angular anatomical parameters of upper airway and craniofacial area in CBCTs of individuals with moderate or severe apnea: a cross-sectional study

Abstract

Objective: The objective of this study is to determine the relationship between location of upper airway respiratory collapse using drug-induced sleep endoscopy (DISE) technique and craniofacial morphology using cone beam computed tomography scans (CBCT) in patients diagnosed with moderate to severe obstructive sleep apnea (OSA).

Methods: In this prospective cross-sectional study, 64 adults with OSA (33 moderate and 31 severe) who were candidates for surgery were included. Before surgery, all patients underwent nocturnal polysomnography, DISE and CBCT scan. Measurements of craniofacial anatomical variables were performed on CBCT images. Associations among age, sex, BMI, neck circumference, CBCT measurements (3D, 2D, linear, and angular), DISE findings at various airway levels, and polysomnography indices (apnea-hypopnea index, average oxygen saturation, and minimum oxygen saturation) were analyzed using regression, ANOVA, Tukey's post-hoc test, t-test, chi-squared test, and correlation coefficients.

Results: At the lateral wall of the velum, (1) obstructions increased AHI; (2) there were significant associations between obstructions and CBCT variables: V-RPA, V-RLA, and L-RPA (P < 0.05). Other DISE findings were mostly not associated with CBCT or polysomnography measurements, except obstructions at oropharynx-lateral which increased AHI. However, DISE-Sum was significantly correlated with AHI and oxygen saturations (P ≤ 0.009). It was also correlated with several CBCT measurements (P < 0.05); Regression analysis showed associations between DISE-Sum and Go-Me, C3H, TGL, and TGH (adjusted R² = 0.44). For AHI, CBCT measurements explained almost all the variance (adjusted R² = 0.97). For minimum oxygen saturation, CBCT variables including C3H, PNS-P, HRGN, SNB, and L-RLA showed predictive capability (adjusted R² = 0.31).

Conclusions: CBCT measurements can predict OSA severity. The newly introduced DISE-Sum is highly correlated with polysomnography and CBCT findings. Individual DISE scores may be associated with AHI and CBCT only at one or two airway levels. BMI is associated with OSA. Age and neck circumference are associated with OSA only before controlling for BMI.

Keywords: 3-dimensional (3D); Apnea-hypopnea index (AHI); Average oxygen saturation; Cone-Beam computed tomography (CBCT); Craniofacial; Drug-induced sleep endoscopy (DISE); Minimum oxygen saturation; Obstructive sleep apnea; Polysomnography (PSG); Sex dimorphism; Upper airway.