Cone-beam computed tomographic characteristics in degenerative temporomandibular joint disease patients with chewing side preference

This study is aimed at assessing the Cone-beam computed tomographic (CBCT) characteristics of temporomandibular joints (TMJ) in degenerative temporomandibular joint disease (DJD) patients with chewing side preference (CSP). CBCT images of 98 patients with DJD (67 with CSP and 31 without CSP) and 22 asymptomatic participants without DJD were measured retrospectively to compare the osteoarthritic changes and the morphology of TMJ. Quantitative analysis of the TMJ radiographic images was performed to present a comparison between the three inter-group groups and between the two sides of the joints. The frequencies of the articular flattening and surface erosion occur more often in the preferred side joints of DJD patients with CSP than the contralateral side. In addition, the horizontal angle of condyle, the depth of glenoid fossa (DGF), and the inclination of articular eminence (IAE) were larger in DJD patients with CSP than that in asymptomatic participants (p<0.05). Also, the condylar anteroposterior dimension of preferred side joints was significantly less than that of non-preferred side (p=0.026), while the width of condyles (p=0.041) and IAE (p=0.045) was greater. DJD patients with CSP appear to have a higher prevalence of osteoarthritic changes, with the morphological changes such as flat condyle, deep glenoid fossa, and steep articular eminence, which might be considered the characteristic imaging features. This study found that CSP is a predisposing factor for the development of DJD, and attention should be paid to the existence of CSP in DJD patients during the clinical practice.


Introduction
Degenerative temporomandibular joint disease (DJD) is the most common portion of temporomandibular disorders (TMD), involving the subchondral bone, cartilage, synovial membrane and other structures [1,2]. The main clinical manifestations of DJD are a series of symptoms and signs such as joint noise, pain, and limited mandibular movement. These clinical evidence of DJD was observed in approximately 35% of the population [3] and 8 to 16% of the TMDs patients [4][5][6]. DJD is characterized by joint remodeling, abrasion, and deterioration of articular tissue with concomitant osseous changes in the condyle, articular fossa, and eminence [1,7]. The etiology of degenerative changes in TMJ is complex, associated with the cooperation of multiple Xiaoting Zhai and Dongzong Huang have contributed equally to this work. factors, such as inflammation, trauma, or aging, more often due to extensive load [2,8]. Temporomandibular joint (TMJ) withstands the compressive pressure during mastication, although it is the most flexible joint. Compared to hyaline cartilage-covered joints, the TMJ condyle is covered by soft fibrocartilage and has a larger range of motion and therefore a greater potential for degenerative changes [9,10].
Chewing side preference (CSP) is a bad habit of chewing with unilateral dentition, observed in about half of the population [11]. Some studies suggested that CSP may act as a potential contributor to the development of TMDs [12][13][14] and determine the patterns of load distributions on the TMJs [15]. The presence of long-term unilateral mastication induces asymmetric overloading of TMJ [16], accelerating the impact of articular cartilage, subchondral bone, and synovium. Remodeling and reconstruction of articular cartilage and bone is a biological response to mechanical load, maintaining a functional balance between joint and occlusion, which may in turn aggravate joint load [17,18]. When excessive or prolonged overload exceeds TMJ endurance, degenerative joint disease may occur [2,19]. In short, CSP may contribute to functional overload in the unilateral joint, which may be related to the structural morphological changes in the TMJ, including cartilage, condyle, glenoid fossa, and articular eminence.
According to the recommended "Diagnostic Criteria for TMD (DC/TMD)" [7], radiographic examination is required for a definitive diagnosis of DJD. With the advantage of lower radiation dose and at the same time with high diagnostic value, cone-beam computerized tomography (CBCT) is considered to be the modality of choice for qualitative analysis of the osseous structure of the TMJ [17,20]. Previous studies have demonstrated the relevance of radiometric findings for TMD and CSP [18,[21][22][23]. However, it remains unclear whether CSP is one of the contributing factors to DJD and is associated with osteoarthritic changes, and whether these DJD patients with CSP have different radiological morphological characteristics of TMJ. At the same time, it remains to be determined whether CSP is only related to asymmetric loading and not to morphological remodeling. Therefore, this study is aimed at investigating the potential relationship between CSP and DJD by comparing the morphology of TMJ, including the condyle, articular eminence, and glenoid fossa, as well as the joint space and position of the condyle.

Participants
The patients who sought treatment of DJD and underwent CBCT examination in the Department of Stomatology at Chinese People's Liberation Army (PLA) General Hospital engaged in the study. Considering the effects of aging on bone morphology [10,24,25], the inclusive criteria were as follows: age between 16 and 45 years old, willingness to participate in the study, and patients who were diagnosed with DJD according to DC/TMD [7] but had not received treatment. The exclusion criteria were as follows: age less than 16 years and above 45 years, acute traumatic injury, temporomandibular joint ankylosis and oncology, rheumatoid arthritis, craniofacial syndromes, pregnancy, severe malocclusion, previous joint surgery, radiotherapy, and orthodontic treatment.
The presence of CSP was determined using a self-report questionnaire and a gum test [21,26]. Participants were asked to place a piece of chewing gum in the center of the dorsum of the tongue, and the position of the gum during the first mastication cycle was recorded. The results of the questionnaire and gum test were combined to determine the preferred side of chewing.
A total of 98 patients were enrolled in this study. Depending on the presence or absence of CSP, 31 cases diagnosed with DJD without CSP were assigned to the DJD group, and 67 cases presenting with both DJD and CSP were assigned to the DJDC group. Another 22 healthy volunteers without DJD and CSP were enrolled as a control group. The age, gender, and CSP composition of the three examined groups are listed in Table 1.
The sample size calculation for comparing groups was done using the software G-power, with α of 0.05 and power (1-β) of 0.80, and Cohen's effect size of each group from a pilot sample and previous articles [17,[21][22][23]. We performed the calculation for the imaging findings of flattening, erosion, subcortical sclerosis, osteophyte, and subchondral cyst. The necessary sample size (n) of each osteoarthritic change was 26, 46, 55, 37, and 17, respectively (Shows in supplementary material: SI. Table 1). In this article, we had a final sample size of 62 in the DJD group and 134 in the DJDC group.
All participants received and signed informed consent forms. The research was conducted with the ethical principles of the Helsinki Declaration and obtained the consent of the Medical Ethics Committee of Chinese PLA General Hospital (S2020-327-01).

Imaging procedures and measurements
All the 120 participants (240 joints) received CBCT examination conducted on a NewTom 3G flat panel-based CBCT machine. The parameters of the exposition were 100kV, 4mA, and an exposition time of 13s. The established field of view was 8 cm × 5 cm. During the CBCT scan, the participants were sitting upright with their heads erect and their mandibles in the maximum intercuspation position, with the orbital-auricular plane parallel to the horizontal plane. All CBCT images were blindly interpreted by an experienced doctor and a maxillofacial radiologist, who have undergone uniform and standardized training. All the measurements were judged separately. In cases where the determination was not the same, a consensus assessment was made based on patient information, including patient history and clinical examination. To test researcher concordance, an agreement matrix was calculated, and to test the inter-rater reliability, Cohen's kappa coefficient (κ) was conducted. Consensus data was used for radiographic classification and later analysis.
All the measurements were performed with the Mimics Research 21.0 Software, a software tool for visualizing and measuring medical images (such as CT and MRI). Firstly, the mediolateral axis of the condyle was determined in the maximum cross-sectional plane. The coronal plane is then adjusted to a corrected oblique coronal position, parallel to the mediolateral axis of the condyle, and sagittal plane perpendicular to the mediolateral axis of the condyle. The obtained images in the corrected coronal and sagittal plane were further measured.
The presence of osteoarthritic changes was determined from the radiographic symptoms of degenerative joint disease [10,17,20,27,28]: (1) flattening (a flat bony contour deviating from the convex form); (2) erosion (a blurred and discontinuous area of the cortex and subcortical bone); (3) subcortical sclerosis (increased density of cancellous bone under the cortical bone); (4) osteophytes (bone outgrowths and protrusion on the surface or margin of the condyle); and (5) subchondral cyst (cystic lesions with osteolytic, clear borders and mostly sclerotic margins).
Measurements of the morphology of the condyle were evaluated in the axial and coronal slices, while the morphology of the glenoid fossa and the joint space were evaluated in the sagittal slices. In the axial image, the width of condyle was the maximum mediolateral dimension from the mesial and distal end of the condyle [29]. The condylar anteroposterior (A-P) dimension was a perpendicular line to the mediolateral dimension of condyle, from the most anterior to the most posterior point of the condyle. The horizontal angle (H angle) of condyle was formed by the intersection of mediolateral axis of the condyle and the line perpendicular to the median sagittal axis [17,21]. In the coronal images, the width of condyle was similar to that of axis slice, from the mesial and distal end of the condyle, and the vertical angle (V angle) of condyle formed by the intersection of mediolateral axis of the condyle and the line parallel to the horizontal line.
In the sagittal images, the acquisition and measurement of the morphology of the glenoid fossa followed the same procedure as reported in our previous work [21,22]. Inclination articular eminence (IAE) was measured as the angle formed by the intersection of two lines: a line-through the highest point of the glenoid fossa to the lowest point of the articular eminence; and b line-from the lowest point of articular eminence and the midpoint of external ear foramen. The depth of glenoid fossa was the perpendicular distance between the highest points of glenoid fossa to b line.

Statistical analysis
The SPSS 25.0 software was used to perform the statistical analysis. The Kolmogorov-Smirnov test (n>50) or Shapiro-Wilk test (n<50) was performed to check the distribution. The Brown-Forsythe test was employed to check the equality of the group variances. When the correlated quantitative data were normally distributed, a paired test was employed to compare the bilateral joints within each group. Otherwise, the Mann-Whitney U test was performed. For the DJDC group, the comparisons were performed between the preferred side (P-side) and  Female  11  19  28  22  Male  11  12  6  11  Total  22  31  34  33 the nonpreferred side (NP-side) of the joints to evaluate the influences of CSP. When three groups were to be compared, the differences among three groups were compared with the analysis of variance (ANOVA). Post hoc Tukey tests were performed when the differences between the groups were significant. The difference was considered statistically significant when p<0.05 for all the measurements included.  Table 2 and SI. Table 3).

The osteoarthritic changes of DJD patients with/ without CSP
The distribution of osteoarthritic changes presented no significant difference within DJD and DJDC groups (p3>0.05). Among all of the examined TMJs, the articular flattening (61.7% of all TMJs) was the most common type of osteoarthritic changes; and the frequency of surface erosion (55.6%) was present in more than half of the examined TMJs, while other osteoarthritic changes were far behind: subcortical sclerosis (36.2%) and osteophyte (24.5%), with the least proportion of subchondral cyst (14.8%).
In the DJD group, there were no significant differences between the right and left sides of TMJs in five types of osteoarthritic changes (p1>0.05), and the osteoarthritic changes were arranged sequentially at the same pattern. However, in the DJDC group, the flattening and erosion observed in the P-side joints were more than those of the NP-side: flattening and erosion were observed in 71.6% and 67.2% of the P-side joints, respectively, compared with 55.2% and 49.3% of the NP-side (χ 2 =3.893, p2=0.048; χ 2 =4.418, p2=0.036). The frequency of subcortical sclerosis, osteophyte, and subchondral cyst among the bilaterally TMJs presented no significant differences (p2>0.05) ( Table 2) Intergroup and intragroup analysis of the morphological differences Table 3 and Fig. 2 show the results of the analysis regarding the CBCT characteristics of the condyle, glenoid fossa, and condyle position. The ANOVA revealed differences between the Control, DJD and DJDC groups, while the paired t test demonstrated the differences between the bilateral side joints of DJDC group.
There were no statistically significant differences regarding the distribution of the condyle CBCT characteristics among the three groups (p2＞0.05), except the horizontal angle of condyle in axial view (p2=0.041). The horizontal angle of condyle in axial view was significantly greater in DJDC group compared to DJD group (p=0.027). In addition, intragroup results demonstrated that the average of condylar anteroposterior dimension of P-side joints in DJDC group was significantly less  than that of NP-side (p1=0.026). Conversely, the width of P-side condyles in coronal view was greater than that of NP-side (p1=0.041). No statistically significant differences were found in other comparing data (p1>0.05). Significant differences in the CBCT characteristics of the glenoid fossa were observed between the 3 groups including the IAE (p2=0.033) and the DGF (p2=0.014). The DJDC group showed significantly greater articular eminence inclination than the control group (p=0.018). Meanwhile, the DGF was significantly higher in DJDC compared to Control group (p=0.012), and it was also significantly higher in DJD compared to Control (p=0.024). In addition, intragroup results showed that the IAE of preferred side joints in the DJDC group was significantly greater than that of non-preferred side (p1=0.045) but showed no statistical difference in the depth of both sides of glenoid fossa (p1>0.05).
Sagittal joint space 120° differed significantly among three groups (p2=0.041): the DJDC group was smaller than the control group (p=0.015). No statistically significant differences were found in the other sagittal space between the three groups and in the comparison of the two-sided sides of the DJDC group (p1>0.05).

Assessment the frequency of sagittal condylar position
In the control group, the centric and posterior position comprised the highest percentage equally (40.9%), which was more than twice the percentage of anterior position (18.2%). In contrast, the frequency of posterior condyle led the way at 59.7% in the DJD group and 53.7% in the DJDC group, followed by central condyle at 30.6% in the DJD group and 35.8% in the DJDC group. These numbers were generally much higher than those for anterior position, which is approximately one out of ten in the DJD group at 9.7% and in the DJDC group at 10.4%. Although there was an uptrend in the frequency of the posterior condylar position and a downward trend in the frequency of the anterior and central condylar positions with the diagnosis of TMJ, the comparison of the condylar positions indicated no statistically significant difference between the 3 groups (χ 2 =4.513, p2=0.341). Table 4 shows the comparison of the condylar positions of the two-sided joints for the three groups. The right and left sides of the joint did not show significantly different condylar position distributions between the Control and DJD Fig. 2 The comparisons of the morphological characteristics. A Anteroposterior dimension of condyle; B horizontal angles of condyle; C width of condyle; D depth of glenoid fossa; E inclination of articular eminence; F joint space 120° in sagittal view groups (p2>0.05). Although the P-side joints in the DJDC group appear to have more posterior and less anterior condyles than the NP-side joints, the difference is not significant (χ 2 = 2.119, p1=0.347).

Discussion
Degenerative temporomandibular joint disease (DJD) is a progressive disease of the cartilage, bone, and supporting tissues of the TMJ, resulting from biomechanical and biochemical factors associated with decreased capacity for adaptation and excessive physical stress [2]. Increased load in the TMJ may stimulate remodeling, which is an essential biological response to match the functional demands, ensuring homeostasis of joint morphology and function [19]. In the present study, we found that the frequency of flattening and erosion is statistically greater in the preferred side joints of the DJDC group than in the preferred side joints. Meanwhile, a flat, broad condyle and a steep eminence were observed in the preferred side joints of DJD patients with CSP. This result indicates that CSP might make a difference in the morphology of bilateral TMJ among DJD patients, contributing to the onset and progression of subchondral bone degeneration.
Of all patients who presented to the clinic seeking treatment for DJD, 70.4% were female and only 29.6% were male; this finding is in agreement with previous studies where nearly the same rates have been reported [3,31,32]. The presence of more frequent DJD in females than in males has been attributed primarily to hormonal differences between the sexes [10]. The estrogen hormone may have a potential effect on the remodeling of the mandibular condyle [1,9]. In addition, some studies suggested that age is a predisposing factor as both the frequency and severity of the degenerative disease appear to increase with aging [24]. To reduce the interference factor of aging, we set an age limit of 16 to 45 for the inclusion criteria.
Osteoarthritic changes in DJD are often manifested as irregular and thickened cortical contours such as flattening, erosions, subcortical sclerosis, osteophytes, and subchondral cyst formation. Other changes in CBCT images include glenoid fossa, articular eminence and narrowing joint space [20,29,33]. In addition to delineating these changes, imaging findings can assist in the staging of the disease and in monitoring the progression of the changes over time [34]. In the present study, the most common findings of TMJ osteoarthritis in DJD cases were a high rate of articular flattening (61.7%) and erosion (55.6%), while the frequency of subcortical sclerosis (36.2%), osteophyte (24.5%), and subchondral cyst (14.8%) were far behind. These rates are in relatively good agreement with the results of previous studies performed by Bianchi et al. [20], Comert et al. [31], and Derwich et al. [17] on patients with TMJ degenerative arthritis. However, there is some inconsistency with Krisjane et al. [4], who believe that in the asymptomatic DJD patients, the fraction of erosion and subcortical sclerosis is similar, while the fraction of flattening is significantly higher. These conflicting findings may result from differences in age groups, ethnic diversity, or osteoarthritis diagnostic criteria [3,24]. Different imaging manifestations of osseous changes may represent different stages of the disease. According to the DC/TMD criteria, the sole presence of mild flattening or cortical sclerosis may be merely an adaptive physiological change due to increased joint load, aging, or remodeling, which are considered a stage of bone repair [7], whereas erosion on the condyle surface may be an early progression of degenerative changes, and osteophytes, short and irregular condyle shapes, or cystic degeneration may indicate late changes in the TMJ [2,7,35]. In this study, it was found that the proportion of flattening and erosion in the preferred side joints of the DJD group was significantly higher than that in the contralateral side joints and in both side joints of DJD patients without CSP. In addition, many researchers have considered that one side of the joint affects the other side [33], suggesting that CSP may be a promoting factor for the onset and development of DJD.
The majority of published studies have focused solely on the identification of TMJ osteoarthritic changes in CBCT images [17]. In previous studies, we have found that the osseous morphology of TMJ in asymptomatic participants or TMD patients with CSP has its own specific features [21][22][23]. However, there are only a few studies that have assessed the TMJ morphology in relation to DJD and CSP.
In this study, we observed a significant increase in the value of the horizontal condylar angle in the DJD patients with CSP compared to the DJD group and the control group. Other studies have shown that joints with higher values of the horizontal condylar angle may be the result of a remodeling process and tend to develop articular degeneration [36,37]. The increased horizontal condylar angle may result in abnormal compression of the articular cartilage and glenoid fossa. Furthermore, it is worth noting that the average condylar anteroposterior dimension of the preferred side of joints in the DJDC group was significantly smaller than that of the contralateral side, while the width of the condylar in the coronal view demonstrated the opposite result, which implied a flatter shape of the condyle in preferred side joints of DJD patients. The relationship between condylar morphology and osteoarthritis has been discussed in the literature, but whether CSP is involved or not remains unknown. Derwich et al. [17] presented a marked decrease in the condylar anteroposterior dimension with the increase in the number of osteoarthritic changes. Cho et al. [38] showed that, compared to the control group, condylar head and ramal in DJD patients were shorter and the condyle presented a more inclination. In addition, Jiang et al. [21] reported that participants with CSP have a wider condyle than participants without CSP. An explanation that was previously suggested for the morphologic alteration of condyle is that the working TMJ primarily acts as a fulcrum during chewing, while the nonworking TMJ acts as a gliding joint [39]. Unilateral chewing implies asymmetries in joint dynamics and load distribution. Individuals with CSP will repeat masticate on the preferred side of the joint instead of alternating with the balanced side, resulting in excessive load and subsequent adaptive changes in the joint [40,41]. Consequently, the horizontal condylar angle increases, and the condylar head flattens.
The glenoid fossa and articular eminence were bony structures of the maxilla that were exposed indirectly to functional loads during chewing and whose shape is in constant remodeling to match the movements of the condyle and the joint disc. In the study presented, a deep glenoid fossa with a steep eminence was observed in patients with DJD. Furthermore, the IAE of the preferred side joint was significantly greater than that of the nonpreferred side in DJD patients with CSP, but not in DGF. Similar results were obtained by other studies. Santana et al. [40] observed that TMD patients with CSP showed steeper condylar path angles and flatter lateral anterior guidance on the symptomatic side of the joint compared to the contralateral side. In our previous work, we observed that, regardless of the diagnosis of TMDs presented, the articular eminence in individuals with CSP was steeper than that of individuals without CSP [21,22]. In contrast, some studies found no significant differences related to measurements of articular eminence among patients with DJD [17,42]. The disagreement between those results may be a consequence of the differences in the comparison targets for which we selected healthy individuals for comparison. Articular eminence plays an important role in the load distribution of the TMJ by determining the trajectories of the condyle and the articular disc during the masticatory movement, while the DGF represents the difficulty at which the condyle begins to slide forward. Regarding the reason for deep glenoid fossa and steep in the preferred side joints of DJD patients, we speculated that reduced condylar motion and the absence of lubrication during glide can lead to overloading of the TMJ and consequent remodeling [39,40]. A deep glenoid fossa with a steep eminence means that during the gliding motion, the condyle has to move further downward than in a flat, shallow fossa, which in turn causes a change in the condylar morphology.
The condylar position can be determined from the joint space measurements, and the values of these parameters indicate the displacement of the condyle but are not necessarily related to the disc position [36]. Although many studies have examined the joint space, a high level of heterogeneity has been found in the literature investigating the relationship between condylar position and disease. In the presented study, the posterosuperior joint space was smaller in the DJDC group than in the other groups. Although the difference was not significant, the patients with DJD appeared to have more posterior and less anterior condyles than the healthy participants, consistent with Derwich et al. [17]. Although some literature has suggested that the posterior condylar position in DJD patients may be related to a concurrent anterior disc displacement, no firm conclusions have been reached yet [5,35].
Unilateral chewing may produce asymmetric compression and shear forces capable of adaptive morphological changes in the condyle and corresponding maxillary bone structure. A flat condyle, deep glenoid fossa, and steep articular eminence in the preferred side of the joint mean that this lateral joint is more difficult to disperse the mechanical loading during the movement, which acts as the primary load-bearing fulcrum. This partly explains why the frequency of osteoarthritic changes in the preferred side of joint in DJD patients is higher than those in the contralateral side.
However, this study has some potential limitations. Firstly, as a cross-sectional observation, the causal relationship between CSP and CBCT imaging properties of DJD cannot be definitively determined. Secondly, the age of the participants was limited to the 16-45 age range. The correlation between TMJ features in older participants and the span of CSP could not be verified. Given that DJD is a multifactorial disease, CSP alone may not be sufficient to cause DJD directly but may act as one of the contributing factors to DJD, allowing for special anatomic features within TMJ.

Conclusion
Within the limits of the presented study, the preferred side joints of DJD patients with CSP presented a higher prevalence of osteoarthritic changes in the developing and advanced stage. In addition, a flat condyle, a deep glenoid fossa, and a steep articular eminence might be considered as one of the characteristic imaging presentations of the DJD patients with CSP; CSP is a predisposing factor for the development of DJD.
Author contribution Xiaoting Zhai contributed to the statistical analysis, original draft preparation, and editing; Dongzong Huang contributed to the design and data acquisition and critically revised the manuscript; Yifan Hu and Xiaoxing Li contributed to the data acquisition and statistical analysis; Jiazhu Wang and Junli Ma contributed to design, data acquisition, and methodology; Hongbo Li, Min Hu, and Hongchen Liu contributed to conceptualization, validation, and supervision; Hua Jiang contributed to the conceptualization, validation, and resources and critically revised the manuscript. All authors have read and agreed to the published version of the manuscript.
Funding This article publication charges are supported by the National Natural Science Foundation of China (81970957).
Data Availability Data supporting reported results can be found in the dataset generated during this study.

Ethics approval
The study was conducted with the ethical principles of the Helsinki Declaration and obtained the consent of the Medical Ethics Committee of Chinese PLA General Hospital (S2020-327-01).