RESUMO
ABSTRACT: Historically, the accuracy of imaging teeth by computed tomography (CT) has been suboptimal and deemed inadequate for surgical planning of orthognathic procedures. However, recent advances in CT hardware and software have significantly improved the accuracy of imaging occlusal anatomy. This technical note describes a quantitative means of evaluating the accuracy of CT-based modeling of teeth. Three-dimensional models of the dentition were created from a CT scan obtained of a craniomaxillofacial skeleton. Multiple reconstruction algorithms and modeling parameters were applied. The dentition of the same skeleton was scanned using a handheld optical scanning device to serve as the "gold standard." Semi-automated registrations of CT and optically acquired models were performed and deviation analysis was conducted. On average, the deviation of the CT model with the optical scan measured 0.19 to 0.25 mm across the various reconstruction and modeling parameters, with a mean of 0.22 mm. Computed tomography underestimated contours at cusp tips, while overestimating contours in occlusal groves. The use of bone reconstruction algorithms and decreased model smoothing resulted in more accurate models, though greater surface noise. Future studies evaluating the clinical effectiveness of CT-based occlusal splints should take this finding into account.
