J. THEODORE PHILLIPS, MD, PhD: Hello, and welcome to this discussion on brain atrophy and multiple sclerosis. I’m your host, Dr. Ted Phillips, from the Multiple Sclerosis Center at Texas Neurology in Dallas. I’m pleased to be joined today by Dr. Rob Bakshi, Director of Clinical MRI MS Imaging at the Brigham and Women’s Hospital in Boston and Dr. Robert Zivadinov, Associate Professor of Neurology at the State University of New York at the Buffalo School of Medicine and Biomedical Sciences and Director of the Buffalo Neuroimaging Analysis Center. Welcome to both of you.

DR. ROHIT BAKSHI: Thank you.

DR. ROBERT ZIVADINOV: Thank you.

J. THEODORE PHILLIPS, MD, PhD: MRI has become a key tool in the longitudinal monitoring of patients with MS, including assessment of brain and spinal cord atrophy. Brain atrophy measurements are emerging as sensitive surrogate markers in therapeutic trials in MS. Some recent studies have shown that brain atrophy progresses significantly on an annual basis and bears a close relationship to clinical impairment. Today we hope to help the practicing neurologist better understand the methods for quantifying regional and whole brain atrophy in MS.
Robert, I’d like to start off with you, since this whole program’s going to be about atrophy in MS, brain atrophy. First off, although fairly apparent, at least superficially, what atrophy means, what’s our current level of understanding with regards to at a pathophysiologic basis? What comprises brain atrophy? When we say “brain atrophy,” what does that mean, really?

DR. ROBERT ZIVADINOV: First of all, I would like to say that atrophy became a primary endpoint and a secondary endpoint in recent studies and of great interest in the last couple of years. Although many, many years ago, actually, a century ago, there were the first histopathological findings of brain and spinal cord atrophy in patients with multiple sclerosis.
Really, in the last 20 to 25 years, with early computerized tomography studies, the atrophy was clinically recognized in patients with multiple sclerosis. But the real evolution in measurement of brain atrophy begins in the last 10 to 12 years with the introduction of magnetic resonance imaging as an important instrument in monitoring patients with multiple sclerosis. And we are learning, from year to year, even more about brain atrophy.
The pathophysiology of brain atrophy surely has been better understood with some recent pathological studies and correlation studies between MRI and pathology and it is current understanding that the brain atrophy represents a conglomeration of different pathological substrates that we can find in patients with multiple sclerosis, firstly including demyelination and finally, as a final event, axonal loss.
There are many, many different processes which are contributing to brain atrophy, but our understanding is, very shortly, that the process seems to be different in different pathological stages of the disease. In the early phases, probably, the inflammation is one of the most important pathophysiological events which is contributing to the development of central nervous system atrophy, where in the later stages, probably some other mechanisms more related to neurodegeneration, and I’m referring here to Wallerian degeneration, to gliosis, are contributing to central nervous system atrophy.

J. THEODORE PHILLIPS, MD, PhD: Okay. We talk about whole brain atrophy and regional atrophy. Rob, is it important at this stage of the game to distinguish between white matter atrophy, gray matter atrophy, with regards to the brain and spinal cord atrophy? I mean, what’s our stage, our level of understanding and whether even it’s important to distinguish regionally with regards to atrophy?

DR. ROHIT BAKSHI: Well, it’s an excellent question. There really is a lot of confusion out there because there are so many different ways to measure atrophy and, as you already mentioned, atrophy affects the brain and the spinal cord, so it can be certainly be confusing to try to understand all the various data.
But let me try to make it as simple as possible. MS is a diffuse disease and a widespread disease which affects much of the central nervous system, and the same is true for the process of atrophy. So the spinal cord is commonly involved. The gray matter of the brain is involved, as is the white matter. So it’s very easy to be able to lump those together, for example, and measure whole brain atrophy quantitatively, or visually by looking at films, and follow patients over time to get an overall sense of the atrophic process.
But when one then tries to correlate specific clinical symptoms in patients, for example, physical disability, ambulatory impairment or various aspects of neuropsychological impairment, these gross general measures show weaker correlations with specific symptoms than if we look specifically at regional or structure-specific atrophy.
I’ll give you one good example. If we look at a correlation between whole brain atrophy and physical disability which is measured on the EDSS scale, which is overall disability, we see moderate correlations, in the range of 0.3, 0.4, meaning we’re explaining perhaps 10 to 20% of the variants in EDSS. So we’re not really there yet.
But then, if we specifically hone in and we decide, for example, let’s only look at ambulatory impairment in patients with MS, which is one aspect of disability, and let’s focus in only on atrophy of the spinal cord -- for example, at the highest cervical level. We now see much better correlations. We see our values more in the range 0.7, 0.8, especially in patients with progressive disease, primary progressive and secondary progressive MS.
The same holds true for cognitive impairment. If you want to specifically look at one aspect of cognition, such as informational processing speed, the overall whole brain atrophy correlates with that only, again, in the moderate range. But if we specifically look at atrophy, say, in the central part of the brain with third ventricular enlargement or thalamic atrophy or atrophy in and near the caudates, we see much more robust correlations with specific aspects of cognitive impairment.
So that’s the long answer to your question. The shorter answer is that both whole brain or gross spinal cord atrophy and regional atrophy measurements all have their place in following patients with MS.

J. THEODORE PHILLIPS, MD, PhD: So in view of that, and this question is to both of you, recognizing the contributions that both of you have made to this area in particular, there’s obviously more than one way to measure atrophy, whether it be whole brain or regionally. Is that standardized now? Does everybody do it the same way? Do you two guys do it the same way?

DR. ROBERT ZIVADINOV: Yes. You are absolutely right. There are probably so many different ways, you know, to look and classify central nervous system atrophy. You know, many measures in the first place looked on so-called absolute volumes of the brain, and then in recent years we understood that probably if you normalize the whole brain volume to the total intracranial volume, then you will get and assess the patient for interindividual differences and normalize that brain. That can avoid some inconsistencies if there are different patients or different subjects which have different sizes of the head and also can help when you put the patient in your MRI scanner, because the patient is usually moving, so having a normalized measure can help in comparing the differences.
However, when you are referring, really, to methods of measurements, then we have, I think, two big classes. Surely, the global measures -- and when I’m saying the global measures, I’m referring to whole brain as well as different compartments of the brain, principally gray matter and white matter. And then some regional measures, as Dr. Bakshi already mentioned, which are really becoming of great interest in the last couple of years.
We still don’t know which region is probably the best one in the follow-up of the patients and we probably have to do many more studies to understand how over time, in the short and long term, these global and regional measures are evolving.
On the other hand, I think that there are more than 80 different software programs which are capable to measure whole brain atrophy, and unfortunately because everybody is using their own software, it’s very difficult to compare results between different clinical trials or between different research studies.
Recently, there have been some groups which introduced more acceptable and accurate measures, and actually allowed these measures to be used for free. However, quantitative measurement of brain and spinal cord atrophy is not very easy to be done in normal clinical practice and high quality laboratories are required to perform such a job.
Obviously, the person should be trained, and as Dr. Bakshi referred before, we can visually observe over time, over long-term time, if the patient develops brain atrophy. However, probably we are interested more to be accurate from visit to visit, and just looking on the scans we cannot recognize the brain atrophy of less than 1% change. So maybe later I will refer more to different disease types and different changes over time in brain atrophy.

J. THEODORE PHILLIPS, MD, PhD: Do the brain atrophy measurements, through whatever algorithm they are determined by, provide a very good correlation between either whole brain or certain regional atrophy measurements with the more conventional MRI measures like gadolinium or T1/T2? Or are we talking, these are all separate and not very well interrelated measures?

DR. ROHIT BAKSHI: Well, the answer to that is we now have a lot of data. Several groups have looked at this, first in cross-sectional studies. How well do T2 lesions or T1 hypointense lesions or gadolinium-enhancing lesions correlate in a cross-sectional way with various atrophy measurements? And then, secondly, how well do these baseline lesion assays predict changes in volume or progressive atrophy? The answer that’s come out of those studies is very interesting, and the bottom line is that there is only really a moderate correlation between lesions and tissue destruction going on throughout the brain, whether it be in white matter or gray matter.
I think what it’s telling us is that the overall process of atrophy is not simply tissue loss within overt lesions. There is a diffuse process occurring in MS which affects areas that otherwise appear normal on T2-weighted images or even on flare-based images, and this pathology is widespread in white matter and gray matter. So it’s made us rethink MS as more of a multifactorial disease in which many pathologic processes contribute to the overall disease burden we see clinically.

J. THEODORE PHILLIPS, MD, PhD: So actually, just to go back to that and expand on that a little bit, Rob, are you saying that atrophy measurements may actually reflect another process separate from, different from, perhaps linked to, but different from the inflammatory process?

DR. ROHIT BAKSHI: We don’t know the answer to that yet. One can certainly speculate that perhaps inflammation is only one aspect of a cascade of events that is occurring in the brain, and overt lesions are only affecting one aspect of that overall process. Another possibility is that there are low levels of inflammation that are less obvious on MRI that are occurring in normal-appearing areas which contribute greatly as well to atrophy. We really simply don’t know yet.
But the take-home message here is that simply assessing our patients only by counting their lesions or assessing their lesion load clinically is not sufficient to give us the overall picture of their disease.
One excellent example of this is a series of studies that have been done by various groups with a common theme, which is that when we compare side by side lesion measurements and atrophy measurements against disability and cognitive impairment, the atrophy measurements stack up very well and are the most clinically relevant in correlating with and predicting clinical impairment. So the bottom line is that atrophy is important and certainly complementary to our lesion measurements.

J. THEODORE PHILLIPS, MD, PhD: Go ahead, Robert. I’d like to hear what you see.

DR. ROBERT ZIVADINOV: Right. But what we do know is that even if we are suppressing completely the inflammation in multiple sclerosis patients, we are not able to suppress completely the brain atrophy development, and several studies now are emerging which show that when you completely suppress, or suppress almost 90%, and you significantly reduce T2 and T1 lesion load, you are not able to reduce the brain atrophy progression. We do not know yet the reason, what is causing that, but probably there are different other mechanisms from inflammation, as I said before, which are probably contributing in different disease stages to undergoing brain atrophy development. And I will mention some of them.
Surely, when the inflammation in the first part of disease is destroying myelin sheath, the axons are remaining by themselves without the myelin, and that with time can cause so-called sub-lethal axonal injury, which means that there seems to be a delay between the time when the inflammation occurs and the time when the brain atrophy occurs, and some of that is really the loss of trophic support of the axons. That’s one mechanism.
Other mechanisms include, surely, direct cytotoxic T lymphocyte action on axons which does not require inflammation. And then, we have seen, as Dr. Bakshi showed in several studies, that also deposition of iron in the gray matter, and especially in the deep gray matter, can contribute to brain atrophy development. So these are all very, very new things that we are learning in the last couple of years.
On the other hand, I would say that really several studies now showed that in the long term, brain atrophy seems to be a better predictor of disability and narrow psychological dysfunction than are the T2 or T1 lesion measures. And partly the reason for that is that there is widespread damage in the brain, and atrophy reflects just the change of the number of the pixels in the brain which are the units we are measuring. And as long as there are units which represent the tissue, then we will find probably better correlation with those clinical measures.

J. THEODORE PHILLIPS, MD, PhD: So along those lines, inasmuch as several of the atrophy studies, longitudinal atrophy studies, have been in the context of investigational therapeutic trials, do you see in the future brain atrophy measurements as becoming actually more than just a research investigational tool, but something that will have direct import to the neurologist in practice? Obviously, it’ll take more standardization and so forth. To do atrophy, do you need a different type of scanner to do that, or is this all a software thing?

DR. ROHIT BAKSHI: It’s actually gotten much more user-friendly in the past few years. Originally -- say, five years ago or more -- the brain atrophy measurements, especially whole brain atrophy, were fairly inaccessible to the typical neurologist practicing in the community, or even in an academic center.
But now there are software tools that are available off the shelf you can purchase. You can install them at your MRI center or at your practice location, and standard MRI scanners can send images to the software program, and you can then produce a brain parenchymal fraction measurement, for example, fairly easily in a fairly limited amount of time -- say, 15 minutes to up to 30 minutes of technician processing time.

J. THEODORE PHILLIPS, MD, PhD: And the brain parenchymal fraction is?

DR. ROHIT BAKSHI: So brain parenchymal fraction is one of the mainstream measures that are available and widely used now to assess, or at least estimate, whole brain atrophy. And brain parenchymal fraction is referred to as such because it’s normalized, or it’s fractionated against the size of the person’s skull or the size of their head. So it’s normalized for that individual, and makes it particularly robust for longitudinal monitoring.

J. THEODORE PHILLIPS, MD, PhD: So with the increased availability of software to do this, and not needing special modifications made to one’s local scanner, is this something at this point in time that neurologists in practice, especially those with an interest in MS, should consider, start looking at in day-to-day practice? Is this something that is important to help monitor their patients, either on or off therapy? I’d like to hear from both of you all.

DR. ROBERT ZIVADINOV: Right. You’re completely right. I think that we are pretty close to that point, but we are still not there, and there is a lot of work to be done still. Surely, all these software programs that Dr. Bakshi mentioned have to become more user-friendly, and we found in our recent studies that the patient has to be followed always on the same scanner, because the inter-scanner changes are very important to measurement of brain atrophy. And also, if you change the software program over time, that can also affect the brain atrophy measurement.
So we, in our clinic, we have now a program which is assessing brain atrophy measurement on a clinical routine basis, and it’s an experimental project. And we evaluated more than 400 patients longitudinally over the last two years, and we are now in the process of understanding how these measure are correlating not with disability in terms of expanded disability status or some other fancy clinical measures, but really in terms of clinical improvement, worsening or stability from their last visit, when the patient was seen.
Moreover, we are assessing how these changes in quantitative measurement of brain atrophy are correlating with the qualitative neuroradiological report in terms of atrophy which has been, you know, reported by the neuroradiologist or not. So can we detect, really, the subtle, invisible changes in the brain which contribute to patient deterioration in all clinical areas, or we cannot do that.
So this is something, when we will be to answer these questions, then I think we’ll be able to propose these measures, because you have to also remember that it’s not easy to interpret the changes of brain atrophy or T1 and T2 lesion measures over time in the clinical routine practice. And the reason for that is that not all the measures are following the same directions, which means you can have an increase in T1 and T2 lesion load and new gadolinium-enhancing lesions, and, surprisingly, you don’t have brain atrophy, but you have more brain volume. And the reason for that is that overall inflammation can increase the amount of the brain volume. On the other hand, you can have improvement in T1 and T2 measures, but you have a substantial decrease in brain parenchymal fraction.
So when we will be able to understand how these correlate in real clinical routine practice, then we will be able to probably develop instruments which, in my view, should be done at the source of acquisition, which means directly at the scanner, and the patient will be then able to get these measures as a standardized protocol.

J. THEODORE PHILLIPS, MD, PhD: Rob, a final comment from you about how close we are to actually being able to practically use atrophy measurements in clinical practice to follow patients with MS on or off of therapy?

DR. ROHIT BAKSHI: I think we should be using them. We certainly use them in our center. I think that while there clearly are software tools available and fancy computerized methods, there’s really nothing to be able to replace a neurologist taking care of a patient to look at the films clinically, compare the baseline film to the followup film and look for progressive atrophy on a qualitative basis. Because we can see fairly large changes just by our trained eye.
And certainly, if somebody has progressive atrophy such that their corpus callosum is shrinking or their third ventricle is enlarging or their spinal cord is shrinking, that clearly is probably going to impact on their clinical progression, and it should be incorporated into how we think about MS for our patients.

J. THEODORE PHILLIPS, MD, PhD: Great. Well, we’re about out of time for this program. I want to thank my colleagues, Dr. Robert Zivadinov.

DR. ROBERT ZIVADINOV: Thank you.

J. THEODORE PHILLIPS, MD, PhD: Thank you. And Rob Bakshi for joining us today. This has been a great discussion. You know, I think that one of the most dramatically evolving and interesting fields in all of medicine, and MS in particular, is the neuroradiology and neuroimaging part. I think there’s a lot more to come from this, and I’ll be looking forward to, in the future, contributions from both of you.
Thanks to the listening audience for joining us for this program. I’m Dr. Ted Phillips.

Case Study: Brain Atrophy

J. THEODORE PHILLIPS, MD, PhD: Hello, and welcome to the case study portion of our program on brain atrophy and multiple sclerosis. I’m your host, Dr. Ted Phillips. I’m pleased to be joined today by Dr. Rob Bakshi, Director of Clinical MRI MS Imaging at Brigham and Women’s Hospital in Boston, and Dr. Robert Zivadinov, Associate Professor of Neurology at the State University of New York at Buffalo School of Medicine and Biomedical Sciences and the Director of the Buffalo Neuroimaging Analysis Center. Welcome to both of you.

DR. ROBERT ZIVADINOV: Thank you.

DR. ROHIT BAKSHI:Thank you.

J. THEODORE PHILLIPS, MD, PhD: Rob, can you please start us off with a description of the case that you have for us today?

DR. ROHIT BAKSHI:I’d be happy to. This is a 30-year-old woman who was diagnosed with MS in 1993. For the next several years she had multiple relapses. She had a very good recovery, with IV methylprednisolone being given for the relapses, and she was intermittently placed on interferon therapy. However, approximately five years after her diagnosis she had two pregnancies, and during that time interferon therapy was discontinued.
Shortly thereafter, she presented to our clinic with a complaint of poor word finding and also some difficulty with comprehension. Despite these cognitive complaints, a hard neurologic exam looking for physical findings was unremarkable. She had no evidence for weakness, numbness or ambulatory dysfunction. Yet clearly, she had complaints of difficulty performing her usual job as a customer service representative.
An MRI was performed, and I am showing you here the flare axial images that are representative of the findings. You can see there’s a mild hyperintense lesion load in the right hemisphere consistent with MS, and a mild to moderate lesion load in the left hemisphere. Overall, the lesion load is not very impressive, despite the cognitive complaints that she presented with.
I know show you T1 sagittal images which are excellent images for being able to look at the corpus callosum and other midline structures, and you can see that there’s evidence for tissue loss. Remember that this woman is in her 30s. And in the next slide, I’ve placed beside this image a normal control age-matched to the MS patient which highlights the corpus callosum atrophy in our patients.
In addition, showing you the flare axial images now, side by side with an age-matched normal control to emphasize the fact that this patient has substantial brain volume loss. Notice the widening of the third ventricle in this slide. In the next slide, you see the prominence of the lateral ventricles in the patient versus the normal control, and also the widening of the cortical sulci.
In the next slide, again, you see similarly that there is prominence of the ventricles and there is enlargement of the cortical sulci.
Again, in the next slide, you see more dorsally the prominence of the cortical sulci diffusely. All these findings are consistent with diffuse brain atrophy.
The patient then underwent neuropsychological testing by a neuropsychologist in our clinic, and she had substantial abnormalities on the neuropsychological exam. She had word-finding difficulty, and she had personality features, as well, and other neurobehavioral features, such as a labile affect and some euphoria and irritability. And when we interviewed her spouse, the spouse also reported, really, more significant cognitive complaints than even the patient had admitted to. And, as I already mentioned, this was affecting her employability and her job performance.
Turning now to the last slide, you’ll see a graph that illustrates that this patient had substantial objective neuropsychological impairment compared to normal controls. The normal controls are shown in the slide side by side with the patient, and you see there are several areas where the patient is far below the 50th percentile, which would be normal. Again, this highlights the fact that the patient’s cognitive complaints were indeed verified by objective testing.
So, to summarize, this is a patient who has relapsing-remitting MS who has very little in the way of disability in terms of physical disability, but clearly has major neuropsychological impairment. And the MRI, despite showing fairly mild findings in terms of lesion load, has substantial evidence for diffuse brain atrophy.

J. THEODORE PHILLIPS, MD, PhD: Very good and interesting. How often do you see somebody like this? Is this a typical sort of scenario, Robert?

DR. ROBERT ZIVADINOV: This is really a typical scenario, more often even than we think. This is really a very young patient, as Dr. Bakshi mentioned, who has no disability or minimal disability, but has, when tested with appropriate instruments, moderate to severe cognitive dysfunction, including some personality changes. And typically, you don’t see a lot of the lesions widespread in the brain. As we saw in this case, there were really minimum to mild hyperintensive lesion load, usually distributed in a periventricular area, but what was prominent was surely the brain atrophy, and typically, we saw more of the brain atrophy distributed in some regions.
I remember the slide which showed clearly a thinning of the corpus callosum, which is one important area for cognitive dysfunction. And on the other hand, there was also shrinking of the sulci in the cortical superior frontal parietal areas of the brain, which we know are correlating very well with cognitive dysfunction.
On the central slides and on the flare and T1-weighted images, we showed prominent widening of the third ventricle and also enlargement of the lateral ventricles, which we also know are one of the best predictors of cognitive and personality changes. There are several reasons why the central atrophy could be so important for cognitive dysfunction, but probably the most important one is that so many structures of the deep gray matter which are very involved in cognitive function are situated in those strategic areas, and they can contribute to cognitive dysfunction.
And obviously, this is another typical case for the reason that she had, with interferon therapy, some attenuation of the symptoms but then obviously she interrupted the therapy because of her two pregnancies, and then the attacks were repeated, and some more cognitive deterioration developed.

J. THEODORE PHILLIPS, MD, PhD: Rob, the slides that you presented impressed me on a variety of counts. But one thing that really occurred to me was that the slides that you showed were showing atrophic measurements, or showing evidence of atrophy, and that you could see this. It didn’t require sophisticated software analysis to pop out, and it makes me think that even if you don’t have the software to do this, these are the kinds of comparisons that you could make in the office. You know, you mentioned the ventricular width. I guess the key thing is having a prior film to compare this to and looking at it yourself. I mean, is that valid, that these are things that you can usually see? You may not be able to quantitate it, but you can actually see it?

DR. ROHIT BAKSHI:Absolutely. I think there are two real important points here. The first point is that neurologists, as we take care of patients with MS, we should integrate reviewing of films into the routine care in our own office. So if the patient comes for a followup, we should have their films in front of us. We should have access to them either by CD or by hard copies, and we should review them in the course of seeing the patient. Simply just reading the report, the written report of the MRI, is not sufficient, because we have to really give it a thorough evaluation for evidence of things such as atrophy, which are clearly important clinically, and I think this case nicely illustrates that.
I think secondly, the degree of atrophy we can detect is, surprisingly, fairly small amounts of atrophy. If we keep in mind, number one, what the baseline scan looked like and, number two, what a normal control of that same age should look like. We all have experience in our clinic. Many of us see patients who are normal, who don’t have MS. We get normal films on them, and we get an idea by looking at many films what the third ventricle should look like in a 30-year-old. What should the cortical sulci look like? And we can compare that body of knowledge to the current patient that’s in front of us.

J. THEODORE PHILLIPS, MD, PhD: Yes, I think this case really nicely illustrates that, and also that you can get significant atrophy and yet still be physically reasonably intact. And that, I think, is another interesting message here, as well.
Any last kind of parting summary comments from both of you all? Robert?

DR. ROBERT ZIVADINOV: I think that we are surely getting into the moment when we will be better able to measure brain and spinal cord atrophy, to understand from more longitudinal long-term studies what is the importance of brain atrophy in MS. Surely, just the measurement of brain atrophy is not an answer on all the changes we are seeing in multiple sclerosis patients, but is an important measure to contribute to our understanding of MS.
I think that with time, in the next few years, these measures will become even more user-friendly, and we will be able to offer them in clinical routine practice. I think that that’s our really ultimate goal.

J. THEODORE PHILLIPS, MD, PhD: Rob?

DR. ROHIT BAKSHI:I think that cognitive impairment is very relevant in our patients, and we shouldn’t be fooled by a patient who has no physical disability, and a patient who has a relatively mild lesion load and lesions look fairly stable, we should always keep in mind that there is a fairly close relationship between cognitive impairment and quality of life, and that both of those are related at least moderately well to atrophy, brain atrophy.

J. THEODORE PHILLIPS, MD, PhD: Well, I’d like to thank both of you all for joining us, and Rob, particularly, for contributing this case. I’d like to thank Robert Zivadinov for joining us, and his helpful thoughts regarding this case. Dr. Rob Bakshi for presenting the case.
I’d like to also thank the listening and viewing audience for their attention. I hope this has been informative to you. I’m Dr. Ted Phillips.