Three-dimensional assessment of virtual clear aligner attachment removal: A prospective clinical study

Introduction

In digital dentistry, virtual attachment removal (VAR) optimizes clear aligner therapy by enhancing efficiency for refinements and enabling prefabricated retainer production through the removal of attachments from a digital scan before the clinical removal of clear aligner attachments. This prospective clinical study aimed to evaluate the accuracy of VAR in the maxillary arch.

Methods

A total of 110 teeth were analyzed from a sample of 54 maxillary scans from 25 subjects. Models with attachments were virtually debonded using Meshmixer (Autodesk, San Rafael, Calif) and superimposed over the control group in MeshLab. Vector Analysis Module (Canfield Scientific, Fairfield, NJ) was used to calculate and analyze 3-dimensional Euclidean distances on the buccal surfaces between the superimposed models. Statistical analysis was performed using SPSS (version 23.0, IBM, Armonk, NY). The Shapiro-Wilkes (α = 0.05) test determined a nonnormal distribution of results. The Kruskal-Wallis (α = 0.05) was used to determine differences between different tooth types and the number of attachments.

Results

The VAR protocol showed no statistical differences in the root mean square between different tooth segments with an overall tendency for inadequate attachment removal. No difference between the groups was found regarding the number of attachments when used as a main factor.

Conclusions

The VAR technique is precise enough for the fabrication of retainers from printed dental models in a clinical setting and is not affected by the number of attachments on the tooth.

Highlights

•

Virtual attachment removal (VAR) is precise enough for the fabrication of orthodontic retainers.
•

VAR accuracy was not affected by the number of attachments or different tooth types.
•

VAR accuracy may be influenced by the size and location of the attachment and/or flash.

Teeth have been shown to move hours after removing fixed appliances. Therefore, the retention of the final tooth position is essential for maintaining orthodontic treatment results. This requires the use of retention appliances, which can be fabricated from many different materials and have many different shapes.

Lately, the most used orthodontic retention appliance is the thermoformed plastic (Essix; Dentsply Sirona, Charlotte, NC) retainer. A recent survey in Australia has found that 92% of orthodontists include a thermoformed retainer as part of their retention protocol after clear aligner therapy (CAT). For years, the traditional method for fabrication of Essix retainers was to take an alginate dental impression at the braces or aligner removal appointment and proceed to make a plaster stone model used as a template to fabricate the retainer. This method presents many inefficiencies and inconveniences, such as long or multiple patient visits, uncomfortable impression procedures, messy laboratory materials, delayed retention appliance delivery, distortion of the plaster model, and/or stone model breakage. ^,

The global integration of digital dentistry has spurred the adoption of intraoral scanning and 3-dimensional (3D) printing for dental arches, transforming dental offices. ^, This technological shift offers enhanced patient experiences, material and time efficiency, and improved infection control and eliminates physical storage needs. ^, ^, This trend has fueled the popularity of CAT, favored for esthetic appeal, dietary flexibility, easy maintenance, and heightened convenience. Orthodontists are drawn to clear aligners for reduced office visits ^, and the ability to treat a wider range of malocclusions compared with past limitations of CAT. The growing interest in CAT is evidenced by the increasing number of options for both doctors and patients.

Leveraging the benefits of digital dentistry, orthodontic providers can prefabricate retainers before treatment conclusion using virtual bracket or attachment removal (VBR or VAR). This involves obtaining a digital dental scan before treatment completion, followed by the virtual removal of brackets or attachments through computer software. VAR is widely used in many orthodontic settings, including clear aligner companies, dental laboratories, and private practices. Notably, VAR finds frequent use in CAT refinements. In private practice, orthodontists can employ VAR through diverse options, such as Vivera (Align Technology Inc, San Jose, Calif), offering VAR with retainer purchases, or using software options with VAR capabilities such as EasyRx 3D (EasyRx LLC, Atlanta, Ga), uDesign (version 7.0; uLab Systems, Inc, Memphis, Tenn), or Clear Aligner Studio (3Shape, Copenhagen, Denmark) if retainers are made in-house.

VAR can be done at no cost with Meshmixer (Autodesk, San Rafael, Calif), a free computer-aided design and manufacturing software that can be used for general 3D mesh manipulation. The accuracy of this process using Meshmixer was recently evaluated for traditional metal bracket removal and was shown to be accurate enough for the fabrication of clinical retainers from 3D-printed models.

To the best of our knowledge, no published study has examined the accuracy of virtually removing clear aligner attachments or VAR. These attachments are auxiliary devices comparable to brackets and bonded to the enamel of a tooth similarly to help facilitate tooth movement. However, unlike brackets, attachments come in many different shapes, sizes, and numbers per tooth. With CAT commanding up to 20% of the global orthodontic market and increasing in popularity, orthodontists should expect more retention of patients who were treated with clear aligners. It is important to understand the accuracy of VAR to analyze potential factors that may lead to negative clinical implications during and after active orthodontic treatment.

This study aimed to measure the accuracy of VAR in the maxillary arch and compare this accuracy across different tooth types and the number of attachments per tooth. The hypothesis is that VAR will be accurate to use for the fabrication of clinically acceptable orthodontic retainers.

Material and methods

The prospective clinical investigation was conducted at the University of Maryland School of Dentistry, Division of Orthodontics, under the Institutional Review Board no. HP-00098656. Informed consent and assent were obtained from the patients and legal guardians of the participants in the study who met the inclusion and exclusion criteria. The inclusion criteria were as follows: (1) orthodontic patients from the University of Maryland School of Dentistry who were about to start or end treatment and/or refinement with CAT, (2) patients with bonded attachments on the maxillary arch, and (3) patients who agreed to take at least 2 intraoral scans. Exclusion criteria included (1) patients who were not being orthodontically treated with clear aligners, (2) patients who did not have any attachments or plan to have any attachments bonded to teeth as part of CAT, (3) scans that did not fully capture adequate anatomy for superimposition or measurement, or (4) scans taken >30 days apart.

The sample consisted of 54 maxillary scans from 25 subjects, which equated to 216 teeth. Maxillary teeth were divided into 5 different tooth types for measurement: central incisors, lateral incisors, canines, premolars, and molars. Central incisors were separated from lateral incisors because of the large difference in size and anatomy. First and second premolars, as well as first and second molars, have comparatively similar buccal surface anatomy that no differentiation was required. A sample size calculation was run with a power of 90% and α = 0.05. For a statistically meaningful sample, it was determined that 22 teeth in each tooth type (110 teeth total) needed to be measured. Because the data were collected for 216 teeth with attachments, a randomization protocol was used to identify the 22 teeth per tooth type that would be analyzed ( www.randomizer.org/ ).

Each subject agreed to have at least 2 digital scans of the maxillary dentition, 1 with attachments and the other without attachments. The maxillary arch was exclusively used as its unique anatomic soft-tissue landmarks on the palate allowed for easier superimposition. Scans without the attachments comprised the control group, and scans with the attachments were defined as the experimental group ( Fig 1 ). All attachments were previously planned as a part of the digital treatment plan. The control and experimental scans were taken less than 30 days apart to ensure minimal natural tooth movement.

All digital scans were acquired with an iTero Element 2 (Align Technology Inc, San Jose, Calif), which is the scanner available at the University of Maryland orthodontic clinic. The scanner has been shown to have a scanning trueness of 0.0435 mm and a precision of 0.0512 mm. The scans were processed, automatically uploaded, and stored on the servers of myaligntech.com , in which they could be exported as standard tessellation language (STL) files. These files were de-identified from patient information and saved to a digitally encrypted folder.

The VAR protocol is nearly identical to that in a similar study that evaluated VBR using Meshmixer. The experimental group scans were uploaded to Meshmixer. Each attachment was selected via the surface lasso or brush selection tool. If extra composite or flash was seen on the scan, it was also included in the digital selection. The boundary was smoothed using the smooth boundary tool. The content inside this boundary was erased using the default program values of the erase and fill tool ( Fig 2 ). Teeth that had bonded buttons for intermaxillary elastics were sometimes present. In this instance, the buttons were virtually removed for superimposition, but those teeth were excluded from any future measurements. Only the teeth with attachments and buttons were manipulated in Meshmixer. After completion, a new file was created and renamed to acknowledge the completion of VAR.

Superimposition of the control and experimental VAR STL files was completed and verified with MeshLab (version 2022.02) using the iterative closest point algorithm with an accepted error of <0.1 mm. The superimposition protocol was very similar to that in another study also using MeshLab. After uploading the STL files into the software, the option to “unify all vertices” was unselected. The align tab was selected in the toolbar, and the first scanned digital file was “glued” in place so the second digital file could be superimposed over the top using an iterative closest point algorithm. To increase the likelihood of accurate alignment, 7 points were selected each time, although the software only requires 4 points. These points included 4 on the teeth (ie, molar, premolar, canine, and incisors) and 3 points on the soft-tissue palate (left, middle, and right rugae). Points were selected on the basis of their ease of identification. For example, the greatest convexity of rugae and distinct occlusal anatomies were frequently used ( Fig 3 ). The STL files were frozen in MeshLab to maintain positional vertices for superimposition, saved, and uploaded to the Vectra Analysis Module (VAM) (Canefield Scientific, Fairfield, NJ) for 3D assessment of VAR accuracy. In addition, VAM was used for the validation of superimposition using color-coded maps ( Fig 4 ) and by measuring the surface difference at a nonspecific spot on the lingual surface of a central incisor. Superimposition differences ranging 0.0-0.1 mm were accepted.

The VAM software measures the surface difference between the superimposed control and experimental group digital models. The surface differences were measured 1 tooth at a time. It was sometimes difficult to determine which areas of the teeth had attachments removed, so the entire buccal surface of the tooth was selected for measurement with the paint area selection tool. The color surface by distance mode was used for the generation of a color-coded map (±300 μm visualization range) on the selected buccal surface ( Fig 5 , D ). The following values were automatically calculated: minimum difference, maximum difference, and root mean squared (RMS). These measurements were recorded and secured in a Microsoft Excel file (Microsoft Corporation, Redmond, Wash).

A calibrated author (I.S.) was trained on how to use all 3 software programs and was completely responsible for virtually removing attachments in Meshmixer, superimposing the control and experimental groups in MeshLab, and analyzing the surface differences in VAM. Intrarater reliability was assessed by remeasuring 30% of the teeth in VAM 2 weeks after the initial measurement.

Statistical analysis

Minimum, maximum, and RMS values were all recorded and analyzed. Descriptive statistics were performed with SPSS (version 23.0, IBM, Armonk, NY). Intrarater reliability (Cronbach α) was determined. The Shapiro-Wilkes test determined a nonnormal distribution of results, and the Kruskal-Wallis (α = 0.05) was used to determine statistical differences between different tooth types and the number of attachments.

Results

The intrarater reliability was determined to be high (>0.98) for all 3 measured metrics of minimum, maximum, and RMS. The medians and interquartile ranges can be seen in Table I . The Kruskal-Wallis (α = 0.05) showed P values all >0.05, indicating that there were no statistically significant differences between different tooth types in the minimum ( P = 0.282), maximum ( P = 0.493), or RMS ( P = 0.425) values. Molar teeth had the largest median values for minimum (−0.090 ± 0.096 mm), maximum (0.279 ± 0.270 mm), and RMS (0.095 ± 0.101) ( Table I ).

Table I

Only gold members can continue reading. Log In or Register to continue