Using CT Scans to Assess Early Postnatal Development of the Bony Eustachian Tube and Hyolaryngeal Apparatus: A Test of Reliability

Abstract

Human infants, like non‐human primates, can breathe and swallow simultaneously, due to similarities in the anatomy of the aerodigestive tract (ADT). Morphometric studies lack standardized methods to investigate the descent of the hyolaryngeal apparatus in relation to the skull base and bony Eustachian tube – a process taking place throughout the first three years of postnatal life. This study tests the hypothesis that 3D landmark coordinates taken on sagittal CT scans can be reliably used to assess ADT and craniofacial structures via 3D geometric morphometrics. We used the computer program Osirix® to identify 14 landmarks on head and neck sagittal CT scans obtained from 22 monkeys (Macaca mulatta) as a model to test accuracy. After four training sessions on images apart from the study sample, two observers assessed the study sample. Each observer performed three assessments per specimen and no specimen was evaluated twice in the same day. We used Generalized Procrustes Analysis, involving optimal superimposition and scaling by Centroid Size, and measured the procrustes distances to interpret data. Inter‐observer errors were significantly high (> 5%), indicating that plotting landmarks on sagittal CTs allows for low reproducibility of measurements via 3D analysis. ANOVA of inter‐observer distances per landmark yielded a significant result (p<0.0001), indicating that some landmarks are less reliable, and a series of t‐tests with Bonferroni‐Dunn adjustment revealed that the craniometric point Prosthion was the least accurate (p<0.05). Tracking the descent of the hyolaryngeal complex relative to craniofacial structures in early postnatal life requires a careful methodological approach. This study indicates that using 3D landmarks obtained from CT sagittal slices to ascertain measurements of the hyoid bone and craniofacial structures situated in different planes yields intrinsic errors of measurements that may jeopardize reproducibility of geometric morphometric analysis. Next step of this project is to use volume rendered 3D CT scans and collect coordinates for the same landmarks to determine interobserver variability, assess accuracy and validate reproducibility.Support or Funding InformationIcahn School of Medicine's Summer Student Investigator Program Award/CNPq ‐ Brazilian National Counsel of Technological Development