Thank you very much. It's an honor to be here. So I'm gonna talk a little bit about the 6D setup and monitoring for spine SRS treatment. Here are my disclosures. And I'm gonna start off by a few illustrative cases. And so this is a gentleman who had melanoma at an oligometastatic disease at T11, with epidural extension of disease displacing the cord laterally. And we treated them on a protocol with spinal radiosurgery alone. He was deemed inoperable. And at my institution, our preferred approach is to treat with a simultaneous integrated boost technique, but also to dose escalate when we consider the histology to be radioresistant.
And so here's an example of that, where we treated the gross disease to 24 Gy. A clinical target volume per consensus guidelines to 16 Gy. And we use a nine-field step and shoot IMRT technique and ExacTrac as our primary image guidance tool for alignment. And so he received single-fraction radiosurgery and he had an excellent response to treatment. You can see complete decompression of the canal and just some scar tissue in the region.
But it doesn't always work out that well. Here's a case where I treated a patient with a large metastatic chordoma in the lumbar spine. And we treated that again with a similar approach with 24 Gy in a single fraction to the metastasis 16 Gy to a clinical target volume. And initially, it responded quite well. And so here's a one-year comparison where the disease has nearly completely resolved at that level. It's always important to follow your patients though. And about two and a half years later, we noted that the disease had recurred in the epidural space. And I bring this example up to illustrate how important it is to optimize a dose, especially when diseases near the epidural space.
And then finally here's a case of a patient with metastatic renal cell carcinoma and disease at C1 that we can see here involving the arch. He had some pain. We commonly, you know, treat these patients with a stabilization procedure prior to radiation, if he was deemed at risk. This was a borderline case, and so we went forward with spinal radiosurgery alone. Again, 24 Gy to the GTV, 16 Gy to the CTV.
At my institution. We also don't use a PTV or PRV around the cord. Here's an example of the DVH one with our SIB approach, but have a separation between GTV and CTV dosing and the normal structures over here to the left. He had a complete response to treatment. His pain completely resolved. This was durable. This is a two-year image. But interestingly, I bring this up because we always get concerned about the risk of biologic dose, escalation, 24 Gy causing fractures and destruction of bone. Here's a case where 3 months after the treatment, the C1 arch, which was missing due to the metastasis reconstituted. So he self-stabilized following radiosurgery, and he never needed a occipital cervical fusion.
As Josh mentioned, patient selection is very important. When we consider using spinal SBRT it can be a very powerful tool, but it comes with risk. And so again, I highlight the NCCN guidelines. This is the 2018 version. The 2019 version is very similar about this illustrates the accepted criteria that can be considered.
So we treat patients with oligometastatic disease or if they've had recurrence after prior radiation, or if they have radioresistant disease to the spine with an indication to treat. Now, alignment's extremely important when we do spinal SBRT. The plan can largely be driven by normal tissue constraints, namely the spinal cord. And so there tends to be a sharp penumbra in that area.
And we did this study about 10 years ago where we took spinal SBRT cases and we introduced artificial shifts offline. To illustrate that if we are off by 2 millimeters, the dose to the spinal cord can be elevated by up to 20%. So it's very important to have optimal immobilization and image guidance to deliver this type of treatment safely.
On top of that, we've illustrated that optimizing dose to the GTV is very important. So we can't be overly conservative with our cord constraints for it can lead to a local failure. And so we did an analysis of our experience looking at over 200 patients, identifying those that suffered a marginal failure and comparing them to a control cohort. And what we demonstrated from this study is that if we achieve a GTV minimum dose of 33.4 Gy in a BED 10 values that we can optimize local control. And so we've translated that to practical recommendations,14 Gy in a single fraction, if we're doing single-fraction spinal SBRT or 21 Gy over 3 fractions. We use that as a goal for the minimum dose delivered to the GTV. And Josh's group has also illustrated similar results prior to this.
And so what do we contour? At my institution, we do the GTV based on MR imaging and then the CTV per consensus guidelines. And in our practice, we don't use a PTV nor a PRV, but this isn't a standard approach. Other institutions certainly may use a PTV, PRV, and single target prescription with excellent results. But as I've mentioned before, patient immobilization is critical. And so using a setup that is designed for stereotactic purposes is very important. This is the type of setup that we use involving a long stereotactic cradle and the body-fixed system.
And in our early experience, we delivered spinal SBRT with CT-on-rails. Dr. Eric Chang started the program back in late 2002. And so this was the image guidance that we had available at that time. And we demonstrated using that tool that we could deliver the spinal SBRT treatment with sub-millimeter accuracy as judged by a CT obtained immediately after the treatment and the treatment positioning. The problem is that each fraction took about an hour and a half to two hours. So it was a very challenging workflow. And this is our CT-on-rail system that we used to deliver spinal SBRT early. Now, as you can see, we introduced ExacTrac to the system later on, and that was to help with our IGRT workflow.
So how has the ExacTrac system then evaluated for IGRT and spinal SBRT? This is a report out of Duke University in which they looked at both phantoms and patients and evaluated ExacTrac treatments and used an offline 6D-6D fusion as the gold standard in their evaluation. And so here's the data from the phantom study. And as we can see, the alignment was quite good when compared to the CT, when we use ExacTrac. So sub-millimeter accuracy in the translational dimensions. And then again, less than one-millimeter rotational error. So that looked good.
The patient data didn't look quite as good but still reasonable. And so in a root mean squared analysis, the translational error was less than 2 millimeters and the rotational error at about 1 degree. But the range in the rotational error went from anywhere from -3 degrees to +2 degrees. And so the data looked good, but there was some modest error noted. So we performed a prospective clinical trial at MD Anderson in which we evaluated...one component of the trial was to evaluate various ExacTrac fusion algorithms.
So one potential approach is to do bony fusion, another one at that time to do a fiducial-based fusion approach. And so we actually enrolled patients prospectively, implanted gold fiducials in the level above and below the intended target and treated the patient, and then evaluated the data with an offline 6D-6D fusion after the fact to assess for residual error in both translational and rotational dimensions. And what we found is that there was no difference between the bony fusion approach and the fiducial-based approach. They agreed very well.
So in the study, we demonstrated there was no need to implant fiducials in order to do spinal SBRT accurately using ExacTrac. And then our evaluation of the bony fusion algorithm. And again, this was ExacTrac 5.5, which was an older version. But in our evaluation of that, the true rotational error was better than what was demonstrated previously, but could still exceed 1 degree in about 20% of measurements. And so based on that, I know here... And here's the data from the study itself. So based on that, this is our general workflow and image guidance approach for spinal SBRT.
And so we first set the patient up and in the treatment room with skin-based alignment, and then we get an ExacTrac with correction and verification. But after we do that, we also still get a cone-beam CT to ensure that there isn't any residual rotational error. And then on top of that, we get orthogonal imaging to confirm that we're at the intended level. And then right prior to the start of the radiation treatment itself, right before beam on, we get one more ExacTrac, and then we continue with the treatment.
And then again, we typically use a nine-field IMRT approach. And so in between each field, we get a snapshot with one of the detectors. But as we're getting the snapshot, it's not true real-time imaging and so we are getting images between each of the 9 fields, so roughly 10%, but that is one limitation of the ExacTrac system as we have it. And there's also can be limited interpretation of images due to the small field of view. In addition, the X-ray tubes have some heat capacity limitations.
And so we're excited about the future of ExacTrac and the next generation of ExacTrac technology which uses... This is called ExacTrac Dynamic. These slides are from Brainlab. And so this is ExacTrac Dynamic, which has several improvements. One is the use of this integrated thermal camera which can provide real-time imaging while the patient's on the table using surface rendering and thermal data. This isn't yet available in the United States. But this is something that I think may have potential.
And in internal data from Brainlab, they've demonstrated preliminarily that can be highly accurate with sub-millimeter accuracy in their assessment. It does offer a larger field of view. And then, of course, the thermal camera is on the entire time during the treatment, even while the beam is on.
They also have new X-ray tubes, which allow for more frequent imaging and increase heat capacity. So you almost get this fluoro-type image that's possible with the X-ray tubes. And the new flat panel detectors can offer improved image quality. There's an enhanced readout speed. And so the quality of the images obtained is likely to be improved.
There's dynamic patient monitoring. And so the system is more integrated with the linear accelerator. And so users can select either surface triggered X-rays, monitoring unit triggered X-rays, or gantry angle triggered X-rays and the system can flow more naturally while the patient's on the table.
And so here's just a brief comparison of the two. So we have the system on the left, which has ExacTrac, which certainly is an efficient and accurate system. We've used it for over eight years in our spinal SBRT cases. We still do get CT verification to confirm that there is no rotational error and our intra-fraction monitoring is done with these snapshots between gantry angles, as we're doing step and shoot IMRT.
The next generation has the potential of improving the IGRT workflow. This all has to be verified, but it's likely to be just as efficient, at least as efficient and accurate. There'll be a question of whether we need to get CT verification afterwards as the imaging might render that needless, and then the intra-fraction monitoring might be an area where we'll be able to have an advance here with the 4D thermal camera and Dynamic X-ray monitoring.
So in summary, at MD Anderson, we've treated hundreds of SSRS patients with ExacTrac-based image guidance within the setting of single fraction treatment and no PRV or PTV use. And this sub-millimeter accuracy allows for optimal GTV dosing which is very important for minimizing likelihood of local failure. And at the same time, it allows for minimal risk to the critical neural structures. And next generation technology has the potential for allowing real-time motion monitoring further reducing the risk of toxicity. So thank you for your time.
