The effect of different CT scanners, scan parameters and scanning setup on Hounsfield units and calibrated bone density: a phantom study

This study investigated the effect of different CT scanners, reconstruction protocols, scan positions, and air gaps on HU and/or calibrated calcium hydroxyapatite concentrations (CaHA). Phantoms containing bone-like materials were scanned on four different CT scanners. The effect of slice thickness, field of view (FOV) and reconstruction kernel on HU was investigated. Using clinical scan and reconstruction protocols and a calibration phantom, HU and CaHA were determined for different positions, and air gaps between phantom and calibration phantom. Changing reconstruction kernel considerably affected the HU (range −31 HU to 64 HU), whereas slice thickness and FOV did not. Inter-scanner differences in HU were <88 HU and decreased to <47 CaHA after calibration. Different positions of the phantom resulted in differences (on average up to −26 HU and −24 CaHA). An air gap of 2.5 cm atop the calibration phantom resulted in errors up to −41 CaHA. If absolute HU and calibrated CaHA are needed, different reconstruction kernels and changes in position within the FOV as well as air gaps should be avoided.