The minority of patients with spinal metastases require surgery. In a study of 31 consecutive patients who had an underlying carcinoma and who initially had back pain (Galasko and Sylvester, 1978), we found that 11 patients had a benign lesion causing their back pain, seven had painful metastases that did not warrant surgical intervention, three had spinal instability with symptoms of sufficient severity to warrant surgical intervention, and ten had evidence of malignant spinal cord or cauda equina compression.

The majority of painful metastases can be treated with radiotherapy, usually combined with adjuvant treatment, such as chemotherapy or endocrine therapy, and frequently combined with bisphosphanates. Recent studies suggest that zoledronic acid, a third-generation bisphosphanate, may be the most useful in patients with skeletal metastases (Hortobagyi, 2005). These forms of treatment are reviewed in other chapters. Radiopharmaceuticals also have been used.

This chapter concentrates on the surgical management of spinal instability and/or malignant spinal cord or cauda equina compression. Surgery also may be occasionally indicated in patients with spinal metastases that have not responded to other forms of treatment and are still painful. Surgery also includes biopsy of a lesion for which the diagnosis is not known, but this aspect is not considered further in this chapter.

A variety of surgical techniques are available. As indicated earlier, preoperative embolization may be essential. The stabilization can be anterior, posterior, or combined anterior and posterior. More recently, minimal-access surgery (Huang et al., 2006) and endoscopically assisted decompression was undertaken (McLain, 1998). McLain concluded that endoscopic-assisted decompression can reduce morbidity and may aid surgery, improving the surgeon’s vision, providing light and magnification as well as a more direct view.

Spondylectomy has been advocated for patients with a single metastasis and in whom the primary tumor has been successfully excised (Sakaura et al 2004, Tomita et al 1994, Abe et al 2001).

Vertebroplasty can be used for painful metastases that have not responded to other treatment or if the instability can be controlled by vertebroplasty without the necessity for instrumentation, as may be the case with a painful compression fracture secondary to spinal metastasis (Alvarez et al., 2003; Cotton et al., 1996; Weill et al., 1996; Gaitanis et al., 2005; Hadjipavlou et al., 2005; Yamada et al., 2004; Barragen—Campos et al., 2006; Vrionis et al., 2005; Dudeney et al., 2002; Burton et al., 2005). This involves the percutaneous injection of bone cement into lytic lesions in the vertebral body to provide support and prevent collapse of the vertebral body. Basically, it takes advantage of interventional radiologic techniques. It has been used in the treatment of osteolytic lesions, including myeloma and spinal metastases. Although a number of articles have been published on this procedure, few have clearly defined the outcome measures, and virtually none has provided a long-term follow-up in terms of the clinical indications, expected complications, and inherent risks in patients with skeletal metastases.

Approximately 80% of the forces to the spinal column are transmitted through the anterior column (Lowery and Harms, 1997). Metastases are more common in the vertebral body as opposed to the posterior elements, and this predisposes to kyphotic collapse, as the metastatic tissue is structurally less able to transmit these forces than is normal cancellous bone. Not all patients are suitable for anterior surgery because of the widespread nature of their metastases, and some patients, because of their poor medical fitness, may not be suitable for major surgery, even though it is indicated. Tokuhashi (1994, 2005, (Table 13.1) made an attempt at a scoring system to determine whether palliative stabilization or a more radical palliative resection should be undertaken. The system uses six parameters: the general condition of the patient; the number of extraspinal bone metastases; the number of metastases in the vertebral body; metastases to major internal organs (lungs, liver, kidneys and brain); the primary site
of the cancer; and the severity of spinal cord palsy. They concluded that the total score obtained for each patient could be correlated with the prognosis but that the prognosis could not be predicted from a single parameter. They advised an excisional operation in patients with a good prognosis and a palliative procedure in those with a poor prognosis. The latter was aimed only at securing support with no or only partial resection of the lesion. To assess the general condition of the patient they used the criteria established by Karnofsky (1967). The Tokuhashi score tends to segregate those who would benefit from conservative treatment and those who would benefit from surgery. It is a useful tool but has not been fully validated and is only a guide to selection. The Karnofsky score (Karnofsky, 1967) is a functional score that can be used to assess changes in outcome after major surgery for spinal metastases, as is the Eastern Cooperative Oncology Group (ECOG) performance status (Oken et al., 1982).

Tomita et al. (2001) based their surgical strategy on a prognostic scoring system that used the grade of malignancy, visceral metastases (no metastases, treatable, untreatable), and skeletal metastases (solitary or isolated, multiple). The surgical strategy was based on a prognostic score. They divided their patients into four groups depending on the prognostic score. Those with the best prognosis were treated with wide or marginal excision, the next group by marginal intralesionary excision, the third group by palliative surgery with stabilization, and those with the worst prognosis, with nonoperative supportive care. The mean survival in their first group was 38.2 months; in the second group, it was 21.5 months; and in those treated with palliative surgery and stabilization, 10.1 months. The mean survival time of patients treated with terminal care was 5.3 months. It is my view that if a patient is in excruciating pain, spinal stabilization is warranted, even if the survival is no more than 2 to 3 months.

Taneichi et al. (1997) attempted to assess whether impending vertebral collapse could be assessed and prophylactic stabilization carried out. They suggested that a significant risk of collapse of thoracic vertebrae exists with a 50% to 60% tumor involvement of the vertebral body with no destruction of other structures or 25% to 30% involvement of the vertebral body with costovertebral joint destruction, and in the thoracolumbar and lumbar spine, if 35% to 40% involvement of the vertebral body or 20% to 25% of the vertebral body with pedicle destruction was found.

Posterior stabilization is less invasive, and the postoperative management usually requires less monitoring and is less intensive. It is usually a smaller surgical insult and can be carried out on less medically fit patients. As has already been indicated, most of the metastatic deposits are anterior and, therefore, the posterior structures at the level of pathology are relatively strong, consisting of pedicles and laminae unaffected by the disease process. If stabilization is combined with a decompressive procedure, these relatively intact structures (spinous processes, interspinous ligaments, and laminae) require removal for the decompression to be undertaken. Thus, in a posterior decompression, the spine is destabilized further by sacrificing intact bony elements to decompress indirectly the neural elements.

Posterior decompression alone, without stabilization, is contraindicated because of the further destabilizing effect. If a posterior decompression is thought to be appropriate, it should always be combined with posterior instrumentation to restore stability. It is usually possible, when carrying out posterior stabilization, to use pedicle screws, sublaminar hooks, or sublaminar wires. However, implant failure will occur, either by implant breakage or implant pull-out unless either a sound bony fusion occurs or the patient dies of the malignancy. Posterior stabilization also is mechanically inferior as a deficient anterior column is not reconstructed. Therefore, force must be transferred to the posterior elements. At least two spinal segments must be stabilized on each side of
the pathologic level, so in total, five levels or more require instrumentation when combined with a one-level posterior decompression. Posterior stabilization alone is inadequate for lesions at L-5 and usually L-4. Posterior surgery has a number of drawbacks compared with anterior surgery.

Nevertheless, posterior stabilization alone rarely fails, if carried out properly and patient selection is correct.