Abstract
BACKGROUND: This study assessed the linear accuracy of three-dimensionally (3D)-printed mandibular models from cone-beam computed tomography (CBCT) scans with two voxel sizes. MATERIALS AND METHODS: In this in vitro study, five dry human mandibles underwent CBCT with 0.2- and 0.3-mm voxel sizes. The images were converted to STL format, and the distances between (I) mental foramen (MF) and alveolar ridge crest, (II) MF and inferior border of the mandible (IBM), and (III) alveolar crest and IBM at the midline, as well as the (IV) left central incisor socket depth, (V) left second premolar buccolingual socket width, and (VI) right third molar buccolingual socket width were measured on the CBCT scans, 3D-printed models, and dry mandibles. Two observers recorded the measurements twice, 1 week apart. We analyzed the data using the intraclass correlation coefficient and Pearson's correlation test. Statistical significance was set at P < 0.05. RESULTS: Since the interobserver agreement was high, the mean data was used for the comparisons. The linear accuracy was high for MF-IBM, MF-alveolar crest, and alveolar crest-IBM distances, and second premolar and third molar buccolingual socket width. CBCT scans demonstrated reliable accuracy for left central incisor socket depth measurement, but a lack of significant correlation was found between the 3D-printing and gold-standard measurements of this variable. CONCLUSION: The linear accuracy of CBCT scans taken with 0.3- and 0.2-mm voxel sizes was comparable, and they may be used for the fabrication of linearly accurate 3D-printed models of mandible. 3D-printed models demonstrated high precision in all measured parameters except socket depth.