3D Treatment Planning and verification with hand calculations

3D Treatment Planning and verification with hand calculations Randy Holt, PhD, CEO Pacific Crest Medical Physics, Inc Chico, CA

Dosimetric Calculations Calibration and Measurement Conditions Patient Conditions r r

Primary Fluence Fluence - the beam the leaves the machine and reaches the patient All MV x-rays come from the Target. Not all x-rays produced by the target reach the patient. Some x-rays reach the patient after scattering off parts in the collimating head. Approximately 90% directly from the Target 5% scatter off the Primary Collimator 4% scatter off the Flattening Filter 1% scatter off the Jaws,Blocks,MLC

Absorbed Dose = Primary + Scatter Primary Dose - Beam fluence enters the patient (or other medium) and with each cm of distance travelled, some percentage is converted into ionizing radiation and that energy is absorbed by the patient or medium. Scatter Dose -Some dose is scattered away from the primary path, and is either absorbed, scattered back or exits the patient.

Beam Profile - dose in water at given depth, d Dose from the fluence exiting the collimators is measured using an ionization chamber in a water filled phantom. The measurements along a plane orthogonal to the beam axis are called a Beam Profile Beam Profiles are characterized by, Energy, Depth, Field size, and orientation in the collimator, e.g. x, y, diagonal Location along the axis of measurement OAR off axis ratio is a profile normalized to the central axis value E.g. 6MV OARx (x, d, FS)

Depth Dose measurements down ray-lines from surface, or orthogonally The Central Axis Depth Dose (DCAX) is the most commonly measured and used for calculations models. PDD or Percent Depth Dose is the Depth Dose normalized by the maximum dose and converted in to a percentage. PDD (d) = D(d)/Dmax* 100% Like profiles, PDD is characterized by Depth Energy, and Field size SSD Source Surface Distance to beam entrance E.g. 6 MV PDD (d, FS, SSD)

PDD characteristics f 1 Surface dose Build-up region D max which occurs at depth d max Attenuation All components in the PDD curve are affected by Beam energy, Lower energy means more absorption per cm, or more attenuation Prior attenuation of the primary, u (units of 1/cm) Back Scatter fluence, Ks, which depends on depth and Field Size, Distance to the beam entrance, f 1 P(d,r,f 1 ) = 100 * [(f 1 + d max )/(f 1 +d)] 2 * e u(d- d max)) * K s r d

Inverse Squares Law Fluence (and dose) vary with 1/r 2 D (a) = D(b) *(b/a) 2 Example: If you measured 20 cgy dose at 100 cm from the source, and then moved the chamber (and phantom) to 120 cm, what would you measure? A. 24 cgy B. 16.6 cgy C. 13.9 cgy D. 28.8 cgy

Effect of SSD on PDD 120 6 MV at various SSD For two SSD, f 1 and f 2 P(d,r,f 2 ) = 100 * [(f 2 + d max )/(f 2 +d)] 2 * e u(d- d max)) * K s P(d,r,f 1 ) = 100 * [(f 1 + d max )/(f 1 +d)] 2 * e u(d- d max)) * K s 100 80 Then P(d,r,f 2 )/ P(d,r,f 1 ) = [(f 2 + d max )/(f 2 +d)] 2 / [(f 1 + d max )/(f 1 +d)] 2 Rearrange to the Mayneord F-factor, which can be used to convert any PDD measured at one SSD to another SSD F = [(f 2 + d max )/(f 1 + d max ) * (f 1 +d) / (f 2 +d)] 2 When f 2 > f 1, then F >1 When f 2 < f 1, then F <1 Or PDD increases with an increase in SSD 60 40 20 0 0 5 10 15 20 25 30 35 SSD 80 SSD 100 SSD 120

Mayneord F factor example The PDD measured at SSD= 100, for a 6 MV beam of FS=10 at depth 7 is 86.1%. What would the PDD at 80 SSD be? A. 84.0 % B. 87.6% C. I can t tell

Mayneord F factor example The PDD measured at SSD= 100, for a 6 MV beam of FS=10 cm at depth 7 cm is 86.1%. Assuming dmax(6mv) = 1.6 cm, what would the PDD at 80 SSD be? A. 84.0 % B. 87.6% C. I still can t tell

Mayneord F factor example The PDD measured at SSD= 100, for a 6 MV beam of FS=10 cm at depth 7 cm is 86.1%. Assuming dmax(6mv) = 1.6 cm, what would the PDD at 80 SSD be? A. 84.0 % B. 87.6% C. None of the above D. I still can t tell F = [ (80 + 1.6) / (100 +1.6) * (100 + 7) / (80 + 7)]^2 = 0.976 And so 86.1% * 0.975 = 84.0%

Effect of Energy on PDD The higher the energy, The lower the attenuation The deeper the dmax The shallower the build up curve Which can be useful for sparing skin dose 1.2 1 PDD Build up region - PDDs 10x10 Field Size with energy xcgy/100 MU 0.8 0.6 0.4 0.2 6X 10X 15X 0 0 1 2 Depth 3 (cm) 4 5 6

Effect of Field Size on PDD A larger field size means Increased scatter dose at depth Increased Surface Dose Decrease in dmax (+/- a few mm) 1.2 6X PDD with Field Size - Buildup region xcgy/100 MU 1 0.8 0.6 0.4 0.2 0 5x5 10x10 20x20 Depth (cm) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 A larger field size also means more entrance fluence But this effect is normalized out by PDD

Dose with Field Size 100 SSD, 100 MU Remember the Normalization of Dose PDD? That Dose difference with FS didn t go away, we just chose to ignore it for a bit Where does it come back in? xcgy/100 MU 120 100 80 60 40 20 0 6X Dose with depth and Field Size 5x5 10x10 20x20 0 5 10 15 20 25 30 35 40 Depth (cm)

Output Factor - Dose at d ref with Field size and Energy Normalized to the Reference Field Size 10x10 cm (standard by convention) Usually measured at d ref = 5 or 10 cm depth, to avoid measurement errors in build-up region due to electron contamination May be corrected later for differences in PDD(d ref, FS) to define dose at d max 1.15 1.1 1.05 1 0.95 0.9 0.85 OF (Sc,p) with square field size Also called Sc,p 0.8 0 5 10 15 20 25 30 35 40 45 6 MV

Sc,p with energy Higher energy means LESS attenuation, and so less dose is scattered not much difference, though a few % at largest field sizes Very small fields sizes not shown. Very small field sizes require special measurement techniques and corrections. 1.2 1.15 1.1 1.05 1 0.95 0.9 0.85 0.8 OF (Sc,p) with square field size 0 5 10 15 20 25 30 35 40 45 4 MV 6 MV 10 MV 15 MV

OF has two components, Sc,p = Sc * Sp Sc is the output due to Collimator Head Scatter 3 major components, Target, Primary collimator, Flattening Filter Effect of PC and FF depend on Field Size r F 1 =SAD r d ref Sc, measured inair, but with a build-up cap Sc,p measured in water or water equivalent medium

Sc and Sc,p can be used to derive Sp Sp = Sc,p / Sc Sp is the dose contribution from scatter from all points in the phantom In theory Sp goes to 0 for zero cm 1.15 1.1 1.05 1 0.95 0.9 0.85 0.8 0.75 Sc,p, Sc, Sp = Sc,p/Sc, 6MV field size 0.7 0 5 10 15 20 25 30 35 Sc Sp Sc,p

Sc and Sp for various energies Average values from hundreds of institutions Quality-Assurance Check of Collimator and Phantom Scatter Factors, Tailor, Followill, et al JMP, Year : 2014 Volume : 39 Issue : 3 Page : 184-191

Equivalent Square Sc,p 3 4 5 7 10 15 20 30 40 6 MV 0.88 0.906 0.928 0.963 1 1.039 1.067 1.104 1.119 Linac Collimators define a square or rectangular field Often, only square field sizes are measured, especially for PDD or TPR. Can use Sterling s Formula (aka Equivalent Squares) S = 4 x Area / Perimeter Also S = 2*L*W/(L+W) Works for Sc, Sc,p and Sp Example. For a 3 cm x 10 cm field, calculate the EqSqFS, and Sc,p S = 4*(3*10) / (3+3+10+10) = 4.6 cm OF(4.61) = 0.919

Sc,p with non-square fields 6 MV Field size X [cm] - > Y jaw (cm) v 3 4 5 7 10 15 20 30 40 Because X and Y jaws are at different distances to the primary collimator and flattening filter, they have slightly different contributions to Sc (and by extension to Sc,p). The larger the FS difference 3 0.88 0.891 0.899 0.908 0.915 0.92 0.923 0.926 0.926 4 0.892 0.906 0.915 0.927 0.936 0.943 0.946 0.95 0.951 5 0.901 0.917 0.928 0.942 0.952 0.961 0.966 0.97 0.971 7 0.913 0.932 0.946 0.963 0.977 0.988 0.995 1.001 1.003 10 0.924 0.946 0.962 0.983 1 1.015 1.024 1.032 1.035 15 0.932 0.956 0.974 0.999 1.02 1.039 1.052 1.063 1.067 20 0.936 0.961 0.98 1.007 1.031 1.053 1.067 1.082 1.086 30 0.941 0.967 0.987 1.016 1.043 1.069 1.086 1.104 1.11 40 0.943 0.97 0.99 1.02 1.048 1.076 1.095 1.113 1.119 the larger the difference. Recall the EqSqFS predicted for a 3 x 10 FS, S = 4.6 Sc,p = 0.919

Why Break Down Sc,p into Sc and Sp? Because after the Linac Jaws, we add on Blocks or MLCs Because the patient doesn t always fill the area covered by the jaws r blocks So Sc accounts for dose variation until it passes the Jaws Sp accounts for the scatter inside the patient And then we can also break down the PDD even further to remove the effects of scatter as well as 1/r 2

Tissue Medium Ratio PDD with 1/r 2 and SSD removed Also called TPR, Tissue Phantom Ratio and Tissue Maximum Ratio Inverse squares is removed because the measurement chamber is always at one position (usually SAD, or isocenter) Requires a special water phantom that lifts the whole 300 kg water tank (smoothly!) while keeping the chamber in the exact same place w.r.t. the linac r F 1 =SAD r TPR (d,r) = D (d) / D(d ref ), TMR(d, r) = D(d) / D max

TMR with Field Size TMR does not include inverse squares or SSD or Sc effects All FS are normalized to 1 at d max Does include scatter changes with depth, which is why you see more dose at depth with larger FS 1.2 1 0.8 0.6 0.4 0.2 0 6 MV TMR with field size FS = 5 cm 7 10 12 14 20 30 0 5 10 15 20 25 30 Depth (cm)

TMR with energy By now it should be no surprise that higher energy penetrates deeper. 1.2 1 0.8 TMR with Energy 0.6 0.4 0.2 0 0 5 10 15 20 25 30 35 6 X, 10x10 15X, 10x10

TMR from PDD and back TMR (d, r d ) = PDD(d,r,f)/100 * [(f + d)/( f + d ref )] 2 * Sp(r dref ) / Sp (r d ) Where r dref = r d * (f+d)/f PDD(d,r,f) = 100 * TMR (d, r d ) * [(f + d ref )/( f + d)] 2 * Sp(r) / Sp (r dref ) But PDD is missing something, actually a few things.

Dosimetric Calculations Calibration Conditions Patient Conditions r r

Radiation Prescription and MU Total Prescribed Dose (cgy or Gy) Prescribed Dose per Fraction (cgy or Gy) Number of fractions, n Energy Prescription point Field Weighting (for > 1 fields) We want to find the MU (or time) required to deliver the Rx dose to the prescription point MU (or time) = Rx dose dose rate at that point

Dose rate factors (we already know most of these) SSD Dose Rate = K * PDD(d, r) * Sc (r c ) * Sp (r p ) * SSD factor SAD Dose Rate = K * TMR(d, r d ) * Sc (r c ) * Sp (r p ) * SAD factor K : the calibration of the machine. Universally we like to use 1 rad (cgy) per MU at dmax for a 10x10 field size. SSD factor = [SCD / (SSD + d ref ) ] 2 SAD factor = [SCD / SAD] 2, SCD is Source to Calibration point ( Usually SAD + dmax) Still missing a few things need to include about effects of wedges, block trays and the collimator edges and flattening filter

Setup: Non-isocentric (SSD) or Isocentric (SAD) SSD = 90 cm r SSD = SAD = 100 cm d = 10 cm d = 10 cm r Patient surface is at the axis of rotation Calc point is at the axis of rotation

Dose rate factors anywhere in the patient Dose Rate = K*TMR(d, r d )*Sc (r c )*Sp (r d )*ISF* OAR (d, x) * TF *WF(d, r d,x) ISF: the inverse squares factor at the dose calculation point relative to the calibration point K ISF = [(SAD + d ref ) / SPD] 2, SAD is fixed (i.e 100 cm), dref is usually dmax SPD is Source to Calculation point disatance For Isocentric calc s ISF = [(SAD + d ref ) / SAD] 2 OAR: Off-axis ratio, accounts for imperfectly flat beam, varies w/ d, energy and FS TF: Tray factor (only when blocks are used), varies per energy WF: Wedge factor, varies with FS and energy, and distance along wedge direction d : Effective or radiological depth

Open Field COLLIMATORS only Not truly flat

OAR at different depths and distances from CAX, FS = 20 x 20,6MV 6 MV OAR with Depth 1.04 1.02 1 0.98 0.96 0.94 0.92 0.9 0.88-10 -8-6 -4-2 0 2 4 6 8 10 1.5 5 10 20 30

Blocked Field COLLIMATORS and a Fixed Block to further shape the field MLCs are just a fancy way to avoid making blocks but since they have no Tray, they have NO TF Block reduces the scatter dose inside the patient Block Tray reduces all dose passing through it, by 2-6 %, energy dependent

Adding wedges to reduce the dose linearly across the field WEDGES attenuate the beam along the one entire axis Wedges can be oriented in any of the 4 cardinal directions, and can range from 15 to 60 degrees But can also be rotated into any direction by rotating the collimator Wedges can be and often are used with Blocks or MLCs The Dynamic Wedge (EDW) is a way of slowly closing a jaw to mimic a wedge, EDW have their own wedge factors

Why use wedges? Account for missing tissue, make dose homogeneous

COMPENSATORS are hand built custom beam modifiers that can change the 2D fluence Creating wedges with COMPENSATORS (e.g. Ellis Block) Compensators are used to correct for uneven surfaces They have a special WF (x,y) MLCs can be used to mimic the Compensator tray

Effective depth due to inhomogeneous tissue one final correction to the dose calculations Bone, metal implants, and air or lungs absorb dose differently than water Effective depth, d, also called radiological depth can account (to a first order approximation) for this change in medium d is only applied to the dose attenuation in TMR TMR(d, r d ) p e = 1.0 p e = 0.4 p e = 1.4 d 1 d 2 d 3 d = d 1 * p e (1) + d 2 * p e (2) + d 3 * p e (3) Bone will increase d Lung or air will decrease d

I thought this section was called 3D planning?

From one to five dimensions 1 D: Single line of dose calculated, usually on central axis, but can also be off-axis. Usually limited to a few points. Requires only distance to surface, depth of desired calculation point and off-axis location. May include known regions of inhomogeneity. 2 D: Patient contours in a single transverse plane, can be done with a stiff contour wire, or a single slice of CT data. Dose distribution in Targets and organs is incomplete. 2 D: Port film (MV or kv radiograph) for generating field placement, field size and blocks 2.5 D: Several transverse planes, limited to co-planar beam arrangements (no couch rotations). 3D: Multiple stacked slices that can be rotated and interpolated into any orientation or view. 3D modelling of beam entry and exit points. Has the ability to contour discrete volumes and compute dose entry from any direction. Can compute a Digitally Reconstructed Radiograph (DRR) for any beam angle, simulating the portal image for that beam. 3.5 D: Uses multiple 3D series, such as PET or MR along with CT. Used for contouring, but not usually for treating. 4 D: Adds the time element, taking gated CT phases and using them to track organ and target motion during treatment. Can be used for generating a more correct target for planning, but can also extend to 4D or gated treatment delivery.

From one to N dimensions 1 D: Single line of dose calculated, usually on central axis, but can also be off-axis. Usually limited to a few points. Requires only distance to surface, depth of desired calculation point and off-axis location. May include known regions of inhomogeneity. 2 D: Patient contours in a single transverse plane, can be done with a stiff contour wire, or a single slice of CT data. Dose distribution in Targets and organs is incomplete. 2 D: Port film (MV or kv radiograph) for generating field placement, field size and blocks 2.5 D: Several transverse planes, limited to co-planar beam arrangements (no couch rotations). 3D: Multiple stacked slices that can be rotated and interpolated into any orientation or view. 3D modelling of beam entry and exit points. Has the ability to contour discrete volumes and compute dose entry from any direction. Can compute a Digitally Reconstructed Radiograph (DRR) for any beam angle, simulating the portal image for that beam. 3.5 D: Uses multiple 3D series, such as PET or MR along with CT. Used for contouring, but not usually for treating. 4 D: Adds the time element, taking gated CT phases and using them to track organ and target motion during treatment. Can be used for generating a more correct target for planning, but can also extend to 4D or gated treatment delivery. 5 D: a popular Grammy award winning pop, R&B group from the 60 s.

3D Simulation - Imaging Equipment For 3D planning, most radiation therapy departments use CT scanning for patient simulation. This is due to: Excellent spatial localization of patient anatomy, including patient contours and inhomogeneity Good differentiation between bone, soft tissue, air and fat Three dimensional data Rapid acquisition No need for patients to remain on site after CT scanning has occurred MRI images may be fused to CT images, but have problems including less accurate 3D spatial mapping, small magnet bore and lengthy scan times. PET suffers from poor resolution. Plain imaging (particularly with fluoroscopy) allows for determination of volume movement but only in two dimensions.

Simulation - Localization Equipment Lasers are used in the simulation and treatment room to assist in patient positioning. These lasers allow for accurate determination of the linac isocenter as well as assisting with patient straightening. Lasers in both the treatment and simulation room are regularly checked for quality assurance. Tolerance of under 2 mm is accepted for QA purposes. Lasers may be located on the roof, walls and the machine itself. Lasers in the simulation room are also used for patient marking, usually with permanent tattoos on the skin. These tattoos allow the patient to be more easily positioned during treatment, although care must be taken as skin is highly mobile relative to deep structures. The Linac also has a field light and Optical distance indicator (ODI)

Simulation - Immobilization Equipment Immobilization equipment is designed to be comfortable and reproducible. Comfort helps patients remain still during treatment. Reproducibility allows the patient to be placed in the same position on every day which is vital for multi fraction treatments. In general, custom immobilization devices for individual patients provide increased comfort and increased reproducibility. Kneefix Perhaps the most simple piece of equipment, the Kneefix is a shaped foam block that lies underneath the patient s knees in the supine position. It provides comfort and helps the patient to maintain their position. Alternatively, the kneefix can be used to support the ankles during prone treatments. Vacuum Bag The vacbag is filled with polystyrene beads. It is easily shaped to fit the patient contour. When the air is removed from the bag, it becomes fixed in position and allows the patient The vacbag provides a comfortable, custom immobilization device that is relatively cheap and can be re-used. Thermoplastic Mask A mask that fits to the patient s head contour may be created by using a low melting point perforated plastic mold. This is lightweight, and the perforations allow the patient to breath and see without difficulty. It reduces the movement of the patient s head for more accurate treatments. Slanted Breast board, Prone board The breast board is a device with numerous settings to cater for individual patients. It contains two arm and two wrist supports to comfortably raise the patients arms above their head. The head rest and bumfix allow the patient to be positioned accurately on the treatment couch in a reproducible position. Despite not being custom made for every patient, the custom settings on the device allow for individual patient positioning and comfort. The prone board is used to allow face down treatment.

ICRU Reports 50 and 62 and 83 ICRU report 83, which discusses volumes and guidelines when performing IMRT, advances some of these concepts further The International Committee on Radiation Units and Measurements (ICRU) provides numerous guidelines for working with radiation. The ICRU reports 50 and 62 provide specific information for guiding radiation treatment. These recommendations suggest various volumes which are useful for planning radiation treatments, as well as a reference point where the dose should be calculated. GTV - Gross Tumor Volume. The GTV is the volume that contains the visible or clinically detectable tumor. This may be on clinical examination or on imaging. This is the smallest of all volumes and is not present in every plan (e.g. adjuvant radiotherapy following excision of the main tumor mass). CTV - Clinical Target Volume. The CTV is the volume which has been determined to require radiation treatment. This includes the GTV as well as areas of clinical risk such as lymph node groups or the region about the GTV that may have microscopic involvement. The CTV should be included in every plan. ITV - Internal Target Volume. The ITV includes a margin to account for physiological patient movements that are unable to be accounted for during treatment. This may include movement of the gut, beating of the heart or respiration. The margin required is known as the internal margin (IM) and may vary in height, breadth and depth based on the location within the body. The ITV is a newer concept that attempts to divide treatment inaccuracies into internal patient factors and external factors. If a method to reduce the effect of internal movements is used (e.g. respirator gating) then the ITV can be substantially reduced. PTV - Planning Target Volume. The PTV is an expansion from the ITV to account for external treatment inaccuracies. These may vary based on the department and the treatment site for instance a treatment inaccuracy of 7 mm for body treatments and 3 mm for head and neck treatments. This distance is the external margin (EM). Improving the external factors which lead to treatment inaccuracies may reduce the external margin and allow for smaller PTV expansions. OAR - Organs At Risk. Organs at risk are volumes placed on organs which are susceptible to radiation. They place constraints on the beam arrangement and dose that may be delivered. OARs may have different radiation tolerances based on the tissue involved. Additional margins of safety on organs at risk are often used. ICRU Reference Point. The ICRU recommends reporting the dose at a single point within the PTV. The point should be clinically relevant, easily defined, and placed in a region of uniform dose (away from steep dose gradients or inhomogeneities if possible). The point should be at the center of the PTV and at the intersection of the beam axes if possible.

GTV, CTV, ITV, PTV, OAR and Reference Point

2D And 3D Planning 2D planning 2D planning refers to calculation of dose distribution considering two dimensions only. It is faster and often possible to perform by hand. It has the downside of much less accuracy, as the contribution of dose from structures above and below the 'slice' are assumed to be similar to the slice being used. 3D planning 3D planning takes into account the variations in dose distribution caused by scatter from multiple planes. A computer is typically required due to the increased complexity of dose calculation. Current computers are able to compute dose in three dimensions much more quickly than a human planner could acheive for a 2D plan. The important anatomical patient data to collect for treatment planning includes: Patient surface contour Target volumes Organs at risk Inhomogeneities The most commonly used method of acquiring patient data is CT scanning, although MRI, PET

Patient Data Acquisition CT scanning is the ideal method of determining patient contour, as a definite change in radiation attenuation occurs at the air tissue junction. CT also provides accurate 3D spatial information and good resolution. MRI also provides good contrast between the air and patient surface. It is limited by inhomogeneities in the magnetic field. PET provides poor discrimination of patient contour due to inadequate resolution. Older methods include use of a contour plotter which used wires and paper to measure the distance from a point. Internal Structures CT scanning provides a balance of good resolution with poor discrimination between different soft tissue types. It provides very useful data on inhomogeneities such as bone, lung or air spaces. MRI has provides increased discrimination between soft tissues such as muscle or brain. It also provides some information on inhomogeneities. MRI is limited by long scanning times which may lead to movement artefacts in areas such as the lung. PET assists in localising structures which may be included in the GTV. FDG typically localises to areas of high glucose metabolism (such as tumours) and can indicate soft tissue masses that are involved with malignancy. PET must be used with CT or MRI for treatment planning to take place due to its poor resolution.

Choice of beam and modifiers- Choice of Beam Energy Choice of Beam Energy There are six typical beam energies to choose from: Superficial/orthovoltage/kilovoltage beams Low megavoltage photon beams ( 60 Co, 4 MV) Intermediate megavoltage photon beams (6 MV, 10 MV) High megavoltage photon beams (15 MV, 18 MV, 25 MV) Low megavoltage electron beams (6, 9 MeV) High megavoltage electron beams (12, 16, 20 MeV)

Choice of beam and modifiers- Choice of Beam Energy - Kilovoltage photon beams Superficial x-rays have their d max at skin surface, and dose falls of rapidly beyond this point. They are not suited for treatment to depths over 5 mm. Superficial x-rays are best used for superficial lesions that do not extend into deeper tissues. Examples where superficial x-rays may be useful are: Squamous cell carcinoma in situ (Bowen's Disease) Superficial basal cell carcinoma / squamous cell carcinoma Non-malignant conditions such as keloid scars If the lesions invade into the subdermal layers, electrons are often a better choice for treatment. Some regions (cheek, lips, eyelids) require internal shielding to be placed to prevent early / late side effects. Internal shielding requires coating with a tissue equivalent material to prevent backscattering onto the internal mucosal layer.

Choice of beam and modifiers- Choice of Beam Energy - Megavoltage photon beams The main differences between the different megavoltage beams are: Skin sparing is more pronounced for beams > 10 MV But, for beam above 10 MV, neutrons are produced in the linac head. d max moves deeper as beam energy increases Increased dose at depth as beam energy increases Widening of penumbra as beam energy increases Note that 60 Co beams (the lowest MV energy) have a wider penumbra than linac beams due to the finite source size Photon beams are ideally suited to treating deeply seated targets. Lower MV energies are more useful in thinner regions (such as the head, neck and limbs) whereas higher energy photons are more easily able to deliver dose to central targets in the abdomen and pelvis. The lung is a special case, as the widened penumbra seen with higher MV energies are more pronounced in lung inhomogeneity. The build up region in tissues surrounded by lung is also longer as photon energy increases. For these reasons, it is often preferable to use lower megavoltage beams (6 MV) in lung treatments which allow field sizes to be smaller. Photon beams are less suited when a lesion is not as deeply seated, and perhaps located near to critical structures. In these scenarios, an electron beam may be preferable.

Choice of beam and modifiers- Choice of Beam Energy - Megavoltage electron beams The main differences between the different electron beams are: Skin sparing becomes lower as beam energy increases (almost nonexistent for energies of 16 MeV and above) Therapeutic depth (the depth at which the beam is still useful) increases as beam energy increases d max initially increases with increasing energy Even though the 20 MeV beam has a shallower d max than the 16 and 12 MeV beams, but the therapeutic range is still deeper Higher energies show a broadening of penumbra at depth. This is due to increased lateral range of the electrons, which becomes apparent after they have lost there initial energy in the superficial tissues. Lateral isodose lines are constricted above 50% and broaden below 50%. Electron beams are suited to superficial treatments. They are able to treat to deeper depths than superficial x-rays, but also show a rapid dose fall off with minimal dose to structures beyond the treated area. Like superficial x-rays, electrons penetrate dense tissues (bone, metal) poorly and experience backscattering, making them unsuitable for treatment of tumours within bony cavities (such as the cranial vault).

Choice of beam and modifiers- Choice of Beam Energy - Summary of beam energies Kilovoltage photon beams are best suited for very superficial lesions of the skin which demonstrate minimal invasion. They outperform electrons in areas such as the face where there are critical structures in close proximity. Megavoltage photon beams are chosen when deep targets require treatment. They are minimally affected by bony inhomogeneity. Higher megavoltage energies are better for deeper targets in the abdomen and pelvis. In the thorax, lower megavoltage energies are better due to a thinner penumbra and faster buildup when lung inhomogeneity is present. Megavoltage electron beams are suited for lesions that are relatively superficial and not blocked by bony inhomogeneity. Lower energies have minimal depth penetration and are suitable for superficial lesions that invade deeply. Higher energies are suitable for treatments that are aimed at deeper structures, but suffer from a widening of the penumbra at depth.

Choice of beam and modifiers- Choice of Field Size Once a planning target volume has been decided, beams must be chosen to cover the volume adequately. The important step in this process is recognizing that all beams have a penumbra, and this penumbra must be accounted for when choosing a field size. The penumbra is dependent on the type of radiation used and the energy. Kilovoltage beams have a relatively small penumbra but a margin of at least 0.5 cm should be used to surround the PTV. Megavoltage beams have a sharp penumbra (about 0.7 cm) which becomes larger with increasing beam energies. Electron beams have sharper penumbras at low energies and shallow depths. High megavoltage beams show significant broadening of the penumbra at depth which must be accounted for if a deep structure is the target.

Choice of beam and modifiers - Choice of Beam Arrangement Kilovoltage and electron beams are typically used as direct fields. Electron beams are sometimes junctioned with other electron beams or megavoltage beams, leading to junctioning problems described elsewhere. With megavoltage beams, there are several options for multiple beam arrangements. A single field is rarely used as the dose distribution given is rarely of clinical use. There are some situations where a single field is preferable, for instance in supraclavicular fossa treatments or single palliative spine fields. Parallel opposed fields are commonly used for palliative treatments, as well as in some other circumstances. They deliver a high dose to most structures covered by the fields, making them less suitable when there are organs at risk present. They are best used for: Palliative treatments of the spine, thorax, brain and numerous other sites. This is because the dose delivered is unlikely to result in long term problems due to poor life expectancy. Treatment of limbs. There are few critical structures in the limbs, so long as the entire width of the limb is not treated. Treatment of para-aortic nodal regions. Some tumors (seminoma) metastasize preferentially to the para-aortic nodes. Due to the proximity of the kidneys laterally, it is often best to use parallel opposed fields to avoid dosing the kidneys at all. The relatively low dose used in this treatment assists in limiting toxicity. A wedge pair is a special type of treatment used for irregular contours or in an attempt to treat a relatively superficial lesion without dosing deeper structures. Multiple fields are the treatment of choice when the target volume is located near critical structures. Multiple fields allow the dose to be concentrated on the target volume, with a lower dose spread out to neighboring tissues. This lower dose field leads to less deterministic side effects, but may increase the risk of stochastic effects in the treated organs.

Opposed fields

Profiles though the axis of two opposed fields, 6X, for varying thicknesses of patient Note increasing Hot Spots as thickness increases For 10 cm thickness, max is < 1%, at 40 cm thickness is 61%

Opposing fields different energies For this 40 cm thick (!) patient, increasing the energy from 6MV to 10 MV to 15 MV reduces the hot spot from 61% to 35% To reduce hot spot further, need to add beams

Four field beam Box arrangement Now, hot spots are reduced, but there is dose going into the patient from the lateral directions. Not always tolerable

Eight fields Even more conformal, but now there is some dose everywhere.

360 Beams. AKA Rotational arc therapy Basically, just turn the beam on and rotate the gantry Usually calculated as a series of discrete steps, from 1-5 degrees

Arc therapy for different energies Full arcs, 3 energies 40 35 30 10x10 6X 10x10 10X 10x10 15X 25 Title 20 0-25 -20-15 -10-5 Title 0 5 10 15 20 25 6X has slightly higher dose at surface, 4-5% Which can be good or to be avoided. 15 10 5

Wedges, drop the dose on one side of the beam

Different wedges available From 0 deg to 60 deg Can mix and match wedges to get somewhere in between 60 degrees I.e. Some open and some wedged, from same direction Can also orient wedges in different directions, i.e From AP beam right to left, and from PA beam, from inf to sup Relative dose at 10 cm deep Dose Profiles for various wedge, 100x10, 6X 10x10 W00 120 10x10 W15 10x10 W30 100 10x10 W45 10x10 W60 80 60 40 20 0-20 -15-10 -5 0 x_axis 5 10 15 20

Wedges correct for oblique beam entry With Gantry entering at 45 degrees, note the sloping effect of the field

Add a wedge to the same beam A 45 degree wedge almost completely counteracts a 45 degree beam incidence

3F beam arrangement to spare one side Often used to spare the spinal cord or some other critical structure Note the nonuniformity of the beam across the central target region

3F Wedged arrangement Wedges from lateral directions counteract and balance the dose from the AP direction More uniform target dose Can put small hot spots in the lateral tissue

Example of beam weighting, 50/50, norm to isocenter At Isocenter, each beam delivers has 150 cgy. Can use that info to compute MU s What else do we need to know to compute MUs? What is the dose just under the skin (AP and PA) at dmax?

Example of beam weighting, 2:1 (67/33), norm to isocenter At Isocenter, the AP beam deleivers 200 cgy and the PA beam delivers 100 cgy Can use that info to compute MU s What is the dose just under the skin (AP and PA) at dmax?

Blocked fields, No Divergence

Blocked Fields with Divergence Divergence highlights the property of Similar Triangles L2 = L1*H2/H1 H1 H2 L1 If you move your target from isocenter to 10 cm deeper, how much bigger or smaller will the block aperture from each portal direction need to be? Can you think of a way to set up your patient and plan so that you could use the same block for each direction? L2

SSD vs SAD SAD is so easy, and the gantry rotates on the isocenter. When would you choose SSD? Reduce hot spot different inverse squares curves. In order to fit a target into a single field E.g. You are trying to treat a 42 cm long femur, and want to give 2 cm margins for both positon and dose bowing at the ends. Your jaws only open to 40 cm (nominal). What Source to Femur distance do you need to encompass the entire femur in a single portal? At the CAX, the Femur is 5 cm deep in the tissue. What would the SSD be isocenter or ODI reading be? Can you think of a way to do it SAD? Divergence highlights the property of Similar Triangles L2 = L1*H2/H1 L1 L2 H1 H2

Long field blocking options SAD = 100 SAD = 100, CA = 45 SSD= 100, can still fit using MLCs

Junction or matched beam arrangements Needed for when desired FS is larger than maximum FS (>40 x40) Requires two or more isocenters Ideally, only along one axis to minimize patient moves

Overlap and under at field edges With two diverging fields, but from one direction, there will always be a region where dose is overlapped (hot) or underlapped

Minimize underlap by raising beams to SSD With careful matching of adjacent fields, underdose can be avoided

But that increases total overlap

Opposed fields can reduce hot spots

But not completely eliminate cold spots

Think in 3D! By rotating the patient couch 90 degrees, and rotating the gantry, you can eliminate divergence at the abutting beam edges

Using non-divergent field edges Single iso center method. Beams from one direction are limited not extend over the isocenter. An orthogonal pair is added to deliver matching dose Also called half-beam blocked matching With no divergence, hot spots and cold spots are minimal

A 3F HBB match method For one direction, typically the anterior, only use a single field Usually used to treat a shallow target from one direction, and minimize dose to an organ behind the field E.g. spinal dose.

Choice of beam and modifiers - Bolus Bolus is typically tissue equivalent material may serve two purposes: Compensating bolus is used when there is contour irregularity on the patient's surface. The bolus smooths out the contour, and prevents hot spots and cold spots from developing in undesirable locations. Build up bolus is used when skin sparing is not desirable. Skin sparing is a feature of megavoltage photon beams as well as electron beams with energies under about 15 MeV. In some situations, such as treatment of skin tumors which are invading deeply, dose is required at skin surface as well as at depth. Alternatively, skin dose may be desirable following the resection of some tumors (such as following mastectomy for breast cancer, or some sarcoma treatments). Bolus thickness is based on the build up region of the depth dose curve, which is typically larger for lower electron energies and higher photon energies.

Flat Bolus raises dose to surface, can also use shaped bolus (e.g. 3D printed, or wax, or playdough-like jelly) to smooth out surface irregularities Bolus calcs are easy just account for change in entrance SSD and depth. Eg If normal setup is SSD 94 cm, calc depth 6cm, then 1 cm bolus makes it 93 SSD cm, calc depth 7 cm.

Plan comparison and evaluation Hot Spots Cold Spots Dose uniformity Dose in target vs. dose outside of volume Dose Volume Histogram Other things to look for: Blank or missing dose usually indicates an incomplete body contour Dose out where no dose belongs Usually indicates accidental PTV contours Review of Blocking (especially when using auto-block generation) Entrance and exit dose specifically skin dose Dose where there are organs but no contoured volumes Sometimes smeared dose over an organ is okay, but sometimes it s better off to sacrifice a small portion of an organ rather than give an average dose to the whole volume Serial vs parallel organs

Plan comparison and evaluation - Dose Volume Histogram The DVH is actually an Integrated Dose Volume Histogram, or cumulative DVH. Requires a) a delineated 3D volume and b) 3D dose computed over the full volume First, choose a bin size and number of bins, typically done in increments relative to the Prescribed dose (E.g. 1% of Rx dose, and computed out to Maximum dose). Let s assume you have a maximum dose of 120% or the Rx, then you would need 120 bins. Also, need to know the size of each volume element being counted (determined by the calculation matrix granularity) Then, for all Voxels inside the volume, count the number of voxels between 0% and <1%, then between 1% but < 2%, and so on. This gives a histogram of the dose inside that volume. To convert into a cumulative DVH, start at the highest dose bin (Bin 120 in our example). Then add that with the next bin down. I.e integrate from the right. So the new bin 119 is now Bin 119 + bin 120. Bin 118 is Sum (bins>=118). Each bin now represents the number of voxels in the target that have relative dose >= the relative dose. Normalize count in each bin by total number of voxels, this results in a % volume vs % dose graph. Can also convert between relative-absolute Volume and relative-absolute Dose Target volumes have an ideal shape of full dose out to the Rx line, and no dose after that. At a glance you can gauge the hot spots, cold spots and dose uniformity. OAR volumes have an ideal shape of little or no dose, and no hot spots. DVH is fast and easy to learn and understand, but DOES NOT TELL YOU WHERE THE HOT SPOTS or COLD SPOTS ARE

DVH. 1. Count the number of voxels in a given dose range Dose Bin (cgy) 1200 Voxel Count 300 323 600 284 900 104 1200 87 1500 154 1800 777 2100 1107 2400 725 2700 723 3000 974 3300 1139 3600 3 DVH of whole body (one slice) 1000 800 600 400 200 BUT- Only count if they are also inside the volume of interest 0 300 600 900 1200 1500 1800 2100 2400 2700 3000 3300 3600

DVH. 2. Accumulate all doses above the range 7000 6000 5000 4000 Cumulative DVH of whole body (one slice) Dose Bin (cgy) Voxel Count Cumulative Voxels 300 323 6400 600 284 6077 900 104 5793 1200 87 5689 1500 154 5602 1800 777 5448 2100 1107 4671 2400 725 3564 2700 723 2839 3000 974 2116 3300 1139 1142 3600 3 3 3000 2000 1000 0 300 600 900 1200 1500 1800 2100 2400 2700 3000 3300 3600 Also known as a reverse or backwards integration, since you start from high and work down

DVH.3 Normalize to 100% Volume, and 100% Rx Shorthand Notations V100 = The % Volume getting 100% Rx (33% here) D100 = The % Dose where 100% of the volume is being treated (10% here). V50 is?? D95 is?? 120 100 80 60 40 Normalized Cumulative DVH of whole body (one slice) 20 0 10 20 30 40 50 60 70 80 90 100 110 120

DVH4. Absolute Volume can be useful E.g. How many CC are getting 50% of the Rx dose?

Other Statistics usually computed with DVH Neat trick: Assume 1 cc = 1g, then, since we know that Gy = J/kg Total integral dose to whole body (J) = Average dose (Gy) * Mass (kg) Shown here: Volume 19.27 kg, Mean Dose = 0.061 Gy, ID= 1.17 J

Ideal Dose Constraints Want to push Target up at Rx line % Volume Covered 100 80 60 Rectum Bladder Prostate Want all OAR to have be pushed low here 40 20 0 0.00 0.25 0.50 0.75 1.00 1.25 Dose Want to push Target dose above RX line down

Typical DVH Results % Volume Covered 100 80 60 Rectum Bladder Prostate 40 20 0 0.00 0.25 0.50 0.75 1.00 1.25 Dose

Lung SBRT; 5 x 1000 cgy

Same Graph, Absolute Dose and Volume

Plan normalization % IDL or other Global value DVH metrics, i.e. 100%Rx covers 95% PTV Target or Whole Body min, Mean, Max Single Point or location All methods effectively just heat the plan up (more MUs) or cool the plan down (less MUs) But they are convenient methods to tweak a plan without having to recalculate each field Important to understand how your TPS works by multiplication or by division