Radiation dose and image quality comparison during spine surgery with two different, intraoperative 3D imaging navigation systems

Abstract Careful protocol selection is required during intraoperative three‐dimensional (3D) imaging for spine surgery to manage patient radiation dose and achieve clinical image quality. Radiation dose and image quality of a Medtronic O‐arm commonly used during spine surgery, and a Philips hybrid operating room equipped with XperCT C‐arm 3D cone‐beam CT (hCBCT) are compared. The mobile O‐arm (mCBCT) offers three different radiation dose settings (low, standard, and high), for four different patient sizes (small, medium, large, and extra large). The patient's radiation dose rate is constant during the entire 3D scan. In contrast, C‐CBCT spine imaging uses three different field of views (27, 37, and 48 cm) using automatic exposure control (AEC) that modulates the patient's radiation dose rate during the 3D scan based on changing patient thickness. hCBCT uses additional x‐ray beam filtration. Small, medium, and large trunk phantoms designed to mimic spine and soft tissue were imaged to assess radiation dose and image quality of the two systems. The estimated measured “patient” dose for the small, medium, and large phantoms imaged by the mCBCT considering all the dose settings ranged from 9.4–27.6 mGy, 8.9–33.3 mGy, and 13.8–40.6 mGy, respectively. The “patient” dose values for the same phantoms imaged with hCBCT were 2.8–4.6 mGy, 5.7–10.0 mGy, and 11.0–15.2 mGy. The CNR for the small, medium, and large phantoms was 2.9 to 3.7, 2.0 to 3.0, and 2.5 to 2.6 times higher with the hCBCT system, respectively. Hounsfield unit accuracy, noise, and uniformity of hCBCT exceeded the performance of the mCBCT; spatial resolution was comparable. Added x‐ray beam filtration and AEC capability achieved clinical image quality for intraoperative spine surgery at reduced radiation dose to the patient in comparison to a reference O‐arm system without these capabilities.


| INTRODUCTION
Pedicle screw malposition can lead to various complications such as vascular and visceral structure damage, or dural lesions and radiculopathy, which might require revision surgery. 1 Image-guided spine surgery using intraoperative three-dimensional (3D)-based navigation increases clinical accuracy of pedicle screw placement compared to free-hand or fluoroscopy-guided placement. Consequently, this improves patient safety. [2][3][4][5] In order to perform a spine surgery with navigation, an intraoperative 3D scan of the spinal region of interest is performed to use during navigation for localization of the instruments in the various 3D planes. At the end of the procedure, a second intraoperative 3D scan can be performed to assess the correct placement of introduced hardware as an alternative to radiographs or postoperative CT. Regardless of which type of camera navigation is used, it was demonstrated that the image quality of the intraoperative 3D scan can influence the accuracy of navigation. 6,7 Proper management of patient radiation dose is necessary. While use of 3D imaging-based navigation decreased periprocedural radiation dose to staff, patient radiation dose increased. 8 Better configuration of imaging equipment may achieve ALARA (as low as reasonably achievable) patient radiation doses. 9 This is particularly important for the pediatric population with a higher risk of radiationinduced cancer. 10,11 A mobile O-arm system is a commonly used intraoperative 3D imaging system during spine surgery. 12,13 Other intraoperative 3D imaging systems have been widely used such as the Iso-C 3D and the AiroCT. 12,13 Intraoperative 3D imaging improves the accuracy compared to preoperative 3D registration as well as the time per screw placement, which significantly reduces radiation dose to the patients. 14 Other studies have investigated the patient's radiation dose from available systems and focused on the care needed to insure patient safety. 15,16 While some radiation dose reduction during intraoperative 3D navigation compared to C-arm fluoroscopyguided spine surgery may occur, 17 complex spine deformity remains a surgery where radiation dose to the patient is still significant when performed with intraoperative 3D navigation compared to other conventional techniques. 18,19 A hybrid operating room (OR) is a surgical suite containing an integrated C-arm system within the room construction. Hybrid ORs are used in vascular surgery and have shown radiation dose reduction to the patient 20 The benefit of using imaging equipment within a hybrid OR for pedicle screw placement and assessment with intraoperative 3D imaging has been documented. 21,22 Such benefits include the seamless integration of the C-arm and the surgical table within the OR suite, the robotic movement of the C-arm, and the ability to intraoperatively replace/adjust misplaced screws that could potentially require revision surgery after discovery during follow-up CT imaging after closure. However, these studies have not investigated the corresponding radiation dose.
This study compares the management of radiation dose and simple analysis (measurements readily made in the field) of image quality of the planning phase of 3D scanning for navigation planning with a hybrid OR robotic ceiling-mounted 3D cone-beam CT unit (hCBCT) 21 to that of a mobile O-arm unit (mCBCT) as a reference.

2.A | Imaging devices
The different design technologies found in an O-arm O2 (Medtronic, Littleton, MA) and an XperCT Augmented Reality Surgical Navigation System (Philips Healthcare, Best, the Netherlands) are compared ( Fig. 1). Both systems use a flat panel detector, which enables twodimensional fluoroscopy as well as 3D CBCT. The mCBCT system is mounted on casters while the hCBCT unit is a robotized ceilingmounted C-arm system. The source to detector distance (SID) during 3D imaging is fixed at 119 cm on both systems. Both systems' hardware specifications are detailed in Table 1. The large focal spot was selected, 0.7 and 1.2 mm, respectively on the hCBCT and mCBCT during 3D CBCT imaging.
Both systems use several preprogrammed anatomical locations for example, head, chest, abdomen, etc., for 3D acquisitions. Since this study focuses on lumbar spine imaging, the "lower torso/hip" and "Spine" protocols were chosen for the mCBCT and hCBCT, respectively. Two different parameters are used for 3D acquisitions on the mCBCT: four patient thickness [small (S), medium (M), large F I G . 1. Hybrid operating room with a surgical table and a ceiling-mounted motorized C-arm Philips AlluraClarity system with augmented reality navigation camera mounted in the C-arm detector and displayed on the hanging medical monitor (on the left). Medtronic O-arm O2 system with infrared navigation camera system and medical imaging display carts (on the right). Courtesy of Philips Healthcare and Medtronic.  The AEC modulates the x-ray tube current, which varies the dose rate to the patient for each angle of rotation. This maintains a relatively uniform detector dose rate despite varying patient thickness due to the rotation of the x-ray beam about the patent. Units without AEC capability, require the operator to manually select a constant dose rate prior to a scan resulting in a relative constant dose rate to the patient and varying dose rate to the detector due to the elliptically shaped patient. The chosen dose level by the operator either improves image quality by reducing noise (more dose), or reduces image quality (less dose). [23][24][25] Table 1 summarizes the difference in acquisition protocols and system specifications between the mCBCT and hCBCT units. 26,27 Note that for other applications than spine (e.g., pelvis) imaging, the mCBCT offers the largest CBCT diameter reconstruction of 40 cm.
This study was conducted in a hybrid OR room equipped with a surgical table (Alphamaquet 1150, Maquet AG, Switzerland) connected to a motorized ceiling-mounted C-arm flat detector system hCBCT. The mCBCT system was wheeled into the hybrid OR room to allow imaging with the same OR table. are used to measure spatial resolution. Figure 3 illustrates the spatial T A B L E 1 3D cone-beam CT scanning parameters and reconstructed volumes of the various tested spine protocols on Philips hCBCT and O-arm mCBCT for spine intraoperative imaging. arrangement of these test objects, the object's size, and the corresponding images. Because of the small size of the ACR phantom, we used the small patient thickness (S) protocol on the mCBCT with a kV of 120 and mAs of 64, 128, and 188 for the LD, SD, and HD imaging protocol, respectively ( Table 1). The kV on the hCBCT was also at 120 with a mAs set by the AEC as 85, 93, and 107 with the 48, 37, and 27 cm FoV imaging protocol, respectively. The cylindrical axis of the ACR phantom was iso-centered when imaged with both systems. The boundary between Section 2 and 3 of the ACR phantom was centered on the central ray of the cone beam.

2.C | Radiation dose
The

2.G | Statistical analysis
A weighted linear regression of the measured HU in the various materials with known HU in the ACR phantom was performed. The 95% confidence interval (CI) of the estimated slope and intercept were calculated as well as the adjusted-R 2 of the regression. A Pvalue below 0.05 was considered for statistical significance. All calculations were performed by using a statistical software package (Matlab, version R2015a; Mathworks, Natick, MA).

3.A | Radiation dose
The patient dose index range for the small, medium, and large CIRS phantoms imaged by the mCBCT for LD, SD, and HD was 9.

3.C | Hounsfield unit accuracy
The HU accuracies of all CBCT performed on the ACR phantom are shown in Fig. 7 for all protocols on both systems as well as the desired measurement line with a slope of 1 and an intercept of 0. 3.D | Uniformity, noise measurement, contrast, and spatial resolution from the ACR phantom Uniformity, noise measurements, and spatial resolution are listed in    Table 1) also leveraged reductions in noise levels.
The measured uniformity of the mCBCT images was less than that of the hCBCT. The average energy of the x-ray beam at greater depths in the patient increased due to attenuation of low energy x rays. This increased the value of HU at the periphery of large diameter objects. 37 Clinically, these artifacts translate into over estimation of HU values of low-attenuating structures (e.g., soft tissue) between high-attenuating structures (e.g., bone). Beam-hardening artifact correction algorithms are implemented as part of the reconstruction process only on the hCBCT to improve uniformity.
Low contrast resolution of 6 HU on the ACR phantom could not be visualized on either system. This corroborates with a reported failure of the mCBCT to visualize 10 HU. 37 An increase in acquired projections can decrease noise and improve the visibility of low contrast soft tissues. For example, using 620 projections on the hCBCT unit for head imaging or a smaller FoV for spine imaging enabled a low contrast resolution of 3-5 HU. 21 Spatial resolution on the mCBCT was 10 lp/cm, which was Intraoperative imaging of screw position is imperative to identify and move incorrectly placed screws without costly revision surgery. 39 The diagnostic imaging performance of both units remains less than that of conventional CT. 40 To date, intraoperative imaging with the mCBCT has not eliminated postoperative imaging with conventional CT. 8 Further dose reduction is needed to make it attractive without compromising image quality, especially for cases requiring intervention at multiple levels of the spine like scoliosis. With dose reductions and improved image quality of technologies similar to those of hCBCT, clinical studies are needed to evaluate whether the patient radiation dose is low enough and imaging quality high enough to match the total periprocedural x-ray radiation dose from x-ray-guided surgery or free-hand with x ray.
Despite the fact that our phantom study had the advantage of experimental reproducibility at different dose settings and configurations, translation of the results to surgical settings is not straightforward. Quantitative metrics such as CNR, contrast, and spatial resolution are not easily correlated to the quality of an imaged patient spine. Hardware solutions using an angulated C-arm to avoid x-ray beam and metal implants collinearity 21 as well as software for metal artifact reduction algorithms exist, 41

CONFLI CT OF INTEREST
None of the authors were paid for this work.