ACTRIMS 2026 | The limitations of brain volume measured by MRI as an outcome measure in MS trials

Well, I think the obvious answer about why it’s important is because it just seems so obvious. It has face validity. You know, if you don’t have a brain, you can’t function very well. And the less brain you have, it would seem the less well you can function. And in a lot of pathology reports, you can see that the brain, as it does get destroyed, leads to disability. So it’s a very natural, full of face validity to use it. And it’s probably a correct measure of outcome, but we can’t do pathology on everybody. So what we’re left with is an MRI. And an MRI measuring the brain and brain volume seems to be the right way to go. It’s a beautiful technique. And what I discussed at the meeting, though, what are some problems with it and why it makes it not a good outcome measure. I focused on two sets of problems, one being the reliability of change in an MRI of brain volume. And the second, in the first drug treatments for MS, they noticed that people who were treated with the active drug relative to a placebo at the first MRI following the baseline after they were treated, they had more brain atrophy than their counterpart. And this led to a behavior of re-baselining and adjusting because what was happening or what is thought to be happening, and I think it’s correct, is that the inflammation that is in the brain by these potent anti-inflammatories is greatly reduced fairly rapidly. And that takes the water out of the brain that’s there to help bathe the inflammation. And it looks like the brain is shrinking. But the brain itself really isn’t shrinking. It’s the absence of water. There may be some shrinkage in the brain. But the bottom line was it was called pseudo-atrophy. It was false atrophy. And then this is then used subsequently as the baseline and then measures of change were measured from that point. And in my view, that process is flawed. It has a greater chance of increasing false positive results. One of the things I showed is what we think is happening is we get a curve that reduces the inflammation and then we see the treatment benefits with two lines for future brain atrophy progressing from there showing the treatment effect. The trouble is you’re really measuring the rate of inflammation reduction. And even with placebos, inflammation recedes naturally. So you’re really looking at two different rates of inflammation reduction, and that can lead to results that make the treatment group look like it has less atrophy than the control group, but it’s false. They are just declining at the same overall rate, but in the first year or first six months, you’ve given a kick to the anti-inflammatory agent or the treatment effect. And we don’t know what would happen if you had a remyelinating agent, et cetera. We don’t have any experience. Maybe brain volume will come back and be useful under those circumstances. But with the data we have today, it’s just not suitable to use this rebaselining. And the rebaselining may actually lead to some false positive results that make a treatment look like it’s better when in fact it’s not.

This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.