In-vivo accuracy of tooth surface reconstruction based on CBCT and dental MRI-a clinical pilot study.

Objectives Guided implant surgery (GIS) requires alignment of virtual models to reconstructions of three-dimensional imaging. Accurate visualization of the tooth surfaces in the imaging datasets is mandatory. In this prospective clinical study, in vivo tooth surface accuracy was determined for GIS using cone-beam computed tomography (CBCT) and dental magnetic resonance imaging (dMRI). Materials and methods CBCT and 3-Tesla dMRI were performed in 22 consecutive patients (mean age: 54.4 +/- 15.2 years; mean number of restorations per jaw: 6.7 +/- 2.7). Altogether, 92 teeth were included (31 incisor, 29 canines, 20 premolars, and 12 molars). Surfaces were reconstructed semi-automatically and registered to a reference standard (3D scans of stone models made from full-arch polyether impressions). Reliability of both methods was assessed using intraclass correlation coefficients. Accuracy was evaluated using the two one-sided tests procedure with a predefined equivalence margin of +/- 0.2 mm root mean square (RMS). Results Inter- and intrarater reliability of tooth surface reconstruction were comparable for CBCT and dMRI. Geometric deviations were 0.102 +/- 0.042 mm RMS for CBCT and 0.261 +/- 0.08 mm RMS for dMRI. For a predefined equivalence margin, CBCT and dMRI were statistically equivalent. CBCT, however, was significantly more accurate (p <= .0001). For both imaging techniques, accuracy did not differ substantially between different tooth types. Conclusion Cone-beam computed tomography is an accurate and reliable imaging technique for tooth surfaces in vivo, even in the presence of metal artifacts. In comparison, dMRI in vivo accuracy is lower. Still, it allows for tooth surface reconstruction in satisfactory detail and within acceptable acquisition times.