Assessment of the Mandibular Condyles Using Cone Beam Volumetric Tomography
- 1. Department of Health and Behavioural Sciences, The University of Queensland, Australia
Abstract
Objectives: (1) To examine the horizontal condylar angle (HCA) at different heights through the condyle using cone beam volumetric tomography (CBVT). (2) To investigate the relationship between the HCA and degenerative joint disease (DJD), occlusion, condylar dimensions, joint spaces, age, and gender.
Study design: CBVT volumes of 264 joints were orientated according to a specific protocol. The left and right joints were examined independently. A mixed-effects model was used for statistical analysis.
Results: Significant differences were found between the mean HCA measured at different heights (p<0.001). The mean HCA at 4.0mm, 6.0mm, and 8.00mm were 68.30°, 64.51°, and 61.30°, respectively. A significant difference was found between the right and left sides (p<0.001). No other significant correlations were identified. Conclusions: The HCA is dependent on measurement height. Results from this study may be useful as a comparable reference for future studies correlating with clinical findings.
Keywords
• Cone beam
• Mandibular condyle
• Temporomandibular joint
INTRODUCTION
Temporomandibular (TM), joint imaging forms an integral part of the assessment and management of TM joint dysfunction/ disorder (TMD). The significance of radiographic features associated with degenerative joint disease (DJD), and its relationship to clinical symptoms has been an area of interest [1- 6].
Plain film imaging and tomography have been traditionally used to evaluate TM joint disorders. However, these methods rely heavily on positioning, radiographic projection, and operator technique [2,7-11], The detection of degenerative osseous changes and the interpretation of plain film studies have been reported to be suboptimal [6,12-16]. Previous studies have used a myriad of landmarks for linear and angular measurements, whereby some reference points may be influenced by beam angulation and patient positioning during image acquisition [10,17-20]. Although useful as baseline assessment of TM joint disorders, the interpretation of plain film imaging and tomography is also compromised by anatomical superimpositions. Computed tomography (CT), has shown promising results in depicting the osseous components of the TM joint, with high accuracy and excellent inter-observer reliability reported [6,21,22]. However, this imaging modality is higher in cost, more difficult to access and higher in radiation compared to conventional imaging [23,24].
The use of magnetic resonance imaging (MRI), for the detection of osseous changes within the TM joint has been controversial, as the reported accuracy and interpretation reliability has been variable between different studies [6,21,25-27]. On the other hand, MRI assessment of soft tissue components of the TM joint has been shown to be excellent [6,21,26-28].
Cone beam volumetric tomography (CBVT) has many applications in dentistry and is becoming invaluable due to its relatively low radiation dose and high spatial resolution compared to multi-detector CT [24,29-32]. For the diagnosis of osseous changes related to DJD in the TM joints, mixed results have been presented with reported sensitivity ranging from 0.23 to 0.91, and specificity from 0.67 to 1.0 [23,33-35]. Linear and angular measurements obtained from CBVT datasets have been shown to be of slight underestimation, but the dimensional accuracy was deemed to be within acceptable limits [30,31,36- 41].
CBVT volumes can be manipulated and reconstructed in multiplanar views without any loss of resolution due to isotropic voxels [32], however orientation of cone beam volumes has not been standardised and tested for reliability. Lagravere et al. [42], raised the question that orientation of the datasets may directly affect the measurements derived from the reformatted images, and suggested several stable anatomical landmarks for plane orientation. Kim et al. [43], also emphasised the importance of midsagittal plane (MSP), determination especially for the quantitative assessment using three-dimensional datasets.
In addition to osseous changes within the TM joint, the horizontal condylar angle (HCA), has also been studied. The HCA refers to the angulation of the long axis of the mandibular condyle in relation to a reference plane. Several authors have explored the significance of this angle, however, the lack of clearly defined landmarks has resulted in difficulties in comparison between different studies. The transmeatal line [18,19,44], frontal/ coronal plane [22,45-47], and MSP [48-53], have all been used as reference planes. Possible correlations with DJD and occlusion have been suggested [47,52-55].
A review of literature revealed a lack of research investigating the longitudinal axis of the mandibular condyle in relation to the MSP at different levels using CBVT in the axial plane. This study aims to orientate the CBVT data sets in a systematic manner, and utilise stable osseous landmarks to provide a consistent and reproducible method for measurement and comparison. This study was designed to determine if the angulation of the HCA changes at different heights through the condyle, and whether there is a significant correlation between this angulation and DJD, occlusion, condylar dimensions, joint spaces, gender or age.
MATERIALS AND METHODS
Ethics approval was obtained from the University of Queensland Dental School Research Committee for this study.
De-identified CBVT datasets were obtained from a private radiology practice (MaxilloFacial Radiology). A total of 145 deidentified datasets (94 females, 51 males) in DICOM format from 213 randomly selected cases were included in this study. The purpose for scan acquisition, joint diagnosis and original radiology reports were not available to the investigators. Poor quality scans with obvious movement, excessive artefact or excessive noise were excluded from this study. Datasets of patients with craniofacial developmental anomalies including bifid condyles, obvious signs of previous TM joint trauma, ongoing orthodontic treatment, edentulous ridges or lack of posterior occlusal support were also excluded. A total of 264 joints (130 left and 134 right), were examined independently by assessor RL (Dento-Maxillofacial Radiology postgraduate student).
All CBVT scans included in this study were obtained using an i-CAT cone beam unit (Imaging Sciences International, Hatfield, PA). Scanning of the patients was carried out at 120 kilovolts (kV), and 5 milliamperes (mA). The resultant volume was 232 millimetres (mm) in width and 170 mm in height, with a total of 324 basis images at 0.4mm voxels. Radiographic assessment and measurement Each CBVT DICOM dataset was imported into and viewed using a DICOM Viewer (Planmeca Romexis Software). Images in the axial, sagittal and coronal planes were generated within the program. All slices were calibrated to 1.0mm thickness spaced at 1.0mm intervals.
Orientation of each volume was carried out according to a specific protocol by Lagravere et al. [42], in the coronal, sagittal and axial views using identifiable and stable landmarks as described previously. In the coronal plane, the line drawn between the right and left superior-lateral border of the external acoustic meatus (r-SLEAM and l-SLEAM) was orientated parallel to the true horizontal plane (Figure 1). In the sagittal plane, the dataset was orientated parallel to the horizontal plane using the Frankfurt plane (Figure 2). A modified technique from those described by Lagravere et al. [42], was used to determine the MSP. In the axial plane, the line drawn between the foramina spinosum (ELSA), was orientated parallel to the horizontal plane. A midpoint bisecting this line joined to the anterior nasal spine (ANS), was used to represent the midsagittal plane (MSP) (Figure 3).
The right and left sides of each patient were examined separately. Angle’s molar relationship (Class I, Class II and Class III occlusions) was classified from reconstructed sagittal images (Figure 2). Any side with one or both missing first molars were excluded from this study.
Each joint was evaluated according to the method described by Alexiou et al. [56]. Joints with degenerative changes were characterised by the presence of condylar flattening, sclerosis, erosions, osteophytes or subchondral cysts as viewed in the coronal and sagittal planes [23,56,57]. To avoid misinterpretation,only osseous changes evident in two different planes or on two consecutive slices were recorded by assessor RL.
In the sagittal and coronal planes, the most superior point of the right condyle (SC) on the right side was identified. Axial slices at 4.0mm, 6.0mm and 8.0mm inferior to the SC were generated (Figure 4). For each height, a line drawn from the most medial and lateral poles of the mandibular condyle was used to represent the longitudinal axis of the mandibular condyle. The horizontal condylar angle (HCA), was defined as the angle made by the longitudinal axis of the mandibular condyle and the MSP (Figure 5). The areas of interest in the axial and corrected sagittal views were magnified for measurement. Angular measurements were made using the Romexis software where the angulations are displayed in degrees to the nearest 0.01°. For each joint included in the study, one HCA for each height was recorded (HCA-4, HCA6, HCA-8).
Oriented volumes were transferred to the TM joint screen. Condylar dimensions were measured for both the right and left sides. In the axial view, the maximum mediolateral (ML), dimension was determined by the line connecting the most medial and most lateral points of the condyle. The median anteroposterior (AP) dimension was the line bisecting and perpendicular to the maximum mediolateral diameter (Figure 6). Linear measurements are displayed in millimetres within the computer software to the nearest 0.01mm.
One mid-condyle corrected sagittal (longitudinal) view per side through the temporomandibular joint was generated by the examiner. The anterior (AJS), superior (SJS) and posterior joint spaces (PJS) were measured as described by Ikeda and Kawamura [58] (Figure 7).
Randomly selected volumes (10% of useable data), were re-measured at two additional time points within a six-month period. All collected data was tabulated in spreadsheet form.
Statistical analysis
A mixed-effects model was used instead of simple linear regression, where an additional parameter known as the random effect was included to account for the correlation of measurements within each patient. An initial mixed effects model was created to predict the HCA using 11 primary variables (Height, Side, AP, ML,AJS, SJS, PJS, DJD, Occlusion, Age and Gender). A number of mixed effects models were also built in order to explore the predictive factors for DJD, condylar dimensions, and joint spaces. Correlation of measurements (HCA, AP, ML, SJS, AJS, PJS) between the two sides within the same patient was examined. Investigation of the differences between patients with different Angle’s class molar relationships was also carried out. For the analysis of intra-examiner measurement errors, Intra-Class Correlation Coefficient (ICC) was used for all continuous variables (condylar height, dimensions, joint spaces).
RESULTS
Horizontal Condylar Angle (HCA)
Statistically significant differences were found between the HCAs measured at different heights (p0.05). The mean HCA in Class III occlusion was significantly larger than those in Class I and Class II occlusions (p=0.02). Class I and Class II occlusions did not differ significantly (p>0.05). The median AP dimension of the condyles increased with increasing HCA (p0.05) were found between the HCA and: joint spaces, age, and gender. A statistically significant difference was observed in the mean HCA between the right and left sides, where the right side was found to be larger by approximately 3 degrees (p<0.001).
Condylar Dimensions
The relationships between joint dimensions and other study variables are summarised in Tables 2 and 3. The median AP and maximum ML dimensions were smaller in joints with DJD compared to radiographically normal joints (AP: p=0.001, ML: p=0.03). Males had significantly larger maximum ML dimensions (p<0.001). The maximum ML dimension increased with age. No statistically significant differences were found between the right and left mean maximum ML dimensions (p=0.74) and mean median AP dimensions (p=0.40).
Joint Spaces
Overall, the mean SJS was greatest, followed by the AJS and PJS (Tables 4, 5 and 6). Males had significantly larger SJS compared to females (p<0.001). The mean PJS on the right side was larger than the left (p=0.02). The mean AJS on the right side was smaller than the left (p=0.04).
Other
The results confirmed that DJD is an age-related condition, where the incidence increases with age (p<0.001).
Reproducibility
In this particular study, reproducibility was defined as the possibility of achieving the same results if the same observer and the same method are used to collect multiple sets of results. For the assessment of intra-rater reliability for all continuous variables, Intra-Class Coefficient (ICC), analysis showed an extremely high level of reproducibility (Table 7).
DISCUSSION
Horizontal Condylar Angle
The HCA measured at different heights was shown to be statistically significant in the current study. Angulations were taken at specified heights from the most superior point of the condyle so the measurements could be compared consistently between different joints. From the results obtained, it is reasonable to expect that the angles between studies can vary significantly if the landmarks are not standardised, or if the orientation planes are altered, resulting in measurements being taken at different condylar heights, especially between different imaging modalities. The mean HCA values obtained from this cohort for all joints (61.30-68.30°) lie within the range described by Yale59. Results from the study by Sato et al. [45], revealed slightly higher values but this was thought to be due to an ethnicity difference. For normal joints, the mean HCA values (65.11°), were comparable with previous studies [18,19,46-48,50-52,54,55,60].
Joints with DJD in this study were found to have a slightly smaller mean HCA (62.86°) compared to normal joints, but this was not found to be statistically significant. This finding agrees with previous SMV studies which also assessed osseous changes radiographically [19,45]. On the other hand, two MRI studies found significant differences between the two groups where joints with DJD had a significantly smaller mean HCA [54,55]. Inconsistencies in the results between studies using different imaging methods may be explained by the lack of a comparable measurement and/or observational standard. Angle measurements taken from CT and MRI images are from arbitrary transverse sections, and therefore results from these imaging methods against SMV radiographs and CBVT images may differ significantly. Westesson et al. [54], suggested that proliferative bone remodelling associated with DJD may result in a change in HCA. Kurita et al. [55], also postulated that resorption of the lateral pole of the condyle may be associated with DJD and subsequently a decrease in HCA, however the exact mechanism was unknown. The absence or presence of DJD signs in this study were classified according to Alexiou et al.[56], however the severity and type of degenerative changes were not recorded specifically and evaluated as part of this investigation. This may have an effect on the comparison between normal joints and joints with DJD, as those with subtle osseous changes and joints with moderate to marked changes may differ in their HCA, but the distinction may be masked in this particular study where they were placed within the same category.
Limited research is available on the relationship between HCA and Angle’s classes of occlusion. The mean HCA values reported in the studies by Rodrigues et al.[52,53], were marginally higher (Class I: 69.96-70.10°, Class II: 67.43-67.80°, Class III: 71.25- 72.88°), than those observed in the present cohort. Angular measurements from the present study were averaged from three height measurements for each joint, and therefore may explain the differences in these figures. Similar to the trend described by Rodrigues et al [52,53], the mean HCA for Class II occlusion in this study was found to be slightly smaller compared to Class I occlusion, whereas the mean HCA for Class III occlusion was found to be significantly larger than those in Classes I and II. The explanation for this observation is unclear, however Katsavrias and Halazonetis [61], suggested that functional loading may be vary for different classes of occlusion, especially for Class III where condyle and fossa shapes are dissimilar to Class I and II occlusions, and thus the condylar angulation
The relationship between the HCA and condylar dimensions has not been explored previously. In the current study, an increase in the median AP dimension of the mandibular condyle was found to be associated with an increase in HCA. The association between maximum ML dimension and HCA was found to be weak but statistically significant. Although statistically significant correlations were found in this study, the clinical significance of these findings is yet to be elucidated.
No statistically significant correlations were found between HCA and joint spaces in the present study. At the time of writing, no known research has been carried out to investigate the direct relationship between HCA and joint spaces. However, several studies have examined the relationship between HCA and internal derangement (ID) [18,54,55,60]. ID was not specifically assessed in the current study, however joints with ADD are associated with increased AJS in the closed position as the articular disc is more anteriorly placed between the mandibular condyle and the articular eminence. Therefore, results from ID studies in relation to the HCA may be extrapolated. Findings from the MRI studies by Westesson et al. [54], and Kurita et al. [55], suggested that anterior disc displacement (ADD), is associated with smaller HCA. On the other hand, results from the studies by Westesson and Liedberg [18], and Sulun et al. [60], did not reveal any significant mean HCA differences between normal joints and those with ADD. Differences in these results from may be due to the use of different transverse planes for angular measurement.
No significant associations were observed between HCA and age or gender in this study. This is consistent with findings from previous studies by several authors [45,46,54,60,62].
Condylar Dimensions
For normal joints, the mean maximum ML dimension were comparable to those reported by other CBVT [36], and CT [46,47], studies, however other studies using SMV [11,19], and direct visual assessments [63], reported marginally higher figures. Differences in the results could be explained by slight underestimation of measurements from reconstructed CBVT [39- 41], and CT [22], images. The mean median AP dimension found in the current study was similar to that reported by Krenkel and Grunert [11], where the AP dimension was also measured at the mid-condylar point. This measurement was distinctly smaller than the maximum AP dimension reported in the other studies [36,63], and this disparity may be explained by the difference in measurement landmarks. The shape of the condyle varies greatly in the axial plane [19], and therefore the median AP measurement would not necessarily correspond to the maximum AP dimension.
Mean ML and AP dimensions of joints with DJD in the present study were found to be significantly smaller than those that were radiographically normal. Kurita et al. [57], also found smaller ML dimensions in joints with resorptive defects as a part of DJD. On the other hand, Ebner et al. [19], found no differences between normal joints and those with early or moderate to marked osseous changes radiographically. Both resorptive and proliferative changes are seen in DJD, and therefore the disease stage may influence the significance of dimensional measurements. Joints with early resorptive changes are more likely to be associated with smaller joint dimensions, whereas those with more marked proliferative changes may have larger joint dimensions due to articular surface flattening [57]. Specific types of degenerative changes were not analysed within this study, therefore it was difficult to determine the exact association between condylar dimensions and DJD. Findings from the current study may be partly attributed to a higher number of joints with resorptive degenerative changes in this particular cohort than those with proliferative changes. Further imaging studies using subclassifications of degenerative joint changes may be of value in assessing the significance of joint dimensions in relation to DJD.
The maximum ML dimension was found to increase with age in the present study. It is reasonable to expect the ML dimension to increase with age due to progressive flattening of the condylar articular surface and proliferative osteoarthritic changes over time, as DJD is an age-dependent disease [56,64]. Christiansen et al. [46], found no correlations between condylar dimensions and age in a cohort with normal joints, but these authors mentioned a positive correlation with age in diseased TM joints in a previous study [65]. Results from the study by Ishibashi et al. [63], observed a trend of gradual decreasing condylar dimension especially after the fifth decade. Comparisons with results from previous studies are challenging as different inclusion criteria were used. Differences in study findings may be explained by the assessment of different age cohorts and also joints of different disease status.
Results from the current study revealed that males had significantly larger maximum ML dimensions. This finding was also reported by Christiansen et al. [46], and Goldman and Taylor [66], however two other studies with Asian subjects found no significant differences in condylar dimensions between genders [63,67]. The reasoning behind these findings is unclear, however in a study using human skulls, Hinton [68], found males to have consistently larger craniofacial dimensions and differences were also found between various ethnic groups. Hinton suggested possible genetic influences and also functional demands during growth as contributing factors to these observations [68].
Joint Spaces
Overall, the mean SJS was found to be significantly larger than PJS and AJS. This is consistent with figures reported in previous studies [46,58,69-72]. No significant differences were found between the overall mean AJS and PJS. The values observed for SJS, PJS and AJS for the group with DJD in this study were smaller than those reported by other studies [66,70]. This difference may be attributed to the fact that only the osseous changes related to DJD was assessed in the present study and not disc displacement. No correlations were observed between any of the joint spaces and DJD. A number of previous tomographic studies found no differences in joint spaces between asymptomatic and symptomatic joints, and that condylar positioning is widely variable in asymptomatic joints [8,71,73]. Pullinger et al. [8], suggested that radiographic analysis alone cannot accurately determine the presence or absence of disc displacement associated with ID. However, the study by Kinniburgh et al. [70], observed a significant difference in joint spaces between normal joints and those with ID. This may be explained by the fact that not all asymptomatic joints are radiographically normal, and not all symptomatic joints are radiographically abnormal [8,74]. Also, joints with anterior disc displacement do not necessarily equate to osseous changes within the mandibular condyle.
Males were found to have significantly larger SJS compared to females. This finding correlates well with those reported by a number of previous studies [70-72]. Kinniburgh et al. [70], theorised that larger SJS in males may be explained by thicker soft tissues within the joint. An overall difference in condyle and glenoid fossa between genders may also explain this observation [68].
Occlusion of the patients at the time of scan was unknown, however a recent CBVT study has shown that the difference in joint space measurements between centric relation (CR), and maximal intercuspation (MI), is negligible [75]. Using a sample of asymptomatic young adult patients with normal occlusion, and Angle’s classes of occlusion, these authors found that although there is a wide variation in condyle position in both CR and MI, the measurement differences were statistically insignificant.
Symmetry
Asymmetry in the HCA between the condyles found in the current study was also observed by several authors [48,59,66]. In two separate studies, HCA symmetry between the two sides was only found in 33% and 48% of the subjects [59,66]. Similar to the findings reported by several authors [11,52,53], no statistically significant differences were found in the mean maximum ML and mean median AP dimensions between the two sides. The mean AJS on the right side was found to be smaller than the left side, and the mean PJS on the right side was larger than the left. These results are in agreement with those reported by Cohlmia et al. [71]. This feature is possibly due to habitual posturing or favouring of one side during function, but the exact cause is yet to be determined [71,76]. It is reasonable to expect a degree of asymmetry within a patient in both normal and abnormal occlusions due to developmental asymmetry. The possibility of unilateral anterior disc displacement may also explain these findings. Although statistical significant differences were reported, Weinberg et al. [73], found no clinical significance from such results.
Reliability
High ICC values suggest that the intra-examiner error is low and the method of volume orientation is reliable and repeatable. The method of CBVT volume orientation used in this study may be considered as a baseline for developing a standardised reference for quantitative assessment.
Limitations
Correlation of radiographic features with clinical findings was not assessed in this retrospective study. Prospective studies examining the significance of the HCA in relation to clinical signs and symptoms in TM joint dysfunction is recommended. Changes in HCA within an individual over time in relation to the progression of DJD will require longitudinal studies for further assessment.
CONCLUSION
This is the first study to explore the significance of mandibular condylar angles measured at different heights through the condyle. Findings show that HCA is dependent on measurement height and may explain inconsistencies in the results from previous studies. HCA measurements from this study may be useful as a comparable reference for further CBVT studies correlating with clinical findings.
A trend of smaller mean HCA was observed in joints with DJD in this study, however further investigation using sub-classifications of DJD differentiating resorptive and/ or proliferative degenerative changes would be of value in further assessing the significance of HCA in relation to DJD and possibly the mechanisms involved. Positive correlations were identified between HCA and Class III occlusions, and HCA and condylar dimensions, however the clinical significance of these findings may be determined by prospective studies including clinical assessments. No significant correlations were identified between HCA and joint spaces. An assessment including disc status may provide further information regarding this finding as only osseous changes were assessed in the present study. No correlations were found between the HCA and age or gender.
ACKNOWLEDGEMENTS
In Memory of Dr Rebecca Liu who very sadly passed away before having this research published. Thanks also to Dr. Mark Griffin for assistance with the statistical analyses. Thanks also goes to MaxilloFacial Imaging for providing de-identified CBVT datasets for this research