Successful orthodontic treatment is based on extensive

Transcription

1 ORIGINAL ARTICLE Three-dimensional dental measurements: An alternative to plaster models Hend Mohammed El-Zanaty, a Amr Ragab El-Beialy, b Amr Mohammed Abou El-Ezz, c Khaled Hazem Attia, d Ahmed Ragab El-Bialy, e and Yehya Ahmed Mostafa f Cairo, Egypt Introduction: The aim of this study was to compare the accuracy of dental measurements taken with calipers on plaster dental casts and those from computed tomography scans of the dentition with a dental measurement program. Methods: The sample consisted of plaster dental models of 34 orthodontic subjects. Dental arch measurements, including mesiodistal widths of teeth, arch widths, arch lengths, arch perimeters, and palatal depths were made with the calipers. The patients were also scanned with computed tomography, and measurements were made digitally with a 3-dimensional-based dental measurements program (3DD, Biodent, Cairo, Egypt). Results: The results showed strong agreement in most measurements between the conventional method and the 3DD in the 3 planes of space. The mesiodistal measurements of the maxillary right and left second premolars, left central incisor, and right first molar, and the mandibular left and right central incisors, right canine, and left first premolar had fair agreement. Conclusions: Excellent agreement between the measurements with the conventional and 3DD methods in the 3 planes of space was found; 3DD can be an alternative to conventional stone dental models. (Am J Orthod Dentofacial Orthop 2010;137:259-65) Successful orthodontic treatment is based on extensive diagnosis and treatment planning. Dental models, photographs, radiographs, and clinical examinations provide essential information for diagnosis. 1 Study casts are a standard component of orthodontic records and are fundamental to diagnosis, case presentation, treatment planning, evaluation of treatment progress, and record keeping. Tooth size, crowding or spacing, overjet, overbite, and Bolton analysis are typically measured manually on models. 1-3 photographs and radiographs are now routinely incorporated into electronic files, but digital dental models are not as widely used. Electronic storage of patient information, including study models, eliminates problems of physical storage, retrieval, From Cairo University, Cairo, Egypt. a Resident, Orthodontic Department, Faculty of Oral and Dental Medicine. b Assistant lecturer, Orthodontic Department, Faculty of Oral and Dental Medicine. c Professor, Orthodontic Department, Faculty of Oral and Dental Medicine. d Assistant professor, Orthodontic Department, Faculty of Oral and Dental Medicine. e Assisstant professor, Systems and Biomedical Engineering Department, Faculty of Engineering. f Professor and head, Orthodontic Department, Faculty of Oral and Dental Medicine. The authors report no commercial, proprietary, or financial interest in the products or companies described in this article. Reprint requests to: Yehya Ahmed Mostafa, Orthodontic Department, Faculty of Oral and Dental Medicine, Cairo University, 52 Arab League St, Mohandesseen, Giza, Egypt; , yehya.mostafa@aaomembers.org. Submitted, July 2007; revised and accepted, April /$36.00 Copyright Ó 2010 by the American Association of Orthodontists. doi: /j.ajodo maintenance, and office management. Documentation of treatment progress and communication between professional colleagues can be enhanced by digital records. 3-7 Attempts have been made to transform dental plaster models into 3-dimensional (3D) virtual models. 4,8-12 OrthoCAD (Cadent, Carlstadt, NJ) is a commercially available system that transforms casts or impressions into 3D digital models. 13 Tomassetti et al 14 evaluated various measuring methods, including OrthoCAD, to analyze the Bolton tooth size discrepancy. OrthoCAD values were found to be less correlated to the baseline values established by the average of 3 repeated measurements with a Boley gauge. The SureSmile (OraMatrix, Dallas, Tex) process begins with a direct clinical 3D scan of the patient s dentition with the OraScanner (OraMatrix), a light-based imaging device that projects a precise patterned grid onto the teeth. The orthodontist can then diagnose and plan treatment on the computer screen, using software tools to measure tooth and arch dimensions, and create symmetric and asymmetric archforms. 15 Several studies were conducted to compare 3D virtual models with plaster models. Stevens et al 16 compared the analysis of the gold-standard plaster model with the digital counterpart of the e-model (GeoDigm) for tooth sizes and occlusal relationships, specifically the Bolton analysis and the peer assessment rating index. Strong agreement between the measurements indicated that the digital models are a good replacement for conventional plaster models. 259

2 260 El-Zanaty et al American Journal of Orthodontics and Dentofacial Orthopedics February 2010 Fig 1. 3D virtual models showing some of the measurements taken. Whetten et al 17 investigated the difference between plaster models and the virtual 3D model (e-model) in treatment planning for Class II patients. They concluded that digital orthodontic study models (e-models) were a valid alternative to traditional plaster models in treatment planning for these patients. The aims of this study were to test the accuracy of a 3D-based dental (3DD) measurement program (Biodent, Systems and Biomedical Engineering Department, Faculty of Engineering, Cairo University, Cairo, Egypt) that uses computed tomography (CT) scans and to compare it with manual measurements made on plaster models. MATERIAL AND METHODS Our sample consisted of 34 adults, equally divided between men and women. Their ages ranged from 20 to 25 years. Approval for this study was obtained from the hospital Board of Research Ethics. All subjects had erupted permanent dentitions from the first molar of one side to the first molar of the other side, with no orthodontic appliances. Measurements were recorded on both the conventional model and the 3D virtual model from the CT scan with a blind protocol. For the conventional plaster models, the maxillary and mandibular casts of the 34 subjects were measured with Vernier calipers (OISt Orthodontics, Aston, Pa) calibrated to the nearest 0.1 mm. For the 3D virtual models, each subject s head was scanned with a DCT device (Light Speed Pro, General Electric medical CT scan machine, Wakesha, Wis) at an axial section of 1.25 mm. Effective milli-amperage was based on the preliminary scanogram, 80 mas, spiral scanning of 120 Kv, high resolution mode. The reconstruction interval was 0.6 mm with table feed of 17.5 mm per rotation. The total rotation time was 15 to 20 seconds. During the CT scanning, the subject was given a prefabricated splint (1 mm) to keep the maxillary and mandibular teeth separated; this was done to reproduce the occlusal anatomy and prevent blurring of the dental images. The inclination of the gantry was set parallel to the occlusal plane of the maxilla. All data were saved with a DICOM extension. For the 3D reconstruction, custom-made 3D computer software (Biodent) was used to reconstruct the 3D volume from the axial CT scans and separate the maxillary and mandibular dentitions. This software is based on the concept of window center and window width (brightness and contrast) produced by CT scanners. Measurements of the 3D volume were made with the measuring analysis tool of the 3DD program. For accuracy and ease of measurements, the images were enlarged on the screen by using a built-in magnifying tool. From the occlusal view, the following dental arch measurements were taken for both the conventional casts and the 3D virtual models by the same observer (H.M.Z.) (Fig 1). 1. Mesiodistal tooth width from the first permanent molar of 1 side to the first permanent molar of the other side.

3 American Journal of Orthodontics and Dentofacial Orthopedics El-Zanaty et al 261 Volume 137, Number 2 Table I. Analysis of method errors of maxillary and mandibular arch widths Maxillary arch width (mm) Interlateral incisor * 0.985* Intercanine * 0.987* Interfirst molar * 0.995* Mandibular arch width Interlateral incisor * 0.973* Intercanine * 0.980* Interfirst molar * 0.988* Table II. Analysis of method errors of maxillary and mandibular arch length Maxillary arch length (mm) Anterior * 0.973* Posterior * 0.981* Mandibular arch length (mm) Anterior * 0.979* Posterior * 0.946* 2. Arch width, maxillary interlateral incisor width, mandibular interlateral incisor width, maxillary intercanine width, mandibular intercanine width, maxillary interfirst molar width, and mandibular interfirst molar width. 3. Maxillary anterior arch length, mandibular anterior arch length, maxillary posterior arch length, and mandibular posterior arch length. 4. Maxillary and mandibular arch perimeters. 5. Maxillary palatal depth (from the distal view). Replication of the measurements (for reliability testing) was done by the same observer (H.M.Z.). These intraobserver measurements were made a week after the first set of measurements for both the conventional and the 3D model. Statistical analysis The data were collected, tabulated, and analyzed. A reproducibility index, called the concordance correlation coefficient (), introduced by Lin, 18 was used. It evaluates the agreement between 2 readings (from the same sample) by measuring the variation for the 45 line through the origin (the concordance line). 19,20 The measurements are accurate and precise. 3-D Method Method Fig 2. representation of the maxillary arch perimeter. The and the Pearson correlation coefficient () were used to evaluate the equivalence and association, respectively, between the conventional and digital measurements. Reproducibility for intraexaminer measurements was assessed with the and statistical test. RESULTS The results and comparisons between the conventional and digital models are shown in Tables I through V. The results were compared for conformity and equivalency by using the and tests, respectively. Excellent agreement was obtained between both measuring modalities for arch width. For maxillary arch width, the ranged from to 0.992, and the was to Mandibular arch widths were to for the and to for the (Table I). For arch length, there was excellent agreement between the measurements with the 2 measuring modalities, with values of and for maxillary anterior and posterior arch lengths, respectively. The findings for the same parameter were and 0.981, respectively. Comparable findings were observed with the and tests: and for mandibular anterior and posterior arch lengths, respectively (Table II). For the maxillary arch perimeter, the and values were and 0.996, respectively (Fig 2), and the value was for the mandibular arch perimeter with both statistical tests (Table III). Measurements of the mesiodistal widths of most maxillary and mandibular teeth showed strong

4 262 El-Zanaty et al American Journal of Orthodontics and Dentofacial Orthopedics February 2010 Table III. Analysis of method errors of maxillary and mandibular arch perimeters Arch perimeter Maxillary * 0.996* Mandibular * 0.979* 3-D Method Table IV. Analysis of method errors of mesiodistal tooth widths Maxillary left mesiodistal tooth width (mm) Central incisor * 0.828* Lateral incisor * 0.812* Canine * 0.822* First premolar * 0.806* Second premolar First molar * 0.882* Maxillary right mesiodistal tooth width (mm) Central incisor * 0.878* Lateral incisor * 0.898* Canine * 0.846* First premolar * 0.773* Second premolar First permanent molar Mandibular left mesiodistal tooth width (mm) Central incisor Lateral incisor * 0.854* Canine * 0.786* First premolar Second premolar * 0.836* First permanent molar * 0.838* Mandibular right mesiodistal tooth width (mm) Central incisor Lateral incisor * 0.827* Canine First premolar * 0.894* Second premolar * 0.885* First permanent molar * 0.85* *P.0.75, excellent agreement; 0.4.P\0.75, fair agreement Method Fig 3. representation of the mandibular right canine. Table V. Analysis of method errors of maxillary palatal depth Palatal depth (mm) Maxillary * 0.933* agreement between the conventional and 3DD modalities (Table IV). The values between the 2 modalities were to for the test and to for the test. The mesiodistal measurements of the following teeth showed fair agreement between the 2 measuring modalities: maxillary left second premolar (, 0.704;, 0.712), maxillary right second premolar (, 0.674;, 0.699), maxillary right first molar (, 0.718;, 0.746), mandibular left central incisor (, 0.659;, 0.704), mandibular right central incisor (, 0.571;, 0.617), mandibular right canine (, 0.723;, 0.723) (Fig 3), and mandibular left first premolar (, 0.716;, 0.725). Results of the palatal depth measurements in Table V showed excellent agreement among the 2 computation modalities (, 0.916;, 0.933). Intraobserver error assessment in Tables VI and VII showed excellent correlation between the intraobserver measurements for the conventional and digital model measuring techniques. The intraobserver values for the conventional models ranged from to for the and to for the. The values for the digital models were to for the and to for the. DISCUSSION Study casts are still the cornerstone of the diagnostic armamentarium of orthodontists. 21 Recent technological advancements, such as OrthoCAD and e-models, have allowed the generation of digital models with versatility of manipulation and saving of 3D volume. 13 E-casts generated by OrthoCAD require a high-quality impression, bite registration, and

5 American Journal of Orthodontics and Dentofacial Orthopedics El-Zanaty et al 263 Volume 137, Number 2 Table VI. Intraobserver error on conventional models T1 T2 Parameter (mm) Maxillary interlateral incisor width * 0.992* Mandibular interlateral incisor width * 0.993* Maxillary anterior arch length * 0.979* Mandibular anterior arch length * 0.998* Maxillary arch perimeter * 0.999* Mandibular arch perimeter * 0.961* Maxillary left central incisor * 0.944* Maxillary right canine * 0.933* Mandibular right second premolar * 0.913* Mandibular right first molar * 0.999* T1, First observation; T2, second observation 1week later. Table VII. Intraobserver error on digital casts T1 T2 Parameter (mm) Maxillary interlateral incisor width * 0.992* Mandibular interlateral incisor width * 0.994* Maxillary anterior arch length * 0.971* Mandibular anterior arch length * 0.979* Maxillary arch perimeter * 0.998* Mandibular arch perimeter * 0.999* Maxillary left central incisor * 0.963* Maxillary right canine * 0.959* Mandibular right second premolar * 0.962* Mandibular right first molar * 0.954* T1, First observation; T2, second observation 1week later. shipping to the company. Then, transfer of the cast into the e-model via scanning is accomplished with special equipment. Highly sophisticated tools and devices for the Sure- Smile (OraMatrix) procedure are needed. 15 Despite all procedures, the information from conventional and digital casts is limited to the surfaces of the teeth. 22 On the other hand, with CT scans, the 3D algorithm allows the operator to visualize the roots of the teeth (Fig 4), the surrounding bone, and soft tissues (Fig 5), adding to the orthodontist s diagnostic capabilities. Since the development of CT, 3D imaging techniques have gained more applications in dentistry and orthodontics. The demand for the digitization, manipulation, and full use of the patient s virtual 3D model is increasing. The long-term goal of orthodontic digital treatment planning with 3D digital models is inevitable without the precision of the digital diagnosis phase. Hence, the purpose of this study was to compare the accuracy of dental measurements on 3D volume, by using 3DD software, with those recorded manually with a Boley gauge on conventional models. To obtain the highest accuracy possible and to overcome the high radiation dose (the most important problem in conventional CT scans in dental practice), we used a CT slice thickness of 1.25 mm. 23,24 Our results showed that arch width, length, and perimeter in both jaws had excellent agreement between the conventional models and 3DD with no exceptions, as shown in Tables I through III. The graphic representation of the for the maxillary arch perimeter is depicted in Figure 2; the trend line (red) and the line of perfect agreement (blue) nearly coincide, and the points show extremely low scatter. On the other hand, Figure 3 shows the lower values for the mandibular right canine; thus, some scattering of the points from the trend line is obvious. Strong agreement between the conventional model and the 3DD resulted for the measurements of mesiodistal width of the maxillary and mandibular teeth. The fair

6 264 El-Zanaty et al American Journal of Orthodontics and Dentofacial Orthopedics February 2010 According to Stevens et al, 16 as long as precise measuring points are selected on the computer screen, it is reasonable that digital measurements are more valid than those made by calipers on casts. However, the operator s skill and learning curve in precisely measuring accurately with the computer mouse on the screen are contributing factors in the accuracy of all computerbased digital casts. Assessment of intraobserver error by using the and the conformity and equivalency tests, respectively, showed excellent agreement for both conventional and 3DD models (Tables VI and VII). Fig 4. Changing the window width and window level to view different structures in the same data set (note the roots of the teeth). CONCLUSIONS Dental measurements obtained from the 3DD are comparable with those from conventional models in the 3 planes of space. This technology has the added benefits of eliminating the need for taking impressions and the time needed for making models. With conebeam CT becoming mainstream in orthodontic practice, are we going to an era whereby a single step of data acquisition can compute all our diagnostic information? Fig 5. 3D volume showing the soft and hard tissues. agreement between the measuring methods for some teeth could be attributed to the difficulty in identifying the precise contact areas of adjacent teeth. The blending of the contact areas of adjacent teeth in addition to the unfamiliarity of the observers with measurements of the 3D images on a computer screen could be another causative factor Excellent agreement between the measuring modalities under investigation was shown in the palatal depth parameter. REFERENCES 1. Quimby M, Vig K, Rashid R, Firestone A, Mayers M. The accuracy and reliability of measurements made on computer-based digital casts. Angle Orthod 2004;74: Ryden H, Bjelkhagen H, Martensson B. Tooth position measurements on dental casts using holographic images. Am J Orthod 1982;81: Joffe L. OrthoCAD: digital casts for a digital era. J Orthod 2004; 31: Peluso MJ, Josell SD, Levine SW, Lorei BJ. casts: an introduction. Semin Orthod 2004;10: Redmond WR, Redmond WJ, Redmond MJ. Clinic implications of digital orthodontics. Am J Orthod Dentofacial Orthop 2000; 117: Redmond WR. casts: a new diagnostic tool. J Clin Orthod 2001;35: Takada K, Lowe AA, DeCou R. Operational performance of the Reflex Metrograph and its applicability to the three-dimensional analysis of dental casts. Am J Orthod 1983;83: Redmond WJ, Redmond MJ, Redmond WR. The OrthoCAD bracket placement solution. Am J Orthod Dentofacial Orthop 2004;125: Sohmura T, Kojima T, Wakabayashi K, Takahashi J. Use of an ultrahigh-speed laser scanner for constructing three-dimensional shapes of dentition and occlusion. J Prosthet Dent 2002;84: Sohmura T, Wakabayashi K, Lowmunkong R, Hojo H, Kusumoto N, Okuda H, et al. 3D shape measurement of dental casts using medical x-ray CT. Dent Mater J 2004;23: Wang B, Li S, Zhou L. Development of a video system for threedimensional cephalometry and dental cast analysis. Zhonghua Kou Qiang Yi Xue Za Zhi 2000;35: Yamamoto K, Hayashi S, Nishikawa H, Nakamura S, Mikami T. Measurements of dental cast profile and three-dimensional tooth

7 American Journal of Orthodontics and Dentofacial Orthopedics El-Zanaty et al 265 Volume 137, Number 2 movement during orthodontic treatment. IEEE Trans Biomed Eng 1991;38: Freshwater M, Mah J. 3D digital dental casts using laser technology. J Clin Orthod 2003;37: Tomassetti JJ, Taloumis LJ, Denny JM, Fisher JR. A comparison of 3 computerized Bolton tooth-size analyses with commonly used method. Angle Orthod 2001;71: Mah J, Sachdeva R. Computer-assisted orthodontic treatment: the SureSmile process Am J Orthod Dentofacial Orthop 2001;120: Stevens DR, Flores-Mir C, Nebbe B, Raboud DW, Heo G, Major PW. Validity, reliability and reproducibility of plaster vs digital study casts: comparison of peer assessment rating and Bolton analysis and their constituent measurements. Am J Orthod Dentofacial Orthop 2006;129: Whetten JL, Williamson PC, Heo G, Varnhagen C, Major PW. Variations in orthodontic treatment planning decisions of Class II patients between virtual 3-dimensional models and traditional plaster study models. Am J Orthod Dentofacial Orthop 2006; 130: Lin LI. A concordance correlation coefficient to evaluate reproducibility. Biometrics 1989;45: Lin LI. Assay validation using the concordance correlationcoefficient. Biometrics 1992;48: Lin L, Torbeck LD. Coefficient of accuracy and concordance correlation coefficient: new statistics for methods comparison. PDA J Pharm Sci Technol 1998;52: Almeida MA, Phillips C, Kula K, Tulloch C. Stability of palatal rugae as landmarks for analysis of dental casts. Angle Orthod 1995;65: Macchi A, Carrafiello G, Cacciafesta V, Norcini A. Three-dimensional digital modeling and setup. Am J Orthod Dentofacial Orthop 2006;129: BjerklinK, Ericson S. How a computerizedtomography examination changed the treatment plans of 80 children with retained and ectopically positioned maxillary canines. Angle Orthod 2006;76: Vannier MW, Hildebolt CF, Conover G, Knapp RH, Yokoyama- Crothers N, Wang G. Three dimensional dental imaging by spiral CT. A progress report. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 1997;84: Houston WJ. The analysis of errors in orthodontic measurements. Am J Orthod 1983;83: Mutinelli S, Cozzani M, Manfredi M. Denatal arch analysis system. Prog Orthod 2004;5: Zilberman O, Huggare JA, Parikakis KA. Evaluation of the validity of tooth size and arch width measurements using conventional and three-dimensional virtual orthodontic casts. Angle Orthod 2003;73:301-6.