Are MRI measures valid outcomes for clinical trials?

One of the problems in developing therapies for MS is that it takes a lot of time, money, and subjects to prove efficacy, because the usual clinical endpoints (relapse and disability) require a long observation period before any significant effect can be seen. That's why so many current clinical trials also include periodic MRIs which evaluate things like enhancing lesions and lesion load. It could be very helpful for conducting clinical trials in the future if MRI outcomes could be substituted for clinical outcomes since changes in MRI can be seen more quickly.

The question is, how well do MRI outcomes correlate with clinical outcomes? Two recent studies asked that question and came up with opposite answers. One study examined data from 23 clinical studies (6,591 subjects) and concluded that enhancing lesions were strongly correlated with relapse rate. However, the other study, which analyzed placebo group data from the Sylvia Lawry Centre database, concluded that enhancing lesions and lesion load were not good surrogates for relapses and disability progression, respectively.

I'm not sure which conclusion is right, but wouldn't it be great if these two groups could get together to compare methodologies and try to reconcile their conflicting findings?