METHODS: 10 skulls were imaged using CBCT (NewTom-3G) and conventional cephalometry. CBCT volume data were exported in DICOM format into Dolphin 3D v1 pre-release where orthogonal and perspective lateral cephalometric radiographs were created from 3D virtual models. Cephalometric landmark identification and linear (9) and angular (5) measurements were made in Dolphin at 3 different times. Caliper measures of midsagittal landmarks (ANS-Me, N-Me, N-ANS) were made directly on skulls. Perspective and conventional image measurements were corrected for known magnification. Differences in repeated measurements were measured by MANOVA. Matched pairs of midsagittal skull and radiographic measurements were analyzed using a Bonferroni adjustment for multiple comparisons. Differences between averaged measurements for the imaging modalities were assessed using ANOVA.
RESULTS: Repeated measurements were not different for any modality (p>.05). Midsagittal linear measurements were not different from skull measurements for any radiographic modality (p>.017). ANOVA demonstrated differences between Skulls (p<.0001) and different landmarks (p<.0001). Image modality (p=.11) and interactions between modality and skull or landmark were not different (p>.05).
CONCLUSIONS: CBCT can reproduce conventional cephalometric geometry with similar precision and accuracy. Measurements on different skulls varied as expected for unrelated individuals. Different measurements also produced different values as expected. Although a skull by modality interaction was not seen in this study, asymmetric skulls may influence measurements with different modalities. Additional studies utilizing CBCT data from patients are needed to confirm CBCT cephalogram reliability for comparisons with pre-existing cephalometric databases.