Transcription
Hello everybody, and welcome back. This is part two in a three-part series looking at dose and CT imaging. Our focus today is going to be on estimating the absorbed dose that a patient receives during a CT examination.
Now, we saw in part one, it was the movement of an electron, a charged particle, through tissues that was responsible primarily for the dose in CT imaging. And that movement of the electron causes energy deposition in the tissues through a process known as linear energy transfer. We have x-rays that are being emitted from an x-ray source. They head towards the patient and interact with the patient tissues. The photoelectric effect and Compton scatter cause the release of those electrons that are ultimately going to provide the dose to the patient. That interaction is what's known as exposure, and we can measure exposure in Roentgens or Coulombs per kilogram.
Now, in air, dosemeters can convert the exposure that we have calculated into a value known as air kerma, or kinetic energy released per unit mass. That will give us how much energy has been released in our dosimeter, and that's going to become incredibly important later when we try and estimate absorbed dose in CT imaging, and I'll bring that up again.
Interactions that happen in the human body, there's going to be energy deposition, Joules per kilogram, or Grays. But that's much more difficult to actually measure. That energy deposition is what's known as the absorbed dose. And the absorbed dose can then be converted into a dose that's known as the equivalent dose, where we account for the type of radiation that we're using to deposit that energy in tissues. And different radiation types are responsible for varying degrees of DNA damage and ultimately long-term cancer risk.
Now, in CT imaging, we're using x-rays. This conversion from absorbed dose to equivalent dose is converted by a weighting factor of one. So, the absolute value of absorbed dose and equivalent dose remains the same. However, the units change. Lastly, we can account for the tissue sensitivities to ionizing radiation. Different tissues are more or less sensitive to ionizing radiation and have differing likelihood of causing cancer over the lifetime of that patient after the examination has occurred.
So, how do we go about calculating, or at least estimating, absorbed dose in CT imaging? I've said in the first part, it's very difficult to actually measure absorbed dose. Well, what we do is we create an estimate known as the Computed Tomography Dose Index, or the CTDI. This is again measured in Grays. Here, we're calculating energy per unit mass of tissue.
Now, I'm going to say this a couple of times in the talk, but this is perhaps the most important thing to remember. This is an index. An index is something that you look up. You go to the back of the book, you find the thing you're looking for, and then you reference that index. The same with CTDI. These are pre-calculated values that are specific to the CT machine. It's got nothing to do with the actual patient. It's got everything to do with an experiment that is run in that specific machine for specific parameters: a specific filament current, a specific kVp, and a specific filtration. So, this says very little about what's actually happening in our patient. It talks much more about the CT parameters that we've chosen, and it allows us to compare CT machines to one another: the amount of energy that these CT machines are releasing towards what's known as a phantom.
Now, what exactly is a phantom? Well, to calculate the Computed Tomography Dose Index, what we do is we place a phantom within that CT machine, and it's made of PMMA, polymethyl methacrylate. Now, we run our CT as we would usually, but instead of the patient being in the CT machine, this phantom is in the CT machine. And these phantoms have differing diameters depending on the CTDI index that we are trying to calculate. The standard phantoms are 32 cm in diameter or 16 cm in diameter, trying to mimic the abdomen or mimic the head or the pediatric population.
Now, we saw that the absorbed dose depends on the diameter of the patient. The smaller the diameter of the patient, the higher the absorbed dose per unit mass in that patient. That's why we need to use different size phantoms here. Now, this index is going to create a list of values that we can look up in a table. You'll see that before placing your patient within the CT machine, you can set the parameters, and you will get a CTDI value.
Now, how do we go about calculating the CTDI value? Well, within this phantom, we place two dosimeters. These are the same dosimeters that we were looking at when we were looking at exposure. They contain gas, they've got electrodes on either side, and they can measure the current that's being produced after x-ray exposure and convert that current per kilogram into a Joules per kilogram figure. That Joules per kilogram figure is going to be a proxy for our absorbed dose. Remember, though, that this is absorbed dose in air. There's no patient tissue here.
As x-rays pass through this, they're going to be attenuated by this PMMA phantom, but they're not going to be attenuated the same as they would be in tissues. So, this is all an estimate. Secondly, the width of this, or the diameter of this phantom, is standard 32 cm. Patients come in all shapes and sizes, and 32 cm is actually quite a large diameter. Many of your patients are going to be much smaller than a 32 cm diameter. That means that our CTDI value that we calculate, this proxy for absorbed dose, is going to way underestimate in these larger phantoms if you've got patients that are smaller than these phantoms. All important things to remember.
So, what do we do? We place this phantom within the CT machine, and we run the CT machine like we would our normal patient. The CT machine is going to rotate around this patient. Now, the first value we're going to calculate is what's known as the CTDI100. These dosimeters that we have placed at the periphery and at the center of our phantom are 10 cm long, 100 mm long. That's why it's called CTDI100.
If we were to look at this side on, we can see the beam is passing through these dosimeters, and it's going to rotate around the phantom here. Now, look how this dosimeter was closer to the x-ray source at 0 degrees. It was furthest away from the x-ray source at 180 degrees. It's mimicking what we're doing with our patient. What we can then do is plot the dosimeter readings on a graph here. This distance here represents the entire length of the dosimeter that we placed here. At the central axis of our beam, here, we're going to get the most radiation dose, and as we head out to the sides, we're going to get less and less radiation dose.
Now, you might think the edges of these dosimeters aren't going to get any radiation dose, but remember that the x-rays that are incident on this patient are going to scatter, and that scatter is going to mean that dose is applied even outside the width of the x-ray beam. So, we can plot that dose on the graph here for our peripheral dosimetry. Here, I've drawn it in purple. See how most of the dose is within the width of the x-ray beam? Now, we can calculate the total dose here by calculating the area under this graph.
To calculate the area under the graph, we use this formula. This is the CTDI100 for the peripheral dosimeter. That's important for the peripheral dosimeter. What we do is we normalize for the width of the beam. We take the number of detectors that we have in our beam and the total width of each detector, so we're getting our exposure through the entire width of the beam, and we get an integral of all of the exposure measured in the Z axis here, along the longitudinal axis of our patient. And this just means that we're adding up all the dose 50 mm from the isocenter this way and 50 mm from the isocenter this way. So, we're adding up all of these values here and getting the area under the curve. That's going to get us a total dose that this dosimeter has measured during one rotation.
We can do the same for the central dosimeter here, and again, plot those values. You see how the beam gets slightly wider when we're further out here, because that fan beam has got some cone angle, and again, we can create the CTDI100 for the central dosimeter here. The same thing, we're calculating the area under the graph. Now, that's what's going to give us our CTDI100 values for both the peripheral and the central dosimeters. Again, these are air kerma values. They're not true tissue absorption values.
Now, we know that when the CT machine rotates around the patient, the dose is going to be unequally distributed depending on the location, whether it be near the periphery or whether it be near the center. There's going to be more dose imparted at the periphery than there is at the center. And what we can do is weight these CTDI values to kind of approximate that distribution of the dose. And that weighting is what's known as the Computed Tomography Dose Index Weighted, CTDIw. For a single rotation of the x-ray source, we can take the peripheral reading that we had and contribute to our weighting two-thirds of that reading. We can take the central reading that we had and contribute the remaining third of that reading. So, we take two-thirds of the peripheral reading and one-third of the central reading, and that's going to give us our CTDI weighted value. That allows us to compensate for that varying distribution of dose because of the rotation of the x-ray source around the patient. So, this estimates the dose for a single rotation to the phantom.
Now, remember, the patient moves when we are rotating the x-ray source. Now, I'm going to see the show the x-ray source moving here, but remember, it's actually the patient moving. The x-ray source stays in the same location, but effectively, this is what happens when we expose a patient to x-ray radiation during our acquisition. Notice how that dose now is being spread along the Z axis here. In our CTDI weighted example, we calculated the dose for a single rotation, but as that x-ray source is rotating, we're exposing different parts of the patient because the patient is moving through.
Now, remember that ratio to how quickly the patient moves through this imaging machine and how quickly we rotate the x-ray source around that patient, as well as the beam width, is known as pitch. The faster the patient moves through the CT machine, the higher the pitch value. Remember, the pitch is a function of the table speed, the rotation time of our x-ray source, and the width of the beam at the isocenter. Lower pitch values are going to expose more of the patient. We're going to have higher dose values here. As the pitch increases, our dose value is going to decrease, and remember that dose is inversely proportional to pitch. You see that doubling the pitch will halve the dose here.
Now, why am I going over this? Well, we calculated a CTDI value weighted for one rotation, but we haven't taken into account this pitch, this movement in the Z axis. And what we can do is account for pitch by creating a value that's known as our CTDI volume value. We're getting the volume or the weighted volume in the Z axis as the patient moves through the CT machine. So, this is our CTDIv. And what we do, it's very simple. We take our CTDI weighted value and divide it by the pitch. So, the higher our pitch, the lower our CTDI volume value is going to be because we're spreading that dose over a further region.
What we've got here now is a proxy for absorbed dose for a single slice in our CT machine. Obviously, we're exposing the patient to multiple rotations of that x-ray machine around them. And now, if we want to calculate the total absorbed dose that they receive during a whole examination, we need to account for the scan length of the patient. This value here, the CTDI volume value, is independent of scan length. It hasn't taken into account how long our scan, our coverage of our scan.
In order to do that, we calculate a value known as the Dose Length Product. For one rotation of our CT machine, say we cover this much of the patient's tissue, for each subsequent rotation, we're going to be adding that absorbed dose, that same CTDI volume dose. And what we do is we take our CTDI volume and we multiply it by the scan length in centimeters. That's going to give us the Dose Length Product, the absorbed dose for the entire coverage of our CT scan.
Now, if we're trying to estimate risk for the patient, we're trying to tell them how much stochastic risk they've got, we want to take into account the entire dose that they're receiving over the full coverage of the scan. So, our Dose Length Product has estimated absorbed dose across the scan length. But how do we then say, well, that absorbed dose equals this amount of risk? And this is probably the most difficult thing to do in CT scanning. First, we've calculated Dose Length Product using CTDI, which has used an artificial phantom. It's calculated doses based on air kerma on two spots in that phantom. That phantom is a rigid size. It hasn't changed size depending on the type of patient that we put in the machine. We haven't accounted for the different tissue types. If I'm scanning the head versus scanning the abdomen, those organs have very different tissue sensitivities to radiation.
So, how do we convert this absorbed dose into that effective dose in CT that we mentioned in our last talk? What is the effective dose, that if their whole body was exposed to that amount of radiation, it would equal the same amount of stochastic risk? We're going to cover four separate ways to kind of calculate or at least estimate effective dose, and I'm a bit hesitant to go through these because at best, these are an estimate. If we're trying to attribute risk to a specific patient, it becomes very difficult, and the accuracy is not good because all of this takes into account data from population data, not from individual data. So, we can often apply these to populations or if we're comparing studies, the risk of different studies to one another, but it's very difficult to attribute risk to a specific patient. So, you need to keep that in mind.
Perhaps the most crude way we can do it is what's known as a Dose Length Product conversion. Before we've placed the patient in the CT scanner, we've selected our tube current, we've selected our kVp, we've selected the filtration, it's going to give us a and we've selected the pitch. It's going to give us a CTDI volume value. We know our scan length that we want to get, so we're going to get a DLP value. We can get a DLP value without placing the patient in the machine at all, so that shows you that best, this is an estimate. We're looking up an index to create this.
What we can do, though, is take the DLP that's given to us by the CT machine and apply a coefficient, what's known as a K factor. Now, you'll see that different examinations, different CT examinations have different K factors. You can take the DLP and multiply it by this K factor to get a very estimated effective dose in this CT machine. To create these K factor values, what they've done is they've taken all the organs that are going to be exposed, looked at the equivalent doses, and looked at the type of radiation, and looked at the tissue sensitivities of each of those organs, summed them all together to get an effective dose for these specific examinations, again, over a population of patients, not a specific patient. We haven't taken into account the size of patients. All of this is averaged out, and again, these are very broad brushstroke estimates here, but it gives us a ballpark figure. Is does the CT head give you more risk than the CT abdomen or less? And we can use these DLP conversions here.
The next thing we can do is use what's known as size-specific corrections. Can measure the diameter of the patient, the lateral diameter of the patient, and use a table to look up what the effective diameter of that patient would be. I'd highly recommend that you read this report. It's a very easy read. It's only about seven pages, and it covers a lot of what we're covering in these talks very well, and perhaps most importantly, it highlights some of the shortcomings of trying to estimate effective dose. In fact, they state in this report, you can't with confidence on an individual basis give patients an actual risk value. You can give patients ranges of estimates of risk, but where they fall within that range is very difficult to calculate.
Now, these calculations of the conversion factors also used what's known as mathematical modeling and calculation. A famous one is what's known as the Monte Carlo simulation, where you take each voxel within the patient and try and mathematically model how x-rays would interact with each of the tissues within the patient, and this got more and more accurate over time. There's been many generations of Monte Carlo simulations where we've taken actual CT scans, we've mapped out the shapes and sizes of various different organs, we've tried to calculate doses that each one of those organs will occur within different size patients and within different parameters of our CT machine. This also takes into account things like backscatter. You know, the skin gets dose from x-rays passing through it, but then they can be backscatter and further dose that we haven't calculated for in our previous modeling. Again, though, this is based on population data and a lot of estimation.
Lastly, we can calculate what's known as a lifetime estimated risk, the LER, where we can take whatever effective dose that we've calculated, however we've calculated it, and multiply by a risk per Sievert value, which takes into account the patient's age, the patient's sex, and that tries to give us a better lifetime risk. Now, there's many shortcomings of this, and I hesitate to put it in here because a lot of the figures that you see online, X number of cancers are caused by CT scanners every year, and you see it in the hundreds of thousands. All of that data is extrapolated from very murky data, and a lot of it is from this B7 report, but I include it here because this is something that can come up, and again, I will link to this report, and you can see how they've come about generating these risk estimates.
So, now we've seen that we can estimate the absorbed dose using CTDI and using the phantoms. We can convert that CTDI100 value into a weighted value and a volume value, and we can spread that volume value over the Z axis over our scan length and create what's known as a Dose Length Product. We can then use that estimated absorbed dose to try and calculate effective dose here in Sieverts. That's going to give our patients an idea of what their risk is for a specific examination, whether it's based on their type of examination as well as their anthropometry, or whether we've used mathematical calculations to try and be more accurate with those effective dose values that we've calculated.
Now, I've mentioned here that effective dose is at best an estimate, and it's difficult to tell a patient what their risk is. However, in clinical practice, there are many precautions that we can take, many steps that we can take to reduce the absolute dose that the patient is going to receive. And in the next talk, I'm going to go through systematically the factors that influence dose in CT scanning and how we can change those factors in order to reduce the patient dose while still getting the images that we need. So, I'll see you all in that talk. Until then, goodbye everybody.