3D volumetric analysis and clinical implications of the upper airway and sinuses

Introduction

The postural relationships of the head, jaws and tongue are established immediately at birth with the patency and opening of the airway. These postural relations change to meet physiological demands and maintain airway patency. The intimate anatomical and functional proximity of craniofacial and upper respiratory structures influences facial and dental growth. A healthy respiratory function is essential for the harmonious growth and development of the maxillofacial structures. Craniofacial skeletal development can be affected by abnormal nasal/oral respiration patterns during growth. The abnormal influences from abnormal respiration might affect dental arch form, occlusion, dental eruption and mandibular and maxillary growth direction. Researchers have extensively investigated the relationship between pharyngeal space and craniofacial morphology variations.

The upper airway is the pharyngeal and paranasal air sinuses ( Fig. 39.1 ).

The pharyngeal airway has three major regions:

1.
Nasopharynx,
2.
Oropharynx, and
3.
Hypopharynx.

The upper airway is supported by various skeletal structures—the cranial base at the top, the cervical spine at the back, the nasal septum at the front and top, and the maxilla, mandible and hyoid bone at the front. However, other structures, such as the nasal conchae, adenoids, soft palate, tongue, pharyngeal and lingual tonsils, can encroach on the upper airway.

Paranasal sinuses are air-containing cavities lodged inside the skull and facial bones, which assure harmony in facial growth and make the bony skull lighter. In addition, sinuses carry out physiological functions of humidification and warming of inhaled air. The sinuses also have an inbuilt mechanism for protection against trauma. Four paired air sinuses border the nasal cavity: the maxillary sinus, frontal sinus, ethmoidal sinus and sphenoidal sinus. The volumetric analysis of the upper respiratory tract represents a valuable diagnostic tool in identifying obstructive sites in the pharyngeal airway. This unique approach allows orthodontic professionals to collaborate with other medical specialities, improving patients’ overall health and treatment outcomes. The potential benefits of such interdisciplinary cooperation are significant and warrant further exploration.

Volumetric evaluation of the upper respiratory tract is vital for therapy planning, evaluation of treatment effectiveness and comprehension of the underlying mechanisms associated with many diseases and deformities.

Imaging modalities of upper airway space

A lateral cephalogram is a commonly used radiological imaging technique to evaluate the airway. It is a simple and affordable tool. The airway is evaluated by measuring the narrowest anteroposterior points in the nasopharynx and oropharynx, which limits the understanding of volumetric changes in the upper airway. Estimating the precise volume of the upper airway based on lateral cephalogram has several challenges and constraints due to image distortion, magnification and the overlapping of bilateral craniofacial structures. These inherent limitations of two-dimensional (2D) radiographs have led to the development of three-dimensional (3D) airway analysis methods ^, ( Fig. 39.1 ).

Because of these limitations, researchers are exploring alternative imaging modalities offering more comprehensive three-dimensional assessments of the pharyngeal airway. The volumetric analysis of the craniofacial region can be obtained through different methods appended below.

Non-radiation methods
- a.
  Acoustic reflection
- b.
  Magnetic resonance imaging (MRI)
Radiation methods
- a.
  Fluoroscopy
- b.
  Computed tomography (CT)
- c.
  Cone beam computed tomography (CBCT)

These 3D imaging methods allow us to evaluate the surface and sub-surface volume of the upper airway accurately. Each method has its inherent advantages and disadvantages.

Non-radiation methods

Acoustic reflection is a non-invasive technique based on the analysis of a sound wave reflected from the airway. It does not involve radiation exposure or provide a high-resolution anatomic representation of the airway or soft tissue structures.
The MRI technique is highly suitable for airway imaging, especially in individuals with sleep apnoea, due to its superior upper airway and soft tissue resolution. Additionally, MRI enables precise assessment of cross-sectional area and volume and allows imaging in multiple axial, sagittal and coronal planes. MRI has an equal resolution but much greater contrast than CT. MRI is performed in a supine position and requires longer operating time. MRI is prone to motion artefacts that affect the quality of imaging. Dynamic imaging is possible with ultrafast MR imaging. Fast MRI can obtain two images per 1 s, which allows multi-dimensional views and visualisation of the dynamic shape of the pharyngeal airway during inspiration and expiration.

Radiation methods

Fluoroscopy is a medical imaging procedure that captures dynamic images of the upper airway. Although it is a valuable tool, it lacks the sensitivity required to measure minute changes in airway dimensions or the detailed motion of the soft tissue structures surrounding the upper airway. Furthermore, this modality is limited in producing cross-sectional images in the axial or sagittal planes. As such, other imaging modalities may be necessary to obtain a comprehensive evaluation of the upper airway.
CT is obtained as 1–2 mm slices in axial or coronal planes or both from the patient in the supine position, which is then combined and reconstructed into 3D images. The exposure to radiation limits the ability to undergo a series of CT tests. Dynamic imaging (images in 50 ms) can be performed with an electron beam (ultrafast computed tomography).
CBCT is more frequently used to assess the upper airway and paranasal air sinuses. It allows for direct measurement of airway volume, allowing precise evaluations and treatment planning. Advantages of CBCT include 3D images with 1:1 reproduction, isotropic voxel size, less acquisition time and a lower radiation dose than multi-detector CT (MDCT). Limitations on the use of CBCT include low contrast range, limited soft tissue visualisation, increased noise and scattering and inability to be used for estimation of Hounsfield units (HU). Although this modality has limitations, CBCT offers a better risk-to-benefit ratio in evaluating upper airway dimensions, particularly in diagnosing and managing obstructive sleep apnoea (OSA) patients.

Radiological anatomy of upper airway

The airway is identified as an oblong, irregularly shaped, low-density (dark) region located anterior to the cervical portion of the vertebral column.

The upper airway is divided into four sub-regions:

1.
Nasal cavity,
2.
Nasopharynx,
3.
Oropharynx, and
4.
Hypopharynx.

The nasopharynx stretches from the skull’s base to the hard palate’s level, while the oropharynx extends from the hard palate to the hyoid bone level. On the other hand, the hypopharynx’s boundaries start extending from the level of the hyoid bone to the caudal cricoid cartilage ( Fig. 39.2 ).

The nasal cavity is separated into three distinct chambers or turbinates on either side by three osseous processes, namely inferior, middle and superior nasal conchae. The nasal septum is identified along the midline of the nasal cavity and is often not fully ossified. Any deviation in the nasal septum contributes to the asymmetry between the right and left nasal cavities. The nasal structures are best visualised in the coronal view at the level of maxillary molars.

The nasopharynx communicates anteriorly with the nasal cavity through the posterior nasal choanae and the middle ear cavity via the eustachian tubes posterolaterally and with the oropharyngeal cavity inferiorly. The base of the skull forms the superior boundary of the nasopharynx, and the inferior boundary is the soft palate and Passavant’s ridge. The lateral boundaries are the superior constrictor muscles/visceral fascia. The adenoid tissues are situated in the middle of the roof of the nasopharynx. The torus tubarius is a soft tissue process on either side of the nasopharynx that separates it from the eustachian tube. Just posterior to the eustachian tube lies the pharyngeal recess or fossa of Rosenmüller, which is included in the evaluation ( Fig. 39.2 ).

Oropharynx: Below the nasopharynx are two areas: the oropharynx and the oral cavity. The oral cavity is located below the nasal fossa and the maxillary sinuses. The oropharynx is situated directly below the nasopharynx and at the back of the oral cavity. The soft palate creates the upper boundary of the oropharynx.

The oropharynx, a segment of the throat, commences at the anterior circumvallate papillae and encompasses the posterior one-third of the tongue. The lateral wall comprises lymphoid tissue denominated palatine tonsils. The posterior oropharyngeal wall is contiguous to the second and third cervical vertebrae. Soft tissue masses, the tonsils and the bilaterally symmetrical faucial pillars are situated on each side of the oropharyngeal airway.

The hypopharynx is a part of the throat that extends from the hyoid bone to the lower level of the cricoid cartilage. It can be divided into three regions: the posterior wall, the pyriform sinus and the post-cricoid region. The pyriform sinus, shaped like a pear, is located on either side of the pharynx and bordered laterally by the thyrohyoid membrane and thyroid cartilage. The hypopharynx’s posterior and lateral walls blend with the cricopharyngeus. The post-cricoid region is the front wall of the lower hypopharynx and serves as the junction between the hypopharynx and larynx.

Segmentation methods

Constructing a patient-specific 3D model of the airway and paranasal air sinuses from a CT/CBCT image requires an accurate segmentation of a region of interest (ROI). The segmentation process is the construction of 3D virtual surface models to best match the volumetric data. Segmentation of a specific structure and removing all other structures of non-interest will aid in better visualisation and the computation of volumes. The segmentation process of the airway can be performed using manual method or automatically.

Manual segmentation

Manual segmentation requires the operator to outline the boundaries of a particular structure on each slice in coronal, axial and sagittal views. The operator marks several points, and the software draws a line between each point. After the line segmentation, the data are transformed into a 3D reconstruction image. The extension of an object in a CT dataset is measured parallel to the object’s coordinate system in the x-, y- and z-directions. The software automatically performs the task of determining the first and last voxel in each coordinate direction. The computation of the object’s volume requires the tally of voxels that belong to each segment, followed by the multiplication of this result by the volume of a single voxel in cubic centimetres. The result obtained by the manual segmentation method is more accurate and allows for better operator control. Inaccuracies may occur because of the inherent imprecision of human vision, rendering the process laborious and meticulous. The manual technique is more time-consuming in comparison to automatic and semiautomatic methods. Hence, using the manual segmentation method is almost impractical for clinical applications.

Automated/semiautomated

Segmentation techniques in medical imaging can be broadly divided into three classes :

1.
Structural techniques: Example, 3D edge detection techniques, isosurfaces and level sets.
2.
Stochastic techniques: Example, thresholding approaches and clustering algorithms.
3.
Hybrid approaches include region growing and artificial neural networks.

No segmentation method provides acceptable results for every type of medical data set. Some methods can be applied to various data sets, but methods specialised for a particular case situation are expected to give better results. Often, a segmentation approach could consist of more than one segmentation algorithm applied one after the other. Automatically segmenting the upper airway region, specifically the nasal cavity and paranasal sinuses, from a CBCT scan is challenging due to noise artefacts and intricate anatomical features. The automated methods do not allow the necessary operator control to correct the segmentation per 2D slice. So, the automated segmentation results need to be more accurate when compared to manual methods. Though the existing software programs have plugins to segment the airway, these are segmentation tools, such as threshold-based and region-growing segmentation techniques.

Thresholding

The thresholding approach is a simple and effective technique for segmentation in volumes with good contrast between regions. This technique is used as the first step towards segmentation of the volume of a structure. The threshold value of a particular ROI, consisting of single or multiple values, is used. When a single value is used, all voxels with intensities more than the threshold value are grouped into one segment and voxels less than the threshold value are grouped into another. If extended to multiple thresholds, a region is defined by two threshold values: a lower limit and an upper limit. The region confined between these threshold values is segmented. Using a single threshold value can generate errors, especially in volume analysis, but it is undoubtedly more reproducible than using a dynamic threshold. The threshold interval selection also influences the upper airway segmentation and volume measurement.

The segmentation process is fundamental to:

1.
Fixed threshold-based segmentation
2.
Interactive threshold-based segmentation

In fixed threshold-based segmentation, the observer chooses a constant threshold value for use throughout the patient’s data. On the other hand, interactive threshold-based segmentation allows the observer to adjust the threshold value according to his perception of the patient’s data. While interactive threshold-based segmentation is highly reliable, it has poor accuracy due to the variations in the operator’s visual discrimination of airway boundaries. The outcomes of manual method can be affected by various factors such as lighting conditions, operating fatigue, grey-scale ability and visual acuity.

Fixed thresholding, on the other hand, eliminates operator subjectivity in boundary selection. However, one of the main drawbacks of thresholding techniques is that depending on the grey threshold, values similar to air, such as thin mucous tissues or secretions, or identical to air, such as noise or air surrounding the patient, may be erroneously added to the segmentation.

Thresholding techniques are sensitive to both noise and intensity inhomogeneities. As a result, they are not feasible for MRI and ultrasound volumes.

Region growing

Region growing technique is the simplest among the hybrid methods. This technique requires the placement of seed points by the user in the ROI. The software creates a segmented region based on a seed point placed. The main drawback of this technique is that it requires manual interaction to place the seed point ( Fig. 39.3 ). If the ROI is not selected correctly, the region will grow into other regions and be segmented accordingly. It is also sensitive to noise and partial volume effects, causing the extracted region to have holes and disconnections. A region-growing feature to segment airway regions using a 3D tool is practical and popular in clinical practice.

Commonly used 3D tools to visualise the data acquired from CT/CBCT include:

•
Anatomage (Santa Clara, CA 95054, USA)
•
AMIRA (AMIRA, Mercury Computer System Inc., Berlin, Germany)
•
Dolphin (Dolphin Imaging and Management Solutions 9200 Oakdale Ave. Suite 700 Chatsworth, CA 91311 USA)
•
Maxilim (Medicim NV, Mechelen, Belgium, Sint-Nikaas, Belgium)
•
Mimics (Materialise HQ, Technologielaan 15, 3001 Leuven Belgium)
•
Vitrea (Vital Images Inc., Plymouth, MN, USA)

Dolphin 3D provides a quick upper airway segmentation, but the threshold interval units (grey levels) are incompatible with other imaging software. Mimics software offers quick and easy airway segmentation. It has the best segmentation control and sensitivity. ITK-Snap is a non-commercial, open-source software that is not as user-friendly as Dolphin 3D and In Vivo Dental.

Segmentation of upper airway

Automatic segmentation of the upper airway and paranasal sinuses has been studied extensively. All these studies have segmented the upper airway and paranasal sinuses as a single region; segmentation of the sub-regions (maxillary sinus, frontal sinus, ethmoidal sinus, sphenoidal sinus, nasal cavity, nasopharynx, oropharynx and hypopharynx) is a tedious task. Later, the semiautomatic segmentation techniques for evaluating the volume of airway sub-regions were proposed. These studies used anatomical boundary definitions to select a volume of interest (VOI) and locate seed points in selected VOI. A seed point is further converged towards the boundaries in the selected VOI. The boundary definitions of each sub-region depend on the pre-defined anatomical landmarks. The exact definitions of upper airway remain undefined, as the anatomical limits vary greatly. The anatomical and technical boundary definitions of pharyngeal airway sub-regions are given in Table 39.1 and of paranasal air sinuses in Table 39.2 . All the required landmarks exist on the mid-sagittal plane ( Table 39.3 ). Hence, the mid-sagittal plane must be identified from the CBCT scan data. The boundary limits on the mid-sagittal plane are shown in Fig. 39.4 .

TABLE 39.1

Anatomical boundary definitions of upper airway sub-regions

Regions	Limits	Definitions for the region of interest	Landmarks
Nasopharynx	Anterior Posterior Upper Lower Lateral	A line perpendicular to the FH plane passing through PNS in sagittal view. A line perpendicular to the FH plane passing through C2 _sp. Soft tissue contour of the pharyngeal wall. Line parallel to FH plane passing through PNS, in sagittal view. Soft tissue contour of the pharyngeal lateral walls	1. PNS 2. C2 _sp
Oropharynx	Anterior Posterior Upper Lower Lateral	A line perpendicular to the FH plane passing through PNS in the sagittal view. A line perpendicular to the FH plane passing through C2 _sp. Line parallel to FH plane passing through PNS, in sagittal view. Line parallel to FH plane passing through C3 _ai. Soft tissue contour of the pharyngeal lateral walls.	1. PNS 2. C3 _ai 3. C2 _sp
Hypopharynx	Anterior Posterior Upper Lower Lateral	A line perpendicular to the FH plane passing through PNS. A line perpendicular to the FH plane passing through C2 _sp. The line parallel to the FH plane passes through C3 _ai. Line parallel to FH plane passing through C4 _ai. Soft tissue contour of the pharyngeal lateral walls.	1. PNS 2. C2 _sp 3. C3 _ai 4. C4 _ai
Nasal cavity	Anterior Posterior Upper Lower Lateral	A line connecting ANS to the tip of the nasal bone to N. A line extending from Sella to PNS. Line connecting N to S. A line extending from ANS to PNS. The plane is perpendicular to the FH plane, passing through the lateral walls of the maxillary sinus.	1. PNS 2. ANS 3. S 4. N 5. Tip of nasion

ANS, Anterior nasal spine; N, nasion; PNS, posterior nasal spine; S, sella.

TABLE 39.2

Anatomical boundary definitions of paranasal air sinuses

S. no.	Regions	Lateral	The sagittal plane perpendicular to the FH plane passes through the lateral walls of the maxillary sinus	Landmarks
1.	Frontal sinus	Anterior Posterior Upper Lower Lateral	A line perpendicular to the FH plane passing through Nasion. A line perpendicular to the FH plane passing through PNS. — Line parallel to FH plane passing through Nasion. The sagittal plane perpendicular to FH passes through the lateral walls of the maxillary sinus.	1. PNS 2. Nasion
2.	Sphenoid sinus	Anterior Posterior Upper Lower Lateral	A line perpendicular to the FH plane passing through PNS. A line perpendicular to the FH plane passing through C2 _sp. — Line parallel to FH plane passing through PNS. The sagittal plane perpendicular to the FH plane passes through the lateral walls of the maxillary sinus.	1. PNS 2. C2 _sp (superior-posterior point of second cervical)
3.	Ethmoidal sinus	Anterior Posterior Upper Lower Lateral	A line perpendicular to the FH plane passing through Nasion. A line perpendicular to the FH plane passing through PNS. Line parallel to FH plane passing through Nasion. Line parallel to FH plane passing through PNS. The sagittal plane perpendicular to the FH plane passes through the lateral walls of the maxillary sinus.	1. Nasion 2. PNS
4.	Maxillary sinus	Anterior Posterior Upper Lower Lateral	The sagittal plane is perpendicular to the FH plane, passing through the most anterior point of the maxillary sinus. The sagittal plane is perpendicular to the FH plane, passing through the most posterior point of the maxillary sinus. The sagittal plane parallel to the FH plane, passing through the uppermost point of the maxillary sinus. The sagittal plane parallel to the FH plane, passing through the lower point of the maxillary sinus. The axial plane perpendicular to the FH plane passes through the lateral walls of the maxillary sinus.	—

PNS, Posterior nasal spine.

TABLE 39.3

Definition of landmarks in three planes of space

S. no.	Landmark	Landmark definitions	Sagittal	Coronal	Axial
1.	C2 _sp	Superior–posterior extremity of the odontoid process of C2	Superior–posterior extremity of the odontoid process in the mid-sagittal plane	Superior and middle point on odontoid process	The superior first visible point on the odontoid process
2.	C3 _ai	Most anterior–inferior point of the body of C3	Anterior–inferior point of the body of C3 in the mid-sagittal plane	Inferior and middle point on C3	Inferior first visible point on C3
3.	C4 _ai	Most anterior–inferior point of the body of C4	Anterior–inferior point of the body of C4 in the mid-sagittal plane	Inferior and middle point on C4	Inferior first visible point on C4

Only gold members can continue reading. Log In or Register to continue

Stay updated, free dental videos. Join our Telegram channel

Tags: Bala Chakravarthy Neelapu, Kharbanda's Orthodontics: Diagnosis and Management of Malocclusion and Dentofacial Deformities, OP Kharbanda; Harish K Sardana; Rajiv Balachandran; Karthik Sennimalai

May 10, 2026 | Posted by drzezo in Orthodontics | 0 comments

Pocket Dentistry

Fastest Clinical Dentistry Insight Engine

3D volumetric analysis and clinical implications of the upper airway and sinuses

Introduction