Slides
Transcript
Thanks again for the introduction. Okay, now I’m going to talk about the DSM technique. I have no disclosures.
So, I want to explain what DSM is. I also want to talk about the fluoroscopy and the cone beam CT. This is another technique you can do in the setting of the DSM, and then I want to give some examples of the DSM for different leak types.
So, the DSM technique is usually performed in angiography suite. This is how it looks like in Freiburg. We have two X-rays here, and this technique originally comes from the digital subtraction angiography – a technique that is used with a catheter to demonstrate, for example, intracranial vessels. And we, as radiologists, adapted this technique for the DSM technique. Therefore, we do a native, one single X-ray – this is called the mask – and then while we inject the contrast agent into the spinal channel, we do furthermore images, X-rays, and then we subtract one from each other. This, of course, is performed automatically. And then we only see, in a dynamic way, how the contrast is running into the image. This helps us to see the pathology.
The pro of this technique is especially the high temporal resolution, and we have something like a live leak observation. I’m going to show you some examples later on. The radiation dose is reported to be moderate compared to CT myelography, which is often a little bit more. The drawbacks are that this technique is very sensitive to movement and breathing artifacts, and we only do images in mono or biplane acquisition. And we have limited coverage because of the length of the detector.
This is the fluoroscopy. This can be performed in the setting of DSM, for example, after the DSM, and this is like a video with lower quality. You can make a diagnosis with it. For example, here you see very small contrast running into the epidural space in a patient with a lumbar puncture.
Then we have the cone beam CT. We create these images by rotating one X-ray, and then we get these cross-sectional images, which will offer a lot more anatomical information. This is similar to CT myelography but limited again by the detector – the coverage is limited again.
Now, I want to talk about some side effects with this technique. When you inject the contrast agent. It is quite common that you may evolve a little bit of headache, sometimes even severe. You may get nausea and vomiting, but this is quite rare, and seizure is like very, very rare. All these side effects are triggered by the contrast agent when this is coming into the head, and you can avoid it with the positioning of the patient. For example, to lift the head in the lateral decubitus positioning or to lift the head in the prone position. And then the contrast agent will stop at the level of the cervical spine and will not go into the head.
And I can tell you, when a lot of contrast agent will go into the head, this can be really a bad experience for the patient. So, keep the contrast out of the head.
Before I’m going to show you some examples, I want to start with a classification. We have the type one leak. This is a ventral dural tear. This is probably the most common one, and these patients have epidural fluid on MRI of the spine. We have the lateral dural tear, also fluid positive on MRI of the spine. Then, we have the CSF-venous fistula. This is an abnormal connection between the intrathecal space and paravertebral veins, and this is the only leak type with no epidural fluid on the spine. And then I want to take the chance to mention the sacral dural leaks that we regularly find in Freiburg in approximately 5 to 6% of patients, and these are again fluid positive.
So, starting with the type one leak, the ventral dural tear. The patient is in the prone position, and we put some pillows under the hips because of the reason that the contrast can overcome the kyphosis and can spread along the spine. And we know that these leaks are very high-flow leaks. We found that it takes like 1.6 seconds for the contrast to go through the dural tear, so very fast. And we perform one frame per second or more with a breath-hold technique.
This is what it looks like. The contrast is running into through the intrathecal space, and exactly here’s the leak. And it is running very fast in the epidural space, which means the leak is quite big at the level of T12 – L1. And we mostly have an underlying bony spur like in this patient. Sometimes we need to do a higher frame rate in indeterminate cases. Like in this one, for example, we see the contrast agent at the first time connected between two levels, and we do not know where it originated from. And then we performed the same scan again with four frames per second, and now it is clear that it comes from the lower level.
Then we have a problem in a critical area of the upper spine because we have a lot of superimpositions here with the muscles and bones, and we cannot even see the contrast agent in this area, especially between C6 and T3. Therefore, you can perform a dynamic CT myelography with more radiation power but also more radiation dose exposure. We try to refine this technique a little bit.
Here again, we have a lot of superimpositions, but when you put the patient in a swimmer’s position, lift one shoulder, you have a free projection into the spinal channel. And we do this with a pumping wedge, and this technique works quite fine. This is an example where we could demonstrate the leak at the upper side of the spine. But this, of course, doesn’t work in every patient. Here we missed the leak at the DSM, and we performed a cone beam CT, which revealed the site of the leak.
For type two leaks, the patient should be positioned in a lateral decubitus position. These leaks can be high to flow, low to flow leaks. And we suggest doing an imaging run between 60 and 90 seconds, and nobody can hold their breath for such a long time, of course. So, the patient has to go on breathing, and this is now really important. It’s important to talk to the patient and explain. I always say, “Now we will start the examination. Please go on breathing, but shallowly, and try to breathe evenly and not to move anymore.” Then you may have a chance to get very good or high-quality images to make the diagnosis.
In this example, you can see an arachnoid herniation through a lateral dural tear that is filling, and the leak is this epidural contrast in the epidural space.
Sometimes, it’s not possible to demonstrate the leak, like in this patient. We missed it with a DSM after 60 seconds. We performed the cone beam CT and we saw a very subtle leak at this site. The type three leaks you can perform in exactly the same manner as with type two leaks. They mostly appear very early, like in the first minute, and so we suggest doing a 60 to 90 second run at one frame per image, and again, go on breathing.
This is an example of a CSF-venous fistula. We see a meningeal diverticulum, which is filling with contrast. We see the contrast agent running into the veins, and we have a typical wash-out phenomenon. We found that cross-sectional imaging is probably more sensitive to detect a CSF-venous fistula, and we demonstrated this with CT myelography. There was another research paper recently published by Dr. Madhavan, who confirmed this with a cone beam CT that followed the DSM myelography, and they demonstrated a higher sensitivity.
We in Freiburg sometimes do a special cone beam CT with very high spatial resolution of 0.14 millimeters. This is kind of a challenging technique because the cone beam CT takes 14 seconds, and the patient has to hold their breath for 14 seconds, and it is also very sensitive to movement artifacts. But you get high-quality images with a lot of anatomical details, and we are sometimes able to see very small CSF-venous fistulas. Then you can do an examination with a DSM to find a sacral dural tear. The patient is in the lateral decubitus position on that side where the leak is most likely to appear, and we do it in the reverse Trendelenburg position to get the contrast agent to the lower spine. We perform a 60 to 90 – second run, and the patient should go on breathing. And we see here fluid. But these leaks, in my experience, are sometimes very slow-flow leaks, so an alternative would be to do a CT myelography.
So, my take-home messages are: DSM is a very nice imaging tool, especially because it gives us a live observation of the leak and gives us therefore more confidential to make the diagnosis. But the DSM technique is challenging. And I think the most crucial aspects are the planning, positioning, and also communication with the patient to make good images. I personally think it particularly suits the type one and type two leaks, and it is important that the patient holds their breath with the examination for the type one leak and goes on breathing for the type two and three leaks. The cone beam CT can be a very valuable tool to make the diagnosis.
Thanks to the team again, and especially to Professor Beck, who did a great job to lead our team in Freiburg. Thank you very much.