Linear accuracy and reliability of cone beam CT derived 3-dimensional images constructed using an orthodontic...

To compare accuracy of linear measurements made on cone beam computed tomographic (CBCT) derived 3-dimensional (3D) surface rendered volumetric images to direct measurements made on human skulls.

Twenty orthodontic linear measurements between anatomical landmarks on 23 human skulls were measured by observers using a digital caliper. The skulls were imaged with CBCT and Dolphin 3D (version 2.3) software used to generate 3D volumetric reconstructions (3DCBCT). The linear measurements between landmarks were computed by a single observer three times and compared to anatomic dimensions using Student's t-test (P < or = .05). The intraclass correlation coefficient (ICC) and absolute linear and percentage error were calculated.

The ICC for 3DCBCT (0.975 +/- 0.016) was significantly less than for skull (0.996 +/- 0.007) measurements. Mean percentage measurement error for 3DCBCT (2.31% +/- 2.11%) was significantly higher than replicate skull measurements (0.63% +/- 0.51%). Statistical differences between 3DCBCT means and true dimensions were found for all of the midsagittal measurements except Na-A and six of the 12 bilateral measurements. The mean percentage difference between the mean skull and 3D-based linear measurements was -1.13% (SD +/- 1.47%). Ninety percent of mean differences were less than 2 mm, and 95% confidence intervals were all less than 2 mm except for Ba-ANS (3.32 mm) and Pog-Go(left) (2.42 mm).

While many linear measurements between cephalometric landmarks on 3D volumetric surface renderings obtained using Dolphin 3D software generated from CBCT datasets may be statistically significantly different from anatomic dimensions, most can be considered to be sufficiently clinically accurate for craniofacial analyses.