Page 38 - Craniomaxillofacial Implant Surgery - Jeroen P.J. Dings

36
Chapter 2
All absolute differences of linear measurements obtained from CBCT and MDCT reconstructions and the reference standard show statistically significant overestimation (Table 1). However, systematic difference of linear measurements with the gold standard on CBCT images was statistically significant influenced by different brightness and contrast settings (P = .0015). No such influence was found for linear measurements on MDCT images (P = .51).
Data was also examined for absolute differences between the different imaging modalities, contrast settings and anatomical locations (Table 2). No statistical significant differences with the gold standard were found for in CBCT imaging at the anterior nasal spine for both contrast settings. Linear measurements on CBCT images at the supraorbital ridge only proved statistically significant different at one contrast setting. With regard to linear measurements on MDCT images, no statistically significant differences with the gold standard were found at the lateral orbital ridge and temporal bone region.
Differences in interobserver measurement error between pooled CBCT and pooled MDCT images proved statistically significant for the anterior nasal spine, piriform aperture and inferior orbital ridge with MDCT imaging showing less variation in linear measurements (Table 2).
DISCUSSION
We assessed the reliability and accuracy of linear measurements on cross-sectional images linked to the 3D hard-tissue surface representations from CBCT and MDCT image data sets. High reliabilities in this study allowed further comparisons with the average of measurements for each imaging modality and different brightness and contrast settings (ranging from 0.98 to 0.99).
In contrast to the findings of Wikner18, our findings demonstrated statistically significant submillimeter overestimation for linear measurements on digital CBCT (0.39-0.53 mm) and MDCT images (0.57-0.59 mm) in comparison to physical measurements with a caliper. Inaccuracy of caliper measurements in this study was 0.15 mm although this uncertainty can be considered clinically insignificant. Literature shows not to overestimate spatial resolution in MDCT and CBCT volumes with a maximal accuracy in the range of half a millimeter.15