The study protocol was reviewed and approved by the Institutional Review Board of Kaohsiung Medical University Hospital, Taiwan. This was a retrospective study conducted initially using 153 samples collected from patients who visited the Department of Orthodontics between January 2009 and April 2014. We included samples obtained from adults (mean age: 23.9 ± 2.72 years) with clear CBCT images. The exclusion criteria were a history of radiotherapy, oral and maxillofacial trauma/fracture, implant treatment, large restorations or amalgams, orthodontic treatment, and orthognathic surgery. A total of 67 cases were eventually selected. The assessments were focused on the mandibular molars and retromolar areas.
During the CBCT examinations (KaVo eXam; KaVo, Biberach, Germany), the participants maintained their heads in a natural position. The exposure settings were 26.9 s and 120 kV, the voxel size was 0.25 mm, and the field of view was 16 x 13 cm. Digital Imaging and Communications in Medicine (DICOM) data were measured using eXam Vision (KaVo eXam Vision; KaVo, Biberach, Germany). Plane A was defined as the occlusal plane and the Fiducial Line A was set as the occlusal line from the incisal edge of the mandibular incisor to the mesiobuccal cusp of the mandibular first molar. The right or left side was chosen randomly. The Fiducial Line B was a mesiodistal line that was equivalent to the average height of the alveolar crest of the mandibular first and second molars. The Fiducial Line C was the angle bisector of the axis of the mandibular first and second molars. Thus, even when the second molar tipped severely, the angle deviation would not be substantially affected. The Fiducial Line C was aligned with the direction of the apical movement in the axial view. Plane B was perpendicular to the Fiducial Line B and tangential to the most distal point of the lower second molar (Fig. 1). Measurements of the submandibular fossa and ridge width were manipulated in Plane B and moved along Line B.
In the first part of this study, we assessed the submandibular fossa. The submandibular fossa is a depression on the lingual surface of the body of the mandible inferior to the mylohyoid line. The most prominent superior and inferior points of the lingual concavity corresponding with the submandibular fossa in the para-axial slices were determined, and an initial line was drawn joining them. A second line was then drawn from the deepest point of the fossa perpendicular to the first line (Fig. 2). The slice in which the fossa was deepest was chosen to represent the maximum depth of the submandibular fossa. A definite submandibular fossa of depth less than 1 mm is considered an irregular anatomical structure. In addition, the boundary of the fossa was described relative to Plane B. The most superior and inferior points of the submandibular fossa were also determined to the average alveolar crest height of mandibular first and second molars. The length of the root was also measured from the average crest height to the root tip.
The second part of this study was based on Plane B and its parallel planes. The first step involved measurement of the total ridge width, which contains the buccal and lingual cortex and the alveolar housing, as well as the root width. These measurements were taken at the level of the average alveolar crest of the mandibular first and second molars and at the subcrest 2, 4, 6, 8, 10, and 12-mm levels. The measurements were repeated on parallel planes with distalization of 0.5, 1.0, 1.5, 2, 2.5, 3, 3.5, and 4 mm (Fig. 3). These measurements were also taken on the plane containing the widest root. All these planes were perpendicular to the Fiducial Line B.
The third part of this study evaluated the anatomical limit distal to the mandibular second molars (Fig. 4). Plane C was perpendicular to the Fiducial Line C and was at the level of the average crest height of the mandibular first and second molars. On Plane C, the available distalization distance was measured from the most distal part of the root to the nearest mandibular cortex. This distance was measured at the buccal and lingual sides. Likewise, the same process was performed on planes parallel to Plane C, with each plane 2 mm apical to Plane C. Along Line C, the measurements were performed at the subcrest 2, 4, 6, 8, 10, and 12-mm levels.
Statistical analysis was performed using SPSS for Windows (version 20; IBM, Armonk, NY, USA). The participants were classified by sex (male vs female), side assessed (left vs right), root-cortex contact, wisdom teeth (present vs absent), and malocclusion category. The chi-square test was used to compare the contact and non-contact condition, and the t-test was used to compare the ridge width and available distalization distance. The analysis of malocclusion category was performed using the Kruskal-Wallis test. Statistical significance was set at p < 0.01. The measurements were performed by one author and repeated 2 weeks later. The method error was examined using the Dahlberg formula . The measurement error for the total ridge width, alveolar housing width, and distalization distance were 0.12, 0.2, and 0.3 mm, respectively.