Intraoperative Image-Guided Navigation for Lumbar Interbody Fusion

Introduction

Minimally invasive spine surgery has gained much popularity in recent years owing to the reductions in patient morbidity, length of hospital stay, and costs. Although these short-term outcomes have seen marked improvements, there has been little improvement in the long-term outcomes when comparing minimally invasive lumbar interbody fusion (MIS-LIF) to open techniques.

When first developed, MIS-LIF required extensive fluoroscopy to ensure accurate interbody cage placement and extensive imaging for the percutaneous placement of pedicle screws. The consequent accumulation of radiation from many of these minimally invasive procedures may result in dangerous radiation dosages to the surgeons who perform these procedures. As such, there has been an increased development and use of navigation-based techniques that rely on the use of intraoperatively acquired images with subsequent image registration allowing for navigation of interbody cage and percutaneous screw placement. This approach exposes surgeons to much less radiation while maintaining comparable accuracy.

In this chapter, we review various advanced imaging modalities related to both accurate interbody cage and associated pedicle screw placement in the lumbar spine.

Fluoroscopy

Prior to the advent of advanced imaging modalities relying on computer-aided image processing and registration, fluoroscopy was utilized to ensure proper cage placement. This method requires successive anterior-posterior and lateral C-arm images to ensure that the cage is inserted orthogonal to the disk space. In patients with deformity or multilevel degenerative disease, ensuring a perfect orthogonal position can require tilting the table to acquire appropriate images; consequently, with the repetitive imaging, there can be significant radiation exposure. Additionally, the accuracy of cage and pedicle screw placement has been a concern, particularly in comparison to more open techniques where visualization is much easier.

Given the increased exposure to large amounts of low-level radiation that can result from high case volumes, several studies have aimed to quantify the average surgeon radiation exposure during MIS-LIF cases. Regarding lateral lumbar interbody fusion (LLIF) cases, Taher et al. found that during eighteen cases fusing a mean 2.4 levels, average total fluoroscopy time was 88.7 seconds, including fluoroscopy at the beginning of the case to ensure accurate positioning. Of note, significant increases in radiation exposure were noted in unprotected areas when compared to the dosimeter located under the lead apron of the primary surgeon. Bindal et al. observed an average fluoroscopy time of 101 seconds during minimally invasive transforaminal interbody fusion (MIS-TLIF), with radiation exposures that were generally improved in a later study by Funao et al., who used a one-shot fluoroscopy technique in an attempt to lessen or reduce surgeon radiation exposure. Other similar low-dose fluoroscopy protocols have been developed to decrease radiation exposure during MIS-TLIF cases.

The significance of such radiation exposures to the surgeon is unclear, although various authors have suggested that exposures may have a more critical impact on younger surgeons who are beginning their practice and have a lifetime of fluoroscopy-dependent spine procedures ahead of them. With this in mind, Taher et al. calculated that 2700 LLIF procedures theoretically could be performed each year without exceeding standards for “safe” occupational radiation exposure. Although this may be true, an interest in reducing surgeon radiation exposure persists.

Stereotactic Navigation

To alleviate the concerns of increases in surgeon radiation exposure and of the placement accuracy of both pedicle screws and interbody cages, there has been a recent push toward the development of technologies that utilize imaging to register an image at the start of the procedure to be used as a reference for navigating instruments. Radiation exposure to surgeons and ancillary personnel is thereby theoretically reduced, as the images taken at the beginning of the procedure for image registration do not require the close proximity of staff. As the procedure progresses and is fully navigated, surgeon visibility improves, ideally also improving the accuracy of placement and addressing both concerns.

Imaging modalities that have been used for the generation of these reference images include intraoperative C-arm fluoroscopy and computed tomography (CT) scans via either an O-arm or another intraoperative CT scanner ( Fig. 5.1 ). In MIS-LIF cases, there is the added benefit of using image registration methods that can be performed after positioning and draping, to decrease navigation error owing to the changes of patient positioning. One general drawback of using navigated instrumentation is increased set-up time, although this may not be a significant issue as time expense can be avoided later in the procedure by obviating the need for repeated fluoroscopy.

Modern 3D Image Acquisition Systems

Navigated interbody and pedicle screw placement requires intraoperative image registration using a dynamic reference frame to allow for effective three-dimensional triangulation of instruments. In general, positioning depends on the approach (supine for posterior approach interbody fusion, left lateral with knees and hips flexed for lateral interbody fusion) and sterile draping occurs following induction of anesthesia as during a non-navigated case. Notably, however, draping must include the site of dynamic reference frame placement, and these frame placement sites include adjacent spinous processes, the iliac crest, and skin during the posterior approach cases and the anterior and posterior superior iliac spines during lateral approaches.

Placement of the dynamic reference frame during posterior minimally invasive approaches to interbody fusion has been investigated in detail as many options exist for bony fixation, including the iliac crest and cephalad spinous processes, during posterior approach cases. Cho et al. investigated the use of cutaneously fixed dynamic reference frames to avoid significant issues that can occur with traditional fixation to bony landmarks. For example, if the reference frame is affixed to the spinous process at the level of the surgical site or on the posterior superior iliac spine, there is a higher possibility of decreasing the already small working field and possibly causing metal artifact. And despite the fact that most reference markers are made from titanium, artifact can commonly occur. However, placing the reference frame at a separate spinous process necessitates a separate incision. Cho et al. found that using a dynamic reference frame affixed to the skin overlying the sacral hiatus allowed acceptable navigation of pedicle screw accuracy during mini-open TLIF. This finding held up despite the potential drawbacks associated with increased dynamic reference frame distance from the pathology and possibly lower stability than fixation to bone.

After sterile placement of the dynamic reference frame, intraoperative three-dimensional imaging is taken, performed via fluoroscopy (i.e., Siemens Iso-C ^3d , ARCADIS Orbic 3D) or CT scans (i.e., Medtronic O-arm). Images are then registered using the associated navigation system. During this early image registration phase, surgeons and ancillary staff can opt for more stringent radiation protection protocols to minimize radiation exposure. These include moving far from the field and using more extensive lead shielding. Following image registration, the surgical procedure can continue as previously performed, replacing active C-arm action shots with direct navigation of all instruments.

Interbody Cage Placement

Several studies have investigated the use of computer-aided navigation during interbody cage insertion. For example, Drazin et al. describe LLIF via an O-arm linked to the Stealth-Station TREON System (Medtronic Sofamor Danek). Their technique involved the use of a dynamic reference frame attached to a pin that is inserted into the posterior superior iliac spine. Use of the O-arm intraoperatively allows for sterile draping of the patient and full positioning prior to image acquisition and registration, thereby decreasing the chance of error from altering the orientation of the dynamic reference frame to the patient’s spine ( Fig. 5.2 ). Park described a similar technique for LLIF that also relied on the use of an O-arm and a similar tracking system, although it included minor changes in retractor positioning and reference frame location. Park advocated the use of the anterior superior iliac spine for placement of the reference frame, although it may slightly increase the risk of injury to the lateral femoral cutaneous nerve. In a series of eight patients, however, Park did not observe any instances of iatrogenic nerve injury. Additionally, on postoperative fluoroscopy, all cages were noted to be placed within the anterior three quarters of the disk space in question, indicating acceptable accuracy as well.

Similar accuracy was demonstrated in a cadaver study comparing conventional fluoroscopy with navigation using registered fluoroscopic images for a direct lateral approach to cage insertion. Interestingly, study authors noted that the lowest accuracy was observed at the L1-2 level, which was located the farthest from the dynamic reference frame affixed to the anterior superior iliac spine. Apart from accuracy, the study also investigated the time needed to successfully position interbody cages. Although significantly more time was needed during the initial setup of the navigation system, subsequent steps required less time with navigation, resulting in similar overall operative time.

Pedicle Screw Placement

When initially developed, navigation for pedicle screw placement was severely limited by poor image quality and underdeveloped computational software. However, over the years, navigation for pedicle screw placement has been explored extensively in the context of open procedures. Studies have generally demonstrated improved accuracy when comparing navigated techniques with images registered through fluoroscopy, conventional CT scans, or O-arm technology ( Fig. 5.3 ). In particular, Kosmopoulos and Schizas demonstrated that when pooling 130 studies of over 35,000 pedicle screws, median screw placement accuracy was 95.2% in navigated cases as compared to 90.3% in non-navigated cases, a conclusion supported by other reviews. Most of these studies have investigated non-minimally invasive surgery, where even freehand pedicle screw technique can be utilized. Many minimally invasive interbody fusion cases, however, rely on percutaneous pedicle screw placement for posterior stabilization.

Percutaneous pedicle screw placement is an important adjunct to minimally invasive interbody fusion as it allows for posterior stabilization without extensive paraspinal muscle dissection, as would be observed in open fusion cases. This can significantly reduce postoperative wound pain, particularly when considering the reduced morbidity from avoiding muscle dissection from transverse processes, as would be performed for traditional posterolateral lumbar spine fusion surgery. One drawback however, of percutaneous pedicle screw placement is facet violation, which is more likely in this instance because the facet joint is not directly visualized. Additionally, there is an increase in future risk of symptomatic adjacent segment degeneration and a resultant need for further revision surgery. Other concerns include the risk of anterior guide wire migration through the vertebral body, particularly in severely osteoporotic patients, which may increase the risk of intestinal or great vessel injury.

To address these concerns, navigation techniques have been developed for use in minimally invasive lumbar spine surgery. Robotic innovations, such as the Mazor Robotics Renaissance guidance system, uses preoperative CT images as guides to plan for the surgical procedure with much more precision before beginning the surgery ( Fig. 5.4 ). Intraoperatively, the robotic system can be placed in a specific location and can achieve a specified trajectory, allowing instrument placement by the surgeon with more precision ( Fig. 5.5 ). The Renaissance Guidance System has been studied in spinal instrumentation procedures. Percutaneous pedical screw placement, for example, was examined, illustrating higher accuracy of placement with much lower radiation exposure. Kantelhardt et al. demonstrated a mean decrease of 43 seconds of intraoperative x-ray exposure in robotic-guided procedures as compared with the conventional imaging methods as well. Additionally, using the robotic system for minimally invasive cases can decrease damage to the surrounding healthy tissue and lead to faster recovery.