[Evaluation of image quality of two different three-dimensional cone-beam-scanners used for orthopedic surgery in the bony structures of the pelvis in comparison with standard CT scans]

Abstract

Aim: This study evaluated the image quality of two different cone beam CT scanners used in the operation theatre in pelvic trauma surgery in relation to their radiation dosage. Furthermore, the assumption that a higher dosage would result in better image quality was analysed by using the different acquisition scanner modes.

Material and methods: We scanned the acetabulum (n=4) and iliosacral joints (n=4) of two human cadavers with a conventional CT and with two mobile cone beam CT scanners (Siemens Arcadis Orbic 3D and Ziehm Vision Vario 3D). With the two cone beam CT scanners (3D-BV), we used 6 different acquisition modes with different radiation dosages. The axial views of all scans were exported and blinded. Subsequently, the images were evaluated by 7 medical doctors with regard to identifiability of cortical structures (acetabular joint, fovea capitis femoris, cortical bone of the femur head, iliosacral joint, and sacral foramina), and the quality of the cancellous structure of the femur head. The evaluation was performed on axial views by using a defined five-point score. The interrater quality was statistically analysed according to Cohen with the kappa coefficient. In addition, the Wilcoxon test was used to identify significances between the 21 paired results of the evaluators. For determination of the signal-to-noise ratio, a Catphan 600 reference block with two different test elements (Teflon, PMP) was used.

Results: Overall, the image quality of the conventional CT scans received the best score. Comparing the two 3D cone beams, the image quality of the Siemens Arcadis Orbic 3D in high-dosage mode received the best score (median: 2.40), the Ziehm Vision Vario 3D in low-dose mode without large patient key received the lowest score (median: 3.16). The differences in the 21 paired results of the two different acquisition modes were significant in 17 cases (p < 0.05) but the size of difference when comparing the different acquisition modes was almost always small. The interobserver agreement in one acquisition mode was low (kappa 0.008-0.134). The overall evaluation results of the same acquisition mode diverged by up to 2 score points. We noted a higher signal-to-noise ratio in the high dosage mode than in the low dosage mode.

Discussion: When using intraoperative 3D imaging with the cone beam CT technique for pelvic injury, image acquisition in low-dose mode is adequate in terms of signal-to-noise ratio and image quality. The image quality does not correlate linearly with a higher radiation dosage. Therefore, the pelvic trauma surgeon using this technique is encouraged to gather his own experience with low dose modes thereby reducing patient radiation exposure.