Accuracy and reliability of landmark-based, surface-based and voxel-based 3D cone-beam computed tomography...

To evaluate and compare the accuracy and reliability of 3 different methods of three-dimensional cone-beam computed tomography scans (3D CBCTs) superimpositions: landmark-based, surface-based and voxel-based.

Pre- and post-orthodontic treatment CBCTs (T1 and T2) of 20 subjects with a mean age of 11 years were obtained. Seven points on the zygomatic arch and supraorbital region were selected to perform landmark-based superimposition. Surface-based and voxel-based superimpositions were performed using the anterior cranial base as a reference. Each superimposition method of T1 and T2 scans was repeated twice to assess the reliability. Accuracy of each technique was tested by superimposing duplicated sets of T1 scans. A total of 11 landmarks on the anterior cranial base, maxilla and mandible were located, and deviations of these landmarks on superimposed data were quantified to assess reliability and accuracy of all superimpositions.

There were no significant differences from zero when duplicated sets of T1 scans were superimposed using surface-based and voxel-based methods. Statistical significant differences were detected in several parameters when evaluating the accuracy of the landmark superimposition. Superimposition of T1 and T2 scans for testing the reliability revealed intraclass correlation coefficients greater than 0.90 for all measurements except for ACP-x and PNS-y of landmark-based method as well as ANS-x of voxel-based method.

Surface-based and voxel-based superimposition methods using the anterior cranial base as a reference structure were accurate and reliable in detecting changes in landmark positions when superimposing. Landmark-based superimposition method was reliable but less accurate than the other methods.