Optimizing MRI Protocols for Epilepsy Diagnosis

Neuroimaging is a tool to classify and ascertain the etiology of epilepsy in people with first or recurrent unprovoked seizures. In addition, imaging may help predict the response to treatment. To maximize diagnostic power, it is essential to order the correct imaging sequences.

In this episode, Aaron Berkowitz, MD, PhD, FAAN speaks with Christopher T. Skidmore, MD, author of the article “Neuroimaging in Epilepsy,” in the Continuum February 2025 Epilepsy issue.

Dr. Berkowitz is a Continuum® Audio interviewer and professor of clinical neurology at the University of California, San Francisco

Dr. Skidmore is  an associate professor of neurology and vice-chair for clinical affairs at Thomas Jefferson University, Department of Neurology in Philadelphia, Pennsylvania.

Additional Resources

Read the article: Neuroimaging in Epilepsy

Subscribe to Continuum: shop.lww.com/Continuum

Earn CME (available only to AAN members): continpub.com/AudioCME

Continuum® Aloud (verbatim audio-book style recordings of articles available only to Continuum® subscribers): continpub.com/Aloud

More about the American Academy of Neurology: aan.com

Social Media

facebook.com/continuumcme

@ContinuumAAN

Host: @AaronLBerkowitz

Guest: @ctskidmore

 Full episode transcript available here

Dr Jones: This is Dr Lyell Jones, Editor-in-Chief of Continuum, the premier topic-based neurology clinical review and CME journal from the American Academy of Neurology. Thank you for joining us on Continuum Audio, which features conversations with Continuum's guest editors and authors who are the leading experts in their fields. Subscribers to the Continuum journal can read the full article or listen to verbatim recordings of the article and have access to exclusive interviews not featured on the podcast. Please visit the link in the episode notes for more information on the article, subscribing to the journal, and how to get CME.

Dr Berkowitz: This is Dr Aaron Berkowitz, and today I'm interviewing Dr Christopher Skidmore about his article on neuroimaging in epilepsy, which appears in the February 2025 Continuum issue on epilepsy. Welcome to the podcast, Dr Skidmore. Would you please introduce yourself to our audience? 

Dr Skidmore: Thank you for having me today. I'm happy to talk to you, Dr Berkowitz. My name is Christopher Skidmore. I'm an associate professor of neurology at Thomas Jefferson University in Philadelphia. I'm a member of the Jefferson Comprehensive Epilepsy Center and also serve as the vice chair of clinical affairs for the department. 

Dr Berkowitz: Thank you very much for joining us and for this fantastic article. It's very comprehensive, detailed, a very helpful review of the various types of brain pathology that can lead to epilepsy with very helpful images and descriptions of some of the more common findings like mesial temporal sclerosis and some of the less common ones such as cortical malformations, heterotopia, ganglioglioma, DNET. So, I encourage all of our listeners to read your article and take a close look at those images. So, hopefully you can recognize some of these findings on patients’ neuroimaging studies, or if you're studying for the right or the boards, you can recognize some of these less common congenital malformations that can present in childhood or adulthood with epilepsy.

In our interview today, what I'd like to do is focus on some practical tips to approaching, ordering, and reviewing different neuroimaging studies in patients with epilepsy. So to start, what's your approach when you're reviewing an MRI for a patient with a first seizure or epilepsy? What sequence do you begin with and why, how do you proceed through the different sequences and planes? What exactly are you looking for? 

Dr Skidmore: It's an important question. And I think to even take a step back, I think it's really important, when we're ordering the MRI, we really need to be specific and make sure that we're mentioning the words seizures and epilepsy because many radiology centers and many medical centers have different imaging protocols for seizure and epilepsy patients as compared to, like, a stroke patient or a brain tumor patient. I think first off, we really need to make sure that's in the order, so that way the radiologist can properly protocol it. Once I get an image, though, I treat an MRI just like I would a CAT scan approach with any patient, which is to always approach it in the same fashion. So, top down, if I'm looking at an axial image. If I'm looking at a coronal image, I might start at the front of the head and go to the back of the head. And I think taking that very organized approach and looking at the whole brain in total first and looking across the flare image, a T2-weighted image and a T1-weighted image in those different planes, I think it's important to look for as many lesions as you can find. And then using your clinical history. I mean, that's the value of being a neurologist, is that we have the clinical history, we have the neurological exam, we have the history of the seizure semiology that can might tell us, hey, this might be a temporal lobe seizure or hey, I'm thinking about a frontal lobe abnormality. And then that's the advantage that we often have over the radiologist that we can then take that history, that exam, and apply it to the imaging study that we're looking at and then really focus in on those areas. But I think it's important, and as I've illustrated in a few of the cases in the chapter, is that don't just focus on that one spot. You really still need to look at the whole brain to see if there's any other abnormalities as well.

Dr Berkowitz: Great, that's a very helpful approach. Lots of pearls there for how to look at the imaging in different planes with different sequences, comparing different structures to each other. Correspondent reminder, listeners, to look at your paper. That's certainly a case where a picture is worth a thousand words, isn't it, where we can describe these. But looking at some of the examples in your paper, I think, will be very helpful as well. So, you mentioned mentioning to the neuroradiologist that we're looking for a cause of seizures or epilepsy and epilepsy protocols or MRI. What is sort of the nature of those protocols if there's not a quote unquote “ready-made” one at someone 's center in their practice or in their local MRI center? What types of things can be communicated to the radiologist as far as particular sequences or types of images that are helpful in this scenario?

Dr Skidmore: I spent a fair amount of time in the article going over the specific MRI protocol that was designed by the International League Against Epilepsy. But what I look for in an epilepsy protocol is a high-resolution T2 coronal, a T2 flare weighted image that really traverses the entire temporal lobe from the temporal tip all the way back to the most posterior aspects of the temporal lobe, kind of extending into the occipital lobe a little bit. I also want to see a high resolution. In our center, it's usually a T1 coronal image that images the entire brain with a very, very thin slice, and usually around two millimeters with no gaps. As many of our neurology colleagues are aware, when you get a standard MRI of the brain for a stroke or a brain tumor, you're going to have a relatively thick slice, anywhere from five to eight millimeters, and you're actually typically going to have a gap that's about comparable, five to eight millimeters. That works well for large lesions, strokes, and big brain tumors, but for some of the tiny lesions that we're talking about that can cause intractable epilepsy, you can have a focal cortical dysplasia that's literally eight- under eight millimeters in size. And so, making sure you have that nice T1-weighted image, very thin slices with no gaps, I think is critical to make sure we don't miss these more subtle abnormalities.

Dr Berkowitz: Some of the entities you describe in your paper may be subtle and more familiar to pediatric neurologists or specialized pediatric neuroradiologists. It may be more challenging for adult neurologists and adult neuradiologists to recognize, such as some of the various congenital brain malformations that you mentioned. What's your approach to looking for these? Which sequences do you focus on, which planes? How do you use the patient 's clinical history and EEG findings to guide your review of the imaging?

Dr Skidmore: It's very important, and the reason we're always looking for a lesion---especially in patients that we're thinking about epilepsy surgery---is because we know if there is a lesion, it increases the likelihood that epilepsy surgery is going to be successful. The approach is basically, as I mentioned a little bit before, is take all the information you have available to you. Is the seizure semiology, is it a hyper motor semiology or hyperkinetic semiology suggestive of frontal lobe epilepsy? Or is it a classic abdominal rising aura with automatisms, whether they be manual or oral automatisms, suggesting mesial temporal lobe epilepsy? And so, take that clinical history that you have to help start to hone your eye into those individual locations. But then, once you're kind of looking in these nonlesional cases, you're also then looking at the EEG and where their temporal lobe spikes, where their frontal lobe spikes, you know, using that and pulling that information in. If they saw a neuropsychologist pulling in the information in from the neuropsychological evaluation; if they have severe reductions in verbal memory, you know, focusing on the dominant temporal lobe. So, in a right-handed individual, typically the left temporal lobe. And kind of then really spending a lot of time going slice at a time, very slowly, because in some of these vocal-cortical dysplasias it can be just the blurring of the gray-white margin. What I find easiest is to identify that gray-white margin and almost track it. Like, you use the mouse to kind of track it around and say, can I outline the exact border of the gray white margin in the frontal lobe that I'm interested in or the temporal lobe that I'm interested in, kind of looking for those subtle abnormalities.

Often as neurologists, we don't have the luxury of being able to immediately reformat. As I mentioned, our T1 volume acquisition study is done in the coronal plane, but sometimes you might want it in the axial plane. And so, I might reach out to the radiologist and say, hey, can you reformat this in the axial plane because I'm interested in the frontal lobe epilepsy and it's a little bit better at looking at it in that plane? And I'll have them reformat and put it back on the pack so I can look at it in that manner. And so that's a, kind of another strategy is to take what you have, but also then go back to the radiologist and say, I need to look at it this a different way. Can you reformat it for me? Looking for that gray-white matter junction is the nice way to pick up for kind of subtle cortical dysplasias. And then when you see an abnormality, to be able to put the T1, the T2, and the flare image all up next to each other and use the technology built into most of our browsers to put on what's called the localizer mode, where I can highlight a specific spot that I'm seeing on the T1 and then very easily quickly see, what does it look like on the T2? What does it look like on the flare? To kind of quickly decide, is it a true abnormality or am I only seeing it on one slice because of an artifact on that one imaging sequence? And I think that's the biggest kind of key is to make sure, is it an artifact or is it not an artifact? That's kind of the most common thing that we, I think, get confused with. 

Dr Berkowitz: So, some very helpful pearls there in terms of reviewing the imaging, being in dialogue with our neuroradiology colleagues to think about potentially reacquiring certain images on certain planes or looking at the images with our neuroradiology colleagues to let them know more about the clinical history and where we're sort of zooming in about possible abnormalities. 

Dr Skidmore: I would just add in there that when looking at especially the mesial temporal structures, because of a lot of artifacts that can be present in an individual MRI machine, it's not uncommon that the mesial temporal structure will appear brighter because of an MRI magnet artifact. And so, it's a good key to look at the hippocampus compared to the insula. And so, the hippocampus and the insula should have similar signal characteristics. You're seeing the hippocampus is bright, but the insula  ipsilateral to it’s normal intensity. That would suggest that that's probably a true hyperintensity on the flare-weighted image as opposed to if both are bright, unless you're suspecting a hemispheric abnormality, it's more likely to be a kind of artifact in the MRI machine.

Dr Berkowitz: Okay. Those are really helpful tips, not just to analyze the hippocampus and medial temporal lobe itself---let's remember our anatomy and the circuit of Papez---and to look at associated structures for supporting evidence of a possible abnormality in the hippocampus itself. It looks like there may be something subtle. We can use some additional information from the image to try to decide if that is real or artifactual, and of course correlating with the clinical picture and EEG. I'd like to talk briefly now about some other imaging modalities that you discuss in your paper, the use of functional imaging such as PET, SPECT and fMRI. Let's talk a bit about each of these. When would you order a PET scan for a patient with epilepsy? What would you be looking for and how would you be using that to make clinical decisions? 

Dr Skidmore: Yeah, so these functional imaging modalities are really utilized when we're evaluating somebody that's not responding to medications. So, they're medically intractable, and we're wondering, could they be a candidate for epilepsy surgery? And so, most of these imaging modalities are really relegated to the world of epileptologists at surgical epilepsy centers. I wanted to include them, though, in the article because I do think it's important for general neurologists to understand kind of what they are, because invariably a patient sees me and then they go back to their general neurology and be like, hey, Doctor Skidmore said I had this PET scan abnormality. What do you think? So, I think it's a good idea for general neurologists to kind of understand them. So, probably the oldest that we've utilized is the FDG PET scan, basically looking at fluorodeoxyglucose and the brain's utilization of glucose. As we all remember, again, glucose is the primary molecule for energy and ATP production in the brain. And so basically, by injecting radioactive glucose in the interictal state--- so not during a seizure but in between seizures---areas of the brain that are not taking up the radiotracer will show as being hypometabolic. So, low metabolism. And hypometabolic regions in the interictal state have been associated with onset regions for epileptic seizures.

Let's say you have a patient clinical history, you think they have temporal of epilepsy, EEG suggests temporal of epilepsy, but the MRI is nonlesional, meaning there's no abnormality that anybody could appreciate even at a 3 Tesla scanner. We'll get an FDG PET scan and see, is there hypo metabolism in that temporal lobe of interest? And if there is, well, that's been shown through several published papers, that's just as valuable as having an abnormality on the MRI. And so, we often again use these PET scans, especially in nonlesional cases, to try to find that subtle cortical dysplasia. Now you have your nice epilepsy protocol MRI, it says it's nonlesional. You get your PET scan, it shows hypometabolism in a region of the frontal lobe, let's say, in a in a frontal lobe epilepsy case. And then often we go back, we kind of talked about strategy of how you find those subtle lesions. Then you go back and say, well, look, this gyrus specifically on the PET scan said it's abnormal. You end up looking for really subtle, very tiny abnormalities that, even with somebody that's skilled, often at first review gets missed. So, that's how we use the PET scan.

The SPECT scan is done typically in the ictal state. So, now somebody’s in an epilepsy monitoring unit often, where you're injecting radio tracer at the exact moment that somebody starts having a seizure. And we know when there's increased seizure activity, the increased seizure activity---let's say it's from my right temporal lobe---is going to increase cerebral blood flow transiently to the right temporal lobe. And then if that seizure discharge spreads from the right temporal lobe maybe to the entire right hemisphere and eventually becomes a focal to bilateral tonic chronic seizure by spreading to the other side, the entire brain is going to be hypoperfused at that point. So, if you want to, as soon as the seizure starts, inject that radio tracer to see, where is the blood flow earliest in the seizure? And then we might do an interictal SPECT when you're not having a seizure. Look at, all right, what's the normal blood flow when somebody's not seizing? What's it like when they're having a seizure? And then the area that has increased activity would- might suggest that's where the seizure started from. But we have to be very careful because again, some seizures can spread very rapidly. So, if you delay injecting an injection ten, fifteen, twenty seconds, the seizure could have already propagated to another region of the brain, giving you a false positive in another location. So, you have to be very careful about that modality.

I think what's most exciting is the functional MRI because the functional MRI, for many, many centers, is replacing a very old technique called the WADA test. So, in the WADA test, typically you place a catheter angiogram into the internal carotid artery and transiently introduce a sedative medication to put, let's say, the left hemisphere to sleep because you wanted to see what functions were still active in the right hemisphere. And then the surgeon would move the catheter or the right internal carotid artery, and you inject a sedative on that side after the left hemisphere is recovered and see what the left hemisphere can do. And we used that for language dominance, we used that for memory dominance. And while most individuals did fine with angiograms, unfortunately complications do occur and there's injury to the artery, there could be strokes that can- that have happened, which can be quite devastating for the patient. And so, functional MRI is a nice, noninvasive way for us to map out language function, motor function, sensory function, visual function, and is starting to show some usefulness also for mapping out kind of memory function, dominant memory function, meaning verbal memory compared to visual memory. To be able to do those things noninvasively becomes really important because, if we're talking about epilepsy surgery, we want to make you seizure-free but neurologically intact. And so, we need to understand the relationship between where we think the seizures are coming from and where eloquent cortex is so we can properly counsel you and avoid those regions during any planned surgery. Those are the three most common functional imaging modalities that we're using now to supplement the rest of the presurgical work. 

Dr Berkowitz: Very helpful. So, these are studies, PET, SPECT, and fMRI, that would really be obtained predominantly in patients in whom epilepsy surgery was being considered to have more precise lesion localization, as well as with the fMRI to get a better sense of how to provide the safest maximal resection of epileptogenic tissue while preserving functions.

Dr Skidmore: That's a perfect summary. 

Dr Berkowitz: Fantastic. This has been a really helpful interview with Dr Skidmore and a really fantastic article. As I said, a picture is worth a thousand words, so I definitely encourage you to read the article and look at the images of some of the conditions we've been talking about and some of these findings that can be seen on interictal PET or ictal SPECT to get a sense of the visual aspects of what we've been discussing. So again, today I've been interviewing Dr Christopher Skidmore about his article on neuroimaging and epilepsy, which appears in the most recent issue of Continuum on Epilepsy. Be sure to check out Continuum audio episodes from this and other issues. And thank you so much to our listeners for joining us today. 

Dr Skidmore: Thank you for having me.

Dr Monteith: This is Dr Teshamae Monteith, Associate Editor of Continuum Audio. If you've enjoyed this episode, you'll love the journal, which is full of in-depth and clinically relevant information important for neurology practitioners. Use this link in the episode notes to learn more and subscribe. AAN members, you can get CME for listening to this interview by completing the evaluation at continpub.com/audioCME. Thank you for listening to Continuum Audio.