Development, Validation, and Application of a 3D print out of the Nasopharyngeal area from Cone Beam Computed Tomography in Children

Abstract

Cone Beam Computed Tomography (CBCT) is increasingly being used in dentistry to explore certain types of craniofacial abnormalities. Sleep-disordered breathing represents a significant burden to individuals and society. Hence, the opportunistic use of CBCT imaging among dentists to screen for upper airway obstruction has increased. Adenoid hypertrophy (AH) is one of the abnormalities linked to upper airway (UAW) obstruction that could lead to obstructive sleep apnea (OSA). The specific use of CBCT imaging to assess AH and the related nasopharyngeal space has been already proposed, but there is still a lack of agreement on its diagnostic value. Dolphin software is widely used in North American offices due to its friendly-user capabilities. This software has an upper airway analysis function which could be used to generate an STL model of a specific upper airway section. To our knowledge, 3D print-out depictions of the nasopharyngeal area has not been explored yet. This thesis project aims to develop and validate 3D depictions of the nasopharynx including different degrees of AH. The 3D depictions (3D-picture and 3D-prototype) were produced based on a pool of CBCT scans of patients with the nasopharyngeal area already examined via Nasoendoscopy (NE). The design and development of the 3D depictions of pharyngeal adenoidal obstruction examples included two different representations of the nasopharyngeal airway, a lumen depiction (LU) and an adenoid mass in relation to the lumen depiction(AD). The developed methodology showed excellent reliability, ICC = 0.982 (95% CI, 0.939-0.995) for LU and ICC = 0.995 (95% CI, 0.981-0.998) for AD. The generated 3D volume surfaces (LU and AD) were converted into STL files and distinct types of prototypes were fabricated (LU 3D and AD 3D). Otolaryngologist-head and neck surgeons (OHNS) evaluated the 3D depictions, the visualization consisted of LU and AD in 2D pictures and in 3D printed prototypes. Therefore four depictions were assessed in total, as follow: LU 2D depiction (surface picture), AD 2D depiction (surface picture), LU 3D depiction (prototype), AD 3D depiction (prototypes). One 3D depiction (LU 2D) failed to show validity as a subjective measurement; however, the other three 3D depictions (AD 2D, LU 3D, and AD 3D) were capable of suggesting validity as a subjective measurement. More precisely, the AD 3D depictions were adequate in detecting AH in this sample. High sensitivity and specificity were achieved 100% and 70%, as well as adequate positive predictive value (PPV) and negative predictive value (NPV) - 66% and 97% respectively. The developed 3D prototype may be a practical and readily available alternative for the assessment of the adenoidal obstructed area. It may also be useful in the future for educational purposes.