Diagnostic accuracy of impression-free digital models

Diagnostic accuracy of impression-free digital models

American Journal of Orthodontics and Dentofacial Orthopedics, 2013-12-01, Volume 144, Issue 6, Pages 916-922, Copyright © 2013 American Association of Orthodontists

Introduction

Impression-free techniques might eliminate the potential shortcomings of digital dental models. Chairside scanners offer the advantage of obtaining digital dental models directly from the patient without the need for dental impressions. The aim of this study was to evaluate the accuracy of 3-dimensional digital models acquired from a chairside intraoral scanner compared with both manual and cone-beam computed tomography measurements of the same dental anatomy.

Methods

The study sample comprised 60 dry skulls. Each skull had the maxillary and mandibular arches scanned with a Cadent iTero scanner (Align Technology, San Jose, Calif) and had a cone-beam computed tomography scan taken with a CS 9300 unit (Carestream Health, Atlanta, Ga). Linear measurements in all 3 dimensions of the space in each dental arch together with tooth-size arch-length analysis for both the maxillary and mandibular arches were carried out manually on the dry skulls with calipers and digitally on the scanned 3-dimensional models and cone-beam computed tomography images. Intraclass correlation (ICC) analysis was performed for all variables tested in the study groups, with the manual measurements on the dry skulls as the gold standard. The Bland-Altman analysis was also applied to the data to graphically display the agreement of the diagnostic measurements obtained from these methods.

Results

Measurements from the iTero models demonstrated near-perfect agreement (ICC, 0.91-0.99) with the caliper measurements. Cone-beam computed tomography measurements had moderate to high levels of agreement (ICC, 0.65-0.99) compared with the caliper measurements.

Conclusions

Direct digital acquisition of the dental arches with a chairside scanner provided almost 1-to-1 diagnostic information of the investigated anatomy and was superior to the cone-beam computed tomography measurements.

Through advances in biomaterials, robust and dimensionally accurate dental plaster models are possible and have long been the gold standard in orthodontics. In the new digital era, various technological advancements have made their way into dental practices. The past decade has seen the advent of digital models with acceptable quality, allowing the orthodontic record to become completely digitized.

For orthodontics, the most important expectation from a digital model system lies in its diagnostic accuracy and reliability. Although the consensus is that measurements with digital models compare well with those derived from plaster models, several studies that investigated complex measurements such as space available, irregularity index, and Bolton analysis indicated that mean differences between the plaster and digital models can exceed 1.5 mm. This much difference might not be clinically acceptable. However, there is also contrary evidence in the literature that supports the validity of digital models for the aforementioned measurements.

Many professionals obtain their digital models through the use of proprietary services. Traditional impressions and plaster models must be submitted to the selected company so that they can be scanned and the digital models can be generated and made available for downloading. Submission of models or impressions by mail can result in breakage of the models or distortion of the impressions during shipment. Even if no breakage or distortion occurs during shipment, the impression materials have an inherent amount of distortion at their time of use. Moreover, there are technique-specific differences in the production phase of digital models such as surface scanning algorithms. All of these might account for the statistical variability related to the accuracy of digital dental models.

Perhaps elimination of the conventional impression step and generation of dental models directly from the 3-dimensional (3D) dental anatomy will yield more accuracy. One way of doing this is to use the DICOM files produced by cone-beam computed tomography (CBCT) scanners. An advantage of this technology is that it allows for the visualization of supragingival structures as well as impacted teeth, bone levels, and joints. Studies using various scanners have been published showing that diagnostically accurate measurements can be made from CBCT scans. Perhaps the greatest drawback of using these scanners to acquire digital models is the required radiation exposure to the patient and the equipment expense to the practitioner. Some authors have also expressed concern over the absence of gingival tissues in the resultant files and the inability to accurately depict buccal alveolar bone.

For those wishing to gain an accurate depiction of gingival tissues and avoid exposing the patient to radiation, an intraoral scanner might be more appropriate. Intraoral scanners, now widely available from several manufacturers, are self-contained in rolling units, making them portable around the office. In this study, we evaluated the diagnostic accuracy of an intraoral scanner and CBCT as sources of digital models from natural dentitions for orthodontic diagnosis compared with direct caliper measurements of the investigated dental anatomy.

Material and methods

The study sample comprised 60 dry skulls with intact dentition chosen from a unique collection at the University of Texas Health Science Center in Houston.

The Carestream CS 9300 (Carestream Health, Atlanta, Ga) was used to capture the CBCT images ( Fig 1 ) of the skulls with exposure parameters of 90 kV, 4 mA, and 300-μm voxel size and volume dimensions of 17 × 13.5 cm. The image reconstruction time was approximately 5 minutes. Each arch was then scanned using the reference model setting on the Cadent iTero scanner running software version 4.0.5.31 (Align Technology, San Jose, Calif). The raw images were available for chairside viewing within 2 minutes after completion of the scanning. These images were then sent via the Internet to Align Technology, where they were reprocessed and made available for downloading within 48 hours as a stereolithography file ( Fig 2 ). This file format is an industry standard native to stereolithography computer-aided design and can be easily viewed and manipulated in various software applications without loss of quality.

Three-dimensional reconstructions of the dental arches with CBCT.
Fig 1
Three-dimensional reconstructions of the dental arches with CBCT.
Dental models generated by the iTero scanner.
Fig 2
Dental models generated by the iTero scanner.

With digital calipers, a tooth-size arch-length analysis was performed in each arch. In addition to crowding measurements, 3 teeth in each arch were measured in the anteroposterior, transverse, and vertical dimensions ( Fig 3 ). For the transverse measurement, each tooth was measured at the center of its long axis at its most convex points on the buccal and lingual surfaces ( Fig 3 , A ). The anteroposterior measurement was defined as the most mesiobuccal contact point to the most distobuccal contact point ( Fig 3 , B ). In the vertical plane, each tooth was measured on the buccal surface at the middle of its long axis from the height of the crestal bone to the cusp tip ( Fig 3 , C ). In the maxillary arch, measurements were made on a molar, a premolar, and a canine. In the mandibular arch, measurements were made on a molar, a premolar, and an incisor ( Table I ).

Linear tooth measurements used in this study: A, transverse; B, anteroposterior; C, vertical planes.
Fig 3
Linear tooth measurements used in this study: A, transverse; B, anteroposterior; C, vertical planes.
Table I
Linear measurements used in the study
Label Description
UM MD Mesial distal width of upper permanent molar
UM V Vertical height of upper permanent molar
UM BL Buccolingual width of upper permanent molar
UP MD Mesiodistal width of upper permanent premolar
UP V Vertical height of upper permanent premolar
UP BL Buccolingual width of upper permanent premolar
UC MD Mesiodistal width of upper canine
UC V Vertical height of upper canine
UC BL Buccolingual width of upper canine
LI MD Mesiodistal width of lower permanent incisor
LI V Vertical height of lower permanent incisor
LI BL Buccolingual width of lower permanent incisor
LP MD Mesiodistal width of lower permanent premolar
LP V Vertical height of lower permanent premolar
LP BL Buccolingual width of lower permanent premolar
LM MD Mesiodistal width of lower permanent molar
LM V Vertical height of lower permanent molar
LM BL Buccolingual width of lower permanent molar

The CBCT files were opened using the InVivo Dental software (version 5.1.9; Anatomage, San Jose, Calif), and measurements were made in millimeters on sectioned views of a gray-scale 3D reconstruction using the software’s built-in tool for measuring the distance between 2 defined points. Scanner stereolithography files were opened in Studio Basic (version 0.1; nettfabb, Lupburg, Germany). The measurements were made on unsectioned solid shaded models using the application’s built-in ruler tool.

Statistical analysis

Intraclass correlation analysis (ICC) was performed for all variables tested in the study groups, with the manual measurements from the dry skulls serving as the gold standard. ICC values were computed using the 2-way random effect model (absolute agreement definition). SPSS statistical software (version 21; IBM, Armonk, NY) was used to analyze the data. Additionally, the Bland-Altman analysis was performed with XLSTAT Mac (version 2012; Addinsoft, New York, NY) to visually demonstrate the agreement of the maxillary and mandibular crowding values between the manual caliper measurements and the 2 methods tested. The within-observer repeatability was evaluated with ICC values by repeating all measurements from 10 randomly selected skulls 1 month later.

Results

Descriptive statistics for the measurements are presented in Table II . Intraobserver repeatability was excellent with ICC values between 0.97 and 0.99. Table III displays the ICC values for linear and crowding measurements for both CBCT and the iTero models compared with the manual caliper measurements. The ICC values for the linear measurements in the comparison of caliper measurements with CBCT measurements ranged between 0.88 and 0.99, displaying near-perfect agreement, except for the buccolingual measurement of the mandibular incisor (ICC, 0.65). The ICC values of the comparison of caliper measurements with the measurements made on the iTero models ranged between 0.92 and 0.99, showing a much higher level of agreement. However, both methods had ICC values greater than 0.98 for maxillary and mandibular crowding compared with manual measurements, indicating strong evidence of their diagnostic capability.

Table II
Descriptive statistics for the measurements used in the study
Variable (mm) Dry skulls CBCT iTero
Mean SD Mean SD Mean SD
UM MD 9.86 0.64 9.84 0.69 9.83 0.67
UM V 7.86 1.49 7.73 1.53 7.81 1.47
UM BL 10.67 0.66 10.62 0.74 10.69 0.63
UP MD 7.02 0.89 7.07 0.95 6.94 0.91
UP V 9.25 1.85 9.22 1.9 9.23 1.86
UP BL 9.17 0.56 9.23 0.56 9.19 0.62
UC MD 7.26 0.69 7.34 0.71 7.2 0.63
UP V 12.21 2.21 12.25 2.29 12.1 2.21
UC BL 7.68 0.89 7.66 0.89 7.61 0.9
LI MD 5.76 1.42 5.6 1.48 5.51 0.41
LI V 11.46 1.82 11.56 2.03 11.32 1.63
LI BL 5.86 0.41 5.98 0.56 5.86 0.44
LP MD 7.21 1.28 7.28 1.27 7.16 1.29
LP V 8.56 1.6 8.49 1.54 8.55 1.57
LP BL 7.96 0.71 8.04 0.64 7.98 0.68
LM MD 10.66 0.7 10.75 0.66 10.61 0.72
LM V 7.6 0.98 7.29 1.04 7.56 1.01
LM BL 10.03 0.62 9.98 0.59 10.01 0.61
Mandibular crowding 4.18 6.28 4.36 6.25 4.25 6.28
Maxillary crowding 2.22 6.36 2.56 6.9 2.47 6.38
Table III
ICC values between the manual and study group measurements
Variable Caliper measurements-CBCT measurements Caliper measurements-iTero model measurements
UM MD 0.91 0.98
UM V 0.97 0.99
UM BL 0.89 0.96
UP MD 0.96 0.99
UP V 0.99 0.99
UP BL 0.91 0.94
UC MD 0.92 0.96
UP V 0.98 0.99
UC BL 0.95 0.97
LI MD 0.88 0.92
LI V 0.93 0.97
LI BL 0.65 0.92
LP MD 0.97 0.99
LP V 0.97 0.99
LP BL 0.89 0.96
LM MD 0.9 0.98
LM V 0.87 0.92
LM BL 0.91 0.94
Mandibular crowding 0.99 0.99
Maxillary crowding 0.98 0.99

The results of the Bland-Altman analysis are summarized in Table IV . For the estimation of maxillary and mandibular crowding, the mean difference between the caliper and iTero model measurements ( Fig 4 ) was 0.16 mm, with a 95% confidence interval (CI) of 0.04 to 0.27 mm. Limits of agreement between the 2 methods were −0.9 to 1.3 mm. Figure 5 displays the agreement between the caliper and CBCT measurements for crowding. The mean difference was 0.26 mm, with a 95% CI of 0.06 to 0.46 mm. A slightly higher disagreement was found for this comparison, with the agreement limits ranging from −1.7 to 2.2 mm.

Table IV
Summary of the Bland-Altman analysis for the estimation of crowding in both dental arches (mm)
Parameter Caliper measurements–CBCT measurements Caliper measurements–iTero measurements
Bias (mean difference) 0.26 0.16
SD 1.03 0.58
CI bias (95%) 0.06-0.46 0.04-0.27
Limits of agreement −1.76-2.28 −0.99-1.31
Bland-Altman plot for the comparison of crowding between the caliper and the iTero model measurements.
Fig 4
Bland-Altman plot for the comparison of crowding between the caliper and the iTero model measurements.
Bland-Altman plot for the comparison of crowding between the caliper and the CBCT measurements.
Fig 5
Bland-Altman plot for the comparison of crowding between the caliper and the CBCT measurements.

Discussion

With nearly every other aspect of the health records becoming digital, it stands to reason that dental models must also cross this divide. Although the diagnostic information from plaster models that are converted to digital files is highly accurate, this approach is simply a bridge and might not be the long-term solution. Eliminating the physical impressions altogether is ultimately where technology is leading.

Attempts have been made to generate digital models directly from CBCT scans, and the accuracy of CBCT data in terms dental diagnostic measurements is highly reliable, as evidenced in our study. However, it is not practical to expose our patients to unnecessary radiation to obtain digital models. Since not all patients require a CBCT scan as part of their diagnostic process, taking one for the sole purpose of acquiring digital models is a clear violation of the “as low as reasonably achievable” principle and goes against the recommendations of the American Dental Association.

Intraoral scanners hold promise because they emit only optical radiation and have proven to be useful to restorative dentists wishing to eliminate physical impressions. Seelbach et al tested 3 leading intraoral scanners and showed that even things as intricate as fixed prosthodontic restorations can be fabricated equally as well with intraoral scanners as with a 2-step putty-and-wash technique. However, their study, like many others, used materials that failed to truly test the accuracy of the iTero scanner because of the physics behind confocal technology.

The iTero scanner is the only intraoral scanner that uses a parallel confocal laser scanner to capture 3D data with the video camera built into the wand to aid the clinician in positioning the laser intraorally. Confocal imaging is based on a beam of light that passes through a pinhole and then reflects off its target object. Only in-focus light can pass back through the filtering mechanism to the sensor. During a scan, the motor in the scanning wand adjusts the filter to different depths of field, and only the points at the exact predetermined depth are incorporated into the data, eliminating potential sources of bad information in the model. For a 14 ×18-mm field, the iTero can clearly obtain 100,000 points of laser light in 300 different intervals of 15 mm of depth. The reflected light that passes back through the filter gives the exact location of each point on the object within 0.05 mm, even without the use of powder.

The 18th-century physicist Augustin-Jean Fresnel described the way when light moves between materials (as it does as it enters and exits prismatic enamel), both reflection and refraction occur. Most studies conducted to date to evaluate the iTero scanner’s diagnostic accuracy used a plastic or chromium-cobalt dentoform. Metals such as chromium-cobalt present total reflection of light, a property not shared by enamel unless it is covered in scanning powder. Plastics reflect and refract light, but the refractive index of each plastic is different and might not be the same as tooth enamel. Any energy lost to refraction will not be reflected back to the sensor and could affect the scanner’s ability to accurately determine the depth of the object it is scanning. Therefore, any study seeking to evaluate the iTero scanner must be performed on natural dentition or at least on a surface with a similar refractive index as tooth structure. We used only natural dentitions to ensure that the prismatic properties of enamel and its effects on the accuracy of the scanner would not be overlooked.

We found after testing on natural tooth structures that the iTero scanner can make linear and crowding measurements that are clinically acceptable for orthodontic diagnosis and treatment planning, and to a finer degree than CBCT measurements. All linear and crowding measurements for both the digital models generated from CBCT and the iTero agreed within a clinically acceptable range for diagnosis and treatment planning purposes, with the exception of the buccolingual measurement of the mandibular incisor in the CBCT scans (ICC, 0.65). The lowest ICC value for the iTero scanner (0.92) was also obtained from the buccolingual measurement of the mandibular incisor. One possible explanation for this is that because of the small mesiodistal width and steep curvature on the buccal and lingual surfaces of the mandibular incisors, deviating from the center of the long axis even slightly on such an oblong tooth can greatly affect the recorded measurement.

Bland-Altman plots for the comparison of caliper measurements with both CBCT and iTero measurements graphically show the systematic errors. In both comparisons, only a few study samples provided crowding measurements outside the agreement limits. In general, the investigated methods consistently provided similar measures to caliper measurements. However, from a clinical perspective, measurements made on the iTero models were superior to the CBCT measurements as shown by the higher level of agreement with the manual caliper measurements. The limits of agreement were small enough for us to be confident that the iTero scanner can be used to perform a diagnostic task in place of the manual caliper measurements for clinical purposes.

More detailed anatomy was shown on the iTero scans compared with the CBCT images ( Figs 1-3 ). This could be of particular interest to clinicians seeking to more easily detect an anomalous morphology when planning treatment for a patient or using the scans in an indirect bonding setup. Lightheart et al expressed concern over this type of detail in surface anatomy because pits and fissures less than 0.5 mm in size might not be accurately portrayed in a CBCT model. Similar topographic or volumetric studies would be needed to discern such resolution from intraoral scanners. These studies would be best carried out in vivo to account for factors such as saliva and shadows from ambient light created intraorally. Further studies should also incorporate other chairside 3D scanners to determine whether one technology outperforms the others before intraoral scanners can be regarded as an acceptable alternative to traditional dental impressions for orthodontic purposes.

Conclusions

Our study shows that measurements obtained from CBCT scans and the iTero models produce interchangeable results with manual measurements, making both methods sufficient for use in orthodontic diagnosis and treatment planning. Ultimately, the iTero models had slightly higher correlations to the skull measurements and outperformed the CBCT measurements in some areas.

All authors completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and the following were reported: Clark D. Colville was paid by Align Technology to serve on a clinical advisory board and as a member of its speakers’ bureau, and Jeryl D. English received a 2012 Align Technologies Research Award; the money was paid to University of Texas Health Science Center, Houston.

References

  • 1. Fleming P.S., Marinho V., Johal A.: Orthodontic measurements on digital study models compared with plaster models: a systematic review.Orthod Craniofac Res 2011; 14: pp. 1-16.
  • 2. Akyalcin S.: Are digital models replacing plaster casts?.Dentistry 2011; 1: pp. e102.
  • 3. Quimby M.L., Vig K.W., Rashid R.G., Firestone A.R.: The accuracy and reliability of measurements made on computer-based digital models.Angle Orthod 2004; 74: pp. 298-303.
  • 4. Rheude B., Sadowsky P.L., Ferriera A., Jacobson A.: An evaluation of the use of digital study models in orthodontic diagnosis and treatment planning.Angle Orthod 2005; 75: pp. 300-304.
  • 5. Stevens D.R., Flores-Mir C., Nebbe B., Raboud D.W., Heo G., Major P.W.: Validity, reliability, and reproducibility of plaster vs digital study models: comparison of peer assessment rating and Bolton analysis and their constituent measurements.Am J Orthod Dentofacial Orthop 2006; 129: pp. 794-803.
  • 6. Tomassetti J.J., Taloumis L.J., Denny J.M., Fischer J.R.: A comparison of 3 computerized Bolton tooth-size analyses with a commonly used method.Angle Orthod 2001; 71: pp. 351-357.
  • 7. Goonewardene R.W., Goonewardene M.S., Razza J.M., Murray K.: Accuracy and validity of space analysis and irregularity index measurements using digital models.Aust Orthod J 2008; 24: pp. 83-90.
  • 8. Leifert M.F., Leifert M.M., Efstratiadis S.S., Cangialosi T.J.: Comparison of space analysis evaluations with digital models and plaster dental casts.Am J Orthod Dentofacial Orthop 2009; 136: pp. 16.e1-16.e4.
  • 9. Mullen S.R., Martin C.A., Ngan P., Gladwin M.: Accuracy of space analysis with eModels and plaster models.Am J Orthod Dentofacial Orthop 2007; 132: pp. 346-352.
  • 10. Torassian G., Kau C.H., English J.D., Powers J., Bussa H.I., Salas-Lopez A., et. al.: Digital models vs plaster models using alginate and alginate substitute materials.Angle Orthod 2010; 80: pp. 474-481.
  • 11. Mah J.: The evolution of digital study models.J Clin Orthod 2007; 9: pp. 557-561.
  • 12. Lightheart K.G., English J.D., Kau C.H., Akyalcin S., Bussa H.I., McGrory K.R., et. al.: Surface analysis of study models generated from OrthoCAD and cone-beam computed tomography imaging.Am J Orthod Dentofacial Orthop 2012; 141: pp. 686-693.
  • 13. Lagravère M.O., Carey J., Toogood R.W., Major P.W.: Three-dimensional accuracy of measurements made with software on cone-beam computed tomography images.Am J Orthod Dentofacial Orthop 2008; 134: pp. 112-116.
  • 14. Tarazona B., Llamas J.M., Cibrian R., Gandia J.L., Paredes V.: A comparison between dental measurements taken from CBCT models and those taken from a digital method.Eur J Orthod 2013; 35: pp. 1-6.
  • 15. Ganguly R., Ruprecht A., Vincent S., Hellstein J., Timmons S., Qian F.: Accuracy of linear measurement in the Galileos cone beam computed tomography under simulated clinical conditions.Dentomaxillofac Radiol 2011; 40: pp. 299-305.
  • 16. Torres M.G., Campos P.S., Segundo N.P., Navarro M., Crusoé-Rebello I.: Accuracy of linear measurements in cone beam computed tomography with different voxel sizes.Implant Dent 2012; 21: pp. 150-155.
  • 17. Leung C.C., Palomo L., Griffith R., Hans M.G.: Accuracy and reliability of cone-beam computed tomography for measuring alveolar bone height and detecting bony dehiscences and fenestrations.Am J Orthod Dentofacial Orthop 2010; 137: pp. S109-S119.
  • 18. Patcas R., Müller L., Ullrich O., Peltomäki T.: Accuracy of cone-beam computed tomography at different resolutions assessed on the bony covering of the mandibular anterior teeth.Am J Orthod Dentofacial Orthop 2012; 141: pp. 41-50.
  • 19. Bland J.M., Altman D.G.: Measuring agreement in method comparison studies.Stat Methods Med Res 1999; 8: pp. 135-160.
  • 20. Baumgaertel S., Palomo J.M., Palomo L., Hans M.G.: Reliability and accuracy of cone-beam computed tomography dental measurements.Am J Orthod Dentofacial Orthop 2009; 136: pp. 19-28.
  • 21. Nguyen E., Boychuk D., Orellana M.: Accuracy of cone-beam computed tomography in predicting the diameter of unerupted teeth.Am J Orthod Dentofacial Orthop 2011; 140: pp. e59-e66.
  • 22. Sherrard J.F., Rossouw P.E., Benson B.W., Carrillo R., Buschang P.H.: Accuracy and reliability of tooth and root lengths measured on cone-beam computed tomographs.Am J Orthod Dentofacial Orthop 2010; 137: pp. S100-S108.
  • 23. Damstra J., Fourie Z., Huddleston Slater J.J., Ren Y.: Accuracy of linear measurements from cone-beam computed tomography-derived surface models of different voxel sizes.Am J Orthod Dentofacial Orthop 2010; 137: pp. 16.e1-16.e6.
  • 24. American Dental Association. ADA statement: diagnostic radiation procedures must be used sparingly to reduce dental radiation risk. Available at: http://www.ada.org/5081.aspx . Accessed February 6, 2013.
  • 25. Logozzo S., Franceschini G., Kilpela A., Caponi M., Governi L., Blois L.: A comparative analysis of intraoral 3d digital scanners for restorative dentistry.Internet J Med Technol 2011; 5:
  • 26. Seelbach P., Brueckel C., Wöstmann B.: Accuracy of digital and conventional impression techniques and workflow.Clin Oral Investig 2013; 17: pp. 1759-1764.
  • 27. Garino F., Garino B.: The iOC intraoral scanner and Invisalign: a new paradigm.J Clin Orthod 2012; 46: pp. 115-121.
  • 28. Crew H.: The Wave Theory of Light – Memoirs by Huygens, Young and Fresnel.1900.American Book CompanyNew York
  • 29. Ender A., Mehl A.: Full arch scans: conventional versus digital impressions—an in-vitro study.Int J Comput Dent 2011; 14: pp. 11-21.
  • 30. Phark J., Oliveira A.L., Sartori N., Giro E.M.A., Kahn R., Duarte S.: Marginal fit of all-ceramic crowns using conventional and digital impressions [abstract].J Dent Res 2012; 91: pp. 991.
  • 31. Kim M., Song H., Heo S., Han J., Koak J., Lim Y., Kim S., Kim S., Yeo I., Kim S.: Accuracy of an intraoral digital impression using parallel confocal imaging [abstract].J Dent Res 2012; 91: pp. 995.

Was this article helpful?

Diagnostic accuracy of impression-free digital models Sercan Akyalcin , Benjamin E. Cozad , Jeryl D. English , Clark D. Colville and Stephen Laman American Journal of Orthodontics and Dentofacial Orthopedics, 2013-12-01, Volume 144, Issue 6, Pages 916-922, Copyright © 2013 American Association of Orthodontists Introduction Impression-free techniques might eliminate the potential shortcomings of digital dental models. Chairside scanners offer the advantage of obtaining digital dental models directly from the patient without the need for dental impressions. The aim of this study was to evaluate the accuracy of 3-dimensional digital models acquired from a chairside intraoral scanner compared with both manual and cone-beam computed tomography measurements of the same dental anatomy. Methods The study sample comprised 60 dry skulls. Each skull had the maxillary and mandibular arches scanned with a Cadent iTero scanner (Align Technology, San Jose, Calif) and had a cone-beam computed tomography scan taken with a CS 9300 unit (Carestream Health, Atlanta, Ga). Linear measurements in all 3 dimensions of the space in each dental arch together with tooth-size arch-length analysis for both the maxillary and mandibular arches were carried out manually on the dry skulls with calipers and digitally on the scanned 3-dimensional models and cone-beam computed tomography images. Intraclass correlation (ICC) analysis was performed for all variables tested in the study groups, with the manual measurements on the dry skulls as the gold standard. The Bland-Altman analysis was also applied to the data to graphically display the agreement of the diagnostic measurements obtained from these methods. Results Measurements from the iTero models demonstrated near-perfect agreement (ICC, 0.91-0.99) with the caliper measurements. Cone-beam computed tomography measurements had moderate to high levels of agreement (ICC, 0.65-0.99) compared with the caliper measurements. Conclusions Direct digital acquisition of the dental arches with a chairside scanner provided almost 1-to-1 diagnostic information of the investigated anatomy and was superior to the cone-beam computed tomography measurements. Through advances in biomaterials, robust and dimensionally accurate dental plaster models are possible and have long been the gold standard in orthodontics. In the new digital era, various technological advancements have made their way into dental practices. The past decade has seen the advent of digital models with acceptable quality, allowing the orthodontic record to become completely digitized. For orthodontics, the most important expectation from a digital model system lies in its diagnostic accuracy and reliability. Although the consensus is that measurements with digital models compare well with those derived from plaster models, several studies that investigated complex measurements such as space available, irregularity index, and Bolton analysis indicated that mean differences between the plaster and digital models can exceed 1.5 mm. This much difference might not be clinically acceptable. However, there is also contrary evidence in the literature that supports the validity of digital models for the aforementioned measurements. Many professionals obtain their digital models through the use of proprietary services. Traditional impressions and plaster models must be submitted to the selected company so that they can be scanned and the digital models can be generated and made available for downloading. Submission of models or impressions by mail can result in breakage of the models or distortion of the impressions during shipment. Even if no breakage or distortion occurs during shipment, the impression materials have an inherent amount of distortion at their time of use. Moreover, there are technique-specific differences in the production phase of digital models such as surface scanning algorithms. All of these might account for the statistical variability related to the accuracy of digital dental models. Perhaps elimination of the conventional impression step and generation of dental models directly from the 3-dimensional (3D) dental anatomy will yield more accuracy. One way of doing this is to use the DICOM files produced by cone-beam computed tomography (CBCT) scanners. An advantage of this technology is that it allows for the visualization of supragingival structures as well as impacted teeth, bone levels, and joints. Studies using various scanners have been published showing that diagnostically accurate measurements can be made from CBCT scans. Perhaps the greatest drawback of using these scanners to acquire digital models is the required radiation exposure to the patient and the equipment expense to the practitioner. Some authors have also expressed concern over the absence of gingival tissues in the resultant files and the inability to accurately depict buccal alveolar bone. For those wishing to gain an accurate depiction of gingival tissues and avoid exposing the patient to radiation, an intraoral scanner might be more appropriate. Intraoral scanners, now widely available from several manufacturers, are self-contained in rolling units, making them portable around the office. In this study, we evaluated the diagnostic accuracy of an intraoral scanner and CBCT as sources of digital models from natural dentitions for orthodontic diagnosis compared with direct caliper measurements of the investigated dental anatomy. Material and methods The study sample comprised 60 dry skulls with intact dentition chosen from a unique collection at the University of Texas Health Science Center in Houston. The Carestream CS 9300 (Carestream Health, Atlanta, Ga) was used to capture the CBCT images ( Fig 1 ) of the skulls with exposure parameters of 90 kV, 4 mA, and 300-μm voxel size and volume dimensions of 17 × 13.5 cm. The image reconstruction time was approximately 5 minutes. Each arch was then scanned using the reference model setting on the Cadent iTero scanner running software version 4.0.5.31 (Align Technology, San Jose, Calif). The raw images were available for chairside viewing within 2 minutes after completion of the scanning. These images were then sent via the Internet to Align Technology, where they were reprocessed and made available for downloading within 48 hours as a stereolithography file ( Fig 2 ). This file format is an industry standard native to stereolithography computer-aided design and can be easily viewed and manipulated in various software applications without loss of quality. Fig 1 Three-dimensional reconstructions of the dental arches with CBCT. Fig 2 Dental models generated by the iTero scanner. With digital calipers, a tooth-size arch-length analysis was performed in each arch. In addition to crowding measurements, 3 teeth in each arch were measured in the anteroposterior, transverse, and vertical dimensions ( Fig 3 ). For the transverse measurement, each tooth was measured at the center of its long axis at its most convex points on the buccal and lingual surfaces ( Fig 3 , A ). The anteroposterior measurement was defined as the most mesiobuccal contact point to the most distobuccal contact point ( Fig 3 , B ). In the vertical plane, each tooth was measured on the buccal surface at the middle of its long axis from the height of the crestal bone to the cusp tip ( Fig 3 , C ). In the maxillary arch, measurements were made on a molar, a premolar, and a canine. In the mandibular arch, measurements were made on a molar, a premolar, and an incisor ( Table I ). Fig 3 Linear tooth measurements used in this study: A, transverse; B, anteroposterior; C, vertical planes. Table I Linear measurements used in the study Label Description UM MD Mesial distal width of upper permanent molar UM V Vertical height of upper permanent molar UM BL Buccolingual width of upper permanent molar UP MD Mesiodistal width of upper permanent premolar UP V Vertical height of upper permanent premolar UP BL Buccolingual width of upper permanent premolar UC MD Mesiodistal width of upper canine UC V Vertical height of upper canine UC BL Buccolingual width of upper canine LI MD Mesiodistal width of lower permanent incisor LI V Vertical height of lower permanent incisor LI BL Buccolingual width of lower permanent incisor LP MD Mesiodistal width of lower permanent premolar LP V Vertical height of lower permanent premolar LP BL Buccolingual width of lower permanent premolar LM MD Mesiodistal width of lower permanent molar LM V Vertical height of lower permanent molar LM BL Buccolingual width of lower permanent molar The CBCT files were opened using the InVivo Dental software (version 5.1.9; Anatomage, San Jose, Calif), and measurements were made in millimeters on sectioned views of a gray-scale 3D reconstruction using the software's built-in tool for measuring the distance between 2 defined points. Scanner stereolithography files were opened in Studio Basic (version 0.1; nettfabb, Lupburg, Germany). The measurements were made on unsectioned solid shaded models using the application's built-in ruler tool. Statistical analysis Intraclass correlation analysis (ICC) was performed for all variables tested in the study groups, with the manual measurements from the dry skulls serving as the gold standard. ICC values were computed using the 2-way random effect model (absolute agreement definition). SPSS statistical software (version 21; IBM, Armonk, NY) was used to analyze the data. Additionally, the Bland-Altman analysis was performed with XLSTAT Mac (version 2012; Addinsoft, New York, NY) to visually demonstrate the agreement of the maxillary and mandibular crowding values between the manual caliper measurements and the 2 methods tested. The within-observer repeatability was evaluated with ICC values by repeating all measurements from 10 randomly selected skulls 1 month later. Results Descriptive statistics for the measurements are presented in Table II . Intraobserver repeatability was excellent with ICC values between 0.97 and 0.99. Table III displays the ICC values for linear and crowding measurements for both CBCT and the iTero models compared with the manual caliper measurements. The ICC values for the linear measurements in the comparison of caliper measurements with CBCT measurements ranged between 0.88 and 0.99, displaying near-perfect agreement, except for the buccolingual measurement of the mandibular incisor (ICC, 0.65). The ICC values of the comparison of caliper measurements with the measurements made on the iTero models ranged between 0.92 and 0.99, showing a much higher level of agreement. However, both methods had ICC values greater than 0.98 for maxillary and mandibular crowding compared with manual measurements, indicating strong evidence of their diagnostic capability. Table II Descriptive statistics for the measurements used in the study Variable (mm) Dry skulls CBCT iTero Mean SD Mean SD Mean SD UM MD 9.86 0.64 9.84 0.69 9.83 0.67 UM V 7.86 1.49 7.73 1.53 7.81 1.47 UM BL 10.67 0.66 10.62 0.74 10.69 0.63 UP MD 7.02 0.89 7.07 0.95 6.94 0.91 UP V 9.25 1.85 9.22 1.9 9.23 1.86 UP BL 9.17 0.56 9.23 0.56 9.19 0.62 UC MD 7.26 0.69 7.34 0.71 7.2 0.63 UP V 12.21 2.21 12.25 2.29 12.1 2.21 UC BL 7.68 0.89 7.66 0.89 7.61 0.9 LI MD 5.76 1.42 5.6 1.48 5.51 0.41 LI V 11.46 1.82 11.56 2.03 11.32 1.63 LI BL 5.86 0.41 5.98 0.56 5.86 0.44 LP MD 7.21 1.28 7.28 1.27 7.16 1.29 LP V 8.56 1.6 8.49 1.54 8.55 1.57 LP BL 7.96 0.71 8.04 0.64 7.98 0.68 LM MD 10.66 0.7 10.75 0.66 10.61 0.72 LM V 7.6 0.98 7.29 1.04 7.56 1.01 LM BL 10.03 0.62 9.98 0.59 10.01 0.61 Mandibular crowding 4.18 6.28 4.36 6.25 4.25 6.28 Maxillary crowding 2.22 6.36 2.56 6.9 2.47 6.38 Table III ICC values between the manual and study group measurements Variable Caliper measurements-CBCT measurements Caliper measurements-iTero model measurements UM MD 0.91 0.98 UM V 0.97 0.99 UM BL 0.89 0.96 UP MD 0.96 0.99 UP V 0.99 0.99 UP BL 0.91 0.94 UC MD 0.92 0.96 UP V 0.98 0.99 UC BL 0.95 0.97 LI MD 0.88 0.92 LI V 0.93 0.97 LI BL 0.65 0.92 LP MD 0.97 0.99 LP V 0.97 0.99 LP BL 0.89 0.96 LM MD 0.9 0.98 LM V 0.87 0.92 LM BL 0.91 0.94 Mandibular crowding 0.99 0.99 Maxillary crowding 0.98 0.99 The results of the Bland-Altman analysis are summarized in Table IV . For the estimation of maxillary and mandibular crowding, the mean difference between the caliper and iTero model measurements ( Fig 4 ) was 0.16 mm, with a 95% confidence interval (CI) of 0.04 to 0.27 mm. Limits of agreement between the 2 methods were −0.9 to 1.3 mm. Figure 5 displays the agreement between the caliper and CBCT measurements for crowding. The mean difference was 0.26 mm, with a 95% CI of 0.06 to 0.46 mm. A slightly higher disagreement was found for this comparison, with the agreement limits ranging from −1.7 to 2.2 mm. Table IV Summary of the Bland-Altman analysis for the estimation of crowding in both dental arches (mm) Parameter Caliper measurements–CBCT measurements Caliper measurements–iTero measurements Bias (mean difference) 0.26 0.16 SD 1.03 0.58 CI bias (95%) 0.06-0.46 0.04-0.27 Limits of agreement −1.76-2.28 −0.99-1.31 Fig 4 Bland-Altman plot for the comparison of crowding between the caliper and the iTero model measurements. Fig 5 Bland-Altman plot for the comparison of crowding between the caliper and the CBCT measurements. Discussion With nearly every other aspect of the health records becoming digital, it stands to reason that dental models must also cross this divide. Although the diagnostic information from plaster models that are converted to digital files is highly accurate, this approach is simply a bridge and might not be the long-term solution. Eliminating the physical impressions altogether is ultimately where technology is leading. Attempts have been made to generate digital models directly from CBCT scans, and the accuracy of CBCT data in terms dental diagnostic measurements is highly reliable, as evidenced in our study. However, it is not practical to expose our patients to unnecessary radiation to obtain digital models. Since not all patients require a CBCT scan as part of their diagnostic process, taking one for the sole purpose of acquiring digital models is a clear violation of the “as low as reasonably achievable” principle and goes against the recommendations of the American Dental Association. Intraoral scanners hold promise because they emit only optical radiation and have proven to be useful to restorative dentists wishing to eliminate physical impressions. Seelbach et al tested 3 leading intraoral scanners and showed that even things as intricate as fixed prosthodontic restorations can be fabricated equally as well with intraoral scanners as with a 2-step putty-and-wash technique. However, their study, like many others, used materials that failed to truly test the accuracy of the iTero scanner because of the physics behind confocal technology. The iTero scanner is the only intraoral scanner that uses a parallel confocal laser scanner to capture 3D data with the video camera built into the wand to aid the clinician in positioning the laser intraorally. Confocal imaging is based on a beam of light that passes through a pinhole and then reflects off its target object. Only in-focus light can pass back through the filtering mechanism to the sensor. During a scan, the motor in the scanning wand adjusts the filter to different depths of field, and only the points at the exact predetermined depth are incorporated into the data, eliminating potential sources of bad information in the model. For a 14 ×18-mm field, the iTero can clearly obtain 100,000 points of laser light in 300 different intervals of 15 mm of depth. The reflected light that passes back through the filter gives the exact location of each point on the object within 0.05 mm, even without the use of powder. The 18th-century physicist Augustin-Jean Fresnel described the way when light moves between materials (as it does as it enters and exits prismatic enamel), both reflection and refraction occur. Most studies conducted to date to evaluate the iTero scanner's diagnostic accuracy used a plastic or chromium-cobalt dentoform. Metals such as chromium-cobalt present total reflection of light, a property not shared by enamel unless it is covered in scanning powder. Plastics reflect and refract light, but the refractive index of each plastic is different and might not be the same as tooth enamel. Any energy lost to refraction will not be reflected back to the sensor and could affect the scanner's ability to accurately determine the depth of the object it is scanning. Therefore, any study seeking to evaluate the iTero scanner must be performed on natural dentition or at least on a surface with a similar refractive index as tooth structure. We used only natural dentitions to ensure that the prismatic properties of enamel and its effects on the accuracy of the scanner would not be overlooked. We found after testing on natural tooth structures that the iTero scanner can make linear and crowding measurements that are clinically acceptable for orthodontic diagnosis and treatment planning, and to a finer degree than CBCT measurements. All linear and crowding measurements for both the digital models generated from CBCT and the iTero agreed within a clinically acceptable range for diagnosis and treatment planning purposes, with the exception of the buccolingual measurement of the mandibular incisor in the CBCT scans (ICC, 0.65). The lowest ICC value for the iTero scanner (0.92) was also obtained from the buccolingual measurement of the mandibular incisor. One possible explanation for this is that because of the small mesiodistal width and steep curvature on the buccal and lingual surfaces of the mandibular incisors, deviating from the center of the long axis even slightly on such an oblong tooth can greatly affect the recorded measurement. Bland-Altman plots for the comparison of caliper measurements with both CBCT and iTero measurements graphically show the systematic errors. In both comparisons, only a few study samples provided crowding measurements outside the agreement limits. In general, the investigated methods consistently provided similar measures to caliper measurements. However, from a clinical perspective, measurements made on the iTero models were superior to the CBCT measurements as shown by the higher level of agreement with the manual caliper measurements. The limits of agreement were small enough for us to be confident that the iTero scanner can be used to perform a diagnostic task in place of the manual caliper measurements for clinical purposes. More detailed anatomy was shown on the iTero scans compared with the CBCT images ( Figs 1-3 ). This could be of particular interest to clinicians seeking to more easily detect an anomalous morphology when planning treatment for a patient or using the scans in an indirect bonding setup. Lightheart et al expressed concern over this type of detail in surface anatomy because pits and fissures less than 0.5 mm in size might not be accurately portrayed in a CBCT model. Similar topographic or volumetric studies would be needed to discern such resolution from intraoral scanners. These studies would be best carried out in vivo to account for factors such as saliva and shadows from ambient light created intraorally. Further studies should also incorporate other chairside 3D scanners to determine whether one technology outperforms the others before intraoral scanners can be regarded as an acceptable alternative to traditional dental impressions for orthodontic purposes. Conclusions Our study shows that measurements obtained from CBCT scans and the iTero models produce interchangeable results with manual measurements, making both methods sufficient for use in orthodontic diagnosis and treatment planning. Ultimately, the iTero models had slightly higher correlations to the skull measurements and outperformed the CBCT measurements in some areas. All authors completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and the following were reported: Clark D. Colville was paid by Align Technology to serve on a clinical advisory board and as a member of its speakers' bureau, and Jeryl D. English received a 2012 Align Technologies Research Award; the money was paid to University of Texas Health Science Center, Houston. References 1. Fleming P.S., Marinho V., Johal A.: Orthodontic measurements on digital study models compared with plaster models: a systematic review.Orthod Craniofac Res 2011; 14: pp. 1-16. 2. Akyalcin S.: Are digital models replacing plaster casts?.Dentistry 2011; 1: pp. e102. 3. Quimby M.L., Vig K.W., Rashid R.G., Firestone A.R.: The accuracy and reliability of measurements made on computer-based digital models.Angle Orthod 2004; 74: pp. 298-303. 4. Rheude B., Sadowsky P.L., Ferriera A., Jacobson A.: An evaluation of the use of digital study models in orthodontic diagnosis and treatment planning.Angle Orthod 2005; 75: pp. 300-304. 5. Stevens D.R., Flores-Mir C., Nebbe B., Raboud D.W., Heo G., Major P.W.: Validity, reliability, and reproducibility of plaster vs digital study models: comparison of peer assessment rating and Bolton analysis and their constituent measurements.Am J Orthod Dentofacial Orthop 2006; 129: pp. 794-803. 6. Tomassetti J.J., Taloumis L.J., Denny J.M., Fischer J.R.: A comparison of 3 computerized Bolton tooth-size analyses with a commonly used method.Angle Orthod 2001; 71: pp. 351-357. 7. Goonewardene R.W., Goonewardene M.S., Razza J.M., Murray K.: Accuracy and validity of space analysis and irregularity index measurements using digital models.Aust Orthod J 2008; 24: pp. 83-90. 8. Leifert M.F., Leifert M.M., Efstratiadis S.S., Cangialosi T.J.: Comparison of space analysis evaluations with digital models and plaster dental casts.Am J Orthod Dentofacial Orthop 2009; 136: pp. 16.e1-16.e4. 9. Mullen S.R., Martin C.A., Ngan P., Gladwin M.: Accuracy of space analysis with eModels and plaster models.Am J Orthod Dentofacial Orthop 2007; 132: pp. 346-352. 10. Torassian G., Kau C.H., English J.D., Powers J., Bussa H.I., Salas-Lopez A., et. al.: Digital models vs plaster models using alginate and alginate substitute materials.Angle Orthod 2010; 80: pp. 474-481. 11. Mah J.: The evolution of digital study models.J Clin Orthod 2007; 9: pp. 557-561. 12. Lightheart K.G., English J.D., Kau C.H., Akyalcin S., Bussa H.I., McGrory K.R., et. al.: Surface analysis of study models generated from OrthoCAD and cone-beam computed tomography imaging.Am J Orthod Dentofacial Orthop 2012; 141: pp. 686-693. 13. Lagravère M.O., Carey J., Toogood R.W., Major P.W.: Three-dimensional accuracy of measurements made with software on cone-beam computed tomography images.Am J Orthod Dentofacial Orthop 2008; 134: pp. 112-116. 14. Tarazona B., Llamas J.M., Cibrian R., Gandia J.L., Paredes V.: A comparison between dental measurements taken from CBCT models and those taken from a digital method.Eur J Orthod 2013; 35: pp. 1-6. 15. Ganguly R., Ruprecht A., Vincent S., Hellstein J., Timmons S., Qian F.: Accuracy of linear measurement in the Galileos cone beam computed tomography under simulated clinical conditions.Dentomaxillofac Radiol 2011; 40: pp. 299-305. 16. Torres M.G., Campos P.S., Segundo N.P., Navarro M., Crusoé-Rebello I.: Accuracy of linear measurements in cone beam computed tomography with different voxel sizes.Implant Dent 2012; 21: pp. 150-155. 17. Leung C.C., Palomo L., Griffith R., Hans M.G.: Accuracy and reliability of cone-beam computed tomography for measuring alveolar bone height and detecting bony dehiscences and fenestrations.Am J Orthod Dentofacial Orthop 2010; 137: pp. S109-S119. 18. Patcas R., Müller L., Ullrich O., Peltomäki T.: Accuracy of cone-beam computed tomography at different resolutions assessed on the bony covering of the mandibular anterior teeth.Am J Orthod Dentofacial Orthop 2012; 141: pp. 41-50. 19. Bland J.M., Altman D.G.: Measuring agreement in method comparison studies.Stat Methods Med Res 1999; 8: pp. 135-160. 20. Baumgaertel S., Palomo J.M., Palomo L., Hans M.G.: Reliability and accuracy of cone-beam computed tomography dental measurements.Am J Orthod Dentofacial Orthop 2009; 136: pp. 19-28. 21. Nguyen E., Boychuk D., Orellana M.: Accuracy of cone-beam computed tomography in predicting the diameter of unerupted teeth.Am J Orthod Dentofacial Orthop 2011; 140: pp. e59-e66. 22. Sherrard J.F., Rossouw P.E., Benson B.W., Carrillo R., Buschang P.H.: Accuracy and reliability of tooth and root lengths measured on cone-beam computed tomographs.Am J Orthod Dentofacial Orthop 2010; 137: pp. S100-S108. 23. Damstra J., Fourie Z., Huddleston Slater J.J., Ren Y.: Accuracy of linear measurements from cone-beam computed tomography-derived surface models of different voxel sizes.Am J Orthod Dentofacial Orthop 2010; 137: pp. 16.e1-16.e6. 24. American Dental Association. ADA statement: diagnostic radiation procedures must be used sparingly to reduce dental radiation risk. Available at: http://www.ada.org/5081.aspx . Accessed February 6, 2013. 25. Logozzo S., Franceschini G., Kilpela A., Caponi M., Governi L., Blois L.: A comparative analysis of intraoral 3d digital scanners for restorative dentistry.Internet J Med Technol 2011; 5: 26. Seelbach P., Brueckel C., Wöstmann B.: Accuracy of digital and conventional impression techniques and workflow.Clin Oral Investig 2013; 17: pp. 1759-1764. 27. Garino F., Garino B.: The iOC intraoral scanner and Invisalign: a new paradigm.J Clin Orthod 2012; 46: pp. 115-121. 28. Crew H.: The Wave Theory of Light – Memoirs by Huygens, Young and Fresnel.1900.American Book CompanyNew York 29. Ender A., Mehl A.: Full arch scans: conventional versus digital impressions—an in-vitro study.Int J Comput Dent 2011; 14: pp. 11-21. 30. Phark J., Oliveira A.L., Sartori N., Giro E.M.A., Kahn R., Duarte S.: Marginal fit of all-ceramic crowns using conventional and digital impressions [abstract].J Dent Res 2012; 91: pp. 991. 31. Kim M., Song H., Heo S., Han J., Koak J., Lim Y., Kim S., Kim S., Yeo I., Kim S.: Accuracy of an intraoral digital impression using parallel confocal imaging [abstract].J Dent Res 2012; 91: pp. 995.

Related Articles

Leave A Comment?

You must be logged in to post a comment.