The aim of the research is to determine if cone beam computerized tomography reconstructed panoramic radiography viewed in two different focal trough shapes, and three different focal trough widths changes the ability to identify high risk radiologic signs associated with an intimate anatomic relationship between third molars and the Inferior Alveolar Nerve. The basic assumption of this research is that a customized focal trough shape with a wider focal trough width at the third molar region will produce more high risk radiologic signs present when there is a relation between the third molar and the Inferior Alveolar Canal.

A retrospective sample of 50 mandibular third molar teeth being less than 3mm from the Inferior Alveolar Canal were picked from the files of the i-CAT CBVCT here at the University of Louisville Radiology Department. The 50 teeth were reconstructed into panoramic images with two different arch shapes (customized and average form), and three different focal trough widths at the third molar region (10, 20, and 40mm). Six high risk radiological panoramic signs of the third molar related to the Mandibular canal will be evaluated at each tooth and determined to be present or absent by two independent observers. Each observer will also utilize the CBVCT cross sectional analysis of the mandible, measure the distance in mm, the mandibular canal to the closest part of the mandibular third molar tooth. With these observations we will be able to determine the presence or absence of the high risk radiological signs and actually see if the teeth that have the high risk radiological signs are associated with the Inferior Alveolar Nerve, related to the focal trough image layer widths. This analysis will be compared in the two focal trough image layer shapes to determine if a average form of dentition focal trough shape will present more radiographic markers than the customized focal trough shape.

A Ordinal Logistic Regression will be performed to evaluate the relative impact of the predictor variables (radiographic signs, arch forms, and focal trough widths) to the outcome variables (distance of tooth to nerve-groups). Descriptive analysis will also be performed on each tooth to describe the comparison of radiological signs present, group that each tooth falls in related to distance of mandibular nerve in mm to the root of the third molar, whether nerve is buccal, lingual, or central to the root of the tooth, and if the nerve runs through the root, or the root is notched by the nerve.

For all tooth images, radiographic signs will be determined whether they are present or absent using two observers as independent experts to determine accuracy. Ordinal logistic regression analysis will be used to measure the outcome of radiographic signs present with three categories (Cat. 1, 0-1mm, Cat. 2, 1.001-2mm, Cat. 3, 2+mm). Intraclass correlation coefficient will be used as a measure of agreement to measure both inter and intra rater variability. With respect to intra-observer variability, inter-observer variability, there were significant differences in intra-observer and inter-observer variability.