19 Image-guided Surgery in the Craniovertebral Junction and Upper Cervical Spine
Scott D. Boden pointed out: “The challenge and obligation of any new technology must be to solve an unsolved problem, enable the physician to perform an otherwise undoable task, or significantly facilitate the performance of a common task. We must all ensure that we do not encourage the triumph of technology over reason.”1 This is particularly true in the field of spinal surgery with its rapid development of navigational techniques over the past decade. This technological evolution not only brought about improved safety and accuracy of various techniques but also allowed less experienced surgeons to perform technically more demanding procedures. Despite this, the majority of experts correctly emphasize that image-guided spinal surgery is a supplement to, not a replacement for, the individual surgeon’s experience and judgment.
All procedures involving the upper cervical spine (UCS) and craniovertebral junction (CVJ) are technically demanding. The presence of vitally important structures and the unique anatomy of this region call for experience, excellence, and the most sophisticated technical solutions available. Numerous surgical techniques for treating pathologies of the CVJ and UCS have been described in the literature.2–11 Therefore, the introduction of new types of imaging and guidance techniques does not substantially contribute to further development of such approaches. However, there are certain types of operations where the extent of pathology is not directly visible, the resection volume of pathological tissue cannot be predicted, or the implant trajectory is out of sight of direct vision. In such situations, simple fluoroscopy is very helpful; however, its mono- or biplanar nature cannot provide the same three-dimensional (3D) control. Hence, new techniques that use image and real-time computer guidance have the potential to significantly improve the safety and accuracy of such procedures.
When it comes to spatial orientation, it is important to have a comprehensive knowledge of not only the target structures, but also all the potential surrounding dangers. Historically, neurosurgeons have always tried to reach deeply located lesions via the safest and least destructive route. Frame stereotaxy12 and frameless navigational systems13,14 proved to be indispensable in this task. Navigation in cranial surgery is simplified by the initially stable relationship of the brain to the skull. It allows for accurate surface registration prior to the intervention. Numerous attempts to use similar techniques in the spine were relatively successful in the past;12 however, the mobile nature of spinal elements prevented the use of stereotaxy in this region. The development of modern imaging techniques (computed tomography [CT], magnetic resonance imaging [MRI], and isocentric fluoroscopy), together with improved computer software solutions, opened the door to a new way of thinking about spinal surgery in three dimensions.15–18
The most suitable type of computer guidance system used in image-guided surgery is dependent on the task at hand. The requirements are different in, for example, oncologic procedures, where localization, tumor extent, and demarcation of the tumor’s borders have to be outlined, and the extent of resection may need to be continuously monitored, than for trauma cases, where mobile fragments are registered based on preoperative scans, but any deformity correction cannot be followed in real time without data re-registration.
It is important to realize that inaccuracies in instrumentation of the UCS or CVJ may result in three types of morbidity—neurological, vascular, and mechanical.
Relevant Anatomy
A detailed knowledge of the anatomy prior to surgery is necessary to avoid unacceptable surgical aggression or implant misplacement. All imaging modalities of the UCS are used to assess the individual anatomy and its variations prior to surgery. Plain radiographs (including lateral and transoral views) can demonstrate bony anomalies and/or deformities. CT, especially thin-cut studies with 2D or 3D reconstructions ( Fig. 19.1a ), are essential for bony anatomy analysis. Furthermore, the use of specific CT guidance protocols allows for virtual preoperative planning that is capable of assessing the feasibility and safety of screw implantation (e.g., implant length and width relative to the accepting bone). MRI ( Fig. 19.1b ) is more frequently used to depict soft tissue anatomy and pathology (e.g., disks, ligaments, and neural structures). Dynamic or cinegraphic MRI can demonstrate the relationship of the spinal envelope to its contents in flexion, extension ( Fig. 19.2 ), or other movements. For example, a damaged transverse ligament can be directly visualized.
Occipital Bone
Osseous anatomy of the occipital bone is important whenever occipitocervical fusion is being considered. Most of the currently available constructs use occipital plates with either lateral (squamous part of bone) or me-dial (occipital ridge) screw holes. It is important to pay attention to the thickness of the bone, as well as the position of the confluence of the sinuses.19–21 Medially placed occipital screws provide stronger purchase due to thicker bone. Screws can be optimally placed in the skull with the aid of image guidance (CT-based cranial navigation, intraoperative CT, or isofluoroscopy); however, its routine use is probably not necessary for this task.
Clivus
Virtual navigation of the clival region is frequently used for skull base tumor localization and for intraoperative monitoring of the extent of tumor resection. Whenever clival screws are considered, the position, angle, and thickness of the clivus needs to be assessed prior to surgery. This structure can be virtually navigated from cranial registration. Navigation can also help to localize the clival edge as an anatomical landmark in complex transoral surgeries.
Atlas
The first cervical vertebra has no body, and the only strong structures for potential implant anchorage are the lateral masses. They can be included in anterior22,23 or posterior atlantoaxial transarticular fixation,4 or they can act as isolated anchors in different constructs.2,24,25 Although the vertebral artery can potentially be violated during anterior exposures, implant-related injury is usually limited to posterior techniques.26–28 This risk may be increased due to vertebral artery course anomalies.29–32 Venous bleeding from vessels surrounding both C2 nerve roots can be avoided by subperiosteal dissection2 and better positioning of the patient on the surgical table.25 Recent publications warn of other “hidden” risks during lateral mass screw placement,26,33 especially if bicortical screw purchase is necessary to strengthen the construct.26 The carotid artery and hypoglossal nerve are also at risk. The hypoglossal nerve lies 2 to 3 mm lateral to the middle of the anterior aspect of the C1 lateral mass.33 The internal carotid artery varies in location but can lie within 1 mm of the potential screw exit point.26 Therefore, it is recommended to place such screws in the posteromedial sublaminar part of the lateral mass; also, the trajectory should be slightly medial. In general, the so-called screw entry safe zones (both anterior and posterior) have been established,2,23–25,34–37 and preoperative thin-cut CT and/or CT/MRI angiography may be helpful in selected cases with suspected vascular anomaly.
Axis
The unique anatomy of the second cervical vertebra represents the most challenging structure in the UCS. In particular, the variability of the vertebral artery course and its modification under pathological conditions are the subject of many articles.29,38–40 A high-riding vertebral artery groove is described in up to 23% of patients.29,31,37–41
The axis, in comparison to the subaxial vertebrae, has the largest pedicles but the least developed lateral masses; hence, the pedicle appears more suitable for screw placement.42–44 This assumption is not supported by the work of Resnick et al.36 and Yoshida et al.45 Measuring the distances of the vertebral artery groove from the possible screw trajectory, they found that the C2-pedicle screw placement has nearly the same potential risk of vertebral artery injury as transarticular screw placement, and they recommended a preoperative 3D computerized evaluation to choose the best surgical technique. Numerous anatomical exclusion criteria for safe transarticular screw placement have been described. Abou Madawi et al.31,39 stated that a 3.5-mm screw cannot be introduced if the internal height of the C2 lateral mass, measured on CT reconstructed images, between the highest point of the vertebral artery groove and the superior facet surface is < 2 mm ( Fig. 19.3 ). This statement is questionable, as the screw trajectory does not necessarily intersect the measured area, and it can be adjusted for adequate placement.46
Measuring the dimensions of the isthmus of the pars interarticularis in the real screw trajectory modeled on 3D software is more reliable.40,47,48 Mandel et al.49 analyzed C2 vertebrae of 205 human cadavers and found that 10% of specimens had an isthmus < 5 mm wide and 5 mm high. This is less than was expected from related clinical articles. Bloch et al.47 studied 17 cadaveric spines and determined that the risk of injury to the vertebral artery can be decreased to 5.9% with the use of a navigational system if the entry point and screw trajectory are modified individually based on preoperative 3D planning. Under these circumstances, the isthmus height can be 4.0 mm for safe placement of a 3.5-mm screw. We also have to consider that size differences depend on the individual’s gender, race, and even the side on which the procedure is being done.29,37,38,46,47,49,50
In conclusion, the height of the internal isthmus should be > 5 mm on CT sagittal reconstructions and 5 mm lateral from the internal spinal canal border ( Fig. 19.4 ). This indirectly implies that the isthmus width is > 5 mm. In challenging cases, the possible trajectory should be tested in a 3D spatial program changing the entry point and screw path.48 If a C2 pedicle screw is considered, the size and angle of the individual pedicle should be measured on a CT reconstruction directly and the trajectory planned accordingly.36,49 In questionable or borderline situations, CT angiography should be done to localize the course of the vertebral artery exactly. The space occupancy ratio of the vertebral artery/bony groove was measured by Cacciola et al.29 as 79% (34–100%).30 Anterior safe zones of the axis were defined by Kandziora et al.35 and Koller et al.23
Technique Description
Virtual Image-guided Surgery
After a positive experience with cranial image guidance, Foley and Smith introduced virtual image-guided surgery to the spine in 1994.13,51 The main problem was the lack of correlation between reliable skin surface registration markers (fiducials) and bony structures of the spine because of skin and spine movement. This was solved with the registration of bony anatomical landmarks of the dorsal spine in association with dynamic reference array (DRA) directly attached to the target vertebra.13,14,52,53 Most surgeons at the time were using virtual image-guided surgery for lumbar pedicle screw placement.13,14,54–56 Later, it became evident that in the lumbar or thoracic spine, implant navigation could be more accurate than traditional methods,30,55,57 and the technique also began to be used in more delicate structures of the UCS.15,16,58–60
In principle, according to the specific protocol, the anatomical data have to be obtained by thin-sliced (1.0–1.5 mm) CT or MRI and transmitted to the computer workstation. DRA must be firmly connected to the target area so as not to hinder the operation. The electro-optical camera connected to the computer workstation then registers the position of the navigated vertebra and the instruments being tracked. The space orientation of tracked entities is marked by either passive arrays (reflective spheres) or active arrays (light-emitting diodes). The virtual picture is shown on the computer output display. The surgery itself follows the virtual preoperative plan and can be controlled in 3D on the screen.
All available systems work with images obtained either before the surgery or during the operation after exposure of the target structures. This means that the anatomical data set is obtained prior to intervention or implant introduction. Intersegmental movement during positioning of the patient or the surgery itself can change the virtual data set. Therefore, only one vertebra can be virtually guided, and any change in the DRA connection has to be updated with the help of new manual or automated registration.
Planning in Cervical Surgery
The acquired set of anatomical data is so comprehensive that the workstation software allows the surgeon to make a plan for any intervention. In the UCS and CVJ, 3D analysis usually focuses on the feasibility of C2 screw acceptance, the course of the vertebral artery, the localization of the spinal cord and brainstem, the position and thickness of the clivus, and the planning for tumor approach and re-section. Such planning allows the surgeon to determine the feasibility of complex procedures. Virtual planning is often recommended even when further navigation is not possible or required.38,39,48
Preoperative CT-based Virtual Image-guided Surgery
This is the most accurate method of virtual spinal bone navigation, especially effective in anatomically difficult regions. If the registration process can be done precisely enough (1.5-mm accuracy), image guidance can be used for all known procedures of screw introduction in the UCS and CVJ. It is most frequently used in transarticular C1–C2 fixations15,16,58,61,62 and the published results describing precision of screw placement are better with navigation than using conventional techniques.30,63–65
Major Drawbacks
A specific protocol calling for preoperative CT is necessary, which increases the cost if CT adequate for diagnosis already exists but is not compatible with the image-guided system. The time for registration is long, there is a learning curve, and only one vertebra can be navigated; therefore, a separate registration process for each vertebral level is unavoidable.
Intraoperative CT-based Virtual Image-guided Surgery
The CT scanner is located in the operating room, and the registration process is automated. The scans are obtained in the final surgical position. Inaccuracy caused by inter-segmental movement can be minimized.51 Postinstrumentation images allow for an intraoperative check of the position.
Major Drawbacks
Drawbacks include the cost of the mobile scanner, ergo-nomic problems (draping, patient manipulation, etc.), the necessity of a special transparent table, and the fact that movable parts of the spine (fracture fragments) cannot be visualized in changed positions without new registration.
Fluoroscopy-based (Two-dimensional) Virtual Image-guided Surgery
This is a method combining fluoroscopy, familiar to all spine surgeons, with image guidance techniques.14,51,66 The accuracy of instrumentation and virtual guidance is increased in only one plane at a time. Its usefulness was not described in UCS and CVJ surgery. Nevertheless, the radiation exposure is much less than in classical fluoroscopy.
Major Drawbacks
The guidance depends on the quality of images obtained by fluoroscope. This can be a problem in shadow areas of the upper thoracic spine or in obese patients. Precise images of the complex UCS anatomy cannot be obtained regularly. Bone structure is also not clearly visible in patients with osteopenia and deformity.
Related posts:
9 Congenital and Developmental Anomalies of the Craniovertebral Junction
11 Atlantoaxial Transarticular Fixation Techniques
16 Transoral Surgery
12 Occipitocervical Stabilization
20 Intraoperative Neurophysiological Monitoring in Surgery for the Craniovertebral Junction
22 Techniques of Occipitocervical Fixation
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
