RELATIONSHIP BETWEEN GRAY LEVELS IN CONE BEAM COMPUTED TOMOGRAPHY AND HOUNSFIELD UNITS IN COMPUTED TOMOGRAPHY

Abstract

<h3>Background</h3> A deficiency of studies on clinical or close clinical materials of biologic nature exists in the conversion of gray values in cone beam computed tomography (CBCT) to more logically useful Hounsfield units (HU) in multidetector CT. <h3>Objective(s)</h3> The aim of this study was to assess the HU of various tissue materials in CT with their corresponding gray levels in 3 different CBCT units and compare their correlation individually, to compare the actual HU (aHU) values for each material with derived HU (dHU) values from gray levels using the interpolation method, and to evaluate its reliability. <h3>Study Design</h3> An in vitro study was conducted using a human skull model, which was coated with a soft tissue equivalent and embedded with 6 premeasured titanium implants. It was imaged with 3 different CBCT scanners and 1 CT scanner. The exported DICOM (Digital Imaging and Communications in Medicine) data were viewed in Carestream 3-D imaging software. The data were analyzed using Pearson's correlation test to evaluate the correlation between observed aHU values from CT and the corresponding gray levels from each CBCT unit. The method of Mah et al. was used to calculate dHU from gray levels and compare them with aHU. <h3>Results</h3> Among the CBCT machines used, the highest correlation for implants and bone was with Kodak 9500; for teeth, Sirona Galileos; and for soft tissue and air space, Planmeca Promax. The dHU values had significant percentage differences from aHU values. <h3>Discussion/Conclusions</h3> 1. Although there are differences in gray values between same materials in different machines, every CBCT scanner showed good correlation between gray levels and the aHU from CT machine for at least 1 material. 2. For clinical or close clinical materials of a biologic nature, dHU using the current algorithm does not reliably correlate with the aHU of the materials.