RESUMO
Otitis media (OM) is among the most common of childhood illnesses. It has long been hypothesized that children under age two are predisposed to OM due to differences in the anatomy of the Eustachian tube (ET), including the angle of the ET. OM in later childhood is less common but does occur, begging the question, are there shape differences in the ET that persist underlying later occurrences of OM? To answer this question, a novel method, which applied geometric and morphometric shape analysis to landmarks obtained from MRI data, was used. MRI scans were performed on 16 children (5 control, 3 cOME, and 8 rAOM) between 2011 and 2015. Sixteen landmarks representing the shape of the ET, cranial base, and palate were analyzed. The results of a Procrustes ANOVA indicate that the shape of the ET varies significantly (p < 0.01) between the OM and control groups. The shape differences between the OM group and the control are a medial and low attachment site of the tensor veli palatini (TVP) muscle, a posterior and high torus tubarius, and an anteriorly projected palate. These results support previous findings that a relatively horizontal ET is associated with a predisposition for OM. This study used a novel approach to examine anatomical differences in children with and without OM. First, the data set is unique in that it includes MRI scans of children with a confirmed OM diagnosis. Second, the use of MRI scans in craniofacial anatomy OM research is novel and allows for the collection of soft tissue landmarks and the visualization of soft tissue structures. Third, geometric morphometric shape analysis is a statistical method that captures shape differences, offering a more universal picture of nuanced changes within the entire set of landmarks, in contrast to more traditional linear and angular measurements used in prior OM studies examining craniofacial anatomy.
RESUMO
OBJECTIVES: Past studies using traditional morphometric approaches have reported a handful of differences in craniofacial dimensions between individuals with and without otitis media (OM). In this study, a geometric morphometry (GM) approach was used to determine if craniofacial shape is different among children with no history of OM and a history of recurrent acute OM (RAOM) at two different ages. METHODS: Nineteen standard landmarks were identified on lateral cephalometric radiographs from 79 children (41 Control, 38 RAOM) at 4 years and 52 children (27 Control, 25 RAOM) at 6 years of age. Following Procrustes superimposition of the landmark coordinate data, comparisons of group differences in overall size and shape were performed. Discriminant function analysis and principal component analysis were used to determine which, if any, aspects of shape variation distinguished RAOM from Control groups. RESULTS: At 4 years of age, craniofacial size and shape were significantly different between RAOM and Control groups (p<0.05). Shape differences were evident in the relative positions of the mandible, cranial base, external acoustic meatus, sphenoid and palate. Those shape differences were not found in the 6-year old group. CONCLUSIONS: At 4 years of age, the RAOM and Control groups have distinct craniofacial morphologies, but by 6 years of age these differences have largely disappeared. This is consistent with the clinical observation that excess RAOM risk resolves around 6 years of age and the hypothesis that this resolution is partially a result of age-related craniofacial changes.
