Journal of Neuro-Oncology 13: 231-237, 1992. © 1992 Kluwer Academic Publishers. Printed in the Netherlands.
Clinical Study
Comparison of myelography combined with postmyelographic spinal CT and MRI in suspected metastatic disease of the spinal canal
Susanne Helweg-Larsen, 1 Aase Wagner, 2 Lasse Kjaer, 2 Anni Johnsen, 3 Jens Boesen, 3 Torben Palner 2 and Per Soelberg SOrensen 1
~Dept. of Neurology, Rigshospitalet, Copenhagen, 2Dept. of Magnetic Resonance, Hvidovre Hospital, Copenhagen, ~Dept. of Neuroradiology, Rigshospitalet, Copenhagen
Key words': metastatic disease, spinal cord, MRI, myelography Abstract
Comparison between myelography (MY) and magnetic resonance imaging (MRI) was carried out in 36 patients with clinical suspicion of spinal cord or root compression due to metastatic disease in the spinal canal. In 3 patients metastatic lesions were visualized on MY but not on MRI, while there were no cases with a negative MY and a positive MRI. In 44% of the cases MY alone or combined with postmyelographic CT (pm-CT) showed a larger tumor extension than did MRI, while the opposite occurred in 25%. As for detection of bony metastases and tumor masses localized outside the spine there was no difference between MRI and MY + pm-CT. The results indicate that the choice between MRI and MY + pm-CT still can be based on the availability and quality of the procedure at a given institution.
Introduction
Compression of the spinal cord due to intraspinal metastases is next to cerebral metastases the most common neurological complication in patients with solid tumors, estimated to occur in about 5% of cancer patients [1, 2]. The condition constitutes one of the most important neurological emergencies and requires immediate treatment with radiation therapy or surgical intervention. Before treatment can be instituted, it is mandatory to determine the extent of the intraspinal mass lesion. Diagnostic imaging of intraspinal masses has relied on myelography (MY) [3, 4], including cervical contrast injection in case of a complete subarachnoid block after lumbar injection. In a recent publication we have shown, that in case of a total block after lumbar injection a postmyelographic spinal computer tomography (pm-CT) usually gives enough information about the extent of intraspinal
masses to make a cervical contrast injection unnecessary [5]. Numerous reports have described the potential of magnetic resonance imaging (MRI) as a valuable method for imaging of the spine and spinal cord, and have compared MRI with MY, eventually combined with pm-CT [6-15], but only some of these studies, included patients with a malignant disease and suspected intraspinal metastases [6- 9, 11, 14]. Comparison of MRI and conventional MY in a series of 31 cancer patients with compression myelopathy showed that MY was superior to MRI [16], while in two other studies with 22 and 21 patients, respectively [17, 18], MRI appeared superior to MY. In still another study of 70 patients no significant difference could be shown between the two diagnostic modalities [19]. Although recently published studies with gadolinium-DTPA enhanced MRI have shown that the use of a paramagnetic agent adds further informa-
232 tion to the imaging of spinal tumors [20, 21], there seems to be different opinions about which examination to choose for the diagnosis of intraspinal metastases. The study was undertaken in order to give additional information about the diagnostic yield of MY combined with pm-CT as compared with MRI in the evaluation of cancer patients with suspected metastatic disease of the spinal canal.
Patients and methods
Thirty six patients (27 women, 9 men, aged 25-87 years, median 67 years) with a known malignant disease of the breast (n = 20), the lung (n = 5), the prostate gland (n = 3), or other organs (n = 8) were referred for MY because of clinical suspicion (radiating back pain, myelopathy, or radiculopathy) of spinal cord compression or root compression in the spinal canal, i.e. intraspinal metastatic disease. In 25 patients a pm-CT was performed immediately after MY, while for technical reasons 11 patients did not have a pm-CT. Within 1-8 days (median 3 days) after MY all 36 patients underwent a spinal MR-examination. MY was carried out after lumbar injection of 10 ml iohexol (Omnipaque R300), with the patient recumbent on the left side in Trendelenburg position under constant fluoroscopic control. Pm-CT was performed on a Siemens Somatom DR 2 scanner. Scan time was 7 s per image and slice thickness was 8mm. Scan parameters were 520mAs., 125 kV, and zoom-factor 5.5. The area of the CT scan was defined from the lower level of the spinal Table 1. Number of patients with total and partial intraspinal contrast block on myelography (MY) compared with postmyelographic CT (pm-CT) in 25 patients.
grade of block on pm-CT grade of block on MY
total partial none (normal pm-CT)
total
partial
none (normal MY)
2 9 0
0 11 2
0 0 1
canal block seen on MY to a normal subarachnoid space cranial to the myelographic block. A supplementary cervical MY was found necessary in 4 patients. For this purpose, an additional 3-4ml iohexol was injected by a lateral C 1-2 puncture. Complete subarachnoid block on MY was defined as a total block of the contrast medium above the tumor after the patient had been placed in Trendelenburg position for 5-10 rain. Partial block was defined as a thinning or a defect in the contrast in relation to the tumor. In addition to evaluation of the degree of block, the pm-CT included an assessment of the bony metastases, i.e. whether the metastatic growth was confined to the vertebral body, had penetrated the posterior margin of the vertebral body, or had extended into the pedicles or neural arches. Tumor outside the spine (pre-/ paravertebral and paraspinal) was also described. MRI was performed on a Siemens Magnetom unit operating at a field strength of 1.5T. Each examination was carried out in the sagittal and axial planes by means of a surface-coil. The procedure included an overlapping multislice technique based on a double spin echo (SE) pulse sequence (TR/TE 1.800/28, 90 ms). Slice thickness was 6 mm and matrix 256 x 256. Moreover, a zoom factor of 1.2 was applied. The examination time was 4560 rain, but in some cases the poor condition of the patient demanded a modified examination strategy. The area investigated by MRI included the region suspected to harbor intraspinal pathology as judged by the clinician. If the intraspinal tumor completely obliterated the subarachnoid space the compression was defined as total, while it was defined as partial if the subarachnoid space was only partly obliterated. Description of the bony metastases included the vertebral bodies, the pedicles, Table 2. Number of patients with intraspinal metastatic lesions on myelography (MY) compared with MRI.
MRI
positive negative
MY positive
negative
30 3
0 3
233 and the laminae. T u m o r masses localized pre-/paravertebrally and paraspinally were also described. Evaluation of the myelograms and pm-CT scans were made by two neuroradiologists, and the M R images were evaluated by two MR-specialists. The evaluations were performed blindly in regard to the other diagnostic modalities. Comparison of the examination modalities was based on non-parametric statistics (the sign test).
Results
MY showed a complete block of the contrast medium in 15 cases and a partial block in 18 cases. Three patients had a normal myelogram. When MY was combined with pm-CT in 25 patients, the distribution of total and partial block changed as demonstrated in Table 1. There was no passage of the contrast medium on pm-CT in only 2 of the 25 patients. The subarachnoid space was totally obliterated in 24 cases on MRI and partially obliterated in 6 cases. In 6 cases there was no intraspinal affection at all. Comparison of MY with MRI (Table 2) showed that in 3 patients an extradural lesion without cord compression could be detected on MY but not on MRI, but one of these 3 patients had bony metastases on both pm-CT and MRI. There were no cases with a negative MY and a positive MRI. An example of complete obliteration of the subarachnoid space on MY and MRI, but with contrast passage on pm-CT is shown in the Figure. The extension of the intraspinal affection on MY or on MY + pm-CT as compared with the extension on MRI is seen in Table 3. In 16 of the 36 patients MY/MY + pm-CT showed a larger exten-
sion than MRI, while in 9 patients the MRI demonstrated a larger extension. In 11 cases no difference was found between the two modalities (p > 0.2). Bony metastases in the spine were detected on MRI in 21 patients and on pm-CT in 20 of the 25 patients in whom CT was performed (Table 4a). The number of vertebral bodies affected was higher on MRI in 11 patients, higher on pm-CT in 2 patients, and equal in 12 patients (p = 0.05) (Table 4b). T u m o r localization outside the spine was detected in 13 cases on MRI and in 16 cases on pm-CT (Table 5a). In Table 5b is seen the extension of the metastases in cranial/caudal direction. In 11 patients the extension was larger on pm-CT than on MRI, while in 8 the extension was larger on MRI (p > 0.2). No differences in the results could be correlated to differences in the origin of the primary tumors.
Discussion
In the present study we found 3 cases of intraspinal extradural tumor growth that was seen on MY but not on MRI, while the opposite did not occur. In all 3 cases MY showed small defects in the contrast
Table 4a. Number of patients with bony spinal metastases on MRI and postmyelographic CT (pm-CT) in 25 patients with intraspinal metastases.
pm-CT
positive negative
MRI positive
negative
20 1
0 4
Table 3. Extension of intraspinal metastases on myelography (MY) alone or combined with postmyelographic CT (pm-CT) as compared with MRI.
Table 4b. Comparison of the number of vertebral bodies showing metastases on MRI and postmyelographic CT (pm-CT) in 25 patients.
tumor extension
No. of patients
tumor extension
No, of patients
MY/MY + pm-CT > MRI MY/MY + pm-CT = MRI MY/MY + pm-CT < MRI
16 11 9
pm-CT > MRI pm-CT = MRI pm-CT < MRI
2 12 11
234 Table 5a. Number of patients with tumor localization outside the spine on postmyelographic CT (pm-CT) compared with MRI.
pm-CT
positive negative
MRI positive
negative
10 3
6 6
Table 5b. Extension of metastases outside the spine in cranial/ caudal direction as described on postmyelographic CT (pm-CT) compared with MRI.
tumor extension
No. of patients
pm-CT > MRI pm-CT = MRI pm-CT < MRI
11 6 8
medium encompassing only one nerve root - but in concordance with the neurological symptoms and signs. These results confirm previous studies which indicate that small epidural masses and root compressions are better detected by MY than MRI [16, 19]. Similar observations have been made in case of disk herniation, where MRI can not be considered equivalent to MY or MY + pm-CT for the detection of small or lateral disk herniations [8]. The reason for that is partly the relatively large slice thickness generally applied in the MRI studies [6, 9, 11, 22], where lesions smaller than 0.5-1.0cm may be missed on MRI. Furthermore, just like nerve roots, small subarachnoid metastatic nodules can move with different patient positions or with CSF pulsation, and these movements may degrade the delineation of the lesions. For the same reason comparison of lesions in follow-up MRI can be difficult, since the lesions may appear in different
positions on consecutive examinations. Recent evidence indicates that identification of small lesions on MRI is markedly improved by using the paramagnetic contrast agent gadolinium-DTPA [20, 21, 23, 24]. In the diagnosis of intraspinal metastases it is important to delineate the extension of the tumor in order to determine the correct radiation port. In our study tumor extension was larger on MY/ MY + pm-CT in 44%, while the opposite occurred in 25%. This difference was not statistically significant. A few studies have compared MRI with MY [1518, 25] in patients with metastatic cancer, but none have made direct comparisons of the extension of the intraspinal lesion. Our findings may, at least partly, be explained by better spatial resolution in the CT-examinations. The use of a larger slice thickness in CT might lead to partial volume effects and consequently overestimation of the tumor extension. On the other hand, the accuracy of the tumor delineation by MRI may be influenced by respiratory artifacts, CSF pulsation, and the absence of Tcweighted imaging in this study. Some authors have stressed the ability of MRI to show the upper limit of the intraspinal process in case of total block of the contrast medium on lumbar M Y a problem we have shown can be solved by pm-CT in most cases [5]. Our data indicate that MRI is equal to MY + pm-CT in detecting spinal metastases. This is in agreement with earlier reports by Colman [26] and Sarpel [14]. Moreover, Avrahami [27] found focal or diffuse vertebral MRI abnormalities confirmed to be metastases by needle biopsy in 21 patients without any abnormalities on CT. We could confirm previous observations [7] that there is no significant difference between MRI and MY + pm-CT in detecting metastases outside the
Fig. 1. Myelography, postmyelographic CT (pm-CT), and MRI in a 66 year old woman with breast cancer and intraspinal metastatic disease at the level Th9. a) Lateral myelogram demonstrating complete subarachnoid block (arrow). b) Pm-CT showing intracorporal tumor masses invading the spinal canal displacing the spinal cord to the left. Arrows indicate faint amount of subarachnoid contrast medium. c) Sagittal MRI demonstrating vertebral collapse and extensive intraspinal tumor invasion (TR/TE 1.800/90 ms). d) Axial MRI showing intraspinal tumor invasion and cord displacement to the left (TR/TE 1.800/28 ms).
235
236 spine. It has been pointed out that MRI is better than MY alone in detecting paraspinal metastases [17, 28]. Better spatial resolution on CT is probably the reason that CT is as good as MRI in the present study. In conclusion, we could not confirm a higher diagnostic yield of MRI than of MY combined with pm-CT. In fact, MY + pm-CT was better in depicting small epidural masses. In cooperative patients excellent spinal images can be made with MRI, and MRI is non-invasive, free of radiation exposure, and can be done on an outpatient basis. Patients who are not able to remain still due to pain frequently have technically inadequate examinations [27] and are better evaluated with MY. MY also enables chemical and cytologic study of the spinal fluid. MRI is a diagnostic modality in evolution. Further technical development including fast imaging techniques will reduce the examination time. The use of paramagnetic contrast agents will also improve the accuracy of the examination, and eventually MRI may become superior to MY + pm-CT in diagnosing intraspinal metastases. We suggest that in general the choice of MRI or MY/MY + pm-CT for imaging of the spine should be based upon the availability, imaging time, and quality of the procedure at a given institution. The ability of the patient to remain still is particularly important in MRI.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Acknowledgements This work was granted by the Danish Cancer Society, Torben Linnemanns Foundation for Cancer Research, the Danish Hospital Foundation for Medical Research, region of Copenhagen, the Faroe Islands, and Greenland, Foundation for Research in Neurology and FondsbOrsvekseler Henry Hansen and wife Carla Hansen, born Westergaards foundation.
12.
13.
14.
15.
References 1. Bach FA, Larsen BH, Rohde K, B~rgensen SE, Gjerris F, BOge-Rasmussen T, Agerlin N, Rasmusson B, Stjernholm
16.
P, Sorensen PS: Metastatic spinal cord compression. Occurrence, symptoms, clinical presentations and prognosis in 398 patients with spinal cord compression. Acta Neurochir 107: 3%43, 1990 Barron KD, Hirano A, Araki S, Terry RD: Experiences with metastatic neoplasms involving the spinal cord. Neurology 9: 91-106, 1959. Constans JP, de Divitiis E, Donzelli R, Spaziante R, Meder JF, Haye C: Spinal metastases with neurological manifestations. Review of 600 cases. J Neurosurg 59:111-118, 1983 S0rensen PS, BCrgensen SE, Rohde K, Rasmusson B, Bach F, BCge-Rasmussen T, Stjernholm P, Larsen BH, Agerlin N, Gjerris F: Treatment and prognosis of metastatic spinal cord compression. Cancer 65: 1502-1508, 1990 Boesen J, Johnsen A, Helweg-Larsen S, S0rensen PS: The diagnostic value of spinal computer tomography in patients with intraspinal metastases causing complete block on myelography. Acta Radiologica 32: 1-2, 1991 Aichner F, Poewe W, Rogalsky W, Walln6fer K, Willeit J: Magnetic resonance imaging in the diagnosis of spinal cord disease. J Neurol Neurosurg Psychiatry 48: 1220-1229, 1985 Karnaze MG, Gado MH, Sartor KJ, Hodges FJ: Comparison of MR and CT myelography in imaging the cervical and thoracic spine. AJNR 8: 983-989, 1987 Masaryk T J, Modic MT, Geisinger MA, Standefer J, Hardy RW, Boumphrey F, Duchesneau PM: Cervical myelopathy: A comparison of magnetic resonance and myelography. J Comput Assist Tomogr 10: 184-194, 1986 Modic MT, Weinstein MA, Pavlicek W, Starnes DL, Duchesneau PM, Boumphrey F, Hardy RJ: Nuclear magnetic resonance imaging of the spine. Radiology 148: 757-762, 1983 Modic MT, Masaryk T, Boumphrey F, Goormastic M, Bell G: Lumbar herniated disk disease and canal stenosis: prospective evaluation by surface coil MR, CT, and myelography. AJR 147: 75%765, 1986 Norman D, Mills CM, Brant-Zawadzki M, Yeates A, Crooks LE, Kaufman L: Magnetic resonance imaging of the spinal cord and canal: Potentials and limitations. A JR 141: 1147-1152, 1983. Packer R J, Zimmerman RA, Bilanuik LT: Magnetic resonance imaging in the evaluation of treatment-related central nervous system damage. Cancer 58: 635-640, 1986 Quencer RM, Sheldon J J, Post MJD, Diaz RD, Montalvo BM, Green BA, Eismont FJ: Magnetic resonance imaging of the chronically injured cervical spinal cord. AJNR 7: 45%464, 1986 Sarpel S, Sarpel G, Yu E, Hyder S, Kaufman B, Hindo W, Ezdinli E: Early diagnosis of spinal-epidural metastasis by magnetic resonance imaging. Cancer 59: 1112-1116, 1987 Sze G, Abramson A, Krol G, Liu D, Amster J, Zimmerman RD, Deck MDF: Gadolinium-DTPA in the evaluation of intradural extramedullary spinal disease. AJNR 9: 153163, 1988 Hagenau C, Grosh W, Currie M, Wiley RG: Comparison
237
17.
18.
19.
20.
21.
22.
23.
of spinal magnetic resonance imaging and myelography in cancer patients. J Clin Oncol 5: 1663-1669, 1987 Smoker WRK, Godersky JC, Knutzon RK, Keyes WD, Norman D, Bergman W: The role of MR imaging in evaluating metastatic spinal disease. AJNR 8: 901-908, 1987 Williams MP, Cherryman GR, Husband JE: Magnetic resonance imaging in suspected metastatic spinal cord compression. Clinical Radiology 40: 286-290, 1989 Carmody RF, Yang P J, Seeley GW, Seeger JF, Unger EC, Johnson JE: Spinal cord compression due to metastatic disease: Diagnosis with MR imaging versus myelography. Radiology 173: 225-229, 1989 Clarke CE, Hawnaur J, Prvulovich E, Yuill GM, Bannister CM: Case report: Magnetic imaging with gadoliniumDTPA enhancement in the differentiation of benign from neoplastic cord cavitation. Clinical Radiology 41: 214-216, 1990 Rothwell CI, Jaspan T, Worthington BS, Holland IM: Gadolinium-enhanced magnetic resonance imaging of spinal turnouts. British Journal of Radiology 62: 1067-1074, 1989 Han JS, Benson JE, Yoon YS: Magnetic resonance imaging in the spinal column and craniovertebral junction. Radiol Clin North Am 22: 805-827, 1984 Fredericks RK, Elster A, Walker FO: Gadolinium-en-
24. 25.
26.
27.
28.
hanced MRI: A superior technique for the diagnosis of intraspinal metastases. Neurology 39: 734-742, 1989 Zimmerman RA, Bilaniuk LT: Imaging of tumors of the spinal canal. Radiol Clin North Am 26: 965-1007, 1988 Li KC, Poon PY: Sensitivity and specificity of MRI in detecting malignant spinal cord compression and in distinguishing malignant from benign compression fractures of vertebrae. MRI 6: 547-556, 1988 Colman LK, Porter BA, Redmond III J, Olson DO, Stimac GK, Dunning DM, Friedl KE: Early diagnosis of spinal metastases by CT and MR studies. J CAT 12: 423-426, 1988 Avrahami E, Tadmor R, Dally O, Hadar H: Early MR demonstration of spinal metastases in patients with normal radiographs and CT and radionuclide bone scans. J CAT 13: 598-602, 1989 Godersky JC, Smoker WRK, Knutzon R: Use of magnetic resonance imaging in the evaluation of metastatic spinal disease. Neurosurg 21: 676-680, 1987
Address for offprints: Susanne Helweg-Larsen, M.D., University Department of Neurology, Rigshospitalet, Blegdamsvej 9, DK 2100 Copenhagen ~, Denmark
Clinical Study
Comparison of myelography combined with postmyelographic spinal CT and MRI in suspected metastatic disease of the spinal canal
Susanne Helweg-Larsen, 1 Aase Wagner, 2 Lasse Kjaer, 2 Anni Johnsen, 3 Jens Boesen, 3 Torben Palner 2 and Per Soelberg SOrensen 1
~Dept. of Neurology, Rigshospitalet, Copenhagen, 2Dept. of Magnetic Resonance, Hvidovre Hospital, Copenhagen, ~Dept. of Neuroradiology, Rigshospitalet, Copenhagen
Key words': metastatic disease, spinal cord, MRI, myelography Abstract
Comparison between myelography (MY) and magnetic resonance imaging (MRI) was carried out in 36 patients with clinical suspicion of spinal cord or root compression due to metastatic disease in the spinal canal. In 3 patients metastatic lesions were visualized on MY but not on MRI, while there were no cases with a negative MY and a positive MRI. In 44% of the cases MY alone or combined with postmyelographic CT (pm-CT) showed a larger tumor extension than did MRI, while the opposite occurred in 25%. As for detection of bony metastases and tumor masses localized outside the spine there was no difference between MRI and MY + pm-CT. The results indicate that the choice between MRI and MY + pm-CT still can be based on the availability and quality of the procedure at a given institution.
Introduction
Compression of the spinal cord due to intraspinal metastases is next to cerebral metastases the most common neurological complication in patients with solid tumors, estimated to occur in about 5% of cancer patients [1, 2]. The condition constitutes one of the most important neurological emergencies and requires immediate treatment with radiation therapy or surgical intervention. Before treatment can be instituted, it is mandatory to determine the extent of the intraspinal mass lesion. Diagnostic imaging of intraspinal masses has relied on myelography (MY) [3, 4], including cervical contrast injection in case of a complete subarachnoid block after lumbar injection. In a recent publication we have shown, that in case of a total block after lumbar injection a postmyelographic spinal computer tomography (pm-CT) usually gives enough information about the extent of intraspinal
masses to make a cervical contrast injection unnecessary [5]. Numerous reports have described the potential of magnetic resonance imaging (MRI) as a valuable method for imaging of the spine and spinal cord, and have compared MRI with MY, eventually combined with pm-CT [6-15], but only some of these studies, included patients with a malignant disease and suspected intraspinal metastases [6- 9, 11, 14]. Comparison of MRI and conventional MY in a series of 31 cancer patients with compression myelopathy showed that MY was superior to MRI [16], while in two other studies with 22 and 21 patients, respectively [17, 18], MRI appeared superior to MY. In still another study of 70 patients no significant difference could be shown between the two diagnostic modalities [19]. Although recently published studies with gadolinium-DTPA enhanced MRI have shown that the use of a paramagnetic agent adds further informa-
232 tion to the imaging of spinal tumors [20, 21], there seems to be different opinions about which examination to choose for the diagnosis of intraspinal metastases. The study was undertaken in order to give additional information about the diagnostic yield of MY combined with pm-CT as compared with MRI in the evaluation of cancer patients with suspected metastatic disease of the spinal canal.
Patients and methods
Thirty six patients (27 women, 9 men, aged 25-87 years, median 67 years) with a known malignant disease of the breast (n = 20), the lung (n = 5), the prostate gland (n = 3), or other organs (n = 8) were referred for MY because of clinical suspicion (radiating back pain, myelopathy, or radiculopathy) of spinal cord compression or root compression in the spinal canal, i.e. intraspinal metastatic disease. In 25 patients a pm-CT was performed immediately after MY, while for technical reasons 11 patients did not have a pm-CT. Within 1-8 days (median 3 days) after MY all 36 patients underwent a spinal MR-examination. MY was carried out after lumbar injection of 10 ml iohexol (Omnipaque R300), with the patient recumbent on the left side in Trendelenburg position under constant fluoroscopic control. Pm-CT was performed on a Siemens Somatom DR 2 scanner. Scan time was 7 s per image and slice thickness was 8mm. Scan parameters were 520mAs., 125 kV, and zoom-factor 5.5. The area of the CT scan was defined from the lower level of the spinal Table 1. Number of patients with total and partial intraspinal contrast block on myelography (MY) compared with postmyelographic CT (pm-CT) in 25 patients.
grade of block on pm-CT grade of block on MY
total partial none (normal pm-CT)
total
partial
none (normal MY)
2 9 0
0 11 2
0 0 1
canal block seen on MY to a normal subarachnoid space cranial to the myelographic block. A supplementary cervical MY was found necessary in 4 patients. For this purpose, an additional 3-4ml iohexol was injected by a lateral C 1-2 puncture. Complete subarachnoid block on MY was defined as a total block of the contrast medium above the tumor after the patient had been placed in Trendelenburg position for 5-10 rain. Partial block was defined as a thinning or a defect in the contrast in relation to the tumor. In addition to evaluation of the degree of block, the pm-CT included an assessment of the bony metastases, i.e. whether the metastatic growth was confined to the vertebral body, had penetrated the posterior margin of the vertebral body, or had extended into the pedicles or neural arches. Tumor outside the spine (pre-/ paravertebral and paraspinal) was also described. MRI was performed on a Siemens Magnetom unit operating at a field strength of 1.5T. Each examination was carried out in the sagittal and axial planes by means of a surface-coil. The procedure included an overlapping multislice technique based on a double spin echo (SE) pulse sequence (TR/TE 1.800/28, 90 ms). Slice thickness was 6 mm and matrix 256 x 256. Moreover, a zoom factor of 1.2 was applied. The examination time was 4560 rain, but in some cases the poor condition of the patient demanded a modified examination strategy. The area investigated by MRI included the region suspected to harbor intraspinal pathology as judged by the clinician. If the intraspinal tumor completely obliterated the subarachnoid space the compression was defined as total, while it was defined as partial if the subarachnoid space was only partly obliterated. Description of the bony metastases included the vertebral bodies, the pedicles, Table 2. Number of patients with intraspinal metastatic lesions on myelography (MY) compared with MRI.
MRI
positive negative
MY positive
negative
30 3
0 3
233 and the laminae. T u m o r masses localized pre-/paravertebrally and paraspinally were also described. Evaluation of the myelograms and pm-CT scans were made by two neuroradiologists, and the M R images were evaluated by two MR-specialists. The evaluations were performed blindly in regard to the other diagnostic modalities. Comparison of the examination modalities was based on non-parametric statistics (the sign test).
Results
MY showed a complete block of the contrast medium in 15 cases and a partial block in 18 cases. Three patients had a normal myelogram. When MY was combined with pm-CT in 25 patients, the distribution of total and partial block changed as demonstrated in Table 1. There was no passage of the contrast medium on pm-CT in only 2 of the 25 patients. The subarachnoid space was totally obliterated in 24 cases on MRI and partially obliterated in 6 cases. In 6 cases there was no intraspinal affection at all. Comparison of MY with MRI (Table 2) showed that in 3 patients an extradural lesion without cord compression could be detected on MY but not on MRI, but one of these 3 patients had bony metastases on both pm-CT and MRI. There were no cases with a negative MY and a positive MRI. An example of complete obliteration of the subarachnoid space on MY and MRI, but with contrast passage on pm-CT is shown in the Figure. The extension of the intraspinal affection on MY or on MY + pm-CT as compared with the extension on MRI is seen in Table 3. In 16 of the 36 patients MY/MY + pm-CT showed a larger exten-
sion than MRI, while in 9 patients the MRI demonstrated a larger extension. In 11 cases no difference was found between the two modalities (p > 0.2). Bony metastases in the spine were detected on MRI in 21 patients and on pm-CT in 20 of the 25 patients in whom CT was performed (Table 4a). The number of vertebral bodies affected was higher on MRI in 11 patients, higher on pm-CT in 2 patients, and equal in 12 patients (p = 0.05) (Table 4b). T u m o r localization outside the spine was detected in 13 cases on MRI and in 16 cases on pm-CT (Table 5a). In Table 5b is seen the extension of the metastases in cranial/caudal direction. In 11 patients the extension was larger on pm-CT than on MRI, while in 8 the extension was larger on MRI (p > 0.2). No differences in the results could be correlated to differences in the origin of the primary tumors.
Discussion
In the present study we found 3 cases of intraspinal extradural tumor growth that was seen on MY but not on MRI, while the opposite did not occur. In all 3 cases MY showed small defects in the contrast
Table 4a. Number of patients with bony spinal metastases on MRI and postmyelographic CT (pm-CT) in 25 patients with intraspinal metastases.
pm-CT
positive negative
MRI positive
negative
20 1
0 4
Table 3. Extension of intraspinal metastases on myelography (MY) alone or combined with postmyelographic CT (pm-CT) as compared with MRI.
Table 4b. Comparison of the number of vertebral bodies showing metastases on MRI and postmyelographic CT (pm-CT) in 25 patients.
tumor extension
No. of patients
tumor extension
No, of patients
MY/MY + pm-CT > MRI MY/MY + pm-CT = MRI MY/MY + pm-CT < MRI
16 11 9
pm-CT > MRI pm-CT = MRI pm-CT < MRI
2 12 11
234 Table 5a. Number of patients with tumor localization outside the spine on postmyelographic CT (pm-CT) compared with MRI.
pm-CT
positive negative
MRI positive
negative
10 3
6 6
Table 5b. Extension of metastases outside the spine in cranial/ caudal direction as described on postmyelographic CT (pm-CT) compared with MRI.
tumor extension
No. of patients
pm-CT > MRI pm-CT = MRI pm-CT < MRI
11 6 8
medium encompassing only one nerve root - but in concordance with the neurological symptoms and signs. These results confirm previous studies which indicate that small epidural masses and root compressions are better detected by MY than MRI [16, 19]. Similar observations have been made in case of disk herniation, where MRI can not be considered equivalent to MY or MY + pm-CT for the detection of small or lateral disk herniations [8]. The reason for that is partly the relatively large slice thickness generally applied in the MRI studies [6, 9, 11, 22], where lesions smaller than 0.5-1.0cm may be missed on MRI. Furthermore, just like nerve roots, small subarachnoid metastatic nodules can move with different patient positions or with CSF pulsation, and these movements may degrade the delineation of the lesions. For the same reason comparison of lesions in follow-up MRI can be difficult, since the lesions may appear in different
positions on consecutive examinations. Recent evidence indicates that identification of small lesions on MRI is markedly improved by using the paramagnetic contrast agent gadolinium-DTPA [20, 21, 23, 24]. In the diagnosis of intraspinal metastases it is important to delineate the extension of the tumor in order to determine the correct radiation port. In our study tumor extension was larger on MY/ MY + pm-CT in 44%, while the opposite occurred in 25%. This difference was not statistically significant. A few studies have compared MRI with MY [1518, 25] in patients with metastatic cancer, but none have made direct comparisons of the extension of the intraspinal lesion. Our findings may, at least partly, be explained by better spatial resolution in the CT-examinations. The use of a larger slice thickness in CT might lead to partial volume effects and consequently overestimation of the tumor extension. On the other hand, the accuracy of the tumor delineation by MRI may be influenced by respiratory artifacts, CSF pulsation, and the absence of Tcweighted imaging in this study. Some authors have stressed the ability of MRI to show the upper limit of the intraspinal process in case of total block of the contrast medium on lumbar M Y a problem we have shown can be solved by pm-CT in most cases [5]. Our data indicate that MRI is equal to MY + pm-CT in detecting spinal metastases. This is in agreement with earlier reports by Colman [26] and Sarpel [14]. Moreover, Avrahami [27] found focal or diffuse vertebral MRI abnormalities confirmed to be metastases by needle biopsy in 21 patients without any abnormalities on CT. We could confirm previous observations [7] that there is no significant difference between MRI and MY + pm-CT in detecting metastases outside the
Fig. 1. Myelography, postmyelographic CT (pm-CT), and MRI in a 66 year old woman with breast cancer and intraspinal metastatic disease at the level Th9. a) Lateral myelogram demonstrating complete subarachnoid block (arrow). b) Pm-CT showing intracorporal tumor masses invading the spinal canal displacing the spinal cord to the left. Arrows indicate faint amount of subarachnoid contrast medium. c) Sagittal MRI demonstrating vertebral collapse and extensive intraspinal tumor invasion (TR/TE 1.800/90 ms). d) Axial MRI showing intraspinal tumor invasion and cord displacement to the left (TR/TE 1.800/28 ms).
235
236 spine. It has been pointed out that MRI is better than MY alone in detecting paraspinal metastases [17, 28]. Better spatial resolution on CT is probably the reason that CT is as good as MRI in the present study. In conclusion, we could not confirm a higher diagnostic yield of MRI than of MY combined with pm-CT. In fact, MY + pm-CT was better in depicting small epidural masses. In cooperative patients excellent spinal images can be made with MRI, and MRI is non-invasive, free of radiation exposure, and can be done on an outpatient basis. Patients who are not able to remain still due to pain frequently have technically inadequate examinations [27] and are better evaluated with MY. MY also enables chemical and cytologic study of the spinal fluid. MRI is a diagnostic modality in evolution. Further technical development including fast imaging techniques will reduce the examination time. The use of paramagnetic contrast agents will also improve the accuracy of the examination, and eventually MRI may become superior to MY + pm-CT in diagnosing intraspinal metastases. We suggest that in general the choice of MRI or MY/MY + pm-CT for imaging of the spine should be based upon the availability, imaging time, and quality of the procedure at a given institution. The ability of the patient to remain still is particularly important in MRI.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Acknowledgements This work was granted by the Danish Cancer Society, Torben Linnemanns Foundation for Cancer Research, the Danish Hospital Foundation for Medical Research, region of Copenhagen, the Faroe Islands, and Greenland, Foundation for Research in Neurology and FondsbOrsvekseler Henry Hansen and wife Carla Hansen, born Westergaards foundation.
12.
13.
14.
15.
References 1. Bach FA, Larsen BH, Rohde K, B~rgensen SE, Gjerris F, BOge-Rasmussen T, Agerlin N, Rasmusson B, Stjernholm
16.
P, Sorensen PS: Metastatic spinal cord compression. Occurrence, symptoms, clinical presentations and prognosis in 398 patients with spinal cord compression. Acta Neurochir 107: 3%43, 1990 Barron KD, Hirano A, Araki S, Terry RD: Experiences with metastatic neoplasms involving the spinal cord. Neurology 9: 91-106, 1959. Constans JP, de Divitiis E, Donzelli R, Spaziante R, Meder JF, Haye C: Spinal metastases with neurological manifestations. Review of 600 cases. J Neurosurg 59:111-118, 1983 S0rensen PS, BCrgensen SE, Rohde K, Rasmusson B, Bach F, BCge-Rasmussen T, Stjernholm P, Larsen BH, Agerlin N, Gjerris F: Treatment and prognosis of metastatic spinal cord compression. Cancer 65: 1502-1508, 1990 Boesen J, Johnsen A, Helweg-Larsen S, S0rensen PS: The diagnostic value of spinal computer tomography in patients with intraspinal metastases causing complete block on myelography. Acta Radiologica 32: 1-2, 1991 Aichner F, Poewe W, Rogalsky W, Walln6fer K, Willeit J: Magnetic resonance imaging in the diagnosis of spinal cord disease. J Neurol Neurosurg Psychiatry 48: 1220-1229, 1985 Karnaze MG, Gado MH, Sartor KJ, Hodges FJ: Comparison of MR and CT myelography in imaging the cervical and thoracic spine. AJNR 8: 983-989, 1987 Masaryk T J, Modic MT, Geisinger MA, Standefer J, Hardy RW, Boumphrey F, Duchesneau PM: Cervical myelopathy: A comparison of magnetic resonance and myelography. J Comput Assist Tomogr 10: 184-194, 1986 Modic MT, Weinstein MA, Pavlicek W, Starnes DL, Duchesneau PM, Boumphrey F, Hardy RJ: Nuclear magnetic resonance imaging of the spine. Radiology 148: 757-762, 1983 Modic MT, Masaryk T, Boumphrey F, Goormastic M, Bell G: Lumbar herniated disk disease and canal stenosis: prospective evaluation by surface coil MR, CT, and myelography. AJR 147: 75%765, 1986 Norman D, Mills CM, Brant-Zawadzki M, Yeates A, Crooks LE, Kaufman L: Magnetic resonance imaging of the spinal cord and canal: Potentials and limitations. A JR 141: 1147-1152, 1983. Packer R J, Zimmerman RA, Bilanuik LT: Magnetic resonance imaging in the evaluation of treatment-related central nervous system damage. Cancer 58: 635-640, 1986 Quencer RM, Sheldon J J, Post MJD, Diaz RD, Montalvo BM, Green BA, Eismont FJ: Magnetic resonance imaging of the chronically injured cervical spinal cord. AJNR 7: 45%464, 1986 Sarpel S, Sarpel G, Yu E, Hyder S, Kaufman B, Hindo W, Ezdinli E: Early diagnosis of spinal-epidural metastasis by magnetic resonance imaging. Cancer 59: 1112-1116, 1987 Sze G, Abramson A, Krol G, Liu D, Amster J, Zimmerman RD, Deck MDF: Gadolinium-DTPA in the evaluation of intradural extramedullary spinal disease. AJNR 9: 153163, 1988 Hagenau C, Grosh W, Currie M, Wiley RG: Comparison
237
17.
18.
19.
20.
21.
22.
23.
of spinal magnetic resonance imaging and myelography in cancer patients. J Clin Oncol 5: 1663-1669, 1987 Smoker WRK, Godersky JC, Knutzon RK, Keyes WD, Norman D, Bergman W: The role of MR imaging in evaluating metastatic spinal disease. AJNR 8: 901-908, 1987 Williams MP, Cherryman GR, Husband JE: Magnetic resonance imaging in suspected metastatic spinal cord compression. Clinical Radiology 40: 286-290, 1989 Carmody RF, Yang P J, Seeley GW, Seeger JF, Unger EC, Johnson JE: Spinal cord compression due to metastatic disease: Diagnosis with MR imaging versus myelography. Radiology 173: 225-229, 1989 Clarke CE, Hawnaur J, Prvulovich E, Yuill GM, Bannister CM: Case report: Magnetic imaging with gadoliniumDTPA enhancement in the differentiation of benign from neoplastic cord cavitation. Clinical Radiology 41: 214-216, 1990 Rothwell CI, Jaspan T, Worthington BS, Holland IM: Gadolinium-enhanced magnetic resonance imaging of spinal turnouts. British Journal of Radiology 62: 1067-1074, 1989 Han JS, Benson JE, Yoon YS: Magnetic resonance imaging in the spinal column and craniovertebral junction. Radiol Clin North Am 22: 805-827, 1984 Fredericks RK, Elster A, Walker FO: Gadolinium-en-
24. 25.
26.
27.
28.
hanced MRI: A superior technique for the diagnosis of intraspinal metastases. Neurology 39: 734-742, 1989 Zimmerman RA, Bilaniuk LT: Imaging of tumors of the spinal canal. Radiol Clin North Am 26: 965-1007, 1988 Li KC, Poon PY: Sensitivity and specificity of MRI in detecting malignant spinal cord compression and in distinguishing malignant from benign compression fractures of vertebrae. MRI 6: 547-556, 1988 Colman LK, Porter BA, Redmond III J, Olson DO, Stimac GK, Dunning DM, Friedl KE: Early diagnosis of spinal metastases by CT and MR studies. J CAT 12: 423-426, 1988 Avrahami E, Tadmor R, Dally O, Hadar H: Early MR demonstration of spinal metastases in patients with normal radiographs and CT and radionuclide bone scans. J CAT 13: 598-602, 1989 Godersky JC, Smoker WRK, Knutzon R: Use of magnetic resonance imaging in the evaluation of metastatic spinal disease. Neurosurg 21: 676-680, 1987
Address for offprints: Susanne Helweg-Larsen, M.D., University Department of Neurology, Rigshospitalet, Blegdamsvej 9, DK 2100 Copenhagen ~, Denmark
