Strategies for Segmenting the Upper Airway in Cone-Beam Computed Tomography (CBCT) Data

Abstract

The wide availability, low radiation dose and short acquisition time of ConeBeam CT (CBCT) scans make them an attractive source of data for compiling databases of anatomical structures. However CBCT has higher noise and lower contrast than helical slice CT, which makes segmentation more challenging and the optimal methods are not yet known. This paper evaluates several methods of segmenting airway geometries (nares, nasal cavities and pharynx) from typical dental quality head and neck CBCT data. The nasal cavity has narrow and intricate passages and is separated from the paranasal sinuses by thin walls, making it is susceptible to either over-or under-segmentation. The upper airway was split into two: the nasal cavity and the pharyngeal region (nasopharynx to larynx). Each part was segmented using global thresholding, multi-step level-set, and region competition methods (the latter using thresholding, clustering and classification initialisation and edge attraction techniques). The segmented 3D surfaces were evaluated against a reference manual segmentation using distance-, overlap-and volume-based metrics. Global thresholding, multi-step level-set, and region competition all gave satisfactory results for the lower part of the airway (nasopharynx to larynx). Edge attraction failed completely. A semi-automatic region-growing segmentation with multi-thresholding (or classification) initialization offered the best quality segmentation. With some minimal manual editing, it resulted in an accurate upper airway model, as judged by the similarity and volumetric indices, while being the least time consuming of the semi-automatic methods, and relying the least on the operator's expertise.

Three-dimensional structure of the basal lamella of the middle turbinate

“…We are not aware of any other 3DSlicer-based surgically oriented studies of the middle turbinate’s anatomy. However, diverse workgroups already reported on this software’s feasibility to analyze the paranasal sinuses and the skull base for anatomical 17 , 18 , biophysical 19 , 20 or forensic 21 purposes. Raappana et al described a 3DSlicer-based planning workflow for transnasal pituitary surgery, focusing on the relevant parasellar anatomy 22 .…”

A machine-learning-based method for automatizing lattice-Boltzmann simulations of respiratory flows

“…Obviously, not only increases the H U value difference at locations with strong H U gradients, but also introduces high H U fluctuations off the interface. Applying a standard threshold of = −550 H U [32] to this filtered data set leads to false segmentations. This is, e.g., the case in the vicinity of ι = 85 or ι = 195 in Fig.…”

Three-Dimensional Image Segmentation of Upper Airway by Cone Beam CT: A Review of Literature

“…These groups outline the portions of the image with intensities corresponding to air, tissue and bone. (18) Kabaliuk N. et al (18) reported that multithresholding is unreliable for CBCT data, as noise is likely to gather in the image and cause oversegmentation. This is particularly true for bones that contain multiple inner structures, such as the nasal cavity, which can lead to extensive variation of the bone pixel intensity.…”

“…also affects performance, with metal error intensity levels dropping into the thresholding range of bone and appearing in CBCT images as pronounced, "star-like" streaks. (10,20) As the nasal cavity has several complicated, narrow passageways and adjacent paranasal sinuses, it is particularly vulnerable to either undersegmentation (true borders between tissue and air are unnoticed by the segmentation algorithm, and discrete air-filled holes are fused into one) or oversegmentation (the algorithm yields false borders which are not anatomically present) (18) . Therefore, the range of the gray threshold should not be fixed, neither for the pharyngeal airway nor for specific software, because the pixel grey value depends not only on the tissue contrast but also on additional factors.…”