Augmented Reality Surgical Navigation in Spine Surgery to Minimize Staff Radiation Exposure

Erik Edström, MD, PhD; Gustav Burström, MD; Artur Omar, PhD; Rami Nachabe, PhD; Michael Söderman, MD, PhD; Oscar Persson, MD; Paul Gerdhem, MD, PhD; Adrian Elmi-Terander, MD, PhD


Spine. 2020;45(1):E45-E53. 

In This Article

Materials and Methods

Patient Data

This study was approved by the local ethical committee. All patients signed informed consent. Twenty consecutive patients undergoing spine surgery with posterior instrumentation between January and October 2017 were included in the study. The median age was 18.5 [16–72]. The median body-mass index of the patients was 19.9 [15–23.3]. The median weight and height were 60 kg and 173 cm, respectively.

The Hybrid Operating Room With Intraoperative Three-dimensional Augmented Reality Surgical Navigation System

The surgical procedures were performed in a hybrid OR using a ceiling-mounted robotic C-arm (Allura Clarity Flexmove, Philips Healthcare, Best, the Netherlands), with an integrated navigation system implementing ARSN (Augmented Reality Surgical Navigation).[10] Figure 1 describes a hybrid OR with an integrated ARSN system in a robotic C-arm. 2D fluoroscopy imaging with 3.75 x-ray pulses/second was used for spinal level identification and iso-centering of the region of interest prior to 3D cone-beam computed tomography (CBCT) imaging. An intraoperative CBCT image was acquired for navigation planning and screw placement and another CBCT was performed to confirm correct screw placement prior to wound closure, replacing the postoperative CT.[12] There were three types of CBCT protocols available, covering a small (12.6 × 12.6 cm2), medium (17.3 × 17.3 cm2), and large (25.2 × 19.5 cm2) field-of-view. The small field-of-view protocol used 482 x-ray pulses to generate the CBCT image (the other protocols used 302 pulses), yielding a better reconstructed image quality at the expense of an increased patient radiation exposure. The large field of view protocol covering the most spinal levels and using fewer x-ray pulses is referred to as a low-dose protocol. All three protocols had a fixed tube kilovoltage of 120 kV, and a tube current-time product ranging from 50 to 325 mAs, modulated by an automatic dose rate control to achieve a similar image quality independent of the patient size.[13] Table 1 summarizes the settings used for the different CBCT acquisition protocols.

Figure 1.

A, Hybrid operating room showing the ceiling suspended robotic C-arm. B, The Augmented Reality Surgical Navigation cameras are integrated at each side of the x ray detector displaying the real-time video of the surgical field augmented with the navigation path of the instrument held by the surgeon. C, Display on one of the suspended medical monitors of the real-time augmented reality of a navigated device along a predefined planned path on the intraoperative cone beam CT.

Occupational OR Staff Dose

Each of the OR staff members present during the surgical procedures wore an active personal dosimeter (APD) on the chest. Each APD was calibrated in terms of the personal dose equivalent Hp(10) measured in units of μSv. The APDs were connected to the staff dosimetry system DoseAware Xtend (Philips, Best, the Netherlands), which is an integrated part of the OR x-ray system. The staff dosimetry system provided real-time information on the radiation exposure of the staff (presented on a medical monitor), and stored the value measured from each separate x-ray exposure in a DICOM operator dose structured report (ODSR).[14] The ODSR files were used to extract the presented staff exposure data. It should be noted that two of the performed procedures were excluded from the analysis, as complete ODSR files were missing.

A transparent lead shield wall (2 mm lead) was placed in the corner of the OR, to shield the staff from scattered x rays when imaging the patient (Figure 2). The imaging technologist used a remote system control unit placed behind a separate transparent 2 mm lead shield (Figure 2). To evaluate exposure reduction by implementing this kind of radiation protection strategy, the exposure measured by the staff APDs was compared with the exposure measured by a reference APD attached to the C-arm. The reference APD measurements correspond to a worst-case staff exposure situation: the staff being directly exposed to scattered x rays. Lead shielding and distance were the only means of radiation protection used. The staff did not wear lead aprons during any of the surgical procedures.

Figure 2.

Hybrid operating room setup during a surgical procedure. The operating room (OR) in green, with the motorized C-arm (C), FlexMove-monitor (FM), and control module (CM) in light blue. Two lead glass shields in dark blue shielding the technician operating the control module, as well as the technical room (yellow). Orange striped areas indicate zones close to the edges of the lead shield not fully protected from radiation. In gray, OR table and anesthesiology equipment. Dashed light blue line indicates area of movement for the C-arm.

Patient Radiation Dose

Patient radiation exposure characteristics were extracted from DICOM radiation dose structured reports (RDSR). The RDSR provides detailed examination-specific data (e.g., the x-ray tube kilovoltage, beam size, C-arm angulation, beam filtration, etc.), as well as radiation dose indicators such as dose-area product (DAP) and reference point air kerma (AK), for each individual x-ray exposure (i.e., fluoroscopy or CBCT acquisition).

The examination-specific data contained in RDSR was used to estimate effective patient doses (ED) in accordance with Report 103 by the International Commission on Radiological Protection.[15] The effective doses were calculated by Monte Carlo simulations, using the method outlined by Omar et al.[16] The height, weight, and age of each patient were considered in the selection of an appropriate computational patient model, and a computational model of the OR table was used to account for the attenuation of each x-ray beam projected through the table top.

Statistical Analysis

Descriptive summary statistics are expressed as mean ± standard error of the mean, median [min–max range], or frequency (percentage), as appropriate. Pearson correlation (ρ) was considered to evaluate correlation between two variables. Difference between groups was analyzed with Mann–Whitney U test. Statistical analysis was performed using Matlab (MathWorks, Natick, MA). P < 0.05 was considered statistically significant.