Treatment and prognosis of epidural spinal cord compression, including cauda equina syndrome


INTRODUCTION — Neoplastic epidural spinal cord compression (ESCC) is a common complication of cancer that causes pain and sometimes irreversible loss of neurologic function. The degree of thecal sac compression required for the designation of ESCC has been variably defined; we and others consider any radiologic evidence of indentation of the thecal sac to be evidence for ESCC. In adults, the tip of the spinal cord usually lies at the L1 vertebral level; below this level, the lumbosacral nerve roots form the cauda equina, which floats in cerebrospinal fluid (CSF). Since the pathophysiology of compression of the thecal sac at the level of the cauda equina does not differ significantly from that of more rostral compression, compression of the cauda equina is still generally referred to by the slightly inaccurate name of ESCC.

Early diagnosis, usually by magnetic resonance imaging or myelography, is essential but the diagnosis is often delayed. In one series of 301 consecutive patients with malignant spinal cord compression, for example, the median delay to treatment in the patients with known malignancy was two months from the onset of back pain and 14 days from the onset of symptoms of spinal cord compression. Most importantly, the majority of patients deteriorated by at least one grade in motor or bladder function during the delay from the initial symptoms of ESCC.

Back pain is a precursor to spinal cord injury in almost all patients with spinal metastases. A prospective study which evaluated 87 such patients radiologically identified three groups: myelopathy; radicular pain; and back pain with normal neurologic findings. The respective likelihood of epidural metastases in the three groups was 78, 61, and 36 percent. A similar stratification was seen with the likelihood of more than 75 percent block on myelography: 93, 53, and 33 percent, respectively. Thus, radiologic testing for early diagnosis should be performed in all cancer patients who develop otherwise unexplained and persistent back pain.

THERAPY — The goals of therapy for ESCC include pain control, avoidance of complications, and a rational attempt to preserve or improve neurologic functioning utilizing techniques appropriate to the patient's burden of disease, life expectancy, and values.

General principles — Management include the administration of corticosteroids in all patients, followed either by surgery and/or radiotherapy (RT). Aggressive surgical management with anterior decompression appears to give better results than RT alone in patients who are candidates for surgical intervention. Compared to RT alone, surgery followed by RT is associated with better neurologic outcomes. Chemotherapy may be beneficial in patients with chemosensitive tumors.

Symptomatic treatment of ESCC often begins prior to more definitive therapy and consists of the following general principles.

  bullet Pain management — Patients with ESCC are frequently in severe pain, often limiting the ability to perform a thorough neurologic examination. Corticosteroids usually improve the pain within several hours, but most patients will require opiate analgesics to tolerate the physical examination and necessary diagnostic studies.
  bullet Bedrest — There is generally no need to confine the patient to bed; patients are quite adept at avoiding maneuvers that trigger their pain and there is no risk that movement will worsen the neurologic status.
  bullet Anticoagulation — Patients with cancer are in a hypercoagulable state.  Although the value of prophylaxis against venous thromboembolism has not been studied in patients with ESCC, it would seem reasonable to give prophylactic subcutaneous heparin or sequential compression devices to nonambulatory patients who are at moderate to high risk.
  bullet Prevention of constipation — Autonomic dysfunction from the spinal lesion, limited mobility, and analgesics may contribute to the development of constipation, ileus, and occasionally perforation of an abdominal viscus, the symptoms of which may be masked by corticosteroids. Consequently, an aggressive bowel regimen is indicated.

  bullet Spinal bracing — External spinal bracing is sometimes employed but has not been carefully studied. It should be considered only in patients with pain refractory to the standard management.
 

Corticosteroids — High dose corticosteroid therapy is generally considered to be part of the standard regimen for ESCC, despite limited documented evidence of benefit and a significant risk of serious side effects. Several studies have suggested that lower doses can be effective but they have not been assessed in randomized trials.

  Clinical trials — The beneficial actions of corticosteroids in restoring neurologic function in patients with ESCC were first reported in the late 1960s. Animal studies documenting salutary effects in various models of ESCC soon followed.

Only one randomized clinical trial has addressed the utility of corticosteroids in ESCC. In this study, 57 patients with carcinoma (two-thirds with a primary in the breast) and myelographically confirmed ESCC were randomized to receive either no dexamethasone or dexamethasone (96 mg intravenously followed by 24 mg four times daily for three days and then tapered over 10 days). Stratification factors included primary tumor type and the extent of neurologic dysfunction; all patients received standardized RT. A significantly higher percentage of patients in the dexamethasone group remained ambulatory both at the conclusion of therapy (81 versus 63 percent) and at six months (59 versus 33 percent). Significant side effects were seen in three patients (11 percent in the steroid group).

Steroid-induced complications with this regimen were more prominent in another series of 28 patients undergoing RT: side effects were noted in 28 percent, one-half of which were considered serious and life-threatening. As a result, the high dose regimen was abandoned and the dose was lowered to 16 mg daily tapered over two weeks; there were no serious side effects seen in 38 patients and the ambulatory outcome was similar to that with higher dose therapy.

In a further attempt to identify the optimal corticosteroid dose, 37 patients with ESCC were randomized to an initial dexamethasone bolus of 10 mg or 100 mg intravenously, both followed by 16 mg daily orally. The average pain score decreased significantly from 5.2 to 3.8 at three hours, 2.8 at 24 hours and 1.4 at one week. There were no differences in pain control or neurologic outcome between the two groups.

It remains uncertain if patients with less severe disease need steroid therapy. One small phase II trial showed that patients with back pain but no myelopathy and less than 50 percent narrowing of the spinal canal by epidural tumor could successfully undergo RT without receiving steroids.

  Recommendations — The physician has a choice between the "high" dose of dexamethasone with proven efficacy and a relatively high rate of serious side effects and a "low" dose with notably fewer side effects but no randomized controlled data to support its use.

I reserve the high-dose regimen for patients with paraparesis or paraplegia and then halve the dexamethasone dose every three days. Patients with pain but minimal neurologic dysfunction receive a bolus of 10 mg followed by 16 mg daily initially in divided doses. The dose is gradually tapered once definitive treatment is well underway. Patients with small epidural lesions and a normal neurologic examination and those with relative contraindications to therapy may forgo the use of corticosteroids.

Surgery — The role of surgery in the management of ESCC has evolved. For many years, posterior decompression of ESCC via laminectomy was the initial approach to the patient with neurologic compromise. However, retrospective comparisons of case series of patients treated with laminectomy with or without RT versus RT alone revealed no advantage to the surgical approach. Similarly, a small randomized trial found no difference in outcome between patients undergoing laminectomy followed by RT versus RT alone.

The failure of laminectomy to improve outcome substantially is not surprising since the bulk of tumor is usually located in the vertebral body anterior to the thecal sac. Decompressing the spine posteriorly provides little access to anterior tumor and may further destabilize the spine.

Radical resection plus RT — Improvements in spinal instrumentation led to the investigation of more aggressive tumor resections followed by spinal reconstruction if necessary for patients with ESCC. Promising results were obtained by curetting out the tumor from the vertebral body and epidural space and then either bone grafting or using methylmethacrylate and instrumentation to achieve anterior stabilization. In view of the limited life expectancy of these patients and the frequent need for postoperative RT, methylmethacrylate has important advantages over bone grafting, since RT can begin one week following methylmethacrylate but must be deferred for at least six weeks to allow for fusion after bone grafting.

The superiority of initial tumor debulking was shown in a controlled trial comparing direct circumferential surgical decompression followed by RT (30 Gy over 10 days, starting within 14 days of surgery) with the same RT alone in 101 patients with metastatic ESCC. Enrolled patients had a known cancer diagnosis (patients with lymphoma or a primary spine tumor were excluded), a single level of cord compression, and could not be paraplegic for more than 48 hours. Both groups received the same initial dose of corticosteroids (dexamethasone 100 mg initially, then 24 mg every six hours), and either began radiation or underwent surgery within 24 hours of presentation.

The study was stopped after a planned interim analysis, when surgery followed by RT alone resulted in a significantly higher ambulatory rate compared to RT alone (84 versus 57 percent), and retained the ability to walk significantly longer than those treated with RT alone (median 122 versus 13 days). In addition, 10 of 16 patients who were initially unable to walk regained the ability to ambulate after surgery, compared to 3 of 16 treated with RT alone. Three of ten patients who failed to respond to RT regained the ability to walk after salvage surgery. The use of corticosteroids and opioid analgesics was also reduced in those treated surgically.

A higher complication rate with postradiation surgery was also noted in a series of 110 patients undergoing aggressive tumor resection for ESCC, 47 of whom had failed to respond to prior irradiation, and 44 percent of whom were nonambulatory. Postoperative complications occurred in 48 percent, and were related statistically to age over 65 years, prior treatment, and presence of paraparesis.

  Summary — In summary, results of a randomized trial indicates that radical resection followed by irradiation increases the likelihood of regaining the ability to walk and of maintaining ambulation following treatment, while those undergoing RT alone will have a lower functional status. However, careful selection is required to identify patients with an adequate life expectancy and good medical status who are candidates for this aggressive approach.

Radiation therapy — Prior to the use of radical surgery, RT was the initial treatment of choice for most patients with ESCC. It is still used for patients who are not considered surgical candidates, and for patients who have undergone surgical decompression.

The radiation portal is usually about 8 cm wide, centered on the spine, and extends one to two vertebral bodies above and below the epidural metastasis. The port is widened when the patient is known to have paraspinal extension of tumor.

  Dose and schedule — A variety of schedules have been used, ranging from single large fractions (eg, 8 Gy) to protracted courses (eg, 40 Gy divided in 20 fractions). At least three major studies looked at the RT dosing schedule:

  bullet In a randomized trial, 300 patients with an estimated life expectancy of less than six months were randomly assigned to either short course (two fractions of 8 Gy) or split course (30 Gy in 8 fractions RT. No differences were seen in either efficacy or toxicity, although concerns have been raised that both regimens were suboptimal.

  bullet One prospective study compared 30 Gy in 10 fractions (110 patients) versus 40 Gy in 20 fractions (104 patients), assigning treatment based upon appointment availability. Although not randomly assigned, the two groups were comparable in all parameters. There were no significant differences in either posttreatment motor function or the percentage of patients regaining ambulation.

  bullet A retrospective analysis of over 1300 patients looked at five treatment schedules ranging from a single treatment of 8 Gy to a total of 40 Gy in 20 fractions. All five regimens gave similar functional results, although the more protracted schedules were associated with fewer recurrences within the radiation field.

For patients with a relatively short life expectancy due to disease burden, short courses of radiation (eg, one or two fractions of 8 Gy) afford similar palliation, without the inconvenience of a more protracted course of radiation. However, surgical decompression followed by radiation appears to be superior in terms of regaining and maintaining neurologic function, in patients who are surgical candidates

  Response to treatment — RT leads to resolution of back pain in most patients, while pretreatment neurologic function is the strongest, and in some reports the only predictor of posttreatment neurologic function. In most series, 80 to 100 percent of patients treated while still ambulatory remain ambulatory at the conclusion of therapy. Approximately one-third of patients who are nonambulatory because of paraparesis regain the ability to walk with treatment, as do 2 to 6 percent of paraplegic patients. The likelihood of being ambulatory posttreatment is higher among patients whose motor deficits developed more slowly (over longer than two weeks versus less than one week prior to therapy), and, in nonambulatory patients, among those whose treatment is begun less than 12 hours after loss of ambulation. Among patients who require a urinary catheter before therapy, 20 to 40 percent will become catheter-free.

The second strongest predictor of response in patients unable to walk is the underlying tumor type. Patients who are paraparetic or even paraplegic have a much better chance of recovery if they have a radiosensitive tumor such as lymphoma, multiple myeloma, or breast or prostate cancer. Patients with these favorable histologies also are less likely to suffer local relapse of their epidural disease.

The degree of subarachnoid block produced by the tumor is also a predictor of posttreatment neurologic function; complete block is a poor prognostic sign.

The radiographic response to RT for ESCC was evaluated in a series of 64 patients who underwent repeat myelography between 20 and 170 days (median, 34 days) after the initiation of therapy. Regression of the epidural lesion, stable disease, and progression were noted in 47, 37, and 16 percent, respectively.

Median survival following diagnosis of ESCC is approximately six months. The outcome is better in walking patients and approximately one-half of patients surviving one year are still ambulatory at that time. RT is generally very well tolerated. When large segments of the spine are irradiated, bone marrow suppression and gastrointestinal toxicity (mucositis) may supervene.

  Recurrent ESCC — In a retrospective series, approximately 10 percent of patients treated with RT for ESCC eventually develop local recurrence. This value increases with time so that as many as one-quarter to one-half of patients surviving for more than one year experience local relapse. A prospective study followed 103 patients diagnosed between 1984 and 1988. Recurrent spinal epidural metastases (SEMs) occurred in 20 percent after a median interval of seven months, and a second recurrence occurred in 11 percent. Recurrent SEMs occurred as often at the same initial level (55 percent) as at a remote level (45 percent). About half of two-year survivors and nearly all three year survivors developed recurrent SEMs. Ambulatory status could be preserved in most patients even after the second recurrence.

Chemotherapy and surgery should be considered in patients with recurrent ESCC. In addition, there are data suggesting that a second course of spinal radiation can be given reasonably safely to patients with no other satisfactory options. Although repeat irradiation may result in a cumulative dose exceeding the reported radiation tolerance of the spinal cord (45 Gy in 2 Gy fractions), radiation myelopathy in this setting is an apparently infrequent occurrence. This may be attributable to repair of sublethal radiation damage between courses and/or the generally short survival of patients receiving repeat irradiation (median, five months) compared to the latency of radiation myelopathy. In addition, some radiation oncologists have argued that spinal cord tolerance has been defined too conservatively and probably is closer to 60 Gy in 2 Gy fractions.

Systemic therapy — There are no inherent reasons why chemotherapy cannot be used to treat ESCC. Unfortunately, the great majority of patients with ESCC have tumors that are not chemosensitive. In those patients with a sensitive tumor, chemotherapy is an attractive option because it can also treat tumor deposits elsewhere in the body.

Sensitive tumors in which ESCC has been treated successfully with chemotherapy include Hodgkin's disease [46], non-Hodgkin's lymphoma, neuroblastoma , germ cell neoplasms , and breast cancer . Hormonal manipulation has occasional documented benefit in ESCC from prostate cancer and breast cancer.

Novel approaches — Other therapies have been used in selected patients, including embolization and stereotactic radiosurgery or Intensity Modulated Radiation Therapy (IMRT). Embolization of highly vascular neoplasms has occasionally been used as an adjunct to surgery. In one case report, a patient with renal cell carcinoma had two episodes of geographically distinct ESCC successfully treated with transarterial embolization utilizing microcoils and alcohol particles.

Another approach has been to develop an extracranial system for delivering stereotactic radiosurgery. In one study, five patients with paravertebral metastases (including two with ESCC) which recurred after radiation were treated with stereotactic radiosurgery. The dose to the spinal cord was calculated to be less than or equal to3 Gy. Two patients died from systemic metastatic disease. In the three survivors, there was been CT- or MR-documented regression of the treated tumor with a decrease of thecal sac compression at a median follow-up of six months. No radiation toxicity was reported. Larger patient numbers and longer follow-up are required to support these encouraging observations.

PROGNOSIS — As noted above, median survival following the diagnosis of ESCC is approximately six months. The outcome is better in walking patients and approximately one-half of patients surviving one year are still ambulatory at that time. The prognosis is better in breast or prostate cancer and significantly worse in lung cancer. These general conclusions are supported by the following observations:

  bullet Studies suggest a median survival of nine to ten months for radiosensitive lesions (breast and prostate cancer, lymphoma, and multiple myeloma) compared to three months for lung cancer.

  bullet The median survival for patients ambulatory prior to RT is eight to ten months compared to two to four months for those who are nonambulatory. For those who remain nonambulatory at the conclusion of RT, the median survival is only one month.

Similar factors are found when survival after spinal metastasis is examined (irrespective of whether the patient has ESCC). Patients with radiosensitive tumors and a single spinal metastasis do best, while patients with lung cancer, multiple vertebral metastases, or visceral or brain metastases fare poorly.

In addition to improved therapies for established disease, two strategies in patients with known malignancy at risk for ESCC may reduce the incidence and severity of this complication: avoidance of treatment delay, and the administration of bisphosphonates.

Treatment delay — As mentioned above, the single most important prognostic factor is pretreatment neurologic status. In a review of 1392 patients presenting between 1963 and 1982, only 32 percent were ambulatory at the onset of therapy. Similar results were noted in a study published in 1998 in which only 33 percent of patients were ambulatory and 53 percent catheter-free at the time of therapy, indicating that little improvement has been made in the early detection of ESCC. These patients had back pain for a median of two months before ESCC was diagnosed and there was a ten day delay between the onset of neurologic symptoms and the start of therapy. The majority of patients deteriorated by at least one grade in motor or bladder function during the delay.

Thus, educating the patient and family as well as responsible health care providers about symptoms that warrant immediate evaluation, particularly otherwise unexplained back pain, might improve the neurologic outcome.

Bisphosphonates — Bisphosphonates such as pamidronate are of proven benefit in reducing pathologic fractures and bone pain in patients with multiple myeloma or breast cancer with lytic bony lesions. Although these studies were not designed to look at ESCC as an endpoint, it is possible that reducing bony progression and pathologic fractures will impact favorably on ESCC.

SUMMARY AND RECOMMENDATIONS — Prompt diagnosis and immediate treatment are critically important in the preservation of neurological function in patients with epidural spinal cord compression (ESCC). The use of corticosteroids in restoring neurologic function remains an important component of the initial management of these patients, even though the optimal dose and schedule remain uncertain.

In preliminary results of a randomized trial, initial surgical resection followed by radiation improved the chances of regaining and maintaining ambulation after treatment, for patients with ESCC due to metastatic non-hematologic malignancies. Furthermore, the complication rate of surgery is significantly lower if it is undertaken initially, rather than as a salvage procedure for the patient who fails initial radiotherapy (RT).

For patients in whom RT remains the treatment of choice, multiple studies showed that short courses of treatment (eg, a single dose of 8 Gy) are as effective as more protracted courses in providing palliation.Furthermore, systemic therapy continues to be important in those with sensitive tumors.

Corticosteroids — Corticosteroids are an integral part of the initial management of patients with ESCC.

  bullet We recommend moderate doses of steroids (eg, dexamethasone as a bolus of 10 mg intravenously followed by 16 mg daily orally in divided doses) for patients with minimal neurologic symptoms.

  bullet For those with paraparesis or paraplegia, we recommend high dose dexamethasone (eg, 96 mg intravenously followed by 24 mg four times daily for three days and then tapered over 10 days)
 

Surgery and RT — Surgery and/or radiation are the primary approaches to treat tumor compressing the spinal cord.

  bullet We recommend decompressive resection for patients with metastatic ESCC whose tumors are not reliably radiosensitive or chemosensitive and who have a limited disease burden and relatively favorable prognosis  Aggressive surgery should be followed by RT (ie, 30 Gy in 10 fractions). We do not recommend shorter courses of radiation in conjunction with decompressive surgery
  bullet Chemotherapy and hormonal therapy are useful in the treatment of ESCC in patients with sensitive tumors (eg, hematologic malignancies, germ cell tumors, breast cancer, prostate cancer).The advantages of surgery over RT to decompress the spinal cord are uncertain in these patients.

  bullet For patients with extensive systemic disease limiting survival to a few months or with severe fixed neurologic deficits, RT offers meaningful palliation in selected patients. We recommend short courses of radiation (eg, one or two treatments of 8 Gy) rather than a more protracted treatment schedule

  bullet Patients who develop recurrent ESCC can be retreated with RT