A voxel-based CBCT superimposition of upper molar distalisation using clear aligners in conjunction with Class II elastics

Clear aligner therapy (CAT) has significantly revolutionised contemporary orthodontics by offering aesthetic and comfort advantages over fixed multibracket therapy (FMB).¹ With advancements in materials and the introduction of auxiliaries, clinicians have increasingly utilised CAT to address complex malocclusions. Maxillary molar distalisation is now considered a viable approach for the treatment of Class II malocclusions.² The findings of retrospective studies inferred that upper molar distalisation using CAT offers effective vertical control.^2,3 Notwithstanding, the discrepancy between an anticipated digital configuration and the actual tooth movement remains a clinical challenge. The existing literature on molar distalisation using CAT lacks a clear consensus, with the predictability of tooth movement varying from 55% to 87%.^4,5 The variability can be attributed to differences in research methodologies and in the reference points utilised, as well as discrepancies in the application of Class II elastics.^6,7

Clinically, Class II elastics are commonly applied to improve anchorage.⁸ Ravera et al. combined Class II elastics with clear aligners which led to a 2.25 mm distalisation of the upper first molar, with minimal crown tipping and vertical movements reported in the superimposed lateral cephalograms.² The study focused on Class II elastics that were applied after the molar distalisation phase, and the evaluation conducted at the completion of the entire treatment rather than specifically at the end of molar distalisation. Additionally, Amm et al. suggested that the early incorporation of Class II elastics alongside the clear aligners before arch levelling as a method to accelerate the sagittal correction of a Class II malocclusion.⁹ The use of Class II elastics in FMB chiefly produces dentoalveolar effects, while facilitating maxillary incisor extrusion and mandibular molar mesialisation.¹⁰ Nevertheless, the use of clear aligners as an overlay appliance was found to partially mitigate the vertical side effects elicited by the Class II elastics.¹¹ Later studies utilised a biomechanical assessment through a finite element technique (FEM) to investigate the effects of integrating clear aligners with Class II elastics during upper-molar distalisation. This process entailed immediate shifts in the initial phases, while force directions consistently varied in practice.^12,13

Cone-beam computer tomography (CBCT) has attracted extensive attention owing to its precise capabilities, lower radiation exposure, and minimal image distortion.^14,15 The utilisation of stable anatomical structures as references for three-dimensional reconstruction and superimposition enables an evaluation of orthodontic treatment outcomes and an analysis of skeletal and dental changes.¹⁶ There has been an increase in the use of voxel-based superimposition in CBCT volumes, as it has been found to be valid, reliable, and repeatable.¹⁷

The present study therefore aimed to assess the displacement of the upper and lower teeth using clear aligners and Class II elastics for molar distalisation by using a voxel-based CBCT superimposition to enhance treatment strategies for clinicians.

The retrospective analysis followed the guidelines of the Helsinki Declaration and received approval from the Ethics Committee of The Air Force Medical University (IRB-REV-2022079). Written permission was provided by all study participants. A sample size of 30 was calculated based on a standard error of 5% and a confidence interval of 95%. A total of 30 individuals, consisting of 13 males and 17 females, were therefore selected for the study. The average age was 25.4 years, with ages ranging from 18 to 35 years. The inclusion criteria identified adults with a Class II malocclusion and without prior orthodontic treatment. The participants were treated by non-extraction methods, using clear aligners along with Class II elastics for maxillary molar distalisation. Patients with syndromes, alveolar bone defects, and those receiving pharmaceutical medications that potentially alter tooth movement were excluded from the investigation.

Each patient was treated using Angelalign® clear aligners to move the maxillary molars by a 0.25 mm step distance, and Class II elastics were placed between the upper canines and lower first molars. The total designed distal molar change was between 1 mm to 5 millimeters. Twenty-four patients had horizontal rectangular attachments bonded on the upper first molars, eighteen patients had vertical rectangular attachments and eighteen patients had no molar attachments. Patients were reminded to follow the treatment guidelines by wearing the aligners for at least 22 hours daily. Therapy involved a sequential distalisation approach, involving a staging of 50% applied in all cases. The patients were also directed to consistently bilaterally wear 3.5 oz Class II elastics. Additionally, each patient underwent two CBCT scans for treatment evaluation, with scans conducted at the beginning of treatment (T0) and at the end of the distalisation phase (T1). The CBCT scans were performed using equipment from GE Healthcare, USA, with a full field of view and a centered rotation. The DICOM data collected at both T0 and T1 were later transferred to Dolphin Imaging software (Dolfin, Chatsworth, CA, USA).

There were ten landmarks selected as reference points for the maxillary and mandibular voxel-based superimposition (Figure 1, Table I).^18,19 The second volume’s head orientation was adjusted based on the skull’s 3D position when overlaid on the base volume. T2U was the label assigned to the second volume adjusted using maxillary superimposition, while T2L was the volume adjusted using mandibular superimposition and exported. The DICOM data at time points T0, T2U, and T2L were imported into Mimics 20.0 Software (Materialise, Leuven, Belgium) for three-dimensional reconstruction using a threshold set above 600HU. In each 3D model, the maxillary and mandibular regions were separately segmented and saved in STL format.

Figure 1.

Voxel-based superimposition using Dolphin software. A. Landmarks for maxillary superimposition; B. Landmarks for mandibular superimposition.

The maxillary teeth and surrounding structures in T0 and T2U were both inputted into Geomagic Studio 2014 software (Raindrop GEOMAGIC, NC, USA) for measuring the upper jaw. Similarly, the mandibular bone and dentition at T0 and T2L were simultaneously imported for mandibular measurements. A horizontal plane parallel to the line connecting OrL and OrR passed through ANS and PNS, while a sagittal plane was defined as the plane passing through N, ANS, and PNS, and a coronal plane was defined as the plane passing through ANS and perpendicular to both the horizontal and sagittal planes.^18,19 The sagittal plane was used to measure the disto-mesial tipping angles of the molars and the labio-lingual tipping angles of the incisors between each vector at T0 and its corresponding vector at T2, as shown in Figure 2. In addition, the rotation angle of the molars was assessed on the horizontal plane using the line between the mesio-buccal cusp and the disto-lingual cusp as landmarks. The bucco-lingual tipping angles of the molars were measured in the coronal plane. Five points on the upper and lower first molars were identified on the T0 and T2 models to determine the achieved amount of tooth displacement. The arch width was measured by the facial axis (FA) points of the first molars.

Figure 2.

Measurement of tooth movements. A. Disto-mesial tipping angles of the upper molars and labio-lingual tipping angles of the upper incisors; B. Disto-mesial tipping angles of the lower molars and labio-lingual tipping angles of the lower incisors; C. Rotation angle of the upper molars; D. Rotation angle of the lower molars; E. Displacement of the upper molars; F. Displacement of the lower molars.

Data analysis was performed using IBM SPSS Statistics for Windows, version 26.0 (IBM, Armonk, NY, USA). Continuous variables in the study were normally distributed and were expressed as means and standard deviations (SD). The connection between anticipated and actual movements was illustrated through scatter plots, Pearson’s correlation values, and basic linear regression examinations. Multiple linear regression models were applied to investigate the impact of various factors. A p-value less than 0.05 was deemed to be statistically significant and 20% of the models underwent a re-analysis by both the original examiner and a second examiner after an interval of 2 weeks following the initial measurements.

A total of 60 upper molars and 60 upper incisors were examined to evaluate the accuracy of molar distalisation and the level of anchorage loss involving the upper incisors. In addition, 60 lower molars and 60 lower incisors were evaluated to determine the actual movement of the mandibular dentition. Reproducibility was evaluated using the intraclass correlation coefficient (ICC), which showed strong agreement with a value of 0.96 for examiner reliability and 0.94 for reliability between examiners. Table II presents the descriptive statistics of the actual tooth movements. The upper molar aligner distalisation efficiency in conjunction with the Class II elastics was 60% ± 7% of the designed movement. The distalised molars displayed distal and buccal tipping and accompanying disto-buccal rotation. The upper incisors showed labial tipping of 3.84° ± 1.1° due to anchorage loss. The lower first molars exhibited mesial and lingual tipping, as well as a mesial shift of 0.64 mm ± 0.27 mm, due to the force exerted by the Class II elastics. There was a slight increase of 1.68 mm in upper arch width at the first molar level, which was not statistically significant (p > 0.05), while the lower arch width at the first molar level had a minor decrease of 0.63 mm, also not statistically significant (p> 0.05).

It was noted that Pearson’s correlation coefficients revealed a moderate-to-high correlation between the predicted upper molar distalisation and achieved distalisation (r = 0.9917, p < 0.0001), actual upper molar distal tipping (r = 0.5676, p < 0.0001), actual upper molar rotation (r = 0.7113, p < 0.0001), actual upper incisor tipping (r = 0.7903, p < 0.0001), actual lower molar medialisation (r = 0.9073, p < 0.0001), and actual lower incisor tipping (r = 0.8394, p < 0.0001), as presented using scatter diagrams and simple linear regression equations (Figure 3). The results of the simple linear regression analysis indicated that a 1 mm prediction in molar distal movement corresponded to an average achievement of 0.64 mm in molar distalisation, 2.42° in molar distal tipping, 2.24° in molar rotation, 2.39° in upper incisor labial tipping, 0.30 mm in lower molar mesialisation, and 1.02° in lower incisor tipping.

Figure 3.

Scatter diagram, Pearson’s correlation coefficients, and simple linear regression analyses of the predicted upper molar distalisation and achieved tooth movement.Table III. Multiple linear regression models for predicted and achieved tooth movements

Table III outlines the influence of predicted upper molar distalisation, gender, attachment, and initial crowding on the achieved tooth movement using multiple linear regression models. Male and female patients showed comparable values in all measurements. Patients with a horizontal or vertical rectangular attachment on the upper molars exhibited an inverse correlation with molar distal tipping (horizontal and vertical attachments: p < 0.05) and molar rotation (vertical attachments: p < 0.0001) compared with those with no attachment. The initial crowding had a negative correlation with upper molar distalisation success and a positive correlation with upper molar distal tipping, upper incisor tipping, and lower incisor tipping. Mild crowding showed significance at p < 0.05 for upper molar distal tipping, while moderate crowding had significance at p < 0.05 for upper incisor tipping.

The use of clear aligner therapy cannot fully achieve designed tooth movement goals due to material limitations and the biomechanical response of the teeth.²⁰ Herein, discrepancies between predicted tooth movements and actual treatment outcomes of upper molar distalisation were measured to help clinical practitioners to understand clear aligner therapy outcomes. Previous studies have evaluated the effectiveness of clear aligners for molar distalisation by superimposing cephalometric radiographs² and superimposing digital models of teeth^5,6 or palatal rugae.^7,21 D’Antò et al. reported that molar distalisation efficacy varied between 68.0% and 79.9% based on the superimposition of digital models using untreated teeth as reference points, wherein anchorage teeth mesialisation is typically overlooked.⁶ Moreover, Simon et al. documented a molar distalisation accuracy of 87% in the superimposition of plaster cast models, using untreated teeth as reference points. However, this approach was relatively unreliable due to anchorage loss.⁵ An accuracy of 73.8% was reported in a comparison between predicted and achieved upper molar distalisation using the palatal rugae for the superimposition of digital models, despite the absence of Class II elastics during the distalisation process. However, there was no stable landmark for mandibular superimposition using oral scan digital models.

CBCT is widely used in orthodontics owing to its accurate 3D outcomes. A voxel-based comparison allows superimposition on stable skeletal anatomic landmarks and accurately matches structures using high grayscale levels, which renders greater accuracy and reliability for assessing treatment effects.^17,21 In the present study, a skeletal superimposition technique revealed a 60% ± 7% decrease in upper molar distalisation efficiency when using clear aligners and Class II elastics, which suggested a need for greater over-correction in clinical applications. By proposing a linear regression coefficient, the achieved tooth movement can be inferred from the designed amount of molar distalisation. Due to anchorage loss, the average labial tipping of 2.39° in the upper incisor and 1.02° in lower incisor tip were achieved when 1 mm molar distalisation was predicted. This implies that a suitable torque design needs to be incorporated in order to improve anchorage control. Furthermore, the mesialisation of the lower molars by 0.30 mm, despite a predicted 1 mm molar distalisation, was attributed to a counterforce associated with the Class II elastics. Therefore, the correction of a Class II malocclusion is not solely dependent on the distalisation of maxillary molars, but also on a small amount of mesialisation of the mandibular molars. If the mandibular molar mesialisation is not expected, a corresponding anchorage consideration should be designed.

A slight increase in the width of the upper dental arch was observed at the first molar level, possibly due to the buccal tipping of the distalised molars. Recent research has indicated that using clear aligners for moving the maxillary molars can effectively widen the arch, potentially creating more room to move the anterior teeth back.^22,23 However, the objects in the studies were designed for dental arch expansion, and the achieved arch width increase was more than the original design. Therefore, the upper arch width increase induced by molar distalisation was present in small quantities. Additionally, a slight reduction in the width of the lower dental arch was observed at the first molar position, possibly due to the pressure from the Class II elastics and the lingual tilt of the mandibular first molars.

The tooth movement trend is consistent with that of previous finite element (FEM) studies; nonetheless, direct numerical comparisons are challenging, given that FEM exclusively involves instantaneous movement in the initial stages without a consideration of complicating factors.^12,13

In the present study, multiple factors influencing the achieved tooth movement were explored using multiple linear regression models. The study subjects were all adult patients; consequently, growth and development considerations were not applicable. Of note, gender effects on tooth movement were not significant. Patients who had attachments either horizontally or vertically bonded on their upper molars exhibited reduced tipping and rotation of distalised molars in comparison to those without attachments. This aligns with findings from earlier research, in which the group lacking attachments displayed uncontrolled tipping, whereas the attachment groups showed better control over bodily movement.^24,25 Gao et al. found no significant difference in the effects of horizontal and vertical rectangular attachments on molar distalisation.²⁶ Furthermore, Garino et al. discovered that attachments play a crucial role in improving posterior anchorage during the distalisation and anterior retraction stages of orthodontic treatment.²⁷

Initial crowding was demonstrated to substantially impact anchorage loss during premolar extraction space closure during fixed orthodontic treatment.²⁸ Dai and colleagues reported that clear aligner therapy of initial crowding was correlated with anchorage loss in extraction patients, whereby greater crowding resulted in mesial tipping and translation of the first molars.²⁹ However, as highlighted in the present study, in non-extraction patients, increased initial crowding corresponded to reduced upper molar distalisation. In addition, initial crowding was positively correlated with upper molar distal tipping, plus upper and lower incisor tipping. The findings collectively indicate a higher demand for over-correction and anchorage preparation in patients requiring molar distalisation within a mild or moderately crowded dentition.

Nevertheless, the present study had limitations that require consideration. The investigation did not include an assessment of patient compliance. Additionally, factors related to age, periodontal health, and tooth root length, which can influence tooth movement, were not considered. Future studies should aim to increase the sample size and further explore other anchorage control methods used in conjunction with clear aligners.

The present study explored the achieved tooth movement using clear aligners and Class II elastics used for upper molar distalisation via a voxel-based CBCT superimposition. The results quantitatively revealed distal and buccal tipping along with a disto-buccal rotation of the upper molars, labial tipping of the upper and lower incisors, and mesialisation of the lower molars. The upper arch width displayed a non-significant increase while the lower arch width showed a non-significant decrease. Molars with attachments exhibited less distal tipping and rotation. Initial crowding was inversely correlated with the achieved upper molar distalisation but positively correlated with tipping.