Mary K. Edwards 1

Martin R. Farlow2

James C. Stevens2

This article appears in the July/August 1986 issue of AJNR and the September 1986 issue of AJR.

Received August 1, 1985; accepted after revi­sion October 14, 1985.

1 Department of Diagnostic Radiology, Indiana University Medical Center, 926 West Michigan Street, Indianapolis, IN 46223. Address reprint re­quests to M. K. Edwards.

2 Department of Neurology, Indiana University Medical Center, Indianapolis, IN 46223.

AJNR 7:595-598, July/August 1986 0195-6108/86/0704-0595 © American Society of Neuroradiology

Multiple Sclerosis: MRI and Clinical Correlation

595

Fifty-three consecutive patients with suspected multiple sclerosis were studied to determine if the extent of disease apparent on MRI correlated with the clinical severity of the disease. MRI images were evaluated and compared with an assessment of the patient's disability using three neurologic rating scales. The severity of the disease seen on MRI showed a strong statistically significant correlation (p = 0.0001) with two of the three methods of clinical evaluation and a significant correlation (p < 0.01) with the third rating scale. The severity of disease shown on MRI correlated only weakly (p = 0.05) with the length of time the patients had been symptomatic. Normal controls did not show any abnormality characteristic of multiple sclerosis on MRI or on neurologic exam.

MRI has been recognized as a major advance in the diagnosis and evaluation of patients with multiple sclerosis (MS) [1-4] . Many more lesions are detected by MRI than by CT, and MRI has proved to be helpful in establishing the diagnosis of MS in cases where the clinical diagnosis is not definite [5-7] . Although MRI has made MS easier to diagnose at an early stage [7] , the correlation between MRI results and the clinical course and disability has been questioned [8, 9].

Subjects and Methods

Fifty-three patients with suspected MS were evaluated in the Department of Neurology at Indiana University using three different clinical scales. The patients were evaluated according to the McAlpine criteria as having either possible , probable, or definite MS [10]. The McAlpine criteria were converted to a numerical scale ranging from 1 (possible) to 3 (definite). Patients who did not qualify as having MS according to the McAlpine criteria were excluded from the study. The patients were also graded using the expanded disability status scale proposed by Kurtzke [11], which grades the patient 's disability from 0 (normal) to 10 (death due to MS) at one-half step intervals. The Kurtzke scale is designed to grade the patient 's functional disability. Cerebellar dysfunction , cranial nerve deficits, and difficulty ambulating are among the functions graded by the Kurtzke scale. Finally , the patients were assigned a score from o (death) to 100 (normal) according to the neurologic rating scale of Sipe et al. [12J . The neurologic rating scale of Sipe is an additive scale based on the neurologic examination . Duration of the disease at the time of imaging varied from 1 month to 31 years. Twelve normal volunteers were also examined.

MRI was performed on a resistive scanner operating at a field strength of 0.15 Tusing two spin-echo (SE) pulse sequences. A pulse repetition rate (TR) of 2000 msec was used with an echo delay time (TE) of 120 msec, producing a T2-weiQhted scan. A second sequence was performed with a TR of 500 msec and a TE of 30 msec, producing a T1-weighted scan . Multislice technique was used , allowing simultaneous acquisition of 15 images at a slice thickness of 1 cm, for a TR of 2000 msec. At a TR of 500 msec, 11 simultaneous images at a 1.5-cm slice thickness were acquired. All images were taken in the axial projection, with additional images in the sagittal projection taken when lesions were identified within the brainstem.

All images were evaluated by three independent observers, who noted the number and

596 EDWARDS ET AL. AJNR :7, July/August 1986

1 2

A B

Fig. 1.-MRI grade 1 MS. Normal axial SE 120/ 2000 image.

Fig. 2. -MRI grade 2 MS. Single lesion with increased signal intensity (arrow) on axial SE 120/ 2000 image.

c Fig. 3.-MRI grade 3 MS. A-C, 5 periventricular lesions of increased signal (arrows) on axial SE 120/2000 image.

size of lesions on the SE 120/2000 images as well as the degree of atrophic change on the SE 30/500 images. Images were graded using a scale of 1-5. Normal images were graded 1 (Fig. 1). Images showing only one to three lesions were assigned a score of 2 (Fig. 2). Images with four to eight lesions were graded 3 (Fig . 3). Images with nine to 13 lesions and only one or two confluent lesions were graded 4 (Fig. 4). Images with 14 or more lesions showing extensive confluence of lesions in the periventricular region , with atrophic changes , were given the grade of 5 (Fig . 5).

The MRI grading scale was proposed by the senior author of this paper and explained to the other two authors using Figures 1-5. The scans were then evaluated independently by the three authors. When disagreement among the three observers occurred, the grade given by two of the three was used for statistical analysis.

The severity of the involvement seen on MRI was then correlated statistically with the clinical severity of the disease, as well as with the length of time the patient had been symptomatic. The statistical analysis included calculation of the p and r values for each correlation.

Results

There was an excellent level of agreement among the three observers in the determination of the severity of disease with

MRI ; they disagreed on only eight of the 53 images, and in those eight cases the observation of one observer differed from the other two by only one grade. The results of the study are summarized in Table 1. Seventeen patients had normal MRI exams. The number of patients in the remaining four MRI grades, in order of increasing severity, were eight, eight, nine, and 11 .

A strong statistically significant correlation was found be­tween the severity of disease shown on MRI and the extent of disease diagnosed by all three methods of clinical evalua­tion. The Kurtzke and Sipe criteria showed very similar cor­relations with MRI. The correlation of disease found by MRI to the results of the Kurtzke scale had a p value of 0.0001 and an r value of 0.558 , while correlation to the Sipe criteria had a p value of 0.0001 and an r value of 0.595 . Correlation was not as strong with MRI and the McAlpine criteria , with a p value of < 0.01 and an r value of 0.440. Only weak correla­tion existed between MRI and the duration of symptoms (p = 0.05 , r = 0.271). No abnormalities were detected on MRI or on neurologic exam in the 12 normal controls.

AJNR :7, July/August 1986 MRI OF MULTIPLE SCLEROSIS 597

Fig. 4.-MRI grade 4 MS. A, Lesion in midbrain (arrow) on axial SE 120/2000 image in patient with nine other lesions. B, single confluent lesion in left periventricular region (arrow) on axial SE 120/2000 image.

Fig. 5.-MRI grade 5 MS. A and B, multiple confluent high signal intensity lesions in the periven­tricular region on axial SE 120/2000 images. C and D, Advanced atrophy with marked widening of ven­tricles and sulci on axial SE 30/500 images.

Discussion

A

A

c

It is recognized that MRI is a sensitive technique for de­tecting the characteristic plaques of MS. Although it seems

B

B

o

reasonable to assume that the manifestation of advanced disease on MRI should correlate with severe clinical disability, this relationship has not been previously established . In fact , reports have appeared indicating that MRI correlates poorly

598 EDWARDS ET AL. AJNR:7, July/August 1986

TABLE 1: Comparison of MRI Findings to Neurologic Examination and Duration of Symptoms

Average Neurologic Score

No. of Patients for Patients in Each MRI Average MRI Grade

(n = 53) Grade Duration of Symptoms (years)

McAlpine Sipe Kurtzke

1 17 1.5 85 2.5 5.4 2 8 2.1 79 2.9 7.5 3 8 2.0 81 3.4 9.6 4 9 2.2 72 4.5 9.7 5 11 2.6 59 5.6 12.4

with lesion localization [9] , and may have no correlation with severity of disease [8]. In our study, a highly significant relationship was demonstrated between MRI and the severity of clinical disease. This may be due to the larger sample of patients in our study compared with previous reports. The use of the Kurtzke and Sipe criteria may have also contributed to the improved statistical correlation. The Kurtzke and Sipe criteria probably correlated more closely to the MRI than did the McAlpine criteria because the McAlpine criteria, using only three grades, is designed primarily to diagnose MS and is a relatively insensitive measure of neurologic dysfunction.

Although statistically significant Ip = 0.0001), the correla­tion was only moderate (r = 0.4-0.6) between the MRI results and the neurologic status of the patient. The lack of strong correlation can be explained by several limitations. Our study did not include the spinal cord , which is frequently the site of symptomatic lesions [9]. Another frequent site of sympto­matic abnormality in MS is the optic nerve. MRI has been disappointing because of its inability to detect demyelinating lesions of the optic nerve [9] , and not one was seen in this series. Another limitation of MRI is its inability to distinguish acute demyelinating plaques from chronic lesions. Johnson et al. demonstrated that acute plaques appear much larger than subacute or chronic lesions of MS, and the same lesion may appear quite different as the disease progresses [13] . In addition, MRI may show many lesions in neurologically silent regions as well as many small lesions causing little neurologic deficit. The sensitivity of lesion detection is also limited by the quality of the images obtained on our resistive magnet sys­tem. These limitations may account for the lack of strong correlation between MRI and the clinical examination.

Because of these limitations, MRI should be used cau­tiously as a prognostic indicator of clinical disease in the individual patient. Although MS has a waxing and waning course, there is a relentless tendency toward progressive involvement of cerebral tissue as the disease advances [14 , 15]. In spite of the limitations mentioned above, the significant correlation found in this study confirms the ability of MRI to monitor the progression of the disease.

The finding of a strong statistically significant correlation between the severity of disease shown on MRI and accepted neurologic rating scales gives added validity of the use of MRI in the long-term evaluation of patients with MS. More

advanced disease is usually found at autopsy than suspected by the clinical examination of the MS patient [13-14]. There­fore, to effectively monitor the response to therapy, an objec­tive method of measuring the progress of individual lesions is necessary [16]. MRI may be an acceptable tool for monitoring the progress of the disease and the response to drug therapy. The ability to observe asymptomatic lesions and to objectively measure the extent of disease involvement may prove that MRI is of significant value in the evaluation of MS patients.

ACKNOWLEDGMENT

We wish to acknowledge the assistance of Pao-Io Yu in preparing the statistical analysis.

REFERENCES

1. Young IR , Hall AS, Pallis CA, Legg NJ, Bydder GM , Steiner RE. Nuclear magnetic resonance imaging of the brain in multiple sclerosis. Lancet 1981;2: 1 063-1 066

2. Young IR , Randell CP, Kaplan PW, James A, Bydder GM, Steiner RE. NMR imaging in white matter disease of the brain using spin-echo sequences . J Comput Assist Tomogr 1983;7:290-294

3. Lukes SA, Crooks LE , Aminoff MJ, et al. Nuclear magnetic resonance imaging in multiple sclerosis. Ann Neurol 1983;13: 592-601

4. Buonanno FS, Kistler JP, Lehrich JR, Noseworthy JH, New PFJ, Brady T J. 'H nuclear magnetic resonance imaging in multiple sclerosis. Neurol Clin 1983;1 :757-764

5. Steiner RE. The Hammersmith clinical experience with nuclear magnetic resonance. Clin Radio/1983;34:13-23

6. Gebarski SS, Gabrielsen TO, Gilman S, Knake JE, Latack JT, Aisen AM. The initial diagnosis of multiple sclerosis: clinical impact of magnetic resonance imaging. Ann Neurol 1985;17:469-474

7. Herndon RM , Rudick RA. Multiple sclerosis: the spectrum of severity. Arch Neuro/1983;531-532

8. Crisp DT, Kleiner JE, DeFillip GJ, Greenstein JI, Liu TH , Sommers D. Clinical correlations with magnetic resonance imaging in mul­tiple sclerosis (abstr). Neurology 1985;35: 137

9. Jackson JA, Leake DR , Schneiders NJ, et al. Magnetic resonance imaging in multiple sclerosis: results in 32 cases. AJNR 1985;6 :171-176

10. McAlpine D, Lumsden CE, Acheson ED. Multiple sclerosis: a reappraisal. London: Churchill Livingstone, 1972:202

11. Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS) . Neurology 1983;33 : 1444-1452

12. Sipe JC, Knobler RL, Braheny SL, Rice GPA, Panitch HS, Old­stone MBA. A neurologic rating scale (NRS) for use in multiple sclerosis. Neurology 1984;34: 1368-1372

13. Johnson MA, Li DKB, Bryant OJ , Payne JA. Magnetic resonance imaging: serial observations in multiple sclerosis . AJNR 1984;5 : 495-499

14. McDonald WI , Halliday AM. Diagnosis and classification of mul­tiple sclerosis Br Med Bull 1977;33 :4-8

15. McFarlin DE , McFarland HF. Multiple sclerosis. N Engl J Med 1982;307 : 1246-1251

16. Lebow S, Anderson DC, Mastri A, Larson D. Acute multiple sclerosis with contrast-enhancing plaques. Arch Neurol 1978;35 :435-439