First Year Exam: TPS and Treatment Planning Flashcards
What does AAA stand for?
Anisotropic Analytical Algorithm
What kind of kernels does AAA use?
Pencil beam kernel
What is a kernel? How are they made?
Kernels are a known distribution of dose deposition of secondary electrons set in motion by the primary interaction(s).
They are pre-calculated using monte carlo codes.
What is the difference between a pencil beam kernel and a point kernel? Which is more accurate? Which is faster?
A pencil beam kernel calculates the fluence of the primary beam on the patient skin and sums the pencil beam kernels from skin to medium.
A point kernel uses point-by-point dose results at each interaction site/voxel. It involves first figuring out where the interaction occurs, then superimposing the point kernel onto the geometry.
Point kernel is more accurate. Pencil beam is faster.
What is the difference between convolution and superposition?
Convolution is a mathematical operation that shows degree of overlap between two functions (Ex. TERMA and Kernel). It does not involve heterogeneity corrections.
Superposition uses density to scale the kernel prior to the convolution. Allowing it to account for heterogeneity.
What are the inherent limitations of kernels?
They have radial symmetry and do not account for side scatter very well. They do account for forward scatter well however. They are pretty simplistic.
What is the Linear Boltzmann Transport Equation?
A function that describes the macroscopic behavior of radiation particles traveling through matter.
What assumptions does the LBTE make?
Particles only interact with the medium and not with each other
No magnetic fields
In broad terms, what does LBTE mean?
It’s simply balancing the particles to say
Particles transported + those absorbed = total number of source particles
What is the goal of Acuros?
To discretize the LBTE and directly solve for it
How do you improve the accuracy of your monte carlo calculation?
Simulate a larger number of particles and histories
How do you improve accuracy of your acuros calculation?
Minimize discretizations
What is a “history”?
A single particle and all of its descendants
what is the equation for noise in a monte-carlo calc?
1/sqrt(N) where N is the number of simulated particle histories
What is a phase space file?
It’s a precalculated file describing radiation emanating from the machine head. It can be used for all calculations to follow, assuming that radiation from machine head does not change.
What does monte-carlo calculate, that most other calc algorithms do not?
Monte-carlo calculates dose to medium
All other algorithms calculate dose to water
What are the pros of using Dw over Dm?
All historical prescribing is based on Dw
Accelerator reference dosimetry is based on Dw
What are the pros of using Dm over Dw?
Dm is more clinically relevant
Accounts for heterogeneity in material by considering the density AND the effective Z, whereas Dw calculations only account by scaling density of water.
How do Dw and Dm account for hetereogeneity? How much of an error do you expect from these two methods?
Dw scales the density of water, ignoring the material type
Dm scales the density of the material and considers the effective Z
In tissue, error is 1-2%
In Bone, error can be as high as 15%
What is the general rule of thumb for determining calc grid size?
<= 3x3 cm2 field, you use 1-2 mm
> 3x3cm2 field, you use 2-3 mm
Give the general workflow of an acuros calculation
A) Transport fluence from source model to patient
B) Calculate scattered photon fluence in patient
C) Calculate scattered electron fluence in patient
D) Dose calculation
Step a is performed for all field orientations. Steps b-d are only performed once.
Up to what energy beams can Acuros be used to calculate?
25 MeV (well below our clinical use)
True or false, Acuros utilizes the same source model as AAA?
True
True or false, for Acuros dose can be calculated to either medium or water?
True
Recall for Acuros, you need to know the material anyway, meaning you already know the effective Z.
What assumptions does Acuros make?
Photons can create electrons but electrons cannot create photons (ignores Bremsstrahlung)
Particles only interact with the medium and not with each other
No magnetic field
What parameters does Acuros discretize in solving the LBTE?
Space
Energy
Angle
With Monte-Carlo, at what point in the electron journey does the algorithm begin to track the history?
When the electron leaves the bending magnet
How does monte-carlo figure out mass density?
It bins HU values into different tissues types
What is the transport cutoff for acuros and monte carlo?
1 keV for photons and 200 keV for electrons
What limitations do both acuros and monte carlo share?
Reliance on NIST interaction cross-section data
CT conversion table (HU to mass density)
Reliance on mapping mass density to material (inherent error)
Assumptions of LBTE (no magnetic field, particles don’t interact with each other)
What limitations are inherent to only Monte Carlo?
Statistical noise depending on number of particles simulated
Assumptions are made for clinically relevant calculation times
What limitations are inherent to acuros only?
Discretization size for space, energy and angle results in loss of accuracy
True or false, acuros simulates individual particles
False, therefore it also produces no statistical noise
What effect if responsible for Sc(10x15) not being equal to Sc(15x10)?
Collimator exchange effect
How do model-based dose calc algorithms take into account inhomogeneities?
By scaling kernels with density
How does Acuros modulate its dose calculation densities?
Higher densities in field, lower densities out of field
(makes sense, it’s trying to be as accurate as possible with the field doses. Doesn’t care as much about all the other scatter and leakages)
True or false, for both Acuros AND Monte-Carlo you need to know the material in your scan
True
Monte-carlo is an example of a ________ calculation, Acuros is _________.
Monte-carlo is an example of a stochastic calculation, Acuros is deterministic.
What is the general workflow to go from CT number to known material?
CT# –> mass density –> some known material
Both steps in the process have their own mapping
Each map has some inherent error
The total error of the process is a product of the errors of the individual steps
What is the equation for dose, using TERMA and Kernel?
Dose = the convolution of TERMA and Kernel
That is, the degree of overlap of the total energy released in matter and the pre-calculated monte carlo distributions
At what energies (lower or higher), do photon dose calculation algorithms work best at?
Lower energies
What kind of interactions are photon calc algorithms best at modeling? Why is this an issue for electron calculations?
Best at modeling primary fluence interactions
They typically approximate secondary interactions. For electrons, secondary interactions are much more important.
What is the MU limit per arc in SRS and SBRT?
10800
Per field MU limit in VMAT?
2000
What is the minimum MU/degree for a 3D arc plan?
0.1 MU/degree
Or 36 MU for a full arc
Which algorithms use electron density curves? Which use mass density curves?
AAA uses electron density curves (HU–>electron density)
Acuros and eMC use mass density curves (HU–> mass density) then figure out the material type based off the mass density.
What is the limit of the mass density curves used for Acuros and eMC?
They only have built in mapping for materials up to an HU corresponding to 3 g/cm3 (or if you extend the curve, then 19.33 g/cm2). Anything above 2.2 g/cm3, and you need to manually assign the material type to the voxels.
What is the density of gold? Why is that important to know?
19.32 g/cm3
It’s important because it is the highest density material that you can see in a body
How does a mass density curve figure out what material to assign if the HU falls between two materials?
It interpolates a mixture of the two material, just for the sake of macroscopic cross section calculations.
What happens if your calculation grid is smaller than your CT slice thickness for Acuros?
Acuros will interpolate the CT values in between and calculate a grid that is better resolution that its respective CT.
True or false, you can change the energy, space and angle discretization settings for Acuros manually?
False
They are defined internally. You can’t change them.
What is the maximum allowed electron density in AAA?
15
What energy ranges can Acuros XB be used for?
4 MV to 25 MV
What calc grid sizes is Acuros used for?
1 to 3 mm
True or false,
Monte Carlo also solves the LBTE?
True and False…
It doesn’t try to directly solve it. But in doing its calculation, it does actually find a solution to the equation.
True or false,
Monte Carlo takes into account charged particle interactions between the primary and secondary electrons?
Kind of…
It approximates
What are the five steps to Acuros XB dose calculation?
- Create physical material map
- Transport components from photon beam source model into the patient
- Transport the scattered photon fluence in the patient
- Transport the electron fluence in the patient
- Calculate desired dose mode (Dw or Dm, Acuros can do either)
Why do Acuros and Monte Carlo absolutely need to know material type and mass density?
Because they use the macroscopic cross section to figure out probability of interactions
What spatial discretization does acuros perform?
Uses a dose grid resolution equal to your user requested size within the PVOI
Uses a dose grid resolution 2 times larger than your user requested size outside of the PVOI
PVOI is the volume where the dose is estimated to be 10-15% of max dose or greater
How many photon energy cross section groups are available in Acuros?
25 total groups
Remember, Acuros discretizes energy in order to use one of the groups
What structures does eMC take into account? What structures does it ignore?
Accounts for added bolus and all voxels within body structure
Ignores support structures
How does eMC transport model, macro monte carlo (MMC) actually work?
It has a pre local geometry probability distribution functions for small spheres. This is calculated in a uniform material only once
Then it performs a macroscopic calculation by transporting the particles through the CT in macroscopic scales based on the PDFs generated in the local calculation
What materials and sphere densities are used for the local geometry PDFs?
Broken into 5 different materials and 5 difference diameter spheres.
Materials include air, lung, water, lucite and bone
Diameters include 1,2,3,4,6 mm with increasing diameters for increasing electron energies
What is the general electron transport workflow for eMC?
Use spheres to determine electron direction and step size
Reduce step size if near boundary
Stop step size at boundary, and transform into step size for new material at other side of boundary (accounts for heterogeneity)
What are some methods of dose smoothing that MC simulations may use to reduce statistical noise?
Gaussian Dose Smoothing
Median Dose Smoothing
What are the two types of optimization strategies? Which one do we use with Eclipse?
Analytical and Statistical
We use statistical with Eclipse
Give a brief explanation of the analytical optimization strategy
You tell the software your desired dose distribution, and it does a reverse process as CT in which it figures out the fluences needed to reach this desired distribution
Like CT, however, it can arrive at results with negative or imaginary fluences since not all distributions are doable
Give a brief explanation of the statistical optimization strategy
It’s a process that aims to minimize a cost function that describes the difference between desired and current solution doses. You assign a priority to each constraint, and it works to minimize the differences given the priorities.
Is VMAT step and shoot or sliding window?
Sliding Window
What’s the difference between step and shoot and sliding window?
With step and shoot, you move the MLCs, turn on the beam, turn off beam, then move MLCs with beam off
With sliding window, you have the beam on as you move the MLCs
What are some advantages to IMRT/VMAT over 3D?
Superior gradient falloff
More conformal dose distributions
Better control over low (for static) and intermediate (for VMAT) doses
What are some disadvantages to IMRT/VMAT vs 3D?
Longer treatment times
More leakage
Heterogenous target coverage with localized hotspots
Increased uncertainty in beam deliverability and modelling
What are some advantages to IMRT vs VMAT?
Better low dose spread
Usually less complex to model (since gantry doesn’t move)
What are some advantages to VMAT vs IMRT?
Quicker treatment times
Less MU
Less leakage dose
Can optimize distribution from many beam angles
What is one inherent limitation of VMAT that IMRT does not have to deal with?
There are minimum and maximum gantry speeds and dose rates. meaning, if you have in angle in the plan that’s really good at delivering dose and you want to deliver as much dose as possible at that one specific angle, you’ll have trouble with VMAT since you need the gantry to be slightly moving at all times, and there’s only so much you can increase the dose rate by.
Meaning, there is a maximum MU you can deliver at a given angle.
How does TPS account for most beam-modifying accessories for AAA?
Applies a user-defined transmission factor to model change in fluence when a beam gets shaped by the accessory
What LINAC are the head beam fluence and energy phase space models calculated and derived for in Eclipse?
Based on a Varian Clinac and adapted to machine specific
There is pre-calculated Monte Carlo code for the phase space model and energy fluence, but the parameters that shape this code are adapted using the structure and material composition of your machine specific accelerator head.
How can you access various transmisison factors, head materials, output factors, etc.?
Through the RT Administration
What is it about the beam that most beam modifiers actually modify?
The Fluence (and to some extent the energy)
So really, the energy fluence
List some conditions that require urgent radiation treatment within hours of diagnosis
Acute cord compression SVC Syndrome Bronchial Obstruction Tumor Bleeding Increased brain pressure (mass effect)
What is the normal tissue objective (NTO) value indicating?
An indication that tells your optimizer how you want dose to fall off as you move away from PTVs
What is the advantge of DIBH for left sided breast cancer treatment?
Increase separation between breast and heart
Expand lung volume (makes volumetric dose constraints easier to meet for lung)
What structures do you use an ITV for?
Targets that are suspected to move
These include targets in the lung AND in the liver
How is a DRR produced?
Determine a virtual source position
Trace ray lines through CT data to a virtual plane, at a distance of the imaging panel for the accelerator
Sum attenuation coeffieicnets along each line to produce an image at the virtual plane
What are control points characterized by?
Gantry angle
MLC leaf positions
MU weight
How does RapidArc achieve dose modulation? Which method is preferred?
Either change the dose rate (by changing timing of electron injection into acceleartor guide)
or
Change gantry speed
Changing the dose rate is more preferred, because the ganry has a certain amount of inertia
How are control points used in VMAT?
You optimize using 1 control point per 2 angles of gantry rotation
During delivery, your MLCs and dose rates move from control point to control point with the beam on (if sliding window) or beam off (if step and shoot)
How are control points used in IMRT?
If step and shoot - you jump from point to point and fire at each control point
If sliding window - you jump from point to point and fire between each control point
In a statistical optimization algorithm, what is the process used to avoid falling into local minima?
Simulated Annealing
How does Simulated Annealing work?
Hint: Acceptance probability
The algorithm decides on an “acceptance probability”, which is initially high when you begin optimization.
This means, that if the change in cost function is positive, you are allowed to take that change if it falls under the acceptance probability (this helps you escape the initial local minima
For later iterations, you want to lower the acceptance probability, since at that point you’re trying to fine-tune and try to settle down the algorithm to avoid jumping out of the global minima
What is the go-to biological parameter that can be used for treatment planning? And what does it mean?
gEUD - generalized equivalent uniform dose
It is the uniform dose that, if delivered over the same number of fractiosn as the non-uniform dose distribution of interest, yields the same radiobiological effect
How do you use gEUD?
Give it a high value (8-40) if you want to limit max dose of a structure
Give it a low value (around 1) if you want to improve low/intermediate dose
What is the major limitation of using gEUD by itself, for serial organs, parallel organs, and targets?
For serial organs, it only focuses on max dose and doesn’t work well to limit low and intermediate doses
For parallel organs, it results in very hot, hotspots
For targets, it can result in very high hotspots
What is the general recommendation for use of biological models for optimization according to TG 166
Use gEUD with your usual volume constraints. Don’t use gEUD or any biological parameter by itself
When did biological treatment planning plug ins first become available to Eclipse?
Version 10
What are the things that you ACTUALLY need to check on a monthly basis during TPS QA
CT Data input and image quality. This includes…
Accurate SSD's Spatial resolution Contrast resolution Geometric scaling CT number consistency Electron density consistency
Remember: dose calc accuracy is an annual test. We just do it monthly, but it’s recommended to be annual
What information do you need for commissioning of a TPS?
Treatment machine data
Algorithm verification
Calculation verification (through use of phantoms)
DVH accuracy
Import/Export accuracy (including patient orientation)
Image quality
What are the advantages to 3D vs IMRT/VMAT?
Quicker treatment time Less MU (less leakage) Better low dose spread Generally more uniform target coverage Less beam deliverability uncertainty Less modelling complexity
What type of algorithm is a Collapsed cone algorithm and does it utilize heterogeneity corrections?
Point kernel
YES it does utilize heterogeneity corrections
What are two systems you should probably know that utilize collapsed cone algorithm?
Pinnacle and Tomotherapy
What values does the Modified BATHO Power Law utilize?
TMR ratios and densities of inhomogeneous material
What are the basic limitations of the equivalent path length method for homogeneity correction?
It cannot take into account inhomogeneities that exist lateral or beyond point of interest. Only those in the path of the beam prior to the inhomogeneity
What is the main benefitof a fast fourier technique? How does it allow for that benefit?
It’s fast… clearly
It allows of fast calculation by performing convolution in Fourier space which simplifies calculations
What is the main downside to fast fourier technique?
It can only be applied for true convolutions, (not superpositions), so no heterogeneity corrections are allowed unless they’re applied to the dose instead of the kernel (which is less accurate than the other way around)
How does collapsed cone convolution speed up calculation? What is the downside?
By ignoring multiple scattering events
Reduces accuracy, but actually not by much. It’s one of the more accurate approximations out there
What are the three sources of radiation that contribute to a AAA and Acuros calculation?
- Primary photon source (primary bremsstrahlung photons from target that don’t interact with the head)
- Extra-focal photons which come from interactions in the head
- Contaminating electrons
What three things do macroscopic cross sections depend on?
Material
Radiation type
Energy
What is the name of the process of determining where sources of radiation from head reach the patient’s body?
Ray Tracing
Define GTV, CTV, ITV and PTV.
Gross tumor volume (GTV) - visible tumor volume in images
Clinical target volume (CTV) - GTV + subclinical/invisible/microscopic invasion of cancer
Internal target volume (ITV) - CTV + IM (internal margin for organ motion)
Planning target volume (PTV) - ITV + SM (setup margin for setup error)
What is the conformality index? What is the ideal value? What is an acceptable deviation?
Ratio of the volume covered by the reference isodose over the total target volume designated as PTV. So…
CI = VRI(% or Gy)/TV
Ideal: CI = 1
Acceptable: CI > 0.95
What calc grid size do we use with Acuros?
0.125 cm
What calc grid size do we use for AAA, non-SRS
0.25 cm
What calc grid size do we use for AAA, SRS or SRT?
0.1 cm
What calc grid size do we use for eMC?
0.25 cm
What statistical uncertainty do we assign to eMC?
What is the limit?
2%
3% limit
What CT sim slice thickness do we use for the following procedures.
HDR:
Stereotactic:
Everything else:
HDR: 0.625 mm
Stereotactic: 1.25 mm
Everything else: 2.5 mm
Which report defines GTV, CTV and PTV?
ICRU 50
What are the clinical applications of PET in treatment planning?
Staging and/or determing the GTV for a boost volume or nodal volume
What are the clinical applications of MRI in treatment planning?
Better visualization of structures. typically used for brain (T2, axial), spine (T1&T2 axial or saggital) and prostate (T2)
Why is a higher hot spot acceptable in SBRT, SRS and SRT?
Because you want quicker falloff and aren’t concerned with too much dose in the GTV (and sometimes PTV)
After finishing an optimization, what information per angle is given? What happens to this information afterwards?
Fluence
From the required fluence, MLC motions are determined.
In a breast setup, which arm is raised? The treated side or opposed site? Why?
Always raise treated side arm
Reduces motion and variability of breast, and allows for tangents that won’t pass through arm
How are wires used in some facilities for breast plans (not ours typically)
Used to mark border of breast
Why is the Sclv field in breast plans typically tilted 10-15 degrees away from patient midline?
To avoid spinal cord and esophagus
What is the clinical significance of the skin sparing strip for extremity plans?
To avoid lymphedema
What beam orientations are electron beams typically treated?
En face
What material are the cutout blocks used in our clinic?
Copper
Why do we use prone treatments for rectal patients?
To get better margin around tumor volume and to displace the small bowel anteriorly
What is the purpose of filling the bladder in prostate patients?
Pushes the prostate to the same place every day and gets more of the bladder out of the field
What is the purpose of emptying the rectum in prostate patients?
Keeps rectum out of treatment area and keeps prostate in place during treatment
For patients with tumors in the oral cavity, what is the purpose of a bite block? (2 purposes)
- To keep the oral tounge down and out of the treatment field
- open the commisure of the lips to prevent a bolus effect in the lip folds
Will decreasing normalization value increase or decrease hot spot?
Increase
for normalization, you’re dividing all dose by that value, so decreasing the normalization value increases all doses
What affect does smaller fields in electrons have on PDD curve?
Makes everything shallower, and increases Dmax
What isodose lines do physicians usually prescribe to for electrons?
90% or 80%
What is the pro and con of prescribing to 80% vs 90%
Con: higher hotspot (20% hot)
Pro: Larger coverage area
Which modality has higher integral dose? Tomo or Linac?
Tomo
Which modality has sharper dose falloff? Tomo or Linac? How does it achieve this?
Tomo because it treats slice-by-slice instead of the entire field
What are some scenarios where we used to use Tomo back when we had it?
Hippocampal sparing (better dose carving)
Chestwall (tangent beams can get the superficial dose)
Long plans with multiple targets
What are some selection specifications you have to make in Tomo that you don’t have to make in traditional Linac?
Modulation factor and pitch
Why should you never place stereotactic patients on tomo?
Because imaging won’t allow for a precise alignment
Does increasing energy typically increase or decrease hot spot?
Increasing energy tends to decrease hotspot
What is one thing you can do to bring dose down in an AP-PA treatment (besides FiF or wedge).
(Think to planning exam)
Add a 3rd field
What is typical dose fall off for eclipse (%/mm)
Roughly 5%/mm
Fill in the blank…
VMAT and Tomo cause _____ low dose than 3D plans
VMAT and Tomo cause more low dose than 3D plans
What is the old method for heterogeneity corrections and dose algorithms?
Correction-based
What are correct-based algorithms based off of? (3)
- Measured dose distributions in water
- Corrections for beam modifiers
- Depth, field boundaries, path lengths
What is the current method for dose algorithms?
Model-based
What are model based calculation algorithms based off of?
Beam intensity (energy fluence)
In what medium is a kernel calculated?
Homogenous water phantom
What is a “radiologic path length”?
Correction of distance using electron density relative to water
What does AAA do to HU values higher than the maximum HU value defined in the CT curve?
Truncates them to the maximum electron density value
Does AAA tend to over or underestimate dose?
Underestimate
For VMAT plans, how do we position the rails?
Put rails all the way in (less angles that they’ll end up hitting)
For static fields, how do we position the rails?
If posterior obliques, put the rails all the way in
If PA field, move the rails all the way out
If you can’t model all components of the treatment setup and couch, what should you do?
- Have avoidance strategies to minimize effects of unmodeled setups
- Perform measurements to see what the expected %differences are
What is the approximate immobilization device attenuation error (in %)?
1-4% for most immobalization devices
5% for Calypso array
What is the approximate attenuation effects of carbon fiber couches when treating oblique vs straight forward?
2-6%
What is the approximate attenuation effects of higher density couch tops?
Can be as high as 15%
What is the approximate attenuation effect of the calypso couch top?
1%
What is the range of buildup discrepancies on skin dose if you don’t take couches into account?
1.5 - 7x more buildup on skin
How does the percent contribution of out of field dose change for the following as you move further from field edge?
- Patient scatter
- Leakage
- Collimator scatter
- Patient scatter - contribution fraction decreases as you move further out of field
- Leakage - increases
- Collimator scatter - stays roughly the same
In comparing IMRT vs 3D, how do doses compare in and out of the treatment field?
IMRT has higher doses out of the treatment field (more leakage and scatter), but lower doses within the treatment field
For Proton therapy, what do you think contributes the majority of out of field dose?
Neutron contamination
At around how far from field edge can differences between measured and calculated doses from photon TPS calculations start to be noticeable?
What typically causes this difference?
Is it underestimating or overestimating?
3 cm from field edge
Underestimation of head leakage, patient scatter and collimator scatter
So an important thing to note, in general the dose is UNDERESTIMATED when you’re 3 cm away. That’s a big deal…
Approximately what magnitude of reduction of integral dose is proton therapy able to produce vs IMRT?
~2-3x less
What are some methods that can be used in TP and simulation that can help reduce out of target dose? (9 total, try to come up with as many as possible)
- Reduce target volume
- Use FFF instead of flattened field
- Beam energy considerations
- Avoid wedging
- Utilize MLCs (they reduce head leakage)
- Utilize beam angle optimization
- Utilize jaw tracking
- Ancillary patient shielding
- Image only what’s necessary
What energy is optimal for reducing scatter? Why is this the case?
10X is optimal
If you go higher, you get neutron contamination
If you go lower, you need more MUs to get an equivalent dose coverage, which results in more scatter
What is a general rule of thumb regarding whether or not TPS calculated MEAN dose is accurate or not?
If majority of organ is in 5% isodose line, mean dose is accurate
If not, mean dose isn’t accurate but a point dose may be
What is a general rule of thumb regarding whether or not TPS out of field dose is accurate or not?
If it’s beyond 3 cm from the field edge (50% line) or the 5% isodose line (for IMRT), then you assume it’s no longer accurate
In what situations does photon therapy actually have better out-of-field dose than proton therapy?
The further from the field you get, because proton contributes a lot of neutron contamination
To decrease uncertainty in a monte-carlo calculation by a factor of 3, how much does the computation time need to increase?
9x
Time is proportional to change in uncertainty squared
How is electron uniformity index defined, and what is an acceptable value?
Ratio of area of 90% isodose to 50% isodose
UI > 0.7 is acceptable
So note: for electrons, the uniformity index is a measure of penumbra
What are fiducial markers (typically BBs) made out of?
Gold
What are BBs being placed in body used for?
Setup in which treatment targets are in regions of low contrast, making it difficult for OBI to visualize them
What is the most commonly used practice of BBs implants in body?
What are two other treatment sites where you might see other forms of fiducial markers used in the body?
Prostate
But you can also see them in stents for pancreas or clips for breast lumpectomy
What is the most dominant photon interaction in radiation therapy, and what is it dependent on?
Compton Scatter
In the therapeutic energy range, it’s dependent on electron density, hence why we need HU to electron density curves
For electron monte carlo, how many simulated particles histories per voxel do you need to achieve an accuracy of 1%?
10,000 particle histories per voxel
NOTE: It’s particle histories per voxel NOT total particle histories
What areas of the LINAC are encompassed in the phase space file modeling?
The phase space file only pre-calculates particles exiting components of the LINAC head prior to the jaws an MLCs. The reason being is because the jaws and MLCs are patient specifici, so you can’t have one file for all patients.
Which of the following would result in a more well defined dose distribution with sharper isodoses?
Step and Shoot or Sliding Window IMRT?
Step and Shoot
In sliding window, having the beam on as you move MLCs results in blurring
Fill in the blanks…
When d < dmax, you are ___________ dose if you don’t account for couch
When d > dmax, you are ___________ dose if you don’t account for couch
Fill in the blanks…
When d < dmax, you are __underestimating__ dose if you don’t account for couch
When d > dmax, you are __overestimating__ dose if you don’t account for couch
What is roughly the attenuation of the calypso couch?
1%
What is roughly the drop in tumor dose for IMRT and VMAT for most couches if you don’t account for them?
2-3% tumor dose drop
What is roughly the attenuation of carbon fiber couches if you don’t account for them?
2-6%
What is the MU verification typical delivered dose error limit?
5% is what’s commonly used
2% is what’s recommended from TG 114
According to TG-158, what is the approximate out of field TPS dose estimation error at dmax and at surface? What causes a lot of the error? What is the implication for pacemaker nanodot measurements?
The error is caused by poorly modeled secondary source fluences (electron contamination, leakage), and approximations of internal scatter
From dmax to surface there is a BUILD DOWN effect, where the dose from out of field sources can increase up to a factor of up to 5x from dmax to surface.
From 3 - 10 cm from field edge, the dose can be a little over 1% CAX dose.
IMPLICATIONS FOR PACEMAKERS: Since nanodots are used to measure surface dose above the pacemaker, you expect the nanodot doses to be somewhere around 6x larger than the TPS estimated dose to the pacemaker.
Which method requires less MUs?
Step and Shoot vs Sliding Window
Step and shoot generally requires 25% less MU than sliding window
This is because step and shoot has less dose blurring and can make the distribution tighter with less modulation
Which of the following contributes the lowest amount of out-of-target dose?
Cyberknife, GammaKnife, Linac-Based SRS
What are the causes of TPS under-estimating out of field dose?
Underestimation of….
- Head leakage
- Patient scatter
- Collimator scatter
According to TG-114, when should you perform a MUV?
Preferrably before first fraction is treated
Or at the very least, before 3 fractions or 10% of the treatment is delivered (whichever is less)
Besides Lung and Liver, what other organs may be affected by respiratory motion? (5)
Esophagus Pancreas Breast Prostate Kidneys
They can all be susceptible to respiratory motion, some more than others, but none as drastic as lung and liver
Between TERMA and KERMA, which is the more encompassing value? What does it include that the other doesn’t?
TERMA is more encompassing
It includes energy losses that DO NOT get transferred to the medium (mainly scattered compton photons). KERMA ignores these and only focuses on the energy transferred to the medium.
Which technique is more susceptible to setup error, IMRT or 3D?
IMRT due to higher dose conformality and steeper gradients
What is the difference between treated volume and irradiated volume?
Treated volume is the volume enclosed by the prescription isodose line
Irradiated volume is the volume enclosed by the 50% isodose line
How much transmission is a cerrobend block allowed?
Less than 5%
This is 4.3 HVLs minimum
What’s the difference between TERMA and KERMA?
TERMA includes energy that is not transferred in medium (Scattered compton photons), KERMA does not. KERMA only includes energy that is transferred to the medium.
Thinking about it in terms of attenuation coefficient…
TERMA corresponds to the mass attenuation coefficient. That is fraction of primary beam lost
KERMA corresponds to mass transfer coefficient. That is fraction of primary beam lost and transferred to medium
Why for AAA lung treatments is there a recommended limit to field energy and for acuros there isn’t?
AAA does a poor job of modeling side scatters and downstream scatters relative to the path of individual beamlets. As a result, recommendation is to use 6X or lower for AAA lung since scatter contribution increases as energy increases
Acuros handles heterogeneity very well, as result you are not limited to energy use
How does AAA calculate dose to a volume ?
Beamlets, angle of incidence, and effective path length
Each beamlet has an associated kernel that is derived from monte-carlo in water. The effective path length will strecth and shrink the kernel accordingly, as will the geometry of the angle of incidence
AAA is a analytical solution that solves for dose to a voxel for every voxel using a lookup table that depends on effective path length and angle and calculates the dose contribution of each beamlet/kernel to the specific voxel
How exactly does AAA pre-calculate kernels?
It calculates a set of monoenergetic kernels using monte-carlo in water
Then to calculate the polyenergetic energy spread kernel, it’s a spectrum-weighted sum of the monoenergetic kernels
How exactly does AAA model lateral heterogeneity for kernels?
Density scaling along radial “spokes” that radiate laterally from the pencil beam central axis
Hence the anisotropic part
How do you adjust beamlet size for AAA calculation?
The beamlet edge-to-edge width is the user set dose calc grid size. So you just adjust the grid size
What is the known uncertainty of AAA in lung for 6MV and 10 MV or higher? Does it underestimate or overestimate? Are errors better or worse for smaller fields? Are errors better or worse for less dense lung?
6 MV or lower - < 3% underestimation. Worse for smaller fields, worse for less dense tissue
10 MV or higher - < 7% overestimation. Worse for small fields, worse for less dense tissue
What are the two terms that AAA is convoluting in its dose calc?
TERMA and Kernels
Briefly describe the difference between convolution and superposition as they pertain to AAA?
Kernel is a dose distribution map that, given TERMA, described dose deposition for fluence for a given beamlet. This process of mapping the two together is convolution
Superimposing is the process of summing all beamlet contributions to all voxels
