Array-projection geometry and depth discrimination with Tuned-Aperture Computed Tomography for assessing the relationship between tooth roots and the inferior alveolar canal

Objectives. The purpose of this study is to compare Tuned-Aperture Computed Tomography (TACT) image sets made with linear vertical, linear horizontal, conical, and x-shaped x-ray projection arrays with regard to observer accuracy in (1) measuring the distance from the apex of a tooth root to the middle of the inferior alveolar canal (IAC), (2) measuring the shortest distance from the surface of a tooth root to the surface of the IAC, and (3) determining whether the root is buccal or lingual to the IAC. The same relationships were also examined by means of pairs of images and the buccal object rule. Study design. Two artificial mandible sections with simulated IACs were fabricated. The same human mandibular premolar root was used in both models to prevent development of learning cues to differentiate between models. The models were imaged from both sides, resulting in 4 orientations. An optical bench was designed to precisely orient different beam projection arrays for production of TACT image sets and pairs of images for use with the buccal object rule. Twelve dentists participated as observers to independently assess the relationships between the tooth root and the IAC with regard to the 4 orientations. Results. For measuring the distance from the tooth apex to the middle of the IAC, TACT image sets made from conical and x-shaped arrays proved significantly more accurate than TACT image sets made from linear vertical or linear horizontal projections or pairs of digital images used with the buccal object rule (P < .05). For measuring the shortest distance between the tooth surface and the IAC surface, TACT image sets made from linear vertical, conical, and x-shaped projection arrays were significantly more accurate than measurements made through use of the buccal object rule or TACT image sets made through use of a linear horizontal projection array (P < .05). The following percentages of mistakes were made in determining whether the tooth root was buccal or lingual to the IAC: buccal object rule (58%), linear horizontal array (31%), linear vertical array (2%), conical array (0%), and x-shaped array (0%). Conclusions. For TACT, images acquired with conical and x-shaped beam projection arrays are preferred to those acquired with linear arrays for assessing the relationship between tooth roots and the IAC. TACT was found to be significantly more accurate than standard application of the buccal object rule for the relationships investigated.