Over the past 20 years MRI techniques have given wonderful insights into the pathogenesis of multiple sclerosis. We have learned how lesions evolve and resolve, that measures of neuronal and axonal loss show that multiple sclerosis is not just a disease of myelin, and that over time the brain atrophies. The suspicion that the normal appearing white cell matter between the lesions of multiple sclerosis is abnormal has been confirmed. The contribution of MRI to clinical trials of novel therapeutic agents has been decisive and much more reproducible than effects on clinical measures of disability.

The Milan Neuroimaging Research Group has been remarkably productive over the past few years, studying the application of new MRI techniques to multiple sclerosis. The paper by Rovaris et al in this issue (pp 723–727)1 adds to their impressive output. Analysis of brain and cervical cord magnetisation transfer ratio (MTR) histograms in individual patients with relapsing and remitting, secondary progressive, and primary progressive multiple sclerosis has been performed in individual patients. Magnetisation transfer ratio analysis is a technique which has been developed recently.2 In a tissue the amount of magnetisation transfer depends on the capacity of macromolecular protons to exchange magnetisation with surrounding water protons and may be quantified by calculating MTR. The main influences on the value in the white matter of the brain are thought to be the myelin and water content. Loss of myelin or increased tissue water results in a reduction of MTR. A lucid explanation of the complex issues may be found in the article by van Buchem et al.3

In previous MRI studies of the cervical cord in multiple sclerosis spinal cord atrophy has been shown to correlate with physical disability to a greater extent than brain pathology. Stevensonet al 4 demonstrated increasing atrophy over a period of follow up.

The paper by Rovaris et al 1indicates that both brain and spinal cord MTR measures were significantly associated with a higher probability for patients to have secondary progressive multiple sclerosis and to have difficulty in walking. Unfortunately the method does not allow an accurate estimate of the relative contributions of brain and spinal cord pathology to disability to be made. Magnetisation transfer analysis is of interest but is likely to be confined to research assessments and is unlikely to be widely applied in clinical trials. With the continuing development of MRI hardware and software, this is a rapidly evolving field. Standards for collaborative studies are under continuous appraisal and review.