Spinal cord MRI should always be performed in clinically isolated syndrome patients: Commentary

One of the present difficulties facing neurologists and neuroradiologists alike is accommodating the increasing demands for MRI scanning in patients with multiple sclerosis (MS). We have a number of JCV-positive patients receiving natalizumab and who thus need frequent MRI scans (at present 4-monthly in our hospital). We also have the new recommendations from the expert committee, led by Fred Lublin, on the assessment of the course of MS, recommending annual MRI scans for all patients with relapsing MS. ¹ Certainly this will place significant demands upon our MRI service in a busy University Hospital and I can foresee that we will have to seek special funding in order to employ more radiographers for weekend work. The Lublin committee’s recommendations are welcome since I think this single paper will do much to improve the level of care for people with relapsing MS. ¹ Neurologists now have to make a decision, at least annually, as to whether a patient has active or inactive disease and thus decide whether to continue with the same treatment or escalate therapy.

However, these recommendations come with a significant cost to health services. Thus neurologists and neuroradiologists have to make decisions as to whether to scan, what to scan, the frequency of scanning, and scanning sequences in patients in order to perform cost-effective procedures.

The operative word in the title of this debate is ‘always’; both the protagonist and the antagonists accept that there are specific indications for spinal cord MRI in patients with a clinically isolated syndrome. These are: (a) when the anatomical site of the presenting symptom is in the spinal cord; and (possibly) (b) when the MRI of the brain is inconclusive, most commonly when white matter lesions are not completely typical of demyelination. However, outside these criteria there is no consensus. Frederik Barkhof argues that, because spinal cord imaging may be of value in satisfying the McDonald criteria and because it may have some prognostic value in relation to disability accrual, it is better to do a routine spinal MRI along with the cerebral sequences at the time of the first imaging procedure. He and his colleagues have produced a very nice paper in support of this point. ² However, in their analysis they indicate that in order to have one positive spinal cord MRI it is necessary to scan approximately seven patients. Similarly, in another small series, the rate of abnormal spinal cord MRI in patients with the clinically isolated syndrome was 14%. ³ Clearly, in many clinical situations, if 1 in 7 of routine tests were abnormal, which might change treatment decisions, this would argue for routine testing. However, MRI scanning is expensive and the addition of a spinal cord sequence will add to the total scanning time. Given the above-mentioned pressures upon MRI scanning departments I really do not feel that this yield (14% abnormal) warrants the increased scanning time (and, thus, longer waiting lists for others). Of course, if there were unlimited resources (time and money), then their recommendation is entirely sensible and reasonable. However, in most health systems, there are pressures to focus on tests/treatments with higher potential health benefits. Thus, while recognising the logic of Frederik Barkof’s position, in our situation, and I am sure this is the same with others, we do not perform, and will not be performing, routine spinal cord MRI scans in our patients with a clinically isolated syndrome, for all the reasons mentioned above.

There is certainly time for a debate on the utilisation of MRI scanning generally in relation to the assessment of disease activity in our patients. In the limited number of studies which have analysed the predictive nature of the annual MRI scan, it has been demonstrated that an increase in the T2 lesion load has relatively low specificity for predicting subsequent disability over the next few years.^4–6 It is clear that MRI scanning is not good enough to give us a full picture of disease activity, because it does not adequately convey the extent of grey matter disease. We need to move on with biomarkers to assess adequate disease suppression, of which the most relevant appear to be cerebrospinal fluid neurofilaments. ⁷

In conclusion, I would take the side of Alex Rovira and Mar Tintore; there are specific indications for spinal cord imaging in clinically isolated syndrome but the routine use of spinal cord imaging in these patients is not warranted.