INTRODUCTION

A plethora of environmental and genetic factors are involved in a complex interplay controlling maxillofacial morphological growth. Thus, in clinical orthodontics, an understanding of maxillary and mandibular growth is very important.1 Occlusal pressure generated by the masticatory muscles influences mandibular size and morphology, notably affecting mandibular body length and ramus height.2 It has also been reported3 that the inclination of the masseter is involved in the mesiodistal growth of the mandible. Furthermore, the direction of growth is affected not only by the occlusal force but also by environmental factors such as orthodontic orthopedic treatment,4,5 parafunctional habits,6 and functional malocclusion.7 The involvement of genetic factors is illustrated by the similarity in facial growth patterns within families and the apparently race-dependent variation in the frequency of specific facial types within a population.1

Recent advances in diagnostic imaging techniques have facilitated accurate three-dimensional computed tomography (CT) of maxillofacial morphology. The application of CT is limited in the dental field because the exposure dose is high and resolution in the axial direction is low. Cone-beam computed tomography (CBCT) for dental and maxillofacial use is useful because of its shortened scan time and high-resolution images, especially in the longitudinal direction. CBCT is used in various fields, including oral surgery, implant treatment, endodontics, periodontal treatment, treatment of temporomandibular joint disorder, pediatric dentistry, and orthodontics, and its benefits have been well documented.8–10

Increases in mandibular length, gonial angle, and mandibular plane angle are all characteristic maxillofacial morphological features of mandibular prognathism.11 A significant difference was noted between the volumes of the bilateral heads of the mandible in patients with facial asymmetry.12,13 Regarding the relationship between mandibular volume and lateral cephalogram-based skeletal classification, an inverse correlation was noted only in ANB in dried Indian skulls,14 and no significant difference was noted among the skeletal classes in Japanese females,15,16 although the sample numbers in these reports were small. Moreover, neither the association between male mandibular volume and skeletal classification nor the correlation between mandibular volume and parameters of cephalometric analysis has been evaluated.

The objective of this study was to investigate the association between values derived from lateral cephalometry and the mandibular volume calculated from CBCT images.

MATERIALS AND METHODS

CBCT images and cephalograms were obtained from 58 Japanese men and 60 Japanese women. The subjects were patients at Showa University Dental Hospital. Those with congenital disorders (eg, cleft palate) or systemic disease were excluded from the sample. When the distance between the anterior nasal spine and the tip of the mental spine was 5.0 mm or greater in the axial image on CBCT multiplanar reconstruction, the morphology was judged as showing asymmetry, and the patient was excluded. This study was approved by the Showa University Dental Hospital Ethics Committee and related committees.

To analyze the craniofacial morphology, measurements were made on tracings of lateral cephalograms under standardized conditions and then computed with Power Cephalo software (ReazaNet, Tokyo, Japan). To evaluate the mandible, we measured four parameters for angle analysis (anteroposterior evaluation of the mandibular basal arch relative to the skull base [SNB angle]; anteroposterior relationship between the maxilla and mandibular basal arch [ANB angle]; mandibular plane angle [Mp]; and gonial angle [Go]; all measured in degrees) and three items for distance analysis (Gnathion-Condylion [Gn-Cd]; Pogonion′-Gonion [Pog′-Go]; and Condylion-Gonion [Cd-Go]; all measured in millimeters). The maxillofacial morphology was classified (based on ANB) into skeletal Classes I (−1° ≤ ANB < 4°), II (ANB ≥ 4°), and III (ANB < −1°). Nineteen males and 20 females were Class I, 20 males and 20 females were Class II, and 19 males and 20 females were Class III.

Images were acquired using a cone-beam X-ray CT system for dental and maxillofacial use (CB MercuRay, Hitachi Medico Technology, Tokyo, Japan) installed at the dental hospital radiology department of our University. Subjects were positioned with the Frankfort plane parallel to the floor. The scanning conditions were 100 kVp, 10 mA, F-mode 512 slices/scan (slice width: 377 µm), and 9.6-second acquisition time. Data obtained were reconstructed using Analyze three-dimensional reconstruction software (Biomedical Imaging Resource, Mayo Clinic and Foundation, Rochester, Minn).

In a preliminary experiment, CBCT was performed using an aluminum bar (30 × 30 × 100.5 mm) under the same imaging conditions to examine the errors of volume measurements (Figure 1). Volumetry was performed using Analyze image processing software. The window of the Analyze region of interest (ROI) was displayed, thresholds were specified to define the area within the contour of the aluminum bar in all horizontal slice images, and autotracing was performed. An object map was prepared by overlapping the autotraces, allowing construction of three-dimensional images and completion of volumetry analysis.

View Figure

• Download as Powerpoint
• Download Figure

For extraction of the mandible, we used Analyze image processing software (Biomedical Imaging Resource, Mayo Clinic and Foundation). The thresholds of the area within the perimeter of the mandibular cortex were determined on all slides, and the mandibular contour was autotraced. An object map was prepared by superimposing the autotraces to create a three-dimensional image, from which the volume was measured (Figure 2). Since separation of the upper and lower dental crowns is difficult as a result of artifacts, dental crowns and the mandible were separated, and only the mandible was extracted. One investigator evaluated all cephalograms and CBCT measurements.

View Figure

• Download as Powerpoint
• Download Figure

One-way analysis of variance was employed to compare the mandibular volume among the three classes of maxillofacial morphology. To investigate the relationship between the mandibular measurement parameters and mandibular volume, simple regression analysis and the Pearson's correlation coefficient test were performed, with the significance level set at 5%. To investigate operator errors, the mandibular volumes of five randomly selected samples were measured twice on different days, and the errors were tested using Dahlberg's formula. Statistical analysis was performed using Statcel2 statistical analysis software (OMS Publishing, Saitama, Japan).

RESULTS

The error (discrepancy between the values measured using image processing software and the actual volume) was 0.23%. The intraobserver measurement errors ranged from 1.3% to 2.4% for volumetric measurements. The means, standard deviations, and maximum and minimum values of the cephalometric and CBCT measurement parameters are shown in Table 1.

Table 1. Means and Standard Deviations (SDs) of the Measurements from Cone-Beam Computed Tomography (CBCT) and Lateral Cephalogram Analysis^a

View Fullscreen

Figure 3 shows a comparison of the mandibular volumes in the three maxillofacial morphological classes. No significant difference was noted between males and females in any of these groups (P > .05).

View Figure

• Download as Powerpoint
• Download Figure

The results of correlation analysis between the measured values of the mandibular morphology and mandibular volume are shown in Table 2. There was a significant correlation between mandibular volume and Go in male subjects (P < .05). Furthermore, analysis of cephalometric distances revealed significant positive correlations between mandibular volume and Pog′-Go in males and between mandibular volume and both Cd-Go and Gn-Cd in females.

Table 2. Regression Analyses between Mandibular Morphology and Volume^a

View Fullscreen

DISCUSSION

Our results showed no significant association between mandibular volume and skeletal classification and only a weak correlation with isolated cephalometric values.

The accuracy of CBCT was verified using an aluminum block in the preliminary experiment. The relative error of the values on CBCT from the actual volume was 0.2% (relative error [%]  =  [(value on CBCT − actual value)/actual value] × 100). Pinsky et al.17 investigated the accuracy of volume measurement by CBCT using an acrylic block and found the error to be 2% of the actual volume. The values recorded here are an order of magnitude lower, demonstrating the superior accuracy of our volume measurements using CBCT.

Separation of the upper and lower dental crowns is problematic owing to artifacts, so the dental crowns and mandible in the CBCT image were separated, and the was mandible extracted as reported by Ilknur et al.,13 Nair et al.,15 and Deguchi et al.16 Cortical bone comprising many lamellar structures forms the outer region of bone and was used here as the outer circumference of the mandible for our measurements. Cancellous bone has a trabecular pattern containing many reticular spaces in the inner region. The trabecular thickness is 100–150 µm. However, the voxel size of the acquired CBCT data was 0.376 mm, and the spatial resolution was 0.75 mm in an isotropic direction, making detection of trabeculae challenging. Thus, we calculated the total volume of the mandible, using the cortical bone to demarcate its perimeter. The mandibular volume was measured with the dental roots remaining in the mandible. The average mandibular volume was reported by Ilknur et al.13 to be 56.3 cm³. When subclassified into skeletal Class I and Class II relationships, these volumes were 60.0 and 51.6 cm³, respectively, according to a report from Hashiba,14 and 46.0 and 38.6 cm³ in two reports by Nair et al.15 and Deguchi et al.,16 illustrating appreciable variation between investigators. In the present study, the average mandibular volume was 63.2 cm³ in males and 55.9 cm³ in females. The voxel value on CBCT is not an absolute value, unlike the CT value on systemic CT, but it is a relative value representing bone density.18 When segmentation of an object is performed with different thresholds, the threshold values must be changed.19 For example, since the bone density is low and multiple bones overlap in the mandibular head region, segmentation is more difficult than in other regions.20 In this study, the mandible was displayed in the Analyze ROI window, and the thresholds for the area within the perimeter of the mandibular cortex were defined on all slides. The border between cortical bone and soft tissue was clear, and the perimeter of the mandibular cortex could be traced.21

Hashiba14 reported that the mandibular volume was significantly larger in skeletal Class I than in Class II, whereas Deguchi et al.16 observed no significant differences among the maxillofacial morphology classifications in 30 adult female Japanese. Our findings in 58 adult male and 60 female Japanese were similar to those reported by this latter study. To investigate whether specific cephalometric features are correlated with mandibular volume and maxillofacial morphology, we performed regression analysis, but we found no significant difference in the mandibular volume in relation to any of the angle analysis parameters representing the positional relationship, confirming the conclusions of Hashiba14 that the angles are not directly correlated to the size or volume of the mandible.

Hashiba also reported a significant positive correlation between Ar-Go (which represents the ramus length on lateral cephalography) and total mandible volume but also that the ramus length was most highly correlated to the mandibular body volume, whereas its correlation with the ramus volume was weaker.14 We observed a positive correlation between mandibular volume and certain cephalometric distance parameters in both genders, but the correlation coefficients were only 0.01–0.44, indicating that the correlation was weak. Thus, we conclude that values derived from two-dimensional figures cannot adequately substitute for values from three-dimensional images.

The craniofacial region is one of the most complex regions of the human skeleton and is thus challenging to analyze. Cephalograms are used to describe the morphology and growth of the craniofacial complex and to predict growth, develop treatment plans, and evaluate treatment outcomes. Their main drawback is radiographic projection error, which can magnify or distort the geometric relationships displayed in the image.22 Reliable and accurate evaluation is difficult because of this inherent geometric magnification, distortion, and superimposition of craniofacial structures.23–27 The use of cephalometry as an adjunct to clinical examination as a basis for predicting craniofacial growth is thus questionable.28,29 Imaging is increasingly favored for perceiving the size and form of craniofacial structures in orthodontics and has the advantage of replicating anatomical detail accurately in three dimensions.

The mandible is a complex three-dimensional structure.30 Our research suggested that cephalometric analysis as a perspective chart cannot represent a real mandible. The mandible changes its morphology depending on its dynamic environment and forms trabeculae to achieve maximum strength with minimum material according to biodynamic principles.31 In addition to conventional two-dimensional and serial cephalography, the study of changes in the mandibular volume could be invaluable in exploring the mechanisms that control mandibular morphogenesis during the growth process and in evaluating the effects of orthopedic and orthodontic treatments. If quantitative measurements can be made in a larger number of dimensions, the evaluation of the relationship between quantitative and qualitative changes may be expedited, contributing to better clinical diagnosis and treatment selection.

CONCLUSIONS

• Mandibular volumes among skeletal Classes I (−1° ≤ ANB < 4°), II (ANB ≥ 4°), and III (ANB < −1°) are not significantly different.

• A significant (albeit weak) correlation was noted between mandibular volume and both cephalometric distance parameters and the gonial angle.