INTRODUCTION

To produce Invisalign clear aligners, there is a digital workflow in which the interaction between orthodontists and Align Technology takes place through a virtual tool, ClinCheck software (Align Technology, San Jose, Calif).¹ This platform displays virtual patient models with predictions of sequential treatment steps to be performed with the aligners. Therefore, many studies have been carried out to establish validity between predictability and the clinical results obtained after treatment completion, while Invisalign becomes increasingly used widely among professionals.^2–5

Among the information provided by ClinCheck, the Bolton discrepancy (BD) is displayed. This analysis identifies the proportions of tooth size between the upper and lower arches⁶ and is directly related to the amount of interproximal reduction (IPR) or resin additions that are indicated to achieve ideal occlusion. This analysis can be performed digitally using virtual three-dimensional (3D) models, or it can be assessed by the conventional technique initially described by Bolton, using physical models and calipers.^7,8 For clear aligners, correct measurement of mesiodistal diameter of the teeth is even more important because it provides for the creation of adequate space to achieve desired tooth movements.⁹ Also, it is related to the correct indication of IPR to obtain space when necessary,^10,11 minimizing iatrogenic effects in treatment with clear aligners.

Information is lacking on how the BD analysis is obtained through the ClinCheck tool. Therefore, the aim of the present study was to evaluate the diagnostic performance of ClinCheck as well as Dolphin Imaging (Dolphin Imaging and Management Solutions, Chatsworth, Calif) and 3D Slicer software (Kitware, Clifton Park, NY) to assess BD.

MATERIALS AND METHODS

The study was approved by the Research Ethics Committee of the Clementino Fraga Filho University Hospital (CEP-HUCFF) of the Federal University of Rio de Janeiro (UFRJ) under No. 5.202.503.

RESULTS

Method Error

Measurements of BDs in millimeters using the conventional method, Dolphin Imaging software, and 3D Slicer were reliable and accurate (Table 1). The ICC for repeated measures of the anterior and overall Bolton ranged from 0.86 to 0.96 and from 0.80 to 0.91, respectively.

Table 1. Method Error for the Measurement of Bolton Discrepancy in Millimeters^a

View Fullscreen

Main Findings

Results of the reliability, precision, and accuracy assessments of the measurements of BD in millimeters are described in Table 2. The three methods showed reliable measurements (ICC ≥ 0.7). Those obtained for the anterior Bolton analysis were slightly more reliable than those obtained for the overall Bolton. Similarly, measurements for the anterior Bolton showed higher precision compared with the overall Bolton. Although the methods showed good ICC values (high correlation between assessments), only the measurements obtained for the anterior Bolton using the ClinCheck tool and Dolphin Imaging were accurate (paired t-test P > .05; no proportion bias). These methods had a precision of 0.65 mm and 0.70 mm, respectively. All the other measurements were inaccurate, evidencing proportion bias.

Table 2. Reliability, Precision, and Accuracy of the Methods to Measure the Bolton Discrepancy in Millimeters^a

View Fullscreen

The Bland-Altman plots showed that the measures for the anterior Bolton in 3D Slicer and for the overall Bolton using the three software programs were scattered widely, above zero, demonstrating no consistency with measurements of the conventional method and tending either to (1) decrease the excess in the mandibular arch; (2) increase the excess in the maxillary arch; or (3) in case of minor discrepancies, indicate that the excess was in the upper arch when it was in the lower arch (Figure 1). When the mean differences between each of the tested methods and the conventional method were compared, no significant differences were observed for the anterior Bolton between the ClinCheck and Dolphin Imaging assessments (P > .999), while measurements using 3D Slicer software did show differences with both (P < .01; Figure 2). When the overall Bolton was analyzed, no significant difference was observed among the groups (Figure 2).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

DISCUSSION

Success with Invisalign occurs when the occlusal results predicted by ClinCheck are consistent with those obtained during actual treatment.² Thus, some parameters of this tool have been studied, such as the predictability of overbite correction,^12,13 transverse changes in the arch,^14,15 accuracy of measurements performed in the anterior region such as arch length, intercanine distance, overjet, overbite, dental midline shift,¹ and accuracy of tooth movements predicted by this system.^5,16,17 However, information is lacking on the validity of BD data indicated by ClinCheck, and considering its importance to diagnosis and orthodontic planning procedures, the present study was undertaken.

The gold standard reference for comparison was measurements performed on printed models using stereolithographic technology since the prototype models have demonstrated similar precision and clinical application to plaster models.^18,19 The prototypes were printed with a layer thickness of 50 µm for accuracy.²⁰ The models were arranged at 0° to the printing platform, minimizing the expansion effect in the terminal region of the arch.²¹

The results indicated that, although the measurements obtained were reliable, only the ClinCheck tool and Dolphin Imaging software showed accuracy for BD measurements performed in the anterior region. Authors of two previous studies investigated the diagnostic performance of ClinCheck compared with manual methods.^9,22

Shailendran et al⁹ performed segmentation of the teeth to measure their mesiodistal diameter and assess BD. The findings indicated that ClinCheck tended to underestimate tooth width measurements by an average of 0.36 mm, while for molars, the average was 0.9 mm. As ClinCheck accuracy decreased as the measurements became farther posterior in the arch, a greater probability of underestimation was found in these areas, while more accurate values were found in the anteroinferior region,⁹ a pattern that was also observed in the current study.

Martin et al.²² compared the values indicated by ClinCheck software with measurements performed on plaster models and found higher values provided by the software, with the largest measurement errors also in the molar region. The measurement methodology was like that of the present study, in which the actual position of the teeth in the arch was considered, along with crowding, inclination, and existing dental rotation. Such factors present increased difficulty for accurate measurement of BD²³ but also represent the classic evaluations performed on plaster models, developed by Bolton himself in 1962.⁶

Concerning overall Bolton assessment, ClinCheck, Dolphin Imaging, and 3D Slicer showed moderate reliability, but the values were significantly different than the measurements performed on the printed models, demonstrating that the measurements performed were not accurate. However, the variation was less than 1.5 mm which, according to Proffit et al,²⁴ would have little clinical implication.

The current findings were like those of previous studies, in which authors observed satisfactory agreement between the Bolton analysis for anterior and overall proportions when comparing measurements in 3D models using software and plaster models. In those studies, the digital models were obtained by digitalizing plaster models and not with intraoral scanning, as performed in the present study.^25–28 Regarding ClinCheck, the correlation found in the study by Shailendran et al⁹ was moderate, with an ICC of 0.83 for overall Bolton results, like that found in the present study.

Some explanation may exist as to why the reproducibility of overall Bolton measurements was not better. Acquisition of digital measures may be easier²⁹ and require less time than manual methods,^25,26 but divergence or imprecision may occur in marking the reference points of the measurements, even when performed by the same operator over a short time interval.^23,27 This limitation may be related to the 3D nature of the models which is viewed two-dimensionally,²⁹ making it difficult to identify points, axes, and planes by the software.^23,26,30 For Dolphin Imaging, 3D features are lost since the software uses a static image of the dental arch in the occlusal view to mark the points.

Another factor would be the difficulty of scanning to faithfully reproduce posterior teeth due to patient-related factors, such as salivation.³¹ The software fills in such regions through a mathematical algorithm, leading to changes in tooth shape that generate measurement inaccuracies.¹

Finally, inherent difficulties may exist in measuring posterior regions using scanning technology. Authors of studies comparing sizes in digital vs conventional models have indicated that, despite having high reproducibility, digital models tended to decrease the intermolar width and increase the palatal depth.²³ This systematic error in the reproduction of the molar surface was also described by Zilberman et al.³⁰ For the ClinCheck tool, authors of studies indicated that the software overestimated the alignment and posterior occlusal relationships obtained after treatment.² Further studies are needed regarding the dental information provided by ClinCheck, especially pertaining to the posterior region of the arch.

Considering common clinical application, evidence of the accuracy of the anterior Bolton values using ClinCheck and Dolphin Imaging is encouraging since intervention in this region is more frequent, either by IPR in cases of excess or by addition of resin or by restoration to modify atypical shape of teeth to achieve proper occlusion. However, in the posterior region, IPR is less often indicated because it may increase the chance of food impaction and caries risk but may still be indicated to alter abnormal shape. It is recommended that some caution may be appropriate when assessing and interpreting BD data obtained through orthodontic software.

CONCLUSIONS

• ClinCheck and Dolphin Imaging software demonstrated precision and accuracy for anterior BD analysis measurements, demonstrating validity to be used during orthodontic diagnosis.

• During analysis of overall Bolton relationships, ClinCheck, Dolphin Imaging, and 3D Slicer showed more variability and tended to decrease excess in the lower arch and increase excess tooth size in the upper arch. However, the variations shown did not have clinically relevant implications.