Novel 3D segmentation for reliable volumetric assessment of the nasal airway: a CBCT study.

OBJECTIVE: Nasal airway obstruction affects up to one-third of Americans and is one of the most common complaints by patients to otolaryngologists. Nasal airway obstruction and obstructive sleep apnea syndrome (OSAS) are closely related. The aim of this study was to use the 3D imaging software, ITK-SNAP as a platform to define a gold standard for anatomically accurate boundaries of the nasal airway in 3D CBCT and to create a more reliable and precise 3D CBCT segmentation of the nasal airway for assisting diagnosis, treatment, and monitoring of nasal airway obstruction and OSAS. METHOD AND MATERIALS: After review of the literature to identify established parameters using CBCT and CT technology for the segmentation of the nasal airway, and the existing drawbacks, a gold standard for locating the anatomical boundaries of the nasal airway using CBCT is proposed. This new method aims at standardization of segmentation and quantification, allowing for more reliable comparison between studies. ITK-SNAP software was used to segment three CBCT samples of healthy patients aged 21 to 59 years, who were patients of record, with CBCT obtained for either orthodontic, endodontic, or prosthodontic treatment planning purposes.
Results: The literature search identified 11 studies describing nasal airway parameters utilizing CBCT and CT. A great variation was detected on where the anatomical boundaries for the nasal airway were selected. A new standard in the identification of anatomical boundaries of the nasal airway is proposed for consistent segmentation and quantification using 3D CBCT by using the following landmarks: the inferior ANS-PNS border, the anterior nares border, the posterior sella-PNS border, and superiorly the border in alignment with the base of the skull (excluding the ostia, frontal, ethmoidal, and sphenoidal air cells). The three segmented samples were volumetrically measured, and statistically analyzed. The mean average Hounsfield unit intensity using the CBCT samples in this study was 629 with a standard deviation of 190.
Conclusion: The literature indicates a lack of a gold standard using CBCT technology for the segmentation of the nasal airway. With the proposed standard in this study, it is possible to quantify the nasal airway volume and thereby its reduction. For the general dental practitioner, this is an important aspect during the evaluation of overall airway assessment. This information can be useful in the diagnosis and treatment of airway compromised dental patients. (Quintessence Int 2021;52:154-164; doi: 10.3290/j.qi.a45429).

Artificial Intelligence for the Computer-aided Detection of Periapical Lesions in Cone-beam Computed Tomographic Images.

INTRODUCTION: The aim of this study was to use a Deep Learning (DL) algorithm for the automated segmentation of cone-beam computed tomographic (CBCT) images and the detection of periapical lesions. METHODS: Limited field of view CBCT volumes (n = 20) containing 61 roots with and without lesions were segmented clinician dependent versus using the DL approach based on a U-Net architecture. Segmentation labeled each voxel as 1 of 5 categories: "lesion" (periapical lesion), "tooth structure," "bone," "restorative materials," and "background." Repeated splits of all images into a training set and a validation set based on 5-fold cross validation were performed using Deep Learning segmentation (DLS), and the results were averaged. DLS versus clinical-dependent segmentation was assessed by dichotomized lesion detection accuracy evaluating sensitivity, specificity, positive predictive value, negative predictive value, and voxel-matching accuracy using the DICE index for each of the 5 labels. RESULTS: DLS lesion detection accuracy was 0.93 with specificity of 0.88, positive predictive value of 0.87, and negative predictive value of 0.93. The overall cumulative DICE indexes for the individual labels were lesion = 0.52, tooth structure = 0.74, bone = 0.78, restorative materials = 0.58, and background = 0.95. The cumulative DICE index for all actual true lesions was 0.67. CONCLUSIONS: This DL algorithm trained in a limited CBCT environment showed excellent results in lesion detection accuracy. Overall voxel-matching accuracy may be benefited by enhanced versions of artificial intelligence.