- Email: [email protected]
Long time echo stir sequence magnetic resonance imaging of optic nerves in optic neuritis
EUROPEAN JOURNAL OF RADIOLOGY
@ ELSEVIER
European Journal of Radiology 19 (1995) 155-163
Long time echo stir sequence magnetic resonance imaging of optic nerves in optic neuritis A . T a r t a r o b, M . O n o f r j *a, A . T h o m a s a, T. F u l g e n t e a, C. D e l l i P i z z i b, L. B o n o m o b aDepartment of Neurology, State University of Chieti, Ospedale ex-Pediatrico, Via Martiri Lancianesi, 6, 66100 Chieti, Italy bDepartment of Radiology, State University of Chieti, Chieti, Italy
Received 24 October 1994; revision received 7 December 1994;accepted 12 December 1994
Abstract
Magnetic resonance imaging of optic nerves was obtained in 13 patients with acute optic neuritis and in 13 patients with a previous history of optic neuritis (ON), assessed by clinical, visual fields and visual evoked potentials evaluations. Results of the conventional short tau inversion recovery (STIR) sequence obtained with short time echo (STE-STIR: 22 ms) were compared with long time echo (LTE-STIR: 80 ms) sequence. The conventional STE-STIR sequence revealed lesions in 78.5% of acute ON and in 58.8% of optic nerves affected by previous ON. The LTE-STIR sequence was diagnostic in 92.8% of acutely symptomatic nerves, in 94.1.% of nerves with previous ON. The calculated length of optic nerve lesions was significantly longer in imaging obtained with the LTE-STIR sequence than with the conventional STE-STIR sequences, both in acute and previous ON. Keywords: Magnetic resonance (MR), technique; Magnetic resonance (MR), brain; Nerves, optic; Multiple sclerosis
1. Introduction
Optic neuritis (ON) is common in multiple sclerosis (MS), appearing as a condition that precedes MS or as a symptomatic lesion during the course of 'definite' MS [1,21. Despite the high sensitivity of magnetic resonance imaging (MRI) in detecting brain lesions in patients with MS, the initial attempts to detect ON were disappointing [3-5] because of the chemical shift artifact between optic nerve and orbital fat observed with T 1 - T 2 weighted sequences. The high proton density and short T 1 relaxation time of orbital fat results in a high signal intensity concealing intraorbital structures and contributing to partial volume averaging artifacts. These problems were partly overcome by the introduction of the short tau inversion recovery (STIR) sequence with the conventional short time echo (STESTIR) [6-8]. STE-STIR sequence evidenced increased signal intensity of optic nerves in 70% of acute ON, and
* Corresponding author.
in 20-50% of optic nerves studied when the acute ON had recovered [9,101. STE-STIR sequence, with inversion time (T1) at the 'null point' of fatty tissue, results in a signal void of the orbit fat: the normal optic nerve however, appears relatively hyperintense and clear separation between optic nerve and perioptic cerebrospinal fluid (CSF) is lost on STE-STIR images, because in sequences obtained with short T1 ( < 150 ms), image contrast is dependent on T 1 and T2 [11]. As already discussed by Hendrick and Roff [11]: 'although it is important to remember that image contrast in... short... STIR domain is due to the additive effects of correlated T1, T2 and N[H] inherent contrast', with a TE reaching 80 ms it is possible to obtain 'a more T2 weighted appearance' and to obtain therefore a contrast differentiation between edema/gliosis and cerebral nervous system structures. Based on this theoretical issue, Finn et al. [12] introduced a double time echo STIR sequence (STE: 22 ms and LTE: 80 ms) and obtained an increment of the signal to noise ratio in comparison with the spin echo T2-weighted images of white matter lesions.
0720-048X/95/$09.50 © 1995 Elsevier ScienceIreland Ltd. All rights reserved SSDI 0720-048X(94)00600-H
156
A. Tartaro et al./ European Journal of Radiology 19 (1995) 155-163
In the present study we hypothesized that the use of a double time echo STIR sequence, with an unconventional LTE of 80 ms, might enhance the contrast resolution of optic nerve lesions in ON. We compared the sensitivity of STIR sequences with an STE of 22 ms and an LTE of 80 ms in detecting lesions of optic nerves in 26 patients with acute or previous ON assessed by clinical, neuro-ophthalmological examinations and visual evoked potential (VEP) recordings.
2. Subjects and methods
2.1. Study population Twenty-six patients (age range 21-52 years; mean age 36.1 ± 9.6 years, 16 women) affected by acute or previous episodes of ON were recruited for the study. All patients underwent fundus examination, assessment of visual acuity by means of Snellen Charts, Ishihara Plates, visual field testing, pattern reversal VEP recordings and MRI of brain and optic nerve. Patients were divided into two groups: group I (13 patients) ineluded patients referred to our department because of the sudden onset of unilateral or bilateral (one patient) ocular pain and blurred vision. MRI studies were performed 3-6 days after the onset of symptoms, and were repeated 3 months after the onset of ON in four patients. Group II included 13 patients who had recovered from an ON occurring 4-18 months before MRI. Three of these patients had a history of ON of the other side, 18-60 months before the last ON attack. All the 13 patients of group II could be classified as definite MS according to McAlpine's criteria [11. In both groups of patients normal visual acuity, normal computerized perimetry and normal VEPs excluded the presence of ON. Monocular visual fields were tested by means of a computerized 640 Humphrey perimeter (Zeiss Humphrey) exploring the central 30° of the retina, and were classified as normal or abnormal compared with the normative age-matched data. Full field pattern reversal VEPs were recorded monocularly in all patients following methods described in previous studies [13,14]. Data of patients were compared with those of 50 age-matched controls selected in a control population described in our previous studies [ 13,14].
2.2. Imaging technique The MRI investigation was carried out using a whole body MRI-scanner (Siemens Magnetom) operating at 1.5 tesla. Orbital images were acquired on axial and coronal planes with a circular polarized head coil. Three different imaging sequences were acquired. First, a T1-
weighted imaging spin echo sequence was performed on axial plane with a repetition time (TR) of 500 ms and TE of 15 ms. The slices thickness was 4 mm and the number of slices was 13 with an interslice gap of 1 mm. The matrix size was 192 x 256, two acquisitions, with 150 mm diameter field of view (FOV). Second, as STE-STIR sequences were acquired on a coronal plane, using a 150-ms inversion time (TI), a short TE of 22 ms and a relaxation time (TR) of 2500 ms. Nine slices 4 mm thick with 1 mm interslice gap were obtained. Third, LTESTIR sequences were acquired on the coronal plane, using a 150-ms TI, a long TE of 80 ms and a TR of 2500 ms. Nine slices, at the same scan position, with the same slice thickness and interslice gap as the STE-STIR, were obtained using the LTE-STIR sequence. The best compromise among scan time, signal-to-noise ratio (S/N) and spatial resolution was obtained using a matrix size of 128 x 256 (voxel size of 1.1 x 0.6 x 4 mm 3) one acquisition and 150 mm diameter FOV, in both STEand LTE-STIR sequences. Two further sequences were finally acquired in acute ON patients following the i.v. injection of Gd-DTPA (Gadolinium diethylene triamine pentacetic acid), 0.1 mmol/kg body weight. The T1weighted spin echo sequence was repeated on axial and coronal planes, with the same parameters as mentioned above, performing the imaging approximately 10 min after the injection of Gd-DTPA. The STE-STIR sequence was then repeated on the coronal plane with the same parameters as mentioned above. Brain and optic nerve imaging obtained after Gd-DTPA injection was not different from results described by Youl et al. [9] and will not, therefore, be described in the present paper. According to the criteria of Miller et al. [10], five portions of the optic nerve were considered in our descriptions: anterior (up to the mid orbit), intraorbital (from mid orbit to optic canal), intracanalicular (within the optic canal), intracranial and chiasmal portions. In order to test the reliability of the STE-STIR sequences we examined the optic nerves of 20 controls with normal visual acuity, perimetry and VEPs. No lesion was evidenced on STE- a n d LTE-STIR images of these subjects.
2.3. Data analysis The approximate length of ON lesions was calculated by summing the number of slices (4 ram) showing hyperintensity with the interslice gap (1 nun) in coronal plane. The Mann-Whitney test was performed to compare lesion length measured with STE- vs. LTE-STIR sequences in acute and previous ON. Visual acuity, size of visual field loss, VEPs latency and amplitudes were correlated with lesion length in patients by means of the Spearman's Ranks correlation. All tests were performed using the Primer Statistical Program [15].
A. Tartaro et al./ European Journal of Radiology 19 (1995) 155-163
3. R ~
157
appeared slightly hypointense compared to the cerebral cortex and periencephalic CSF (Fig. la). The LTE-STIR sequence shows instead a central hypointense core of the optic nerve, corresponding to myelinated nerve fibers, and a peripheral hyperintense ring corresponding to the perineural subarachnoidal space (Fig. lb). Intracranial subarachnoidal CSF appears hyperintense compared with the hypointense white matter.
Clinical features, electrophysiological and MRI findings of all patients are summarized in Table 1. Twelve patients of group I had unilateral ON, one patient had bilateral ON with abnormal acuity, visual fields and VEPs. Visual field defects, from a previous ON, were evident in 10 eyes of patients of group II. Visual acuity was below < 6/10 only in four eyes of patients of group II, VEPs were abnormal in all patients of group II. Scans of brain and optic nerve obtained with the 22 ms and 80 ms STE- and LTE-STIR sequence in one control subject are shown in Fig. 1. Using the STE-STIR sequence, the optic nerve-perineural space complex
3.1. Acute ON
With the STE-STIR sequence focal MRI lesions with increased signal of the optic nerve were observed in 11/14 (78.5%) optic nerves affected by acute ON. The
Table 1 Clinical and electrophysiological findings in patients with acute ON, in patients previously affected by ON and the follow-up of four patients of t h e first g r o u p ( a c u t e O N )
Acuity L
Acute
Perimetry
VEPs
TE22
R
L
R
L
R
TE80
L
R
L
R
14,a,o
-
ON
1
L
2
10
+
-
+
-
2
R
5
10
+
-
+
.
14,a,o
.
-
.
.
.
3
R
10
4
-
+
-
+
-
14,a,o
-
14,a,o
4
L
5
10
+
-
+
-
-
-
9,a,o,
-
5
L
1
8
+
-
+
-
24,a,o,i
-
24,a,o,i
-
6
L
1
10
+
-
+
-
29,a,o,i
-
29,a,o,i
-
7
L
4
10
+
-
+
-
19,a,o,i
-
19,a,o
-
8
R
10
5
+
-
+
-
-
9,a
-
9,a
9
R
10
2
-
+
-
+
-
6,0
-
6,0
10
R
9
3
-
+
-
+
-
-
-
9,a
II
R
l0
2
-
+
-
+
-
24,a,o
-
24,a,o
12
R
i0
3
-
+
-
+
-
19,a,i
-
19,a,i
13
B
4
5
+
+
+
+
14,a,o
24,i,c
24,a,o
24,i,c
-
Previous ON 14
L
5
10
-
-
+
-
19,a,o
-
24,a,o
15
L
7
9
-
-
+
-
14,o
-
19,o
-
16
L
6
10
-
-
+
-
-
-
9,a
-
17
L
6
10
-
-
+
-
9,c
-
9,c
-
18
R
10
5
-
+
-
+
-
9,c
-
14,i,c
19
R
10
10
-
-
-
+
-
-
-
6,0
20
L
6
10
+
-
+
-
19,o,h
-
24,o,h
-
21 22
R R
10 10
6 8
-
+ +
-
+ +
-
19,o
-
6,a 24,o
23
B
6
6
+
+
+
+
-
9,0
-
14,o
24
B
5
6
+
-
+
+
6,i
-
6,i
14,i,c
25
B
7
5
+
+
+
+
-
9,i
9,o
9,i
26
B
8
10
+
-
+
+
-
6,a
6,a
9,o
6,o
-
-
9,o 14,a
Follow-up
7
L
9
10
+
-
+
-
-
8
R
10
6
-
-
-
+
.
-
9 12
R R
10 10
l0 6
-
+ +
-
+ +
-
Abnormal
f i n d i n g s a r e i n d i c a t e d b y t h e s y m b o l +. T h e n u m b e r r e p o r t e d in t h e c o l u m n s o f S T I R f o r b o t h T E (22 a n d 80 m s ) i n d i c a t e s t h e l e s i o n
.
. 9,a
.
l e n g t h ( m m ) a n d t h e l e t t e r i n d i c a t e s t h e i n v o l v e d a n a t o m i c a l p o r t i o n o f o p t i c n e r v e (a, a n t e r i o r ; o, i n t r a o r b i t a l ; i, i n t r a c a n a l i c u l a r ; c, i n t r a c r a n i a l ; h, c h i a s m a l ) . W h e n a b n o r m a l
s i g n a l is d e t e c t e d i n o n l y o n e slice ( 4 r a m ) , a m i n i m u m v a l u e o f 6 m m (4 + 2 i n t e r s l i c e g a p s ) is e n t e r e d in t h e t a b l e .
158
A. Tartaro et al./ European Journal of Radiology 19 (1995) 155-163
approximate lengths of optic nerve lesions o f the five anatomic portions are reported, for each individual patient, in Table 1. Swelling of the optic nerve with asymmetrical cross-sections increased signal was detected by the S T E - S T I R in 21.4% of optic nerves affected by acute O N (3/14). The mean length of S T E - S T I R measured acute O N lesions was 14.7 4- 7.7 mm. Increased signal intensity or swelling of the nerve with disappearance o f the normal differentiation between nerve fibers and meningeal sheath was seen in all but one (13/14) optic nerve affected by acute O N (92.8%) using the L T E - S T I R sequence. The mean length of the lesions was 17.5 4- 7.6 mm. Table 1 shows a comparison of results obtained with the STE- and L T E - S T I R sequence in each patient. The location of detected lesions was similar in the two
a
sequences (Table 1): in two optic nerves anterior and intraorbital lesions were observed only with LTE- STIR, in two optic nerves lesions measured with L T E - S T I R were longer than lesions measured with STE-STIR. The Mann-Whitney test comparison of lesion lengths, as measured by STE- vs. L T E - S T I R sequence did not reach statistical significance (U = 3.1, P = 0.058). The Spearman's R a n k correlation between approximated lesion length and visual acuity was -0.58, P < 0.03 for STE-STIR and -0.59, P < 0.01 for L T E - S T I R . Fig. 2 shows a Tl-weighted spin echo sequence o f optic nerves in the axial plane in acute right O N (Fig. 2a), and a comparison with coronal STE (Fig. 2b) and L T E (Fig. 2c)STIR imaging. Fig. 3 shows a comparison between STE (Fig. 3a) and L T E (Fig. 3b)-STIR imaging in left acute ON.
b
Fig. 1. Control. Coronal STIR sequence (TR 2500 ms; TI 150 ms; TE 22-80 msec) of the optic nerve obtained in one control subject (male, 31 years old). (a) STE-STIR: the short TE (22 ms) image shows the anterior portion of the optic nerve (arrow). The optic nerve-perineural sheath complex has approximately the same signal intensity as CSF and cerebral cortex. (b) LTE-STIR: long TE (80 ms) sequence of the same portion of the optic nerve. The perineural CSF can be distinguished from a central core (isointense to white matter) corresponding to the myelinated nerve fibers (arrow). Note definition of the dural ring around the optic nerves (open arrow) and the homogeneity of contrast in the two sides of the brain.
A. Tartaro et al./ European Journal of Radiology 19 (1995) 155-163
a
b
159
c
Fig. 2. Acute fight ON (male, 28 years old). Axial plane Tl-weighted imaging (a) is compared with coronal plane STE (b) and LTE (c)-STIR. Note the resolution of the ON lesion with LTE-STIR (arrow). The left optic nerve has a normal appearance with a dural ring only in LTE-STIR images. In STE-STIR images, although a difference between nerve cross-sections is evident, both optic nerves have the same intensity.
a
b
Fig. 3. Acute left ON (female, 33 years old). (a) STE-STIR: note dishomogeneity of left/right structures in STE-STIR images, masking the left optic nerve hyperintensity (arrow). (b) LTE-STIR: note the homogeneous contrast in LTE-STIR and an evident hyperintensity of the left optic nerve (open arrow).
160
A. Tartaro et al./ European Journal of Radiology 19 (1995) 155-163
a
b
c
d
e
f
Fig. 4. Previous ON (female, 29 years old). (a,b,c) The three STE-STIR sequences on the top half show three separate slices of the optic nerves in intraorbital tracts. Hyperintensity of the left optic nerve in intraorbital tract is barely distinguishable from the intraorbital structures (arrow). (d,e,0 The three LTE-STIR sequences on the bottom half show the evident differences between the two nerves (open arrow).
A. Tartaro et al. / European Journal of Radiology 19 (1995) 15S-163
3.2. Previous ON
In the 13 patients of group II VEPs, visual acuity measurements and visual field perimetry detected signs of previous ON in 17 eyes (nine unilateral, four bilateral). Abnormal VEPs corresponding to the affected eye, were also recorded in the four patients of group I undergoing follow-up MRI (Table 1, bottom hal0. STESTIR evidenced increased signal intensity of the affected optic nerve (compared with the other side) in six of nine eyes affected by unilateral previous ON. In patients affected by bilateral previous ON, the STE-STIR sequence evidenced only two hypcrintense signals of one optic nerve: the diagnostic yield of STE-STIR was, therefore, 66% for unilateral previous ON and 50% for bilateral previous ON, resulting in 58.8% overall detection of previous (unilateral or bilateral) ON. The mean length of ON lesion was 8.4 ± 6.1 mm.
a
161
With the LTE-STIR sequence increased signal intensity was observed in 16/17 optic nerves (94.1%) affected by ON. The lesion length was, on average, 14.6 ± 4.2 mm. Lesions were anterior in four nerves (21.1%), intraorhital in eight (42.1%), intracanalicular in three (15.8%), intracranial in three (15.8%) and in the chiasmal portion in one (5.2%). Table 1 shows a comparison of results obtained with STE- and LTE-STIR sequences. The Mann-Whitney test comparison of lesion length between the STE- and LTE-STIR in patients with previous ON showed significant differences (U = 2.2, P = 0.039). Fig. 4 shows three slices obtained in the intraorbital portion of the optic nerve with STE (Fig. 4a,b,c) and LTE-STIR (Fig. 4d,e,f). Fig. 5 shows imaging of a previous left ON and a demyelinating plaque of the right frontal white matter. In all patients, VEP abnormalities
b
Fig. 5. Previous ON. Left ON (female, 39 years old). The anterior portion of nerve is shown. (a) No abnormality is seen in the short TE images. Both optic nerves display the same thickness and signal intensities. In the white matter of the frontal lobe there is a lesion which appears isointense to CSF and cortex (arrow). (b) In the long TE image the left optic nerve is hyperintense in comparison with the fight optic nerve and the cerebral cortex (open arrow). The signal intensity of the affected nerve displays the same signal intensity as the frontal lobe lesion.
162
A. Tartaro et al./ European Journal of Radiology 19 (1995) 155-163
consisted of the absence of identifiable responses or of latency increments and amplitude reductions of VEP components, as reported in previous studies [9]. The latency and amplitude of VEPs did not correlate with the size of optic nerve lesions in acute ON. The amplitude of VEPs was inversely proportional to lesion size in acute and previous ON. VEP/MRI correlations will be discussed in detail in a subsequent study. 4. Discussion
Our study shows that the LTE-STIR sequence is more effective than the STE-STIR sequence (corresponding to the conventional MRI STIR sequence) in showing acute or previous lesions of the optic nerve in patients affected by ON. The diagnostic yields obtained with STE-STIR sequences in our patients concurs with results reported in literature [7]. The LTE-STIR sequence gives an improved contrast resolution of ON increasing by 14% the detection of abnormality in acute ON and by 35% the detection of previous ON, where the altered signal is probably due to glial reorganization. The detection of ON with STE-STIR approaches therefore the diagnostic yield, reported with a new method of fat-saturation sequences by Miller et al. in 1993 in seven patients [16]. Furthermore, altered signals are consistently detected in more slices with LTE-STIR than with conventional STE-STIR sequence (Table 1, Figs. 3-6). Although we are aware that precise measurement of the optic nerve is considered barely reliable because of the oblique position of the nerve that lies 5-10% off a true perpendicular axis, the finding that significant differences in the length of lesions were observed (Table 1) consistently in favor of LTE-STIR images further indicates that LTE-STIR sequence is an extremely sensitive tool for ON imaging. The conventional STE-STIR technique enables to visualize the optic nerve by removing all signals arising from the orbital fat and thus represents a technical advance for the identification of ON [9,10]. The STE-STIR images are characterized by signal intensities very similar for optic nerve (nerve fibers and perineural space), cerebral cortex, CSF and white matter lesions, so that the presence of edema, demyelination or proliferative glial events is not always identifiable. A further technical problem lies in the non-homogeneous detection of the signal by the head coil, due to the proximity of one hemisphere to the head coil, that render STE-STIR more sensitive to magnetic field dishomogeneties, sometimes resulting in a left-right difference of signal intensity of intracranial and intraorbital structures [3]. This may lead to misinterpretation of MRI images, particularly when comparing signals of the two optic nerves, or of each nerve with other parts of the brain. The LTE-STIR images of normal optic nerves (Fig. l b) by approximating a T2 contrast [11,12] can allow a distinction between meningeal sheath and nerve fibers,
while cerebral cortex, CSF and white matter have different signal intensity and are homogeneously represented in the two sides of the head. In patients of groups I and II, lesions were easily identified, based on a comparison of the two optic nerves and of each nerve with other intracranial structures. Purposely, in Figs. 3-5 we compare results in three patients in whom ON could not be clearly identified by STE-STIR sequence, because of the isointensity of nerve, cortex, CSF, white matter, or because of a left-right difference of intensity of intracranial or orbital structures. In the same patients clear abnormalities of the optic nerve were detected with the LTE-STIR sequence. In conclusion, the LTE-STIR sequence warrants a resolution of ON lesions that is superior to previously described STE-STIR techniques. The identification of optic nerve abnormalities in patients who suffered ON several months before the MRI study, reaches 90% with the LTE-STIR sequence, thus approximating VEP resolution [9]. We suggest that LTE-STIR sequence studies of ON might allow a comparison of clinical, and VEP measurements with anatomical imaging. References [I] McAlpine D.Th¢ benign form of multiple sclerosis. A study based on 241 eases seen within three years of onset and followed up until the tenth year or more of disease. Brain 1961; 84: 86-203. [2] Poser CM, Paty DW, Scheinberg L, McDonald I, Davis FA, Johnson KP, Sibley WA, Silbernerg DH, Tourteliotte WW. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol 1983; 13: 227-231. [3l Atlas SW, Grossman RI, Hackney DB, Goldberg HI, Bilaniuk LT, Zimmermann RA. STIR-MR imaging of the orbit. A JR 1988; 15: 1025-1030. [4] Larsson HBW, Thomesen C, Frederiksen J, Henriksen O, Olesen J. Chemical shift selective magnetic resonance imaging of the optic nerve in patients with acute optic neuritis. Acta Radiol 1988; 29: 629-632. [5] Guy J, Mao J, Bidgood JR, Mancuso A, Quisling RG. Enhancement and demyelination of the intraorbital optic nerve. Fat suppression magnetic resonance imaging. Ophthalmol 1992; 99: 713-719. [6l Daniels DL, Kneeland JB, Shimakawa A, Pojunas KW, Schenk JF, Hart H Jr., Foster T, Williams AL, Haughton VM. MR imaging of the optic nerve and sheath: correcting the chemical shift misregistration effect. AJNR 1986; 7: 249-253. [7] Johnson G, Miller DH, MacManus D, Toffs IS, Barnes D, du Boulay EP, McDonald WI. STIR sequences in NMR imaging of the optic nerve. Neuroradiology 1987; 29: 238-245. [8] Simon J, Szumowski J, Totterman S, Kido D, Ekholm S, Wicks A, Plewes D. Fat-suppression MR imaging of the orbit. AJNR 1988; 9: 961-968. [9] Youl BD, Turano G, Miller DH, Towell AD, MacManus DG, Moore SG, Jones SJ, Barrett G, Kendall DG, Moseley IF, Toffs PS, Halliday AM, McDonald WI. The pathophysiology of acute optic neuritis. An association of Gadolinium leakage with clinical and electrophysiological deficits. Brain 1991; 114: 2437-2450. [10] Miller DH, Newton MR, van der Poel JC, du Boulay GH, Halliday AM, Kendall BE, Johnson G, MacManus DG, Moseley IF,
A. Tartaro et al./ European Journal of Radiology 19 (1995) 155-163 Mc Donald WI. Magnetic resonance imaging of the optic nerve in optic neuritis. Neuroradiology 1988; 38: 175-179. [11] Hendrick RE, RoffU. Image contrast and noise. In: Stark DD, Bredley JR, editors. Magnetic resonace imaging, lind ed., Vol. 1. St. Louis, MO: Mosby Year Book, 1992; 109-144. [12] Finn JP, Kendal BE, Kingsley DMK, HalI-Craggs MA, Connelly A. Myelination, structural abnormalities and disease of white matter an:assessment using high field MRI. In: du Boulay G, Molyneux A, Moseley IF, editors. Proceedings of the XIV Symposium Neuroradiologicum, London. Neuroradioiogy 1991; 33 (suppl): 254-256. [13] Onofrj M, Bazzano S, Malatesta G, Fulgente T. Mapped distribution of pattern reversal VEPs to central field and lateral half-
163
field stimuli of different spatial frequencies. Electroenceph Clin Neurophysiol 1991; 80: 167-180. [14] Onofrj M, Fulgente T, Malatesta G, Ferracci F. Visual Evoked Potentials (VEPs) to altitudinal stimuli: effects of stimulus manipulations on VEP scalp topography. Clin Vision Sci 1993; 8: 529-544. [15] Glantz SA. Statistica per discipline biomediche. 2° edizione. McGraw-Hill Libri Italia srl, 1988. [16] Miller DH, MacManus DG, Bartlett PA, Kapoor R, Morrissey SP, Moseley IF. Detection of optic nerve lesions in optic neuritis using frequency-selective fat-saturation sequences. Neuroradiology 1993; 35: 156-158.
@ ELSEVIER
European Journal of Radiology 19 (1995) 155-163
Long time echo stir sequence magnetic resonance imaging of optic nerves in optic neuritis A . T a r t a r o b, M . O n o f r j *a, A . T h o m a s a, T. F u l g e n t e a, C. D e l l i P i z z i b, L. B o n o m o b aDepartment of Neurology, State University of Chieti, Ospedale ex-Pediatrico, Via Martiri Lancianesi, 6, 66100 Chieti, Italy bDepartment of Radiology, State University of Chieti, Chieti, Italy
Received 24 October 1994; revision received 7 December 1994;accepted 12 December 1994
Abstract
Magnetic resonance imaging of optic nerves was obtained in 13 patients with acute optic neuritis and in 13 patients with a previous history of optic neuritis (ON), assessed by clinical, visual fields and visual evoked potentials evaluations. Results of the conventional short tau inversion recovery (STIR) sequence obtained with short time echo (STE-STIR: 22 ms) were compared with long time echo (LTE-STIR: 80 ms) sequence. The conventional STE-STIR sequence revealed lesions in 78.5% of acute ON and in 58.8% of optic nerves affected by previous ON. The LTE-STIR sequence was diagnostic in 92.8% of acutely symptomatic nerves, in 94.1.% of nerves with previous ON. The calculated length of optic nerve lesions was significantly longer in imaging obtained with the LTE-STIR sequence than with the conventional STE-STIR sequences, both in acute and previous ON. Keywords: Magnetic resonance (MR), technique; Magnetic resonance (MR), brain; Nerves, optic; Multiple sclerosis
1. Introduction
Optic neuritis (ON) is common in multiple sclerosis (MS), appearing as a condition that precedes MS or as a symptomatic lesion during the course of 'definite' MS [1,21. Despite the high sensitivity of magnetic resonance imaging (MRI) in detecting brain lesions in patients with MS, the initial attempts to detect ON were disappointing [3-5] because of the chemical shift artifact between optic nerve and orbital fat observed with T 1 - T 2 weighted sequences. The high proton density and short T 1 relaxation time of orbital fat results in a high signal intensity concealing intraorbital structures and contributing to partial volume averaging artifacts. These problems were partly overcome by the introduction of the short tau inversion recovery (STIR) sequence with the conventional short time echo (STESTIR) [6-8]. STE-STIR sequence evidenced increased signal intensity of optic nerves in 70% of acute ON, and
* Corresponding author.
in 20-50% of optic nerves studied when the acute ON had recovered [9,101. STE-STIR sequence, with inversion time (T1) at the 'null point' of fatty tissue, results in a signal void of the orbit fat: the normal optic nerve however, appears relatively hyperintense and clear separation between optic nerve and perioptic cerebrospinal fluid (CSF) is lost on STE-STIR images, because in sequences obtained with short T1 ( < 150 ms), image contrast is dependent on T 1 and T2 [11]. As already discussed by Hendrick and Roff [11]: 'although it is important to remember that image contrast in... short... STIR domain is due to the additive effects of correlated T1, T2 and N[H] inherent contrast', with a TE reaching 80 ms it is possible to obtain 'a more T2 weighted appearance' and to obtain therefore a contrast differentiation between edema/gliosis and cerebral nervous system structures. Based on this theoretical issue, Finn et al. [12] introduced a double time echo STIR sequence (STE: 22 ms and LTE: 80 ms) and obtained an increment of the signal to noise ratio in comparison with the spin echo T2-weighted images of white matter lesions.
0720-048X/95/$09.50 © 1995 Elsevier ScienceIreland Ltd. All rights reserved SSDI 0720-048X(94)00600-H
156
A. Tartaro et al./ European Journal of Radiology 19 (1995) 155-163
In the present study we hypothesized that the use of a double time echo STIR sequence, with an unconventional LTE of 80 ms, might enhance the contrast resolution of optic nerve lesions in ON. We compared the sensitivity of STIR sequences with an STE of 22 ms and an LTE of 80 ms in detecting lesions of optic nerves in 26 patients with acute or previous ON assessed by clinical, neuro-ophthalmological examinations and visual evoked potential (VEP) recordings.
2. Subjects and methods
2.1. Study population Twenty-six patients (age range 21-52 years; mean age 36.1 ± 9.6 years, 16 women) affected by acute or previous episodes of ON were recruited for the study. All patients underwent fundus examination, assessment of visual acuity by means of Snellen Charts, Ishihara Plates, visual field testing, pattern reversal VEP recordings and MRI of brain and optic nerve. Patients were divided into two groups: group I (13 patients) ineluded patients referred to our department because of the sudden onset of unilateral or bilateral (one patient) ocular pain and blurred vision. MRI studies were performed 3-6 days after the onset of symptoms, and were repeated 3 months after the onset of ON in four patients. Group II included 13 patients who had recovered from an ON occurring 4-18 months before MRI. Three of these patients had a history of ON of the other side, 18-60 months before the last ON attack. All the 13 patients of group II could be classified as definite MS according to McAlpine's criteria [11. In both groups of patients normal visual acuity, normal computerized perimetry and normal VEPs excluded the presence of ON. Monocular visual fields were tested by means of a computerized 640 Humphrey perimeter (Zeiss Humphrey) exploring the central 30° of the retina, and were classified as normal or abnormal compared with the normative age-matched data. Full field pattern reversal VEPs were recorded monocularly in all patients following methods described in previous studies [13,14]. Data of patients were compared with those of 50 age-matched controls selected in a control population described in our previous studies [ 13,14].
2.2. Imaging technique The MRI investigation was carried out using a whole body MRI-scanner (Siemens Magnetom) operating at 1.5 tesla. Orbital images were acquired on axial and coronal planes with a circular polarized head coil. Three different imaging sequences were acquired. First, a T1-
weighted imaging spin echo sequence was performed on axial plane with a repetition time (TR) of 500 ms and TE of 15 ms. The slices thickness was 4 mm and the number of slices was 13 with an interslice gap of 1 mm. The matrix size was 192 x 256, two acquisitions, with 150 mm diameter field of view (FOV). Second, as STE-STIR sequences were acquired on a coronal plane, using a 150-ms inversion time (TI), a short TE of 22 ms and a relaxation time (TR) of 2500 ms. Nine slices 4 mm thick with 1 mm interslice gap were obtained. Third, LTESTIR sequences were acquired on the coronal plane, using a 150-ms TI, a long TE of 80 ms and a TR of 2500 ms. Nine slices, at the same scan position, with the same slice thickness and interslice gap as the STE-STIR, were obtained using the LTE-STIR sequence. The best compromise among scan time, signal-to-noise ratio (S/N) and spatial resolution was obtained using a matrix size of 128 x 256 (voxel size of 1.1 x 0.6 x 4 mm 3) one acquisition and 150 mm diameter FOV, in both STEand LTE-STIR sequences. Two further sequences were finally acquired in acute ON patients following the i.v. injection of Gd-DTPA (Gadolinium diethylene triamine pentacetic acid), 0.1 mmol/kg body weight. The T1weighted spin echo sequence was repeated on axial and coronal planes, with the same parameters as mentioned above, performing the imaging approximately 10 min after the injection of Gd-DTPA. The STE-STIR sequence was then repeated on the coronal plane with the same parameters as mentioned above. Brain and optic nerve imaging obtained after Gd-DTPA injection was not different from results described by Youl et al. [9] and will not, therefore, be described in the present paper. According to the criteria of Miller et al. [10], five portions of the optic nerve were considered in our descriptions: anterior (up to the mid orbit), intraorbital (from mid orbit to optic canal), intracanalicular (within the optic canal), intracranial and chiasmal portions. In order to test the reliability of the STE-STIR sequences we examined the optic nerves of 20 controls with normal visual acuity, perimetry and VEPs. No lesion was evidenced on STE- a n d LTE-STIR images of these subjects.
2.3. Data analysis The approximate length of ON lesions was calculated by summing the number of slices (4 ram) showing hyperintensity with the interslice gap (1 nun) in coronal plane. The Mann-Whitney test was performed to compare lesion length measured with STE- vs. LTE-STIR sequences in acute and previous ON. Visual acuity, size of visual field loss, VEPs latency and amplitudes were correlated with lesion length in patients by means of the Spearman's Ranks correlation. All tests were performed using the Primer Statistical Program [15].
A. Tartaro et al./ European Journal of Radiology 19 (1995) 155-163
3. R ~
157
appeared slightly hypointense compared to the cerebral cortex and periencephalic CSF (Fig. la). The LTE-STIR sequence shows instead a central hypointense core of the optic nerve, corresponding to myelinated nerve fibers, and a peripheral hyperintense ring corresponding to the perineural subarachnoidal space (Fig. lb). Intracranial subarachnoidal CSF appears hyperintense compared with the hypointense white matter.
Clinical features, electrophysiological and MRI findings of all patients are summarized in Table 1. Twelve patients of group I had unilateral ON, one patient had bilateral ON with abnormal acuity, visual fields and VEPs. Visual field defects, from a previous ON, were evident in 10 eyes of patients of group II. Visual acuity was below < 6/10 only in four eyes of patients of group II, VEPs were abnormal in all patients of group II. Scans of brain and optic nerve obtained with the 22 ms and 80 ms STE- and LTE-STIR sequence in one control subject are shown in Fig. 1. Using the STE-STIR sequence, the optic nerve-perineural space complex
3.1. Acute ON
With the STE-STIR sequence focal MRI lesions with increased signal of the optic nerve were observed in 11/14 (78.5%) optic nerves affected by acute ON. The
Table 1 Clinical and electrophysiological findings in patients with acute ON, in patients previously affected by ON and the follow-up of four patients of t h e first g r o u p ( a c u t e O N )
Acuity L
Acute
Perimetry
VEPs
TE22
R
L
R
L
R
TE80
L
R
L
R
14,a,o
-
ON
1
L
2
10
+
-
+
-
2
R
5
10
+
-
+
.
14,a,o
.
-
.
.
.
3
R
10
4
-
+
-
+
-
14,a,o
-
14,a,o
4
L
5
10
+
-
+
-
-
-
9,a,o,
-
5
L
1
8
+
-
+
-
24,a,o,i
-
24,a,o,i
-
6
L
1
10
+
-
+
-
29,a,o,i
-
29,a,o,i
-
7
L
4
10
+
-
+
-
19,a,o,i
-
19,a,o
-
8
R
10
5
+
-
+
-
-
9,a
-
9,a
9
R
10
2
-
+
-
+
-
6,0
-
6,0
10
R
9
3
-
+
-
+
-
-
-
9,a
II
R
l0
2
-
+
-
+
-
24,a,o
-
24,a,o
12
R
i0
3
-
+
-
+
-
19,a,i
-
19,a,i
13
B
4
5
+
+
+
+
14,a,o
24,i,c
24,a,o
24,i,c
-
Previous ON 14
L
5
10
-
-
+
-
19,a,o
-
24,a,o
15
L
7
9
-
-
+
-
14,o
-
19,o
-
16
L
6
10
-
-
+
-
-
-
9,a
-
17
L
6
10
-
-
+
-
9,c
-
9,c
-
18
R
10
5
-
+
-
+
-
9,c
-
14,i,c
19
R
10
10
-
-
-
+
-
-
-
6,0
20
L
6
10
+
-
+
-
19,o,h
-
24,o,h
-
21 22
R R
10 10
6 8
-
+ +
-
+ +
-
19,o
-
6,a 24,o
23
B
6
6
+
+
+
+
-
9,0
-
14,o
24
B
5
6
+
-
+
+
6,i
-
6,i
14,i,c
25
B
7
5
+
+
+
+
-
9,i
9,o
9,i
26
B
8
10
+
-
+
+
-
6,a
6,a
9,o
6,o
-
-
9,o 14,a
Follow-up
7
L
9
10
+
-
+
-
-
8
R
10
6
-
-
-
+
.
-
9 12
R R
10 10
l0 6
-
+ +
-
+ +
-
Abnormal
f i n d i n g s a r e i n d i c a t e d b y t h e s y m b o l +. T h e n u m b e r r e p o r t e d in t h e c o l u m n s o f S T I R f o r b o t h T E (22 a n d 80 m s ) i n d i c a t e s t h e l e s i o n
.
. 9,a
.
l e n g t h ( m m ) a n d t h e l e t t e r i n d i c a t e s t h e i n v o l v e d a n a t o m i c a l p o r t i o n o f o p t i c n e r v e (a, a n t e r i o r ; o, i n t r a o r b i t a l ; i, i n t r a c a n a l i c u l a r ; c, i n t r a c r a n i a l ; h, c h i a s m a l ) . W h e n a b n o r m a l
s i g n a l is d e t e c t e d i n o n l y o n e slice ( 4 r a m ) , a m i n i m u m v a l u e o f 6 m m (4 + 2 i n t e r s l i c e g a p s ) is e n t e r e d in t h e t a b l e .
158
A. Tartaro et al./ European Journal of Radiology 19 (1995) 155-163
approximate lengths of optic nerve lesions o f the five anatomic portions are reported, for each individual patient, in Table 1. Swelling of the optic nerve with asymmetrical cross-sections increased signal was detected by the S T E - S T I R in 21.4% of optic nerves affected by acute O N (3/14). The mean length of S T E - S T I R measured acute O N lesions was 14.7 4- 7.7 mm. Increased signal intensity or swelling of the nerve with disappearance o f the normal differentiation between nerve fibers and meningeal sheath was seen in all but one (13/14) optic nerve affected by acute O N (92.8%) using the L T E - S T I R sequence. The mean length of the lesions was 17.5 4- 7.6 mm. Table 1 shows a comparison of results obtained with the STE- and L T E - S T I R sequence in each patient. The location of detected lesions was similar in the two
a
sequences (Table 1): in two optic nerves anterior and intraorbital lesions were observed only with LTE- STIR, in two optic nerves lesions measured with L T E - S T I R were longer than lesions measured with STE-STIR. The Mann-Whitney test comparison of lesion lengths, as measured by STE- vs. L T E - S T I R sequence did not reach statistical significance (U = 3.1, P = 0.058). The Spearman's R a n k correlation between approximated lesion length and visual acuity was -0.58, P < 0.03 for STE-STIR and -0.59, P < 0.01 for L T E - S T I R . Fig. 2 shows a Tl-weighted spin echo sequence o f optic nerves in the axial plane in acute right O N (Fig. 2a), and a comparison with coronal STE (Fig. 2b) and L T E (Fig. 2c)STIR imaging. Fig. 3 shows a comparison between STE (Fig. 3a) and L T E (Fig. 3b)-STIR imaging in left acute ON.
b
Fig. 1. Control. Coronal STIR sequence (TR 2500 ms; TI 150 ms; TE 22-80 msec) of the optic nerve obtained in one control subject (male, 31 years old). (a) STE-STIR: the short TE (22 ms) image shows the anterior portion of the optic nerve (arrow). The optic nerve-perineural sheath complex has approximately the same signal intensity as CSF and cerebral cortex. (b) LTE-STIR: long TE (80 ms) sequence of the same portion of the optic nerve. The perineural CSF can be distinguished from a central core (isointense to white matter) corresponding to the myelinated nerve fibers (arrow). Note definition of the dural ring around the optic nerves (open arrow) and the homogeneity of contrast in the two sides of the brain.
A. Tartaro et al./ European Journal of Radiology 19 (1995) 155-163
a
b
159
c
Fig. 2. Acute fight ON (male, 28 years old). Axial plane Tl-weighted imaging (a) is compared with coronal plane STE (b) and LTE (c)-STIR. Note the resolution of the ON lesion with LTE-STIR (arrow). The left optic nerve has a normal appearance with a dural ring only in LTE-STIR images. In STE-STIR images, although a difference between nerve cross-sections is evident, both optic nerves have the same intensity.
a
b
Fig. 3. Acute left ON (female, 33 years old). (a) STE-STIR: note dishomogeneity of left/right structures in STE-STIR images, masking the left optic nerve hyperintensity (arrow). (b) LTE-STIR: note the homogeneous contrast in LTE-STIR and an evident hyperintensity of the left optic nerve (open arrow).
160
A. Tartaro et al./ European Journal of Radiology 19 (1995) 155-163
a
b
c
d
e
f
Fig. 4. Previous ON (female, 29 years old). (a,b,c) The three STE-STIR sequences on the top half show three separate slices of the optic nerves in intraorbital tracts. Hyperintensity of the left optic nerve in intraorbital tract is barely distinguishable from the intraorbital structures (arrow). (d,e,0 The three LTE-STIR sequences on the bottom half show the evident differences between the two nerves (open arrow).
A. Tartaro et al. / European Journal of Radiology 19 (1995) 15S-163
3.2. Previous ON
In the 13 patients of group II VEPs, visual acuity measurements and visual field perimetry detected signs of previous ON in 17 eyes (nine unilateral, four bilateral). Abnormal VEPs corresponding to the affected eye, were also recorded in the four patients of group I undergoing follow-up MRI (Table 1, bottom hal0. STESTIR evidenced increased signal intensity of the affected optic nerve (compared with the other side) in six of nine eyes affected by unilateral previous ON. In patients affected by bilateral previous ON, the STE-STIR sequence evidenced only two hypcrintense signals of one optic nerve: the diagnostic yield of STE-STIR was, therefore, 66% for unilateral previous ON and 50% for bilateral previous ON, resulting in 58.8% overall detection of previous (unilateral or bilateral) ON. The mean length of ON lesion was 8.4 ± 6.1 mm.
a
161
With the LTE-STIR sequence increased signal intensity was observed in 16/17 optic nerves (94.1%) affected by ON. The lesion length was, on average, 14.6 ± 4.2 mm. Lesions were anterior in four nerves (21.1%), intraorhital in eight (42.1%), intracanalicular in three (15.8%), intracranial in three (15.8%) and in the chiasmal portion in one (5.2%). Table 1 shows a comparison of results obtained with STE- and LTE-STIR sequences. The Mann-Whitney test comparison of lesion length between the STE- and LTE-STIR in patients with previous ON showed significant differences (U = 2.2, P = 0.039). Fig. 4 shows three slices obtained in the intraorbital portion of the optic nerve with STE (Fig. 4a,b,c) and LTE-STIR (Fig. 4d,e,f). Fig. 5 shows imaging of a previous left ON and a demyelinating plaque of the right frontal white matter. In all patients, VEP abnormalities
b
Fig. 5. Previous ON. Left ON (female, 39 years old). The anterior portion of nerve is shown. (a) No abnormality is seen in the short TE images. Both optic nerves display the same thickness and signal intensities. In the white matter of the frontal lobe there is a lesion which appears isointense to CSF and cortex (arrow). (b) In the long TE image the left optic nerve is hyperintense in comparison with the fight optic nerve and the cerebral cortex (open arrow). The signal intensity of the affected nerve displays the same signal intensity as the frontal lobe lesion.
162
A. Tartaro et al./ European Journal of Radiology 19 (1995) 155-163
consisted of the absence of identifiable responses or of latency increments and amplitude reductions of VEP components, as reported in previous studies [9]. The latency and amplitude of VEPs did not correlate with the size of optic nerve lesions in acute ON. The amplitude of VEPs was inversely proportional to lesion size in acute and previous ON. VEP/MRI correlations will be discussed in detail in a subsequent study. 4. Discussion
Our study shows that the LTE-STIR sequence is more effective than the STE-STIR sequence (corresponding to the conventional MRI STIR sequence) in showing acute or previous lesions of the optic nerve in patients affected by ON. The diagnostic yields obtained with STE-STIR sequences in our patients concurs with results reported in literature [7]. The LTE-STIR sequence gives an improved contrast resolution of ON increasing by 14% the detection of abnormality in acute ON and by 35% the detection of previous ON, where the altered signal is probably due to glial reorganization. The detection of ON with STE-STIR approaches therefore the diagnostic yield, reported with a new method of fat-saturation sequences by Miller et al. in 1993 in seven patients [16]. Furthermore, altered signals are consistently detected in more slices with LTE-STIR than with conventional STE-STIR sequence (Table 1, Figs. 3-6). Although we are aware that precise measurement of the optic nerve is considered barely reliable because of the oblique position of the nerve that lies 5-10% off a true perpendicular axis, the finding that significant differences in the length of lesions were observed (Table 1) consistently in favor of LTE-STIR images further indicates that LTE-STIR sequence is an extremely sensitive tool for ON imaging. The conventional STE-STIR technique enables to visualize the optic nerve by removing all signals arising from the orbital fat and thus represents a technical advance for the identification of ON [9,10]. The STE-STIR images are characterized by signal intensities very similar for optic nerve (nerve fibers and perineural space), cerebral cortex, CSF and white matter lesions, so that the presence of edema, demyelination or proliferative glial events is not always identifiable. A further technical problem lies in the non-homogeneous detection of the signal by the head coil, due to the proximity of one hemisphere to the head coil, that render STE-STIR more sensitive to magnetic field dishomogeneties, sometimes resulting in a left-right difference of signal intensity of intracranial and intraorbital structures [3]. This may lead to misinterpretation of MRI images, particularly when comparing signals of the two optic nerves, or of each nerve with other parts of the brain. The LTE-STIR images of normal optic nerves (Fig. l b) by approximating a T2 contrast [11,12] can allow a distinction between meningeal sheath and nerve fibers,
while cerebral cortex, CSF and white matter have different signal intensity and are homogeneously represented in the two sides of the head. In patients of groups I and II, lesions were easily identified, based on a comparison of the two optic nerves and of each nerve with other intracranial structures. Purposely, in Figs. 3-5 we compare results in three patients in whom ON could not be clearly identified by STE-STIR sequence, because of the isointensity of nerve, cortex, CSF, white matter, or because of a left-right difference of intensity of intracranial or orbital structures. In the same patients clear abnormalities of the optic nerve were detected with the LTE-STIR sequence. In conclusion, the LTE-STIR sequence warrants a resolution of ON lesions that is superior to previously described STE-STIR techniques. The identification of optic nerve abnormalities in patients who suffered ON several months before the MRI study, reaches 90% with the LTE-STIR sequence, thus approximating VEP resolution [9]. We suggest that LTE-STIR sequence studies of ON might allow a comparison of clinical, and VEP measurements with anatomical imaging. References [I] McAlpine D.Th¢ benign form of multiple sclerosis. A study based on 241 eases seen within three years of onset and followed up until the tenth year or more of disease. Brain 1961; 84: 86-203. [2] Poser CM, Paty DW, Scheinberg L, McDonald I, Davis FA, Johnson KP, Sibley WA, Silbernerg DH, Tourteliotte WW. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol 1983; 13: 227-231. [3l Atlas SW, Grossman RI, Hackney DB, Goldberg HI, Bilaniuk LT, Zimmermann RA. STIR-MR imaging of the orbit. A JR 1988; 15: 1025-1030. [4] Larsson HBW, Thomesen C, Frederiksen J, Henriksen O, Olesen J. Chemical shift selective magnetic resonance imaging of the optic nerve in patients with acute optic neuritis. Acta Radiol 1988; 29: 629-632. [5] Guy J, Mao J, Bidgood JR, Mancuso A, Quisling RG. Enhancement and demyelination of the intraorbital optic nerve. Fat suppression magnetic resonance imaging. Ophthalmol 1992; 99: 713-719. [6l Daniels DL, Kneeland JB, Shimakawa A, Pojunas KW, Schenk JF, Hart H Jr., Foster T, Williams AL, Haughton VM. MR imaging of the optic nerve and sheath: correcting the chemical shift misregistration effect. AJNR 1986; 7: 249-253. [7] Johnson G, Miller DH, MacManus D, Toffs IS, Barnes D, du Boulay EP, McDonald WI. STIR sequences in NMR imaging of the optic nerve. Neuroradiology 1987; 29: 238-245. [8] Simon J, Szumowski J, Totterman S, Kido D, Ekholm S, Wicks A, Plewes D. Fat-suppression MR imaging of the orbit. AJNR 1988; 9: 961-968. [9] Youl BD, Turano G, Miller DH, Towell AD, MacManus DG, Moore SG, Jones SJ, Barrett G, Kendall DG, Moseley IF, Toffs PS, Halliday AM, McDonald WI. The pathophysiology of acute optic neuritis. An association of Gadolinium leakage with clinical and electrophysiological deficits. Brain 1991; 114: 2437-2450. [10] Miller DH, Newton MR, van der Poel JC, du Boulay GH, Halliday AM, Kendall BE, Johnson G, MacManus DG, Moseley IF,
A. Tartaro et al./ European Journal of Radiology 19 (1995) 155-163 Mc Donald WI. Magnetic resonance imaging of the optic nerve in optic neuritis. Neuroradiology 1988; 38: 175-179. [11] Hendrick RE, RoffU. Image contrast and noise. In: Stark DD, Bredley JR, editors. Magnetic resonace imaging, lind ed., Vol. 1. St. Louis, MO: Mosby Year Book, 1992; 109-144. [12] Finn JP, Kendal BE, Kingsley DMK, HalI-Craggs MA, Connelly A. Myelination, structural abnormalities and disease of white matter an:assessment using high field MRI. In: du Boulay G, Molyneux A, Moseley IF, editors. Proceedings of the XIV Symposium Neuroradiologicum, London. Neuroradioiogy 1991; 33 (suppl): 254-256. [13] Onofrj M, Bazzano S, Malatesta G, Fulgente T. Mapped distribution of pattern reversal VEPs to central field and lateral half-
163
field stimuli of different spatial frequencies. Electroenceph Clin Neurophysiol 1991; 80: 167-180. [14] Onofrj M, Fulgente T, Malatesta G, Ferracci F. Visual Evoked Potentials (VEPs) to altitudinal stimuli: effects of stimulus manipulations on VEP scalp topography. Clin Vision Sci 1993; 8: 529-544. [15] Glantz SA. Statistica per discipline biomediche. 2° edizione. McGraw-Hill Libri Italia srl, 1988. [16] Miller DH, MacManus DG, Bartlett PA, Kapoor R, Morrissey SP, Moseley IF. Detection of optic nerve lesions in optic neuritis using frequency-selective fat-saturation sequences. Neuroradiology 1993; 35: 156-158.