Dr. Marcel Maya at the 2023 Cedars-Sinai Intracranial Hypotension Conference
Dr. Marcel Maya, Professor & Co-Chair, Department of Imaging, Cedars-Sinai in Los Angeles, CA, presented this talk titled “Spinal Imaging: Diagnostic Algorithm I” at the 2023 Cedars-Sinai Intracranial Hypotension Conference on July 8, 2023. The conference was hosted by Cedars-Sinai with generous support from the Spinal CSF Leak Foundation in Kohala Coast, Hawaii.
Transcript
Dr. Maya discusses “Spinal Imaging: Diagnostic Algorithm I”:
[00:00:09] I want to thank the organizers, Wouter and Marcus for inviting me. It’s a pleasure to be here in person in Hawaii, and I wanna acknowledge the space we’re in. It’s a beautiful space on earth and the likes of which there is none in my opinion. Anyway these are the official state symbols, including the Nene, the hibiscus, the flower, the official state flower.
[00:00:45] These are the taro fields. And the toughest one is this one. It’s the humuhumunukunukuapua’a [the reef triggerfish]. All right. So, um, but on a serious note, I’m here representing really the whole Cedars team. And even though Wouter’s on the second line, of course, he’s a leader and he’s been pushing this for 20 years. And I’ve been lucky enough to be tagging along and helping out.
[00:01:10] And I really appreciate the opportunity, but these are, on a day-to-day basis, people who do the studies, my colleagues in neuroradiology and I want to acknowledge them. Of course, um, the people who do the real work don’t get the top billing are these people in IR. They really are fundamental in producing all these wonderful images the DSMs, and these are tough patients to deal with, and they’re really peerless in, in their area. We have a large body of experience since 2001, over 3,000 patients evaluated and of those over 1,800 have met the CSF leak criteria. Of course, it’s a skewed population. Many of them from out of state, out of different parts of the country, and even international.
[00:02:02] Spinal imaging is important in the sense that it really is important in contributing to the diagnosis as well, because you also have the modified criteria, which includes imaging evidence of CSF leak. And yes, mostly the imaging evidence of CSF leak is on a cranial MRI. But sometimes cranial MRI is negative as we saw, especially the longer the leak symptoms go on, but the spinal imaging may give you the diagnosis.
[00:02:34] And our approached to spinal imaging and SIH is to initially to be noninvasive. And if you get anything out of the next 20 minutes, is that the noninvasive way to go is MRI with MR myelogram. And we rely heavily on that. Some of the patients, many of them go on and respond to conservative treatment and there’s no further need for further imaging, but some of them do need further imaging.
[00:03:03] And that’s when we need to really be specialized, find the leak site and be accurate about it so we can do targeted treatments, whether it’s injections or surgery or whatnot. And for these, we find that dynamic CTM or DSM is a good way to go. At our site, we do the DSMs. So, um, we have a multitude of imaging options.
[00:03:24] But really the ones that we rely heavily on, the MRI with T2 weighted myelographic sequences. And the DSM and we understand, and we’ll see in the next talk, of course, the dynamic CT plays very important part in many different centers.
[00:03:44] Let’s move on. I think this article has been shown many times this morning already, but it does give some support to our approach in terms of the accuracy and sensitivity of DSM in finding the leak site. In a meta review of studies and the sensitivity of T2 weighted, heavily T2 weighted MR spinal imaging for the presence of the leak, even though it’s not very specific about the site.
[00:04:11] Just wanted to show this algorithm where we start out with brain MRI with contrast and brain spine MRI with contrast and the heavily T2-weighted MR model graphic sequence confirms SIH. And they go on to conservative measures, including non targeted blood patches. Only if they have persistent symptoms, they go on to more invasive, depending on whether there is a ventral leak or lateral collection, ventral lateral collection.
[00:04:40] Let’s show some specific examples of how MR spine helps us. Here’s a patient with contrast enhancement study showing the dural thickening and enhancement, just like in the brain, that really is due to CSF leak. Another commonly seen sign is the leakage, also labeled as false localizing sign, that is seen at C1 C2 level.
[00:05:04] And here in this patient, not only that, but in the cervical thoracic junction in the upper thoracic spine. These are leaks that are seen not necessarily indicating the primary site of the leak, but as a phenomenon of the fluid leaking and leaking out at several points of low resistance. This can be seen very nicely with again, the myelographic T2 weighted images in the thoracic and lumbar spine.
[00:05:32] So just recognizing this would get us a long way towards making the diagnosis. For extrathecal collection, the MRI T2 weighted is very good and demonstrates this and we dare say it’s better than CT myelogram. I’ll show some examples of that. That can be seen on the heavy T2 weighted images as well.
[00:05:54] Wouter talked about the type 2 meningeal diverticula and what better than the T2 weighted myelographic sequences to show that demonstrated and various different appearances of these cysts, which may be situated anywhere from the top of the spine to sacrum or extensively as you see in that last case, in that last image.
[00:06:18] Here are a few examples from our study comparing the T2 weighted imaging myelographic sequence Dr. Tay wrote this one, and MRI is much better in demonstrating the CT in terms of the extrathecal collection. Now with the most advanced CT techniques, this may not be the case, but at the time it was written up a year or two ago, this was the case.
[00:06:42] We could not see the extrathecal collection on the post myelogram CT. Here’s another example of a large posterior collection that was very easy to pick up on MRI, but was not really appreciable on the CT myelogram. And a third example of a small leak, which is really very difficult to see on the CT. So we think this, we wrote this paper where we said that MR myelography is not inferior to CT myelographic myelographic techniques for extradural CSF detection.
[00:07:19] Let’s move on to more invasive DSM. And DSM starts out with first description of the DSM in a pseudomeningocele, but with a single case report. And later on, Hoxworth and colleagues demonstrated this first usage of CSF in the spontaneous hypotension patient ventral leak with DSM. We quickly adopted it and then later stumbled upon and reported this, a different mechanism of CSF leak with a fistula, as you can Dr. Shevink mentioned earlier. A major advance in the detection of these fistulas was advocated by Farb when he said that we’re better off scanning these patients on their sides to improve the chances. And we found that to be true, of course. And our yield of CSF leak detection increased from 15 percent to 75 percent when we did that.
[00:08:17] This slide was shown earlier with a positive brain MRI. Obviously, the yield is much higher. Just briefly about our technique. It is done in the biplane injury room with a tilt table. I know many other places which may not have a tilt table, use a device to prop up the patient. We do them under general anesthesia and for various reasons, but the most important one is to achieve very still and clear images without motion artifact.
[00:08:47] Not so much patient comfort. Um, we puncture with a 22 gauge Gertie Marx needle and follow up with a chaser. Here are some examples. This is a patient who has been suffering with CSF leak for long term. And initially we couldn’t diagnose her. We didn’t have the DSM technique at that time.
[00:09:09] And she had siderosis as well. She had surgery with mild improvement. When we did the DSM when she came back with recurrent symptoms, we saw that leak in the mid thoracic spine. I’m showing you this case not only for the leak shown on the lateral, but on the AP you can see nicely the point of the leak right there demonstrating on the left side of the patient’s spine with a bony spur at that location. This is a patient example of a lateral leak. You see a posterior collection and lateral collection, and this is the DSM of the same patient. And you can see the leak coming up around the nerve root, but just in the thecal sac and disseminating laterally and posteriorly.
[00:09:59] Another patient with a leak. This time the collection is posterior. The way we deal with this is we put a lumbar catheter and then flip the patient and do it in supine manner and demonstrate the actual point of the leak. This collection spanned a few segments, so it was important to demonstrate exactly where it was.
[00:10:20] Let’s come to the CSF venous fistula, which is the exciting leaks, which are so curious and difficult to visualize. And this patient had one at T3, as you can see,
[00:10:34] which was treated with endovascular approach with gluing the intercostal access to the epidural and neuroforaminal network with nice follow up imaging. CSF fistulas have many different appearances. I wish all of them had very obvious appearance like these that show here.
[00:10:54] Unfortunately, some of them are really tough to show. I’ll show you a few examples. I Also want to show a few traps that we can encounter with the CT myelo graphic techniques. And, um, this is a patient who had several blood patches, eventually had a diagnosis of a right-sided fistula with a nice demonstration of hyperdense para spinal vein published by Peter Kranz.
[00:11:22] She had a foraminotomy, but she did not have relief. When did we did the DSM on our study, we see that it’s coming from the other side. The leak itself originating from the other side, but the draining to the contralateral paraspinal vein. So that could be a trap. Seeing the paraspinal vein may not be the end all and be all.
[00:11:47] Another example of a paraspinal vein kind of ambushing us is when the leak originates the level below. Whereas this hyperdense vein is seen to level above because of its course. So that may result in some repeat interventions. This patient had a leak, a ventral leak. This was difficult to find on a CTM and you can see a spur there which made them believe there was a leak at T2-3.
[00:12:22] But after surgery, laminectomy, the symptoms did not go away. And on DSM, you can see one level below, two levels below, as a matter of fact, at T4-5. So that’s another trap there. The other interesting thing about the leaks and the CSF fistula is that they can occur anywhere from C1 to sacrum.
[00:12:47] This is an example that brings this home. A 39 year old woman who presented actually with bilateral upper extremity weakness. She did have headaches, but you can see the myelopathy from a cervical syrinx. And brain sag and all the elements of intracranial hypotension on an MR on the posterior fossa.
[00:13:06] Our imaging showed large sacral cysts and on DSM, we’re able to demonstrate a fistula coming out of one of those cysts. She had several interventions, including surgical interventions, percutaneous. But finally, one of the surgeries was successful in eliminating the the fistula at that level.
[00:13:27] She continues to have some issues, but this was at the time successful. This is that the on the opposite end, a patient, a 39 year old, with headaches and tinnitus. And we found this subtle leak in the cervical spine at C2 level. So DSM has its challenges, and there are potential solutions.
[00:13:50] Because of the limited detector size, we cannot cover C1 to S1. So we have to determine where the most likely leak is. And usually it’s the thoracic spine. The other challenge is to have to scan them two days apart for lateral decubitus optimal yield. Um, and the other one is some patients have really big shoulders and it’s tough to examine the thoracic, cervical thoracic junction in those patients, which happens to be a quite a common location for the ventral leaks where the, there’s a herniated disc. And sometimes the slow leaks and very small fistulas may be tough in bigger patients.
[00:14:29] There are some solutions there. The two station technique refers to doing the DSM at the same sitting, examining the upper cervical or thoracic spine and then doing the second injection in the lower part. That could be a solution to that. And complementary CT myelo is obviously helpful, dynamic CT myelogram.
[00:14:49] Lutzen and colleagues recently reported a very nice technique of cone beam CT to follow DSMs, and they show that the fistula could be shown on those images immediately after the DSM itself. And the respiratory maneuvers by the UCSF group with resistant breathing, which we’ll hear about. Peter will talk about them.
[00:15:11] This one such patient with a ventral leak. You can see extra thecal collection. You can see how lordotic he is, and he’s a big guy, and the, it, the DSM really, total wipeout at the cervical thoracic junction. And we repeated it from the cervical spine: no go. But the CT myelogram gave us the answer with a small calcified spur, and that was indeed the site of the leak.
[00:15:39] Some subtle leaks. Um, this one, you can see one of those fistulas, which is right adjacent to the thecal sac, subtle one. So it’s important to look carefully and compare it with the CT myelogram as well. You can see the vein coming up very subtly on that left side where the arrow is trying to point.
[00:16:05] And another one is on the right is here. So I’d like to conclude that MR myelography and DSM is really good for CSF leak in our shop. And it’s, MR myelogram is excellent. First line approach DSM is sensitive and specific for leak site determination, safe and well tolerated radiation. Those compares well with other techniques.
[00:16:28] And I want to upgrade Wouter’s license plates with this one. Thank you very much.