The role of cone-beam computed tomography in the radiographic evaluation of obstructive sleep apnea: A review article

Obstructive sleep apnea (OSA) is a common sleep disorder that leads to serious physiological abnormalities with numerous therapeutic implications. The concept of the apnea-hypopnea index has redefined the understanding of OSA, serving as a standard measure of sleep physiology abnormalities and an indicator of disease severity.1 As such, OSA is now characterized as 5 or more instances of apnea or hypopnea per hour of sleep with accompanying symptoms, or 15 or more episodes per hour of sleep regardless of associated symptoms.2, 3

The human pharyngeal airway lacks bone support. Consequently, the muscles must work to maintain an unobstructed air passage through the upper respiratory tract. These muscles facilitate the bending of the upper airway for speech, swallowing, and respiration. Evidence regarding neural modulation, muscular response, and efficacy suggests that the pharyngeal muscles may contribute to OSA.4

Sleeplessness is a result of intermittent upper airway collapse and low blood oxygen saturation.5 In the initial stages of OSA, patients often cannot differentiate between symptoms of the condition and regular fatigue, primarily because chronic fatigue, a common OSA symptom, is frequently underestimated. As the disease progresses, sleepiness becomes detrimental, interfering with daily life.2 Recent international research indicates that the prevalence of OSA has risen to 2% - 4% in men and 1% - 2% in women within the average-aged population.1 Since OSA is quiet, 20%–30% of middle-aged patients display no clinical indicators.2 The rates of OSA have surged in tandem with global obesity trends and show no signs of decelerating. OSA has been recognized as a serious problem with extensive implications for public health, given its association with conditions such as lung diseases, diabetes, and cardiovascular diseases.2, 6

Polysomnography serves as the fundamental basis for diagnosing OSA. This procedure, also referred to as a sleep study, entails the continuous monitoring of various neurophysiological and cardiorespiratory changes.7 Polysomnography plays a pivotal role in both the diagnosis of OSA and the evaluation of treatment success. However, to arrive at a definitive diagnosis, particularly in children, it is essential to integrate polysomnography with the patient’s clinical history and radiographic evaluation.8

This review article received approval from the Ethics Committee of the Faculty of Dentistry at Cairo University prior to publication. The aim of this review was to examine the radiographic evaluation of OSA, with an emphasis on CBCT as an imaging modality. Furthermore, most tools related to OSA assessment are detailed in the ensuing discussion.

The clinical evaluation process is complex due to the location of the disease, which does not permit direct visualization.9 To identify the anatomical structures that undergo changes during sleep, a complete radiographic evaluation is necessary.10 Two-dimensional (2D) and 3-dimensional (3D) imaging methods can be employed to assess the oropharyngeal airway. These techniques include cephalometric radiography, magnetic resonance imaging (MRI), computed tomography (CT), and cone-beam computed tomography (CBCT).11

In the differentiation of OSA from other conditions, 2D cephalometric scans were long considered reliable. However, without 3D scanning, the complex outline of the airway cannot be properly evaluated.12 Multiple studies have demonstrated that the static cross-sectional area of the pharyngeal airway is smaller in patients with OSA compared to those without the condition.1, 13 Current research indicates that changes to the mandible and maxilla can influence the cross-sectional dimensions of the pharyngeal airway space.14, 15

CT imaging establishes numerous sections in different orthogonal planes, characterized by minimal thickness and imperceptible partial volume artifacts.16 However, the primary disadvantage of CT imaging is the requisite exposure to radiation. Additionally, multiple logistical challenges are associated with sleep monitoring during CT imaging.17 Although MRI can accurately depict the true soft tissue shape of the airway, its application has been limited due to the associated noise and excessively prolonged scan time.

Maxillofacial imaging has undergone a major transition from a 2D approach to a volumetric perspective, largely due to the power of CBCT 3D imaging.18 CBCT is considered the optimal choice for 3D dental-sleep imaging due to its high spatial resolution, which enables differentiation between structures such as bone, empty space, and soft tissues.19

Despite the superiority of CT over CBCT in non-soft tissue contrast, CBCT demonstrates a comparable reduction in radiation dose. While still acknowledging the substantial adverse impacts of OSA, the principle of “as low as reasonably achievable” necessitates the reduction of radiation dose to enable precise diagnosis.

When considering the triad of non-soft tissue contrast, radiation exposure, and the potential impacts of OSA, CBCT therefore emerges as the most reasonable imaging modality. Furthermore, the application of specialized analysis software equipped with advanced thresholding techniques, along with the evolution of artificial intelligence-based computer-aided diagnostic methods, can effectively address the validation of non-soft tissue contrast abnormalities.20

The role of CBCT in OSA can be divided into several tasks. These include evaluating the dimensions of the airway; distinguishing between patients with OSA and their counterparts; correlating clinical parameters with radiographic findings; and, ultimately, validating diagnosis and treatment planning by determining the level of collapse and identifying abnormalities not visible via drug-induced sleep endoscopy.

Airway volume rendering

The development of volume rendering was driven by the difficulty of correlating structures in each orthogonal plane. Volume rendering is defined as a collection of methods designed to enhance visualization through the selective presentation of volumetric data. Two distinct methods are typically employed. The first, termed indirect volume rendering, involves the selection of grayscale intensity levels for the voxels, a process referred to as segmentation. This technique allows for the reconstruction of a 3D volumetric surface. The second method, direct volume rendering, involves the random selection of threshold intensities for individual voxels, while excluding all but the darkest and lightest shades of grey. Both of these techniques can be utilized to segment the airway structure, effectively isolating it from the surrounding structures. Given that CBCT data are initially reconstructed in 3D and offer superior spatial resolution, this technology yields the most accurate 3D segmented airway volume, as depicted in Figure 1.18

Fig. 1
A cone-beam computed tomographic scan displays an automated 3-dimensional segmentation of the oropharyngeal airway volume. The segmentation, generated by InVivo5 software (Anatomage, San Jose, CA, USA), extends from the level of the posterior nasal spine to the level of the most anterior inferior point of the second cervical vertebrae.

• Click for larger image

The primary metrics used to assess the static structure of the upper respiratory tract are typically recorded at the retropalatal or retroglossal site. The volume of the airway can be evaluated using 3 separate measurements: the retropalatal airway volume, the retroglossal airway volume, and the total airway volume. Numerous studies have established CBCT as the gold standard for both linear and angular measurements, comparable to multislice CT. Furthermore, the use of an airway phantom for airway evaluation has indicated that CBCT is accurate and reliable for imaging the upper airway and for measuring airway volume, as depicted in Figure 2.21, 24

Fig. 2
Automated graphs represent a patient with obstructive sleep apnea. These graphs display the cross section area, measured in mm², along with the anteroposterior (AP) and lateral (L) dimensions, measured in mm. These measurements were taken along the predetermined oropharyngeal airway space using Invivo 5 software (Anatomage, San Jose, CA, USA).

• Click for larger image

Soft palate and airway lengths

An elongated soft palate is a local airway component that may contribute to OSA. CBCT studies have demonstrated that individuals with sleep apnea possess larger and longer soft palates than those without the disorder, as depicted in Figure 3. The soft palate has been found to change over time and with disease progression; specifically, research has demonstrated excessive flaccidity and muscle atrophy in the soft palates and uvulas of patients with OSA who snore. Shigeta et al. analyzed the midsagittal cross-sectional area of the soft palate using CBCT and observed that the soft palate was substantially longer in patients with OSA compared to control participants.27

Fig. 3
Image of a patient with obstructive sleep apnea who had a long soft palate, as seen in the midsagittal plane of a cone-beam computed tomographic scan.

• Click for larger image

The distance between the palatal plane and a parallel plane created at the level of the lowest border of C4 is commonly used to estimate the maximum length of the upper airway. In a CBCT study conducted by Abramson et al.,28 the presence of OSA was associated with a greater airway length (76.7 ± 11.1 mm) relative to controls (66.3 ± 10.1 mm). Furthermore, the morphology of the airway differs between patients with OSA and their counterparts. Those with OSA typically have an airway with a concave or elliptical shape, whereas non-OSA controls tend to have an airway that displays a concave, round, or square shape.29

Validation of diagnosis and treatment planning

The oropharynx, which extends from the posterior nasal spine to the tip of the epiglottis, is the primary focus of investigation in the upper airway due to its high susceptibility to obstruction.30 A pioneering CBCT study revealed that CBCT, a cutting-edge 3D imaging technique, offers reliable information about the level of collapse with minimal radiation dosage, a finding that has been validated by drug-induced sleep endoscopy. Notably, drug-induced sleep endoscopy is considered a dynamic method for determining the level of collapse, whereas virtual endoscopy is a static modality that lacks this capability.31

CBCT serves as an effective adjunct imaging tool that aids in the evaluation of upper airway dysfunction. Particularly in patients with OSA who exhibit tongue-palate contact or an extended palate, CBCT images provide definitive results, most notably in the end-expiration views in orthogonal plane interpretation (Fig. 4).32 CBCT should not be solely relied upon for assessing the severity of collapse. Instead, it should be utilized as a potent additional imaging resource that can assist in quantifying upper airway failure.31

Fig. 4
Images cited from Aboelmaaty et al.32 depict the views of the end of expiration (A) and inspiration (B) in a single patient with obstructive sleep apnea. The tongue-palate interaction is observed to reduce the airway volume at the end of expiration.

• Click for larger image

The validity of CBCT in assessing OSA is debatable, given that most CBCT devices capture images of the airway while the patient is either standing or sitting. A recent study performed by Aboelmaaty et al.31 found that the patient’s alignment during image acquisition could impair the expansion capacity of the upper airway in the anterior-posterior plane. Furthermore, they asserted that the airway space is at its minimum size when the patient is awake.

In 2005, Tsuiki et al.33 examined changes in airway size among patients with OSA in both upright and supine positions. The researchers found that the anatomic characteristics of the upper airway are influenced by the tonicity of the musculature, which differs significantly when the patient is in the supine position as opposed to when they are awake. Furthermore, they noted that airway obstruction at the level of the larynx and laryngopharynx cannot be established through static imaging techniques such as CBCT. Instead, such obstructions are better examined through dynamic methods, like drug-induced sleep endoscopy. Consequently, they concluded that the results obtained from CBCT should be interpreted as a relative index of airway obstruction, rather than an absolute value.

Another study, conducted by Camacho et al.34 at Stanford Hospitals and Clinics, indicated that gravity and tissue elasticity likely play major roles in the variations observed in upper airway volume. The retropalatal region, which is particularly susceptible to the effects of gravity when the palate is in the supine position, is the area where changes in the size of the upper airway are most likely to occur. The study demonstrated that when patients were in the supine position, both the total volume and the cross-sectional area of the airway were reduced.

Souza et al.35 asserted that the grayscale value of a voxel is influenced by its position within the image volume. Consequently, it is improbable for 2 voxels with identical tissue density but differing scan positions to exhibit the same grayscale level in the reconstructed image. This phenomenon illustrates the absence of correlation between Hounsfield units and grayscale values. The lack of soft tissue contrast in the CBCT data stems from the constraints on grayscale values and tissue density. As such, the effects of thresholding on spatial positions are impeded.

In a study performed to establish a reliable method for measuring the dimensions of the upper airway, manual segmentation of each CBCT section was suggested as an alternative to single threshold-based segmentation algorithms. However, this approach is generally considered impractical due to the extensive time required to complete it for a single dataset; additionally, the potential for error remains. Thus, the automatic segmentation offered by various software packages is viewed as a more reliable technique. Nevertheless, the accuracy of airway volume measurements is contingent upon the thresholding technique employed by each software package.35

In conclusion, CBCT is an innovative imaging technique within the field of dental sleep medicine, offering a broad spectrum of applications for the diagnosis, evaluation, and staging of OSA. However, regarding OSA assessment, CBCT should not be used in isolation. Instead, it should be integrated with other diagnostic tools, such as clinical assessment and polysomnography, as part of a comprehensive approach by a multidisciplinary team.

CBCT is a low-radiation 3D imaging technology that can be employed to diagnose patients with OSA as an incidental finding, facilitating early referral. In patients with OSA, the average size and capacity of the airway are lower, the airway distance is greater, and the airway passage is more constricted compared to healthy individuals. Moreover, the CBCT technology may contribute to enhancing the prediction of surgical outcomes.