The purpose of this study was to quantify and qualify the 3-dimensional (3D) condylar changes using mandibular 3D regional superimposition techniques in adolescent patients with Class II Division 1 malocclusions treated with either a 2-phase or single-phase approach.
Twenty patients with Herbst appliances who met the inclusion criteria and had cone-beam computed tomography (CBCT) images taken before, 8 weeks after Herbst removal, and after the completion of multibracket appliance treatment constituted the Herbst group. They were compared with 11 subjects with Class II malocclusion who were treated with elastics and multibracket appliances and who had CBCT images taken before and after treatment. Three-dimensional models generated from the CBCT images were registered on the mandible using 3D voxel-based superimposition techniques and analyzed using semitransparent overlays and point-to-point measurements.
The magnitude of lateral condylar growth during the orthodontic phase (T2-T3) was greater than that during the orthopedic phase (T1-T2) for all condylar fiducials with the exception of the superior condyle ( P <0.05). Conversely, posterior condylar growth was greater during the orthopedic phase than the subsequent orthodontic phase for all condylar fiducials ( P <0.05). The magnitude of vertical condylar development was similar during both the orthopedic (T1-T2) and orthodontic phases (T2-T3) across all condylar fiducials ( P <0.05). Posterior condylar growth during the orthodontic phase (T2-T3) of the 2-phase approach decreased for all condylar fiducials with the exception of the posterior condylar fiducial ( P <0.05) when compared with the single-phase approach.
Two-phase treatment using a Herbst appliance accelerates condylar growth when compared with a single-phase regime with Class II elastics. Whereas the posterior condylar growth manifested primarily during the orthopedic phase, the vertical condylar gains occurred in equal magnitude throughout both phases of the 2-phase treatment regime.
This is a follow-up study of Atresh et al, previously published in AJO-DO in August 2018.
Adolescent subjects with Class II malocclusions were treated with the Herbst appliance or Class II elastics with 3D mandibular voxel-based superimposition used to describe condylar changes.
Two-phase treatment using a Herbst appliance accelerated condylar growth when compared to single-phase treatment with Class II elastics.
Posterior condylar growth manifested primarily during the orthopedic phase of the 2-phase regime.
Vertical condylar gains occurred equally in both orthopedic and orthodontic phases of the 2-phase treatment regime.
Untreated Class II malocclusions do not appear to be self-corrective at the occlusal level despite favorable skeletal base change. , Nonextraction Class II treatment regimens can generally be divided into either 2-phase treatment, which involves the use of a functional appliance and subsequent multibracket appliances, or alternatively, a single-phase treatment regime, which involves Class II mechanics (ie, Class II elastics) concurrent with fixed appliance therapy. Two-phase treatment has been associated with an acceleration but not an increase in absolute mandibular length, increased overall treatment times, and a relapse of the dentoalveolar movements between phases when compared with a treatment protocol that does not involve the use of a functional appliance.
Introduced in 1905 by Emil Herbst, the Herbst appliance is a rigid noncompliant intermaxillary appliance designed to keep the mandible in a continuous protrusive position using the maxillary and mandibular dentitions as anchorage units by means of bilateral telescopic arms. This continuous “bite jumping” is maintained while still permitting the opening and lateral excursive movements involved with eating or speaking. , Closure of the mouth can only occur with the mandible in a protruded position. It has been proposed that this protrusion induces condylar growth stimulation as an adaptive response to the forward positioning of the mandible, with the possibility of some degree of mandibular growth redirection.
The noncompliant nature of the Herbst appliance and its continuous bite jumping effect make it the ideal clinical model to study the capacities of functional appliances to produce orthopedic change. One of the main limitations of studies using removable functional appliances has been the role that compliance may play in delivering the therapeutic intervention. Compliance with removable functional appliances may also be limited owing to impairment of speech, sleeping patterns, and family relationships while in treatment and may explain why 34% of patients who were given a Twin-block did not complete their functional appliance treatment. Conversely, the Herbst appliance approximates what occurs in animal studies because the actual period of intervention is relatively short and compliance factors are eliminated. ,
The contemporary literature supports that at least in the short- to medium-term (ie, 30 months) there appears to be greater skeletal contribution to the overall degree of sagittal correction in the 2-phase group. During the orthopedic phase (T1-T2) with the Herbst appliance, evidence supports that there is 1.6 mm of additional sagittal temporomandibular joint (TMJ) displacement in the Herbst group, with similar amounts of vertical TMJ displacement in both single-phase and 2-phase protocols after the orthodontic phase (T2-T3). However, the overall orthopedic contributions to a Class II correction with the 2-phase regime appear to decline and even reverse over time as treated individuals enter late adolescence and beyond. There appears to be more similarities than differences in the long-term between single-phase treatment with Class II elastic use and 2-phase treatment with the Herbst appliance. Thus, the main benefit of the Herbst appliance from a cephalometric perspective may simply be stated as “you get the growth when you need it” in subjects with Class II malocclusions. The acceleration of mandibular growth after continuous bite jumping during the orthopedic phase is believed to allow for approximately equal continued maxillary and mandibular growth during the subsequent orthodontic phase, with the dentoalveolar Class II correction maintained by subsequent interdigitation.
Lateral cephalometric studies are limited for assessing the magnitude of condylar changes owing to projection error and difficulties in landmark identification, especially when changes are small. Methods that seek to minimize the errors associated with landmark identification , , are associated with superimposition errors, which tend to be exaggerated in the condylar area and fail to distinguish the contributions of condylar change from physiological and/or therapeutic glenoid fossa changes. Furthermore, reference frame orientation affects the perceived direction of condylar change, which can affect mandibular projection owing to its relationship to mandibular growth rotation. Although the Pancherz analysis has been used widely in the contemporary 2-dimensional (2D) lateral cephalometric literature, it may not be a valid means of describing condylar growth changes in the overall dentofacial complex owing to its reliance on the mandibular occlusal plane of the initial radiograph. Lateral cephalometric studies also cannot differentiate temporary displacement of the mandible from true mandibular growth. ,
Recently, Souki et al were able to demonstrate the use of cone-beam computer tomography (CBCT) and mandibular voxel-based 3-dimensional (3D) regional superimpositions in growing adolescents, during the active period of Herbst appliance treatment with comparisons with matched untreated controls using a common 3D reference frame. Although there have been some initial forays into the determination of condylar changes that occur during the initial orthopedic phase (T1-T2) of 2-phase treatment, there has been no 3D quantification of these condylar changes in the orthodontic phase (T2-T3) of 2-phase treatment or in single-phase treatment with Class II elastics (T1-T2).
The goals of this study were to quantify and qualify the 3D condylar changes during the initial orthopedic phase (T1-T2) with the Herbst appliance and subsequent orthodontic phase (T2-T3) of a 2-phase treatment regime for adolescents’ Class II Division 1 malocclusions, with the 2-phase treatment regime with the Herbst appliance (T1-T3) compared with a group of matched subjects treated with Class II elastics in a single-phase treatment protocol (T1-T2) using mandibular 3D regional superimposition techniques.
Material and methods
Ethics approval for this retrospective study was obtained from the University of Melbourne Human Research Ethics Committee (HREC 1647867). All subjects who were treated with a 2-phase treatment regime with the Herbst appliance were sourced from the office of a specialist orthodontist from Geelong, Australia. The subjects were selected by searching the practice database for an item code denoting Herbst appliance insertion. The consent to the use of records was taken in accordance with the Victorian Health Records Act 2001 and Federal Privacy Act 1988 when the subjects started their course of orthopedic and/or orthodontic treatment.
A sample size calculation completed using G∗Power version 3.1.92 determined the following for comparisons of condylar effects during the initial orthopedic (T1-T2) and subsequent orthodontic (T2-T3) phases:
Nine subjects would provide 80% statistical power in detecting a 1-mm difference in a posterior direction for the posterior condylar fiducial assuming a standard deviation of 0.87 mm and a significance level of P <0.05 using values obtained by Souki et al during the active period of Herbst appliance treatment and a paired 2-sided t test with the null hypothesis of no statistically significant difference in the magnitude of posterior condylar growth during the orthopedic and orthodontic phases of a 2-phase treatment regime.
Ten subjects would provide 80% statistical power in detecting a 1-mm difference in the superior direction for the superior condylar fiducial assuming a standard deviation of 0.99 mm and a significance level of P <0.05 using a paired 2-sided t test with a null hypothesis of no statistically significant differences in the magnitude of vertical condylar growth during the orthopedic and orthodontic phases using values obtained by Souki et al during the active period of Herbst appliance treatment.
The Herbst sample consisted of 20 consecutively treated subjects who completed a 2-phase regime using a Herbst appliance that consisted of stainless steel crowns fitted to the maxillary and mandibular first permanent molars, with a cantilevered arm extended forward from the mandibular first molar to the level of the mandibular first premolar. The maxillary framework incorporated a Hyrax expansion screw to accommodate the advancement of the mandibular arch, whereas a well-adapted 0.040-inch stainless steel lingual arch connected the left and right mandibular molars. The lower framework also featured occlusal rests on the mandibular first premolars or second primary molars. Activation of the appliance involved an initial activation of 5 mm with subsequent 2-mm activations to bring the incisors into an overcorrected edge-to-edge position. If there was insufficient overjet to achieve the desired amount of mandibular advancement, then a preorthopedic phase incorporating limited fixed appliances was commenced before the start of Herbst appliance treatment. CBCT scans were taken before the start of treatment (T1), 8 weeks after removal of the Herbst appliance (T2), and when the multibracket appliances were removed (T3). The mean age for the subjects with Herbst appliances for the initial radiograph was 12.76 ± 0.89 years. The activation of the Herbst appliance started an average of 3.09 ± 1.87 months after the initial radiograph, at an average age of 13.01 ± 0.86 years. The mean treatment time for the orthopedic phase with the Herbst appliance was 7.79 ± 1.82 months, and the orthodontic phase was 22.08 ± 3.69 months for a total treatment time of 29.87 ± 4.56 months.
The single-phase Class II elastics group consisted of 11 de-identified, matched subjects with Class II malocclusions with pretreatment (T1) and posttreatment (T2) CBCT scans obtained from collections at the University of North Carolina, Chapel Hill, NC; the University of Minnesota, Minneapolis, Minn; and the University of Michigan, Ann Arbor, Mich. The average age of the Class II elastics group at the initial radiograph was 13.52 ± 0.97 years, and these subjects were treated over a period of 22.97 ± 9.24 months.
Subjects in both 2-phase and single-phase regimes were restricted to adolescents presenting with Class II skeletal (ANB >4°) and dental (bilateral Class II molar relationships >4 mm) relationships who were treated near peak pubertal growth as established by cervical vertebral maturation (CVM) between stages 3 and 4. Subjects were excluded if there was any history of early orthodontic treatment other than limited fixed appliances to provide sufficient overjet for Herbst activation, craniofacial syndrome, or incomplete pre- and posttreatment records, which included 1 or more condyles outside of the field of view at any time point.
All CBCT scans used in this study were taken using an i-CAT machine (Imaging Sciences International, Hatfield, Pa) with a 16 × 22 cm field of view and patients in maximum intercuspation. De-identification and downsizing of the Digital Imaging and Communications in Medicine scans to 0.5 × 0.5 × 0.5-mm voxel sizes were achieved using Slicer CMF (version 3.1, www.slicer.org ) and conversion to the gipl.gz file format to decrease computational time for the eventual 3D voxel-based mandibular superimposition stage. Slicer CMF was then employed to perform head orientation of T1 surface models generated following an automated thresholding algorithm to generate full-skull models by means of the intensity segmenter and model maker modules. The transforms module was then used to align the Frankfort horizontal, midsagittal, and transporionic planes to match the axial, sagittal, and coronal planes of the T1 skull surface models. This step ensured that all T1 mandibles were placed in a common coordinate system within the Slicer software as described by Ruellas et al for subsequent vectorial descriptions of condylar direction.
The 3D volumetric models of the mandible were subsequently segmented in ITK-SNAP (version 3.6; open-source software, www.itksnap.org ) with refinements to the Slicer CMF segmentation by means of the active contour and adaptive segmentation tools available in ITK-SNAP. These refined mandibular segmentations were converted subsequently into 3D surface models using the model maker module within Slicer CMF to produce T1, T2, and T3 surface models. ,
Manual approximation of T2 and T3 CBCTs and segmentations onto the T1 mandible were performed by means of cross-sectional views in all 3 planes of space, in Slicer CMF using the transforms module with the aim of decreasing subsequent computational time. Regions of interest (masks) on the T2 and T3 mandibular segmentations were then identified on the approximate T2 and T3 segmentations in ITK-SNAP. The masks are used to define the stable regions of superimposition for the mandibular body voxel-based image registration as seen in Supplementary Figure 1 . A fully-automated voxel-based growing mandibular superimposition was then performed using the growing registration module in Slicer CMF to generate a matrix that allowed for subsequent superimposition of the T2 and T3 mandibular surface models onto the original T1 mandibular surface model. ,
Manual landmark identification then occurred with 3 open windows of ITK-SNAP to visualize all 2 (Class II elastics) or 3 (Herbst) mandibular time points simultaneously in 3D space. The axial, coronal, and sagittal views of the original gray scale images recreated by multiplanar reconstruction and the 3D volumetric model were used in landmark placement. , Five condylar landmarks as seen in Supplementary Figure 2 were placed on each condyle, which could subsequently be identified using the Q3DC tool in Slicer CMF to allow for comparisons with the work by Souki et al. Quantification of point-to-point linear distance was performed using the m esh s tatistics tool in Slicer CMF for each plane of 3D space. , Condylar changes subsequently could be visualized by semitransparency overlays of the T1, T2, and T3 mandibular surface models. , ,
Data analysis was performed with Excel (Microsoft, version 16.13.1, Redmond, Wash) and MATLAB (Mathworks, version 18.104.22.1681655, Natick, Mass). Means, standard deviations, and 95% confidence intervals were calculated for all subjects. Statistical differences were assessed using paired t tests for intragroup differences between the orthopedic and orthodontic phases in the Herbst subjects and unpaired t tests for overall condylar changes between the 2-phase and single-phase regimes. Manual landmark placement and subsequent Q3DC linear measures were conducted using Slicer CMF 3.1. ,
The reliability of intraobserver manual landmark placement error was assessed with the placement of the 5 condylar fiducials on 5 separate days with linear distances assessed using the Q3DC tool ( Supplementary Table I ). A Bland-Altman plot was performed for the Q3DC , and the m esh s tatistics tool of the same manual landmarks can be seen in Supplementary Figure 3 .
The descriptive statistics comparing the Herbst and Class II elastics groups are summarized in Table I from the 2D cephalometry generated from the CBCTs. The 2 groups were well matched for age, duration in multibracket appliances, ramus length, and divergency before the start of treatment. There were significant differences between the 2 groups, with an increased total duration of treatment in the Herbst and multibracket group compared with the Class II elastics group (respectively, 29.87 ± 4.56 months and 22.97 ± 9.24 months). At T1, the Herbst group also presented with greater mandibular retrognathism, shorter effective mandibular length and corpus length, and a greater ANB than the Class II elastics group (respectively, CoGn 112.06 ± 4.78 mm and 116.54 ± 6.07 mm, GoPg 79.68 ± 4.69 mm and 82.47 ± 3.20 mm, ANB 6.27° ± 1.87° and 3.32° ± 2.14°).
|Measurement||Herbst and multibracket appliances||Class II elastics||P value|
|Sex||6 males||14 females||4 males||7 females||N/A|
|Orthopedic phase, mo||7.79||1.82||N/A||N/A||N/A|
|Orthodontic phase, mo||22.08||3.69||22.98||9.24||0.375|
|Total duration, mo||29.87||4.56||22.98||9.24||0.037 ∗|
|Effective mandibular length CoGn, mm||112.06||4.78||116.54||6.07||0.005 ∗|
|Ramus length CoGo, mm||51.22||3.80||52.79||3.90||0.133|
|Corpus length GoPg, mm||79.68||4.69||82.47||3.20||0.008 ∗|
|ANB, o||6.27||1.87||3.32||2.14||0.001 ∗|
The condylar fiducials featured in this study can be seen in Supplementary Figure 2 on 1 of the mandibles used in this study. The semitransparent overlays of the T1, T2, and T3 mandibular superimpositions for 1 representative subject from the Herbst and multibracket appliances groups are shown in Figures 1 , A and B .
Right- and left-side condylar changes appeared to be symmetrical throughout all time points in the Herbst group and the Class II elastics group, with no statistically significant differences ( P >0.05) between the contralateral condyles, and as such they were combined. This allowed us to treat contralateral condyles as a single group for subsequent intra- and intergroup comparisons.
The means, standard deviations, and 95% confidence intervals for the condylar changes occurring in the Herbst and multibracket appliances groups through the 2-phases and the overall Herbst and multibracket appliances in comparison with the Class II elastics group, can be seen in Tables II and III . Linear distances of condylar fiducial changes during the orthopedic phases (T1-T2) found by Souki et al, who studied the effect of a 1-step activation Herbst in 25 subjects over a period of 7.9 ± 0.4 months are also presented in Table II .
|ROI||Coordinates||Orthopedic phase (T1-T2)||Souki et al group (T1-T2)||Orthodontic phase (T2-T3)|
|Mean||SD||95% CI||Mean||SD||Mean||SD||95% CI|
|ROI||Coordinates||Herbst and multibracket appliances (T1-T3)||Class II elastics (T1-T2)|
|Mean||SD||95% CI||Mean||SD||95% CI|