This study aimed to assess sagittal and vertical skeletal and dentoalveolar changes through the use of 3-dimensional imaging in prepubertal Class II malocclusion patients treated with a cantilever Herbst appliance (HA). Condyle-glenoid fossa positional changes were also quantified.
This retrospective cohort study assessed 22 children (11.2 years ± 1.2) consecutively treated with a cantilever HA for 12 months and 11 untreated children (aged 9.3 ± 0.30 years) that served as controls. Cone-beam computed tomography was performed at baseline (T1) and at the end of the observation period (T2). Movements in the regions of interest were measured as linear displacements from cone-beam computed tomography images through algebraic calculations. A Student t test for independent samples was used for group equivalence testing at T1, and the treatment differences between T2 and T1 were evaluated by 2 analyses of covariance, one considering the expected growth unit as a covariate and the other with an annualized factor.
The largest dental movement was a mesial movement of mandibular molars (3.70 mm), whereas the largest skeletal changes consisted of a larger relative length of the mandible (difference of 1.2 mm) in the HA group than in the control group.
Within the study limitations (retrospective cohort, historical control group, and sample size), 3-dimensional imaging suggests that HA corrected Class II malocclusion in a predominantly prepubertal sample through more dental than skeletal changes. The changes were more significant in the sagittal than in the vertical direction. In addition, relative stability in the condyle-fossa relationship was noted.
Changes were assessed using a 3D landmark-based superimposition model.
Largest dental movement was the mesial movement of mandibular molars (3.70 mm).
Largest skeletal changes were an increased mandibular length (1.2 mm).
Herbst appliance corrected Class II malocclusion through more dental than skeletal changes.
The changes were more significant in the sagittal than in the vertical dimension.
The Herbst appliance (HA) is a fixed functional orthopedic appliance used to correct Class II malocclusion associated with mandibular deficiency. It does not depend on patient compliance, and treatment length ranges from 6 to 18 months. In the dentoalveolar region, maxillary teeth tend to move distally, whereas mandibular teeth tend to move mesially. Skeletal changes have also been observed. These changes reflect a posterior force vector on the maxillary dentition and an anterior force vector on the mandibular dentition.
The majority of Herbst studies have been performed using 2-dimensional (2D) cephalometric imaging, an approach that cannot adequately assess the complex interactions of 3-dimensional (3D) changes that occur with craniofacial growth and orthopedic treatment. In a recently published systematic review, it was recommended that the portrayed skeletal and dental changes attributed to the HA be interpreted with caution because of the low quality of evidence and publication bias. The limitations of this analysis were the small number of high-quality studies and uncontrolled consideration of cephalometric magnifications.
Currently, 3D imaging is widely accessible and able to quantify skeletal and dental measurements more accurately because distortions and superimpositions are eliminated. Many studies have attempted to develop reliable methods for assessing data from different time points using 3D imaging. Current methods rely on the voxel-based superimposition of the skull base or landmark-based superimpositions. , ,
Previous 3D studies about the HA have had some limitations, including small sample sizes, control groups treated with fixed appliances, and failure to report the changes in the 3 different spatial planes. This retrospective study is the first to assess 3D skeletal and dentoalveolar changes distinctly in the sagittal and vertical planes produced by a cantilever HA in Class II malocclusion, mostly in prepubertal patients using a landmark-based superimposition 3D method. Changes were compared with an untreated Class II malocclusion group. Condyle-glenoid fossa positional changes were also quantified.
Material and methods
This retrospective study was approved by the local Research Ethics Committee of Positivo University (process no. 2.207.562). The sample consisted of cone-beam computed tomography (CBCT) images from girls aged 9-12 years and boys aged 10-13 years consecutively treated with cantilever HA at the outpatient dental clinic affiliated with a university ( Table I ). As of 2010, the University’s clinic started using CBCT imaging to diagnose patients with significant skeletal discrepancies. All available patients with HA (n = 22) up to 2016 that met the inclusion criteria were considered.
|Initial age, y||11.2 ± 1.2||9.3 ± 0.30|
|Time between CBCT scans, y||1.5 ± 0.4||1.9 ± 0.5|
|CVM stage 1||4||1|
|CVM stage 2||9||8|
|CVM stage 3||3||2|
|CVM stage 4||4||0|
|CVM stage 5||2||0|
The following inclusion criteria were used: Class II molar relationship with at least half cusp on both sides, pronounced overjet (>4 mm), convex facial profile suggestive of mandibular retrognathia, and an improved facial profile when the mandible was positioned forward. Patients subjected to previous orthodontic treatment, tooth agenesis, and history of abnormal bone growth were excluded from the study.
The HA group included 22 patients (12 boys and 10 girls) with a mean age of 11.2 ± 1.23 years at baseline, treated with a cantilever HA ( Fig 1 ). Seven patients required maxillary expansion with a Hyrax appliance for a mean period of 4 months. This appliance was removed before the placement of the HA. The PMA telescopic system (3M Unitek Abzil, São José do Rio Preto, São Paulo, Brazil) was used. All appliances contained Rollo bands (American Orthodontics, Sheboygan, Wis) on the 4 first molars and a cantilever on the mandibular molars. A transpalatal arch was used for the maxillary molars, and a lingual arch with occlusal rests on the deciduous second molars or the mandibular second premolars attached to the mandibular first molars. A construction bite registration was obtained for edge-to-edge incisor relationship, with a mean mandibular advancement of 7.2 mm (max: 10 mm, min: 4 mm) in a single step. The appliance was worn for at least 12 months.
The control group (CG) included 11 patients (8 boys and 3 girls) with a mean age of 9.3 ± 0.30 years at baseline, with the same characteristics as those described for the HA group. These patients underwent an examination at baseline and, for different reasons—especially lack of availability of their parents or legal guardians to bring them to the dental appointments, in addition to financial hardships associated with treatment costs—could not initiate orthodontic treatment. After approximately 18 months, new contact was made with the patients, and they underwent a new examination (including orthodontic records) and were referred to treatment.
CBCT imaging was considered appropriate by the teaching institution for patients with Class II malocclusion with significant skeletal discrepancies.
The patients were grouped on the basis of their skeletal maturation stage, determined by the cervical vertebral maturation (CVM) method proposed by McNamara and Franchi ( Table I ).
All patients underwent a CBCT scan examination at baseline (T1) and the end of the observation period or up to 7 days after Herbst treatment (T2). The time between CBCT scans in the HA group was, on average, 1.5 years with a standard deviation of 0.4 years. This long period is due to a delay of approximately 4 months in initiating treatment, including the time for maxillary expansion. The average time in the CG was 1.9 years, with a standard deviation of 0.5 years.
CBCT scans were performed with standard head positioning (Frankfurt horizontal plane) at these settings: 120 kVp; 8 mA; 0.3-mm voxel size; scan time, 17.8 seconds; field of view of 170 mm × 170 mm; and patient in maximum intercuspation. An i-CAT (model 9140; Imaging Sciences International, Hatfield, Pa) was used. The CBCT images were exported as digital imaging and communication in medicine files.
CBCT images were analyzed using Avizo software (version 8.1; Mercury Computer Systems, Inc, Berlin, Germany). Landmarks were located on the sagittal plane and positioned on the axial and coronal planes 3 times by a single calibrated examiner (K.L.S.). Spherical digital markers (0.5 mm) were placed to determine the center of each point. Supplementary Figure and Table II show the 3D images and the definitions of the reference points and landmarks.
|Dentoalveolar measurements||Skeletal measurements||Temporomandibular joint measurements|
|ELSA/PC16||Maxillary right first molar at the center of the largest cross-sectional PC area||ELSA/A||Distance between ELSA and A point||ELSA/SRC||Superior right condyle|
|ELSA/PC26||Maxillary left first molar at the center of the largest cross-sectional PC area||ELSA/B||Distance between ELSA and B-point||ELSA/SLC||Superior left condyle|
|ELSA/PC36||Mandibular left first molar at the center of the largest cross-sectional PC area||ELSA/RMF||Distance between ELSA and right mental foramen||ELSA/PRC||PRC|
|ELSA/PC46||Mandibular right first molar at the center of the largest cross-sectional PC area||ELSA/LMF||Distance between ELSA and left mental foramen||ELSA/PLC||PLC|
|ELSA/IS11||Incisal edge of maxillary right central incisor||ELSA/ANS||Distance between ELSA and ANS||ELSA/SRGF||Superior right glenoid fossa|
|ELSA/IS21||Incisal edge of maxillary left central incisor||ELSA/PNS||Distance between ELSA and posterior nasal spine||ELSA/SLGF||Superior left glenoid fossa|
|ELSA / II 31||Incisal edge of mandibular left central incisor||ELSA/Pog||Distance between ELSA and pogonion||ELSA/PRGF||Posterior right glenoid fossa|
|ELSA/II 41||Incisal edge of mandibular right central incisor||ELSA/GoR||Distance between ELSA and right gonion||ELSA/PLGF||Posterior left glenoid fossa|
|ELSA/MBA16||Mesial buccal root apex of maxillary right first molar||ELSA/GoL||Distance between ELSA and left gonion||ELSA/ARGF||Anterior right glenoid fossa|
|ELSA/MBA26||Mesial buccal root apex of maxillary left first molar||PRC/A||Distance between posterior right condyle and A point||ELSA/ALGF||Anterior left glenoid fossa|
|ELSA / MA36||Mesial root apex of mandibular left first molar||PLC/A||Distance between posterior left condyle and A point|
|ELSA / MA46||Mesial root apex of mandibular right first molar||PRC/Pog||Distance between PRC and Pog|
|ELSA/A11||Root apex of maxillary right central incisor||PLC/Pog||Distance between PLC and Pog|
|ELSA/A21||Root apex of maxillary left central incisor|
|ELSA/A31||Root apex of mandibular left central incisor|
|ELSA/A41||Root apex of mandibular right central incisor|
|IOF/PC16||Distance between superior most aspect of the right IOF outer border and maxillary right first molar at the center of PC||ANS/RMF||Distance between ANS and right mental foramen|
|IOF/PC26||Distance between the most superior aspect of the left IOF outer border and maxillary left first molar at the center of PC||ANS/LMF||Distance between ANS and left mental foramen|
|MF/PC46||Distance between right mental foramen and mandibular right first molar at the center of PC|
|MF/PC36||Distance between left mental foramen and mandibular left first molar at the center of PC|
The methodology employed in this study consisted of 4 steps: identification of landmarks, coordinate system transformation, a superimposition using optimization calculations, and measurement of displacement of the assessed structures.
First, reference points at the skull base were used to establish the coordinate system and plane orientation. The right and left external auditory canals, right and left foramina ovalia, foramen magnum, and a point equidistant from the points at the center of each foramen spinosum (ELSA ), were identified.
Later, a coordinate system transformation was performed, subtracting the vector that describes the distance between ELSA and the original position (0, 0, 0) and repositioning all the other anatomic structures.
After that, an optimization problem had to be solved for the CBCT images taken at baseline and at the end of treatment to be superimposed on the Cartesian system. The relative distance and the relationship of angles between the landmarks were calculated on each image, and then an algorithm was developed by using linear algebraic equations for the necessary corrections.
Finally, the face was marked to indicate the skeletal and dental changes and the position of the condyle and mandibular fossa. The total variation observed in the treatment period was calculated by the difference between the distances measured in different periods (T2 − T1). Distance (d) in millimeters was determined by the following equation:
The anatomic distances used in the present study are described in Table II .
The normality of the data was assessed by the Kolmogorov-Smirnov test. P <0.05 was considered statistically significant. Data were analyzed by SPSS software (version 20.0; IBM, Armonk, NY).
A Student t test for independent samples was used to compare the HA and CG with baseline (T1) values. Because of the difference in time between CBCTs in both groups, the sample had to be adjusted for the final assessment considering differences in expected growth (T2 − T1). Therefore, 2 statistical approaches were used: (1) analysis of covariance with the expected growth unit (EGU) factor. EGU corresponds to an individualized estimate of growth intensity expected to occur in orthodontically untreated patients of the same sex and age at a specific time interval. (2) Analysis of covariance with changes in annualized measurements to control the time difference between radiographs.
Based on the repeated measurements in 3 different periods by the same evaluator, the intraclass correlation coefficient (ICC) was then estimated. The images were remeasured 15 days after the first measurements. The third evaluation was carried out 30 days after the second one. All data from both groups (HA and CG) and baseline measurements for the 3 coordinates of each point (x, y, and z) were used. Dahlberg formula was used to calculate the random error.
For the precision of landmarks, with few exceptions, ICC values were very close to 100%, indicating good reliability. Overall, the ICCs of the landmarks were greater than 0.962, 0.959, and 0.987 on the x-, y-, and z-axes, respectively. The poorest ICC on the x- and y-axes was for B-point, measuring 0.962 and 0.959, respectively. The poorest ICC on the z-axis was for the right infraorbital foramen, measuring 0.987. The random error of most landmarks was <1 mm. The greatest error on the x-axis was for posterior right condyle, measuring 1.03 mm. The greatest error on the y-axis was for root apex of right lower central incisor, measuring 1.24 mm. The greatest error on the z-axis was for B-point, measuring 1.11 mm.
The descriptive statistics of dentoalveolar changes summarized in Table III suggest that, at baseline, the maxillary and mandibular molars were more anteriorly positioned in the HA group than in the CG, with a statistically significant difference for the crowns and roots. The maxillary incisors were more labially positioned, with a statistically significant difference for the crowns of maxillary and mandibular central incisors. The mandibular incisors were in relatively similar positions. Therefore, one can observe that, at baseline, the Class II malocclusion characteristics were more pronounced in the HA group than in the CG because the mean distance between the maxillary and mandibular molars was 5 mm vs 4 mm, whereas overjet had a difference of 7 mm vs 4 mm, respectively, indicating a more severe Class II malocclusion in HA than in the CG.
|T2 − T1||3.12||0.80||0.29||2.40||0.001 †|
|T2 − T1||1.66||0.36||0.10||1.68||0.008 ‡|
|T2 − T1||2.24||1.28||1.17||2.39||0.278 †|
|T2 − T1||1.15||0.59||0.73||1.58||0.633 ‡|
|T2 − T1||2.7||1.0||0.41||3.08||0.035 †|
|T2 − T1||1.44||0.46||0.14||2.02||0.073 ‡|
|T2 − T1||2.74||1.75||1.35||2.42||0.165 †|
|T2 − T1||1.48||1.12||0.89||1.53||0.352 ‡|
|T2 − T1||3.72||1.35||5.10||2.52||0.060 †|
|T2 − T1||2.01||0.75||3.59||1.61||0.001 ‡|
|T2 − T1||4.12||1.24||4.77||3.15||0.292 †|
|T2 − T1||2.29||0.89||3.29||1.86||0.033 ‡|
|T2 − T1||4.03||0.90||5.26||2.14||0.060 †|
|T2 − T1||2.20||0.58||3.82||1.79||0.002 ‡|
|T2 − T1||4.41||1.31||4.69||2.46||0.614 †|
|T2 − T1||2.40||0.75||3.39||1.93||0.040 ‡|
|T2 − T1||3.64||1.91||1.13||2.12||0.031 †|
|T2 − T1||2.03||1.30||0.75||1.46||0.191 ‡|
|T2 − T1||3.29||1.29||1.14||1.92||0.008 †|
|T2 − T1||1.76||0.71||0.80||1.36||0.089 ‡|
|T2 − T1||3.26||0.89||1.30||2.17||0.062 †|
|T2 − T1||1.79||0.65||0.86||1.50||0.350 ‡|
|T2 − T1||2.89||0.98||1.30||1.81||0.027 †|
|T2 − T1||1.54||0.53||0.93||1.29||0.256 ‡|
|T2 − T1||3.25||0.93||5.22||2.28||0.010 †|
|T2 − T1||1.78||0.60||3.68||1.54||0.001 ‡|
|T2 − T1||3.57||1.28||4.06||1.93||0.364 †|
|T2 − T1||1.95||0.83||2.91||1.46||0.031 ‡|
|ELSA/II 41||T1||78.03||3.06||79.09||4.47||0.485 ∗|
|T2 − T1||3.30||0.86||5.42||2.37||0.007 †|
|T2 − T1||1.81||0.59||3.80||1.57||0.001 ‡|
|T2 − T1||3.53||1.12||4.08||2.25||0.358 †|
|T2 − T1||1.92||0.71||2.93||1.64||0.034 ‡|
|T2 − T1||1.54||0.91||1.57||1.20||0.947 †|
|T2 − T1||0.81||0.46||1.05||0.81||0.355 ‡|
|T2 − T1||2.22||2.04||1.63||1.20||0.235 †|
|T2 − T1||1.28||1.34||1.11||0.77||0.666 ‡|
|T2 − T1||−0.01||0.95||−1.56||1.90||0.001 †|
|T2 − T1||0.01||0.54||−1.08||1.04||0.001 ‡|
|T2 − T1||−0.43||0.65||−1.15||0.93||0.019 †|
|T2 − T1||−0.24||0.38||−0.88||0.78||0.018 ‡|
After treatment, the anterior movement of the crown of maxillary right and left first molars was more limited, with a mean difference of 1.56 mm and 1.30 mm, respectively, compared with the annualized average movement. The mean anterior displacement of the maxillary molars in the HA group was 0.12 mm vs 1.57 mm in the CG.
The mandibular molars in the CG showed a mean anterior movement of 2.01 mm for the left first molar and 2.20 mm for the right first molar. In contrast, the HA group showed a larger anterior movement (3.59 mm and 3.82 mm, respectively). The mean anterior displacement of the mandibular molars, assessed by the ELSA/pulp chamber [PC]36 and ELSA/PC46 measurements in the HA group, was 3.70 mm vs 2.10 mm in the CG.
At T2 − T1, the anterior movement of maxillary incisors was restricted in patients wearing HA. The mandibular right and left central incisors in the HA group revealed the anterior crown movement of approximately 2.0 mm when compared with the CG, with statistically significant differences. The apices of teeth moved 1.93 mm in the CG and 2.92 mm in the HA group, also with statistically significant differences ( Fig 2 ).
In the vertical measurements (infraorbital foramina [IOF]/PC16 and IOF/PC26), the maxillary molars intruded an average of 1.0 mm in both groups with a difference of 0.04 mm in the annualized statistics and 0.28 mm in the EGU statistics, without significant difference.
Regarding vertical measurements, the CG had a mean extrusion of mandibular molars of approximately 0.12 mm (FM/PC36 of 0.01 mm and FM/PC46 of −0.24 mm), compared with extrusion of 0.98 mm (FM/PC36 of −1.08 mm and FM/PC46 of −0.88 mm) in the HA group. There were statistically significant differences in both measurements ( Fig 2 ).
The statistical analysis of the maxillary and mandibular measurements ( Table IV ) suggests that the maxilla was in a similar position at baseline in both groups, whereas the mandible was anteriorly positioned in the HA group, with a difference of approximately 2.75 mm (85.98 mm for the HA group and 83.23 for the CG) at point B and 3.34 mm at the pogonion (Pog). In addition to these 2 points, the distance of ELSA from the right and left mental foramina (RMF/LMF) and the right and left gonion revealed significant differences at T1. At T2 − T1, these parameters showed no statistical difference, but there was some constraint on the maxilla in the HA group because the anterior movement from point A was significantly smaller than the CG (0.51 mm and 1.23 mm, respectively).
Mandibular length, measured by posterior right condyle (PRC)/Pog and posterior left condyle (PLC)/Pog, had a larger significant increase, with a mean difference of approximately 1.20 mm (PRC/Pog: 1.97 mm and 3.25 mm; PLC/Pog: 1.93 mm and 3.03 mm in the CG and HA group, respectively, for annualized changes), with larger, albeit nonsignificant, anterior movement from point B in the HA group ( Fig 3 ).
The measurement from the mental foramen to the anterior nasal spine (ANS/RMF and ANS/LMF) analyzed the anteroinferior facial height. The mean difference was 1.29 mm in the HA group and 0.76 mm in the CG, with an average difference of 0.5 mm in the annualized statistics, with no significant difference ( Fig 3 ).
The condyle-mandibular fossa relationship is shown in Table V . Descriptive statistics showed no statistically significant difference between the groups at T1 and T2 − T1, considering EGU or annualized changes.
|T2 − T1||0.54||1.51||1.16||2.13||0.373 †|
|T2 − T1||0.28||0.79||0.76||1.38||0.279 ‡|
|T2 − T1||0.85||1.20||0.42||1.79||0.309 †|
|T2 − T1||0.41||0.58||0.34||1.32||0.858 ‡|
|T2 − T1||1.03||1.28||1.11||1.84||0.801 †|
|T2 − T1||0.54||0.62||0.72||1.21||0.562 ‡|
|T2 − T1||0.78||0.99||0.44||1.89||0.512 †|
|T2 − T1||0.38||0.45||0.38||1.35||0.977 ‡|
|T2 − T1||0.57||1.32||1.13||2.10||0.404 †|
|T2 − T1||0.30||0.69||0.74||1.36||0.308 ‡|
|T2 − T1||1.12||1.39||0.41||1.76||0.136 †|
|T2 − T1||0.55||0.68||0.34||1.35||0.627 ‡|
|T2 − T1||1.33||1.34||0.89||1.91||0.568 †|
|T2 − T1||0.71||0.65||0.57||1.29||0.936 ‡|
|T2 − T1||0.87||1.17||0.49||1.83||0.522 †|
|T2 − T1||0.44||0.57||0.43||1.33||0.936 ‡|
|T2 − T1||0.42||1.29||1.15||2.32||0.326 †|
|T2 − T1||0.23||0.68||0.77||1.48||0.264 ‡|
|T2 − T1||1.33||1.13||0.37||1.82||0.069 †|
|T2 − T1||0.68||0.54||0.29||1.30||0.362 ‡|
Both points marked on the mandibular condyle (ELSA/superior right condyle and ELSA/PRC) shifted more anteriorly in the HA group. The upper condyle position was previously displaced by 0.55 mm in the HA group and 0.34 mm in the CG, with no statistical difference. The posterior point shifted by an average of 0.58 mm vs 0.46 mm, respectively, without statistical difference.
In the mandibular fossa, 3 points were marked: superior, posterior, and anterior (ELSA/superior right glenoid fossa, ELSA/posterior right glenoid fossa, and ELSA/anterior right glenoid fossa). The upper position of the mandibular fossa displaced anteriorly with a higher average in the HA group (0.54 mm) than in the CG (0.42 mm). The posterior point of the mandibular fossa was displaced anteriorly with a lower average in the HA group (0.50 mm) than in the CG (0.57 mm). The anterior region of the mandibular fossa displaced anteriorly with a higher average in the HA group (0.53 mm) than in the CG (0.45 mm).