Page 40 - Craniomaxillofacial Implant Surgery - Jeroen P.J. Dings

38
Chapter 2
Although our study made use of dried skulls like many other papers, the clinical truth is hampered by the absence of soft-tissue attenuation.3,4 Linear measurements on objects without a simulated soft tissue component may be more accurate due to a more optimal high-contrast resolution and decreased scatter.10,14,26 However, Ganguly et al14 found no effect of soft tissues on the accuracy of measurements. In our study no metallic artifacts were present and, due to stabilization of the skulls, motion artifacts were ruled out. Image acquisition was completed in normal scanning position, although several studies have stressed that linear measurements on CBCT and MDCT images are not significantly influenced by the position and inclination of the object during scanning.2,10,18,22,30 Also, tube current reduction in CBCT and scan mode is shown to have little influence on image quality.5 In the study of Al-Ekrish et al41 lower image contrast only affected low-contrast resolution. However, with regard to the goal of pre- surgical assessment of bone volume only high-contrast resolution is required.35,41
Most measurements in this study comprised only a few millimeters in bone width necessitating a high spatial resolution.35 Spatial resolution, as the size of the acquisition voxel, depends on different reconstruction parameters (i.e. reconstruction algorithm), geometrical aspects and acquisition mode. The optimal exposure settings for obtaining clinically adequate image quality needs to be determined for each CBCT and MDCT device.5,31
A limitation of this study is that only one CBCT and MDCT scanner was used with only one image acquisition protocol and one imaging software package. Chen et al34 showed different software packages offering reliable dimensional measurements but with underestimation compared to gold standard measurements. However, significant differences were observed in volume and cross-sectional measurements using different CBCT and MDCT scanners.20
Conform standard practice settings in our university clinic the spatial resolution in our study was 0.3 mm for the CBCT and 0.5 mm for the MDCT scan. In comparison to MDCT, CBCT voxels are very small (ranging from 0.076 to 0.4 mm) and generally isotropic. These characteristics make linear measurements possible in all planes.37,42,43 Some publications state that voxel size has no significant influence on accuracy of diagnoses or image quality.12,13,35 In contrast, Maret et al39 and Shokri9 et al indicated a slight tendency toward underestimation in volumetric measurements with increasing CBCT voxel size. Maret et al39 elucidated a statistically significant underestimation using voxel sizes of 0.3 mm and beyond. However, a voxel is only a very crude predictor