Point-based superimposition of a digital dental model on to a three-dimensional computed tomographic skull: an accuracy study in vitro.

We investigated the accuracy of point-based superimposition of a digital dental model on to a 3-dimensional computed tomographic (CT) skull with intact dentition. The physical model was scanned by CT to give a virtual skull model, and a plaster dental model was taken and laser-scanned to give a digital dental model. Three different background investigators were recruited and calibrated to make the point-based superimposition, and afterwards were asked to repeat 5 superimpositions each. Five bone-to-tooth measurements for the maxilla and 6 for the mandible were selected to indicate the relation of teeth to skull. Repeated measures were made on the physical model to act as a control group, and on the virtual model to act as the test group. The absolute agreement intra-class correlation coefficient (ICC) was used to assess the intra/inter-investigator reliability; Bland-Altman analysis was used to calculate the general differences, limits of agreement, and precision ranges of the estimated limits. Inter/intra-investigator reliability was excellent with ICC varying from 0.986 to 1; Bland-Altman analysis indicated that general difference was 0.01 (0.25)mm, the upper limit of agreement was 0.50mm and the lower limit -0.47 mm, and the precision range for the upper limit was 0.43 mm to 0.57 mm and for the lower limit -0.54 mm to -0.40 mm. Clinically acceptable accuracy can be achieved using a direct point-based method to superimpose a digital dental model on to a 3-dimensional CT skull.