Surgery involving the cervicothoracic junction (CTJ) is universally accompanied by questions about postoperative stability. This junction must be crossed in treating many pathologic conditions: tumors, trauma, infections, and inflammatory, degenerative, and congenital disorders. Anterior access to this region is complex and associated with morbidity, requiring significant surgical skill. Posterior approaches provide a simpler alternative, but the complex biomechanics requires thoughtful preoperative planning. Laminoplasty is acceptable when preoperative cervical lordosis is present and facet disruption is minimized. However, even when lordotic alignment is normal, the ability of laminoplasty to maintain cervical range of motion (ROM) remains questionable, and doing so is even considered counterproductive by some authors. Furthermore, the more accepted forms of laminoplasty (i.e., open- door and double-door) have a narrow range of indications. Standalone cervical laminectomy in the setting of kyphosis (i.e., “reverse lordosis”), sigmoid deformity (i.e., “swan neck”), straightening of the normal cervical lordosis, or crossing of the CTJ remains controversial and recently has largely been abandoned. Laminectomy involving segments rostral to or including C7, in which there is disruption of the C7-T1 interspinous ligament, is particularly controversial. Many cases of postlaminectomy instability, characterized by progressive kyphotic deformity or, more severely, by cervical stenosis with accompanying neurologic deficits from kyphosis or listhesis, have been reported after this procedure. With improved understanding of the complex biomechanics of the CTJ, particularly the inherent stresses in this region of transition from the mobile cervical spine to the relatively immobile thoracic spine, ways to minimize postoperative structural failure become clearer. The stresses here are only compounded with partial disruption of the posterior elements (interspinous ligaments and ligamentum flavum) as required for effective laminectomy.
Alternatives to formal laminectomy (e.g., arcocristectomy, myoarchitectonic spinolaminoplasty) may provide a solution for those limited cases involving only stenosis, but the long-term benefits have not been confirmed by formal longitudinal study. An improved understanding helps to reinforce surgeons’ comfort with stabilization of this region. Those well versed in spinal neurosurgery now accept that destabilization of the cervicothoracic spine (including posterior element disruption only) typically should be followed with a subsequent procedure to provide adequate stabilization.
A 54-year-old man with a history of cerebral palsy came for treatment of progressive neck pain and right upper extremity radicular pain.
Exam: His neurologic examination revealed bilateral deltoid weakness, spasticity in all four extremities, and pathologic reflexes.
Imaging: Magnetic resonance imaging (MRI) of the cervical spine ( Figures 15-1 and 15-2 ) revealed diffuse central stenosis from the second to the seventh cervical segments secondary to marked spondylosis with both anterior and posterior compression. The patient’s MRI scans also showed loss of the normal lordotic curvature with straightening. Flexion and extension radiographs revealed no hypermobility or instability.
Surgical options include the following: anterior decompression via multilevel diskectomies or corpectomies, laminoplasty, laminectomy without posterior fusion, or laminectomy with posterior fusion (and/or posterior-anterior fusions if there is significant athetoid motion).
Given the patient’s loss of lordosis and circumferential compression, a multilevel cervical laminectomy with lateral mass fusion was recommended. Intraoperatively, his habitus prevented optimal intraoperative imaging. He ultimately underwent a decompressive laminectomy from C3 through C7. Given the disruption of his C7-T1 interspinous ligament, it was deemed that fusion would need to include T1. C7 was skipped to facilitate instrumentation.
Postoperatively, the patient’s myelopathic findings did not progress and his deltoid weakness improved modestly. He did well for several months until he fell from standing height. He noted immediate neck pain. Examination revealed no significant change in his baseline neurologic status. Radiographs ( Figure 15-3 ) revealed bilateral fracture of the T1 screws at the interface between the pedicle and pars interarticularis. Flexion and extension radiographs showed persistent motion (see Figure 15-3 ) despite salvage attempts with a rigid cervical orthosis and strict adherence to treatments with an external bone growth stimulator. Ultimately, his fusion had to be extended to more caudal thoracic segments without anterior supplementation.
This case raises several questions that will be addressed using published evidence:
Would standalone laminectomy have been appropriate in this patient?
Was there a place for laminoplasty in this case?
Should the initial fusion have been longer? If so, how far caudally should it have been extended? What type of instrumentation would have been optimal in this case?
Would anterior stabilization have helped?
Would intraoperative navigation (use of three-dimensional [3D] systems rather than two-dimensional [2D] fluoroscopy) have been helpful in this case?
Laminectomy is performed using established techniques. Subsequent lateral mass fusion is facilitated by wide exposure of facets. An appropriate starting point can be determined by creating an imaginary X over the lateral mass ( Figure 15-4 ). Superior and inferior boundaries are the facet joints, and medial and lateral boundaries of the lateral mass serve as the other boundaries. The ideal starting point is 1 mm medial to the middle of the imaginary X. A “matchstick” bur is used to penetrate the cortex and create a starting point ( Figure 15-5 ). An up-and-out technique is used for the hand drill trajectory ( Figure 15-6 ). A medial to lateral trajectory at 30 degrees avoids injury to the vertebral artery, and a cephalad to caudal trajectory at 20 degrees avoids injury to the nerve root. Before placement of screws, the facet joints of the segments included in the fusion are decorticated ( Figure 15-7 ). The dorsal cortical surfaces are decorticated for onlay arthrodesis as well. Polyaxial screws are placed and can be measured before placement during the hand drill and feeler steps. A medial trajectory risks injury to the vertebral artery ( Figure 15-8 ). Failing to aim cephalad places the nerve root at risk. Starting too far laterally risks fracture of the lateral mass. When extension to the upper thoracic region is necessary, upper thoracic pedicle screws are placed using anatomic landmarks, because visualization with fluoroscopy at these levels can be difficult in the sagittal plane. T1 pedicle screw trajectory is described in Figure 15-9 ; a large lateral mass screw can be used to avoid the need for a transitional rod in the construct.
The caudal trajectory for thoracic pedicle screws decreases from 20 degrees to 5 degrees from T1 to T12.