A systematic review of MRI to diagnose MS raises several issues, says Dr Penny Whiting.
The diagnosis of MS is mainly clinical, based on identifying evidence of dissemination in both time and space. This translates to a patient experiencing two attacks of neurological dysfunction, such as optic neuritis or numbness and tingling of the leg, occurring at different points in time and affecting different parts of the CNS.
The time between a patient first experiencing symptoms and going on to have the second attack can range from months to many years. Not all patients who experience symptoms will go on to develop MS.
The role of MRI in diagnosing MS has become of increasing importance.
The McDonald criteria, published in 2001 and revised in 2005, may allow an earlier diagnosis of MS to be made by combining MRI with clinical data to provide evidence of dissemination in time and space, even if a patient has experienced a single attack of neurological dysfunction (see box).
NICE has adopted these criteria.
At the University of Bristol, we conducted a systematic review to estimate the accuracy of MRI for the early diagnosis of MS in patients presenting with suspected MS. We identified 29 studies that reported relevant data: these differed in design, patients, criteria used to define a positive MRI scan and MRI technology.
The studies showed considerable variation in estimates of the accuracy of MRI. We first investigated reasons for this variation in accuracy by comparing the results of studies that prospectively enrolled patients with MS with those that did not.
The majority of studies examined case-control studies where the researchers selected a group of patients with known MS and a control group of healthy volunteers or patients with other known conditions.
We found that studies of this type grossly overestimated the diagnostic accuracy of MRI.
We therefore restricted further analyses to studies that prospectively enrolled patients with possible MS - the patients in whom MRI has the potential to prove an earlier diagnosis.
These showed that MRI is diagnostic for MS, but that its accuracy is much lower than the estimates of accuracy obtained from case-control studies.
A problem we faced in the second stage of our analysis was that very few studies followed up patients for sufficiently long periods to ascertain that patients who were expected to eventually develop MS had done so.
The duration of clinical follow-up in these studies ranged from seven months to 14 years.
However, studies have shown that it can take more than 10 years from experiencing an initial attack to patients experiencing a second.
Our analysis showed that studies with shorter durations of follow-up tended to overestimate sensitivity and underestimate specificity. There were only two studies that included sufficiently long durations of follow-up (greater than 10 years) for us to be reasonably confident that the majority of patients included in these studies who would eventually go on to develop clinically definite MS had already done so.
Our final analysis was therefore restricted to these two studies. Both of the long-term cohort studies included patients presenting with clinically isolated syndromes. One was restricted to patient presenting with optic neuritis and followed-up these patients for at least 10 years, while the other included patients with a range of presenting symptoms and followed them for 14 years.
Each evaluated threshold is based on the number of non-clinical T2 lesions present on brain MRI. Based on the results of these studies (the likelihood ratios for a positive or negative MRI result), assuming that 60 per cent of patients presenting with possible MS will eventually go on to develop MS, even in the presence of many MRI lesions (>10) the probability of the patient going on to develop MS was just over 80 per cent.
Similarly, in patients with a negative scan for MS (no lesions) the probability of developing MS decreased to 13 per cent in one study and 43 per cent in the other.
This suggests that around 20 per cent of people presenting with symptoms of MS and found to have numerous MS-like lesions on their MRI scan will not develop MS over the next 10-14 years.
Similarly, of those found not to have any lesions on their MRI scan, around 13-43 per cent will go on to develop MS over the same time period.
The clinical implications of our study should be considered. The potential benefits of providing an earlier diagnosis of MS include earlier access to treatment.
However, this can have a negative impact on a patient's quality of life, with increased insurance premiums and discrimination at work. The consequences of false-positive and false-negative diagnoses of MS should also be considered.
False-positive diagnoses could imply the prescription of disease-modifying treatments with unpleasant side-effects.
Neurologists should discuss with the patient the potential diagnosis, treatment and ultimate effect of potential errors.
There are many reasons for requesting an MRI scan in a patient presenting with possible MS, beyond its utility in diagnosis. These can include ruling out differential diagnoses such as brain tumours.
- Dr Whiting is research fellow, MRC Health Services Research Collaboration
Department of Social Medicine, University of Bristol
MCDONALD DIAGNOSTIC CRITERIA FOR MS
Clinical presentation Additional data needed for
Two or more attacks; objective - None
clinical evidence of two or more
Two or more attacks; objective - Dissemination in space,
clinical evidence of one lesion demonstrated by:
OR two or more MRI detected
lesions consistent with MS plus
OR Await further clinical attack
implicating a different site
One attack; objective clinical - Dissemination in time,
evidence of two or more lesions demonstrated by:
OR Second clinical attack
One attack; objective clinical - Dissemination in space,
evidence of one lesion demonstrated by:
(monosymptomatic presentation; - MRI
clinically isolated syndrome) OR two or more MRI-detected
lesions consistent with MS plus
- Dissemination in time,
OR Second clinical attack
Insidious neurological - One year of disease
progression suggestive of MS progression (retrospectively
or prospectively determined)
Two out of three of the
a. Positive brain MRI (9 T2
lesions or four or more T2
lesions with positive visual
b. Positive spinal cord MRI (two
or more focal T2 lesions);
c. Positive CSF
Source: Polman et al. Diagnostic Criteria for Multiple Sclerosis: 2005 Revisions
to the 'McDonald' Criteria. Ann Neuro 2005; 58: 840-6.
- Whiting P, Harbord R, Main C et al. Accuracy of magnetic resonance imaging for the diagnosis of multiple sclerosis: systematic review. BMJ 2006; 332: 875-84.
- National Collaborating Centre for Chronic Conditions. Multiple sclerosis: National clinical guideline for diagnosis and management in primary and secondary care. London: Royal College of Physicians, 2004.
- Beck R W, Trobe J D, Moke P S, Gal R L, Xing D, Bhatti M T, et al. High- and low-risk profiles for the development of multiple sclerosis within 10 years after optic neuritis: experience of the optic neuritis treatment trial. Arch Ophthalmol 2003;121: 944-9.
- Brex P A, Ciccarelli O, O'Riordan J I, Sailer M, Thompson A J, Miller D H. A longitudinal study of abnormalities on MRI and disability from multiple sclerosis. N Engl J Med 2002; 346: 158-64.