OBJECTIVE The purpose of this study was to evaluate the accuracy and reproducibility of linear measurements obtained from three-dimensional reconstructions of computed tomography (CT) scans. METHODS Ten rectangular acrylic blocks were prepared with titanium molybdenum alloy (TMA) markers spaced from 1 to 10 mm, respectively. A plastic sphere was prepared with 10 sets of TMA markers spaced at variable intervals of 1 to 10 mm. Each object was scanned three times at 3-mm slice thicknesses and 1.5 mm with 0.5 mm overlap slice thicknesses, as well as positioned in the CT scanner in two different directions (perpendicular and parallel) to the scanning beam. Intermarker distances of the reconstructed objects were then measured using the measurement tool of the MediCAD software and compared with measurements taken by hand with a vernier caliper. RESULTS Using the 3-mm cut protocol, the data indicated that inconsistency exists between intermarker distance in the scans when the rectangular objects were scanned parallel to the scanning beam. This finding was not seen using the 1.5-mm with 0.5-mm overlap slice thickness protocol. The intermarker distances for objects scanned perpendicular to the scanning beam were consistent but subject to demagnification in the range of 17% to 20% for both scanning protocols. CONCLUSIONS The orientation of the object to the scanning beam and slice thickness protocol appear to have an impact on the accuracy and variability of linear measurements taken in the x, y, z axes.