Inertial Measurement Unit-Assisted Implantation of Pedicle Screws in Combination With an Intraoperative 3-Dimensional/2-Dimensional Visualization of the Spine

Inertial Measurement Unit-Assisted Implantation of Pedicle Screws in Combination With an... Abstract BACKGROUND Inertial measurement units (IMUs) are microelectromechanical systems used to track orientation and motion. OBJECTIVE To use instruments mounted with IMUs in combination with a 3- and 2-dimensional (3D/2D) rendering of the computed-tomography scan (CT) to guide implantation of pedicle screws. METHODS Pedicle screws were implanted from T1 to S1 in 2 human cadavers. A software application enabled the surgeon to select the starting points and trajectories on a 3D/2D image of the spine, then locate these starting points on the exposed spine and apply the IMU-mounted instruments to reproduce the trajectories. The position of the screws was evaluated on the postoperative CT scan. RESULTS A total of 72 pedicle screws were implanted. Thirty-seven (77%) of the thoracic screws were within the pedicle (Heary I), 7 (15%) showed a lateral breach of the pedicle, and 4 (8%) violated the anterior or lateral vertebral body (Heary III). In the lumbar spine and S1, 21 screws (88%) were within the pedicle (Gertzbein 0), 2 (8%) screws had a pedicle wall breach < 2 mm (Gertzbein 1), and 1 > 2 to < 4 mm (Gertzbein 2). In the second cadaver, the position was compared to the intraoperatively shown virtual position. The median offset was 3°(mean 3° ± 2°, variance 5, range 0°–9°) in the sagittal plane and 3° (mean 4° ± 3°, variance 9, range 0°–12°) in the axial plane. CONCLUSION IMU-assisted implantation of pedicle screws combined with an intraoperative 3D/2D visualization of the spine enabled the surgeon to precisely implant pedicle screws on the exposed spine. Image-guided surgery, Inertial measurement unit, Navigation, Pedicle screws, Spine ABBREVIATIONS ABBREVIATIONS CT computed-tomography DICOM digital imaging and communications in medicine IMUs inertial measurement units SD standard deviation Inertial measurement units (IMUs) are microelectromechanical systems used to track orientation and motion in space. IMUs combine accelerometers and gyroscopes that measure acceleration and angular rotation; a magnetometer may be used to improve the IMU’s accuracy. The information delivered by IMUs can be used to determine the orientation in space. Currently, their use is most apparent in smartphones or tablets, in which they empower motion-sensitive applications or adapt the screen from portrait to landscape mode depending on how the device is held. Apart from this, IMUs have been used for a long time in air and spacecraft maneuvering and in safety systems of cars (such as airbag deployment or electronic stability controls for the breaks).1 In medicine, IMUs are used to track body kinematics2-4 or enhance precision in total knee arthroplasty5. As previously reported, IMUs could be applied to guide implantation of thoracic, lumbar, and sacral pedicle screws, and also S2 alar iliac screws.6-8 In this laboratory investigation, we combined IMU-based image guidance with a 3- and 2-Dimensional (3D/2D) rendering of the preoperative computed-tomography scan (CT) to implant pedicle screws in the thoracic and lumbosacral spine. Specifically, a custom-made software enabled the surgeon to plan the starting point and trajectory of the pedicle screw on the 3D or 2D Digital Imaging and Communications in Medicine (DICOM) dataset of a computed tomography scan of the spine and intraoperatively reproduce the spatial orientation of the trajectory with wirelessly transmitting IMUs that were mounted onto the pedicle finder and screwdriver (Figures 1 and 2). For every screw, this was done in two steps, first calibrating the IMU along an exposed anatomic edge, second, matching the highlighted starting point on the 3D rendered surface with the starting point on the exposed spine and matching the spatial orientation of the pedicle finder and screwdriver with the target orientation on the computer screen (Figure 3). FIGURE 1. View largeDownload slide A, IMUs were used to track the spatial orientation of the pedicle finder and screwdriver. The IMUs were mounted on the instruments in 3D printed boxes from where they wirelessly communicated with a custom-made planning and guidance application on a notebook computer. B, The guidance part of the presented technology was initiated with a quick alignment of the IMU along a predefined and well reproducible plane such as perpendicular to the supraspinous ligament to “zero” the IMUs (Figure 3). A 3D-printed box with a notch to fit on the supraspinous ligament and holes for inserting the pedicle finder or screwdriver facilitated this process. Reprinted with permission from Jost GF et al.6 ©JNSPG 2016. FIGURE 1. View largeDownload slide A, IMUs were used to track the spatial orientation of the pedicle finder and screwdriver. The IMUs were mounted on the instruments in 3D printed boxes from where they wirelessly communicated with a custom-made planning and guidance application on a notebook computer. B, The guidance part of the presented technology was initiated with a quick alignment of the IMU along a predefined and well reproducible plane such as perpendicular to the supraspinous ligament to “zero” the IMUs (Figure 3). A 3D-printed box with a notch to fit on the supraspinous ligament and holes for inserting the pedicle finder or screwdriver facilitated this process. Reprinted with permission from Jost GF et al.6 ©JNSPG 2016. FIGURE 2. View largeDownload slide Reference orientation and starting point for the pedicle screw are chosen on a 3D/2D image, which is reconstructed from the CT. A, The red, green, and blue dot define the plane for the reference orientation. In this example, the reference orientation is given by a plumb line (green line and dot) through a line (red) that connects the tips of the spinous processes of T3 (red dot) and T4 (blue dot). This corresponds to an orientation along a plane, which is perpendicular to the supraspinous ligament. The yellow dot has been placed at the presumed starting point for the pedicle screw. In B, sagittal and C, axial view, the starting point can be moved with the cursor and the tilt (yellow line) can be adjusted moving the white bars with the cursor. D, In the guidance mode, a grey stick points towards the starting point. On a target, the present orientation of the IMU-mounted instrument is represented as green dot and the target orientation is represented with a yellow dot. FIGURE 2. View largeDownload slide Reference orientation and starting point for the pedicle screw are chosen on a 3D/2D image, which is reconstructed from the CT. A, The red, green, and blue dot define the plane for the reference orientation. In this example, the reference orientation is given by a plumb line (green line and dot) through a line (red) that connects the tips of the spinous processes of T3 (red dot) and T4 (blue dot). This corresponds to an orientation along a plane, which is perpendicular to the supraspinous ligament. The yellow dot has been placed at the presumed starting point for the pedicle screw. In B, sagittal and C, axial view, the starting point can be moved with the cursor and the tilt (yellow line) can be adjusted moving the white bars with the cursor. D, In the guidance mode, a grey stick points towards the starting point. On a target, the present orientation of the IMU-mounted instrument is represented as green dot and the target orientation is represented with a yellow dot. FIGURE 3. View largeDownload slide A, To zero the IMU of the pedicle finder or screwdriver, the IMU-mounted instrument is inserted into the 3D printed box and placed on the supraspinous ligament and held still perpendicular to the supraspinous ligament and aligned with the axis of the underlying spinous processes. A wireless USB footswitch is pressed to “zero” the IMU. © KTB Studios, LLC. 2017. All rights reserved. Used with permission. B, Zeroing the IMU sets the green dot to the zero position. C and D, The IMU-mounted pedicle finder is moved to the entry point for the pedicle screw and oriented in such a way that the green dot (present orientation of the instrument) will overlie the yellow dot (target position). C, © KTB Studios, LLC. 2017. All rights reserved. Used with permission. The pedicle finder is then advanced into the pedicle of the corresponding vertebra. FIGURE 3. View largeDownload slide A, To zero the IMU of the pedicle finder or screwdriver, the IMU-mounted instrument is inserted into the 3D printed box and placed on the supraspinous ligament and held still perpendicular to the supraspinous ligament and aligned with the axis of the underlying spinous processes. A wireless USB footswitch is pressed to “zero” the IMU. © KTB Studios, LLC. 2017. All rights reserved. Used with permission. B, Zeroing the IMU sets the green dot to the zero position. C and D, The IMU-mounted pedicle finder is moved to the entry point for the pedicle screw and oriented in such a way that the green dot (present orientation of the instrument) will overlie the yellow dot (target position). C, © KTB Studios, LLC. 2017. All rights reserved. Used with permission. The pedicle finder is then advanced into the pedicle of the corresponding vertebra. METHODS The regional review board reviewed and approved the study. Two human cadavers, aged 64 (female) and 75 (male), were used. Preoperative fine-cut CT scans were performed and the DICOM dataset was uploaded to the custom 3D/2D software application for image guidance with IMUs. The application was used on a MacBook Air (Apple Inc, Cupertino, California) running Windows (Microsoft, Redmond, Washington). The software being fully functional, but not optimized yet for user friendliness, it was operated by the programmer. Hardware Setup Wireless IMUs (MTw sensor, Xsens Technologies BV, Enschede, the Netherlands) were boxed in a custom-made 3D-printed housing (Figure 1) and attached to the pedicle finder and screwdriver. A large screen (Samsung, Seoul, South Korea) near the operative table showed the surface anatomy of the spine in 3D or 2D cuts, that is, the axial and sagittal plane (Figure 2). Choosing Entry Point, Reference Position, and Trajectory At first, the rendered surface anatomy was looked at to choose the desired level and side and to set the cursor at the starting point for a pedicle screw (Figure 2A). It was then necessary to allocate a reference or “zero” orientation to the software with 3 dots (Figure 2A): by default, this reference orientation was set along the ventrodorsal axis of the associated spinous process and perpendicular to the supraspinous ligament. Then, the display was switched to the 2D view to set the trajectory in the sagittal and axial view (Figures 2B and 2C). Bars to the side of the screen were moveable to adjust the trajectory in the sagittal and axial plane, directing it more cranially or caudally, or into optimally converging alignment. If necessary, the entry point was also optimized in the 2D view with the cursor. The screen was then switched back to display the surface anatomy on which a grey stick pointed at the entry point (Figure 2D). Implanting the Pedicle Screw First, the surgeon decorticated the spine at the spot that corresponded to the position of the yellow dot or grey stick on the computer screen (Figures 2A and 2D). Second, the IMU-mounted pedicle finder (Depuy Synthes, Johnson and Johnson, New Brunswick, New Jersey) was tucked into a little 3D-printed box with a notch to accommodate the supraspinous ligament (Figure 3A), and then positioned on the spine in accordance with the reference orientation in the 3D view, and held still to zero the IMU. As stated above, the reference orientation was by default a line along the ventrodorsal axis of the associated spinous process, and at right angles to the course of the supraspinous ligament. A USB footswitch (Scythe, Tokyo, Japan) was pressed to freeze this position as reference for the IMU. On the computer screen, the orientation of the pedicle finder was shown with a green dot on a target (Figure 3B). Third, the pedicle finder was moved to the decorticated starting point and tilted in the axial and sagittal plane until the green dot overlapped the yellow dot, which corresponded to the previously planned orientation of the pedicle screw (Figures 3C and 3D). After checking the cannulation of the pedicle with a probe, the surgeon switched to the screwdriver and repeated the steps for referencing and IMU-assisted orientation of the screwdriver (Depuy Synthes) before turning in the screw (MATRIX Spine System, Depuy Synthes). The computer screen showed the relative axial and sagittal tilt angles as numbers and as a virtual trajectory of the screw on the sagittal and axial plane in the 2D mode (Figure 4). With the fully set screw in the second spine and prior to releasing the screw from the screwdriver, a screenshot of the virtual trajectory was stored for postoperative analysis. FIGURE 4. View largeDownload slide Virtual orientation of the pedicle finder or screwdriver with implanted screw is shown as a red line in A, sagittal and B, axial plane. The yellow lines correspond to the planned orientation. This orientation is chosen before implanting the screw. The red line is the system's estimation of the spatial orientation of the pedicle finder or screwdriver. The system presumes that the surgeon has correctly matched the entry point and automatically starts the red line at the previously chosen yellow entry point. FIGURE 4. View largeDownload slide Virtual orientation of the pedicle finder or screwdriver with implanted screw is shown as a red line in A, sagittal and B, axial plane. The yellow lines correspond to the planned orientation. This orientation is chosen before implanting the screw. The red line is the system's estimation of the spatial orientation of the pedicle finder or screwdriver. The system presumes that the surgeon has correctly matched the entry point and automatically starts the red line at the previously chosen yellow entry point. Postoperative Analysis A postoperative CT (Siemens, Munich, Germany) was performed on both cadavers and the position and orientation of the screws were examined and checked for cortical breaches using the multiplanar reconstruction mode of the DICOM viewer Osirix (Pixmeo, Geneva, Switzerland). The accuracy of all screws was classified according to Heary9 in the thoracic spine and Gertzbein and Robbins10 in the lumbar spine by a radiologist. For the second cadaver, screenshots of each screw in both the sagittal and axial plane were stored. Semitransparent screenshots of the screws were overlaid with semitransparent screenshots of the virtual trajectory on a PowerPoint file (Microsoft Office, Redmond, Washington). Then the difference between the direction of the virtual trajectories and the screws in the sagittal and axial plane was measured in degrees using PixelStick (Plum Amazing, Princeville, Hawaii). Of note, this procedure of evaluating angular mismatches has not been formally validated, but appeared accurate. In the second cadaver, the congruency between the planned entry points and the center of the implanted screws was qualitatively assessed on the semitransparent screenshots. Analyses were conducted in R (R Development Core Team, 2010). RESULTS Global Assessment of Screw Positions Both cadavers were instrumented with bilateral pedicle screws from T1 to S1, totaling in 72 screws. Of the thoracic screws, 37 (77%) were Heary Grade I (screw entirely within the pedicle and vertebral body [VB]), 7 (15%) screws were Heary Grade II (screw violated lateral pedicle but tip of the screw in the VB), and 4 (8%) screws were Heary Grade III (screw tip penetrated anterior or lateral VB; Figures 5 and 6). In the lumbar spine and S1, 21 screws (88%) were Gertzbein 0 (screw completely within pedicle), 2 (8%) screws were Gertzbein 1 (pedicle wall breach < 2 mm), and 1 screw was Gertzbein 2 (pedicle wall breach > 2 - < 4 mm). FIGURE 5. View largeDownload slide A, In the first cadaver, both T2 pedicle screws were found to course parallel to the upper endplate as intended, but through the costovertebral joint instead of the pedicle. B, The screw on the right was Heary II, on the left was Heary III. C, The right L1 screw of the second cadaver was just caudal and D, laterally off-centered to the virtual trajectory. It was 4° less converging and 1° more caudally oriented than the virtual spatial orientation. FIGURE 5. View largeDownload slide A, In the first cadaver, both T2 pedicle screws were found to course parallel to the upper endplate as intended, but through the costovertebral joint instead of the pedicle. B, The screw on the right was Heary II, on the left was Heary III. C, The right L1 screw of the second cadaver was just caudal and D, laterally off-centered to the virtual trajectory. It was 4° less converging and 1° more caudally oriented than the virtual spatial orientation. FIGURE 6. View largeDownload slide A and B, Sagittal overview of the postoperative CT scan of the second cadaver A, on the left, and B, on the right. Most screws are parallel to the upper endplates and converging as intended. C-E, Axial view of the C, T1 screws, D, T7 screws, and E, S1 screws. FIGURE 6. View largeDownload slide A and B, Sagittal overview of the postoperative CT scan of the second cadaver A, on the left, and B, on the right. Most screws are parallel to the upper endplates and converging as intended. C-E, Axial view of the C, T1 screws, D, T7 screws, and E, S1 screws. Postoperative vs Virtual Intraoperative Trajectory In the second cadaver, we were interested in analyzing how accurately the intraoperative virtual trajectories corresponded to the trajectories of the implanted pedicle screws. In the sagittal plane, the median offset was 3° (mean 3° ± 2° [standard deviation; SD], variance 5, range 0°-9°). Thirty screws (83%) were within an offset range of 0° to 5° and 6 (17%) screws were within an offset range of 6° to 10°. In the axial plane, the median offset was 3° (mean 4° ± 3° [SD], variance 9, range 0°-12°). Twenty-six screws (72%) were within an offset range of 0° to 5°, 9 screws (25%) were within an offset 6° to 10°, and 1 (3%) had 12° offset. For these aforementioned results, we used the absolute angular differences. The boxplot in Figure 7 shows the dispersion around the zero values and illustrates if screws on the postoperative CT had a more or less converging and a more cranially or more caudally oriented trajectory compared to the intraoperative virtual trajectory. In the second cadaver, we also measured proximity of the final screw position to the planned entry points. Sixteen screws (44%) were centered over the planned entry point. In 18 screws (50%), the planned entry point was within the diameter of the screw, hence a maximum of 2.5 mm from the center of the screw in the thoracic spine (diameter of screws 4 or 5 mm) and 2.5 to 3 mm in the lumbar spine (diameter of the screws 5 or 6 mm). Two screws (6%) started just off the planned entry points (Figure 5). Among the uncentered screws, 6 started lateral, 4 caudal, 2 cranial, 4 lateral and caudal and 4 lateral and cranial to the planned entry point. FIGURE 7. View largeDownload slide This boxplot compares the virtual trajectories, which the application showed on the screen after fully implanting the screws, to the trajectories of the pedicle screws in the postoperative CT. Differences are expressed in degrees. In the axial plane, positive values show how many degrees the screws were more converging than the virtual trajectory and negative values show how many degrees the screws were less converging than the virtual trajectory. In the sagittal plane, positive values show how many degrees the screws were aiming more caudally than the virtual trajectory and negative values show how many degrees the screws were aiming more cranially than the virtual trajectory. The bar in the middle is the 50th percentile, the bottom and top of the box are the 25th and 75th percentile. The whiskers mark the minimum and maximum values. FIGURE 7. View largeDownload slide This boxplot compares the virtual trajectories, which the application showed on the screen after fully implanting the screws, to the trajectories of the pedicle screws in the postoperative CT. Differences are expressed in degrees. In the axial plane, positive values show how many degrees the screws were more converging than the virtual trajectory and negative values show how many degrees the screws were less converging than the virtual trajectory. In the sagittal plane, positive values show how many degrees the screws were aiming more caudally than the virtual trajectory and negative values show how many degrees the screws were aiming more cranially than the virtual trajectory. The bar in the middle is the 50th percentile, the bottom and top of the box are the 25th and 75th percentile. The whiskers mark the minimum and maximum values. DISCUSSION This laboratory investigation demonstrates that a relatively simple technological setup with parts from the world of consumer electronics has the potential to assist in the process of implanting pedicle screws in the thoracic and lumbar spine. The retail prices for IMUs start below 100 USD, which may make it a low-cost alternative to much more expensive devices to assist in placement of screws. A further discussion of its potentials and use is warranted. In extrapolating from previous work on freehand pedicle screws,11 we considered Heary I or II or Gertzbein 0 or 1 as accurate screws. Hence, 67 of 72 (93%) screws were considered accurate. This was achieved through a series of measures to guide the workflow of this IMU-assisted freehand technique. First, the 3D rendering of the preoperative CT visualized the surface anatomy of the spine, and was found to be very helpful to select the correct entry point for each screw on the dorsally exposed spine (Figures 2-4): After coarsely setting the starting point on the 3D rendered spine, the surgeon could switch to the 2D mode to check its correct location, shift it if necessary, and then adjust the trajectory in the axial and sagittal plane. This workflow allowed the surgical team to forego the preoperative planning phase as reported earlier.6,7 In the future, the software application could be improved to choose and set all entry points and trajectories beforehand, or to automatize this entire process and have the application suggest entry points, trajectories, and screw lengths. On the other hand, no time delay was experienced by flipping back and forth between the 3D and 2D images of the spine to choose and adjust the entry points and the trajectories in the axial and sagittal plane. For implanting the screw, the surgeon first “zeroed” the IMU-mounted pedicle finder along the supraspinous ligament and then started the trajectory on the exposed spine at the location which corresponded to the position indicated on the computer screen (Figure 3). “Zeroing” was repeated with the IMU-mounted screwdriver, which was then positioned at the starting point to implant the screw. In 94% of the screws of the second cadaver, the postoperative CT showed this starting point to be centered or within the screw diameter. While advancing the pedicle finder or the screw, the correct orientation of the instruments was given, when the dots in the target of the large-scale screen overlapped. Comparison to Other Freehand Techniques In the literature, the term “freehand placement” distinguishes techniques for screw placement from techniques using navigation. Whereas “freehand placement” relies predominantly on recognizing anatomic landmarks, it does not preclude the use of intraoperative imaging or some other guidance technology. Classically, freehand placement is combined with conventional radiography or fluoroscopy to guide and control the placement of pedicle screws. Using anatomic landmarks only for T1 to L5 screws in a cadaver investigation, a cortical breach rate of 29% was found.12 In our own prior cadaveric study, the rate of pedicle breaches in freehand screws without any further (image) guidance was 20%.8 A misplacement rate of only a few percent in a case series is among the highest reported accuracy for this technique.13 With use of lateral fluoroscopy, Amato et al14 reported an accuracy of 98%. Parker et al11 reviewed their experience with 6816 free-hand placed pedicle screws and found a cortical breach rate of more than 25% of the screw diameter in only 1.7%. Whereas they mainly referred to anatomic landmarks to choose the entry points and trajectory for each screw, their technique did also apply further assistance: lateral radiographies were taken prior and after placing the screws, and in some instances also anteroposterior radiographies to check accuracy. Lumbar and sacral screws were usually placed with an electromyographic stimulation probe to detect a medial cortical wall breach. It was however not reported how many cortical breaches were detected and screws corrected intraoperatively using these adjuncts. As opposed to these 2 exemplary studies, we did not apply any intraoperative imaging. But seeing the 3D surface rendering of the preoperative CT and the underlying anatomy of the pedicle and vertebral body on 2D CT planes immediately before placing the screw and then being assisted by the IMU measurements appeared to give the surgeon at least as much if not more confidence level to place the screw. Thus, whereas freehand techniques with some kind of imaging may guarantee excellent accuracy, the proposed technique using IMU guidance only, appears to be similarly accurate. The prototype device is not sterile and has not been tested on patients. However, implementing this method in a clinical scenario harbors potential to minimize harmful exposure to ionizing radiation to the staff in the operating room without compromising on precision;15 after use of the C-arm to localize and confirm the level to be operated on, IMU-assisted implantation of pedicle screws could take over. IMU-Assistance vs Navigation Given the simplicity of this IMU-assisted surgery, some limitations must be considered: (1) Unlike in contemporary 3D navigation, the starting point is found by confirming the vertebra of interest using anatomic landmarks or an intraoperative fluoroscopy, and matching its location on the 3D rendered spine with the dorsally exposed spine of the patient. Failure to exactly match the entry point on the spine with the entry point on the 3D rendering may lead to a malplaced screw. (2) Apart from the orientation, advancement of the instruments and screw into the pedicle and vertebral body is not monitored and has to be controlled with the depth readings on the instruments. (3) As the accuracy of IMU-based surgery relies mainly on the surgeon's ability to zero the IMU prior to placing the screw by holding it exactly along the plane as chosen and visualized on the screen (for instance along the axis of the spinous process and perpendicular to the supraspinous ligament; Figure 2), errors of orientation may occur if the device is not zeroed correctly or if the spine rotates away under the load of the pedicle finder or screwdriver, because the device does not compensate for such secondary movements. In the current experiment, zeroing was a straightforward process that could quickly be repeated in case of doubt, and, although not formally assessed, was found to be well reproducible. Rotation of the spine caused by the loading force of an instrument is a problem inherent to every navigation technique, and, if unavoidable, requires the surgeon to follow along and “override” the IMU measurement or navigation screen, which may temporarily suggest a mismatch. Failure to do so may contribute to the finding that navigated screws may come to lie less convergent than freehand screws.16 3D navigation has been shown to excel fluoroscopy-assisted placement of pedicle screws and at the same time reduce intraoperative radiation exposure to the staff.17,18 It furthermore enables minimally invasive surgeries such as percutaneous placement of screws without the added radiation exposure that a conventional technique would demand. And it can be applied to any region of the spine. But the cost of 3D navigation is still high and there may be fewer devices available per hospital than parallel running spine surgeries. For many open surgeries, anatomic landmarks and the information that an IMU-assisted technique can provide may be sufficient information to confidently place a pedicle screw in the thoracolumbar spine. The mean angular offset between the intraoperative virtual trajectory and the screw on the postoperative CT was 3° ± 2° for the sagittal plane and 4° ± 3° for the axial plane. Given the rather simple setup of IMU-based image guidance, this data compares favorably to O-arm navigation where the mean angular difference between the virtual and actual image in all screws was 2.16° ± 2.24° on sagittal images and 2.17° ± 2.20° on axial images.19 Another accurate technique uses patient-specific templates to implant screws.20 Like IMU-guidance it requires a preoperative CT but intraoperative C-arm is minimized. A disadvantage is that the 3D printing to produce the templates takes time. Hence patient-specific templates are not available for emergency surgeries, but may be used to fix even demanding deformities.21 For such tasks, the performance of IMU-assistance has not yet been tested. Aside from an earlier solution that combined IMUs with a C-arm to percutaneously implant lumbar pedicle screws22 and alternative medical uses of IMUs,2-5 to our knowledge currently no other IMU device exists that guides inserting pedicle screws. Development of the prototype into a clinically applicable product may enable surgeons to do so. CONCLUSION In combination with a 3D rendering of the surface anatomy of the spine, IMU-assisted implantation of thoracic, lumbar, and sacral pedicle screws is feasible without a preoperative planning phase and could be helpful for surgeons aiming to minimize use of the C-arm but with no access to contemporary navigation devices. Disclosures This study was supported by the Gebert Ruef Foundation, Switzerland (GRS-010/13). This grant was designated to G.F.J and was used for hardware and human cadavers. A patent has been filed: Jost, G. Walti, J. Cattin, Ph. (2014). Controlling a surgical intervention to a bone. European Patent No 2901957 A1. The other authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article. Notes Parts of the study were presented as oral presentation at the Global Spine Congress in Dubai, April 13-16, 2016, which was published as Jost GF, Walti J, Mariani L, Schaeren St., Cattin Ph: A novel approach to navigated implantation of thoracic and lumbosacral pedicle screws using inertial measurement units. Global Spine J, 2016. 06 - GO271. REFERENCES 1. Norhafizan A , Ghazilla R , Khairi N , Kasi V . Reviews on various inertial measurement unit (IMU) sensor applications . Int J Signal Process Syst . 2013 ; 1 ( 2 ): 256 – 262 . 2. Cutti AG , Giovanardi A , Rocchi L , Davalli A , Sacchetti R . Ambulatory measurement of shoulder and elbow kinematics through inertial and magnetic sensors . Med Biol Eng Comput . 2008 ; 46 ( 2 ): 169 – 178 . Google Scholar CrossRef Search ADS PubMed 3. Ferrari A , Cutti AG , Garofalo P et al. First in vivo assessment of “Outwalk”: a novel protocol for clinical gait analysis based on inertial and magnetic sensors . Med Biol Eng Comput . 2010 ; 48 ( 1 ): 1 – 15 . Google Scholar CrossRef Search ADS PubMed 4. Giansanti D , Maccioni G , Benvenuti F , Macellari V . Inertial measurement units furnish accurate trunk trajectory reconstruction of the sit-to-stand manoeuvre in healthy subjects . Med Biol Eng Comput . 2007 ; 45 ( 10 ): 969 – 976 . Google Scholar CrossRef Search ADS PubMed 5. Nam D , Weeks KD , Reinhardt KR , Nawabi DH , Cross MB , Mayman DJ . Accelerometer-based, portable navigation vs imageless, large-console computer-assisted navigation in total knee arthroplasty . J Arthroplasty . 2013 ; 28 ( 2 ): 255 – 261 . Google Scholar CrossRef Search ADS PubMed 6. Jost GF , Walti J , Mariani L , Cattin P . A novel approach to navigated implantation of S-2 alar iliac screws using inertial measurement units . J Neurosurg Spine . 2016 ; 24 ( 3 ): 447 – 453 . Google Scholar CrossRef Search ADS PubMed 7. Walti J , Jost GF , Cattin P . A new cost-effective approach to pedicular screw placement . In: Linte CA , Yaniv Z , Fallavolita P , Abolmaesumi P , Holmes III D , eds. 9th International Workshop, AE-CAI 2014 . Boston : Springer International Publishing Switzerland ; 2014 : 90 – 97 . Google Scholar CrossRef Search ADS 8. Jost GF , Walti J , Mariani L , Schaeren S , Cattin P . Inertial measurement unit-assisted implantation of thoracic, lumbar, and sacral pedicle screws improves precision of a freehand technique . World Neurosurg . 2017 ; 103 : 11 – 18 . Google Scholar CrossRef Search ADS PubMed 9. Heary RF , Bono CM , Black M . Thoracic pedicle screws: postoperative computerized tomography scanning assessment . J Neurosurg . 2004 ; 100 ( 4 Suppl Spine ): 325 – 331 . Google Scholar PubMed 10. Gertzbein SD , Robbins SE . Accuracy of pedicular screw placement in vivo . Spine . 1990 ; 15 ( 1 ): 11 – 14 . Google Scholar CrossRef Search ADS PubMed 11. Parker SL , McGirt MJ , Farber SH et al. Accuracy of free-hand pedicle screws in the thoracic and lumbar spine: analysis of 6816 consecutive screws . Neurosurgery . 2011 ; 68 ( 1 ): 170 – 178 ; discussion 178 . Google Scholar CrossRef Search ADS PubMed 12. Alhabib H , Nataraj A , Khashab M , Mahood J , Kortbeek F , Fox R . Pedicle screw insertion in the thoracolumbar spine: comparison of 4 guidance techniques in the intact cadaveric spine . J Neurosurg Spine . 2011 ; 14 ( 5 ): 664 – 669 . Google Scholar CrossRef Search ADS PubMed 13. Kotil K , Bilge T . Accuracy of pedicle and mass screw placement in the spine without using fluoroscopy: a prospective clinical study . Spine J . 2008 ; 8 ( 4 ): 591 – 596 . Google Scholar CrossRef Search ADS PubMed 14. Amato V , Giannachi L , Irace C , Corona C . Accuracy of pedicle screw placement in the lumbosacral spine using conventional technique: computed tomography postoperative assessment in 102 consecutive patients . J Neurosurg Spine . 2010 ; 12 ( 3 ): 306 – 313 . Google Scholar CrossRef Search ADS PubMed 15. Bamshad Azizi K , Ozgur G , Wilson E et al. Inertial measurement unit for radiation-free navigated screw placement in slipped capital femoral epiphysis surgery . Medical Image Computing and Computer-Assisted Intervention–MICCAI . Munich : Springer ; 2015 . 16. Gelalis ID , Paschos NK , Pakos EE et al. Accuracy of pedicle screw placement: a systematic review of prospective in vivo studies comparing free hand, fluoroscopy guidance and navigation techniques . Eur Spine J . 2012 ; 21 ( 2 ): 247 – 255 . Google Scholar CrossRef Search ADS PubMed 17. Villard J , Ryang YM , Demetriades AK et al. Radiation exposure to the surgeon and the patient during posterior lumbar spinal instrumentation . Spine . 2014 ; 39 ( 13 ): 1004 – 1009 . Google Scholar CrossRef Search ADS PubMed 18. Shin BJ , James AR , Njoku IU , Hartl R . Pedicle screw navigation: a systematic review and meta-analysis of perforation risk for computer-navigated versus freehand insertion . J Neurosurg Spine . 2012 ; 17 ( 2 ): 113 – 122 . Google Scholar CrossRef Search ADS PubMed 19. Miller CA , Ledonio CG , Hunt MA , Siddiq F , Polly DW Jr. Reliability of the planned pedicle screw trajectory versus the actual pedicle screw trajectory using intra-operative 3D CT and image guidance . Int J Spine Surg . 2016 ; 10 : 1 – 13 . Google Scholar CrossRef Search ADS PubMed 20. Hu Y , Yuan ZS , Spiker WR et al. A comparative study on the accuracy of pedicle screw placement assisted by personalized rapid prototyping template between pre- and post-operation in patients with relatively normal mid-upper thoracic spine . Eur Spine J . 2016 ; 25 ( 6 ): 1706 – 1715 . Google Scholar CrossRef Search ADS PubMed 21. Putzier M , Strube P , Cecchinato R , Lamartina C , Hoff EK . A new navigational tool for pedicle screw placement in patients with severe scoliosis . Clin Spine Surg . 2017 ; 30 ( 4 ): E430 – E439 . Google Scholar CrossRef Search ADS PubMed 22. Idler C , Rolfe KW , Gorek JE . Accuracy of percutaneous lumbar pedicle screw placement using the oblique or “owl's-eye” view and novel guidance technology . J Neurosurg Spine . 2010 ; 13 ( 4 ): 509 – 515 . Google Scholar CrossRef Search ADS PubMed Acknowledgments The authors wish to thank Magdalena Müller-Gerbl, Sandra Blache, Roger Kurz, and Peter Zimmermann from the Institute of Anatomy, University of Basel, for their kind collaboration, Selina Ackermann from the University Hospital Basel, for editorial assistance, Kate Galloway from KTB studios for the artwork, and Daniel Zumofen from the Departments of Radiology and Neurosurgery, University Hospital Basel for classifying screw positions. Copyright © 2018 by the Congress of Neurological Surgeons This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Operative Neurosurgery Oxford University Press

Inertial Measurement Unit-Assisted Implantation of Pedicle Screws in Combination With an Intraoperative 3-Dimensional/2-Dimensional Visualization of the Spine

Loading next page...
 
/lp/ou_press/inertial-measurement-unit-assisted-implantation-of-pedicle-screws-in-vx0lbZs6KR
Publisher
Oxford University Press
Copyright
Copyright © 2018 by the Congress of Neurological Surgeons
ISSN
2332-4252
eISSN
2332-4260
D.O.I.
10.1093/ons/opy141
Publisher site
See Article on Publisher Site

Abstract

Abstract BACKGROUND Inertial measurement units (IMUs) are microelectromechanical systems used to track orientation and motion. OBJECTIVE To use instruments mounted with IMUs in combination with a 3- and 2-dimensional (3D/2D) rendering of the computed-tomography scan (CT) to guide implantation of pedicle screws. METHODS Pedicle screws were implanted from T1 to S1 in 2 human cadavers. A software application enabled the surgeon to select the starting points and trajectories on a 3D/2D image of the spine, then locate these starting points on the exposed spine and apply the IMU-mounted instruments to reproduce the trajectories. The position of the screws was evaluated on the postoperative CT scan. RESULTS A total of 72 pedicle screws were implanted. Thirty-seven (77%) of the thoracic screws were within the pedicle (Heary I), 7 (15%) showed a lateral breach of the pedicle, and 4 (8%) violated the anterior or lateral vertebral body (Heary III). In the lumbar spine and S1, 21 screws (88%) were within the pedicle (Gertzbein 0), 2 (8%) screws had a pedicle wall breach < 2 mm (Gertzbein 1), and 1 > 2 to < 4 mm (Gertzbein 2). In the second cadaver, the position was compared to the intraoperatively shown virtual position. The median offset was 3°(mean 3° ± 2°, variance 5, range 0°–9°) in the sagittal plane and 3° (mean 4° ± 3°, variance 9, range 0°–12°) in the axial plane. CONCLUSION IMU-assisted implantation of pedicle screws combined with an intraoperative 3D/2D visualization of the spine enabled the surgeon to precisely implant pedicle screws on the exposed spine. Image-guided surgery, Inertial measurement unit, Navigation, Pedicle screws, Spine ABBREVIATIONS ABBREVIATIONS CT computed-tomography DICOM digital imaging and communications in medicine IMUs inertial measurement units SD standard deviation Inertial measurement units (IMUs) are microelectromechanical systems used to track orientation and motion in space. IMUs combine accelerometers and gyroscopes that measure acceleration and angular rotation; a magnetometer may be used to improve the IMU’s accuracy. The information delivered by IMUs can be used to determine the orientation in space. Currently, their use is most apparent in smartphones or tablets, in which they empower motion-sensitive applications or adapt the screen from portrait to landscape mode depending on how the device is held. Apart from this, IMUs have been used for a long time in air and spacecraft maneuvering and in safety systems of cars (such as airbag deployment or electronic stability controls for the breaks).1 In medicine, IMUs are used to track body kinematics2-4 or enhance precision in total knee arthroplasty5. As previously reported, IMUs could be applied to guide implantation of thoracic, lumbar, and sacral pedicle screws, and also S2 alar iliac screws.6-8 In this laboratory investigation, we combined IMU-based image guidance with a 3- and 2-Dimensional (3D/2D) rendering of the preoperative computed-tomography scan (CT) to implant pedicle screws in the thoracic and lumbosacral spine. Specifically, a custom-made software enabled the surgeon to plan the starting point and trajectory of the pedicle screw on the 3D or 2D Digital Imaging and Communications in Medicine (DICOM) dataset of a computed tomography scan of the spine and intraoperatively reproduce the spatial orientation of the trajectory with wirelessly transmitting IMUs that were mounted onto the pedicle finder and screwdriver (Figures 1 and 2). For every screw, this was done in two steps, first calibrating the IMU along an exposed anatomic edge, second, matching the highlighted starting point on the 3D rendered surface with the starting point on the exposed spine and matching the spatial orientation of the pedicle finder and screwdriver with the target orientation on the computer screen (Figure 3). FIGURE 1. View largeDownload slide A, IMUs were used to track the spatial orientation of the pedicle finder and screwdriver. The IMUs were mounted on the instruments in 3D printed boxes from where they wirelessly communicated with a custom-made planning and guidance application on a notebook computer. B, The guidance part of the presented technology was initiated with a quick alignment of the IMU along a predefined and well reproducible plane such as perpendicular to the supraspinous ligament to “zero” the IMUs (Figure 3). A 3D-printed box with a notch to fit on the supraspinous ligament and holes for inserting the pedicle finder or screwdriver facilitated this process. Reprinted with permission from Jost GF et al.6 ©JNSPG 2016. FIGURE 1. View largeDownload slide A, IMUs were used to track the spatial orientation of the pedicle finder and screwdriver. The IMUs were mounted on the instruments in 3D printed boxes from where they wirelessly communicated with a custom-made planning and guidance application on a notebook computer. B, The guidance part of the presented technology was initiated with a quick alignment of the IMU along a predefined and well reproducible plane such as perpendicular to the supraspinous ligament to “zero” the IMUs (Figure 3). A 3D-printed box with a notch to fit on the supraspinous ligament and holes for inserting the pedicle finder or screwdriver facilitated this process. Reprinted with permission from Jost GF et al.6 ©JNSPG 2016. FIGURE 2. View largeDownload slide Reference orientation and starting point for the pedicle screw are chosen on a 3D/2D image, which is reconstructed from the CT. A, The red, green, and blue dot define the plane for the reference orientation. In this example, the reference orientation is given by a plumb line (green line and dot) through a line (red) that connects the tips of the spinous processes of T3 (red dot) and T4 (blue dot). This corresponds to an orientation along a plane, which is perpendicular to the supraspinous ligament. The yellow dot has been placed at the presumed starting point for the pedicle screw. In B, sagittal and C, axial view, the starting point can be moved with the cursor and the tilt (yellow line) can be adjusted moving the white bars with the cursor. D, In the guidance mode, a grey stick points towards the starting point. On a target, the present orientation of the IMU-mounted instrument is represented as green dot and the target orientation is represented with a yellow dot. FIGURE 2. View largeDownload slide Reference orientation and starting point for the pedicle screw are chosen on a 3D/2D image, which is reconstructed from the CT. A, The red, green, and blue dot define the plane for the reference orientation. In this example, the reference orientation is given by a plumb line (green line and dot) through a line (red) that connects the tips of the spinous processes of T3 (red dot) and T4 (blue dot). This corresponds to an orientation along a plane, which is perpendicular to the supraspinous ligament. The yellow dot has been placed at the presumed starting point for the pedicle screw. In B, sagittal and C, axial view, the starting point can be moved with the cursor and the tilt (yellow line) can be adjusted moving the white bars with the cursor. D, In the guidance mode, a grey stick points towards the starting point. On a target, the present orientation of the IMU-mounted instrument is represented as green dot and the target orientation is represented with a yellow dot. FIGURE 3. View largeDownload slide A, To zero the IMU of the pedicle finder or screwdriver, the IMU-mounted instrument is inserted into the 3D printed box and placed on the supraspinous ligament and held still perpendicular to the supraspinous ligament and aligned with the axis of the underlying spinous processes. A wireless USB footswitch is pressed to “zero” the IMU. © KTB Studios, LLC. 2017. All rights reserved. Used with permission. B, Zeroing the IMU sets the green dot to the zero position. C and D, The IMU-mounted pedicle finder is moved to the entry point for the pedicle screw and oriented in such a way that the green dot (present orientation of the instrument) will overlie the yellow dot (target position). C, © KTB Studios, LLC. 2017. All rights reserved. Used with permission. The pedicle finder is then advanced into the pedicle of the corresponding vertebra. FIGURE 3. View largeDownload slide A, To zero the IMU of the pedicle finder or screwdriver, the IMU-mounted instrument is inserted into the 3D printed box and placed on the supraspinous ligament and held still perpendicular to the supraspinous ligament and aligned with the axis of the underlying spinous processes. A wireless USB footswitch is pressed to “zero” the IMU. © KTB Studios, LLC. 2017. All rights reserved. Used with permission. B, Zeroing the IMU sets the green dot to the zero position. C and D, The IMU-mounted pedicle finder is moved to the entry point for the pedicle screw and oriented in such a way that the green dot (present orientation of the instrument) will overlie the yellow dot (target position). C, © KTB Studios, LLC. 2017. All rights reserved. Used with permission. The pedicle finder is then advanced into the pedicle of the corresponding vertebra. METHODS The regional review board reviewed and approved the study. Two human cadavers, aged 64 (female) and 75 (male), were used. Preoperative fine-cut CT scans were performed and the DICOM dataset was uploaded to the custom 3D/2D software application for image guidance with IMUs. The application was used on a MacBook Air (Apple Inc, Cupertino, California) running Windows (Microsoft, Redmond, Washington). The software being fully functional, but not optimized yet for user friendliness, it was operated by the programmer. Hardware Setup Wireless IMUs (MTw sensor, Xsens Technologies BV, Enschede, the Netherlands) were boxed in a custom-made 3D-printed housing (Figure 1) and attached to the pedicle finder and screwdriver. A large screen (Samsung, Seoul, South Korea) near the operative table showed the surface anatomy of the spine in 3D or 2D cuts, that is, the axial and sagittal plane (Figure 2). Choosing Entry Point, Reference Position, and Trajectory At first, the rendered surface anatomy was looked at to choose the desired level and side and to set the cursor at the starting point for a pedicle screw (Figure 2A). It was then necessary to allocate a reference or “zero” orientation to the software with 3 dots (Figure 2A): by default, this reference orientation was set along the ventrodorsal axis of the associated spinous process and perpendicular to the supraspinous ligament. Then, the display was switched to the 2D view to set the trajectory in the sagittal and axial view (Figures 2B and 2C). Bars to the side of the screen were moveable to adjust the trajectory in the sagittal and axial plane, directing it more cranially or caudally, or into optimally converging alignment. If necessary, the entry point was also optimized in the 2D view with the cursor. The screen was then switched back to display the surface anatomy on which a grey stick pointed at the entry point (Figure 2D). Implanting the Pedicle Screw First, the surgeon decorticated the spine at the spot that corresponded to the position of the yellow dot or grey stick on the computer screen (Figures 2A and 2D). Second, the IMU-mounted pedicle finder (Depuy Synthes, Johnson and Johnson, New Brunswick, New Jersey) was tucked into a little 3D-printed box with a notch to accommodate the supraspinous ligament (Figure 3A), and then positioned on the spine in accordance with the reference orientation in the 3D view, and held still to zero the IMU. As stated above, the reference orientation was by default a line along the ventrodorsal axis of the associated spinous process, and at right angles to the course of the supraspinous ligament. A USB footswitch (Scythe, Tokyo, Japan) was pressed to freeze this position as reference for the IMU. On the computer screen, the orientation of the pedicle finder was shown with a green dot on a target (Figure 3B). Third, the pedicle finder was moved to the decorticated starting point and tilted in the axial and sagittal plane until the green dot overlapped the yellow dot, which corresponded to the previously planned orientation of the pedicle screw (Figures 3C and 3D). After checking the cannulation of the pedicle with a probe, the surgeon switched to the screwdriver and repeated the steps for referencing and IMU-assisted orientation of the screwdriver (Depuy Synthes) before turning in the screw (MATRIX Spine System, Depuy Synthes). The computer screen showed the relative axial and sagittal tilt angles as numbers and as a virtual trajectory of the screw on the sagittal and axial plane in the 2D mode (Figure 4). With the fully set screw in the second spine and prior to releasing the screw from the screwdriver, a screenshot of the virtual trajectory was stored for postoperative analysis. FIGURE 4. View largeDownload slide Virtual orientation of the pedicle finder or screwdriver with implanted screw is shown as a red line in A, sagittal and B, axial plane. The yellow lines correspond to the planned orientation. This orientation is chosen before implanting the screw. The red line is the system's estimation of the spatial orientation of the pedicle finder or screwdriver. The system presumes that the surgeon has correctly matched the entry point and automatically starts the red line at the previously chosen yellow entry point. FIGURE 4. View largeDownload slide Virtual orientation of the pedicle finder or screwdriver with implanted screw is shown as a red line in A, sagittal and B, axial plane. The yellow lines correspond to the planned orientation. This orientation is chosen before implanting the screw. The red line is the system's estimation of the spatial orientation of the pedicle finder or screwdriver. The system presumes that the surgeon has correctly matched the entry point and automatically starts the red line at the previously chosen yellow entry point. Postoperative Analysis A postoperative CT (Siemens, Munich, Germany) was performed on both cadavers and the position and orientation of the screws were examined and checked for cortical breaches using the multiplanar reconstruction mode of the DICOM viewer Osirix (Pixmeo, Geneva, Switzerland). The accuracy of all screws was classified according to Heary9 in the thoracic spine and Gertzbein and Robbins10 in the lumbar spine by a radiologist. For the second cadaver, screenshots of each screw in both the sagittal and axial plane were stored. Semitransparent screenshots of the screws were overlaid with semitransparent screenshots of the virtual trajectory on a PowerPoint file (Microsoft Office, Redmond, Washington). Then the difference between the direction of the virtual trajectories and the screws in the sagittal and axial plane was measured in degrees using PixelStick (Plum Amazing, Princeville, Hawaii). Of note, this procedure of evaluating angular mismatches has not been formally validated, but appeared accurate. In the second cadaver, the congruency between the planned entry points and the center of the implanted screws was qualitatively assessed on the semitransparent screenshots. Analyses were conducted in R (R Development Core Team, 2010). RESULTS Global Assessment of Screw Positions Both cadavers were instrumented with bilateral pedicle screws from T1 to S1, totaling in 72 screws. Of the thoracic screws, 37 (77%) were Heary Grade I (screw entirely within the pedicle and vertebral body [VB]), 7 (15%) screws were Heary Grade II (screw violated lateral pedicle but tip of the screw in the VB), and 4 (8%) screws were Heary Grade III (screw tip penetrated anterior or lateral VB; Figures 5 and 6). In the lumbar spine and S1, 21 screws (88%) were Gertzbein 0 (screw completely within pedicle), 2 (8%) screws were Gertzbein 1 (pedicle wall breach < 2 mm), and 1 screw was Gertzbein 2 (pedicle wall breach > 2 - < 4 mm). FIGURE 5. View largeDownload slide A, In the first cadaver, both T2 pedicle screws were found to course parallel to the upper endplate as intended, but through the costovertebral joint instead of the pedicle. B, The screw on the right was Heary II, on the left was Heary III. C, The right L1 screw of the second cadaver was just caudal and D, laterally off-centered to the virtual trajectory. It was 4° less converging and 1° more caudally oriented than the virtual spatial orientation. FIGURE 5. View largeDownload slide A, In the first cadaver, both T2 pedicle screws were found to course parallel to the upper endplate as intended, but through the costovertebral joint instead of the pedicle. B, The screw on the right was Heary II, on the left was Heary III. C, The right L1 screw of the second cadaver was just caudal and D, laterally off-centered to the virtual trajectory. It was 4° less converging and 1° more caudally oriented than the virtual spatial orientation. FIGURE 6. View largeDownload slide A and B, Sagittal overview of the postoperative CT scan of the second cadaver A, on the left, and B, on the right. Most screws are parallel to the upper endplates and converging as intended. C-E, Axial view of the C, T1 screws, D, T7 screws, and E, S1 screws. FIGURE 6. View largeDownload slide A and B, Sagittal overview of the postoperative CT scan of the second cadaver A, on the left, and B, on the right. Most screws are parallel to the upper endplates and converging as intended. C-E, Axial view of the C, T1 screws, D, T7 screws, and E, S1 screws. Postoperative vs Virtual Intraoperative Trajectory In the second cadaver, we were interested in analyzing how accurately the intraoperative virtual trajectories corresponded to the trajectories of the implanted pedicle screws. In the sagittal plane, the median offset was 3° (mean 3° ± 2° [standard deviation; SD], variance 5, range 0°-9°). Thirty screws (83%) were within an offset range of 0° to 5° and 6 (17%) screws were within an offset range of 6° to 10°. In the axial plane, the median offset was 3° (mean 4° ± 3° [SD], variance 9, range 0°-12°). Twenty-six screws (72%) were within an offset range of 0° to 5°, 9 screws (25%) were within an offset 6° to 10°, and 1 (3%) had 12° offset. For these aforementioned results, we used the absolute angular differences. The boxplot in Figure 7 shows the dispersion around the zero values and illustrates if screws on the postoperative CT had a more or less converging and a more cranially or more caudally oriented trajectory compared to the intraoperative virtual trajectory. In the second cadaver, we also measured proximity of the final screw position to the planned entry points. Sixteen screws (44%) were centered over the planned entry point. In 18 screws (50%), the planned entry point was within the diameter of the screw, hence a maximum of 2.5 mm from the center of the screw in the thoracic spine (diameter of screws 4 or 5 mm) and 2.5 to 3 mm in the lumbar spine (diameter of the screws 5 or 6 mm). Two screws (6%) started just off the planned entry points (Figure 5). Among the uncentered screws, 6 started lateral, 4 caudal, 2 cranial, 4 lateral and caudal and 4 lateral and cranial to the planned entry point. FIGURE 7. View largeDownload slide This boxplot compares the virtual trajectories, which the application showed on the screen after fully implanting the screws, to the trajectories of the pedicle screws in the postoperative CT. Differences are expressed in degrees. In the axial plane, positive values show how many degrees the screws were more converging than the virtual trajectory and negative values show how many degrees the screws were less converging than the virtual trajectory. In the sagittal plane, positive values show how many degrees the screws were aiming more caudally than the virtual trajectory and negative values show how many degrees the screws were aiming more cranially than the virtual trajectory. The bar in the middle is the 50th percentile, the bottom and top of the box are the 25th and 75th percentile. The whiskers mark the minimum and maximum values. FIGURE 7. View largeDownload slide This boxplot compares the virtual trajectories, which the application showed on the screen after fully implanting the screws, to the trajectories of the pedicle screws in the postoperative CT. Differences are expressed in degrees. In the axial plane, positive values show how many degrees the screws were more converging than the virtual trajectory and negative values show how many degrees the screws were less converging than the virtual trajectory. In the sagittal plane, positive values show how many degrees the screws were aiming more caudally than the virtual trajectory and negative values show how many degrees the screws were aiming more cranially than the virtual trajectory. The bar in the middle is the 50th percentile, the bottom and top of the box are the 25th and 75th percentile. The whiskers mark the minimum and maximum values. DISCUSSION This laboratory investigation demonstrates that a relatively simple technological setup with parts from the world of consumer electronics has the potential to assist in the process of implanting pedicle screws in the thoracic and lumbar spine. The retail prices for IMUs start below 100 USD, which may make it a low-cost alternative to much more expensive devices to assist in placement of screws. A further discussion of its potentials and use is warranted. In extrapolating from previous work on freehand pedicle screws,11 we considered Heary I or II or Gertzbein 0 or 1 as accurate screws. Hence, 67 of 72 (93%) screws were considered accurate. This was achieved through a series of measures to guide the workflow of this IMU-assisted freehand technique. First, the 3D rendering of the preoperative CT visualized the surface anatomy of the spine, and was found to be very helpful to select the correct entry point for each screw on the dorsally exposed spine (Figures 2-4): After coarsely setting the starting point on the 3D rendered spine, the surgeon could switch to the 2D mode to check its correct location, shift it if necessary, and then adjust the trajectory in the axial and sagittal plane. This workflow allowed the surgical team to forego the preoperative planning phase as reported earlier.6,7 In the future, the software application could be improved to choose and set all entry points and trajectories beforehand, or to automatize this entire process and have the application suggest entry points, trajectories, and screw lengths. On the other hand, no time delay was experienced by flipping back and forth between the 3D and 2D images of the spine to choose and adjust the entry points and the trajectories in the axial and sagittal plane. For implanting the screw, the surgeon first “zeroed” the IMU-mounted pedicle finder along the supraspinous ligament and then started the trajectory on the exposed spine at the location which corresponded to the position indicated on the computer screen (Figure 3). “Zeroing” was repeated with the IMU-mounted screwdriver, which was then positioned at the starting point to implant the screw. In 94% of the screws of the second cadaver, the postoperative CT showed this starting point to be centered or within the screw diameter. While advancing the pedicle finder or the screw, the correct orientation of the instruments was given, when the dots in the target of the large-scale screen overlapped. Comparison to Other Freehand Techniques In the literature, the term “freehand placement” distinguishes techniques for screw placement from techniques using navigation. Whereas “freehand placement” relies predominantly on recognizing anatomic landmarks, it does not preclude the use of intraoperative imaging or some other guidance technology. Classically, freehand placement is combined with conventional radiography or fluoroscopy to guide and control the placement of pedicle screws. Using anatomic landmarks only for T1 to L5 screws in a cadaver investigation, a cortical breach rate of 29% was found.12 In our own prior cadaveric study, the rate of pedicle breaches in freehand screws without any further (image) guidance was 20%.8 A misplacement rate of only a few percent in a case series is among the highest reported accuracy for this technique.13 With use of lateral fluoroscopy, Amato et al14 reported an accuracy of 98%. Parker et al11 reviewed their experience with 6816 free-hand placed pedicle screws and found a cortical breach rate of more than 25% of the screw diameter in only 1.7%. Whereas they mainly referred to anatomic landmarks to choose the entry points and trajectory for each screw, their technique did also apply further assistance: lateral radiographies were taken prior and after placing the screws, and in some instances also anteroposterior radiographies to check accuracy. Lumbar and sacral screws were usually placed with an electromyographic stimulation probe to detect a medial cortical wall breach. It was however not reported how many cortical breaches were detected and screws corrected intraoperatively using these adjuncts. As opposed to these 2 exemplary studies, we did not apply any intraoperative imaging. But seeing the 3D surface rendering of the preoperative CT and the underlying anatomy of the pedicle and vertebral body on 2D CT planes immediately before placing the screw and then being assisted by the IMU measurements appeared to give the surgeon at least as much if not more confidence level to place the screw. Thus, whereas freehand techniques with some kind of imaging may guarantee excellent accuracy, the proposed technique using IMU guidance only, appears to be similarly accurate. The prototype device is not sterile and has not been tested on patients. However, implementing this method in a clinical scenario harbors potential to minimize harmful exposure to ionizing radiation to the staff in the operating room without compromising on precision;15 after use of the C-arm to localize and confirm the level to be operated on, IMU-assisted implantation of pedicle screws could take over. IMU-Assistance vs Navigation Given the simplicity of this IMU-assisted surgery, some limitations must be considered: (1) Unlike in contemporary 3D navigation, the starting point is found by confirming the vertebra of interest using anatomic landmarks or an intraoperative fluoroscopy, and matching its location on the 3D rendered spine with the dorsally exposed spine of the patient. Failure to exactly match the entry point on the spine with the entry point on the 3D rendering may lead to a malplaced screw. (2) Apart from the orientation, advancement of the instruments and screw into the pedicle and vertebral body is not monitored and has to be controlled with the depth readings on the instruments. (3) As the accuracy of IMU-based surgery relies mainly on the surgeon's ability to zero the IMU prior to placing the screw by holding it exactly along the plane as chosen and visualized on the screen (for instance along the axis of the spinous process and perpendicular to the supraspinous ligament; Figure 2), errors of orientation may occur if the device is not zeroed correctly or if the spine rotates away under the load of the pedicle finder or screwdriver, because the device does not compensate for such secondary movements. In the current experiment, zeroing was a straightforward process that could quickly be repeated in case of doubt, and, although not formally assessed, was found to be well reproducible. Rotation of the spine caused by the loading force of an instrument is a problem inherent to every navigation technique, and, if unavoidable, requires the surgeon to follow along and “override” the IMU measurement or navigation screen, which may temporarily suggest a mismatch. Failure to do so may contribute to the finding that navigated screws may come to lie less convergent than freehand screws.16 3D navigation has been shown to excel fluoroscopy-assisted placement of pedicle screws and at the same time reduce intraoperative radiation exposure to the staff.17,18 It furthermore enables minimally invasive surgeries such as percutaneous placement of screws without the added radiation exposure that a conventional technique would demand. And it can be applied to any region of the spine. But the cost of 3D navigation is still high and there may be fewer devices available per hospital than parallel running spine surgeries. For many open surgeries, anatomic landmarks and the information that an IMU-assisted technique can provide may be sufficient information to confidently place a pedicle screw in the thoracolumbar spine. The mean angular offset between the intraoperative virtual trajectory and the screw on the postoperative CT was 3° ± 2° for the sagittal plane and 4° ± 3° for the axial plane. Given the rather simple setup of IMU-based image guidance, this data compares favorably to O-arm navigation where the mean angular difference between the virtual and actual image in all screws was 2.16° ± 2.24° on sagittal images and 2.17° ± 2.20° on axial images.19 Another accurate technique uses patient-specific templates to implant screws.20 Like IMU-guidance it requires a preoperative CT but intraoperative C-arm is minimized. A disadvantage is that the 3D printing to produce the templates takes time. Hence patient-specific templates are not available for emergency surgeries, but may be used to fix even demanding deformities.21 For such tasks, the performance of IMU-assistance has not yet been tested. Aside from an earlier solution that combined IMUs with a C-arm to percutaneously implant lumbar pedicle screws22 and alternative medical uses of IMUs,2-5 to our knowledge currently no other IMU device exists that guides inserting pedicle screws. Development of the prototype into a clinically applicable product may enable surgeons to do so. CONCLUSION In combination with a 3D rendering of the surface anatomy of the spine, IMU-assisted implantation of thoracic, lumbar, and sacral pedicle screws is feasible without a preoperative planning phase and could be helpful for surgeons aiming to minimize use of the C-arm but with no access to contemporary navigation devices. Disclosures This study was supported by the Gebert Ruef Foundation, Switzerland (GRS-010/13). This grant was designated to G.F.J and was used for hardware and human cadavers. A patent has been filed: Jost, G. Walti, J. Cattin, Ph. (2014). Controlling a surgical intervention to a bone. European Patent No 2901957 A1. The other authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article. Notes Parts of the study were presented as oral presentation at the Global Spine Congress in Dubai, April 13-16, 2016, which was published as Jost GF, Walti J, Mariani L, Schaeren St., Cattin Ph: A novel approach to navigated implantation of thoracic and lumbosacral pedicle screws using inertial measurement units. Global Spine J, 2016. 06 - GO271. REFERENCES 1. Norhafizan A , Ghazilla R , Khairi N , Kasi V . Reviews on various inertial measurement unit (IMU) sensor applications . Int J Signal Process Syst . 2013 ; 1 ( 2 ): 256 – 262 . 2. Cutti AG , Giovanardi A , Rocchi L , Davalli A , Sacchetti R . Ambulatory measurement of shoulder and elbow kinematics through inertial and magnetic sensors . Med Biol Eng Comput . 2008 ; 46 ( 2 ): 169 – 178 . Google Scholar CrossRef Search ADS PubMed 3. Ferrari A , Cutti AG , Garofalo P et al. First in vivo assessment of “Outwalk”: a novel protocol for clinical gait analysis based on inertial and magnetic sensors . Med Biol Eng Comput . 2010 ; 48 ( 1 ): 1 – 15 . Google Scholar CrossRef Search ADS PubMed 4. Giansanti D , Maccioni G , Benvenuti F , Macellari V . Inertial measurement units furnish accurate trunk trajectory reconstruction of the sit-to-stand manoeuvre in healthy subjects . Med Biol Eng Comput . 2007 ; 45 ( 10 ): 969 – 976 . Google Scholar CrossRef Search ADS PubMed 5. Nam D , Weeks KD , Reinhardt KR , Nawabi DH , Cross MB , Mayman DJ . Accelerometer-based, portable navigation vs imageless, large-console computer-assisted navigation in total knee arthroplasty . J Arthroplasty . 2013 ; 28 ( 2 ): 255 – 261 . Google Scholar CrossRef Search ADS PubMed 6. Jost GF , Walti J , Mariani L , Cattin P . A novel approach to navigated implantation of S-2 alar iliac screws using inertial measurement units . J Neurosurg Spine . 2016 ; 24 ( 3 ): 447 – 453 . Google Scholar CrossRef Search ADS PubMed 7. Walti J , Jost GF , Cattin P . A new cost-effective approach to pedicular screw placement . In: Linte CA , Yaniv Z , Fallavolita P , Abolmaesumi P , Holmes III D , eds. 9th International Workshop, AE-CAI 2014 . Boston : Springer International Publishing Switzerland ; 2014 : 90 – 97 . Google Scholar CrossRef Search ADS 8. Jost GF , Walti J , Mariani L , Schaeren S , Cattin P . Inertial measurement unit-assisted implantation of thoracic, lumbar, and sacral pedicle screws improves precision of a freehand technique . World Neurosurg . 2017 ; 103 : 11 – 18 . Google Scholar CrossRef Search ADS PubMed 9. Heary RF , Bono CM , Black M . Thoracic pedicle screws: postoperative computerized tomography scanning assessment . J Neurosurg . 2004 ; 100 ( 4 Suppl Spine ): 325 – 331 . Google Scholar PubMed 10. Gertzbein SD , Robbins SE . Accuracy of pedicular screw placement in vivo . Spine . 1990 ; 15 ( 1 ): 11 – 14 . Google Scholar CrossRef Search ADS PubMed 11. Parker SL , McGirt MJ , Farber SH et al. Accuracy of free-hand pedicle screws in the thoracic and lumbar spine: analysis of 6816 consecutive screws . Neurosurgery . 2011 ; 68 ( 1 ): 170 – 178 ; discussion 178 . Google Scholar CrossRef Search ADS PubMed 12. Alhabib H , Nataraj A , Khashab M , Mahood J , Kortbeek F , Fox R . Pedicle screw insertion in the thoracolumbar spine: comparison of 4 guidance techniques in the intact cadaveric spine . J Neurosurg Spine . 2011 ; 14 ( 5 ): 664 – 669 . Google Scholar CrossRef Search ADS PubMed 13. Kotil K , Bilge T . Accuracy of pedicle and mass screw placement in the spine without using fluoroscopy: a prospective clinical study . Spine J . 2008 ; 8 ( 4 ): 591 – 596 . Google Scholar CrossRef Search ADS PubMed 14. Amato V , Giannachi L , Irace C , Corona C . Accuracy of pedicle screw placement in the lumbosacral spine using conventional technique: computed tomography postoperative assessment in 102 consecutive patients . J Neurosurg Spine . 2010 ; 12 ( 3 ): 306 – 313 . Google Scholar CrossRef Search ADS PubMed 15. Bamshad Azizi K , Ozgur G , Wilson E et al. Inertial measurement unit for radiation-free navigated screw placement in slipped capital femoral epiphysis surgery . Medical Image Computing and Computer-Assisted Intervention–MICCAI . Munich : Springer ; 2015 . 16. Gelalis ID , Paschos NK , Pakos EE et al. Accuracy of pedicle screw placement: a systematic review of prospective in vivo studies comparing free hand, fluoroscopy guidance and navigation techniques . Eur Spine J . 2012 ; 21 ( 2 ): 247 – 255 . Google Scholar CrossRef Search ADS PubMed 17. Villard J , Ryang YM , Demetriades AK et al. Radiation exposure to the surgeon and the patient during posterior lumbar spinal instrumentation . Spine . 2014 ; 39 ( 13 ): 1004 – 1009 . Google Scholar CrossRef Search ADS PubMed 18. Shin BJ , James AR , Njoku IU , Hartl R . Pedicle screw navigation: a systematic review and meta-analysis of perforation risk for computer-navigated versus freehand insertion . J Neurosurg Spine . 2012 ; 17 ( 2 ): 113 – 122 . Google Scholar CrossRef Search ADS PubMed 19. Miller CA , Ledonio CG , Hunt MA , Siddiq F , Polly DW Jr. Reliability of the planned pedicle screw trajectory versus the actual pedicle screw trajectory using intra-operative 3D CT and image guidance . Int J Spine Surg . 2016 ; 10 : 1 – 13 . Google Scholar CrossRef Search ADS PubMed 20. Hu Y , Yuan ZS , Spiker WR et al. A comparative study on the accuracy of pedicle screw placement assisted by personalized rapid prototyping template between pre- and post-operation in patients with relatively normal mid-upper thoracic spine . Eur Spine J . 2016 ; 25 ( 6 ): 1706 – 1715 . Google Scholar CrossRef Search ADS PubMed 21. Putzier M , Strube P , Cecchinato R , Lamartina C , Hoff EK . A new navigational tool for pedicle screw placement in patients with severe scoliosis . Clin Spine Surg . 2017 ; 30 ( 4 ): E430 – E439 . Google Scholar CrossRef Search ADS PubMed 22. Idler C , Rolfe KW , Gorek JE . Accuracy of percutaneous lumbar pedicle screw placement using the oblique or “owl's-eye” view and novel guidance technology . J Neurosurg Spine . 2010 ; 13 ( 4 ): 509 – 515 . Google Scholar CrossRef Search ADS PubMed Acknowledgments The authors wish to thank Magdalena Müller-Gerbl, Sandra Blache, Roger Kurz, and Peter Zimmermann from the Institute of Anatomy, University of Basel, for their kind collaboration, Selina Ackermann from the University Hospital Basel, for editorial assistance, Kate Galloway from KTB studios for the artwork, and Daniel Zumofen from the Departments of Radiology and Neurosurgery, University Hospital Basel for classifying screw positions. Copyright © 2018 by the Congress of Neurological Surgeons This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Operative NeurosurgeryOxford University Press

Published: May 30, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off