Accuracy of Orthodontic Models Scanned With a Cone Beam Computed Tomography

Objectives: The aim of the study is to determine the accuracy of digital models produced by Cone Beam Computed Tomography (CBCT) from standard plaster orthodontic models
Methods: Forty-three single arch study models were used. Twelve points were identified and marked using radio-opaque composite dots to enable three smaller and 3 larger linear measurements (two in each dimension: height, width, and depth). Models were then scanned using a CBCT (I-CAT). Comparison of the measurements (256 total) between the physical model (using a digital caliper) and the 3-D digital model (I-CAT software) was conducted using the paired t-test. Repeated measurements were taken for reliability and were evaluated using Cronbach’s alpha.
Results: Reliability tests produced poor results (alpha <0.7) involving the smaller measurement in each dimension (canine height, mesiodistal of molar, and buccolingual of premolar) and were therefore excluded from comparisons. Reliability of larger measurements was high (alpha >0.9). Statistically significant differences were found in all remaining measurement. T-test results (p-value, mean difference) were: height of case (0.001, 0.49), molar to canine distance (0.02, 0.21), and intercanine distance (0.001, 0.30).
Conclusions: Although statistically significant, mean differences of less than 0.5mm are considered clinically acceptable and may be attributed to measurement error. Therefore, the use of CBCT for production of 3-D digital models provides a viable and accurate option for clinical use and research