Image Quality of New 3D C-Arm for Intraoperative Imaging of the Pelvis is Comparable to that of CT

Objectives: In recent years, 3D image intensifiers have increasingly found their way into operating rooms, improving the quality and accuracy of intraoperative imaging.

While the quality of the volumetric image acquisition has increased, intraoperative 3D imaging is still not regarded as being equivalent to CT in terms of the ability to identify structures with high accuracy and to reduce artifacts which may prevent correct visualization anatomical structures and implant positioning. Therefore, postoperative CT imaging remains the gold standard for postoperative control using high-resolution sectional imaging.

A potential advantage of an intraoperative 3D C-arm with CT-quality imaging may be a reduction in revision surgery rate by accurately controlling implant placement intraoperatively, therfore reducing costs and improving morbidity.

The aim of this study was therefore to evaluate the image quality and implant positioning in pelvic screw osteosynthesis using intraoperative flat-panel detector 3D scans compared to conventional CT.

Methods: In five human cadavers two sacroiliac screws (S1) were inserted free-hand after creating of artificial fractures of the pelvis. Subsequently, a CT scan (Siemens Somatom Force) as well as a 3D scan (Ziehm RF3D) of each pelvis was performed. Using a visual analog scale (VAS) and previously developed image quality assessment points scale (Stuebig T, J Trauma 2007), four orthopaedic surgeons evaluated the images using a standard DICOM viewer. The screw position in relation to several anatomical structures was also assessed. All analyses were conducted using R.

Results and Conclusion: The observers rated the quality of the imnages with a mean VAS of 8.1 (SD 0.5) for CT images and 5.3 (SD 0.9) for RFD3D images (P < 0.05).

Using the points system, the mean quality of CT was 2.3 (SD 2.8) points, while the RFD3D images were rated with a mean of 2.8 (SD 0.5) points (P > 0.05).

Regarding the objective implant and anatomical measurements, we found no significant difference between CT and RFD3D at the right SI joint location (mean difference 0.9, p = 0.2336), at the the left SI joint (mean difference 0.9, p = 0.6635), or at the level of the vertebral body (mean difference 0.3, p = 0.9275).

Conclusions: While the VAS score was higher for CT images, we found no difference in the points system or in objective anatomic measurements between CT and RFD3D images.

These results show that intraoperative imaging using the RFD3D device may be of similar quality to that of CT, which could result in improved intraoperative diagnostic and guidance ability and may help detect incorrect implant placement before the patient has left the OR, therefore reducing costs and resources by avoiding a separate revision surgery.

Stichwörter: 3d imaging; intraoperative imaging; ziehm; ct;