 I think I have 10 minutes maybe to finish my one hour talk, but it's okay. There's a lot of things that were kind of revealed already, so we can do real quick. Okay, just basically this is showing you the difference between we call the 3D. It's kind of a uniform and the MRT what we can do is you can have a different intensity, that's what you see on the film. Can you see it? Hope you can see it. This is where you see on the film each slice, each grade is like one centimeter by one centimeter. This is the old nobles pick up treatment planning you see in this way, this MRT. So MRT basically we go through all this from the conventional. Dr. Yuko has talked about it we use. In the old days we do this conformal already. And then now a day actually we still using compensator. Those in our institution on NINEC especially on the proton therapy. When we have the proton treatment therapy we still using compensator very often. Actually for each patient, all the patients using compensator. Okay, so that's still kind of MRT because you basically adjust the intensity inside the field. You have a different level of the fields. Okay, and then but it's easier to do the MRT using the MLCs and these MLCs to be computerized MLC and the MLC is a different way. I just showing you the pickup or the tomo therapy. That's basically the primary concept and then you can do cyber knife. That's much, much complicated. Okay, we're not going to talk about it. We're also doing use this using a scanning bin. Okay, that's where you're using the scanning bin for proton to the IMPT. Okay, intensive modulation proton treatment. So nowadays we can adjust the bin. Unless you do MPT, you won't be able to get those distribution that they want to. Okay, show you, hopefully I can show you something, see if it works. I want to show you some of this different way to deliver this MRT by using stab and shoot sliding window and the VMAT. Okay, so whatever the last talk you have is something like this. Okay, basic stab and shoot. It's just like he shows like we call the poor man like a four, three segments. Usually we have a much, much more segment to be used for stab and shoot. This is like a, we have like a 14 segment. Okay, starting with this, see this small opening. And this kind of leaves goes anywhere he wants to, okay? And to deliver the dose, okay? That's called a stab and shoot. And this called a sliding window, which means the leaf has to go only one direction. Not stab and shoot, they can go anywhere they want to. Okay, this called a sliding window. Watch out these leaves go only this direction, okay? See the radiation goes only one direction. This is more look like, but that you could just show you the way to generate this leaf sequencing. We call the leaf sequencing to provide this kind of, we call the sliding window delivery, okay? This is the way you can deliver. All this code, now we can even use the VMAT. VMAT meaning during the leaf motion, you can move the carometer or you can move the gantry at the same time. This is just try to, it's very hard to simulate. When you move the gantry and move the simulator, and move the carometer at the same time. See this is opening at a certain gantry angle, then you're just doing this. But supposedly if this is what you want to deliver, you can move the gantry and the carometer at the same time. Where the leaf is motion, okay? This is called the VMAT, or a different name, I'm going to talk about, okay? While we need to do MRT, the motivation I think last slide with, last talk we discussed. What's the main reason from 3D, 2D to 3D, from 3D to MRT? It's basically, you can see the dose distribution from the 3D, and this is the MRT, okay? You can see the dose to the product is much, much lower. If you use the 3D on this lateral side, you end up half spot. But then you get a full dose to your product, okay? Another thing is, watch this dose curve. You can curve even like a concave shape, okay? You cannot achieve this by using 3D unless you use multiple, multiple pins. Just like we did exercise last Friday, you have to have more pins. And you have to do another pullman modulation, okay? So in order to protect these critical structures, then you end up using MRT. Otherwise you cannot get those higher. You get the complications, is that right? The last talk, okay? Basically, what you have is 3D. You end up having this uniform dose basically inside here. But when you do the intensity modulation, you have different shape at a different location. See, in order to avoid your this either spinal cord or your rectum, if you're treating the prostate, you can make this intensity level lower here, see? Intensity level lower here, intensity level lower here. So you can avoid those to the critical structure. At the same time, you get this uniform dose to your target, okay? Real quick, I don't really have time to really, really go. Okay, so that's the summary, okay? For the benefit of MRT, it's over here. Okay, you can read, I don't really have time. Okay, then basically, we do the forward planning. This is what we're gonna do. Basically, we decide being, come from where, and then we calculate the dose. You can move this different angle position. You can give different weighting, try to get a dose. But no matter why, you end up get a uniform dose. The goal for the 3D is get a uniform dose inside whatever your being aperture, okay? But for the inverse planning, this is where we instead of give description from here, we give description from here. Meaning how much dose to target, and how much dose, the maximum dose to the critical structure. You give the description over here, and then you decide what is your feel looks like. So that's called the inverse planning, okay? Real quick, I don't really have time. Okay, so you have a different way to do this, we call the inverse planning, okay? The basic way to do the inverse planning, I think we've discussed the last talk. We use this, we call it like a pencil being. Or we call it a bin leg, or we call it a pencil being. I hope I can, I think Dr. Yookoff has discussed last talk already. I probably spend a little bit more time talking about the DAO, the direct aperture optimization. Okay, so this is called the bin leg. I think we have to go through this already. Basically for each small field, they calculate those goes through this. Okay, for each this small grade, those grade either one centimeter by one centimeter. If your MLO says one centimeter width, but most likely right now we have a five millimeter by five millimeter. So each five millimeter you define your target, then you can define the shape. I think you have this last talk of it. And then you calculate this dose and you calculate, remember the last talk, all these DI different weighting things. And then you end up determining for each dose grade what's the best weighting. Relative weight, so that you can create the dose, okay? Where you see this, then you end up, you create it from this dose grade. You end up to see what is your leaf openings, okay? And then you need this, we call it based on the intensity, this just give you those grade. But actually this intensity could be like a, like a curve, like a continuous curve, correct? This will call the open density matrix or whatever it is, the fullerance. This is the fullerance, it's like a continuous. Then you have to generate this. We call the leaf sequencing last talk, we show you. So I write a kind of sliding window, the last talk. You can leave the leaf where you open. This is called a two step, okay? First you generate ideal for this map that you can deliver those. But then you need to generate this, what we call the deliverable. Deliverable, MLLC, this is called the leaf sequencing. You need a different program, there's a different way to generate. But the problem is once you generate this deliverable MLLC, what happened? You calculate the back, the dose. The dose will be different from whatever you have from the ideal fullerance, correct? That's the problem, okay? So you generate this very small field, this relative is very small, okay? So the final dose is calculated from all this small segment or some people call the control point. Then you'll end up when you see from this ideal fullerance, it's not what you get from your small dose calculation. You lose it. So basically you need to go back and forth to change your prescription. Again, in order to cheat the treatment plan system, cheat the algorithm so that you can get what you want. You're signature it, because once you do this conversion or do this leaf sequencing, you have a problem. I'll show you all these at the beginning when you do these two steps. Do I have, when you're doing these two steps see all these fields, very, very small, relatively very small, see it? These very small openings, this is where the field. When we first start doing this MRT, our physician don't want to try it. Because they say, in the old days, 3D conformal therapy. We always ask our physician, say the smallest field you can have is four by four or three by three. Right now, he say, all the field give to me is less than three by three. How can I treat? You always told me that if you feel smaller than three by three, we have a certainty that we don't know what's the dose. Now all the fields, four by three by three. So at the beginning, they don't want to try. Okay, you need to make sure your dosage is correct. But anyway, watch out, this is starting. This is where we first started. This is called a bin leg. This is called a modulation, see it's like using the sliding window concept, but it's step and shoot. So we end up have like 83, 84 control point. Meaning, for this particular field, we have 42 small fields. We generate this from ideal Florence map. Remember the last talk, we said every 10% dose difference, you divide the 10% so you create a sliding window. But 10%, there's a big difference, you know what I'm saying? So if you want to get more accurate, then you end up using 5% or 4%, but if you reduce the percentage, meaning you need to end up more segments. But the more segments, you end up more smaller fields. You have a jeopardy, the same thing, you don't know what to do. So all these smaller fields can generally to get these final dose, see that's the problem. So we talk about you have this leakage, you don't know this dose is uncertain, is that right? You have this different leakage, you have this how much opening, there's some gap. There's a lot of what we call a mechanical limitation or a mechanical uncertainty to cause you dose, distribution, whatever you see, it's not what you get. So this is the planning process. So you have to go back. If the dose is not good, if the dose is not good, you'll go back to change your bingo or change your prescription. Hopefully after you do this organization, after you do this live sequencing, after you do this final dose calculation, this meet your goal, meet your prescription. If not, you'll go back. If you're not, you'll go back. This is the we call the two step. You end up as a problem. So this is just the history. So instead of doing that, we do this called direct aperture optimization. What happens is once you're generally starting MLC that you can deliver, this is called deliverable segments, then you can somehow, during this process, you can optimize your leaf shape and then your leaf each bean segment where the leaf will be and each segment, what the bean weighting should be, but at the same time, you can calculate the dose. So you are not optimizing the influence. You are optimizing the bean weight and the bean shape. See what I'm saying? This is called a direct aperture. You optimize the aperture instead of you optimize the influence. For instance, somewhere you cannot deliver. But this is something, you waste your machine parameter or you use your maybe compensators. Okay, compensator could be like a one centimeter compensator or could be two centimeter compensator, whatever the size, but that's what you want to deliver. This is deliverable openings, deliverable aperture that you do the optimization. So at the end, what do you have? What do you see is what do you have? Okay. So this is called a direct aperture optimization DAO. I don't know if I have time. I just give you the concept. You can go back to read, okay? The most familiar system I have is called a pinnacle and this is using, we call it DMPO. That's why they call it direct machine parameter optimization, meaning they optimize this direct deliverable MLC instead of just the influence. Because influence cannot deliver. I don't really want to use your, just give you some concept, okay? The reason we can do that is you first define how many segments you want for each pinnacle, okay? So let the system decide what's the best. So begin with this, maybe simplest way, okay? Just begin, maybe if you just say for this certain pinnacle, I only want a three aperture, only three aperture, just like a poor MMRT. But then you let the computer to do it for you. So the system first define this pinnet, okay? And then assume this three aperture has the same identical opening. Three, and then you do the optimization. You move your leaf, okay? Maybe say move in or out. So you can move this particular ending, you choose randomly any particular leaf, okay? You move in or out, meaning you optimize this location of the leaf. And at the same time, you optimize your pin weight and then you determine after they moved, you calculate those. You determine is this move good or bad, okay? If this move bad, meaning you reject. If it is good, you keep, okay? Depends on whether you have this, you meet this MMRT constraint, meaning how fast it goes, okay? Or can you hit each other? And then depending also the cost function, okay? Meaning you make the cost function, you make the penalty lower, you make cost function lower, then you keep continue. If you make the cost function higher, meaning this move is not good, so we reject. Okay, just keep trying, okay? Also talking about an Indian rule. An Indian, maybe we just passed it. So this just give you maybe some constraint. You need to meet this constraint in order to give. This is like an ELECTR, they say you don't wanna hit, that each leaf don't wanna within one centimeter. Even the next leaf, don't wanna, why we don't do that? Because we don't wanna cause this leaf collision, okay? Sometimes you have this calibration error, leaf collapse, and then you have like a car head and head collision, then the kilometer will break down, then you cannot change the patient, okay? So this is a safety. And then you end up several iterations, you end up these three segments, see what I'm saying? And then this is the DAO, I don't really have that, okay? The take-home message is, you don't really need too many aperture, like what I show you, the sliding window, you only need a few openings, few aperture, because this just give you a sample. If it's only like a three aperture per angle, just like I show you, you can end up have a seven different way to deliver, different weight. So this based on this formula, okay? So if you have a three aperture, you have a seven intensity to modulate. If like a four aperture, you have a 15 intensity to modulate, okay? So it's very easy, so this is just binary, okay? So you don't really need more than 10 aperture for each beam angle. If you have more than that, meaning, so this 63 million is how many those difference? 2%, is that right? You don't really need more than that for each beam angle, okay? So if you use 10, it's enough, usually, okay? Unless you end up with a complicated plan. Let's just take a home message. So use this, you don't really end up, you really end up smaller aperture with relatively large field and with better dose distribution. I don't have time. This is the DMPO from the system. The way you can do MLT, see this is the optimization. The way you do MLT, you can DMPO, you can do intensity modulation, that's what I said, you do two step, okay? Once you do this DMPO, this DAO, you don't want to do intensity modulation. You can also do segment weight or this stuff. So basically, the operator gives the system how many total segments for this particular patient. If you have like nine angles, this is like 15 aperture. This is what's the maximum field, what's the maximum number of, minimum number of the MU to do it, okay? This is like operator define the field size. All these parameters we can talk about. Okay, so this is to give you the end result. For this particular being angle, see this relatively smaller number of the segments, only like 13 for this one. This is a complicated one. But for this one, you end up, I'm showing you this, you have a too many segment. How can you tell? Because see this MU, what's the number here? Can you see? Only two, is that right? Why this goes to two? Because when you define at the beginning, you say for each segment has two more than two MU. So this Chenmen plan system, why is two reduced, reduced, reduced? When they do two, they cannot go in down. Actually you're limited organization. They stop at the two. Do you understand what I'm saying? So that's indication if a lot of these two MUs, meaning you end up using, you're asking too many opening, too many segments. This kind of waste, see? Okay. So really quick, we evaluate what's the, whether this is good system. Anyway, this is to give you a summary. Okay, compare with whatever we have before, using two step, four and, and then leave sequencing, and then you compare with the DNPO. Our experience is the plan quality will be much better. If you compare the total cost of function, the penalty, we reduce almost 50%. Much better. And the treatment delivery is much better because total treatment, modern units, we realize it can reduce more than 40%. And the number of segments, we can reduce more than 50%. Okay. What means by reduce total modern units? Meaning your treatment is more efficient. You use less MU, you achieve the same goal. So you give the more, less being on time. So you feed the machine longer life. Correct? What happened to the segments? You have less segment, meaning you have a more, longer life for your MLC. Because MLC controlled by the motor, you cannot save your motor life. Okay? So this can, and use the less segment, meaning you deliver the treatment with less treatment time, more efficient. Okay? I don't think I want to talk about this. So basically then we go to the V-MAT. So I don't think I have time to talk about V-MAT. I think we just finish here because we don't want to over your break time. But I'd be here if you wanna, you have questions, I'd be happy to answer you. But then you're free to have the break, okay? Sorry. Thank you.