Transcript
Well, it's always a pleasure for me to be here. Sometimes I feel like I wanna, you know, get a little teary-eyed because this is a time where I get to see many, many old friends that I've met over the last how many years? 11 years of Novalis involvement. So it's always a lot of fun to be here and I'm gonna present something a little bit different this time, but, and it's also a little intimidating to come after Dr. Kilpatrick.
So I know he's gonna have questions and I don't know how I'll answer them, but in any event, I guess you guys are aware of this game, where's Tim? So now Tim's at UT Southwestern. So I've had a chance to talk with a lot of you guys, but not everybody so which I apologize for, but I wanted to just tell the whole group why UT Southwestern, so we could just get that out of the way, and then we can renew our friendships and stuff.
So Dallas has better weather than Omaha, although it's starting to get a little warm, but it's not Houston. So far all the tornadoes stay north of Dallas. It's been a terrible year for tornadoes. That by the way, was my tornado shelter in Omaha. Dallas has a lot better shopping. Midwestern people are pretty friendly. You'll notice this is the main shopping center in Omaha, by the way. You'll notice they do give free marital advice at the shopping center. I'm not sure if that's needed in Nebraska, but you never know.
Texas cows are definitely bigger than Nebraska cows and on the slightly more serious note, UT Southwestern has four Nobel laureates in medicine. It's more than any institution in the country. Nebraska has three billionaires including this guy that Paula and I would have ice cream with once a week, famous guy from Nebraska. But seriously, UT Southwestern has a very strong historical support for research, basic research, and now translational and clinical research. And the department in particular has a very strong dedication to innovative clinical radiotherapy with a strong emphasis on clinical trials.
Probably several times during this talk, I'm really gonna emphasize, I think the importance of prospective trials because we can take all the data that we've all accumulated in this room and all the Novalises in the world, but if everything's done differently and we just pulled together random data, all we'll get is a big pool of random data. So we really need to think about prospective trials.
Yep. So what I wanna really talk about is SBRT. We like to call it a blade of radiotherapy. It's not hypofractionated, it's a blade of radiotherapy. And the initial clinical success that I think we're all aware of in lung and some other settings has really led to a rapid development of technology and clinical adoption of this. But many centers still lack a real systematic approach and the doses as you'll see are all over the place.
So I'd really like to see, as we embark on SBRT, I'd like to see us do something different than we did with SRS, and that is take a real systematic approach to it, including, as you heard in the previous talk, studies in the preclinical setting. So I'm gonna show you kind of a hodgepodge of that. So those of you who are engaged with this are used to seeing great results like this. And at Nebraska, our standard dose was 5×10Gy, and this is a patient that's three and a half months after 5×10Gy.
So, you know, we're all really amazed when we start this, that it really works as well as it does. But then you get these patients like this who after the same dose and same period of time, nothing looks like it's changed at all. And you get other patients who resolve later and differently. So here's a patient pre-treatment after three and a half months, nothing, and then after five and a half months, you see some significant anatomical changes. So there's all kinds of variation. Here's a patient where we implanted a marker in a lung tumor and on CT imaging after five months looked exactly the same hadn't shrunk at all, but it was cold on PET.
So we have all these questions that we'd like to get better at addressing preferably before we even treat the patient. That is, when are they gonna respond? How are they gonna respond, and what can we do to modify their response if we see something going on? And I'll talk about tumor response, but, of course, with SBRT, like radiosurgery tumor response isn't the only issue. In fact, its complications that are a very significant thing. So I'd just like to remind the group in the initial Karolinska experience, they treated 11 in this case, liver tumors, and they killed three of them with SBRT. So we need to be careful as we embark on these things that there are problems and not just successes.
So the question is I just pulled out a bunch of these. Most of these are prospective studies, but I pulled out a bunch of things from the literature to look and see what people are doing. So this is for lung, and this is for liver. Some well-known organizations or institutions, RTOG, our center, a couple in Japan and in Germany. So these are their prospective studies and the doses that they use for lung. And you can see they're kind of all over the map. Here's an interesting one, 1×26. And for liver again, all over the map. And look at particularly the RTOG because I'm gonna comment on a couple of times, that's 10 fractions with a dose escalation to 10×5Gy. And I'm gonna show you again, it's just a model and I'm gonna show you that this 10×5Gy is a lot different as you can imagine than this 5×10Gy. So these actually are very potent doses.
So you heard about the linear-quadratic model. I don't have to tell you about it. You all understand it, and you all understand from the previous presentation that the linear-quadratic model overestimates the effect in large dose per fraction. So the linear-quadratic model suggests it keeps curving. And as you saw from that presentation and this publication, it actually doesn't keep curving at least in-vitro. I agree, we need, as I said, in my opening slide, we really need some in-vivo studies so that when we take that BED and put those, equate two different dose fractionation schemes to try to come up with an equivalent dose, when we do that it's not accurate. And if you believe that the linear-quadratic model overestimates cell kill, then we're underestimating the effect, sorry, we're overestimating our BED effect when we use the linear-quadratic model.
So there's another model that came before the linear-quadratic model. It's not very fashionable because these are all just models. They're all just curve fittings, as you heard Dr. Kilpatrick say. And this one really has no basis in science, but it's a multi-target model. And if you look at the multi-target model as you go to large doses it asymptotically approaches that. And so it's a good approximation at least of in-vitro cell kill for large doses.
So my colleagues at UT Southwestern said, ''Well, great, the LQ model works good for low doses. The multi-target model works good for high doses, and let's put them together' in a very simplistic way," if you read the paper that came out about a month ago. And so they came up with this universal survival curve, which is a combination of these two models in attempt to try to make some sense out of SBRT type doses.
And they did some studies at UT Southwestern. These are in-vitro again in prostate cancer cells and the data points are the measured survival. And the two lines are the two different models, and you can see that their model fits the data better in-vitro. And then from that, they took some parameters from the literature and came up with a nomogram for SBRT, for non-small cell lung cancer. So you can look at this nomogram and figure out if you're doing one, three, or five fractions, look at your total SBRT dose and come up with an equivalent dose in 2Gy per fraction. So this is a standard mechanism that they do at UT Southwestern now for looking at doses.
So going back to doses, people have used in the literature in some clinical trials, what does that model then say about these physical doses? So again, here are the liver studies, and I wanna point out as I did before these 2, 5f of 10Gy or 10f of 5Gy. If you plug those into the model appropriately, what you find is 5 fractions of 10Gy is about a standard effective dose of 83 versus 65. So there's about a 20Gy equivalent difference in the same physical dose.
So, you know, obviously, this isn't the perfect model. It's maybe not even a very good model, but this really highlights our challenge as we embark on SBRT clinically. And just to look at the lung numbers, so here's the physical doses. These are some of the standard doses we use in Texas, and these are the standard equivalent doses in 2Gy/f.
Right. So just to... I don't have too much animation or movies in my talks anymore because I was getting really excessive with that. So this is my only movie slide, but it's to emphasize again, that we have an opportunity in the preclinical setting to really study this well in-vivo. So I wanna show you some of the work that's been done, very little in-vivo work with regard to large dose per fraction.
Here's two papers that came from UT Southwestern. This one is for prostate cancer. So prostate tumor cells grown in the flank subcutaneously of nude mice. I'm looking at three different doses, looking at the volume as a function of time after radiation. And I just put up here for reference the standard equivalent dose. Obviously, you get more radiation, the tumors respond better. That's not a big surprise. And they looked at PSA as well and saw a better PSA control.
So that's great. All it does is show a dose effect and nothing more. Here's another one that they did for renal cell carcinoma. Again, just at one dose and it was effective in tumor control in these subcutaneous models. Kind of in parallel at the University of Nebraska, well, while we were still there, we looked at developing some subcutaneous models.
The first one we looked at was this colon adenocarcinoma model. So you see the dose distribution. We had two questions at that time. We looked at three doses that were equivalent using the linear-quadratic model, and we wanted to see if they were equivalent. So those dose schemes are equivalent, but they're not equivalent if you use Bob Timmerman's new standard equivalent dose.
We also wanted to look and see if there was a dose rate effect at clinically significant dose rates. So we repeated this at the high and low-dose rate on the Novalis. This is just a show. This is a small nude mouse scan on a conventional CT. There you can see the tumor, we put a bolus over him. I'll show you this in a minute so that we can have full dose to the target. And even though it's not the greatest resolution, just remember this mouse is like this big. And then planned it using eye plan and BrainScan and treated on the Novalis. Here, you see the bolus, it goes over the tumor, so we can give the full dose to the animal. There's the dose distribution again.
And so what we saw, I've got two dose levels here. We actually saw that they weren't equivalent in terms of tumor response, but the 5×7Gy gave a better response than the 1×20Gy as would have been predicted by the universal model, but not by the linear-quadratic model. So that was an interesting observation. We weren't actually looking for that at the time. And you can see the doubling times here so much, almost a twofold difference in the doubling times. Interestingly, no difference in the dose rate effect.
We also looked at the VEGF Expression. We only have one time point for VEGF Expression, and this is mouse VEGF Expression. So this is coming from the mouse stroma. And what we saw, so at 21 days afterwards, what we saw is exactly the opposite of what we would have expected, and that is the good responders in terms of a long doubling time had the highest levels of VEGF Expression. So that's a complete opposite of what we would've expected and the poor responders had lowest levels of VEGF expression.
And so we postulate that in response to radiation, the tumor changes starts to react, the stroma of the animals starts to respond. And so those tumors that feel a big response to radiation early on. So this is the tumor growth. So this would be a rapid growth. Early on, the stroma reacts, expresses VEGF to try to counteract that... One minute. Perfect.
And so what we're actually measuring is 21 days later, the VEGF has done its job. And so the tumor is being killed and we have low levels down there as opposed to tumors that are undergoing slow growth. And I wanna point out that this is again, stromal VEGF. So it's coming from the mouse and not from the tumor in this case. We looked at interaction of small molecule inhibitors, in this case, Sorafenib. It's approved for use in advanced renal cell carcinoma, I think.
So it acts on the VEGF receptor and the platelet-derived growth factor receptor, and also the RAF kinase pathway. So, we like it because it has many mechanisms for interaction. And this is very preliminary data, but as you can see drug alone, radiation alone, and then a profound effect when the two are given together.
So all of these interests that led me to UT Southwestern has led us to put together a program, Project Grant and tumor and normal tissue response to large dose per fraction. As part of that program, Project Grant, we were building a successive, a number of preclinical, image-guided SBRT machines. So this is the first generation, which was just finished a couple of weeks ago. So the idea is, and again, this is a mouse, so he's only that big that you can see his lung.
The idea in this study is to target normal lung with a small beam of radiation high dose. So this is pre, this is our localization image, and then he's moved so the field center is denoted here. So he's moved into the correct position, and then I've got a double exposure. This happens to be a five-millimeter collimator. And we actually determined that we needed small collimators. So we have a three-millimeter collimator in the small animal radiator, very nice sharp profile. We can give 100Gy to the animals in less than 8 minutes.
As part of the program Project Grant, we're not looking at just subcutaneous tumors, but orthotopic tumors, which can be a real challenge. So here's an orthotopic lung model. And I can let my colleagues tell you how this has developed, and we firmly believe we'll be able to target these, use BrainsScan to treat them.
So just to finish up in conclusion, there's a rapid adoption of SBRT. We don't really know what these doses mean that we're using. So we should take advantage of technologies at present like molecular imaging in the preclinical setting to try to really understand that in a robust way. And I think that those of us who embark on these new technologies and adoption of these new clinical technologies should really be encouraged to do it in the scope of prospective studies. And this was Mexico City. This was two Novalises ago, and I'm happy the whole group is back together again. So who knows where the next Novalis Circle will be, but we'll have more pictures as the days go on. Thanks very much.
So I know he's gonna have questions and I don't know how I'll answer them, but in any event, I guess you guys are aware of this game, where's Tim? So now Tim's at UT Southwestern. So I've had a chance to talk with a lot of you guys, but not everybody so which I apologize for, but I wanted to just tell the whole group why UT Southwestern, so we could just get that out of the way, and then we can renew our friendships and stuff.
So Dallas has better weather than Omaha, although it's starting to get a little warm, but it's not Houston. So far all the tornadoes stay north of Dallas. It's been a terrible year for tornadoes. That by the way, was my tornado shelter in Omaha. Dallas has a lot better shopping. Midwestern people are pretty friendly. You'll notice this is the main shopping center in Omaha, by the way. You'll notice they do give free marital advice at the shopping center. I'm not sure if that's needed in Nebraska, but you never know.
Texas cows are definitely bigger than Nebraska cows and on the slightly more serious note, UT Southwestern has four Nobel laureates in medicine. It's more than any institution in the country. Nebraska has three billionaires including this guy that Paula and I would have ice cream with once a week, famous guy from Nebraska. But seriously, UT Southwestern has a very strong historical support for research, basic research, and now translational and clinical research. And the department in particular has a very strong dedication to innovative clinical radiotherapy with a strong emphasis on clinical trials.
Probably several times during this talk, I'm really gonna emphasize, I think the importance of prospective trials because we can take all the data that we've all accumulated in this room and all the Novalises in the world, but if everything's done differently and we just pulled together random data, all we'll get is a big pool of random data. So we really need to think about prospective trials.
Yep. So what I wanna really talk about is SBRT. We like to call it a blade of radiotherapy. It's not hypofractionated, it's a blade of radiotherapy. And the initial clinical success that I think we're all aware of in lung and some other settings has really led to a rapid development of technology and clinical adoption of this. But many centers still lack a real systematic approach and the doses as you'll see are all over the place.
So I'd really like to see, as we embark on SBRT, I'd like to see us do something different than we did with SRS, and that is take a real systematic approach to it, including, as you heard in the previous talk, studies in the preclinical setting. So I'm gonna show you kind of a hodgepodge of that. So those of you who are engaged with this are used to seeing great results like this. And at Nebraska, our standard dose was 5×10Gy, and this is a patient that's three and a half months after 5×10Gy.
So, you know, we're all really amazed when we start this, that it really works as well as it does. But then you get these patients like this who after the same dose and same period of time, nothing looks like it's changed at all. And you get other patients who resolve later and differently. So here's a patient pre-treatment after three and a half months, nothing, and then after five and a half months, you see some significant anatomical changes. So there's all kinds of variation. Here's a patient where we implanted a marker in a lung tumor and on CT imaging after five months looked exactly the same hadn't shrunk at all, but it was cold on PET.
So we have all these questions that we'd like to get better at addressing preferably before we even treat the patient. That is, when are they gonna respond? How are they gonna respond, and what can we do to modify their response if we see something going on? And I'll talk about tumor response, but, of course, with SBRT, like radiosurgery tumor response isn't the only issue. In fact, its complications that are a very significant thing. So I'd just like to remind the group in the initial Karolinska experience, they treated 11 in this case, liver tumors, and they killed three of them with SBRT. So we need to be careful as we embark on these things that there are problems and not just successes.
So the question is I just pulled out a bunch of these. Most of these are prospective studies, but I pulled out a bunch of things from the literature to look and see what people are doing. So this is for lung, and this is for liver. Some well-known organizations or institutions, RTOG, our center, a couple in Japan and in Germany. So these are their prospective studies and the doses that they use for lung. And you can see they're kind of all over the map. Here's an interesting one, 1×26. And for liver again, all over the map. And look at particularly the RTOG because I'm gonna comment on a couple of times, that's 10 fractions with a dose escalation to 10×5Gy. And I'm gonna show you again, it's just a model and I'm gonna show you that this 10×5Gy is a lot different as you can imagine than this 5×10Gy. So these actually are very potent doses.
So you heard about the linear-quadratic model. I don't have to tell you about it. You all understand it, and you all understand from the previous presentation that the linear-quadratic model overestimates the effect in large dose per fraction. So the linear-quadratic model suggests it keeps curving. And as you saw from that presentation and this publication, it actually doesn't keep curving at least in-vitro. I agree, we need, as I said, in my opening slide, we really need some in-vivo studies so that when we take that BED and put those, equate two different dose fractionation schemes to try to come up with an equivalent dose, when we do that it's not accurate. And if you believe that the linear-quadratic model overestimates cell kill, then we're underestimating the effect, sorry, we're overestimating our BED effect when we use the linear-quadratic model.
So there's another model that came before the linear-quadratic model. It's not very fashionable because these are all just models. They're all just curve fittings, as you heard Dr. Kilpatrick say. And this one really has no basis in science, but it's a multi-target model. And if you look at the multi-target model as you go to large doses it asymptotically approaches that. And so it's a good approximation at least of in-vitro cell kill for large doses.
So my colleagues at UT Southwestern said, ''Well, great, the LQ model works good for low doses. The multi-target model works good for high doses, and let's put them together' in a very simplistic way," if you read the paper that came out about a month ago. And so they came up with this universal survival curve, which is a combination of these two models in attempt to try to make some sense out of SBRT type doses.
And they did some studies at UT Southwestern. These are in-vitro again in prostate cancer cells and the data points are the measured survival. And the two lines are the two different models, and you can see that their model fits the data better in-vitro. And then from that, they took some parameters from the literature and came up with a nomogram for SBRT, for non-small cell lung cancer. So you can look at this nomogram and figure out if you're doing one, three, or five fractions, look at your total SBRT dose and come up with an equivalent dose in 2Gy per fraction. So this is a standard mechanism that they do at UT Southwestern now for looking at doses.
So going back to doses, people have used in the literature in some clinical trials, what does that model then say about these physical doses? So again, here are the liver studies, and I wanna point out as I did before these 2, 5f of 10Gy or 10f of 5Gy. If you plug those into the model appropriately, what you find is 5 fractions of 10Gy is about a standard effective dose of 83 versus 65. So there's about a 20Gy equivalent difference in the same physical dose.
So, you know, obviously, this isn't the perfect model. It's maybe not even a very good model, but this really highlights our challenge as we embark on SBRT clinically. And just to look at the lung numbers, so here's the physical doses. These are some of the standard doses we use in Texas, and these are the standard equivalent doses in 2Gy/f.
Right. So just to... I don't have too much animation or movies in my talks anymore because I was getting really excessive with that. So this is my only movie slide, but it's to emphasize again, that we have an opportunity in the preclinical setting to really study this well in-vivo. So I wanna show you some of the work that's been done, very little in-vivo work with regard to large dose per fraction.
Here's two papers that came from UT Southwestern. This one is for prostate cancer. So prostate tumor cells grown in the flank subcutaneously of nude mice. I'm looking at three different doses, looking at the volume as a function of time after radiation. And I just put up here for reference the standard equivalent dose. Obviously, you get more radiation, the tumors respond better. That's not a big surprise. And they looked at PSA as well and saw a better PSA control.
So that's great. All it does is show a dose effect and nothing more. Here's another one that they did for renal cell carcinoma. Again, just at one dose and it was effective in tumor control in these subcutaneous models. Kind of in parallel at the University of Nebraska, well, while we were still there, we looked at developing some subcutaneous models.
The first one we looked at was this colon adenocarcinoma model. So you see the dose distribution. We had two questions at that time. We looked at three doses that were equivalent using the linear-quadratic model, and we wanted to see if they were equivalent. So those dose schemes are equivalent, but they're not equivalent if you use Bob Timmerman's new standard equivalent dose.
We also wanted to look and see if there was a dose rate effect at clinically significant dose rates. So we repeated this at the high and low-dose rate on the Novalis. This is just a show. This is a small nude mouse scan on a conventional CT. There you can see the tumor, we put a bolus over him. I'll show you this in a minute so that we can have full dose to the target. And even though it's not the greatest resolution, just remember this mouse is like this big. And then planned it using eye plan and BrainScan and treated on the Novalis. Here, you see the bolus, it goes over the tumor, so we can give the full dose to the animal. There's the dose distribution again.
And so what we saw, I've got two dose levels here. We actually saw that they weren't equivalent in terms of tumor response, but the 5×7Gy gave a better response than the 1×20Gy as would have been predicted by the universal model, but not by the linear-quadratic model. So that was an interesting observation. We weren't actually looking for that at the time. And you can see the doubling times here so much, almost a twofold difference in the doubling times. Interestingly, no difference in the dose rate effect.
We also looked at the VEGF Expression. We only have one time point for VEGF Expression, and this is mouse VEGF Expression. So this is coming from the mouse stroma. And what we saw, so at 21 days afterwards, what we saw is exactly the opposite of what we would have expected, and that is the good responders in terms of a long doubling time had the highest levels of VEGF Expression. So that's a complete opposite of what we would've expected and the poor responders had lowest levels of VEGF expression.
And so we postulate that in response to radiation, the tumor changes starts to react, the stroma of the animals starts to respond. And so those tumors that feel a big response to radiation early on. So this is the tumor growth. So this would be a rapid growth. Early on, the stroma reacts, expresses VEGF to try to counteract that... One minute. Perfect.
And so what we're actually measuring is 21 days later, the VEGF has done its job. And so the tumor is being killed and we have low levels down there as opposed to tumors that are undergoing slow growth. And I wanna point out that this is again, stromal VEGF. So it's coming from the mouse and not from the tumor in this case. We looked at interaction of small molecule inhibitors, in this case, Sorafenib. It's approved for use in advanced renal cell carcinoma, I think.
So it acts on the VEGF receptor and the platelet-derived growth factor receptor, and also the RAF kinase pathway. So, we like it because it has many mechanisms for interaction. And this is very preliminary data, but as you can see drug alone, radiation alone, and then a profound effect when the two are given together.
So all of these interests that led me to UT Southwestern has led us to put together a program, Project Grant and tumor and normal tissue response to large dose per fraction. As part of that program, Project Grant, we were building a successive, a number of preclinical, image-guided SBRT machines. So this is the first generation, which was just finished a couple of weeks ago. So the idea is, and again, this is a mouse, so he's only that big that you can see his lung.
The idea in this study is to target normal lung with a small beam of radiation high dose. So this is pre, this is our localization image, and then he's moved so the field center is denoted here. So he's moved into the correct position, and then I've got a double exposure. This happens to be a five-millimeter collimator. And we actually determined that we needed small collimators. So we have a three-millimeter collimator in the small animal radiator, very nice sharp profile. We can give 100Gy to the animals in less than 8 minutes.
As part of the program Project Grant, we're not looking at just subcutaneous tumors, but orthotopic tumors, which can be a real challenge. So here's an orthotopic lung model. And I can let my colleagues tell you how this has developed, and we firmly believe we'll be able to target these, use BrainsScan to treat them.
So just to finish up in conclusion, there's a rapid adoption of SBRT. We don't really know what these doses mean that we're using. So we should take advantage of technologies at present like molecular imaging in the preclinical setting to try to really understand that in a robust way. And I think that those of us who embark on these new technologies and adoption of these new clinical technologies should really be encouraged to do it in the scope of prospective studies. And this was Mexico City. This was two Novalises ago, and I'm happy the whole group is back together again. So who knows where the next Novalis Circle will be, but we'll have more pictures as the days go on. Thanks very much.