Assessing the Reliability of Digital Models: Comparing Plaster, Laser-scanned and Cone-beam Computer Tomography Models

Objectives: The aim of this study was to evaluate the reliability and validity of digital models measurements by comparing plaster models, laser-scanned models and models generated from cone-beam computed tomography (CBCT) scans. Plaster models are considered the gold standard of diagnosis in orthodontic treatment, thus laser scanned models and CBCT model measurements will be compared to those of the gold standard.
Methods: The sample of this study included 20 orthodontic subjects that were previously scanned with CBCT and had CBCT generated dental models (Anatomodel- Anatomage, San Jose, CA). Following IRB approval, stone dental models of the same subjects were scanned using Ortho Insight 3D laser scanner (Motionview LLC, Hixson, TN) to generate the laser scanned models. The scan data of the models were exported from the laser scanner in STL format file extension. Dental arch parameters, including arch widths, lengths and tooth size were located and measured digitally on the laser scanned models and on the Anatomodels using Motionview software (Motionview LLC, Hixson, TN) and InvivoDental 5.1 CBCT software (Anatomage, San Jose, Calif) respectively. The stone models were measured using a boley gauge. Intraclass correlations (ICC) were performed on duplicate measures of 10 dental models for each type to assess reliability. ANOVA was used to assess the differences of the selected parameters among the dental models. Statistical significance was set at p< 0.05.
Results: ICC value were >0.80 for all measurements. Statistically significant differences were detected for maxillary inter-molar width, arch length as measured from the upper left first molar to the upper left canine cusp tip, and the maxillary upper right canine tooth length.
Conclusions: Digital dental models can be used for orthodontic diagnosis because they provide similar representations of dental arch anatomy. Although statistically significant differences were found in some parameters, they are considered clinically acceptable.