The purpose of this project was to compare the accuracy and precision of linear measurements made on conventional lateral cephalograms (LC) and simulated LC using cone-beam computed tomography (CBCT) to direct measurements on dentate skulls. The linear distances for 5 mid-sagittal and 4 bilateral measurements between anatomical landmarks on 22 skulls were measured by two observers. The skulls were imaged using a CBCT scanner at 4 scan settings: high resolution (0.2mm) and normal resolution (0.4mm) at 20-sec (306 projections) and 40-sec scans (612 projections). Three simulated LC were generated: Ray sum image reconstructions of the maximum thickness mid-sagittal multiplanar projections, a single frame basis projection, and the scout image. LCs were acquired using a storage phosphor system. Projection images were exported into a cephalometric analysis program and linear measurements determined. Analyses were repeated three times and statistically compared to measured anatomic truth using repeated measures general linear model (p≤0.05). The mean absolute error was used as an index of precision.

While there were statistical differences between skull and modality measurements for 6 of the nine measurements, there were no overall differences between LC and CBCT derived image measurements irrespective of method of construction. CBCT imaging provides clinically reliable 2D simulation images however, dose considerations demand that evidence-based selection criteria should be developed for CBCT in orthodontics that take into account the ALARA principle.