Reliability of hypocycloidal tomography for the evaluation of the distance from the alveolar crest to the mandibular canal.

When implants are to be inserted in the mandible posterior to the mental foramen, tomography is required to determine the exact location of the mandibular canal and assess the dimensions of the body and alveolar process. Previous studies have shown discrepancies between the true distance between the alveolar crest and the upper border of the mandibular canal and that measured from computed tomograms. This study evaluates the reliability of such measurements made from hypocycloidal tomograms. Six examiners, three radiologists and three oral surgeons, independently measured the distance from the alveolar crest to the upper border of the mandibular canal in transverse hypocycloidal tomograms of the mandible. Analysis of variance showed an average standard deviation between observers of 1.33 mm. This variation, which is the total of inter- and intra-observer variation conditional upon patient and radiograph, was mainly due to the latter which amounted to a mean of 1.04 mm. Calculation of lower confidence limits showed that these could be decreased by the use of several observers. Various confidence levels, equal to the probabilities of obtaining lower confidence limits which are below the true values, can be used to choose acceptable levels of risk in overestimating the true distance.