The aim of this study was to evaluate the changes in condylar positions after orthognathic surgery using virtual surgical planning via the BOS system. In this study, there was no statistically significant difference in the change in condyle after surgery in the x- and z-axes. In contrast, a statistically significant difference was observed only in the y-axis. The change in the condylar position on the y-axis after surgery occurred mainly downward, and the change was approximately 1 mm. These results are similar to those of Park et al. [
25], who used the same analysis software (Table
2). Both studies used a voxel-based registration method to assess the accuracy of 3D virtually planned orthognathic surgery. Park et al. [
25] determined the position of the condyle using the intended manual positioning during orthognathic surgery. A significant downward movement of the condyle was observed immediately after orthognathic surgery, but a gradual return to the preoperative condylar position was observed up to 6 months after surgery [
25]. There was no statistically significant difference between the preoperative and 6-month postoperative condylar positions (a difference of less than 1 mm) [
25]. Therefore, Park et al. [
25] concluded that the intended manual condylar positioning might minimize changes in the condylar position. Comparatively, the changes in condylar positions after orthognathic surgery via the BOS system showed less change in the condylar position than when using the intended manual condylar positioning. Therefore, changes in the condylar positions after orthognathic surgery via the BOS system are also clinically acceptable.
Generally, three methods can be applied to assess the accuracy of 3D virtually planned orthognathic surgery: landmark-based, surface-based, and voxel-based registration, depending on the manner in which CT images are superimposed [
20,
26‐
29]. The landmark-based registration method involves manually setting stable anatomic landmarks and superimposing them through point matching, similar to the conventional method of superimposing two-dimensional cephalometric radiographs [
26]. It generates human errors, depending on the landmark setting and interobserver variations [
27,
28]. The surface-based registration method involves manually setting stable anatomic regions and superimposing them by matching the corresponding closest point on the same 3D reference surface based on the interactive closest-point algorithm [
26]. The voxel-based registration method is a relatively recent method used for aligning two CT images based on the grayscale differences of voxels [
29]. Voxels, each with a unique grayscale value that depends on the opacity of the scanned structure, are units of volume with isotropic x, y, and z dimensions [
29]. This method calculates the rotation and translation required to align two CT images based on mathematical algorithms [
26]. It automatically superimposes two CT scans based on volumetric similarities and significantly reduces the possibility of human error by eliminating the need to set cephalometric landmarks multiple times [
20,
28]. Although all three methods are reliable for detecting changes in landmark positions when superimposed, the surface-based and voxel-based registration methods are more accurate than the landmark-based registration method [
20,
26].
A limitation of this study is that only changes in the position of the condyles immediately after orthognathic surgery were observed. The position of the condyle changes over a long period and immediately after orthognathic surgery [
25]. Therefore, further studies on the long-term changes in the condylar position after orthognathic surgery using the BOS system are needed.