Full text of DOI:10.1007/PL00006049

Neuroradiology ꢀ2001) 43: 325±330
Ó Springer-Verlag 2001
DIAGNOSTIC NEURORADIOLOGY
Y.Inoue
Y.Nemoto
K.Ohata
H.Daikokuya
A.Hakuba
T.Tashiro
Syringomyelia associated with adhesive
spinal arachnoiditis: MRI
M.Shakudo
K.Nagai
K.Nakayama
R.Yamada
Abstract We reviewed the MRI of
seven patients with syringomyelia
associated with surgically proven
adhesive spinal arachnoiditis to de-
scribe clinical and MRI findings in
this condition. Using 0.5, 1.0 or
geal thickening was seen on T2-
weighted sagittal images only in one
case. Cord deformity due to adhe-
sion or displacement due to an asso-
ciated arachnoid cyst was seen in all
cases best demonstrated on axial
images. Focal blurring of the syrinx
Received: 25 November 1999
Accepted: 27 March 2000
Y.Inoue ꢀ ) ´ Y.Nemoto ´ H.Daikokuya ´ 1.5 tesla units, we obtained sagittal
)
A.Tashiro ´ M.Shakudo ´ K.Nakayama ´
T1- and T2-weighted spin-echo and wall on axial images was seen in six
R.Yamada
Department of Radiology,
Osaka City University Medical School,
axial T1-weighted images. Addi-
tional sagittal T2-weighted images
were obtained without using mo-
tion-artefact suppression. Contrast
medium was given intravenously to
five patients. The syrinx cavities
were thoracic in five cases, cervico-
thoracic in one, and extended from
C4to L1 in one. No Chiari malfor-
mation or craniovertebral junction
anomaly was demonstrated. Menin-
patients. Flow voids were seen in the
syrinx fluid in all cases on T2-
weighted images obtained without
motion-artefact suppression. No ab-
normal contrast enhancement was
demonstrated.
1
-4-3, Asahi-machi, Abeno-Ku
Osaka 545±8585, Japan
Tel.: + 81-6-66453831
Fax: + 81-6-66466655
K.Ohata ´ A.Hakuba ´ K.Nagai
Department of Neurosurgery,
Osaka City University Medical School,
Key words Syringomyelia ´
Arachnoiditis ´ Magnetic resonance
imaging
1
-4-3, Asahi-machi, Abeno-Ku
Osaka 545±8585, Japan
known to be associated with syringomyelia, the specific
MRI features of syringomyelia caused by this condition
Introduction
Syringomyelia, a tubular cavity within the spinal cord, have not been described. We report clinical and MRI
can be associated with a variety of disorders [1, 2, 3], findings in seven patients with syringomyelia caused by
most commonly the Chiari malformation, followed by spinal arachnoiditis.
intramedullary tumour, then spinal cord trauma, of
which it is a late sequel [1, 2, 3]. Scarring within spinal
meninges is another cause [1, 2, 4]. In such scarring, usu-
Patients and methods
ally termed ªarachnoiditisº, all three enveloping mem-
branes are usually involved and adherent to one
another. While ªpachymeningitis with leptomeningitisº
We reviewed seven patients with syringomyelia associated with
chronic adhesive spinal arachnoiditis, two men and five women,
thus would be a more appropriate designation from the ranging in age from 39 to 67 years. Chief symptoms and signs
were paraparesis in four, hypoaesthesiae in the limbs in three,
pain or numbness of the arm in two, atrophy of the hand in one
and thoracic back pain in one. The cause of the chronic arachnoidi-
tis was determined from the history: previous tuberculous meningi-
pathological standpoint, the term ªarachnoiditisº will
be adopted, following common usage. Arachnoiditis, of-
ten called adhesive spinal arachnoiditis, may develop af-
ter trauma, following pyogenic, tuberculous or luetic
tis in one, postoperative scarring in one, minor trauma in two and
meningitis, after bleeding into the subarachnoid space
spinal anaesthesia in one; the cause was undetermined in two. Mi-
and postoperatively [4]. Although arachnoiditis is well nor trauma represented injury to the back or spine without a neu-

3
26
Table 1 Clinical data, surgical findings, and postoperative outcome
Case
Age
years)/sex
Presumed cause
of arachnoiditis
Interval to
symptoms
Duration of
arachnoiditis
Levels of adhesion
at surgery
Outcome
ꢀ
1
2
3
47 / M
39 / F
57 / M
Unknown
6 years
9 months
6 months
?
C6±T3
C4±L1
T1±3
No change
T7±10
T2±10
T5?±10
Complete recovery
No change
Mild deterioration
Spinal anaesthesia
Minor trauma
11 months
Minor trauma
Unknown
17 years
31 years
T10±11
20 years
?
34 years
7
/ F 44
Operative scar
54/ F
29 years
5
6
7
2 years
10 months
16 years
No change
Deterioration
Moderate improvement
55 / F
45 / F
Meningitis
Table 2 MRI findings
Case Level of
Thickening
Deformity of Focal blurring
Flow void
Contrast
syrinx
of meninges
spinal cord
of syrinx wall
in syrinx
enhancement
1
2
3
4
5
6
7
C3±T2
C4±L1
C1 ±T2
No
No
No
Yes
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
T9±L1
No
No
Yes
No
T8±11
T6±12
T5±L1
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Not given
Not given
rological deficit at the time of the accident. Clinical data are sum-
marised in Table 1. Our erroneous preoperative diagnosis was sy-
ringomyelia probably associated with intramedullary tumour in
cases 6 and 7, seen early in our MRI experience. In the remaining
five cases, our impression was syringomyelia probably associated
with adhesive arachnoiditis. All the patients underwent surgical
Results
The MRI findings are summarised in Table 2. The syr-
inx cavities were easily visible, exceeding 5 mm in diam-
eter on sagittal images and varying in length from 4to 17
procedures and syringomyelia associated with chronic adhesive vertebral segments. Five were thoracic, one was cervico-
arachnoiditis was confirmed.
thoracic and the last extended from C4to the conus me-
dullaris. No Chiari malformation or craniovertebral
junction anomaly was present. Definite thickening of
meninges was demonstrated in one case ꢀFig.1b), in
which it was visible only on sagittal T2-weighted images.
Deformity or displacement of the spinal cord without
Surgery consisted of microsurgical arachnoid dissection with
shunting of the syrinx cavity in cases 5 and 6, shunting followed by
arachnoid dissection for recurrence of syrinx in cases 2 and 3,
arachnoid dissection alone in case 1 and shunting alone in cases 4
and 7.
MRI was performed at 1.5 T in cases 1±3, 1.0 T in case 4and
0
.5 T in cases 5±7. Sagittal T1- ꢀ500±600/15±30 ms, TR/TE) and enlargement was seen in all seven patients, at the cranial
T2-weighted and axial T1-weighted spin-echo images were ob-
tained. T2-weighted images were obtained with fast spin-echo se-
quences ꢀ4500±5000/90 ms eff) in cases 1 and 2, and with spin-
echo sequences ꢀ2000±2500/90±120 ms) in the remaining five cases.
Axial T2-weighted images were also obtained cases 1±3. In all
cases, additional sagittal T2-weighted spin-echo images were ob-
tained without motion-artefact suppression, so that movement of
or caudal end or middle portion of the syrinx cavity, best
demonstrated on axial images ꢀFig.2e). In six cases,
blurring of a portion of the syrinx wall was noted on ax-
ial images ꢀFigs.2f, 3c). All syrinx cavities showed a
the fluid could be observed in the syrinx. The acquisition matrix Fig.1a, b Case 5. a Sagittal T1-weighted image. The spinal cord is V
was 256 to 192 ” 256 with two averages. The scan thickness was enlarged and a cavity is seen to extend from T8 to T11. b Sagittal
4
±5 mm in five patients and 8 mm in two. Gadopentetate dimeglu- T2-weighted image with motion-artefact suppression. Thickening
mine ꢀGd-DTPA; 0.1 mmol/kg) was administered intravenously to of meninges is seen as a linear band of low intensity at the dorsal
five patients. Three neuroradiologists reviewed the images for the aspect of the spinal cord from T7 to T10 ꢀarrows); this was con-
site of syrinx cavities, meningeal thickening, deformity or displace- firmed at surgery
ment of the spinal cord, blurring ꢀi.e., an indistinct appearance) of
Fig.2a±f Case 2. a, b Sagittal T1-weighted images. The spinal
the syrinx wall, flow void within the syrinx fluid and abnormal con-
cord is enlarged, and a syrinx cavity is evident from C4to L1; no
trast enhancement were evaluated. The results of the three review-
meningeal thickening is demonstrated. c, d Sagittal T2-weighted
ers were compared: indeterminate cases were reviewed again and a
images with motion-artefact suppression. Flow void is seen in the
consensus was reached. The MRI and surgical findings were com-
cavity fluid despite the motion-artefact suppression. e Axial T1-
pared. Clinical follow-up was by review of the medical records.
weighted image at T1 level. A cystic space ꢀarrow) seen on the
left, lateral to the swollen cord, was found at surgery to be an
arachnoid cyst. f Axial T2-weighted image at T9 level. The right
dorsal aspect of the cavity wall is thinned and blurred ꢀarrow).
This area was found to have the most severe adhesions at surgery

3
27
1
a
1b
2
a
2
b
2c
2
d
2
e
2f

3
28
Fig.3a±c Case 1. a Sagittal T1-
weighted image. b Sagittal T2-
weighted image without mo-
tion-artefact suppression. A
syrinx cavity is seen from C3 to
T2; flow void is evident within
it. c Axial T1-weighted image at
C6 demonstrates focal blurring
of the cavity wall over its right
dorsolateral aspect ꢀarrows)
a
b
c
flow void: within them on sagittal T2-weighted images
On follow-up MRI, the syrinx cavity had disappeared
without motion-artefact suppression ꢀFigs.2b, 3d). No in case 1, decreased in size in cases 3, 4, 5 and 7, and was
patient who received intravenous contrast medium unchanged in cases 2 and 5. On neurological examina-
showed abnormal enhancement.
tion, complete recovery was achieved in case 1, and there
At surgery, meningeal adhesions were seen, varying was moderate improvement in case 7, with no change in
in length from 3 to 19 vertebral segments ꢀTable 2); in cases 2, 4and 5 and deterioration in cases 3 and 6.
case 7, the extent of the arachnoid adhesions was not
precisely defined. The meningeal adhesions were seen
over the dorsal or dorsolateral aspect of the spinal
cord; the ventral aspect appeared uninvolved. The spi-
Discussion
nal cord parenchyma at the site of the most severe adhe- In addition to the Chiari malformation, intramedullary
sion appeared focally softened or spongy. Further tumours and trauma producing paraplegia or quadriple-
analysis postoperatively suggested that blurring of the gia [1, 2, 3], syringomyelia can be seen with basal menin-
syrinx wall on axial images corresponded to the site of gitis [5]. On MRI of syringomyelia with the Chiari
the most severe adhesions to the spinal cord. The syrinx malformation, low cerebellar tonsils are evident and
cavities were soft and flaccid. Associated arachnoid the cavity is most frequently seen in the lower cervical,
cysts were lysed cephalad to the adhesion in case 4and but occasionally in the midcervical or thoracic spinal
in the midportion of the adhesion in case 2. Displace- cord [6, 7]. On axial images, the syrinx cavity and walls
ment or deformity of the cord demonstrated on MRI usually are well-delineated and distinct. When syrinx
corresponded to the site of adhesion of the cord to the cavities exceed 3 mm in diameter, the fluid within them
dural sac, and the arachnoid cyst in two cases.
usually demonstrates flow void on sagittal T2-weighted

3
29
spin-echo images without motion-artefact suppression ings, we suspect that the blurred zone is a site of commu-
6, 8]. nication between the syrinx cavity and the spinal
Syringomyelia without Chiari malformation often subarachnoid space. Since blurring of the syrinx wall
[
raises the question of an associated intramedullary tu- can also can be demonstrated in cases of post-traumatic
mour. Since the introduction of intravenous MRI con- syringomyelia, it is not specific.
trast medium, this question has been largely resolved,
Definite thickening of the meninges was seen in only
since most intramedullary tumours show contrast en- one patient. The degree and extent of the meningeal
hancement, even histologically low-grade gliomas, un- thickening in adhesive arachnoiditis varies and it may
like those in the brain [9, 10, 11]. Some intramedullary not be thick enough to be demonstrated on MRI.
tumours, however, do not enhance [9, 11]; absence of
enhancement does not, therefore, completely exclude a our cases indicated that the cavity was flaccid, which
tumour. was confirmed at surgery. The flow void is not a distin-
The flow void demonstrated in the syrinx cavity in all
In post-traumatic syringomyelia, the history of trau- guishing characteristic of syringomyelia associated with
ma is clear, with diplegia or quadriplegia and sensory chronic adhesive arachnoiditis, since it is seen in almost
deficits below the level of injury; new symptoms and all cases associated with a Chiari malformation or with
signs can appear due to formation or extension of a cav- basal meningitis [8]. The majority of cavities associated
ity many years after the trauma.
with intramedullary tumours demonstrate no flow void,
In the majority of cases of syringomyelia secondary although rare exceptions have been reported [6], and
to basal arachnoiditis, a history of meningitis may be ob- we have seen a flow void in a cavity associated with a
tained, and MRI demonstrates the cavities to be contin- haemangioblastoma.
uous with the fourth ventricle. No cord deformity has
Contrast enhancement was not seen, probably be-
been evident in our limited experience ꢀunpublished ob- cause the arachnoiditis was in an inactive phase. We do
servations).
not know whether enhancement would have been seen
Surgical management and the postoperative neuro- with active inflammation.
logical outcome of patients with syringomyelia associat-
At least three mechanisms have been proposed for
ed with adhesive arachnoiditis has been reported [12, syrinx formation in cases of adhesive arachnoiditis: vas-
3, 14, 15]. To the best of our knowledge, however, cular compromise [16], cerebrospinal fluid ꢀCSF) block,
1
MRI findings in syringomyelia associated adhesive spi- and fixation of the spinal cord due to adhesive arach-
nal arachnoiditis has not been described in the English- noiditis ꢀmeningeal scarring) [17, 18].
language literature. We demonstrated that the most
Focal scarring of meninges and spinal block seem to
striking findings were: deformity of the cord at the site be the basic pathology. This fixes the spinal cord and
of adhesive arachnoiditis, blurring of a part of the syrinx could obstruct spinal CSF circulation and transient
wall, flow void within the syrinx cavities, and lack of ab- movement of CSF. Altered CSF pressure cranial and
normal contrast enhancement. Unexpectedly, menin- caudal to the block ꢀªupper spinal/lower spinal pressure
geal thickening suggesting the arachnoid scarring itself dissociationº) [18] could force CSF into the spinal cord
was rarely detected.
via Virchow-Robin spaces [19] whether or not there is
Deformity of the spinal cord was demonstrated in all already a cavity. On this theory, syrinx formation and
seven cases at or near the site of arachnoid scarring. growth are well explained. With a blocked subarachnoid
Such deformity can be seen in cases of cord damage after space, CSF below the block could be forced into the
inflammation or infarction, and also may be seen due to cord substance during periods of transiently elevated
compression by extramedullary tumours or arachnoid pressure.
cysts. In these cases, syringomyelia is usually absent. In
According to the vascular compromise theory, the ar-
syringomyelia associated with a Chiari malformation or teries and veins of the spinal cord could be narrowed or
basal arachnoiditis there is usually cord swelling without occluded by surrounding fibrosis, resulting in cord is-
deformity; an atrophic cord can occasionally be seen [6, chaemia manifest as areas of softening and cavitation.
7
, 8]. Deformity of the cord at one end or in the middle Caplan et al. [16] maintain that arterial obliteration
of a syrinx cavity appears to be characteristic of syringo- and focal scarring producing spinal CSF block are both
myelia caused by adhesive spinal arachnoiditis. causes of syrinx formation. Although all three mecha-
Syrinx cavities associated with a Chiari malformation nisms could be interrelated in the formation of syrinx
or basal arachnoiditis are usually well-demarcated on cavities, vascular compromise may not be a necessary
T1-weighted sagittal and axial images. In contrast, focal element, since syringomyelia has been associated with
blurring of the syrinx wall was demonstrated on axial intradural-extramedullary tumours [6, 20, 21], and in
images in six of our seven patients. At surgery, the focal many reported clinical cases of old spinal cord infarct
blurring corresponded the level of the most severe cord as well as in our experience, MRI has demonstrated
adhesion, and may represent softening or rarefaction of only atrophy of the cord and varying degrees of in-
the damaged cord. Correlating surgical and MRI find- creased signal intensity on T2-weighted images [22].

3
30
References
1
. Harding BN. Malformations. In:
Adams JH, Duchen LW ꢀeds) ꢀ1992)
Greenfield's neuropathology, 5th edn.
Arnold, London, pp 521±638
. Adams RD, Victor M ꢀ1993) Principles
of neurology. McGraw-Hill, New York,
pp 1078±1116
. Oakes WJ ꢀ1992) Chiari malformation,
hydromyelia, syringomyelia. In: Wilkins
RH, Rengachary SS ꢀeds) Neurosur-
gery. McGraw-Hill, New York, pp
8. Fukuda T, Inoue Y, Nemoto Y, Mat-
sumura Y, Takemoto M, Shakudo M,
Hashimoto H, Kida A, Shibakiri I, On-
oyama Y, Hakuba A, Yagura H ꢀ1987)
Magnetic resonance imaging of syrinx
15. Klekamp J, Batzdorf U, Samii M, Bothe
HW ꢀ1997) Treatment of syringomyelia
associated with arachnoid scarring
caused by arachnoiditis or trauma.
J Neurosurg 86: 233±240
2
3
cavities: differentiation of syrinxes with 16. Caplan LR, Norohna AB, Amico LL
spinal cord tumor and without tumor.
Nippon Acta Radiol 47: 1609±1611
ꢀJapanese)
ꢀ1990) Syringomyelia and arachnoiditis.
J Neurol Neurosurg Psychiatry 53:
106±113
9. Parizel PM, BalØriaux D, Rodesch G,
Segabarth C, Lalmand B, Christophe C,
Lemort, Haesendoch P, Niendorf HP,
Flament-Durand J, Brotchi J ꢀ1989)
Gd-DTPA-enhanced MR imaging of
spinal tumors. AJNR 10: 249±258
10. Sze G, Stimac GK, Bartlett C, et al
ꢀ1990) Multicenter study of gadopen-
tate dimeglumine as an MR contrast
17. Barnett HJM ꢀ1973) Pathogenesis of
syringomyelic cavitation associated
with arachnoiditis localized to the spi-
nal canal. In: Barnett HJM, Foster JB,
Hudgson P ꢀeds) Syringomyelia. Saun-
ders, London, pp 245±260
18. Williams B ꢀ1980) On the pathogenesis
of syringomyelia: a review. J Roy Soc
Med 73: 798±806
3
593±3616
4
. Barnett HJM ꢀ1973) Syringomyelia as-
sociated with spinal arachnoiditis. In:
Barnett HJM, Foster JB, Hudgson P
ꢀ
eds) Syringomyelia. Saunders, Lon-
don, pp 220±244
5
6
7
. Foster JB, Hudgson P ꢀ1973) Basal
arachnoiditis. In: Barnett HJM, Foster
JB, Hudgson P ꢀeds) Syringomyelia.
Saunders, London, pp 30±49
agent: evaluation in patients with spinal 19. Ball MJ, Dayan AD ꢀ1972) Pathogene-
tumors. AJNR 11: 967±974 sis of syringomyelia. Lancet ii:799±801
11. Enzmann DR, DeLaPaz RL ꢀ1990) Tu- 20. Quencer RM, Gammal TE, Cohen
. Sherman JL, Barkovich AJ, Citrin CM
mor. In: Enzmann DR, DeLaPaz RL,
Rubin JB ꢀeds) Magnetic resonance of
the spine. Mosby, St. Louis, 301±324
Gꢀ1986) Syringomyelia associated with
intradural extramedullary masses of the
spinal canal. AJNR 7: 143±148
ꢀ
1986) The MR appearance of syringo-
myelia: new observations. AJNR 7:
85±795
7
12. Williams B ꢀ1990) Post-traumatic syrin- 21. Slasky BS, Bydder GM, Niendorf P,
. Elster AD, Chen MYM ꢀ1992) Chiari
malformations: clinical and radiological
reappraisal. Radiology 183: 347±353
gomyelia, an update. Paraplegia 28:
296±313
13. Sgouros S, Williams B ꢀ1995) A critical
appraisal of drainage in syringomyelia.
J Neurosurg 82: 1±10
Young IR ꢀ1987) MR imaging with ga-
dolinium-DTPA in the differentiation
of tumor, syrinx, and cyst of the spinal
cord. J Comput Assist Tomogr 11:
845±850
1
4. Sgouros S, Williams B ꢀ1996) Manage-
ment and outcome of posttraumatic sy-
ringomyelia. J Neurosurg 85: 197±205
22. Mawad ME, Rivera V, Clawford S, Ra-
mirez A, Breibach W ꢀ1990) Spinal
cord ischemia after resection of thora-
coabdominal aortic aneurysms: MR
findings in 24patients. AJNR 11:
9
87±991