Do mandibular functional units exhibit inherent asymmetry in adults with a normal degree of chin deviation? A cone-beam computed tomography study

Abstract

This study aimed to investigate the inherent asymmetry of mandibular functional units and to compare the degree of asymmetry between adults with different sagittal skeletal patterns. Ninety CBCTs scans of normodivergent adults with a normal degree of chin deviation, representing different sagittal skeletal patterns (Classes I, II, and III) were included; 30 each. Three-dimensional mandibular models were generated, and volumetric and linear bilateral measurements were obtained for the hemimandible and the condylar, coronoid, ramus, angular, alveolar, body, and chin units. The degree of asymmetry was presented using the asymmetry index (AI) and between groups measurements were compared using ANOVA test and GEE model. The mean AI values were significantly higher than 0% across all measurements for all sagittal skeletal patterns. Volumetric measurements yielded higher overall AI values than linear measurements (3.55 ± 3.6 vs 1.83 ± 1.74, P < 0.001). Negative and positive correlation values were found between mandibular units, thus supporting the theory of compensatory mechanism of the mandible. In all groups, the coronoid process volume and angular unit length exhibited the highest degree of asymmetry, whereas the total body length and hemimandibular volume showed the lowest. No significant differences were found in the AI of volumetric or linear measurements between sagittal skeletal patterns (P > 0.05). Mandibular functional units exhibited a degree of asymmetry in adults from different sagittal skeletal patterns.

Clinical Relevance: This study quantifies the inherent asymmetry and compensatory mechanisms in mandibular functional units of adults with normal chin deviation.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-38624-2.

Introduction

Facial beauty and attractiveness are often linked to facial asymmetry, which refers to differences in size, shape, or position between bilateral facial structures such as the eyelids, eyebrows, nose, and chin^1,2. Mandibular asymmetry typically manifests as a deviation in the chin position relative to the median sagittal plane (MSP)³. Mandibular asymmetry is generally observed in a resting state and more rarely during dynamic facial expression⁴. Relative symmetry, also called subclinical or normal asymmetry, is the unnoticed asymmetry by laypeople⁵. Severe asymmetries worsen the individual’s facial appearance, this may impact smile⁴, facial soft tissue esthetics⁶, and quality of life⁷. Beyond its aesthetic implications, the mandibular asymmetries can contribute to functional complaints including malocclusion like in crossbites, occlusal asymmetry, functional disorders of the masticatory system, excessive tooth wear, and temporomandibular disorders^5,8. As a result, there has been an increased focus on the diagnosis and treatment of facial asymmetries and underlying malocclusions to enhance the patients’ esthetics, oral functionality, and overall quality of life.

Several methods have been reported in the literature for evaluating facial asymmetry, including clinical examination⁴, facial scans¹, radiographic assessments with two-dimensional (2D) panoramic or posteroanterior radiographs, or three-dimensional (3D) images such as cone-beam computed tomography (CBCT)⁹, and, more recently, artificial intelligence-based methods¹⁰. The advent of CBCTs and the development of sophisticated 3D modeling software have revolutionized our ability to evaluate mandibular asymmetry accurately and comprehensively⁹. The traditional method for diagnosing skeletal mandibular asymmetry involves measuring the deviation of central landmarks at the mandibular symphysis, such as the menton (Me), pogonion (Pog), gonion (Go), or genial tubercle (GT) points, to the MSP^11,12. The degree of asymmetry is categorized based on the degree of chin deviation into relative symmetry (0-2 mm), moderate asymmetry (2–4 mm), and severe asymmetry (> 4 mm)^5,13. When determining the necessity and type of intervention for these mandibular asymmetries, it is critical to consider both aesthetic concerns and functional implications. This could necessitate a range of interventions, from less invasive options like botulinum neurotoxin (BoNT) injection and conventional orthodontic treatment to more invasive options such as orthognathic surgery, depending on their impact on the patient’s psychosocial status, overall facial appearance, and potential functional complaints^5,14. In 1988, Habets et al.¹⁵ introduced the asymmetry index (AI) to examine the vertical skeletal asymmetry of the mandible, through ramal and/or condylar height measurements obtained from panoramic radiographs. The AI value of 3% was considered indicative of relative vertical asymmetry; the asymmetry index equation and value were implemented later with more precision in several previous studies^16,17.

Importantly, the AI method proposed by Habets et al. remains highly relevant in contemporary clinical practice. Panoramic radiography offers a low-radiation, cost-effective, and widely accessible means of assessing mandibular asymmetry, particularly condylar and ramal height differences¹⁵. The method has been independently validated in subsequent experimental studies using radiographic phantoms, confirming its reliability under controlled conditions, and has been applied in recent clinical investigations to evaluate the effects of asymmetric malocclusions and functional orthodontic treatment on mandibular and condylar asymmetry^18–20. Thus, panoramic-based asymmetry assessment continues to represent a valuable diagnostic tool in routine orthodontic settings, while three-dimensional volumetric approaches provide complementary insights in research contexts.

The traditional assessment methods, which rely solely on linear measurements, do not provide a comprehensive assessment as they lack volumetric measurements and mainly depend on comparing asymmetry based on the chin deviation, neglecting the assessment of asymmetry between the bilateral mandibular structures. To overcome these limitations, several methods have been developed. For instance, the method proposed by You et al.^21,22, based on several bilateral linear measurements specific for different mandibular units and supplemented by some bilateral volumetric measurements, represents significant methodological advancement. This method relies on the functional matrix theory, first described by Moss and Rankow in 1968²³. This theory comprises two components: the functional matrix (periosteal and capsular matrices) and the skeletal unit (micro- and macro-skeletal units). According to this theory, any change in position and form of the skeletal units results from the effect of the functional matrix, typically muscles. Thus, the mandible is seen as a composite macro-skeletal structure consisting of independent micro-skeletal units. The micro-skeletal units are influenced by the attached muscles, such as temporalis, masseter, and medial pterygoid muscles. Recent findings by Li et al.²⁴ demonstrated that the unilateral injection of BoNT type A (BoNT-A) into the masseter muscle induced mandibular asymmetry by suppressing the growth of the angular process on the injected side due to impaired masseter muscle activity. Vespasiano et al.¹⁷ found that different mandibular segments could compensate for each other to maintain an equilibrium. Hence, while asymmetry may exist between two bilateral mandibular units, the compensatory abilities of other units within each hemi-mandible can lead the mandible to appear symmetrical, with no noticeable chin deviation.

Studies on mandibular asymmetry are primarily conducted to identify the relationships between various factors and the presence of mandibular asymmetry^{13,16,17,25,26}. Patients with different sagittal skeletal patterns, along with a normal degree of chin deviation, constitute the majority in orthodontic practices. However, the degree of asymmetry and the compensatory abilities among these patients remain unclear. Understanding these compensatory abilities is crucial, as it could significantly influence the management and treatment outcomes of orthodontic and maxillofacial procedures. Identifying how patients with mandibular asymmetry adjust to or compensate for their condition can help tailor interventions that are more personalized and effective, thereby potentially improving functional and aesthetic results. Additionally, establishing a baseline for inherent asymmetry will aid in differentiating pathological asymmetry from natural variations. To the best of our knowledge, no previous study has comprehensively evaluated inherent mandibular asymmetry by assessing both volumetric and linear asymmetries of mandibular functional units.

Study aim and hypotheses

This study aimed to investigate the inherent asymmetry of mandibular functional units and to compare the degree of asymmetry among adult participants with different sagittal skeletal patterns with a normal degree of chin deviation.

Among normodivergent adult participants with different sagittal skeletal patterns and with a normal degree of chin deviation, the null hypothesis of this study was:

• Mandibular functional units exhibit inherent symmetry (AI = 0%).

• The degree of inherent asymmetry of mandibular functional units is similar between participants with different sagittal skeletal patterns.

Materials and methods

Study design and setting

The protocol of this retrospective study was reviewed and approved by the Ethics Committee at the Stomatological Hospital of Chongqing Medical University, Chongqing, China (CQHS-REC-2024; LSNo.006). The study was conducted in accordance with the principles of the Declaration of Helsinki²⁷. This study included 90 CBCT scans for adult participants who visited the Department of Orthodontics of the Stomatological Hospital of Chongqing Medical University between January 2020 and January 2023. Informed consent had been obtained from all participants for the use of their records in research.

A priori sample size calculation was conducted using GPower 3.1 software (Universität Düsseldorf, Germany) based on a previous study²⁸. To detect a clinical difference of 2 mm in the chin unit length using analysis of variance (ANOVA) between groups means, the required number of participants was 60 for the three groups (20 each), assuming the power at 80% and the significance level at 0.05. However, we use data from the same mandible which represents a cluster design with 7 units. Considering the clustering effect in the sample size calculation, the required sample size was increased to 90 participants for the three groups due to a design effect of 1.5.

Participants

The inclusion criteria for the study were as follows: (1) participants older than 18 years of age; (2) participants with different sagittal skeletal patterns based on ANB angle: Class I (0° ≤ ANB ≤ 5°), Class II (ANB > 5°), and Class III (ANB < 0°); (3) participants with a normal maxillary sagittal position, indicated by an SNA angle of 82.8° ± 4; (4) normal divergent participants, with an FMA angle between 20° and 32°²⁹; (5) participants exhibiting relative mandibular symmetry, with a normal degree of chin deviation (0-2 mm), measured from both the GT and Me points¹² to a 3 points MSP (nasion-basion-incisive foramen; N-Ba-IF)³⁰ (Supplemantry file 1); (5) participants with fully erupted permanent teeth, excluding third molars, and presenting with mild to moderate crowding. To minimize selection bias, the study included the first 90 participants who were selected based on the inclusion criteria and were age- and gender-matched between groups. Participants were matched with respect to age and gender using a minimization procedure. For each newly enrolled subject, group assignment was determined by minimizing imbalance across predefined demographic strata (age > 25 and < 25 per gender and skeletal classification). Participants were excluded if they had systemic diseases, syndromic or non-syndromic dentofacial conditions, impacted teeth, crossbite, canted occlusion, previous orthodontic or maxillofacial surgical treatments, temporomandibular joint disorders, a history of trauma, or any conditions or lesions influencing the mandibular bone density or level. Additionally, CBCT scans were excluded if they were unmeasurable, incomplete, blurred, or taken for any conditions or operations that could influence the asymmetry of the mandible or different mandibular units.

CBCT examination and mandibular segmentation

All participants underwent CBCT examinations (KaVo Dental GmbH, Bismarckring, Germany) under proper clinical justification as part of their diagnostic records including treatment planning for orthognathic surgeries, temporary anchorage devices placement, digital planning, or comprehensive assessments of the temporomandibular joint, alveolar boundary, and airway. The CBCT scans were performed with the participants in a natural head position, the horizontal plane parallel to the floor, and in maximum intercuspation. The settings were as follows: 5 mA, 120kVp, an exposure time of 8.9 s, a voxel size of 0.4 mm, and a field of view of 16 × 17. All CBCT scans were saved as Digital Imaging and Communications in Medicine (DICOM) files. To ensure blinding during measurements, each patient was assigned a unique ID.

The DICOM files were imported into Mimics software (version 21.0, Materialise NV, Liege, Belgium) to generate 3D models for each mandible, as previously described and validated^17,26. A segmentation mask was created to obtain a 3D model of the skull using the threshold function (Bone CT; 226–3071 Hounsfield units). The mandibular mask was then separated from the whole skull through a semi-automatic segmentation process using the split mask function. Following the method described by Nolte et al.³¹, the clinical crowns of the mandibular teeth were removed from the 3D mandibular object by applying a cut plane passing through the emergence profile from the alveolus of each mandibular tooth. The mandibular canals were then flood-filled using the Smart Fill function (Hole closing distance = 4 voxels), and any scattering on the 3D objects was removed using the Lasso function. Finally, the software’s calculation algorithm automatically calculated the whole volume of the 3D mandibular object.

Linear and volumetric measurements

Initially, all relevant points were identified on the axial, coronal, or sagittal sections, and then adjusted on the mandibular 3D object (Table 1). The software’s algorithms automatically measured the distances of interest in mm. The bilateral linear measurements were performed as previously described and validated by You et al.^21,22 (Fig. 1). For volumetric measurements, each 3D mandibular object was divided into two hemimandibles using a 3-point median cut plane passing through Me, B point, and GT. Subsequently, each hemimandible was segmented into seven distinct units: the condyle, coronoid, ramus, angle, body, alveolar, and chin. This segmentation was based on the cut planes and methods described by Leonardi et al.²⁶ (Fig. 2). Finally, the volume of each unit was recorded. All measurements were repeated and averaged. Further details on points and references used for volumetric measurement are reported in supplementary file 1.

Table 1.

Mandibular landmarks for both volumetric and linear measurements.

Definitions	Abbreviation	Landmark
Most superior point of the condylar head	Consup	Condylion superius
Most medial point of the condylar head	Conmed	Condylion medialis
Most lateral point of the condylar head	Conlat	Condylion lateralis
Most superior point of the coronoid process	Corsup	Coronoid superius
Deepest point on the sigmoid notch	C	C point
Most inferior point on the fossa of the mandibular foramen	F	Fossa of mandibular foramen
Most superior point of the mandibular lingual	Li	Mandibular Lingual
Most posterior point on the mandibular angle	Gopost	Gonion posterius
Most inferior point on the mandibular angle	Goinf	Gonion inferius
Midpoint between Gopost and Goinf on the mandibular angle	Gomid	Gonion midpoint
Point on the line that bisects the mandibular angle into two angles of the same degree and located 1.5 cm above the Gomid point	Go1.5	Gonion-1.5
Point located on the retromolar trigone located where the line that bisects the gonial angle meets the outer bony portion of the trigone	Rm	Retromolar
Entrance of the mental foramen	MF	Mental foramen
Midpoint of mental tubercles	Metub	Mental tubercles
Most anterior midpoint on the symphysis	Pog	Pogonion
Most inferior midpoint on the symphysis	Me	Menton
Midpoint of the greatest concavity on the anterior border of the symphysis	B	Supramentale
Midpoint of the genial tubercle	GT	Genial tubercle

Fig. 1.

Landmarks used for linear measurements: (a) condylar unit length; (b) coronoid unit length; (c) angular unit length; (d) body unit length: (e) chin unit length: (f) condylar width; (g) ramal height; (h), body length. Abbreviations: Con_Sup, condylion superius; Con_Lat, condylion lateralis; Con_Mid, condylion medialis; Cor_Sup, coronoid superius; Go_Mid, gonion midpoint; MF, mental foramen; F, fossa of mandibular foramen; Pog, pogonion; Me, menton.

Fig. 2.

(A) Landmarks and planes used for volumetric measurements. (B) Functional units in both right and left sides after segmentation, (C) Right and left Hemimandibles with all functional units with color coding showing below. Abbreviations: C, sigmoid notch, Go_post, gonion posterius; Go_inf, gonion inferius; Go_1.5, Gonion-1.5 cm; B, supramentale; Me, menton; GT, genial tubercle; Me_tub, mental tubercles; Rm, retromolar.

Asymmetry assessment

To evaluate the asymmetry, both volumetric and linear measurements of the mandible’s left and right sides were analyzed across different groups. This analysis utilized the left, right, and the difference between Left and right (LRDiff) measurements obtained from each patient. The absolute asymmetry index (AI) for each measurement was calculated using the following equation to quantify the degree of asymmetry: Please for this AI equation make space between the words! 

This formula calculates the AI as a percentage, which provides a straightforward method to compare the degree of asymmetry between the two sides of the mandible for both volumetric and linear measurements¹⁵.

Statistical analysis

All measurements were performed by a single investigator. To assess the intraobserver reliability, measurements on 10 randomly selected participants were repeated by the same investigator after a two-month washout period. The intraclass correlation coefficient (ICC) and Bland–Altman Plot were used to evaluate the agreement. Given the normal distribution of the data, continuous data were presented as mean and standard deviation (SD). While dichotomous data were presented as frequencies and counts.

For intragroup comparison, the one-sample t-test was used to determine if there was a significant difference in the AI of all measurements from a hypothetical mean of 0%. For intergroup comparison, ANOVA was conducted, with Benjamini-Hochberg³² correction applied for post hoc comparisons where statistically significant differences were identified. Pearson’s Chi-squared test was used to analyze dichotomous data. To explore the potential association between age, gender, sagittal skeletal patterns (Class I, II, and III) and measured unit with the AI value, the Generalized Estimating Equation (GEE) model was fitted to account for the clustering effect (within-subject correlation) arising from multiple mandibular measurements obtained from the same individual³³. Subjects were treated as the clustering unit, with the unique subject identifier (id). A Gaussian family with identity link was applied, consistent with the continuous nature of the outcome variables. Exchangeable correlation structure was applied for all GEE models assuming constant pairwise correlation among repeated measurements within the same subject. The significance level was set at a two-sided 0.05. All statistical analyses were conducted using R studio software version 4.3.2 (R Foundation for Statistical Computing, Vienna, Austria) and Stata17.0 (Stata Corp, TX, USA).

Results

Participant characteristics showed no significant differences across groups, except for the ANB angle (Table 2). The ICC value showed high agreement, ranged from 0.91 to 0.99 for volumetric measurements and 0.95 to 0.99 for linear measurements (Supplementary File 1).

Table 2.

Characteristics of participants in all groups.

	Class I	Class II	Class III	P value
Age (Years)*	25.3 (3.3)	25.3 (5.7)	25.2 (4.7)	> 0.91
Gender#				> 0.92
Female	15 (50%)	15 (50%)	15 (50%)	–
Male	15 (50%)	15 (50%)	15 (50%)	–
Total	30 (100%)	30 (100%)	30 (100%)	–
ANB (°)*	2.8 (0.8)	7.2 (1.3)	-4.4 (2.5)	< 0.0011
FMA, (°)*	27.16 (2.09)	27.86 (1.43)	27.47 (2.39)	0.41
Chin deviation, (mm)*	1.05 (0.64)	1.09 (0.64)	1.17 (0.64)	0.71

*Data presented as mean (SD), ^#Data presented as Number (%). ¹ One-way ANOVA, ² Pearson’s Chi-squared.

Intra- and inter-group comparison

All groups exhibited a mean AI significantly different from 0% for all volumetric and linear measurements (P < 0.0001 for all). The degree of asymmetry, assessed through the mean values of AI and LRDiff for all volumetric and linear measurements, showed no significant differences between groups (Table 3, Fig. 3, and Supplementary file 2). In general, volumetric measurements yielded higher AI values than linear measurements for the same mandibular unit across all groups. The AI of volumetric measurements ranged from 1.01 to 8 across Classes I, II, and III, with the coronoid process and hemi-mandible volumes presenting the highest and lowest AI in all groups, respectively. The AI of linear measurements ranged from 0.71 to 3.61 across Classes with total body length and angular unit length having the highest and lowest AI in all groups, respectively (Table 3, Fig. 3).

Table 3.

Comparison of the asymmetry index of volumetric and linear measurements of mandibular functional units between groups.

	Measurements, AI (%)	Sagittal skeletal patterns (mean and 95%CI)			P-Value
		I	II	III
Hemi-mandible	Volume	1.01 (0.69, 1.3)	1.44 (1.0, 1.8)	1.25 (0.83, 1.7)	0.271
Functional units
Condylar process	Volume	4.65 (3.4, 5.9)	3.60 (2.7, 4.5)	3.85 (2.6, 5.1)	0.461
	Unit length	1.80 (1.4, 2.2)	1.98 (1.4, 2.6)	1.54 (1.0, 2.0)	0.461
	Width	2.31 (1.0, 2.0)	2.64 (1.6, 3.7)	2.80 (2.0, 3.6)	0.711
Coronoid process	Volume	8.26 (5.9, 11)	8.02 (5.7, 10)	7.77 (5.7, 9.9)	0.951
	Unit length	1.54 (1.1, 2.0)	1.84 (1.3, 2.4)	1.80 (1.3, 2.3)	0.951
Ramal process	Volume	2.86 (2.1, 3.6)	4.01 (2.9, 5.2)	3.50 (2.5, 4.5)	0.481
	Height	1.60 (1.0, 2.2)	1.85 (1.3, 2.4)	1.45 (1.0, 1.9)	0.551
Angular process	Volume	4.70 (1.0, 1.9)	3.52 (2.7, 4.4)	3.76 (2.7, 4.8)	0.451
	Unit length	2.86 (2.0, 3.7)	3.61 (2.5, 4.7)	2.95 (2.1, 3.8)	0.451
Body process	Volume	1.91 (1.3, 2.5)	2.46 (1.8, 3.1)	1.69 (1.3, 2.1)	0.271
	Unit length	1.49 (1.1, 1.9)	1.37 (1.0, 1.7)	1.16 (0.82, 1.5)	0.431
	Total length	0.71 (0.48, 0.95)	0.92 (0.70, 1.1)	0.72 (0.56, 0.89)	0.281
Alveolar process	Volume	3.32 (2.4, 4.2)	3.39 (2.4, 4.3)	3.76 (2.6, 4.9)	0.801
Chin process	Volume	1.78 (1.3, 2.3)	2.23 (1.6, 2.8)	2.42 (1.7, 3.1)	0.341
	Unit length	1.51 (0.96, 2.1)	1.97 (1.4, 2.5)	1.44 (0.83, 2.0)	0.341
Overall
	Volumetric measurements	3.55 (3.28, 3.18)			< 0.0012
	Linear measurements	1.83 (1.70, 195)			< 0.0012

Data are presented as mean (SD) of asymmetry index (%), ¹ ANOVA, ² One-sample t-test.

Fig. 3.

Comparison of the asymmetry index of volumetric and linear measurements of mandibular functional units between groups. Each measurement, labeled as “unit”, encompasses both the unit’s length and volume, except of the ramal unit which includes both the unit’s height and volume.

The difference between the left and right sides was statistically significant in Class III participants in regard to the volumes of the condylar, body, and alveolar processes (MD = 63.60 mm³, 229.60 mm³, and -460.09 mm³; P values = 0.04, 0.01, and 0.02, respectively). Likewise, the volume difference was statistically significant in Class I for the coronoid process (MD = 45.82 mm³, P-value = 0.02). Notably, the linear differences between the left and right sides were not statistically significant throughout the skeletal groups (Supplementary file 2).

GEE findings

The GEE model showed that there was no effect of age and gender on the AI value. The volumetric measures yielded a greater value of AI than the linear measurements. Though the GEE model showed an increase in AI of linear measurements of Class II compared to Class I, this was declined when the AI was calculated from the volumetric measurements. Finally, a correlation was noted between the different parts of the mandible, which was almost negative for the AI of linear measurements of mandibular units compared to the AI of angular unit length. However, different values of negative and positive correlation were also found between the AI of volumetric measurements (Table 4).

Table 4.

GGE model showing the linear relationship between the asymmetry index of volumetric and linear measurements and the covariates.

Asymmetry index	Predictor variables		Coefficient	95%CI	P value
Volumetric versus Linear AI	Linear AI		Ref	−	−
	Volumetric AI		1.72	1.43, 2.01	< 0.001a
Linear measures	Age		0.006	− 0.02, 0.03	0.64
	Gender	Female	Ref	–	–
		Male	− 0.071	− 0.30, 0.16	0.55
	Skeletal pattern	Class I	REF	–
		Class II	0.29	0.16, 0.66	0.04
		Class III	0.004	− 0.27, 0.28	> 0.9
	Linear measures	Angular unit length	Ref	–	–
		Body unit length	− 1.79	− 2.27, − 1.33	< 0.001a
		Body total length	− 2.35	− 2.82, − 1.88	< 0.001a
		Chin unit length	− 1.50	− 1.97, − 1.03	< 0.001a
		Condylar unit length	− 1.37	− 1.83, − 0.90	< 0.001a
		Condylar width	− 0.56	− 1.02, − 0.09	0.02
		Coronoid unit length	− 1.42	− 1.88, − 0.95	< 0.001a
		Ramal height	− 1.50	− 1.97, − 1.03	< 0.001a
Volumetric measures	Age		0.04	− 0.01, 0.09	0.13
	Gender	Female	Ref	−	−
		Male	0.33	− 0.13, 0.79	0.16
	Skeletal pattern	Class I	Ref	−	−
		Class II	0.023	− 0.54, 0.59	> 0.9
		Class III	− 0.06	− 0.63, 0.50	0.83
	Volumetric measures	Angular unit volume	Ref	–	–
		Alveolar unit volume	− 0.51	− 1.39, 0.37	0.26
		Body unit volume	− 1.97	− 2.85, − 1.09	< 0.001a
		Chin unit volume	− 1.85	− 2.73, − 0.97	< 0.001a
		Condyle unit volume	0.04	− 0.85, 0.92	> 0.9
		Coronoid unit volume	4.03	3.14, 4.91	< 0.001a
		Hemi− mandible volume	− 2.76	− 3.65, − 1.88	< 0.001a
		Ramal unit volume	− 0.54	− 1.42, 0.35	0.23

AI, asymmetry index; CI, confidence interval.

^a Statistically significant difference.

Discussion

The present study found that mandibular functional units inherently exhibited a degree of asymmetry in participants from different sagittal skeletal patterns and without noticed asymmetry. The AI of mandibular units in participants with relative symmetry is statistically greater than 0% for all volumetric and linear measurements. Hence, the first part of the null hypothesis was rejected, thus indicating that mandibular asymmetry is more pronounced than previously assumed, even with the absence of apparent facial asymmetry. The degree of asymmetry varied among the units within each group, but not between groups. This emphasizes the nuanced nature of mandibular asymmetry and highlights the importance of a detailed, unit-based analysis for a comprehensive understanding of mandibular characteristics.

This study found that AI values obtained from volumetric measurements were greater than those from linear measurements, a finding confirmed by the GEE model. This discrepancy aligns with the fact that volumetric measurements add a dimension to linear measurements, allowing them to capture the complexity of 3D structure more effectively. Volumetric measurements are also more sensitive to minor variations, providing a wider picture perspective on asymmetry. In contrast, linear measurements, might underestimate the degree of asymmetry^9,26. The GEE model indicated that AI in adult participants was not influenced by age, which suggests that once facial growth ceases, the asymmetry established at that point tends to remain into adulthood without further age-related changes¹³. Additionally, the model indicated that gender had no impact on AI values, reinforcing the idea that asymmetry is a common trait, occurring regardless of gender. These findings align with previous studies^25,34, which concluded that mandibular asymmetries are independent of gender. From a clinical perspective, panoramic-based asymmetry indices remain indispensable for routine diagnosis and follow-up, whereas CBCT-derived volumetric analyses are best reserved for complex cases and research applications.

The GEE model indicated that sagittal skeletal pattern had no impact on AI values. As such, the second part of the hypothesis could not be rejected. These findings suggest that the sagittal skeletal pattern may not be a reasonable factor for the degree of mandibular asymmetry, while aligns with previous studies which found an equal distribution of asymmetries between skeletal Class I, II, and III^5,35. Previous studies have reported that asymmetry is more frequently associated with Class III patients, likely due to the longer and excessive mandibular growth among these individuals, which could pose a risk factor for unbalanced mandibular development^13,36. Thiesen et al.⁵ observed no statistically significant difference in the distribution of relative mandibular symmetry, moderate asymmetry, and severe asymmetry between groups when assessed with the degree of chin deviation. Similarly, a recent systemic review¹¹ evaluated asymmetry through horizontal chin deviation or vertical asymmetry index and reported that the prevalence of mandibular asymmetry ranged from 17.43 to 72.95% among orthodontic and orthognathic patients with different sagittal skeletal patterns, with a high prevalence among Class III patients. In this study, to mitigate the effect of chin deviation on the AI of different units, all patients had relative mandibular symmetry with chin deviation ranging between 0-2 mm. AI is commonly used to investigate the degree and direction of asymmetry between sides. The non-absolute AI can indicate whether one side is larger than the other. However, this distinction will no longer be apparent when calculating the mean values of the non-absolute AI, due to the combination of negative and positive values²⁵. Therefore, this study calculated the absolute value of AI to assess the degree of asymmetry, and directly compared the two sides’ raw measurements to determine if one side is significantly larger than the other. We found that there were no significant differences between the two sides in all groups for all measurements, except for some volumetric measurements, specifically the coronoid process volume in Class I patients and the condylar, body, and alveolar process volumes in Class III patients. These latter results suggested that Class III patients exhibited more relatively asymmetrical mandibular units even in the absence of moderate or severe chin deviation, possibly due to unbalanced growth in those patients. However, the degree of asymmetry was similar between groups when comparing both AI and the difference between the left and right sides.

We found that the highest degree of asymmetry was in the coronoid process volume and angular unit length across all groups, whereas the lowest degree of asymmetry was in hemi-mandibular volume and total body length. From a biomechanical perspective, the greater asymmetry observed in the coronoid process and angular unit may reflect differential loading patterns driven by masticatory muscle activity²⁴. The coronoid process serves as the primary insertion site for the temporalis muscle, whereas the angular region is strongly influenced by masseter and medial pterygoid forces. Kwon et al.⁸ reported that there are no significant differences in the volumes of these muscles between the two sides among patients without facial asymmetry. However, previous experimental and clinical studies have demonstrated that asymmetric or dominant muscle activity can induce localized bone remodeling through mechanotransduction pathways, consistent with the principles of the functional matrix theory and Wolff’s law^23,37. In particular, Li et al.²⁴ reported that suppression of masseter activity via botulinum toxin injection resulted in asymmetric remodeling of the mandibular angular process, supporting a direct muscle–bone interaction. Similarly, Park et al.²⁸ emphasized that functional dominance of masticatory muscles plays a critical role in shaping mandibular morphology. Within this framework, asymmetry of muscle-driven units may be partially offset by compensatory adaptations in other mandibular regions, allowing preservation of overall mandibular symmetry despite localized structural differences. Further future studies are needed to elucidate this complex relationship between muscles activity and the degree of asymmetry between the bilateral units.

Lee et al.³⁸ reported changes in the volume of mandibular angular process following two BoNT-A injections into the masseter muscles of human subjects. Similarly, a recent systematic review³⁹ found that BoNT-A injections into masticatory muscles can cause long-term changes in mandibular bone, such as decreases in bone volume, cortical thickness, and cortical and trabecular density. The complex relationship between the degree of asymmetry and asymmetric muscle activity, as seen in cases of chewing side preference, remains a subject of debate. Heikkinen et al.⁴⁰ reported that the side opposite to the preferred chewing side exhibits a higher chin volume and indicating that functional factors play a more significant role than heredity.

The compensation of asymmetry between the two hemimandibles is essential for mandibular growth harmony¹⁷. The GEE indicated that the AI of volumetric and linear measurements showed a degree of negative and positive correlation, indicating a compensatory mechanism at play within the mandible to maintain overall mandibular symmetry. The compensatory mechanism was also confirmed by the observation that hemi-mandibular volume (derived from volumetric measurements of all units) and total body length (measured as the distance between Go to Me, encompassing angular, body, and chin units) both showed the least degree of asymmetry in all groups. Our findings extend previous observations by You et al.^21,22 and Vespasiano et al.¹⁷ by demonstrating that compensatory modulation occurs not only at the hemimandibular level but also among individual functional units, particularly those under strong muscular influence. This supports the concept that mandibular symmetry emerges from coordinated adaptation across multiple skeletal units rather than uniform bilateral growth. Vespasiano et al.¹⁷ included children aged between 1 to 12 years, and utilized one volumetric, three linear, and one angular measurement to suggest that different mandibular units can adapt with each other to preserve mandibular growth harmony. Similarly, Ramirez-Yañez et al.³⁴ employed panoramic radiographs to derive four linear and two angular measurements, proposing that the mandible adapts to functional requirements by exhibiting side-specific growth. A higher degree of asymmetry in these measurements could lead to true asymmetry with chin deviation, as demonstrated by You et al.^21,22, who observed that hemi-mandibular volume and total body length values were significantly greater on the non-deviated side in facial asymmetry patients with mandibular retrognathism²¹ and prognathism²².

Despite the development of new methods for evaluating mandibular variations, such as statistical shape models⁴¹ and surface-to-surface matching techniques²⁶, these methods were not aligned with the study objectives and had limitations related to generalizability and ability to provide actionable values. Consequently, a landmark-based method was utilized in this study to calculate the AI based on both volumetric and linear measurements. Volumetric and linear measurements are commonly employed for evaluating skeletal structures, with linear measurements derived from 3D human skulls proven to be accurate and reliable⁴². However, volumetric measurements, despite their potential for providing a more accurate and precise evaluation⁴³, are utilized to a lesser extent. This study incorporates both volumetric and linear measurements to offer a comprehensive assessment, aiming to capture differences between the bilateral sides of 3D objects such as the mandible. This approach is useful for understanding the growth behaviors and compensatory abilities within each hemimandible^20,21,26. The linear measurements in this study were based on the fossa of mandibular foramen and mental foramen, these points are considered important reference points located at the junctions of the skeletal units, providing a good basis for measuring the lengths of these units^20,21,28. Although several studies have compared asymmetry among sagittal skeletal patterns using linear measurements from panoramic radiograph^5,25, assessments of volumetric asymmetry are scarce. In contrast to our findings, Mendoza et al.¹⁶ who reported a higher AI for condylar volume in Class II compared to Classes I and III. This could be due to using the non-absolute value of AI in their study, in contrast to our approach.

The skeletal components of Class II and Class III patients exhibit a range of diversity in terms of the spatial and size relationships between the maxilla and mandible^20,21. Celic et al.²⁵ observed higher vertical AI values in high-angle compared to normal- and low-angle skeletal Class I patients. Therefore, this study only included normal divergent patients with a normal maxillary sagittal position, as determined by the FMA and ANB angles, in an effort to eliminate potential biases in the results caused by these factors.

Clinically, recognition of inherent and unit-specific mandibular asymmetry is important for orthodontic and orthognathic treatment planning. Understanding that apparent chin symmetry may mask localized skeletal asymmetries underscores the need for careful evaluation of muscle-related functional units, particularly when planning asymmetric mechanics, functional appliances, or surgical corrections.

Limitations and strengths

The limitations of this study included its retrospective design which may introduce a selection bias. However, a matching technique was applied to balance groups on observed confounders (age, sex). The statistical analysis was also adjusted for confounders and handled the repeated measures per person. Including of normodivergent young adult participants with a normal maxillary sagittal position and without previous orthodontic treatments, may lack the overall assessment of the problem, but the aim of the inclusion criteria was to eliminate the confounding factors related to growth or orthodontic interventions on the findings. Factors prevalent in an older population, such as missing teeth, periodontal disease, or changes in bone density, were not explicitly accounted for in our study population. These conditions could potentially influence asymmetry and, therefore, might affect the generalizability of our results to individuals in late adulthood. Furthermore, the study did not account for the relationship between muscles and mandibular units, i.e., chewing side preference and masticatory muscle dominance, which could be possible confounders. In addition, because the asymmetry index was calculated using absolute values, information regarding the direction of asymmetry (i.e., whether the right or left side predominated) was not retained. While this approach allows robust quantification of the magnitude of asymmetry, it limits interpretation of side-specific dominance. Despite these limitations, this study contributes valuable insights into the structural basis of mandibular asymmetry and its association with skeletal patterns. This could aid in the development of predictive models for growth and treatments outcomes. This study could lay as a groundwork for future investigations into the cause of mandibular asymmetry, its progression overtime, and animal studies aimed at devising treatment strategies that target mandibular units or the attached muscles. The generalizability of the findings is limited to adult participants with different sagittal skeletal patterns, normal vertical pattern, and relative mandibular symmetry with chin deviation less than 2 mm.

Recommendations for future research

Future research should employ longitudinal designs to investigate the development and progression of mandibular asymmetry across different stages of growth and adulthood. Including larger and more diverse populations with varying age ranges, sagittal and vertical skeletal patterns, and degrees of facial asymmetry would improve generalizability and help distinguish inherent, growth-related, and pathological asymmetries. Longitudinal data would also clarify whether compensatory modulation among mandibular functional units remains stable over time or changes in response to functional or environmental influences.

Additionally, future studies should integrate functional and biological assessments to better elucidate underlying mechanisms. Incorporating evaluations of masticatory muscle activity, chewing side preference, and muscle dominance, could clarify the biomechanical pathways driving unit-specific asymmetry. Experimental and translational research, including histological analyses in animal models, may further reveal how asymmetric muscle loading influences bone remodeling at specific mandibular units, thereby strengthening the biological foundation for clinical orthodontic and orthognathic applications.

Conclusions

Based on our study findings, we concluded the following:

• Mandibular functional units exhibited inherent asymmetry among normodivergent participants with different sagittal skeletal patterns and normal degree of chin deviation, for all volumetric and linear measurements.

• Mandibular functional units showed a compensatory mechanism to maintain overall mandibular symmetry.

• Asymmetry index values were higher for volumetric than linear measurements.

• The degree of asymmetry is similar for all measurements between different sagittal skeletal classes.

• The coronoid process volume and angular unit length showed the highest degree of asymmetry in all groups.

• The hemimandibular volume and total body length showed the lowest degree of asymmetry in all groups.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

B.D. contributed to the conceptualization, methodology, investigation, data collection, formal analysis, writing—original draft, writing—review & editing. S.M. contributed to the conceptualization, formal analysis, and writing—review & editing. L.C. contributed to the data collection and writing—review & editing. D.Y. contributed to the data collection and writing—review & editing. Y.H. contributed to the data collection and writing—review & editing. T.Z. contributed to the data collection and writing—review & editing. L.Z. contributed to the conceptualization, methodology, data collection, and writing—review & editing. All authors read and approved the final version of this manuscript.

Data availability

All data generated or analyzed during this study are included in this published article and its supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

This study was approved by the Research and Ethics Committee of the Affiliated Stomatology Hospital of Chongqing Medical University (CQHS-REC-2024; LSNo.006).