This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Husband, D J Articles by Romaniuk, C S Search for Related Content PubMed PubMed Citation Articles by Husband, D J Articles by Romaniuk, C S

MRI in the diagnosis and treatment of suspected malignant spinal cord compression

Clatterbridge Centre for Oncology, Clatterbridge Road, Bebington, Wirral, CH63 4JX UK

	Abstract

Top Abstract Introduction Patients and methods Results Discussion References

 
It remains unclear whether MRI is essential in all patients with suspected malignant spinal cord compression (MSCC), or whether some patients can be treated on the basis of plain radiographic findings and neurological examination. A prospective study was carried out of 280 consecutive patients with suspected MSCC, and the results of neurological examination plus plain radiographs were compared with MRI. 201 patients had MSCC (186 extradural, 5 intradural extramedullary and 10 intramedullary) and 11 patients had thecal sac compression without evidence of spinal cord compression. 25% of patients with MSCC had two or more levels of compression, 69% of these involving more than one region of the spine. A paraspinal mass was noted at the site of extradural spinal cord compression in 28%, and only one-third of these were detected on plain radiography. Focal radiographic changes and consistent neurology were present in 91 (33%) patients who had not had previous radiotherapy. MRI confirmed the presence of MSCC in 89/91 patients (specificity and positive predictive value of radiographic/clinical findings 98%) and the level of disease in all. MRI led to a change in the radiotherapy plan in 53% of patients (21% major change). The sensory level when present was four or more segments below the MRI level in 25/121 (21%) patients, and two or more levels above in 8/121 (7%) patients. Although focal radiographic abnormalities with consistent neurological findings, when present, accurately predicted the presence and level of MSCC, whole spine MRI is indicated in most patients with suspected MSCC because the additional information may alter the management plan. Treatment may be appropriately initiated on the basis of focal radiographic changes and consistent neurology if MRI is contraindicated or delayed, and in patients with a poor prognosis. In patients in whom there are no focal radiographic abnormalities and consistent neurological findings, urgent MRI is mandatory before radiotherapy is commenced.

	Introduction

Top Abstract Introduction Patients and methods Results Discussion References

 
Metastatic malignant spinal cord compression (MSCC) is a major cause of morbidity in patients with cancer [1]. The outcome of treatment is poor, with less than half of patients retaining or regaining the ability to walk, and about 40% requiring a permanent urinary catheter [2, 3]. Delay in diagnosis and treatment is common, and is associated with deterioration in functional status [4] and consequent poor functional outcome [5].

Prior to the advent of MRI, myelography (with or without CT) was the gold standard for diagnosis of MSCC [6]. MRI of the spine is now the imaging modality of choice because it provides non-invasive, multiplanar imaging of the spine with additional information about the presence of bone marrow involvement and paravertebral soft tissue masses [7–9]. Prior to the availability of MRI, it was considered by many that myelography was not required if clinical and plain radiographic information pointed to a lesion at one level [10–13]. Previous radiotherapy to the suspected level was considered an absolute indication for spinal cord imaging because of the need to consider a surgical approach, the risks associated with re-irradiation and the concern that diagnostic accuracy might be different in this group of patients, although little evidence is available to substantiate this latter point. Several studies have suggested that it is mandatory to undertake spinal cord imaging by myelography [14–16] or MRI [17] in all patients with suspected MSCC because of inexact correlation between plain radiographs or sensory levels and the level of spinal cord compression, the presence of two or more levels of compression in some patients, and because spinal cord imaging leads to changes in the radiation fields. It is axiomatic that myelography or MRI will provide additional information in most patients compared with that available from plain radiographs, and it is clear that spinal cord imaging will be necessary in a significant proportion of patients. In one series, 47% of patients were considered not to require myelography [12]. In a series reported previously from this centre [4], 35% of patients were deemed not to require spinal cord imaging. The issue is whether a subgroup of patients can be defined on the basis of clinical and plain radiographic findings in which the additional information provided by MRI does not significantly affect diagnosis, treatment or outcome. When plain radiographs indicate metastatic disease limited to two or three vertebral bodies in the area of symptoms, a radiation port extending two segments above and below these lesions is likely to encompass all epidural disease demonstrated by myelography [14]. This remains a relevant question in view of the continuing limited availability of MRI, the contraindications to MRI in a significant number of patients, and the poor prognosis of many patients with MSCC. In those patients with contraindications to MRI, myelographic skills will be increasingly unavailable because of the virtual obsolescence of the technique [18]. Even when MRI is available, logistical problems and the need for urgent treatment may mean that radiotherapy has to be commenced before MRI is undertaken [17]. Guidelines are required as to which patients, if any, may be properly treated in this manner. The argument for routine spinal cord imaging is heavily dependent on the incidence of multiple levels of compression, but the reported incidence has varied from 3% to 35% [2, 15–17, 19, 20]. Variation in the vertical extent of extradural disease has not been addressed in detail.

A prospective study was carried out (i) to assess the routine use of whole spine MRI in patients with suspected MSCC; (ii) to assess the possibility that a subgroup can be defined in whom spinal cord MRI is not necessary; and (iii) to define the distribution and extent of disease to allow definition of appropriate radiation portals in those patients in whom MRI cannot be carried out. The study has been limited to patients referred for radiotherapy. The role of MRI in pre-operative imaging in patients with MSCC has not been considered.

	Patients and methods

Top Abstract Introduction Patients and methods Results Discussion References

 
362 consecutive patients with suspected MSCC were assessed at the Clatterbridge Centre for Oncology over a 2-year period. 51 patients had undergone MRI at other hospitals prior to referral, showing MSCC in all cases; these patients were excluded because the number of patients scanned at the other hospitals with negative results was not known, which would have biased the assessment of the diagnostic tests. Of the remaining 311 patients, 31 did not undergo MRI for the following reasons: (i) it was felt to be clinically inappropriate (n=17; 14 because of poor general condition); (ii) MR scanner unavailability (n=5); (iii) the presence of orbital metal fragments (n=2); and (iv) inability of the patient to tolerate MRI (n=7). The remaining 280 patients underwent MRI and were studied prospectively. Characteristics of the patients are shown in Table 1.

The spine and spinal cord were imaged using a quadrature spine coil, with a field of view (FOV) of 370–420 mm. The examination consisted of a sagittal T₁ weighted spin echo sequence (TR/TE 500/25 ms, 4 mm slice thickness, 0.4 mm interslice gap, four signal averages) and a sagittal T₂ weighted fast spin echo sequence (TR/TE 2500/130 ms, 4 mm slice thickness, 0.4 mm interslice gap, six signal averages), obtained with a matrix of approximately 205x256 using a rectangular FOV. At levels of suspected spinal cord compression, axial T₁ weighted images (TR/TE 400/20 ms, FOV 250 mm, four signal averages, matrix 180x256; rectangular FOV, 4 mm slices at 0.4 mm interslice gap) were obtained to assess the degree of spinal cord compression. Additional T₁ weighted images following iv contrast medium were obtained if intramedullary or meningeal tumour was suspected.

Patients received treatment as appropriate, based on the clinical findings, plain radiographs and MR images as soon as they were available.

The plain radiographs and MRI studies were reported by two of the authors (CSR and KAG) and subsequently reviewed by DJH. On the plain radiographs, each vertebra was scored for the presence or absence of signs of malignant involvement (loss of pedicle, presence of osteolytic or osteosclerotic metastases, collapse and paraspinal mass). The findings were compared with the radiologists' reports, and any discrepancies were resolved by consensus. In the presence of a focal abnormality on the plain radiograph and neurological signs consistent with spinal cord compression at that level [21], a diagnosis of MSCC was viewed as certain, i.e. a positive test result. These patients formed the "MRI non-mandatory" group, with the exception of those patients who had had previous radiotherapy to the indicated level (the "MRI mandatory because of previous radiotherapy" group). Absence of focal plain radiographic changes of malignant involvement and/or inconsistent neurological findings was viewed as a negative test result for the diagnosis of MSCC and this group formed the "MRI mandatory" group. For the "MRI non-mandatory" group, mock radiotherapy fields were planned, allowing a two vertebral body margin above and below involved vertebrae and extending laterally to cover any paravertebral soft tissue masses visualized on the plain radiographs.

The MR images were then scored by segment for the presence of vertebral metastases, collapse, extradural disease, extradural spinal cord compression, paraspinal mass, intradural extramedullary spinal cord compression and intramedullary metastases. Spinal cord compression was defined as the presence of a mass that displaced, indented or led to complete loss of definition of the spinal cord or cauda equina [8]. Intradural disease that met the above definition was classified as spinal cord compression; presence of nodules within the dural space not meeting the above definition was classified as meningeal carcinomatosis. Intramedullary metastases were viewed as a subgroup of spinal cord compression. A mass that distorted the theca but did not indent or displace the cord was documented as thecal compression [8]. When an area of spinal cord compression or thecal compression extended over more than one adjacent vertebra, it was recorded as one site of compression. The maximal lateral extent of a paraspinal mass from the midline was measured on axial images. The findings were compared with the radiologists' reports and any discrepancies resolved by joint review of the images.

The additional information provided by MRI compared with plain radiography was recorded for all patients. The diagnostic performance of plain radiographs and neurological examination for the diagnosis of MSCC was compared with MRI; and specificity, sensitivity, and positive and negative predictive values were calculated [22, 23]. For the "MRI non-mandatory" group, further mock radiotherapy fields were planned using the MRI information. Errors in treatment that would have been made if treatment had been based on the radiographic and clinical information alone were assessed. The vertical extent of the main level of extradural disease was assessed in units of vertebral body length, and the extension of extradural disease beyond the superior and inferior borders of the involved vertebral bodies was recorded. The sensory level, when present, was compared with the level of compression on the MRI. The number of segments below the level of true compression was calculated from the extradural disease nearest to the sensory level. In the presence of multiple levels of spinal cord compression this correlation was not possible.

In patients without evidence of MSCC, other MRI findings were recorded. The final clinical diagnosis in these patients was subsequently assessed from the clinical records.

	Results

Top Abstract Introduction Patients and methods Results Discussion References

 
Neurological examination and plain radiograph assessment
Focal plain radiographic changes and consistent neurological findings were present in 104/280 (37%) of the patients. 58 of these patients had a sensory level. The plain radiographic diagnosis was based on the presence of vertebral collapse in 84 (81%) patients. 13 patients had received previous radiotherapy to the indicated level so that MRI was viewed as non-mandatory in 91 (33%) patients.

MRI findings
201 patients were found to have MSCC: 186 with extradural spinal cord compression, 5 with intradural extramedullary spinal cord compression (2 with additional intramedullary metastases) and 10 with intramedullary metastases. The remaining 79 patients had no evidence of MSCC; the MRI findings in these patients are shown in Table 2. 11 patients had evidence of thecal sac compression without evidence of spinal cord compression. Three patients with MSCC (all in the "MRI mandatory" group) had evidence of benign spinal disease, which may have contributed to their symptoms (lumbar spinal stenosis in two and prolapsed thoracic intervertebral disc in one). Other clinical diagnoses (not based on MRI findings) to explain the presenting symptoms in patients without MSCC were: meningeal carcinomatosis in one (CSF positive, MRI negative); lumbar–sacral plexus compression in one (diagnosed on pelvic CT scan); vertebral body metastases in four; brain metastases in four; malignant hypercalcaemia in two; carcinomatous neuropathy in four; cerebrovascular disease in two; and peripheral vascular disease in one.

Additional information provided by MRI, beyond that available from plain radiographs, is shown in Table 7 for all patients and for the "MRI non-mandatory" group. Additional findings, other than vertebral body metastases and paraspinal masses, were uncommon in the "MRI non-mandatory" group, with additional levels of extradural spinal cord compression in 5.5% and thecal compression in 6%. None of these patients had isolated thecal compression alone, intradural extramedullary spinal cord compression or intramedullary metastases.

Extent of extradural disease
The vertical extent of extradural disease was measured in units of vertebral body length in 182 patients with extradural spinal cord compression. It was not assessable in four patients with sacral disease. The distribution is shown in Figure 1. 31 of 182 (17%) patients had extradural disease extending over more than two vertebral body lengths. In the "MRI non-mandatory" group, 17 of 87 (20%) patients had disease of two or more vertebral bodies in length. Extension of extradural disease above and below the involved vertebral bodies (as demonstrated by MRI) is shown in Table 11. Extension more than one vertebral body above or below the involved vertebral body occurred in only 1% of patients.

View larger version (9K): [in this window] [in a new window]   Figure 1. Length of extradural disease (EDD) in patients with extradural spinal cord compression. Units are vertebral body units (VBU).

View larger version (8K): [in this window] [in a new window]   Figure 2. Relationship of sensory level to MRI level in (a) all patients and (b) "MRI mandatory" group.

	Discussion

Top Abstract Introduction Patients and methods Results Discussion References

 
MRI confirmed the diagnosis of MSCC in 72% of the patients studied. Although there are many causes of neurological signs and symptoms in patients with cancer [1, 24, 25], this figure is similar to the 72% [19] and 68% [17] reported in previous studies. Of particular note is the finding of a number of other diagnoses in patients without MSCC, and that non-neoplastic findings occurred in 25 (9%) patients (3 in addition to MSCC). MRI was normal in only 7% of patients. Leptomeningeal metastases were noted in 5% of patients, including 16% of those without MSCC.

Other series have not commented on the number of patients with intramedullary metastases, although the overall proportion of 5% is similar to the 3.4% reported [26] in the pre-MRI era.

25% of patients in this study with MSCC had compression at two or more levels. This supports the estimate of recent myelographic (30% [16], 35% [20]) and MRI (23% [19], 26% [17]) studies, but is considerably higher than some earlier estimates (3% [2], 8% [15]). This presumably reflects previous underestimation of the number of levels in patients who underwent lumbar myelography and who had a complete block but did not subsequently undergo cervical/cisternal myelography, and illustrates one of the advantages of MRI. In those patients with two or more levels of spinal cord compression, 70% occured in two or more regions of the spine, supporting the need for whole spine imaging rather than localized imaging. The primary tumour is not helpful in predicting which patients will have more than one site of compression, except that this is uncommon in tumours of haematological origin.

A paraspinal mass was noted at the site of extradural spinal cord compression in 28% of patients, and only one-third of these were visualized on plain radiographs. The only previous study noted an incidence of 18% in patients with extradural spinal cord compression, similarly noted that such masses were most common in patients with carcinoma of the bronchus (because in many cases the mass represents a primary tumour that has invaded the adjacent spine) and noted a low incidence from breast and prostate primaries [27].

The findings of thecal sac compression without evidence of spinal cord compression in 11 (4%) patients was not anticipated. Another recent study found that 6.3% of patients had thecal compression without spinal cord compression [17]. The significance of these findings is not well understood. In eight patients there were neurological abnormalities that correlated with the level of thecal compression, and in five patients there was no other cause of the signs. Thecal sac compression without demonstrable spinal cord compression may cause neurological abnormalities by a vascular mechanism. Thecal sac compression without spinal cord compression under these circumstances should not be viewed as an incidental finding, and would normally warrant treatment.

Lack of correlation between sensory levels and MRI levels is similar to that reported in previous studies [17, 19]. The sensory level was four or more vertebral segments below or two or more above the MRI level in 27% of patients. This indicates the futility of using sensory levels in the absence of focal plain radiographic abnormalities to plan radiotherapy fields, even in an emergency whilst awaiting MRI.

The occurrence of sensory levels below the MRI level is expected, and is the result of known neuroanatomy [6, 28]. The occurrence of sensory levels above the MRI level, however, is inexplicable on the basis of neuroanatomy but was observed in the Bristol study [17] and attributed to error in a retrospective case note review study. This is unlikely to be the explanation in the present prospective study with observation by a single experienced clinician. Posner [29] commented on the occurrence of this phenomenon in MSCC, and similar "distant effects" have been noted with transverse cord lesions [30]. This is therefore likely to be a real phenomenon, possibly due to compression of ascending veins [29] and secondary ischaemia, and should be taken into account when considering whether the neurological findings are consistent with the radiological level of disease.

The finding that MRI contributed significant additional information compared with the plain radiographic findings was anticipated. In the "MRI non-mandatory" group however, additional information other than additional vertebral body metastases was uncommon. In particular, extramedullary and intramedullary spinal cord compression was not found in this group, nor was any non-malignant condition. Two patients had malignant spinal root compression rather than MSCC at the site indicated by the plain radiographs, and management would therefore be unaltered. These two patients were the only ones in whom the diagnosis of MSCC was not correct in the "MRI non-mandatory" group, resulting in a specificity and positive predictive value of 98% for the finding of focal plain radiographic changes and consistent neurology. It is possible, therefore, to define a group representing 33% of all patients and 45% of those with MSCC in whom the diagnosis of MSCC can be made without recourse to MRI or myelography. The correct level would have been treated in all these patients. However, MRI led to a change in the radiotherapy plan in 53% of these patients. In five patients an additional level of MSCC was identified.

Whether asymptomatic additional levels of MSCC should be treated is controversial. Second episodes of extradural spinal cord compression (EDSCC) have occurred in 16% of patients [31]. However, in a further study of patients with second asymptomatic levels of EDSCC who were not irradiated, only 8% had a further episode of EDSCC at this level, and the incidence of a second episode of EDSCC was the same in those with a single level at first presentation and those with a second level that was not treated [20]. It was therefore recommended that a second asymptomatic level of EDSCC should not be treated. However, determining that a second level of EDSCC is in fact asymptomatic can be difficult and a pragmatic approach would be to treat additional levels in patients with a relatively good prognosis when this can be achieved without excessive morbidity.

The correct approach to additional levels of thecal sac compression without spinal cord compression is equally unclear. The same pragmatic approach can be applied as for additional levels of EDSCC.

Failure to treat the whole tumour in 54% of patients with a paraspinal mass (and 21% of all patients with MSCC) unless MRI or CT is performed has not been previously reported but has been inferred [27]. While treating the whole of the localized tumour seems, a priori, to be the correct approach, the outcome for patients with paraspinal mass is poor [27], and under these circumstances it is likely to be the response of the intraspinal disease rather than the extraspinal disease that determines outcome.

The margin above and below extradural disease was less than the conventional two vertebral body lengths [32] in 36% of the "MRI non-mandatory" group, but was less than one vertebral body length in only 3.5%; the whole of the extradural disease was not treated in one patient. However, no data are available to indicate the appropriate treatment margin, and a smaller margin may be adequate if planning is undertaken using MR images. Extradural disease extended over more than one vertebral body length in 53% of patients. Extension above the involved vertebrae occurred in 31% of patients and below the involved vertebrae in 19%. If a margin of two vertebral body lengths above and below extradural disease is required, then a margin of three vertebrae above and below the involved vertebrae on plain radiographs will achieve this in 95% of patients.

MRI clearly contributed additional valuable information in the planning of radiotherapy fields in the "MRI non-mandatory" group. This would support a policy of undertaking whole spine MRI in any patient with suspected MSCC [17]. However, it is not clear whether the changes made because of the MRI will improve the outcome, and therefore it is not certain that a policy of undertaking MRI on all patients will improve the prognosis for patients with MSCC. It is clear from this study however, that the diagnosis and level of MSCC can be safely established in patients in whom focal plain radiographic changes are associated with clinical findings consistent with spinal cord compression at the suspected level. This will allow treatment of those patients in whom MRI is contraindicated and allow treatment to be commenced urgently when MRI is not available. It may also be an appropriate approach in poor prognosis patients. In patients in whom focal plain radiographic and consistent clinical findings are absent, urgent spinal cord imaging is mandatory before radiotherapy is instituted. This implies the need for MRI to be available on an emergency basis in all centres treating MSCC.

	Acknowledgments

 
We are grateful for the invaluable contribution of the radiographers in the X-ray and MRI departments at Clatterbridge Centre for Oncology. Ms H J Moore (Superintendent Radiographer) assisted in patient recruitment as well as supervising the taking of the radiographs. Mrs J Whiteway helped with data collection and prepared the manuscript. Dr D J Husband and Dr C S Romaniuk acknowledge support from the Clatterbridge Cancer Research Trust, who also funded the MR scanner.

Received for publication April 17, 2000. Revision received August 14, 2000. Accepted for publication August 31, 2000.

	References

Top Abstract Introduction Patients and methods Results Discussion References

 

1. Posner JB. Neurological complications of cancer. Philadelphia, PA: FA Davies, 1995.
2. Bach F, Larsen BH, Rohde K, Borgensen SE, Gjerris F, Boge-Rasmussen T, et al. Metastatic spinal cord compression. Occurrence, symptoms, clinical presentations and prognosis in 398 patients with spinal cord compression. Acta Neurochir (Wien) 1990;107:37–43.[Medline]
3. Gilbert RW, Kim J, Posner JB. Epidural spinal cord compression from metastatic tumour: diagnosis and treatment. Ann Neurol 1978;3:40–51.[Medline]
4. Husband DJ. Malignant spinal cord compression: prospective study of delays in referral and treatment. BMJ 1998;317:18–21.[Abstract/Free Full Text]
5. Findlay GFG. Adverse effects of the management of spinal cord compression. J Neurol Neurosurg Psychiatry 1984;47:761–8.[Abstract/Free Full Text]
6. Byrne TN, Waxman SG. Spinal cord compression. Philadelphia, PA: FA Davies, 1990.
7. Li KC, Poon PY. Sensitivity and specificity of MRI in detecting malignant spinal cord compression andin distinguishing malignant from benign compression fractures of vertebrae. Magn Reson Imaging 1988;6:547–56.[Medline]
8. Williams MP, Cherryman GR, Husband JE. Magnetic resonance in suspected metastatic spinal cord compression. Clin Radiol 1988;40:286–90.
9. Sarpel S, Sarpel G, Yu E, Hyder S, Kaufman B, Hindo W, et al. Early diagnosis of spinal epidural metastases by magnetic resonance imaging. Cancer 1987;59:1112–6.[Medline]
10. Involvement of the vertebral column and spinal cord. In: Henson RA, Urich H. Cancer and the nervous system. Oxford: Blackwell, 1982:120–54.
11. Warlow C. Handbook of neurology. Oxford: Blackwell, 1992.
12. Stark RJ, Henson RA, Evans SJW. Spinal metastases. A retrospective survey from a general hospital. Brain 1982;105:189–213.[Free Full Text]
13. Podd TJ, Carpenter DS, Baughan CA, Percival D, Dyson P. Spinal cord compression: prognosis and implication for treatment fractionation. Clin Oncol (R Coll Radiol) 1992;4:341–4.
14. Rodichok LD, Ruckdeschel JC, Harper GR, Cooper G, Prevosti L, Fernando L, et al. Early detection and treatment of spinal epidural metastases: the role of myelography. Ann Neurol 1986;20:696–702.[Medline]
15. Calkins AR, Olson MA, Ellis JH. Impact of myelography on the radiotherapeutic management of malignant spinal cord compression. Neurosurgery 1986;19:614–6.[Medline]
16. Van der Sande JJ, Kroger R, Boogerd W. Multiple spinal epidural metastases; an unexpectedly frequent finding. J Neurol Neurosurg Psychiatry 1990;53:1001–3.[Abstract/Free Full Text]
17. Cook AM, Lau TN, Tomlinson MJ, Vaidya M, Wakeley CJ, Goddard P. Magnetic resonance imaging of the whole spine in suspected malignant spinal cord compression: impact on management. Clin Oncol (R Coll Radiol) 1998;10:39–43.
18. Stevens JM. Imaging of the spinal cord. In: Hughes RAC, editor. Neurological investigations. London: British Medical Journal, 1997:141–78.
19. Offenbacher H. The diagnostic impact of magnetic resonance imaging on the evaluation of suspected spinal cord disease. Wien Klin Wochenschr 1992;104:589–93.[Medline]
20. Helweg-Larsen S, Hansen SW, Sorensen PS. Second occurrence of symptomatic metastatic spinal cord compression and findings of multiple spinal epidural metastases. Int J Radiat Oncol Biol Phys 1995;33:595–8.[Medline]
21. Gurusinghe NT. Spinal cord compression and spinal cord tumours. In: Critchley E, Eisen A, editors. Diseases of the spinal cord. London: Springer-Verlag, 1992:351–408.
22. Counting data. In: Fisher LD, van Belle G. Biostatistics: a methodology for the health sciences. New York, NY: Wiley, 1993.
23. Determining the presence or absence of disease. In: Lang TA, Secic M. How to report statistics in medicine. Philadelphia, PA: American College of Physicians, 1997.
24. Gilbert MR, Grossman SA. Incidence and nature of neurological problems in patients with solid tumours. Am J Med 1986;81:951–4.[Medline]
25. Choucair AK. Myelopathies in the cancer patient: incidence, presentation, diagnosis and management. Oncology 1991;5:25–31.[Medline]
26. Edelson RN, Reck MD, Posner JB. Intramedullary spinal cord metastases. Neurology 1972;22:1222–31.[Free Full Text]
27. Kim RY, Smith JW, Spencer S, Meredith RF, Salter MM. Malignant epidural spinal cord compression associated with a paravertebral mass: its radiotherapeutic outcome on radiosensitivity. Int J Radiat Oncol Biol Phys 1993;27:1079–83.[Medline]
28. Jamieson DRS, Teasdale E, Willison HS. False localising signs in the spinal cord. BMJ 1996;312:243–4.[Free Full Text]
29. Spinal metastases. In: Posner JB. Neurological complications of cancer. Philadelphia, PA: FA Davies, 1995:111–42.
30. Haymaker W. Bing's local diagnosis in neurological disease. St Louis, MO: Mosby, 1969.
31. Kaminski HJ, Diwan VG, Ruff RL. Second occurrence of spinal epidural metastases. Neurology 1991;41:744–6.[Abstract/Free Full Text]
32. Linstadt DE, Spinal cord. In: Leibel SA, Philips TL, editors. Textbook of radiation oncology. Philadelphia, PA: WB Saunders, 1998.

This article has been cited by other articles:

				S. M. Dy, S. M. Asch, A. Naeim, H. Sanati, A. Walling, and K. A. Lorenz Evidence-Based Standards for Cancer Pain Management J. Clin. Oncol., August 10, 2008; 26(23): 3879 - 3885. [Abstract] [Full Text] [PDF]

				N Walji, A K Chan, and D R Peake Common acute oncological emergencies: diagnosis, investigation and management Postgrad. Med. J., August 1, 2008; 84(994): 418 - 427. [Abstract] [Full Text] [PDF]

				D. A. Loblaw, J. Perry, A. Chambers, and N. J. Laperriere Systematic Review of the Diagnosis and Management of Malignant Extradural Spinal Cord Compression: The Cancer Care Ontario Practice Guidelines Initiative's Neuro-Oncology Disease Site Group J. Clin. Oncol., March 20, 2005; 23(9): 2028 - 2037. [Abstract] [Full Text] [PDF]

				D Summers, D A Collie, and P Levack Assessment of MSCC using MRI Br. J. Radiol., October 1, 2001; 74(886): 977 - 977. [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Husband, D J Articles by Romaniuk, C S Search for Related Content PubMed PubMed Citation Articles by Husband, D J Articles by Romaniuk, C S