Purpose To retrospectively assess the accuracy of measurements of temporal bone anatomy made from reconstructed clinical high-resolution computed tomography (HRCT) scans.
Methods Nine HRCT scans were performed on unselected clinical cases in which the subjects had a temporal bone study judged to be normal. The orbitomeatal line was prescribed for the direct axial sections. Variations in head position (rotation at the neck and lateral bending of the neck) were corrected by using the software supplied by the manufacturer. All measurements were done on standard 1-mm axial sections and axial reconstructions obtained from 1-mm coronal slices. The images were viewed at 4000 Hounsfield units (HU) window width and 1000 HU window level. Measurements (n = 3) made on 1-mm direct axial HRCT scans were compared with the measurements made on reconstructed axial HRCT images from the same nine patients. These values were also compared with published cadaver data.
Results The measurements obtained from axial reconstructed and direct HRCT series approximated each other in each of the nine individual studies and also fell within the range of published cadaver values. They demonstrated the expected normal temporal bone variability between individuals.
Conclusion Useful anatomic approximations can be measured in vivo from reconstructed clinical HRCT images. Pitfalls are improper window settings, head tilt, and rotation. This protocol is widely available and can be implemented retrospectively from clinical HRCT scans.