The aim of this study was to evaluate 3-dimensional facial shells by incorporating a population-specific average template with a group of Class III subjects preparing to have orthognathic surgery.
The Class III group included 14 male (MCIII) and 15 female (FCIII) subjects. We used 43 male and 44 female Class I subjects to construct average male (AvM) and female (AvF) faces. Coordinates of 3 points on the facial templates of groups MCIII and FCIII and the templates AvM and AvF were compared. MCIII-AvM and FCIII-AvF superimpositions were evaluated for differences.
Vertical distances (sella to soft-tissue pogonion) were statistically significantly higher for the AvM (9.1%) and MCIII (10.1%) than for the AvF and FCIII, respectively ( P <0.05). The distances of soft-tissue pogonion in the horizontal x-axis were positive in 80% of the FCIII group and 85.7% of the MCIII group. The Class III subjects differed from the average face in the lower two thirds, but, in 50% (MCIII) and 60% (FCIII), they differed also in the upper facial third.
(1) The average and Class III Slovenian male morphologic face heights are statistically significantly higher than those of the female subjects. (2) The Slovenian Class III male and female subjects tend toward a left-sided chin deviation. (3) Differences between Class III patients and a normative data set were determined.
Three-dimensional (3D) imaging in maxillofacial surgery and orthodontics is a fast developing field. Several noninvasive and radiographic methods have been introduced in the last 20 years, and they have proved valid and reliable compared with direct anthropometry. The methods that render 3D imaging possible are photogrammetry, laser acquisition systems, structured light systems, video imaging, computerized tomography, cone-beam computerized tomography, magnetic resonance imaging, and ultrasound. Because of ever improving techniques, the acquisition of 3D data today is safe, affordable, and precise. The software applications are also being reengineered to efficiently handle and analyze these highly precise 3D data formats.
Three-dimensional imaging is now being used for various orthodontic and maxillofacial assessments: 3D treatment planning, preorthodontic and postorthodontic evaluations, preoperative and postoperative evaluations, 3D prefabricated archwires, research, distinction between syndromes involving craniofacial deformities, and more. Soft-tissue prediction software has also been used successfully in patients with skeletal Class III malocclusion treated with bimaxillary surgery. Three-dimensional imaging with a laser scanning system has proven to be reliable, with accuracy within 0.85 mm. A study with a photogrammetric tool for 3D acquisition showed a lower system error: within 0.2 mm. On the other hand also, a recent study showed that the 3D cone-beam computerized tomography measurements were statistically significantly different from measurements performed on ex-vivo skulls in two thirds of the measurements, but the authors concluded that this statistical significance was probably not clinically relevant.
Despite the favoring trends in 3D imaging, 2-dimensional diagnostic methods are still the main tools (lateral and frontal cephalograms, dental panoramic tomograms, intraoral and extraoral photographs) in maxillofacial surgery and orthodontics. This might be a direct result of the lack of 3D evaluation tools to accompany newer imaging modalities.
A Class III malocclusion is a common condition that, along with Class I and Class II malocclusions, has physical, psychological, and social effects on quality of life. Class III patients’ most common features are retrusive maxilla, protrusive maxillary incisors, retrusive mandibular incisors, protrusive mandible, and long lower facial height. Facial asymmetry is a 3D problem that often accompanies other facial deformities. Many analyses compare right and left measurements with a constructed midline reference plane for the estimation of asymmetries. This method, however, has raised concerns, and new methods of asymmetry evaluation are still emerging.
The aim of this study was to evaluate 3D facial shells by incorporating a population-specific average template with a group of Class III subjects preparing for orthognathic surgery. To date, 3D data of such nature have not been used to determine differences between Class III patients and a normative data set.
Material and methods
Two groups from the University Medical Center in Ljubljana, Slovenia, were included in the study. The first group consisted of normal subjects (Class I) at the Division of Stomatology, and the second group consisted of Class III subjects who came for surgery at the Department of Maxillofacial and Oral Surgery.
The inclusion criteria for the Class I group were (1) white descent, (2) between 18 and 30 years of age, (3) no adverse skeletal deviations (a basic orofacial examination was performed to exclude them), (4) normal body mass index of 18.5 to 25, and (5) no gross craniofacial anomalies.
The inclusion criteria for the Class III group were (1) white descent, (2) normal body mass index of 18.5 to 25, (3) diagnosed Class III condition that required combined orthodontic and surgical treatment, and (4) no other forms of pathology (eg, condylar hypolasia). The Class III group was further divided into subgroups by sex.
The study was approved by the Slovenian National Medical Ethics Committee. It was conducted according to the principles of the Helsinki-Tokyo declaration. Informed consent was obtained from all subjects.
The laser scanning system consisted of 2 high-resolution Vivid VI900 3D cameras (Konica Minolta, Tokyo, Japan) with a reported manufacturing accuracy of 0.1 mm, operating as a stereo pair. Each camera emits an eye-safe Class I laser, 690 nm at 30 mW, with an object-to-scanner distance of 600 to 2500 mm and a fast mode scan time of 0.3 seconds. The system uses a one half frame transfer charged couple device and can acquire 307,000 data points. The scanner’s output data are 640 × 480 pixels for 3D and red, green, and blue color data. The data were recorded on a desktop workstation, and, for surface capture, a medium-range lens (Konica Minolta) with a focal length of 14.5 mm was used. The cameras were placed 1350 mm from the subjects. The scanners were controlled with multi-scan software (Cebas Computer, Eppelheim, Germany), and data coordinates were saved in a vivid file format. Information was transferred to a reverse modeling software package, Rapidform 2006 (RF6) (INUS Technology, Seoul, Korea), for analysis.
The images were acquired with the subjects in natural head posture. NHP has proven to be clinically reproducible. The subjects sat on an adjustable chair and were asked to look at an object located centrally between the cameras. Adjustments to the height and angle were made to achieve the NHP and appropriate positioning. The subjects were asked to keep their facial musculature as relaxed as possible and to remain as still as possible during the scan. The image acquisition took approximately 10 seconds for every patient and was repeated if any movement in the head position or mimics was noted.
The images were analyzed by using the RF6 software. Absolute mean shell deviations, standard deviations of errors during shell-to-shell overlaps, maximum and minimum range maps, histogram plots, and color maps were generated. The data were further processed before analysis to obtain an image with preserved shape, surface, and volume by using custom-made macros for the RF6. Surface defects were filled automatically or manually without loss of raw data. The result was 1 composite shell per subject.
The construction of an average face was performed by using a previously validated software subroutine available in the RF6. The Class I group was divided by sex. The results were an average male (AvM) shell and an average female (AvF) shell. The steps required to produce an average face have already been described and are summarized as follows: (1) the images are prealigned to determine the principal axes of rotation; (2) manual corrections are made to positioning; (3) best-fit alignment is done with the built-in algorithm in RF6; (4) the z-coordinates of the images are averaged based on normals to a facial template; (5) the point cloud is triangulated to obtain an average face; (6) defects and unwanted areas are removed, and holes are filled; (7) color texture is applied; and (8) shells are created with 1 positive and 1 negative standard deviation.
All images were oriented in the virtual space to have a NHP before analysis. Sella (S), subspinale (A), and soft-tissue pogonion (Pog’) were chosen as described before and shown in Figure 1 . The surface shell was translated in the 3D space so that S represented the zero point (x, y, and z values were 0, 0, and 0). The values of the other points’ coordinates therefore represented distances from S in the chosen axis in millimeters, and their corresponding positive or negative value sign (the plus sign was omitted for positive values) indicated the directions (ie, positive x, left; positive y, up; positive z, to the front). The coordinates of points A and Pog’ were summated in the following manner. (1) As absolute values to demonstrate the absolute difference—ie, distance from S not taking the direction into account; in this way, by dividing the sum by the number of subjects, average distances of points A and Pog’ from S (zero) for the male and female Class III groups were calculated. (2) With their positive and negative values and divided by the number of subjects to give the mean value of the coordinate, showing also the direction. The differences of the A and Pog’ coordinates of the template AvM and group MCIII (AvM – MCIII) and the differences of the template AvF and group FCIII (AvF – FCIII) were also summated and divided in these 2 ways to give the average distances regarding the average face and the means showing also the direction of the points in the Class III groups compared with the average facial templates (AvM and AvF). Means of the coordinates and means of their differences (AvM – MCIII and AvF – FCIII) were compared and tested for significant differences between the sexes. The differences (AvM – MCIII, AvF – FCIII) of coordinates of points A (ax) and Pog’ (px) were also compared for significant differences.
Superimpositions of the shells from the Class III group were performed with the AvM and AvF shells by using a previously described technique. The morphologic differences between the shells were depicted.
The process of comparing the facial average shells involved a manual alignment of the 5 points on the facial scans (4 points at the outer and inner canthus of the eyes and 1 point on the nasal tip) followed by fine alignment performed automatically by the RF6. Color histogram and surface areas and shapes were the parameters used in the study. The color histogram indicates the difference between the average facial shells: the blue areas show negative values, and the red areas show positive values. Surface areas and shapes were automatically generated by the RF6. These shapes were obtained when a previous tolerance of 0.85 mm was applied to the paired surface shell studies. The areas corresponding to 0.85 mm were deemed to be similar between the 2 shells, whereas the shapes above this tolerance represented differences and were shown as surface shapes and color deviations. The percentage of the areas corresponding to the tolerance of 0.85 mm was calculated by the RF6 and represented the similarity of 2 shells.
The data were tested for significant differences by using the independent-samples 2-tailed Student t test in SPSS for Windows (version 11.0.0, SPSS, Chicago, Ill).
One hundred sixteen subjects were included in this study; 43 male and 44 female subjects constituted the normal group that made up the average templates, and 14 male and 15 female subjects constituted the Class III group.
Coordinates of the points Pog’ (px, py, and pz) and A (ax, ay, and az) with the average distance from S (average of the absolute values) for the groups FCIII and MCIII as well as for the AvM and AvF facial templates are presented in Tables I and II , respectively. Their mean values and corresponding directions are also shown. Tables III and IV show the differences between the coordinates Pog’ and A chosen on the AvF and AvM templates and the coordinates of Pog’ and A chosen on the subjects of groups FCIII and MCIII, respectively. The values of the coordinates of points A and Pog’ of FCIII and MCIII and values of the A and Pog’ coordinates of the average facial templates (AvF and AvM) are also presented in Figures 2 to 7 . Coordinates py and ay of the groups MCIII and FCIII were statistically significantly different ( P <0.05). The other coordinates did not show statistical significance.
|Value on AvM †||−0.83||−103.96||−5.49||0.37||−55.60||3.33|
|Subject||Diff px||Diff py||Diff pz||Diff ax||Diff ay||Diff az|
|Subject||Diff px||Diff py||Diff pz||Diff ax||Diff ay||Diff az|