Magnetization transfer ratio (MTR) is a sensitive parameter to quantify the integrity of myelinated white matter in patients with multiple sclerosis. Lesional MTR decreases in the acute phase due to demyelination, and subsequently shows recovery depending on the degree of remyelination in the absence of axonal loss. Recovery of average lesion MTR therefore might prove a viable outcome measure to assess the effect of remyelinating agents. Our objective was to determine the required sample size for phase II multicentre clinical trials using the recovery of average lesion MTR as primary outcome measure. With 7-monthly MRI scans, the MTR evolution of 349 new enhancing lesions before and after enhancement was assessed in 32 MS patients from 5 centres. Multilevel models were fitted to the data yielding estimates for the variance components, which were applied in power calculations. Sample sizes were determined for placebo-controlled, multicentre trials using lesional MTR recovery post-enhancement as primary outcome measure. Average lesion MTR decreased slightly in the build-up to enhancement, decreased dramatically during enhancement and showed recovery in the period after cessation. The power calculations showed that for a power of 80%, approximately 136 patients per trial (mean number of 6 lesions per patient) are required to detect a 30% increase in lesional MTR post-enhancement compared with placebo, whereas 48 subjects are required to detect a 50% increase in lesional MTR compared with placebo. Recovery of lesion MTR is a feasible outcome measure for future multicentre clinical trials measuring the effect of remyelinating agents.