Three-Dimensional Evaluation of Skeletal Stability following Surgery-First Orthognathic Approach: Validation of a Simple and Effective Method

Background  The three-dimensional (3D) evaluation of skeletal stability after orthognathic surgery is a time-consuming and complex procedure. The complexity increases further when evaluating the surgery-first orthognathic approach (SFOA). Herein, we propose and validate a simple time-saving method of 3D analysis using a single software, demonstrating high accuracy and repeatability. Methods  This retrospective cohort study included 12 patients with skeletal class 3 malocclusion who underwent bimaxillary surgery without any presurgical orthodontics. Computed tomography (CT)/cone-beam CT images of each patient were obtained at three different time points (preoperation [T0], immediately postoperation [T1], and 1 year after surgery [T2]) and reconstructed into 3D images. After automatic surface-based alignment of the three models based on the anterior cranial base, five easily located anatomical landmarks were defined to each model. A set of angular and linear measurements were automatically calculated and used to define the amount of movement (T1–T0) and the amount of relapse (T2–T1). To evaluate the reproducibility, two independent observers processed all the cases, One of them repeated the steps after 2 weeks to assess intraobserver variability. Intraclass correlation coefficients (ICCs) were calculated at a 95% confidence interval. Time required for evaluating each case was recorded. Results  Both the intra- and interobserver variability showed high ICC values (more than 0.95) with low measurement variations (mean linear variations: 0.18 mm; mean angular variations: 0.25 degree). Time needed for the evaluation process ranged from 3 to 5 minutes. Conclusion  This approach is time-saving, semiautomatic, and easy to learn and can be used to effectively evaluate stability after SFOA.


Introduction
The surgery-first orthognathic approach (SFOA) has many advantages. It decreases the total treatment time since the lengthy presurgical orthodontic stage is skipped. It also provides immediate improvement in patients' aesthetic profile, facilitates dental movements in the postsurgical phase by the regional acceleratory phenomenon, and improves patients' quality of life. [1][2][3][4][5] The only debate preventing widespread acceptance of SFOA is postsurgical stability. While many consider it as stable as the conventional approach, 6,7 other studies associated it with greater postsurgical mandibular relapse. 8 Long-term skeletal stability after orthognathic surgery (percentage of treatment change that was retained after 1 year or longer [achieved total movement À amount of relapse/total movement]) is affected by numerous factors, including surgical technique, temporomandibular joint and condyle seating, muscle and soft tissue effects, and dental movements 9 : thus, measuring postoperative movements and relapses is crucial for assessing the outcomes of the used approach as well as to delineate the effects of different factors on the surgical outcomes.
Two-dimensional (2D) cephalometry provides reasonable information about vertical and sagittal bone movements as well as rotations around the x-axis (pitch); however, it fails to comprehend horizontal (side to side) movements or rotations around the y-and z-axes (yaw and roll). Moreover, its results can be easily confounded by image overlap, landmark identification, and the experience of the observer. 10 Several researchers have reported a significant difference between 2D and 3D cephalometric measurements. 11,12 In contrast, 3D evaluation can accurately assess postsurgical stability. However, most of the methods in the literature are difficult to replicate, especially for surgeons. These methods depend on the reidentification of a large number of cephalometric landmarks on postoperative images, measurements of each point with the three planes of space, a process which accumulates identification errors leading to inaccuracies, 13,14 or an automatic image registration process, which is time-consuming and requires at least one additional software other than the one used in planning. 15,16 Thus, 3D evaluation is not yet universally applied.
Rapid and extensive dental movements after surgery are another issue encountered when evaluating the 3D stability of SFOA. Relying on dental landmarks, as suggested by other studies, 13,15,17 will not provide accurate information on the skeletal movement. Moreover, a separate 3D segmentation of the maxilla and mandible and dental superimposition by a 3D dental scan are required to avoid the artifacts caused by dental braces during postsurgical orthodontics. This makes the virtual evaluation process more time-consuming.
In a systematic review, Gaber et al 14 suggested three criteria for accurate 3D evaluation of orthognathic surgery planning and stability: "1) an automatic voxel-based registration based on the cranial base; 2) Automated or semiautomated evaluation of the outcome indicative of changes in 3D-whether it is translational or rotational, based on the different axes (x, y, z); 3) Inter-and intra-observer reliability should be used to validate the results." 14 In another study by Almukhtar et al 18 for 3D assessment of surgical changes following orthognathic surgery, no significant difference was found between surface-and voxel-based registration methods.
The purpose of the present study was to propose and validate a simple, time-saving, semiautomatic, and accurate method for 3D evaluation of orthognathic surgery that could be applied to SFOA. This method meets the criteria suggested by Gaber et al, 14 and is hypothesized to have excellent reliability if the intraclass correlation coefficient (ICC) between intraand interobserver measurements is excellent (ICC > 0.95). 19

Methods
This retrospective study included patients with class 3 malocclusion who underwent bimaxillary surgery (Le Fort I and bilateral sagittal split osteotomy) without any presurgical orthodontic work-up, between 2020 and 2021, with accessibility to a good quality computed tomography (CT) or conebeam computed tomography (CBCT) scans of the head region at three different time points: preoperatively (T0), immediately postoperatively (T1), and at least 1 year after surgery (T2); all scans were captured with the condyles and occlusion seated in centric relation. Exclusion criteria were cleft lip/palate, congenital anomalies, and posttraumatic and temporomandibular joint pathology. The study was approved by the institutional review board (registration number: S2022-2275-0001). The aims of this study were as follows: (1) to describe a simple, time-saving, and effective method for evaluation of postoperative stability following orthognathic surgery; (2) to validate this proposed method by calculating the ICC between the repeated measurements of two independent observers; and (3) to define a novel set of linear and angular measurements that could represent the maxillary and mandibular movements after SFOA using the lowest number of points while avoiding dental artifacts and postsurgical movements. A written informed consent was obtained from the patient for illustrative figures and videos.

Data Acquisition and Software Settings
CT or CBCT images were exported into the Digital Imaging and Communications in Medicine (DICOM) format. The DICOM files of each stage (T0, T1, T2) were imported into PROPLAN CMF v 3.0.0 software (Materialise, Leuven, Belgium) and segmented using bone threshold within the program to form a 3D virtual image of the head skeleton. Each of the three images (T0, T1, and T2) was assigned a different contrasting color for better identification in the alignment and landmarking stages.
For automatization of the measurement process, a set of angles and distances (►Table 1) were predefined in the program. This was done by using the cephalometry wizard within the software or directly by importing an extensible markup language (XML) file, which we developed for this evaluation process (►Supplementary Material, available in online version only).

Alignment
Virtual 3D image (T0) was set as the base model, to which T1 and T2 were aligned. For orientation and measurements, seven landmarks are identified on the T0 image, and subsequently, the program automatically defined three planes of space (horizontal, sagittal, and coronal) as described in ►Table 2 (►Fig. 1).
T2 and T3 virtual models were subsequently aligned to T0 using an automatic functionality of the program. To ensure that the program uses the fixed cranial base (not the osteotomized maxilla and mandible), we marked the external surfaces of both superior orbital ridges and both zygomatic arches (►Fig. 2). Subsequently, a quick and accurate automatic surface-based alignment was made (►Video 1).

Measurements and Recordings
Through an automatic process, the program measured the distances between the landmarks (ANS, PNS, rt M, and lt M) on each virtual model and the three planes of space. The three rotational movements (roll, pitch, and yaw) for the maxilla and mandible of each model were also measured separately by using the predefined angles (►Table 2; ►Fig. 4).
All measurements were exported in a spreadsheet format (MS Excel, Microsoft Corporation).
For each case, the time elapsed from the start of the alignment process till the measurements was recorded in seconds.

Validation
To validate the current method and to evaluate the accuracy of measurements, two independent observers performed the alignment and landmarking for all cases. One of them repeated the same steps after 2 weeks; thereafter, all the measurements were automatically calculated and recorded.

Statistical Analyses
Measurements were expressed as mean absolute differences (MADs), mean and standard deviation (SD) of the surgical movements (T1-T0), and surgical relapse (T2-T1), for both translational and rotational changes.
ICCs (absolute agreement in a two-way mixed-effects model) were used to test the repeatability of measurements by comparing inter-and intraobserver measurements. ICCs ˃ 0.80 and ˃ 0.95 were defined as good and excellent, respectively.
The mean linear movements of both mental foramina landmarks (rt M and lt M) in relation to the three planes defined the mean mandibular translational movement. In addition, the mean movements of the ANS and PNS landmarks defined the mean maxillary translational movements.
All statistical analyses were performed using IBM SPSS Statistics 20 (IBM Corporation, Armonk, NY).

Results
The study enrolled 12 patients (8 females and 4 males) whose mean age at surgery was 20 years (range, 16-25 years).

Duration of the Evaluation Process
The mean time for the process of alignment and landmarking of all three virtual models was 3 minutes 40 seconds ( AE 30 seconds) for the first observer and 4 minutes 10 seconds ( AE 50 seconds) for the second observer.

Validation of the Method
The MADs, SD, and ICCs for intra-and interobserver variations for the maxillary and mandibular measurements are displayed in ►Tables 3 and 4, respectively. None of the mean intraobserver variations exceeded 0.06 mm for linear measurements with ICCs between 0.99 and 1. The highest interobserver MAD was 0.18 mm (SD ¼ 0.05) for linear movements and 0.25 degrees for angular movements (SD ¼ 0.06), with ICCs (0.95-1) demonstrating a very high repeatability between the measurements.
The level of agreement between intra-and interobserver measurements is presented using boxplots in ►Figs. 5 and 6, respectively. The differences were close to zero in intraobserver measurements and not exceeding 0.3 for interobserver differences. Linear intraobserver measurements were subjected to the least observer-dependent errors followed by the angular intraobserver measurements, while the angular interobserver measurements showed the greatest variations.

3D Skeletal Relapse Assessment following SFOA
Since there were no significant differences in the measurements of the two observers, the first observer recordings were used to assess the relapse.
The means and SDs of all landmarks' linear movements in relation to the three planes, as well as maxillary and mandibular rotational movements for both (T1-T0) and (T2-T1), are displayed in ►Tables 5 and 6, respectively.  The mean linear movements of the mandible and maxilla are displayed in ►Table 7, which shows the greatest degree of mandibular relapse in the sagittal movements (2.5 AE 2 mm), while the maxilla was relatively stable 1 year after surgery with the greatest mean relapse of 0.7 mm in the posterior direction.

Discussion
Measuring skeletal stability after orthognathic surgery is a very critical step in evaluating the achieved long-term aes-thetic and functional results, testing the reliability of the surgical-orthodontic approach, and studying factors affecting the amount of postoperative skeletal relapse. 2D cephalometric studies have long been used for these purposes. However, they cannot provide enough information to assess the amount of horizontal relapse (side-to-side movements) as well as the degree of pitch and yaw rotations. On the other hand, 3D evaluations described in the literature are rather complicated approaches that consume a lot of time and include the application of numerous steps. In this study, Table 3 Intra-and interobserver intraclass correlation coefficients (ICCs), mean absolute differences (MAD), and standard deviations (SD) for translational and rotational measurements of the maxilla in both surgical movements (T0-T1) and postsurgical relapse (T1-T2)  we described and validated a simple yet effective method for the 3D evaluation of skeletal stability after orthognathic surgery that could be applied effectively to SFOA. We compared intra-and interobserver measurements of two independent observers, where the difference between the repeated measurements was used to evaluate repeatability. Minimal differences with repetition indicated a small margin of error in the essential steps (alignment and landmarking). The results demonstrated the reliability of this new method as shown by the excellent correlation coefficient range and low variations in the repeated measurements. The linear differences were below 0.3 mm, which is less than the clinically acceptable threshold of repeatability of 0.5 mm proposed in previous studies. 13,16,20 This high accuracy could be attributed to the following: • The alignment process was automatic, skull parts not affected by surgery (zygomatic arches and superior orbital ridges). • All the linear measurements depended on sharp anatomical landmarks (ANS, PNS, rt M, and lt M), which are easily identifiable on the 3D skeletal virtual model, except for point A on the maxilla, which was used only for angular measurements. • The reference planes (horizontal, sagittal, and coronal) were defined one time only on T1 before the alignment of  T2 and T3. This makes all the linear and angular measurements to be dependent only on the five landmarks, unaffected by the orientation planes even if they contain errors. • All measurements and recordings were obtained automatically, using a single software, which prevented all potential human errors.
The key advantage of this new approach is that it saves time, with a range of 3 to 5 minutes for the entire process from aligning the three virtual models to measurements. To the best of our knowledge, with the same level of accuracy reported in the literature, 17 this is the least time-consuming 3D method of evaluating stability in orthognathic surgery, compared with similar studies, which either failed to provide the time duration of the assessment procedure or were considered to be time consuming. Nada et al 16 reported 30 Abbreviations: ANS, anterior nasal spine; PNS, posterior nasal spine; SD, standard deviation. Note: Distance to midsagittal plane was calculated in absolute mean to avoid errors that occur when crossing the midline (þ to right side, À to left side). Abbreviations: ANS, anterior nasal spine; PNS, posterior nasal spine; SD, standard deviation. Note: Distance to midsagittal plane was calculated in absolute mean to avoid errors that occur when crossing the midline (þ: to right side; À: to left side). to 40 minutes per set of two CBCT scans, and Shaheen et al 15 reported a mean time of 20 minutes for two modules.
Automatization of alignment and measurements as well as dependence on only five landmarks makes the procedure less time-consuming. The small number of landmarks also increases the accuracy of measurements. The range of ICCs for mandibular angular measurements (0.98-0.99) was better than those for the maxilla (0.95-1); however, both measurements indicated excellent correlations. The minor difference could be attributed to using three landmarks (ANS, PNS, point A) in assessing maxillary rotations, compared with only two landmarks for the mandible. This finding supports the idea that increasing cephalometric landmarking is associated with an increase in measurement errors and variations with repetition. 21,22 Another advantage of this method is using a single software, which is commonly used in segmentation and surgery planning. This improves the surgeon's familiarity with the software leading to faster and easy assessment. It is also possible to apply this method and its associated measurements (Supplementary Material, available in online version only) to any other planning software, which will make the process of evaluation comparatively easy and effective. Previous studies reported the use of at least two software for evaluation. 13,17 Yet another study used one software for evaluation and the other for planning and segmentation, 15 which increases the complexity and time consumption of the procedure.
This novel method of evaluation is independent of dental landmarks, making it more suitable for SFOA, which is characterized by major dental movements after surgery. Besides, when we rely on dental landmarks for evaluation, there is a need to superimpose scanned dental surfaces to the maxilla and mandible separately to avoid artefacts of dental braces in postoperative CT. The superimposition process requires segmentation of the mandible and maxilla separately, which contributes to an increase in the time required and inaccuracy of the process. 23 Findings of 3D evaluation of skeletal stability after SFOA show high stability of the maxilla postsurgery. In the case of the mandible, there was a mean forward relapse of 2.5 mm ( AE 2), which is consistent with previous findings. 7, 19,24,25 There was also an upward vertical relapse of 1.2 mm ( AE 1.2). This vertical reduction in the SFOA can be explained by a more vertical bite settling after surgery, especially in cases with severe occlusal interference between arches. 26 Moreover, there were side-to-side horizontal mandibular relapses of 2 mm ( AE 1.8). These significant horizontal changes could not be recorded by using conventional 2D imaging, indicating the importance of 3D evaluation in orthognathic surgery.
The presented method had a few limitations. The software used cannot detect subdecimeter (<0.1 mm) measurements. In addition, the angles described defining rotational movements of the maxilla and mandible are novel and nonstandardized, which can only be compared with previous readings of the same angles. Hence, the differences between them do not express the true magnitude of movement.
This study presents a novel method of 3D evaluation of skeletal stability after orthognathic surgery. This method is semiautomatic, time-saving, and easy to learn and could be used effectively to measure stability after SFOA. Only five anatomical landmarks were used to assess the stability of the maxilla and mandible. Inter-and intraobserver measurements proved a high reliability and accuracy of these measurements

Ethical Approval
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The study protocol was approved by the Ethics Committee (registration number: S2022-2275-0001).

Patient Consent
Written patient consent was obtained for publication and academic purposes.