High-Resolution Computed Tomography in Multiple Sclerosis

Lynn Barrett, MD,” Burton Drayer, MD,’? and Chelsu Shin, MDJ-

Eighty-five patients were classified aq having definite (n = 34), probable (n = 18), or possible (n = 33) multiple sclerosis using the criteria of Poser. Each patient had an enhanced computed tomographic examination, and most had cerebrospinal fluid and evoked response studies at the same time. Abnormalities including focal decreased brain density, abnormal enhancement, and cerebrospinal fluid space enlargement were found in 62% of patients (85% of those with definite, 39% of those with probable, and 52% of those with possible multiple sclerosis), and abnormal enhancing areas were demonstrated in 29% (44% of those with definite, 17% of those with probable, and 21% of those with possible disease). In the subgroup of patients with definite multiple sclerosis and recent clinical exacerbation (within the prior 8 weeks), abnormal enhancement was present in 89%. There was a strong correlation between clinical exacerbation and abnormal contrast enhancement. Cerebrospinal fluid studies (IgG, white blood cell count, total protein) had no correlation with exacerbation or abnormal enhancement. Maps of low-density and enhancing areas were similar to those previously described in postmortem studies. Computed tomography thus provides an in vivo, objective, and anatomically specific map of the brain parenchyma and the integrity of the blood-brain barrier that is useful in research studies evaluating the treatment of multiple sclerosis. It is also useful in patients in whom the diagnosis of multiple sclerosis is suspected but not certain on the basis of clinical and laboratory evaluation.

Barrett I., Drayer B, Shin C: High-resolution computed tomography in multiple sclerosis. Ann Neurol 17133-38, 1985

The diagnosis of multiple sclerosis (MS) depends on a careful history taking and clinical examination. Affir- mative data are obtained from evaluation of evoked responses and cerebrospinal fluid (CSF). Computed tomography (CT) has become an important supplemen- tary tool in the diagnosis of MS and may prove useful in therapeutic trials because it provides an in vivo, ana- tomically specific map of the brain parenchyma and the integrity of the blood-brain barrier. A comparison of the accuracy of contemporary scanning with that of other diagnostic tools should prove useful in the design and interpretation of future therapeutic trials. In addition, a study of the ability of high-resolution CT to detect intracerebral lesions in a large group of patients with MS will provide an important data base for comparison with the results of future studies of new imaging technologies (e.g., nuclear magnetic resonance).

Methods A retrospective review was made of 85 patients with known or suspected MS evaluated at Duke University Medical Cen- ter from January 1979 through February 1983. Each patient had an enhanced CT scan; most patients also had an unen- hanced study. Routine enhancement (42 gm of iodine) was

used in 30 patients; the remaining 5 5 individuals had a dou- ble-dose (84 gm of iodine) study with delayed (30- to 60- minute) imaging. Scans were performed on a GE 8800, Siemens Somatom 2, or Pfizer 450 unit using standard radio- graphic factors and 4 to 10 mm collimation.

Patients were classified as having definite, probable, or possible MS on the basis of clinical signs and symptoms, first symptom, and age at onset (Poser classification [GI). En- hanced and unenhanced CT scans were reviewed for evi- dence of low-density areas, enhancing lesions, and size of the ventricles and subarachnoid spaces. In the last 46 scans re- viewed, low-density and enhancing areas were entered onto anatomical maps, each lesion being represented by a single dot.

CSF studies (cells, protein, IgG), CT scans, and evoked response evaluations were compared in patients with defi- nite, probable, and possible MS. Relationships among dura- tion of disease, onset of disease, CT and CSF abnormali- ties, and patient classification were examined.

The subgroup of patients with definite MS who had dou- ble-dose delayed CT scanning was reviewed to determine whether exacerbation had occurred in the eight weeks prior to scanning. Relationships among clinical exacerbation, CT enhancement, and CSF studies were examined. The propor- tion of patients with enhancing lesions was compared in pa- tients who had received single-dose against those who had received double-dose iodinated contrast material. In both

From the Deparrments of ‘Radiology and ?Medicine (Neurology), Duke University Medical Center, Durham NC 27710. Received Sept 8, 1983, and in revised form Apr 24, 1984. Accepted for publication Apr 28, 1984.

Address reprint requests to Dr Drayer, Department of Radiology, Box 3808, Duke University Medical Center, Durham, NC 27710.

33

Table 1. Comparison of Computed Tomographic Abnormalities, Luboratovy Findings, and Patient Profile1 in Patients with Multiple Sclerosis

Findings Definite MS Probable MS Possible MS

Cf ABNORMALITIES

Abnormal CT Low-density areas Enhancing areas Large CSF spaces

29/34” 16/34b 15/34‘ 20134‘

71 18 4118 3/18 5118

17/33 6133 7133 8/33

IABORATORY FINDINGS ~

CSF protein >50 mgll00 ml CSF WBC >5 cells

Abnormal evoked responses CSF IgG 3 1 2 %

6126 10122

5/16 2 5/26

0114 2/10 3/12

15/17

2/28 712 1 312 1

21/31“

PATIENT PROFILE

Age at CT Age at onset Disease duration at CT

40.2 31.3

8.9

42.9 34.2

8.7

37.6 31.9

5.7‘1

Significantly different from other two groups: “p < 0.005 by chi square test; ‘p < 0.05 by chi square test; ‘p < 0.01 by chi square test. dSignificandy different from group with definite M S (p < 0.05; t rest). MS = multiple sclerosis; CT = computed tomography; CSF = cerebrospinal fluid; WBC = white blood cell count.

groups scanning was delayed 15 to 60 minutes after iodine infusion.

Results The 85 patients (55 women, 30 men) were classified as having definite (n = 34), probable (n = 18), and pos- sible (n = 33) MS.

CT abnormalities were present in 62% of patients, with enhancing lesions in 29%, low-density areas in 3196, and enlarged CSF spaces in 39% (Table 1). Of patients with abnormal scans, 49% had enhancing areas, 49% had low-density areas, and 62% had in- creased CSF spaces (Figs 1-3). The enhanced CT scan demonstrated an abnormality not seen on the unen- hanced scan in 47% of patients with positive CT findings. The percentage of patients with CT abnor- malities varied significantly among the clinical groups. In general, patients with definite MS had more CT abnormaltties than those with probable or possible MS. Fully 85% with definite MS had abnormal scans, with enhancing areas in 44%, low-density areas in 4796, and enlarged CSF spaces in 59%. More than 50% of those with possible MS had abnormal scans, with enhancing lesions in 21%, low-density areas in 1896, and enlarged CSF spaces in 24%. The groups with probable and possible MS did not differ statisti- cally.

The mean age at onset among patients in all three groups was 32.2 years; means differed little among the three groups. Mean age at the time of CT scanning was also similar among groups and overall averaged 38.9

years. Mean duration of illness was 7.9 years and was longer in patients with definite MS than in patients with possible MS.

Increased CSF protein levels (more than 50 mg per 100 ml) were found in 12% of patients; the levels did not vary significantly among groups. Increased n m - bers of CSF white blood cells (more than 5 per cubic millimeter) were seen in 3696, and increased IgG (greater than 12%) was noted in only 22%. Again, neither value varied significantly among the clinical groups. Thus, these CSF variables are not sensitive to the staging of clinical disease. The prevalence of abnor- mal evoked responses varied significantly among clini- cal groups, being noted in 96% of those with definite, 88% of those with probable, and 68% of those with possible disease.

The only CT abnormality that correlated with dura- tion of illness was enlargement of the CSF spaces. Pa- tients with enlarged CSF spaces had a longer duration of illness than those with normal CSF space size ( p < 0.005). The only CT abnormality that correlated with age at onset and age at CT scanning was enhancing lesions. Patients with enhancing areas had experienced onset and underwent CT scanning at earlier ages than did patients without enhancing areas (p < 0.05), al- though the duration of illness was the same.

There was no relationship between recent exacerba- tion and abnormal CSF protein, WBC, or IgG findings (Table 2). There was a highly significant correlation, however, between enhancing lesions and recent ex- acerbation, demonstrating that increased blood-brain

34 Annals of Neurology Vol 17 No 1 January 1985

Fig I. Multiple sclerosis, abnormal enhancement. Multiple re- gions of enhancement (arrows) are consistent with disruption of the blood-brain barrier in this patient with definite multiple scle- rosis undergoing exacerbation. Although multiple enhancing ab- normalities are noted in the supratentorial white matter, no le- sion is seen Sy computed tomography in the posterior fossa. The two small, discrete low densities in the midbrain (arrowheads, righthand figure) most likely represent the normal termination of the superior cerebellar peduncles rather than a sclerotic lesion. Al- though most of the enhancing lesions are seen in the white mat- ter, some lesions also involve the gray matter.

barrier permeability is common with clinical exacerba- tion and that CT changes are more useful than CSF findings for demonstrating that the disease is active.

There was no significant difference in the proportion of patients with enhancing lesions demonstrated with standard iodine and with double-dose iodine contrast enhancement techniques (Table 3). In the last 47 CT scans reviewed, 102 enhancing lesions were visualized in 16 patients (Fig 4). Periventricular lesions and le- sions of the central cerebral white matter were the most common, accounting for 36% and 30%, respec- tively. An additional 20% of the enhancing abnor- malities were noted at the junction of the gray and white matter of the centrum semiovale. Other involved areas included the cortical gray matter (5%), brainstem ( 5 % ) , deep gray matter (2%), and superior cerebellar peduncle (2 %). Fifty-five low-density lesions were seen in 15 patients (Fig 5) . Again, periventricular lesions (49%) and lesions of the central cerebral white matter (33%) were the most common. Other involved areas were the gray-white junction ( 5 % ) , brainstem and cerebellum (7%), and internal and external capsules (5%).

Discussion Abnormalities commonly observed with CT scanning in cases of MS include areas of low density, areas of contrast enhancement, and enlargement of the CSF spaces. Anatomical maps of the distribution of enhanc- ing lesions and low-density areas of demyelination and gliosis derived from our CT studies correlate closely

Fig 2. Multiple sclerosis, abnormal enhancement in posterior fissa. lntravenously enhanced computed tomographic scans at the pontine level, showing enhancing abnormalities in the pons (ar- row) and superior cerebellar peduncle (arrowhead). The scan- ning angle has decreased the usual transverse artqact that gener- al(y limits lesion detection in the posterior fossa.

Barrett et al: High-Resolution CT in MS 35

Fig 3. Computed tomographic abnormalities in multiple sclerosis. Lou)-density areas (top, arrows), enlarged subaruchnozd space (bottom), and enhancing lesions (bottom, arrow) were common computed tomographic abnormalities in patient.r with multiple sclerosis.

Table 2. Cowelation Between Recent Clinical Exacerbation and A b n o m l Computed Tomographic Enhancement or Cerebrospinal Fluid Findings a

No Findings Exacerbation Exacerbation

CSF protein >50 mg/ 218 217 100 ml

CSF WBC >5 cells 519 416 CSF IgG a1295 115 116 Enhancing area o n CT 819 2/9b

*All patients had definite multiple sclerosis and were scanned with a double dose of iodinated contrast media (84 gm of iodine). Recent clinical exacerbations were defined as chose occurring within the prior 8 weeks. bSignificantly different from value in exacerbation group ( p < 0.025; chi square test). Abbreviations as in Table 1

Table 3. Demonstration of Enhancing Lesions with Computed Tomographic Scanning Using Standard and Double-Dose lodinated Contrast Medzum

Enhancing Lesions Enhancing Lesions with Standard with Double-

Condition Dose of Iodine Dose Iodine ~~~~~~ ~

Definite MS 5/16 Probable MS 017 Possible MS 117

10118 311 1 6/26

All patients 6/30 19155

No significant differences by chi square test. MS = multiple sclerosis.

with results of postmortem studies IS} with a few im- portant exceptions. CT scans show MS lesions well in the supratentorial cavity, not as well in the posterior fossa, and almost not at all in the spinal cord. CT under- represents peripheral gyral lesions and generally fails to detect optic nerve involvement. Decreased detection of peripheral gyral lesions is due to the small size of gyral lesions (often only a few millimeters) noted in postmortem studies as well as to artifacts created by the inner table of the skull. In postmortem studies about 20% of lesions are found to lie in the posterior fossa, whereas in our study CT detected only 792 of enhanc- ing and low-density lesions in the brainstem and cere- bellum. Anatomical definition in the posterior fossa is degraded by artifacts related to the petrous bones. In addition, “shadow plaques” (incomplete demyelina- tion) are more common in the brainstem than in the cerebrum and may be more difficult to detect on CT scanning than a completely demyelinated plaque. Our findings also confirm the symmetry of plaque distribu- tion stressed in pathological reports 131.

The use of advanced CT scanning equipment has improved the demonstration of intracranial abnor-

36 Annals of Neurology Vol 17 No 1 January 1985

Fig 4. Distribution of I02 enhancing lesions in 16 patients (8 with dejnite, 3 with probable, and 5 with possible multiple scle- rosis). Prriventricular, 36%; centrum semiovale, 309%; gray- white junction, 20%; brainstemlcerebellum, 7%; cortex, 5 9%; deep g w , 2%.

Fig 5 . The distribution of 55 low-density areas in 15 patients 18 with dejnite, 2 with probable, and 5 with possible multiple scle- rosis). Periventricular, 49%; centrum semiovale, 33%; gray- white junction, 5%; brainstemlcerebellum, 7%; internal and ex- ternal capsules, 5 %.

malities. Hershey and colleagues [4] reviewed 66 pa- tients with definite or suspected MS (classification of Rose and colleagues {8]> who were scanned on an early generation CT scanner with standard enhancement (42 gm of iodine). They found some CT abnormality in 29% of patients (21% had low-density areas, 9% had abnormal contrast enhancement, and 20% had en- larged CSF spaces). CT abnormalities were found in 17 of the 35 patients with definite MS and in only 2 of the 31 patients with possible or probable MS [4]. Using high-resolution (GE 8800) CT scanning and a double- dose (84 gm of iodine) delayed enhancement strategy, Vinuela and associates { 1 11 conducted a prospective study of patients with early acute or relapsing MS. Ab- normal enhancement was noted in 82% of patients with definite MS and in 3 1 % with suspected MS.

Our study confirms that CT scanning is a powerful technique for demonstrating pathological changes characteristic of MS. Fully 85% of patients with definite MS had abnormal CT scans. In our study many patients with possible MS had clinical disease that could be localized to one anatomical location and that had occurred concurrently. Although 68% of patients with possible MS had an abnormal evoked response, it often demonstrated abnormalities in the anatomical lo- cale already demonstrated clinically, did not elucidate other anatomical lesions, and therefore did not con-

tribute to a diagnosis of MS. Over 50% of patients with possible MS had abnormal CT scans. In many cases the presence of an abnormal enhancing area or low-density lesion on the CT scan demonstrated ab- normality in an area not suspected on the basis of clinical or laboratory evaluation and therefore contrib- uted to the demonstration of lesions separate in space and time and to a diagnosis of MS. In the patient for whom the diagnosis is suspected but not confirmed by clinical and laboratory evaluation, therefore, CT may provide useful supporting pathological anatomical in- formation.

In contrast to the studies of Vinuela and colleagues { 11) and Sears and co-workers {9], our investigations revealed no difference in the incidence of abnormal enhancement between the standard and the double- dose delayed enhancement strategies. Because our standard enhancement procedure entailed a 15- to 30- minute delay prior to scanning, the incidence of abnor- mal enhancement may depend as much on a delay before scanning as on the contrast dose. Contrast- enhanced scanning demonstrated abnormalities not seen on the unenhanced study in 47% of patients with CT abnormalities, affirming the value of routinely us- ing intravenous contrast enhancement when scanning patients with suspected MS.

Broman [I} described barrier disruption in MS

Barrett et al: High-Resolution CT in MS 37

using a trypan blue-albumin complex, a technique similar to but likely less sensitive than that involving the use of lower-molecular-weight iodinated contrast medium. It is well established that abnormal contrast enhancement correlates closely with disruption of the blood-brain barrier 12). Tourtellotte [ 10) and Walsh [ 121 have extensively examined the role of CSF mark- ers (such as protein) in evaluating the course of MS and the integrity of the blood-brain barrier. Because protein in the bloodstream will not cross an intact blood-brain barrier into the brain but will pass a dis- rupted barrier, Tourtellotte has used a CSF-blood ratio to analyze barrier function. Of the patients he studied, 77% had normal CSF protein and the others had only small elevations; he concluded that barrier abnormality was rarely extensive in MS [lo}. In addi- tion, he found no significant change in CSF protein level with exacerbation and thus concluded that there was no correlation between barrier integrity and the development of new symptoms [lo). Our work sup- ports his data in that the CSF protein level was abnor- mal in only 125% of our patients and had no relation- ship to recent exacerbation ( p > 0.900) but does not confirm his conclusions concerning the blood-brain barrier. Alterations in the blood-brain barrier, as dem- onstrated by abnormal contrast enhancement, were a consistent feature in cases of recent clinical exacerba- tion. There was no correlation between abnormal en- hancement (barrier disruption) and elevated CSF pro- tein level in our series, confirming that measurement of CSF protein is not a sensitive indicator of an al- teration in blood-brain barrier permeability. Our series also affirms that there is no straightforward relationship between CSF IgG and activity of disease, occurrence of clinical symptoms, duration of disease, or blood-brain barrier disruption.

Our analysis of patients with definite MS who underwent double-dose delayed enhanced CT scan- ning showed a highly significant correlation between recent exacerbation and abnormal contrast enhance- ment. The question arises as to whether barrier disrup- tion is merely epiphenomenal or, rather, represents an integral, mechanistically important feature of the dis- ease. A myelinolytic factor circulating in the blood- stream may enter the brain only during temporary epi- sodes of blood-brain barrier disruption resulting from abnormal permeability of capillary endothelial cells or medullary venules. This type of intermittent opening of the barrier, whether immunologically or chemically

mediated, might explain the clinical course of remis- sions and exacerbations as well as the vasogenic edema sometimes associated with active lesions. Stressful epi- sodes such as sickness, fatigue, emotional upset, infec- tion, heat, or trauma can stimulate the central ad- renergic system, which appears to regulate the cerebral microcirculation. Preskorn and associates [ 7) have shown that amitriptyline, an adrenergic agonist, in- creases the permeability of the blood-brain barrier. A similar mechanism of adrenergic stimulation related to stress resulting in intermittent barrier disruption may occur in MS. This adrenergic barrier hypothesis ac- cords with the relative symmetry of brain lesions seen in our study as well as in previous neuropathological studies. Barrier disruption demonstrated by CT scan- ning may thus prove a useful index of active demyeli- nation during the evaluation of patients receiving ex- perimental therapy.

References 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

Broman T: Blood-brain barrier damage in MS: supravital test observation. Acta Neurol Scand 40:21-24, 1964 Drayer BP: Functional applications of CT of the central nervous system. AJNR 2:495-510, 1981 Fog T: The topography of plaques in multiple sclerosis. Acta Neurol Scand 15 (suppl41): 1-162, 1765 Hershey LA, Gado MH, Trotter J L Computerized tomography in the diagnostic evaluation of multiple sclerosis. Ann Neurol

Lumsden CE: The neuropathology of multiple sclerosis. In Vin- ken PJ, Bruyn GW (eds): Handbook of Clinical Neurology, Vol 9. Amsterdam, North-Holland, 1970, pp 217-309 Poser CM: A numerical scoring sysrem for the classification of multiple sclerosis. Acta Neurol Scand 6O:lOO-111, 1979 Preskorn SH, Harrman BK, Irwin GH, Hughes CW: Role of the central adrenergic system in mediating amitriptyline-induced alteration in the mammalian blood-brain barrier in vivo. J Phar- macol Exp Ther 223:388-395, 1982 Rose AS, Ellison GW, Lawrence WM, Tourtellotte WW: Crite- ria for the clinical diagnosis of multiple sclerosis. Neurology (Minneap) 2620-22, 1976 Sears ES, McCammon A, Bigelow R, er al: Maximizing the harvest of contrast enhancing lesions in multiple sclerosis. Neu- rology (NY) 323154320, 1982 Tourtellotte WW: Cerebrospinal fluid in multiple sclerosis. In Vinken PJ, Bruyn GW (eds): Handbook sf Clinical Neurology, Vol 9. Amsterdam, North-Holland, 1970, pp 324-382 Vinuela FV, Fox AJ, Debrum GM, er at: New perspectives in computed tomography of multiple sclerosis. AJNR 3:227-281, 1782 Walsh MJ, Tourtellotre WW: The cerebrospinal fluid in multi- ple sclerosis. In Hallpike JF, Adams CW, Tourtellorte WW (eds): Multiple Sclerosis. Cambridge, Cambridge University Press, 1983, pp 275-358

5:32-39, 1977

38 Annals of Neurology Vol 17 No 1 January 1985