Introduction

Mandibular prognathism with or without maxillary deficiency is one of the most common dentofacial deformities in adults which is usually indicated for combined orthognathic and orthodontic treatments [1]. Most patients regain harmonious dental occlusion, esthetic facial profile, and improved stomatological functions after treatment. Clinically, dental and skeletal stabilities remain the key factors affecting long-term outcomes [2]. Mounting evidence has demonstrated that positional changes and remodeling of condyle following orthognathic surgery are intricately involved in postoperative stability and long-term outcomes [3,4,5]. Indeed, condylar remodeling is an adaptive course influenced by mechanical forces sustained by the temporomandibular joint (TMJ) and the adaptive capacities of condyle. Orthognathic surgery inevitably induced spatial changes of condyle, which in turn might lead to postoperative TMJ symptoms, skeletal deformities relapse, and even condylar bone re.sorption when these changes were detrimental and beyond physical adaption [6, 7]. The possible reasons for these condylar changes include the relationship with surrounding muscles, joint edema, movement of mandibular bone fragments, and the fixation techniques of bone fragments during bilateral sagittal split ramus osteotomy (BSSO) [6, 7]. Undoubtedly, it is critical for surgeons and orthodontists to evaluate condyle position and morphology properly at both intraoperative and postoperative stages, to keep TMJ healthy and maintain treatment outcomes.

The advent of cone-beam computed tomography (CBCT) allows for radiographic evaluations of dentomaxillofacial structures from traditional two-dimensional (2D) measurements to 3D analyses [8]. It has been widely utilized for dentomaxillofacial CT scanning due to its high spatial resolution [9]. Moreover, CBCT aided by 3D reconstruction and calibration software allows for comprehensive and accurate evaluations of dental and skeletal morphological and spatial changes. For example, accurate determinations of 3D morphological and topological changes of condyle were impossible in 2D radiographs [5, 10]. Furthermore, various 3D image superimposing strategies have been proposed to robustly assess positional and remodeling over time with more accuracy [5, 11]. Three methods for 3D superimposition are currently available including landmark-based, surface-based, and voxel-based superimposition. Given the considerable inconsistencies in manually selected landmarks, surface- or voxel-based approaches were preferred with higher accuracy across previous clinical and radiographic studies [12, 13].

With regard to condylar changes after orthognathic surgery, multiple radiographic modalities such as cephalometric radiography and 2D sliced CBCT images have been exploited to assess condylar displacement in patients with various skeletal dentofacial deformities [14, 15]. Although these approaches have advanced our understanding of condyle changes following orthognathic surgery, craniofacial structure overlapping, radiographic projection, and magnification errors might severely compromise the accuracy and reliability of these results. Our reports together with prior works have documented the characteristic changes of condyle in patients with mandibular protrusion, such as medial, superior displacement and outward, superior rotation within 6–12 months after orthognathic surgery [3, 16, 17]. Moreover, both conventional orthodontic-first and surgery-first approaches have similar postoperative condylar positional alterations irrespective of the timing of orthognathic surgery [16]. However, these previous studies have primarily focused on condylar changes less than 12 months after surgery [3, 16, 17]. Until now, to the best of our knowledge, two previous reports have described the long-term 3D positional and morphological changes of condyles after bimaxillary orthognathic surgery in patients with skeletal class III malocclusions with more than 5-years follow-up [18, 19].

Thus, the present study aimed to characterize the long-term positional changes, surface and volumetric remodeling of condyle by 3D quantitative calibrations in patients with skeletal class III malocclusion after bimaxillary orthognathic surgery.

Materials and methods

Study design and patient enrollment

This retrospective clinical and radiographic study was performed in adult patients with skeletal class III malocclusion who underwent orthognathic surgery and orthodontics at Affiliated Stomatological Hospital, Nanjing Medical University from Jan. 2013 to Dec. 2016. The criteria for patient inclusion were listed as follows: (1) adult patients with mandibular prognathism and maxillary deficiency, but without prominent facial asymmetry (mandibular dental midline deviation to the facial midline less than 4 mm); (2) patients underwent bimaxillary surgery including Le Fort I osteotomy for maxillary advancement and BSSO for mandibular setback with or without genioplasty; (3) patients completed total treatment procedure with more than 60 months (5 years) follow-up; and (4) CBCT datasets at preoperatively (T0), postoperatively immediately (T1), 12 months (T2), and 5-year follow-up (T3) were available. Patients would be excluded if he/she had (1) prior history of mandibular trauma and surgery; (2) clinical symptoms or signs of temporomandibular disorders before treatment such as click, noise, and pain; (3) radiographic signs suggesting pathological diseases such as cortical defects, cystic degeneration, and overgrowth of condyle or condylar resorption before treatment; (4) congenital or developmental defects or systemic diseases such as cleft lip and/or cleft palate; and (5) unqualified original CBCT data. This whole study was reviewed and approved by the Institutional Ethics Committee of Nanjing Medical University and performed in accordance with institutional guidelines. Written informed consent was obtained from each patient.

Orthodontic and surgical procedures

All participants underwent conventional orthodontics and bimaxillary orthognathic surgery by the same team led by Prof. Jie Cheng and Prof. Wenjing Chen. Routine presurgical orthodontics for dental compensation, alignment, and leveling as well as postsurgical orthodontic adjustments were performed as scheduled. The detailed procedures for Le Fort I osteotomy and BSSO with Hunsuck’s modification were performed as previously reported [17, 20]. Following the Le Fort I osteotomy, maxilla was down-fractured, mobilized, and advanced as planned. After intermaxillary fixation with an intermediate occlusal splint, the maxillomandibular complex was fixed with four L-shaped miniplates at the piriform aperture and zygomatic buttress. During the BSSO, the mesial and distal segments were separated at both sides and then passively contacted after the placement of a prefabricated surgical splint and transient intermaxillary fixation. Condylar (distal) segment was seated within the glenoid fossa assisted by the surgeon’s hands without specific device. Then, these mandibular segments were fixed using 1 miniplate and 4 monocortical screws. Light elastics were applied for 2 to 3 weeks to limit mandibular anterior movement when indicated. Postoperative orthodontic treatments routinely started 4 weeks after surgery. During the follow-up period, no severe complications or obvious recurrence were recorded.

CBCT data acquisition and processing

The original CBCT data (New Tom VG or New Tom 5G, Verona, Italy) in DICOM format for each patient were retrieved from our Radiographic Data Center. The detailed parameters for the CBCT scan were listed as follows: 5.73 mA, 110 kV, exposure time of 17 s, voxel size of 0.30 mm, axial slice thickness of 0.30 mm, and scanning area of 18 × 16 cm. The CBCT scanning was acquired with the subject patients in the sitting position with maximum intercuspation and the Frankfort plane paralleled to the ground level. CBCT data were acquired at four time points to analyze postoperative changes of the condyle.

As illustrated in Fig. 1, CBCT images were registered based on the cranial base and skull structure by voxel-based registration using 3D slicer 4.11.0 open-source software (www.slicer.org). The CBCT at the T0 stage was set as a reference. The segmentation of mandible was then constructed based on the registered CBCT data by selecting a suitable grayscale cutoff value and remodeling the condyle automatically. Complementarily, the segmentation and remodeling of 3D condyles were performed by ITK-SNAP 3.6.0 (www.itksnap.org) and exported into new GIPL files which generated visual images (VTK files) via the 3D slicer for further evaluation.

Fig. 1
figure 1

Detailed procedures for 3D condyle segmentation and reconstruction. A, B Voxel-based registration with cranial base as a reference by 3D slicer. The CBCT images at the T0 stage are set as references to register with other stages via cranial base. C Original CBCT image. D Image segmentation: selection of a suitable greyscale cut-off value for condyle and automatic segmentation of condyle. E 3D reconstruction of condyle

Full size image

3D measurements of condylar displacement, surface remodeling, and volume alteration

The CBCT images taken at different times were superimposed to fit the cranial base and skull structures. A 3D coordinate system (including the X, Y, and Z axes; X: lateral + and medial − ; Y: anterior + and posterior − ; Z: superior + and inferior −) was established to evaluate the spatial changes of the condyle. The most lateral (C-1) and most medial (C-2) poles of the condylar head were established on 3D models for 3D measurements of condylar locations. The center of the condylar head (CC) was defined as the midpoint between C-1 and C-2. Detailed information regarding these reference lines, points and all measurements are delineated in Table 1. Three-dimensional measurements of condylar position were divided into the amount of bodily shift and rotational changes. Changes in condylar displacement were examined by measuring the coordinate changes of CC including anteroposterior, super-inferior, and mediolateral movements (Fig. 2). Moreover, angular changes were assessed by calculating the angle of the condylar axes and the longitudinal axis passing through C-1 and C-2 on the preoperative and postoperative 3D models in the horizontal, coronal, and sagittal planes (Fig. 3).

Table 1 Definitions of reference points, lines, and measurements
Full size table
Fig. 2
figure 2

Quantitative measurements of condylar positional changes in a 3D coordinate system. A C-1 represents the most lateral point of the condylar head; C-2 represents the most medial point of the condylar head; CC (condylar center) represents the midpoint between C-1 and C-2. B, C The X distance represents mediolateral movement of CC; the Y distance represents anteroposterior movement of CC; the Z distance represents downward and upward movement of CC

Full size image
Fig. 3
figure 3

Quantitative 3D measurements of condylar rotational displacement. A Axial view of condylar rotation. B Sagittal view of condylar rotation. C Coronal view of condylar rotation. D Roll for inferior ( +) and superior ( −) of condylar head. E Pitch for anterior ( +) and posterior ( −) rotation of condylar head. F Yaw for inward ( +) and outward ( −) rotation of condylar head. Landmarks are positioned on the condylar head (2 points on the medial pole and 2 points on the lateral pole). The quantitative 3D cephalometric module of 3D Slicer software is utilized to calculate the translation and rotational displacements

Full size image

The surface registration of pre-and postoperative 3D condylar models was performed in the 3D slicer to determine the extent of surface remodeling of the condylar head. All these superimpositions were conducted using the same scale on the distance color-code map, generating a visual graph for the morphological changes. The abovementioned procedure was automatic which minimize the observer error. To gauge the changes in condylar surface, 6 areas in the superimposed condyle (anterolateral, anterior middle, anteromedial, posterolateral, posterior middle, and posteromedial regions; Fig. 4) were defined as we previously reported [17]. Then the remodeling types (bone resorption, no change, and bone formation) in each area were recorded and depicted in the color maps. To assess the condylar volumetric changes, a fully automated registration of condyle at different time points was performed using the surface registration in the 3D slicer. Condylar slicing was performed by Easy Clip, 3D slicer 4.11.0 with manually selected lowest point of sigmoid notch as the marker point of the horizontal plane. The bone part above the plane was used for calculation of condyle volume (Fig. 5).

Fig.4
figure 4

3D measurements of surface remodeling in condylar head based on the superimposed color map. A Six areas of condylar head are schematically illustrated. B After registration and superimposition of segmented condyle between different stages, condylar remodeling is shown in color maps depicting areas of bone formation or resorption

Full size image
Fig. 5
figure 5

Volumetric measurements of condyle. Condyle was sliced by a plane parallel to the Frankfurt Horizontal plane through the deepest point of the sigmoid notch. Segmented condyles are cut using the easy clip tool (3D slicer software) for further volume assessments

Full size image

For verifications of the reliability and reproducibility of these analyses and inter-observer consistency, all data were measured twice by the same investigator with 2 weeks intervals. Moreover, the assessments were conducted independently in a blind manner.

Statistical analyses

The intra-observer and inter-observer variability was analyzed by Cohen kappa values. Quantitative data were presented as mean ± standard deviation. Paired Student’s t-tests were employed to compare condylar changes among different periods when these data followed normal distributions as indicated. These statistical analyses were performed using IBM SPSS Software Version 22.0 (IBM, Armonk, NY). Two-sided P-values < 0.05 were considered statistically significant.

Results

This retrospective study included 23 eligible patients who satisfied the inclusion criteria (9 males and 14 females), with a mean age of 28.28 ± 3.28 years old (range from 23 to 35). All patients were diagnosed with skeletal class III malocclusion (mandibular hyperplasia and maxillary hypoplasia) and underwent conventional bimaxillary orthognathic surgery combined with pre- and postsurgical orthodontics. All patients successfully completed the whole treatment and received routine follow-up at regular intervals. The minimal follow-up exceeded 60 months (77.9 ± 15.0 months) after orthognathic surgery. No significant complications or posttreatment relapse were observed. Three patients (2 females, 23 and 26 years old; 1 male, 24 years old) had transient, moderate TMJ symptoms including clicking after surgery, and relieved automatically or through physical therapy within 12 months. Meanwhile, we did not detect specific image findings in these 3 patients based on our 3D image analyses.

3D quantitative analyses of positional and angular changes of condyle following orthognathic surgery

Initially, the intra-observer and inter-observer reliabilities of our 3D quantitative calibrations were determined. Our results found that Cohen kappa values were all above 0.85 for the detections of 3D positional changes and remodeling of condyles, thus supporting the reliability and reproducibility of these measurements.

Large numbers of previous studies have detailed the changes of condyle after orthognathic surgery within relatively short follow-up [3, 17, 21]. Here, the present study focused on the long-term changes of condyles in patients with mandibular prognathism receiving orthognathic surgery by comparisons between T0/1/2 and T3 (5 years). Spatial changes of condyles were calculated and compared within 3D coordinate system. Condylar changes in both sides were generally similar and detailed data of right condyle were presented. As detailed in Table 2, the center of the condyle (CC) shifted 0.23 ± 1.50 mm anteriorly, 0.34 ± 0.99 mm medially, and 1.11 ± 1.10 mm superiorly from T1 to T3 postoperatively. From T2 to T3 period, the CC shifted 0.53 ± 1.11 mm anteriorly, 0.42 ± 0.89 mm laterally, and 0.21 ± 1.09 mm superiorly. When compared with preoperative condylar positions (T0–T3), the CC showed an amount of displacement of less than 0.41 mm in all directions (medial–lateral, super-inferior, and anteroposterior) in the long run, which moved 0.41 ± 1.43 mm anteriorly, 0.14 ± 0.79 mm laterally, and 0.27 ± 0.94 mm superiorly. Noticeably, there was an obvious tendency for condyle to return to its original position as evidenced by 3D distance changes. As shown in Table 3, there were significant differences between T0–T3 and T0–T2, T0–T3, and T1–T2 in mediolateral translational changes and significant differences in T0–T3 and T0–T1, T0–T3 and T1–T3, amd T1–T3 and T0–T1 in craniocaudal translational changes. However, no statistical significance existed in the 3D distances of condylar movements between these stages (Table 3).

Table 2 The 3D quantitative positional and angular changes of condyle after orthognathic surgery
Full size table
Table 3 Comparisons of positional changes of condyle between diverse stages
Full size table

In terms of rotational changes, the condyle shifted 1.58 ± 3.11° outward, 1.83 ± 5.08° superiorly, and 4.79 ± 13.75° posteriorly in T1–T3. During the T2–T3 period, the condylar angular changes were reduced relative to T1–T3 period, with 0.53 ± 2.57° outward, 1.23 ± 4.11° superiorly, and 2.56 ± 14.17° posteriorly. Compared with the preoperative condylar position (T0–T3), the long–term rotational changes of condyle were 1.58 ± 3.28° inward, 1.41 ± 6.14° superiorly, and 2.18 ± 16.27° backward (Table 2). Significant differences in rotational changes were only found in yaws between T0–T1 and T1–T2, T0–T1, and T1–T3 (Table 3).

3D quantitative surface and volumetric remodeling of condyle following orthognathic surgery

Given the condyle undergoes adaptation and remodeling over time, the present study next sought to characterize the surface remodeling and volumetric changes of condyle after orthognathic surgery. As shown in Fig. 4 and Table 4, bone remodeling of six areas at condyle head varied substantially between T1 and T3 stages. Generally, bone resorption in the condylar head occurred more frequently than bone formation, whereas selected regions remain stable with few changes. In detail, bone resorption was observed more frequently in the anterolateral area (63.16%) in T0–T3. Meanwhile, the anteromedial (57.90%) surface of condylar head was found to have bone formation between T0 and T3. Additionally, during the T1–T3 period, bone resorption was observed mostly in anterolateral area (68.18%) similarly, and the medial (anteromedial and posteromedial; 68.18%, 59.09%) area tended to experience bone formation.

Table 4 Surface remodeling of condylar head between T1 and T3
Full size table

Lastly, the long-term volumetric changes of condyle before (T0) and after surgery (T2/3) were quantified by 3D Slicer (Fig. 5). No significant differences in volume changes between bilateral condyles were found. The mean volume of right condyles in T0 was 2019.98 mm3, 2014.13 mm3 in T2, and 2009.79 mm3 in T3. Condylar volume was slightly reduced during the follow-up periods, but these differences between stages failed to achieve statistical significance (P = 0.63 for T0–T2; P = 0.15 for T0–T3; P = 0.06 for T2–T3; paired Student’s t tests).

Discussion

Condylar displacement and remodeling after orthognathic surgery are strongly associated with postoperative occlusal stability, skeletal relapse and TMJ health [7, 21, 22]. The present study characterized the 3D quantitative condylar displacement and remodeling in patients with skeletal class III malocclusion treated with bimaxillary orthognathic surgery during the long-term follow-up. Our results improve current understanding of condylar adaptation and remodeling after orthognathic surgery and reiterate that condylar changes induced by orthognathic surgery largely fall within physical adaptive capacities.

Multiple approaches have been utilized to determine the alterations of condyle after orthognathic surgery including lateral cephalometric radiography, 2D CBCT sliced measurement, and 3D calibrations [3, 5]. Given the complex anatomic structure of TMJ, the inherent weakness of both lateral cephalometric radiography and sliced 2D measurements significantly compromise the accuracy and comprehensiveness of characterizations of postoperative condylar alterations. Accumulating evidence has indicated that the 3D CBCT volume superimposition method is more accurate and reliable with decreased operator errors compared with 2D cephalometric analysis [14]. To improve the accuracy and consistency of superimposition and image interpretation, the cranial base was set as the reference and automatic voxel-based and surface registration was applied; both have been validated in previous reports [23, 24]. Furthermore, 3D calibrations were utilized to assess linear, angular and volume alterations of condyle, which have been demonstrated with high reliability and reproductivity [10, 24, 25]. Indeed, our 3D quantitative calibrations of condyle are reliable and reproducible as supported by the intra-observer and inter-observer consistencies as well as previous reports [24, 25].

Consistent with pioneering studies of condylar displacements in patients with mandibular prognathism, the present study along with others reported that the condyle moved laterally and inferiorly during T0–T1 and then moved medially and superiorly in T1–T2 after mandibular setback surgery using rigid fixations [16, 17]. However, Lee et al. [26] reported that the condyle moved anteriorly, medially, and superiorly during 6-month follow-up after a single BSSO with titanium or bioabsorbable bicortical screws fixation. In addition, our data revealed that condyles moved anteriorly in T0–T3, which was contrary to Kim’s report [4], wherein they detect posterior movement of condyle during 17.36 ± 2.65 months follow-up after a single BSSO for mandibular setback. These discrepancies might be associated with diverse surgical techniques and fixations. Moreover, our findings revealed that condyles continuously turned medial and superior to their original positions after surgery during the long-term follow-up (T1–T3). In parallel, the 3D distance of condyle tended to decrease, also suggesting that the condyle moved back to its original position. Indeed, these condylar returning movements after orthognathic surgery were also observed in several previous reports [15, 16, 21]. Noticeably, Han et al. [6] reported postoperative returning movement of condyle displacement after BSSO with different fixation methods and found that stronger rigid fixation might reduce flexibility of functional adjustment of displaced condyle.

Concerning the rotational changes, our results revealed that condyle rotated inward and forward during T0–T1 and tended to return to its original position by outward and posterior rotations postoperatively in T1–T2 and T1–T3 period, which was in line with previous studies [4, 6]. Furthermore, no significant differences in most translational and rotational changes were observed between different periods. Kim et al. [21] reported a similar tendency and claimed that the amount of outward rotation was not statistically significant. In contrast, Han et al. [27] reported an inward rotation of condyle 12 months postoperatively in patients who underwent BSSO with or without Le Fort I osteotomy. These inconsistencies in positional changes between ours and others may be attributed to multiple factors including sample sizes, variation in osteotomies and fixation methods, the amount of mandibular setback, and postoperative follow-up durations [6, 7, 27].

Condylar remodeling especially obvious resorption has significant effects on postoperative skeletal stability following orthognathic surgery. Prior investigations into condylar surface changes have offered essential clues regarding detailed surface remodeling in patients with skeletal II or III malocclusions [17, 28]. Here, the present study provided long-term results of condylar surface remodeling through 3D surface-rendering image superimposition. Our data showed that bone resorption prominently appeared in the anterolateral area, and bone formation was mainly in the anteromedial area, in line with previous observations [17]. Considering that both directions for condylar displacement and rotation were superior, the bone reduction of condylar head may be attributed to the compressive stresses on condyle in the articular fossa which were induced by condylar superior movement. Complementarily, bone formation in condyle head might result from decreased activities of lateral pterygoid muscle during the adaptive remodeling process. However, detailed assessments of condylar movement and its remodeling after orthognathic surgery and identifications of relevant factors responsible for the remodeling are needed.

In addition to positional changes, our results indicated that total volume of condyle remained largely stable with minimal decrease in these patients with skeletal class III malocclusion during the follow-up. Consistently, condylar volume and height experienced a significant decrease from T0 to T3 (6.1 ± 2.1 years) in skeletal class III patients [19]. Moreover, these condylar volume changes mostly occurred as condylar height changes and did not associate with postoperative skeletal movement. Condylar volume remodeling after orthognathic surgery may result from distortions and mechanical loading of condyle due to rigid internal fixations and postoperative tensions in the surrounding muscles, periosteum, or ligament [5, 6].

Of course, our study had several limitations. Firstly, the relatively limited sample size restricts the generalization of our study, which also impedes the detailed investigations in patients stratified by diverse surgical techniques as well as the amount of mandibular setback. We failed to identify a significant relationship between these condylar changes and clinical outcomes such as dental/skeletal stability and TMJ discomfort after surgery. More studies with adequate number of patients are needed to pinpoint the 3D condylar changes and their effects on clinical outcomes as well as potential differences in condylar changes between males and females or among diverse age subgroups. Moreover, although individuals with preoperative discomforts and bony abnormalities of TMJ were excluded, accurate objective evaluations of preoperative and postoperative TMJ symptoms for each patient should be included in further study.

Conclusion

In conclusion, our 3D superimposition and quantitative measurements reveal the characteristic positional changes and remodeling of condyle after bimaxillary orthognathic surgery in patients with mandibular prognathism. Given the importance of condyle morphology and function, clinicians are still recommended to observe condylar changes after treatment at regular intervals.