Okay, so first of all, John Cleveland, very nice. So now we're only three to one. It's just like last year, so there's a chance. All right. Okay, I have a quick conflict of interest slide. I put together some educational objectives about how we're going to learn everything that you need to know about medical physics in 15 minutes, but really what's going to happen is we're going to try to define some terms and maybe answer some questions that you were too afraid to ask. So this is kind of a mixed audience, which is strange, full of radiation oncology residents that hopefully by now know a lot about radiation and neurosurgery residents who maybe haven't had this in their training at all. So hopefully I'll be able to cover both of you and everybody will learn something and maybe even a little bit about radiosurgery physics. So we're going to start at the beginning. So what's this radiation stuff? So radiation is any process in which energy is emitted or propagated, and when we're talking about electromagnetic radiation, it's an alternating set of electric and magnetic fields. And it can be in the form of particles or electromagnetic radiation, and quantum mechanics tells us that we can sort of treat those two things as equivalent. And so in radiosurgery, this is what we're dealing with, and we usually have gamma rays, X-rays, and much less commonly charged particles. So gamma rays and X-rays, remember, are the same thing, basically. It's just where they're coming from that differs. So gamma rays come from the nucleus of an atom, usually because of radioactive decay. X-rays we'll talk about in a minute, but have to do with energy emitted when electrons decelerate and charged particles are charged particles. So physicists have come up with lots of different ways to describe various aspects of radiation, and you can see a bunch of units here, everything from activity to exposure to absorbed dose. It's a little complicated for a 7 in the morning lecture, so this is a little easier to remember. So if you have a hunk of radioactive material, a becquerel is a unit of activity, so it's how brightly it glows. A gray is how much energy you've absorbed, so you can think of it as how brightly it makes you glow, although you're not really radioactive yourself. And sieverts is kind of the biological effect of that, it's how many extra eyes are you going to have at the end of the day. And so obviously for radiosurgery, absorbed dose in units of gray is what we see most often, and that's the amount of energy that's been absorbed, and it correlates pretty well with biological effect. So one of the first terms you'll hear physicists toss around is a unit called an MEV, and that stands for mega electron volts, and it's just a unit of energy. So what happens is that individual photons have some amount of energy that's represented by this formula on the screen, and so the higher the frequency of the photon or the shorter the wavelength, the higher the energy and the deeper the penetration. And equivalently, particles have an amount of energy that's represented by Einstein's well-known formula. And so you can have terminology such as a 6 MEV photon or a 150 MEV proton, so you'll hear us talk about stuff like that. But strangely, and something that I found a little confusing when I was first learning this stuff, is that when we refer to a machine, we actually throw around units called MV and not MEV. And so one question that you might have been afraid to ask is why that is, and the answer is just that when you have a linear accelerator, it actually emits a spectrum of photon energies in the beam, and so you can't just say it's got a 6 MEV linear accelerator, so they use the term MV instead. And you can see, so you can have a 6 MV LINAC and it can emit some 6 MEV individual photons, but on average, the photon energy is much lower. It's usually by a factor of 3 lower. So where do we get these photons that we're going to use for radiosurgery? So one way is through radioactive decay, and sort of the common example for radiosurgery is cobalt-60, which decays through something called beta-minus decay. And so we create cobalt-60 in a reactor by bombarding cobalt-59 with neutrons until it becomes neutron-heavy, and so it wants to go back and lose a neutron, and so it converts a neutron into a proton, and in the process releases a beta particle, which is an electron, and that gets absorbed right away, and an antineutrino, which doesn't really like to interact, so it floats away through the universe. And we create nickel-60 at an elevated energy state, and then that decays to a ground state and releases a couple gamma rays that you can see at the end, and that's what we use to treat patients. And because it's radioactive decay, the activity is always decreasing, and the activity formula is down at the bottom, so it decreases exponentially. And so mainly what that matters for radiosurgery is that if you're using something like a gamma knife, the dose rate decreases a tiny little bit constantly, and so every once in a while you have to reload the amount of cobalt-60. So the example for this is a gamma knife, so just a quick 30-second overview. A gamma knife is actually conceptually a pretty simple device where you've got radioactive sources that would be here and here around sort of like a styrofoam coffee cup-shaped collimator, and there's some motors in the back of the unit, maybe I'll try this laser pointer, there's some motors in the back of the unit that drive the sources back and forth over different-sized beam channels to give us different-sized beams. Most of the rest of the gamma knife is shielding to protect everybody from the cobalt-60, and then the patient is on a treatment table that moves into the machine and can very, very accurately position the patient within the machine. So pretty simple. An interesting fact is that there's only about 20 grams of cobalt-60 in a unit that weighs 20 metric tons, so it's pretty cool. Okay, so the other way we get photons is via x-rays. So x-rays are generated when charged particles interact with the electric and magnetic fields of a nucleus. You can see that represented here, where you've got some incoming electrons, and depending on how close they get to the nucleus, they decelerate at different amounts and bend, and when they bend, they release x-rays at different frequencies. So you can see that for any given event, we're going to get a different frequency photon, which is why, again, we refer to Linux in units of mv instead of mev. So that is called Bremsstrahlung interactions. I bring this up mainly because it's a big word that physicists like to use. I'm convinced so that we sound smart all the time. Okay, so an example of where we use x-rays is obviously a linear accelerator, so we'll talk about that for 30 seconds, except that they're a little more complicated, so 120 seconds. So this picture is meant to represent kind of the simplest possible linear accelerator. So if you had a sort of evacuated glass tube and you had a little heating coil on one side and a metal plate on the other side, and you put a 6 million volt battery across it, then if you heat up this coil, you'll make some electrons, and because of the voltage potential between these two, you can accelerate the electrons across the tube and hit a window, and when you hit this metal window, they'll make x-rays. A real linear accelerator schematically looks something more complicated like this, where you've got something called a pulse modulator that's sort of like a signal generator, and it sends a signal to an electron gun at the top and a microwave power source that creates an oscillating RF wave, and they both meet up in something called the accelerating waveguide, and that is kind of the heart of the LINAC, where we're going to accelerate electrons to super high speed. Then the accelerated electrons go around the bending magnet to point them down towards the patient, and then into something called the treatment head that we'll talk about in a minute, and then to the patient. So for a second, I just want to talk about the waveguide, because it's kind of the interesting part of a LINAC. So here you see a schematic of what's called a traveling wave accelerator, where you've got RF energy coming in, so this is an oscillating electric field, which you can see here, and it comes out the other end, and at the same time, the electron gun injects some electrons, and they basically surf across the LINAC at higher and higher speeds, so I'm sort of oversimplifying for time, but that's basically the heart of a LINAC. And then the other important part is the treatment head, which, if you blow it up, has an X-ray target, so you have electrons that come in, they hit the X-ray target, they create photons that are shaped into a cone by the primary collimator, they get sort of flattened by a flattening filter, there's an ion chamber that sort of measures how much radiation is going through the machine, so we know when to stop it, and then there's collimators that help shape the beam, and then you get down to the patient. For radiosurgery, there's two basic ways to shape a beam, there's radiosurgery cones and multi-leaf collimators. Radiosurgery cones make a nice, sharp, circular field, and multi-leaf collimators can make much more complicated shapes. A term that you'll hear physicists argue about a lot is penumbra, which you can think of as just how sharp the edge of the field, the edge of the beam is. Physicists love to say what machine has better penumbra than other machines, it's usually a religious argument. Once you get photons, how do they interact with matter? So at the energy photons, we typically use the primary mode of interactions called Compton scattering, and the way it works is if you have a photon incoming, it can hit an outer shell, what's called a free electron, and transfer energy to the electron and knock it off the atom, ionizing it, and then deflecting the photon at a different energy in a different direction. And these secondary electrons are the core of what John was talking about last night, because they go on to ionize atoms and create free radicals and cause DNA damage. So what do we mean when we say gray or units of absorbed dose? So photons are indirectly ionizing, so what happens is the photon comes in and it'll travel in the tissue away before it eventually will have an interaction, and that interaction will knock off one of those electrons, and the electron will go on to have lots of little ionization events and give up all of its energy along some kind of crooked path. And so delivering dose to tissue is where we're getting the notion of absorbed dose. So dose is given in units of gray, and one gray is one joule per kilogram. And what's important about that is in bulk, the way photons attenuate is if you have a photon source and you scan along the primary beam, you're going to get a curve that looks like one of these curves, where you have a region called a buildup region, where there's actually not a lot of absorbed dose, because most of the dose gets absorbed downstream of where the photon has an interaction. So this region here is called the buildup region, and it's why we can treat people with photons without just burning a hole right through their skin, right? So most of the absorbed dose is at depth. And the other thing to notice is that as the energy of the photons increases, the depth at which the absorbed dose happens also increases. So those are sort of the two primary things about photon attenuation. So enough about basic physics. So how does this relate specifically to radiosurgery? So radiosurgery has the problem that we're usually trying to treat things that are hard to operate on surgically, and so we're often close to things that are really critical. And there's two examples here where we have a pituitary tumor and the optic pathways are millimeters away, and a spine case where you're trying to treat a vertebral body and the spinal cord is millimeters away. And so to be able to do this, you have to generate really sharp dose gradients. And in radiosurgery, we're usually talking about dose gradients on the order of 10 to even 25% per millimeter. And so if you're going to do this, then that means you have really high requirements for accuracy and precision. And so the way we generate these dose gradients in radiosurgery, no matter how we're doing it, is by spreading out the energy. So we use many beams from many directions, and that can be the 192 beams of a gamma knife, or it can be hundreds of beams that we use in a cyber knife treatment, or we can spread energy out on a linac using arcs. And so another term that we talk about are isocenters, which are just the point that all these beams are aimed at. And so you can have single isocenter treatments, multiple isocenter treatments, or even non-isocentric treatments, depending on what you're doing. So if you're going to generate all these gradients and aim them at an isocenter, you need to be able to aim somehow. So the classic way to aim is to use a stereotactic frame. There's a couple examples here. The stereotactic frame actually sets up a coordinate system we can target off of. And so when we image the patient, we put an indicator box on top that has kind of an end pattern. And so when you slice through it on tomographic imaging, you get dots that you can identify in the image and tell the computer where you are in three-dimensional space. Another way to aim is to use onboard imaging. So you have a treatment planning image, you make a treatment plan. The machine itself has an imaging unit, so you can see an example on a linac here. You get the patient set up, you take an image, you co-register it to the planning image, the computer tells you how different those two images are, and then you move the patient to the correct location so they match the position they were in when you made the treatment plan and then you can treat the patient. A lot of times anatomy is moving, especially if you're doing extracranial cases. And so we have different techniques for immobilizing patients from thermoplastic masks to bite block frames to vacuum cushions. And then we also do things like abdominal compression to sort of minimize the amount of movement. We can even help the patients hold their breath. And we can also do what are called gated treatments, where we're only treating in certain parts of the respiratory cycle that are kind of easy to manage and keep anatomy in sort of a consistent position each time. Radiosurgery treatment planning. So for those that haven't spent a lot of time looking at radiosurgery treatment plans or any treatment plans, this is pretty much what a treatment plan looks like. These lines are called isodose lines, and those are just lines of equal dose. And you can think of it exactly analogous to a contour map of a mountain. So if you drew a profile through this, you would get some mountain-looking-like pattern where it's dose instead of altitude. So usually these lines are depicted as a percentage of some maximum dose. And so we'll make a treatment plan based on percentages of a maximum dose, and then we'll pick some line that we're going to assign a specific amount of absorbed dose to. So in this case, the yellow line is the 50% isodose line. If we give 15 gray to that 50% line, then that predetermines all the other lines, right? So 15 gray to the 50 is equal to 30 gray to the 100 is equal to 22.5 to the 75, et cetera. So to make that plan, there's two ways basically people try to go about it. There's forward planning and inverse planning. So in forward planning, it's up to the person running the planning system to decide on beam placement and beam weighting. Here's a GammaKnife example where you've got a spot and a tumor and they don't really match. So you add another one and then another one and then a bunch more, and eventually you get a prescription isodose line that closely matches the shape of the target. The other way to do it is inverse planning, which is more common on the linear accelerators where instead of manually placing everything, you manually place some things, but mostly you decide on dose-volume constraints. So you say, I want to give 95% of the tumor X amount of dose, and I don't want to give more than 5% of an optic nerve Y% of the dose, and then there's an optimizer in the computer that tries to figure out the best solution. It's just sort of two ways to go about it. Once we have a treatment plan, we need some way to evaluate them. So there's something called a dose-volume histogram that's just a plot of relative volume versus dose, and it kind of takes what's a three-dimensional dose distribution and boils it down into something easier to understand. And so here you can see that the target is getting a lot of dose and the normal tissue structures are getting a little bit of dose. Their DBHs are used to evaluate everything from tumor coverage to dose homogeneity to dose-volume constraints, but you have to remember that it ignores the geometry of the distribution because you've lost that part. But geometry is everything for radiosurgery, so I encourage you to always look at the actual treatment plan also. One thing that boils out of the DBH are what I like to call the VX and DX metrics. So when you read papers, people are always talking about things like the V95 or the V12 or the DO5. So these are just boiling the dose-volume histogram down further into simple metrics, right? So the D numbers are the volume of tissue receiving at least some amount of dose, and that can either be as a percentage or an actual dose in gray. And the D values are the dose received by some amount of structure. So it could be, again, it could be in a percentage or it could be an actual volume. And the problem with these are that often in the literature they don't actually put what unit they're using. So you have to know that V95 is probably a percentage and V12 is actually dose in gray. But you have to be really, really careful because it's inconsistent, so you have to always look for what the units are. Okay, so another way we evaluate plans is through something called a conformity index, which is just a metric that tries to tell us how closely the shape of our treatment plan matches the thing that we're trying to treat. There's a number of different conformity indices that have been adopted over time. One of the common ones in radiosurgery is the PATIC conformity index, and you can see the formula here. But again, it's just trying to say that something like this is more conformal to the target than something like this. It's a pretty simple metric that people use to evaluate plans. So radiation oncology in general, I would say, has a sort of bewildering and increasing amount of acronyms and random jargon that can be really confusing. So there's 3D conformal and IMRT and VMAT and IGRT and SRS and SBRT and SABR all together. And so one of the things in this talk I want to urge you is not to get too worried about it. And remember that SRS and SBRT use all these other things, so they're not all separate. They're all interrelated. So not to get too worried about it. And so for the end of the talk, sources of uncertainty. So radiosurgery is a complicated multi-part procedure that has a large uncertainty chain. And it includes everything from the mechanics of the machine to the biological model to whether we even can define what target we're trying to hit. And so there's a lot of different places uncertainty creeps into the equation. And so even though your computer gives you a really beautiful treatment plan with nice sharp lines and stuff, what you're actually delivering is something much blurrier and it's maybe a little offset and a little blurry. And this is something I got from Stan Benedict at UC Davis, and it's just to remind us that even some things we think of as static, like the spinal cord, actually can maybe move a little bit. So never believe that what you're seeing on the screen is absolutely reality. And so to deal with this, the last set of definitions for this talk is you'll hear people talk about margins, especially in the rest of radiation oncology, but even in SBRT. And so there's a bunch of different types of margins and margin structures called GTV, CTV, ITV, PTV, and you'll hear us talking about this all the time. So for now, just know that these are all trying to account for different parts of the uncertainty equation. And it's not always so formal in SRS and SBRT. So to summarize, you know, it's a jargon-heavy field, and while the details of the physics can be complicated, the general principles of radiosurgery are actually pretty simple, which is why it's been successful. And so if you have questions, don't be afraid of your physicist. Most of us, we don't bite. And finally, and maybe most importantly, remember as you're doing this that any complex procedure has many sources of uncertainty, so always be aware of what you're doing. And then this slide, if you get the slides, I guess, is just something from XKCD that is a handy reference to different levels of dose for different things. So that's all I have for the moment.

medical physics

radiosurgery

radiation oncology

neurosurgery

gamma rays

X-rays