AJR:168, March 1997 807

MR Imaging of Tubercular SpinalArachnoiditis

Aseem Sharma1Mayank GoyalNalin K. MishraVivek GuptaShailesh B. Gaikwad

OBJECTIVE. The purpose of our study was to examine the spectrum of abnormalities

seen on MR imaging in patients with tubercular spinal arachnoiditis.

MATERIALS AND METHODS. A retrospective analysis of MR findings in 22

cases of tubercular spinal arachnoiditis was carried out. The diagnosis had been established

on the basis of clinical features, evidence of associated tubercular meningitis or of tubercu-

lar spondylitis. and CSF analysis.

RESULTS. Nineteen (86%) patients had involvement of more than one spinal region,

with the dorsal regiomi being most commonly involved. CSF showed increased signal inten-

sity on TI-weighted images in 17 (77%) patients. leading to complete loss of cord-CSF

interface in seven patients and shaggy cord outline in 10 patients. As suggested by

increased signal intensity on T2-weighted images. we saw cord involvement in 18 (82%)

patients. Three of these patients had evidence of cord cavitation. Other findings seen on

unenhanced images were CSF loculations in five patients. nodules in subarachnoid space in

six patients. and clumping of cauda equina nerve roots in six patients. Contrast-enhanced

studies were available in 20 patients. Meningeal enhancement was seen in 16 (80%) of 20

patients, and nerve root enhancement was seen in six (30%) patients. Cord enhancement

was seen in four (20% ) of 20 patients. Enhancement was observed along the surface of the

cord in two of these patients. whereas the other two patients showed central enhancement.

Associated findings were tubercular spondylitis in two patients. basal exudate in eight

patients. and intracranial granulomas in five patients.

CONCLUSION. MR imaging revealed several pathologic changes that occur in

patients with tubercular spimial arachnoiditis and, hence. niay play an important role in the

diagnosis of this emitity.

Received July 8, i996; accepted after revisionSeptember 5, 1996.

1�Jl authors: DepartmentofNeuroradiology, Neurosciences

Centre,All India lnstituteofMedicalSciences,Ansan Nagar,New Delhi-i10029, India. Address correspondence toM. Goyal.

AJR1997;i68:807-812

O36i�-8O3XJ97/i683-8O7

© American Roentgen Ray Society

T uberculosis is an inip()rtamit.

potentially treatable cause of spi-

nal arachnoiditis. Tubercular

infection is common in third-world countries

and is showing arm increasing trend in the

West because ofthe high prevalence of AIDS

[1. 21. Frequently associated radiculomyelitis

makes tubercular arachnoiditis distinct from

other causes. which include surgery. sub-

arachnoid bleeding. and intrathecal adminis-

tration of anesthetic agents or myelographic

contrast mnedia. The diagnosis of tubercular

spinal arachnoiditis is usually based on clini-

cal features. CSF analysis. evidence of tuber-

culosis elsewhere in the body (especially

meningitis). and characteristic myelographic

findings. Myelography has played an impor-

tant role in the diagnosis of spinal arachnoidi-

tis. However. recent literature has shown MR

imaging to be useful in diagnosing spinal

arachnoiditis 13-51. Most of these studies

have been carried out in patients with spinal

ar’achnoiditis of mioninfectious causes. The lit-

erature on MR iniaging findings of tubercular

spinal arachnoiditis is scant I I . 6-8]. To our

knowledge. the only study dealing with a size-

able number of patients 171 includes patients

with all forms of intraspinal tuberculosis. includ-

ing arachnoiditis. intraniedullary tuberculonias,

and epidural abscess. The aim of the present

study was to highlight the spectrum of MR

imaging findings seen in patients with a specific

fonn of intra.spinal tuberculosis. tubercular spi-

nal arachnoiditis.

Materials and Methods

MR imaging examinations were pertuirmed in 22

patietits with spinal arachnoiditis of tubercular origin

during the last 3 years at our institution. A retrospec-

tive review t)f MR iniaging studies of these patients

was carried out. Patients included IS niales and

sevemi females, who were 7-70 years old (mean ±

SD. 32.7 ± I 7.0 years tild). None of the patients was

ininiunocompromnised. The diagmiosis of tubercular

spinal arachnoiditis was based on clinical features

(onset of niyelopathy in a known case of tubercular

menimigitis). CSF analysis. evidence of tuberculosis

elsewhere in the body. and iniaging findings.

All MR mmages were obtaimied on a system

equipped with a I .5-T supercomiducting magnet

( Magnetomii: Siernemis. Erlamigen. Germany ). All

studmes mmicluded ‘otgtttal T I -weighted (TR rangelFE

during or after completion of antitubercular

Fig. 1-24-year-old man who experienced giddiness for i year and quadriparesis for 2 months.A, Sagittal Ti-weighted MR image shows loss of normal cord-CSF interface and large CSFloculation in C1-C2 region leading to compression. Note normal signal intensityof Ioculated CSF.B, Sagittal 12-weighted MR image shows widening (arrow) and increased signal intensity of cord.C, Contrast-enhanced MR image shows enhancement of meninges and CSF loculation margins. Note enhancing exudate in prepontine cistern (arrow).

808 AJR:168, March 1997

Sharma et aI.

range, 450-800/15-20) and T2-weighted (2000-

4500/60-90) extended field-of-view (FOV) imagesof the whole spine (FOV, 450-500: slice thickness,

3-4 mm; interslice gap, 0. 1� matrix size, 256-384 x512; two or three excitations). Additional sagittal

and axial images were acquired for the region ofinterest using the surface spine coil. Axial scans

were acquired using a 192-256 x 256 matrix with

rectangular FOV. 4- to 5-mm slice thickness. and

two or three acquisitions. T2-weighted axial scans

were acquired using either fast spin-echo or gradi-

ent-echo sequences. Gadopentetate dimeglumine-

enhanced studies (Magnevist. 0.2 mI/kg: Schering.

Berlin, Germany) were available in 20 patients.

All MR imaging studies were retrospectively

reviewed. Specific points noted were region ofinvolvement, extent of involvement, signal intensity

of CSF on TI-weighted MR images, cord-CSF

interface, presence of CSF loculations, nodular

lesions in the subarachnoid space. nerve root clump-

ing, cord expansion, signal-intensity alterations of the

cord on T2-weighted MR images. pattern of

enhancement after administration of contrast me-

dium. and associated vertebral involvement. CranialCT images. MR images, and chest radiographs also

were studied when available. Clinical data of these

patients were reviewed, and details ofclinical presen-

tation and results of other investigations (CSF. chest

radiography. and CT and MR imaging of the brain)

were noted.

The extent of CSF signal alteration. contrast

enhancement of the meninges. or clumping of nerve

roots (whichever was more extensive) was used to

define the extent of spinal involvement. Loss of

cord-CSF interface on TI-weighted images was dis-

tinguished fromii cord widening by mioting the pres-

ence ofCSF around the cord in T2-weighted images.Evaluation of the signal intensity of the CSF on TI-weighted images was on the basis of visual impres-

sion and comparison with the adjacent spinal cord.

The presence of low signal intensity with well-defined margins within the cord on Tl-weighted

images and corresponding high signal intensity simi-lar to that of CSF on T2-weighted images was taken

as evidence of cord cavitation.

Results

Eight patients who were previously diag-

nosed with tubercular meningitis (based on

clinical features, CSF findings. enhancing

exudate in the basal cisterns on CT or MR

imaging, and coexi�tent tuberculosis else-

where) developed features of myelopathy

therapy. Paraparesis was the most common

presenting symptom in 16 of 22 cases. Other

symptoms included quadriparesis, monopare-

sis, urinary incontinence, backache, and

radiculopathy. The duration of symptoms

ranged from 3 days to 2 years 6 months.

Associated chronic meningitis was present

in I 2 patients (Figs. I and 2). Eight of these

patients were previously diagnosed cases of

tubercular meningitis. In four patients, the

presence of enhancing exudate in the basal

cisterns (revealed on cranial CT and MR

imaging) was taken as evidence of associated

intracranial meningitis. The results of CSF

analysis were available for 20 patients. Nine-

teen of these patients had an elevated CSF

protein level (range, 120 mg/dl to >1 g/dl), a

moderately decreased glucose level, and

increased cells in the CSF (predominantly

lymphocytes). India ink studies were done

and CSF was cultured for all CSF samples to

exclude fungal infection. One patient had nor-

mal CSF. The CSF analysis in this patient,

however, had been done 2 months before the

onset of myelopathy (for suspected tubercular

MVCOb(k’tC!ill!?l tI�ber(’u/osi.s had been cul- CT or MR imaging studies of the brain were

tured fromii the sputum in the past. One patient available for 1 1 patients. Two patients did not

had evidence of hydropneumothorax in the have any intracranial abnormalities, eight

chest radiograph. Two patients had associated patients showed evidence of enhancing exu-

MR Imaging of Tubercular Spinal Arachnoiditis

AJR:i68, March 1997 809

Fig. 2.-28-year-old man presenting with fever andbackache of 1 year’s duration and recent onset ofquadriparesis.A, Sagittal Ti-weighted MR image shows loss of CSFsurrounding cord.B, Sagittal 12-weighted MR image shows wideningand increase in signal intensity. Note compression ofcord by extramedullary lesion at C2 level (arrow).C, Contrast-enhanced MR image revealing enhance-ment of meninges and extramedullary mass.D, Contrast-enhanced Ti-weighted sagittal MR imageofwhole spine shows destruction and enhancement ofDio and Dii vertebral bodies with involvement of in-tervening disk (arrow). Note presence of extraduralabscess leading to compression of cord.E, Contrast-enhanced sagittal MR image through brainshows multiple ring-enhancing lesions in frontal lobe.Enhancing exudate is seen filling prepontine and su-prasellar cisterns.

meningitis). The diagnosis in this patient was

suggested on the basis of imaging findings of

spinal arachnoiditis. chronic meningitis. and

associated pulmonary tuberculosis. Radio-

logic evidence of pulmonary tuberculosis was tubercular spondylitis with epidural abscess date in basal cisterns suggestive of associated

seen in five patients. In two of these patients. (Fig. 2D). which was confirmed at surgery. chronic mneningitis (Figs. IC and 2E), five

Fig. 3-20-year-old man with backache and myelopathy.A, Ti-weighted sagittal MR image of spine shows long-segment cord cavitation.B, Contrast-enhanced parasagittal MR image shows subtle alteration of cord contour suggesting presence of CSF locu-lation (arrow). No enhancement of meninges or loculation is seen.

Sharma et al.

810 AJR:168, March 1997

patients had intracranial granulomas (Fig. 2E).

and one patient had associated ventriculitis.

A long segment of the spine was frequently

involved, spanning more than five vertebral

bodies in all patients. In 18 patients. the

inflammatory process involved more than one

spinal region. Involvement of both the cervi-

cal and dorsal spine was seen in nine patients.

Disease was localized to the dorsal spine in

two patients, involved the dorsal and lumbar

spine in six patients. and involved the luni-

bar or lumbosacral region in three patients. In

two patients, the disease process extended

throughout the length of the spinal column.

Increased signal intensity of the CSF

around the cord on TI-weighted images was

seen in 17 patients, leading to complete loss

of cord-CSF interface in seven patients (Fig.

lA) and shaggy, indistinct cord outline in 10

patients. Five patients had MR imaging evi-

dence of CSF loculations in the subarachnoid

space that were causing compression of the

cord. Two patients had multiple loculations.

The wall of the loculation could be identified

only in one patient in whom enhancement

around the loculation also was present after

adrninistrationofcontrastmedium(Fig. I). In

the remainder of the patients, the margins of

the loculations did not enhance after adminis-

tration ofcontrast medium (Fig. 3).

In six patients. MR imaging showed nodu-

Iar lesions in the subarachnoid space. of which

four lesions caused cord compression. All of

these lesions were better seen on T2-weighted

images, wherein these lesions were sharply

contrasted against the bright signal intensity of

the CSF. One of the lesions enhanced after

injection of contrast medium (Fig. 2).

Thickened and clumped nerve roots in the

cauda equina region were identified in six

patients. We cannot comment on root adhe-

sions in the cervical and dorsal regions

because MR imaging failed to reveal individ-

ual nerve roots in these regions.

Cord changes represented by increased

signal intensity on T2-weighted images were

seen in I 8 patients (Figs. I B and 2B). In three

of these patients, MR imaging showed cord

cavitation (Fig. 3). Expansion ofthe cord was

seen in I 2 patients. Extrinsic cord compres-

sion was seen in I I patients, caused by CSF

loculations in five patients, nodular lesions in

the subarachnoid space in four patients, epi-

dural abscess in two patients. and thickened

meninges in one patient (Fig. 4).

Enhanced studies were available in 20

patients. Sixteen patients showed enhancement

of spinal meninges (Figs. lC. 2C. and 4). four

patients showed cord enhancement. and six

patients showed nerve r(x)t emiliamicemiient. Of

the four patiemits showing cord emiliamicemiiemit.

the surface ot the cord enhamiced imi two

patients (Fig. 5) amid cemitral cord emiliancemiiemit

was seen imi the other two patients (Fig. 6.

Discussion

Spinal arachmioiditis refers to arm intlamiimiia-

tory process of the leptomiiemiinges that has �‘ar-

ious causes, imicluding mmitectiomis, imitrathecal

administration of chemical agents (e.g.. tmies-

thetic agents. miiyelographic comitrast miiedia.

and antibiotics ). subarachmioid liemomi’liage.

trauma, surgical scar. and disk dmsease 9).

Tuberculosis has beemi reported to be the miiost

common infectious cause I 101. Early diagmiosis

of these cases is important because timely

institution of proper medical treatmiiemit niay

ensure good recovery I I I (. Frequemit imivo1�e-

ment of nerve roots and the spinal cord dilThr-

entiates tubercular arachmioiditis fmomii aracli-

noiditis of other causes I I 2 . Accordingly. the

term tubercular radiculoniyelitis has heemi sug-

gested. Although the disease niay o)ccur as �t

primary event, niore than 50’/ of the cases are

associated wmth miiemiingitis o)r. occasiomially.

tubercular spondylitis (imi our series. this value

was 55� ) I I 2). Imi early stages of the disease

� miiemitmigeal imillamiimiiatio)mi leads to con-gestmo)n and i mlflamumvtato)ry exudate. Inflamnma-

tiomi lreq(memitl\ imivolves the spmnal cord and

mierve roots :ind miiay head to formiiatiomi of local-

ized imitramiiedul lamv or extraniedul lary granu-

Iomiias. Adliesiomis o)ccumTimig betweeti (ibm-

coated nerve tOots amid miiemiimiges lead to) root

tetliemimig. blockage of CSF f’lO)��’, amid formna-

tio)Ii 0)1 CSF locuLitiomis. Demise fibrous adhe-

510)115 miiav develop later mithe disease process.

Vascu lam i mivo)Ivemiient. caused either by

mnllamiimiiat io)mi or by comiipression ( by the

tihrotms tisstie ). tiiil\ result itt cord ischernia.

i milarctmom#{236}.amid ca� tat io)ti.

I)iagmiosts 0)1 tubercular arachmioiditms is usu-

ally based on climimcal features, associatiomi of

tubercular miiemiingltms. amid CSF amialysis. Acid-

last hacmllm have rarely heemi idemitmfied I I I I and

s�eme 10)1 seemi mi amiy 0)! ()1t� patiemits. CSF anal-

ysis usujll� reveals ati elevated protein level, a

reduced glucose level. amid ami increase mi the

miimmher o)l cells ( mnaimihv Ivmiiphiocytes). A mieg-

Fig. 4.-63-year-old woman complaining of backache for 1 month and paraparesis for 10 days.A, Contrast-enhanced Ti-weighted sagittal MR image shows prominent enhancement of thickened meningeswith associated cord compression. Hyperintensity of one oflowerthoracic vertebrae (arrow) is caused by mci-dental hemangioma.B, Degree of cord compression is better shown in axial contrast-enhanced MR image.

Fig. 6-45-year-old woman with tubercular arach-noiditis. Axial contrast-enhanced MR image showscentral cord enhancement.

MR Imaging of Tubercular Spinal Arachnoiditis

AJR:168, March 1997 811

Fig. 5.-40-year-old woman who developed para-paresis while being treated for tubercular meningitis.Contrast-enhanced MR image through spine showsenhancement along surface of cord (arrow).

ative India ink study is necessary to) rule out

cryptococcal infectiomi. Radiologic imivestmga-

tions play ami iniportamit role in comioboratimig

the clinical suspiciomi and firniing a diagmiosis.

Myelography had beemi the miiost co)miimiiomily

used techiiique for diagmiosimig arachmioiditms

umitil the advemit of MR imiiaging. Irregularity of

the co)mitrast columiimi amid mierve root abnormali-

ties caused by arachmioiditis are well delineated

by niyelography. However, shortcomings

include low sensitivity in detecting subtle

adhesiomis I 131, imiability to reveal the complete

extent of the disease in case of a myelographic

block, inability to define the cause of CSF

block, and poor detection of cord abnormali-

ties. MR imaging has recently been shown to

be useful in the diagnosis of noninfectious [3-

5] and tubercular arachnoiditis [I, 6, 71. In our

study, MR imaging also was able to delineate

many of the pathologic changes occurring in

patients with tubercular arachnoiditis.

Previous studies have found the thoracic

region to be most commonly involved in

patients with tubercular arachnoiditis [7, 1 1,

12, 14]; we observed thoracic involvement in

19 patients. However, in our study, the

inflammatory process was localized to the

thoracic region in only two patients. In other

patients, the disease extended to the cervical

or lumbar region, or both. This long-segment

involvement observed in our patients is in

accordance with the results of other studies

[I. 6, 7, 1 1]. In view of frequently observed

long-segment involvement, we believe that

sagittal images for the whole spine using

extended FOV should be included in all stud-

ies performed for tubercular arachnoiditis.

In a large number of our patients, the cord

outline was lost or indistinct on Tl-weighted

images because of increased signal intensity of

the CSF surrounding the cord. This finding,

which also has been observed by others [1, 6],

may be related to an increase in the level of

CSF proteins, the presence of inflammatory

exudate on the surface of the cord, or the for-

mation of microadhesions between the cord

and the theca [ 15]. Tubercular arachnoidi#{220}s is

known to cause a marked increase in CSF pro-

tein levels [ 16]. Although some of the patients

with noninfectious arachnoiditis may have an

increase in CSF protein levels [4], an increased

signal intensity of the CSF has not been

described in these patients. In all ofour patients

with CSF loculations, the intensity of the CSF

within the loculations was always normal, even

when the CSF outside showed increased signal

intensity on MR imaging. Similarly, the signal

intensity of the CSF around the cauda equina

roots never was increased, indicating that the

increase in the CSF signal intensity may not be

related solely to the CSF protein levels and that

multiple factors may be responsible for the

appearance of CSF on Ti-weighted MR im-

ages of these patients.

Recognition of CSF loculations in patients

with tubercular arachnoiditis is important as sur-

gical intervention may be required [15]. MR

imaging was able to delineate these lesions

clearly as extramedullary fluid loculations with

signal intensity con’esponding to that ofCSF on

T2-weighted images. Associated cord compres-

sion was also evident. The wall of the locula-

Sharma et al.

812 AJR:168, March 1997

tions often was not clearly seen, and the

presence of contour changes of the cord at the

site ofloculation led to their detection (Fig. 3B).

Thickening and clumping of the nerve roots

have been seen frequently in arachnoiditis.

Demonstration of these changes on myelogra-

phy or MR imaging has been one of the impor-

tant diagnostic signs in evaluation of these

patients. Although we were able to document

these changes in six of our patients with lum-

bosacral involvement, commenting on the pres-

ence or absence of nerve root clumping in a

majority of patients who had cervical or dorsal

spine involvement was impossible. In the

future, high-resolution fast spin-echo sequences

[5] may be used to improve the sensitivity of

MR imaging in detecting these changes.

Five of our patients showed the presence of

nodular lesions in the subarachnoid space repre-

senting either extramedullary tuberculomas or

fibrous tissue. Chang et al. [6] suggested that

differentiation between the two may be possible

on contrast-enhanced MR images. However.

fibrous tissue is known to become vascularized

and may show enhancement after IV adminis-

tration of contrast medium [4]. Thus, it may not

be possible to differentiate extramedullary gran-

ulomas from fibrous tissue on MR images

alone. In our experience, we have frequently

observed resolution of intracranial tuberculo-

mas after antitubercular treatment. If extramed-

ullary granulomas show a similar response.

repeat MR imaging after treatment may be able

to differentiate fibrous tissue from granulomas.

The spinal cord signal intensity changes

seen on the T2-weighted images of 82% of our

patients are in accordance with the involve-

ment of the spinal cord that has frequently been

reported in patients with tubercular arachnoidi-

tis [1, 6, 7, 1 1]. Apart from the presence of

myelitis, cord involvement may occur mdi-

reedy in the form of edema, ischemia, myelo-

malacia, or cavitation. Although patients with

cord cavitation could be differentiated from

patients with other causes of cord involvement,

differentiation of cord edema from infarction or

myelitis was impossible. In addition to detect-

ing intrinsic cord abnormalities, MR imaging

delineated both the presence and the cause of

extrinsic compression of the spinal cord, which

was present in I I of our patients.

Meningeal enhancement has been described

in patients with both postoperative and tubercu-

lar arachnoiditis FI , 4, 6, 7]. The MR images of

80% of our patients revealed enhancement of

the meninges. In view of the absence of any spi-

nal surgery in these patients, we believe this

enhancement represents ongoing meningeal

inflammation and may constitute an early sign

of arachnoiditis on MR imaging. Enhancement

of nerve roots in the lumbar region was

observed in six of our patients. Contrast

enhancement in patients with arachnoiditis is

usually mild [4, 5]. Although no attempt was

made in the present study to grade the degree of

enhancement, prominent enhancement of the

meninges was not a rare occurrence (Figs. 2C

and 4). However. the enhancement was smooth

in contra.st to the irregular. nodular pattern of

enhancement seen in patients with leptom-

eningeal carcinomatosis. Cord enhancement

was observed in 20% of the patients. Both sur-

face and central patterns of enhancement were

observed. Although the enhancement of the sur-

face of the cord may represent pial enhancement

or enhancing surface exudate, enhancement of

the center of the cord possibly is associated with

myelitis, ischemia, or the presence of intramed-

ullary granulomas. However, further studies are

required to evaluate the role of central enhance-

ment of the cord in the differentiation of cord

edema from myelitis or infarction.

Two of our patients had associated tubercu-

lar spondylitis. Of 1 1 patients for whom era-

nial CT or MR imaging studies were available,

positive findings (basal exudate, intracranial

granulomas, and ventriculitis) were observed

in nine patients. Three of these patients were

asymptomatic for intracranial lesions, and the

cranial study was performed after the extended

FOV whole-spine images revealed abnormali-

ties in the posterior fossa. In view of the fre-

quent occurrence of intracranial abnormalities

in these patients, we recommend contrast-

enhanced MR imaging examination of the

brain in such patients.

In conclusion, we believe that contrast-

enhanced MR imaging should serve as the first

line of investigation for the evaluation of sus-

pected tubercular arachnoiditis. Increased sig-

nal intensity ofthe CSF on Tl-weighted images

and enhancement of the meninges (signifying

ongoing inflammation) are signs that may aid in

early diagnosis of arachnoiditis before onset of

the adhesive process. The excellent contrast res-

olution of MR scans helps in detection of cord

changes that are frequently seen in these

patients. Other signs of arachnoiditis seen on

MR imaging include the presence of CSF locu-

lations. subarachnoid nodular lesions, clumping

of cauda equina nerve roots, and enhancement

of spinal cord or nerve roots. The ability of MR

imaging to evaluate the whole spine irrespec-

tive of any myelographic blocks. delineation of

the exact cause of cord compression, and detec-

tion of frequently associated intracranial or ver-

tebral abnormalities are other advantages of

MR imaging over myelography. Because of

freq uent long-segment involvement, extended

FOV images should be acquired for the whole

spine in all patients. Study of the brain should

also be routinely carried out to detect the fre-

quently associated intracranial abnormalities.

References

I . Kumar A, Montanera W, Willinsky R. TerBrugge

WG, Aggarwal S. MR features of tubercular arach-noiditis. J Coinput Assist Tomogr 1993: 17: 127-130

2. Jinkins iR. Gupta R, Chang KH, Radriguez-Carba-

jal i. MR imaging of central nervous system tuber-

culosis. Radial Clin North Am 1995;33:771-7863. Ross iS, Masaryk Ti. Modic MT. et al. MR

imaging of lumbar arachnoiditis. AiR 1987;149:1025-1032

4. Johnson CE, Sze G. Benign lumbar arachnoidi-tis: MR imaging with gadopentetate dimeglu-

mine. AJNR 1990; 1 1:763-770

5. Fin GJ. Stevens JM. Postoperative arachnoiditis

diagnosed by high resolution fast spin echo MRI of

the lumbar spine. Neuroi-adiologv 1995:37:139-1456. Chang KH, Han MH, Choi YW, Kim 10. Han

MC, Kim CW. Tubercular arachnoiditis of thespine: findings on myelography, CT and MR

imaging. AJNR 1989; 10:1255-12627. Gupta RK, Gupta S. Kumar S. Kohli A, Misra

UK, Gujral RB. MRI in intraspinal tuberculosis.

Neuroradio!ogv 1994:36:39-43

8. Kioumehr F, Dadsetan MR. Rooholamini SA, Au

A.Central nervous system tuberculosis: MRI.

Neuroradio!ogv 1994:36:93-969. Whisler WW. Chronic spinal arachnoiditis. In:

Vinken P1. Bruyn GW. eds. Handbook of clinic a!

neurology. vol. 33. Infections of the nervous sys-

tem, Part 1 . Amsterdam: Elsevier-North Holland

Biomedical. 1978:263-274

10. Shaw MDM. Russel IA, Grossart KW. The

changing pattern of spinal arachnoiditis. J Neurol

Neurosurg Psychiatry 1978:4 1 :97-107

1 1. Phadke RV, Kohli A. lain VK. Gupta RK, Kumar

S. Gujral RB. Tubercular radiculomyelitis (arach-

noiditis): myelographic (and CT myelographic)

appearances. Australas Radio! 1994:38: 10-16

12. Wadia NH. Dastur DK. Spinal meningitides with

radiculomyelopathy. I . Clinical and radiologicalfeatures. J Neu,-ol Sci 1969:8:239-260

13. Mooij JJA. Spinal arachnoiditis: disease or coinci-

dence?Acta Neurochir (Wien) 1980:53: 151-160

14. Kozlowski K. Late spinal blocks after tubercular

meningitis. AiR 1963:90:1220-122615. Mark AS. Infectious and inflammatory diseases

of the spine. In: Atlas SW, ed. Magnetic reso-

notice imaging of the brain and spine. Philadel-

phia: Lippincott-Raven. 1996:1207-1264

16. Tandon PN. Tubercular meningitis (cranial and

spinal). In: Vinken P1. Bruyn GW. eds. Hand-

book of c!inica! iu’urolog�; vol. 33. Infections of

the nervous system, Part 1 . Amsterdam: Elsevier-

North Holland Biomedical. 1978:195-262