The

Neuroradiology Journal 24: 511-518, 2011 www.centauro.it

Clinico-Radiologic Profile of Spinal Cord

4th Module/4° Modulo, Agosto 2011

Multiple Sclerosis in Adults

CME Course / Corso ECM

A.k. BAg, B.N. PATEL, S. OSMAN, g.H. ROBERSON
Department of Radiology, University of Alabama; Birmingham, AL, USA

Key words: multiple sclerosis, MRI, spinal cord

SUMMARY – MRI is extremely useful for the assessment of initial disease burden and to identify
the dissemination of the multiple sclerosis (MS) in time and space. Though MRI of the spinal cord
is not used to establish the diagnosis of MS, spinal cord is frequently involved in this disease and
there has been increasing emphasis of the spinal imaging in making clinical decision in the man-
agement of MS. We undertook a retrospective study of patients with diagnosed MS: 1) to identify
radiologic pattern of spinal cord involvement in MS and 2) to correlate radiologic findings with
clinical presentation. We reviewed radiologic records from 2004 to 2009 of patients with abnormal
T2 signal intensity of the spinal cord with radiologic concern of demyelinating disease. Patients
in this cohort who met the Revised McDonald MS Diagnostic Criteria were included in this study.
166 patients were included in the study. There was preference for cervical spinal cord particularly
posterior aspect of the spinal cord. Enhancement of the lesions was rare (4.1%). Mean lesion length
was 18.2 mm. The average number of lesions per patient was 2.04. Sensory symptoms were predom-
inating and most of the patients had relapsing-remitting course. Patients with sensory symptoms,
bladder and bowel involvement and motor symptoms had almost equally distributed lesions among
anterior, posterior and central spinal cord. However, all of the patients presented with posterior
column signs and gait abnormality had involvement of the posterior spinal cord. Radiologic mani-
festation of spinal cord MS is extremely variable and can involve the entire length of the spinal
cord. Clinical symptoms may or may not be associated with radiologic presentation of the lesions.

MRI of the brain is an integral part in diag- tic work-up and to identify dissemination of the
nosis of multiple sclerosis (MS) and to identify disease in time and space which was rapidly im-
dissemination of MS in time and space. No em- plemented, both in daily practice and research 1.
phasis was given to spinal cord lesion in the di- Today spinal cord imaging plays very impor-
agnostic work-up and management of MS in the tant role in the management of MS patients.
20th century. Spinal cord imaging is extremely Specific spinal cord lesion imaging characteris-
helpful in management of patients with MS be- tics have also been defined 4. We conducted this
cause it helps to exclude alternative diagnoses study to identify distribution, characterization
and unlike brain, healthy people do not develop of spinal cord lesions in a larger patient group
white matter lesions in the spine 1. The original and also to find out if involvement pattern has
McDonald criteria, published in 2001, recom- any association with symptoms.
mended that “one spinal cord lesion can be sub-
stituted for one brain lesion” 2. This statement
was confusing and did not provide sufficient “im- Materials and Methods
plementation guidelines” to incorporate spinal
cord imaging in the diagnostic work-up of MS Patient selection
3
. In the revised McDonald criteria, spinal cord
imaging was integrated in the initial diagnos- The study was approved by the local insti-
tutional review board. We ran a system-wise
Paper presented at the XIX Symposium Neuroradiologicum, 2010. Radiology Information Service (RIS) database

511
 Clinico-Radiologic Profile of Spinal Cord Multiple Sclerosis in Adults A.K. Bag

search at the University of Alabama at Bir- due to paucity of lesions in longer thoracic cord.
4th Module/4° Modulo, Agosto 2011

mingham of all the patients who presented be- Lesions at the level of dens (C1) were grouped
CME Course / Corso ECM

tween January 2004 and December 2009 and under C2 lesions.

had abnormal T2 signal of the spinal cord in
spinal MRI with radiologic concern for demy- Clinical parameters evaluated
elinating disease. We then reviewed clinical
records and laboratory results of all these pa- From the clinical records we reviewed the de-
tients. Patients with abnormal T2 signal abnor- mography of patients (age, sex and race), clini-
mality on the spinal cord who met the revised cal presentation, neurologic signs and disease
McDonald MS Diagnostic Criteria in the course course [Relapsing remitting (RR), secondary
of the disease were then included for the study. progressive (SP), Primary progressive (PP) and
progressive relapsing (PR)] of all the patients.
MRI protocol We then evaluated if the involvement pattern
and distribution of the lesion were associated
All MR imaging examinations were per- with clinical presentation and whether there
formed either with a 1.5 T system or 3 T sys- is any association between numbers of lesions
tem with the following imaging protocol sagit- with disease course. Data were analyzed by
tal T1 spin echo (SE), T2 fast spin echo (FSE), nonparametric statistical methods. We used
short tau inversion recovery (STIR) and axial SPSS 13.0 student version to calculate mean
T1 SE, T2 FSE, STIR. Post contrast sagittal and standard deviation.
and axial T1 SE sequences were obtained with
injection of 0.1 mmol/Kg gadoteridol (Prohance,
Bracco Diagnostic Inc., Princeton, NJ). Cervi- Results
cal spinal cord MRI was obtained in all the pa-
tients. Thoracic and lumbar spine MRIs were We identified total 544 patients with abnor-
obtained based on the patients’ symptoms and/ mal T2 signal in the spinal with radiologic con-
or positive neurologic examination. cern for demyelinating disease. After extensive
review of the radiologic, clinical and laboratory
Image analysis records we found 166 patients met the revised
McDonald MS Diagnostic Criteria. Most of the
We analyzed a) involvement pattern of the patients with abnormal T2 signal abnormality
spinal cord (anterior, posterior, central or dif- who did not meet the revised McDonald MS di-
fuse), b) distribution of lesion(s) along the agnostic criteria were due to compressive my-
length of the spinal cord (cervical, thoracic or elopathy secondary to degenerative changes.
lumbar), c) cranio-caudal length of the individ- These 166 patients were regularly followed up
ual lesion, d) total number of lesions per pa- in the MS clinic.
tient and e) enhancement pattern of individual Demographic Characteristics. The median
lesion. Anterior lesions were defined when the age was 42 years with range of 17-75 years.
lesions were limited anterior two third of the Female to male ratio was 12.9:1. Caucasians
cord with no abnormality in the posterior cord were more frequently involved compared to the
(both sagittally and axially). Posterior lesions African American (1.84:1).
were defined when the lesions were limited Clinical Presentation. Most common clini-
within posterior two third of the cord (both sag- cal presentation was sensory 42.77%. Motor
ittally and axially) with no involvement of the symptoms were present in 37.95% of patient.
anterior cord. Other clinical symptoms in decreasing order
Central lesions were defined when lesions of frequency were gait abnormality (21.68%),
were well defined, centrally located with nor- bladder and bowel abnormality (12.65%) and
mal appearance of the periphery of the cord. Lhermitte’s syndrome (3.01%). In 21 (12.65%)
Diffuse abnormalities were defined when le- patients, there were no specific spinal symp-
sions involved the entire cross section of the toms. RR course was the predominant disease
cord or more than two third of the spinal cord progression pattern (in 71.68%), followed by SP
on cross section. in 24.09%. There was no case of PP or PR in
Distribution of the lesions along the length our patient population. Neuromyelitis optica
of the spinal cord was grouped as two-vertebral (NMO) was diagnosed in 4.1% of all patients.
body approach in the cervical spine and four- NMO patients were included in the statistical
vertebral body approach in the thoracic spine analysis. Average number of lesions in patients

512
www.centauro.it The Neuroradiology Journal 24: 511-518, 2011

Chart 1 Longitudinal bar diagram to depict distribution of the lesions along the spinal cord.

4th Module/4° Modulo, Agosto 2011

CME Course / Corso ECM
Note: The X-axis depicts the cervical cord segment.

with RR course pattern were 2.20 (range 1-7) nal cord is given in Chart 1. There was pref-
and in patients with SP disease pattern were erential involvement of the posterior spinal
2.14 (range 1-5) which was not statistically sig- cord in 46.47% of the lesions (Figure 1) (Table
nificant (Table 1). 2). Anterior cord was involved in 27.94% of
Radiologic Findings. There were total 340 cases (Figure 2), central cord was involved in
lesions in 166 patients. The distribution pat- 22.35% (Figure 3) and diffuse involvement was
tern along the length of the spinal cord is given in 3.23% (Figure 4). The mean length of indi-
in the bar diagram (Chart 1). Cervical cord vidual lesion was 18.2 mm (range 3 mm -108
was most commonly involved (46.2%). Tho- mm). The mean number of lesions per patient
racic cord was less commonly involved. There was 2.04. Of the 340 lesions, only 14 (4.11%)
were two lesions in the conus. The distribu- lesions had subtle enhancement (Figure 5).
tion of the lesion along the length of the spi-
Correlation of MRI imaging Findings
Table 1 Clinical presentation of spinal cord MS in 166
patients. and Symptoms
Disease course Patients with sensory, motor and bladder
RRMS 71.68% symptoms had lesions almost equally distrib-
uted among the anterior, posterior and central
SPMS 24.09% cord. All the patients who had posterior column
PPMS 0%
Table 2 Radiologic characteristics of 340 lesions.
PRMS 0%
Location
NMO 4.21%
Posterior 46.47%
Sensory 42.77%
Anterior 27.94%
Motor 7.95%
Central 22.35%
Gait 21.68%
Diffuse 3.23%
Bladder 12.65%
Enhancement 4.4%
Lhermitte 3.01%
Mean 18.2 mm
No Spinal symptom 12.65% Lesion length
[range 3-108 mm]
Note: Relapsing remitting (RR), Secondary progressive (SP), Pri- Average number of lesions
mary progressive (PP), Progressive relapsing (PR), neuromyelitis 2.04
optica (NMO).
per patient

513
 Clinico-Radiologic Profile of Spinal Cord Multiple Sclerosis in Adults A.K. Bag
4th Module/4° Modulo, Agosto 2011

A B
CME Course / Corso ECM

Figure 1 Posterior lesion. On the sagittal T2-weighted sequence (A), there is a T2 hyperintense lesion in the posterior aspect of
the cord at the level if C1 (arrow) which is clearly seen at the right posterolateral aspect of the spinal cord on axial T2-weighted
sequence (B).

Figure 2 Anterior lesion. On this sagittal STIR sequence (A),

the T2 hyperintense lesion is located at the level of C6 which
on axial T2-weighted sequence (B) appears as a pachy lesion at
A the anterior aspect on the left side.

514
www.centauro.it The Neuroradiology Journal 24: 511-518, 2011

4th Module/4° Modulo, Agosto 2011

A B

CME Course / Corso ECM

Figure 3 Central lesion. On this sagittal T2-weighted sequence (A), a T2 hyperintense lesion is seen at the central cord with no
discrete lesion margin. However, there is a rim of normal tissue all around the lesion which is more clearly seen at the axial T2-
weighted sequence (B).

signs had posteriorly located lesions. Some of two vertebral segments, b) unequivocal hy-
the patients with posterior column signs also perintensity on T2-weighted sequence, c) in-
had anterior and central lesions. There is no volvement of only part of the cord cross sec-
difference in involvement pattern along the tion rather than whole cord and d) little or no
cross section of the cord and in distribution swelling of the cord 4,8,10. In our cohort of pa-
pattern along the length of the cord between tients, mean lesion length was 18.2 mm and
RRMS and SPMS. Mean number of lesions per most of the lesions were less than two verte-
patient was similar in RRMS and SPMS. bral body length. All of the lesions were une-
quivocally T2 hyperintense. Unfortunately we
did not looked at the swelling or atrophy of the
Discussion cord associated with MS.
Lycklama et al. 6 suggested different extent
Prevalence of spinal cord abnormalities in of spinal cord involvement in patients with
established MS is very high, even in the early RRMS and PPMS. In our cohort of patients,
stage of the disease 5. Depending upon differ- there was not a single patient with PPMS.
ent series, spinal cord abnormalities may be However, we did not find any difference in the
seen up to 97% of patients when focal and dif- extent and pattern of spinal cord involvement
fuse involvement was combined 6-7. However, in in patients with RRMS and SPMS. Individual
clinically isolated syndrome, the prevalence of lesion characteristics (including enhancement
spinal cord lesion is lower, particularly in the pattern) of RRMS and SPMS were similar in
absence of spinal cord symptoms 8. Asympto- our study patients.
matic cord lesion is not uncommon and is seen Enhancing lesions are less frequent in the
in about 30-40% of patients with clinically iso- spinal cord compared to the brain 11. In our pa-
lated syndrome 9. tient group, incidence of enhancing cord lesion
Typical characteristics of a spinal cord le- is only 4.4%, much lower compared to 17.2%
sion are a) at least 3 mm in size but less than (15/87) in the study conducted by Bot et al 4.

515
 Clinico-Radiologic Profile of Spinal Cord Multiple Sclerosis in Adults A.K. Bag

A
4th Module/4° Modulo, Agosto 2011
CME Course / Corso ECM

Figure 4 Diffuse lesion. On this sagittal STIR sequence (A),

there are two different lesions. The superior lesion at the level
of C1 is discrete posterior/central lesion. The inferior lesion is
rather long involving three vertebral segments and is diffusely
hyperintense with no discrete lesion margin and involves al-
most entire cross sectional area of the spinal cord, better de-
picted by the axial STIR sequence (B).

This difference in incidence of enhancement is sociated with posterior paramedian lesions

due to the fact that the all the study patients involving the dorsal column. Like many other
in the author’s group were newly diagnosed, groups 5,19 in our patient cohort there were si-
which are more likely to have enhancement 11. lent spinal lesions and all spinal lesions were
However, the incidence of enhancing MS lesion not clinically relevant.
in the spinal cord is so low that recent recom- Our study has many limitations. First of all
mendation suggests to use postcontrast T1- the study is retrospective and we used abnor-
weighted scans very selectively only in patients mal T2 signal in the spinal cord to screen pa-
with clinical suspicion of development of new tients. Due to the study design, we could not
spinal lesion or to exclude any other disease 5. establish the prevalence of spinal cord lesions
Imaging findings often do not match with clin- in patients with MS and evaluate how the spi-
ical presentation in MS patients, particularly in nal lesions were helpful in managing the pa-
patients with abnormal brain imaging 12. Simi- tient. We also did not evaluate the associated
larly spinal cord lesions are also weakly asso- swelling or atrophy.
ciated with symptoms 13-15. Possible explanation
of this so called ‘clinico-radiologic paradox’ are:
lack of spatial resolution of conventional MRI 16 Conclusion
and lack of pathologic specificity of these abnor-
mal T2 signal hyperintensity for MS lesions 17. This study describes the distribution of MS
However, newly formed spinal cord lesions lesions within the spinal cord. The study also
seem to have higher association with new MS establishes that distribution and location of le-
symptoms 5,11,18. In our patient cohort there was sions do not correlate with sensory, motor and
no association between lesion location and sen- bladder symptoms. However, patients with
sory, motor and bladder symptoms. However, posterior column signs and symptoms are as-
patients with posterior column signs were as- sociated with posteriorly located lesions.

516
www.centauro.it The Neuroradiology Journal 24: 511-518, 2011

4th Module/4° Modulo, Agosto 2011

C

CME Course / Corso ECM

D

Figure 5 Enhancing lesion. On the sagittal (A) and axial (C)

pre-contrast T1-weighted sequences, there is focal expansion
of the spinal cord at the level of C2-C3 with no appreciable
change of signal intensity. With contrast enhanced sagittal (B)
and axial (D) T1-weighted sequences, there is subtle enhance-
ment at the posterior aspect of the cord (arrows). On the axial
T2-weighted sequence (E), the lesion appears much larger
compared to the enhancing portion of the lesion.

517
 Clinico-Radiologic Profile of Spinal Cord Multiple Sclerosis in Adults A.K. Bag

References
4th Module/4° Modulo, Agosto 2011
CME Course / Corso ECM

1 Polman CH, Reingold SC, Edan G, et al. Diagnostic cri- 12 Barkhof F, van Walderveen M. Characterization of tis-
teria for multiple sclerosis: 2005 revisions to the “Mc- sue damage in multiple sclerosis by nuclear magnetic
Donald Criteria”. Ann Neurol. 2005; 58: 840-846. resonance. Philos Trans R Soc Lond B Biol Sci. 1999;
2 McDonald WI, Compston A, Edan G, et al. Recom- 354: 1675-1686.
mended diagnostic criteria for multiple sclerosis: guide- 13 Wiebe S, Lee DH, Karlik SJ, et al. Serial cranial and
lines from the International Panel on the diagnosis of spinal cord magnetic resonance imaging in multiple
multiple sclerosis. Ann Neurol. 2001; 50: 121-127. sclerosis. Ann Neurol. 1992; 32: 643-650.
3 Korteweg T, Barkhof F, Uitdehaag BM, et al. How to 14 Turano G, Jones SJ, Miller DH, et al. Correlation of
use spinal cord magnetic resonance imaging in the Mc- SEP abnormalities with brain and cervical cord MRI in
Donald diagnostic criteria for multiple sclerosis. Ann multiple sclerosis. Brain. 1991; 114 (Pt 1B): 663-681.
Neurol. 2005; 57: 606-607. 15 Honig LS, Sheremata WA. Magnetic resonance imag-
4 Bot JC, Barkhof F, Polman CH, et al. Spinal cord ab- ing of spinal cord lesions in multiple sclerosis. J Neurol
normalities in recently diagnosed MS patients: added Neurosurg Psychiatry. 1989; 52: 459-466.
value of spinal MRI examination. Neurology. 2004; 62: 16 Bachmann R, Reilmann R, Schwindt W, et al. FLAIR
226-233. imaging for multiple sclerosis: a comparative MR study
5 Bot JC, Barkhof F. Spinal-cord MRI in multiple sclero- at 1.5 and 3.0 Tesla. Eur Radiol. 2006; 16: 915-921.
sis: conventional and nonconventional MR techniques. 17 van Waesberghe JH, Kamphorst W, De Groot CJ, et
Neuroimaging Clin N Am. 2009; 19: 81-99. al. Axonal loss in multiple sclerosis lesions: magnetic
6 Lycklama a Nijeholt GJ, Barkhof F, Scheltens P, et resonance imaging insights into substrates of disabil-
al. MR of the spinal cord in multiple sclerosis: relation ity. Ann Neurol. 1999; 46: 747-754.
to clinical subtype and disability. Am J Neuroradiol. 18 Kidd D, Thorpe JW, Kendall BE, et al. MRI dynamics
1997; 18: 1041-1048. of brain and spinal cord in progressive multiple sclero-
7 Nijeholt GJ, van Walderveen MA, Castelijns JA, et al. sis. J Neurol Neurosurg Psychiatry. 1996; 60: 15-19.
Brain and spinal cord abnormalities in multiple sclero- 19 Papadopoulos A, Gatzonis S, Gouliamos A, et al. Corre-
sis. Correlation between MRI parameters, clinical sub- lation between spinal cord MRI and clinical features in
types and symptoms. Brain. 1998; 121 (Pt 4): 687-697. patients with demyelinating disease. Neuroradiology.
8 Brex PA, Leary SM, O’Riordan JI, et al. Measurement 1994; 36: 130-133.
of spinal cord area in clinically isolated syndromes sug-
gestive of multiple sclerosis. J Neurol Neurosurg Psy-
chiatry. 2001; 70: 544-547.
9 O’Riordan JI, Losseff NA, Phatouros C, et al. Asymp-
tomatic spinal cord lesions in clinically isolated optic
nerve, brain stem, and spinal cord syndromes sugges-
tive of demyelination. J Neurol Neurosurg Psychiatry.
1998; 64: 353-357. Asim K. Bag, MD
10 Dalton CM, Brex PA, Miszkiel KA, et al. Spinal cord Division of Neuroradiology
MRI in clinically isolated optic neuritis. J Neurol Neu- Department of Radiology
rosurg Psychiatry. 2003; 74: 1577-1580. 619 19th St S, WP-150
11 Thorpe JW, Kidd D, Moseley IF, et al. Serial gado- Birmingham, AL-35249-6830, USA
linium-enhanced MRI of the brain and spinal cord in Tel.: +1 205 934 3920
early relapsing-remitting multiple sclerosis. Neurol- Fax: +1 205 975 9262
ogy. 1996; 46: 373-378. E-mail: abag@uabmc.edu

518