1
Electron therapyElectron therapyClass 2: Review questionsClass 2: Review questions
2
Raphex Question: T63, 2002
• In what situation is electron backscatter likely to be a problem?
A. Using 1cm of tissue equivalent bolus on the skin.B. Using a lead intra-oral shield to protect the tongue
when treating a cheek.C. When electrons are incident at an angel greater than
30o.D. When electrons of different energies are matched on
the skin.
•• In what situation is electron backscatter likely to In what situation is electron backscatter likely to be a problem?be a problem?
A.A. Using 1cm of tissue equivalent bolus on the skin.Using 1cm of tissue equivalent bolus on the skin.B.B. Using a lead intraUsing a lead intra--oral shield to protect the tongue oral shield to protect the tongue
when treating a cheek.when treating a cheek.C.C. When electrons are incident at an angel greater than When electrons are incident at an angel greater than
3030oo..D.D. When electrons of different energies are matched on When electrons of different energies are matched on
the skin.the skin.
3
Raphex Question: T63, 2000
• The %DD at 5 cm for a 6MeV electron beam is approximately:
A. 100%B. 90%C. 80%D. 50%E. <5%
•• The %DD at 5 cm for a 6MeV electron beam is The %DD at 5 cm for a 6MeV electron beam is approximately:approximately:
A.A. 100%100%B.B. 90%90%C.C. 80%80%D.D. 50%50%E.E. <5%<5%
4
Raphex Question: T55, 1999
• The internal mammary nodes (IMN) can be included in wide tangents or treated with a direct beam. In the latter case, a combination of photons and electrons can be used. The reason for adding photons, instead of using electrons alone, is:
A. To reduce dose to underlying lung.B. To keep cord dose below tolerance.C. To reduce the skin dose.D. All of the above.
•• The internal mammary nodes (IMN) can be The internal mammary nodes (IMN) can be included in wide tangents or treated with a direct included in wide tangents or treated with a direct beam. In the latter case, a combination of photons beam. In the latter case, a combination of photons and electrons can be used. The reason for adding and electrons can be used. The reason for adding photons, instead of using electrons alone, is:photons, instead of using electrons alone, is:
A.A. To reduce dose to underlying lung.To reduce dose to underlying lung.B.B. To keep cord dose below tolerance.To keep cord dose below tolerance.C.C. To reduce the skin dose.To reduce the skin dose.D.D. All of the above.All of the above.
5
Raphex Question: T62, 2002
• A lateral neck tumor is treated with a single direct electron field. The max depth of the treatment volume is 1.5cm, and the minimum cord depth is 5.0cm. Two techniques are considered: 6MeV electrons 9MeV electrons with 0.5cm bolus
Which of the following is TRUE?A. The cord dose will be at least 50% of prescribed dose in either
case.B. The surface dose will be higher with the 6MeV Beam.C. The dose falloff beyond the tumor will be steeper with 6MeVD. The tumor would be underdosed with 9MeV
•• A lateral neck tumor is treated with a single direct electron A lateral neck tumor is treated with a single direct electron field. The max depth of the treatment volume is 1.5cm, field. The max depth of the treatment volume is 1.5cm, and the minimum cord depth is 5.0cm. Two techniques and the minimum cord depth is 5.0cm. Two techniques are considered:are considered: 6MeV electrons6MeV electrons 9MeV electrons with 0.5cm bolus9MeV electrons with 0.5cm bolus
Which of the following is TRUE?Which of the following is TRUE?A.A. The cord dose will be at least 50% of prescribed dose in either The cord dose will be at least 50% of prescribed dose in either
case.case.B.B. The surface dose will be higher with the 6MeV Beam.The surface dose will be higher with the 6MeV Beam.C.C. The dose falloff beyond the tumor will be steeper with 6MeVThe dose falloff beyond the tumor will be steeper with 6MeVD.D. The tumor would be The tumor would be underdosedunderdosed with 9MeVwith 9MeV
6
Raphex Question: T64, 2000
• Which of the following properties of electron beams are true?
1. The range in tissue in cm is about is about ½ the beam energy in MeV.
2. The distance between the 90% and 20% isodose levels on the axis increases with increasing energy.
3. The width of the 90% isodose decreases with depth.4. As energy increases, skin dose decreases.
A. 1,2,3B. 1,3C. 4 onlyD. All of the above
• Which of the following properties of electron beams are true?
1. The range in tissue in cm is about is about ½ the beam energy in MeV.
2. The distance between the 90% and 20% isodose levels on the axis increases with increasing energy.
3. The width of the 90% isodose decreases with depth.4. As energy increases, skin dose decreases.
A. 1,2,3B. 1,3C. 4 onlyD. All of the above
7
Raphex Question: T46, 2001
• Compared with 6MeV electrons, 16MeV electrons have:
1. A greater surface dose.2. A lower bremsstrahlung tail.3. A broader plateau region.4. A sharper fall-off between 80% and 20% isodose levels
A. 1,3B. 2,4C. 1,2,3D. 4 onlyE. 1,2,3,4
•• Compared with 6MeV electrons, 16MeV electrons have:Compared with 6MeV electrons, 16MeV electrons have:
1.1. A greater surface dose.A greater surface dose.2.2. A lower A lower bremsstrahlungbremsstrahlung tail.tail.3.3. A broader plateau region.A broader plateau region.4.4. A sharper fallA sharper fall--off between 80% and 20% off between 80% and 20% isodoseisodose levelslevels
A.A. 1,31,3B.B. 2,42,4C.C. 1,2,31,2,3D.D. 4 only4 onlyE.E. 1,2,3,41,2,3,4
8
Raphex Question: T57, 2002
• A 12 MeV electron beam has a range of____cm, and a 90% DD at approximately_____cm.
A. 6,4B. 12,4C. 12,6D. 4,3E. 9,6
•• A 12 A 12 MeVMeV electron beam has a range electron beam has a range of____cmof____cm, , and a 90% DD at and a 90% DD at approximately_____cmapproximately_____cm..
A.A. 6,46,4B.B. 12,412,4C.C. 12,612,6D.D. 4,34,3E.E. 9,69,6
9
Raphex Question: T69, 2003
• When a custom electron insert has dimensions smaller than the range of the electrons, all of the following are likely to occur except:
A. The output (cGy/MU) will be reducedB. Surface dose will decrease.C. PDD will decrease beyond dmax.D. Dmax will shift toward a shallower depth.
•• When a custom electron insert has dimensions When a custom electron insert has dimensions smaller than the range of the electrons, all of smaller than the range of the electrons, all of the following are likely to occur except:the following are likely to occur except:
A.A. The output (The output (cGycGy/MU) will be reduced/MU) will be reducedB.B. Surface dose will decrease.Surface dose will decrease.C.C. PDD will decrease beyond PDD will decrease beyond ddmaxmax..D.D. DmaxDmax will shift toward a shallower depth.will shift toward a shallower depth.
10
Electron therapyElectron therapyClass 2: Fundamentals Class 2: Fundamentals
continuedcontinued
11
Field flatness and symmetry
• Plane perpendicular to beam axis• Uniformity index (ICRU): Adose>90% /
Ageom
• Specifies dose<103% of CAX value• Depth: e.g. 95% beyond dmax (AAPM
TG#25)• Area: 2cm inside geometrid edge of fields
larger than 10x10cm• ± 5% (optimally be within ± 3%) TG#25• Symmetry: compares dose profile on one
side of central axis to the other. Max 2% variation (AAPM)
• Flatness and symmetry change: <1% of baseline
•• Plane perpendicular to beam axisPlane perpendicular to beam axis•• Uniformity index (ICRU): Uniformity index (ICRU): AAdosedose>90%>90% / /
AAgeomgeom
•• Specifies dose<103% of CAX valueSpecifies dose<103% of CAX value•• Depth: e.g. 95% beyond Depth: e.g. 95% beyond dmaxdmax (AAPM (AAPM
TG#25)TG#25)• Area: 2cm inside geometrid edge of fields
larger than 10x10cm• ± 5% (optimally be within ± 3%) TG#25• Symmetry: compares dose profile on one
side of central axis to the other. Max 2% variation (AAPM)
• Flatness and symmetry change: <1% of baseline
12
MU calculation
ODDMU
%
EOFSCSEOFFSCSEO cal ,,,,
D Prescribed dose%D Prescribed %O Dose output at R100
Output depends on energy, applicator, field size, SSD and skin collimation.
First consider different cones and field sizes:
Note: All at R100, which is dependent of field size and energy
EOFSCSEOFDDMU
cal
,,%
(more on MU calcs later)
13
Skin blocking
• Useful for:• Small fields• Critical structures close to field• Need to restore beam penumbra (e.g. extended SSD, arc therapy)
14
Effect of skin collimation on dosimetry
• PDD: – mostly scatter in patient
– Use field size determined by skin collimation
• Output factor:– Mostly scatter in air
– Use field size determined by secondary collimator (e.g. cerrobend cutout)
•• PDD: PDD: –– mostly scatter in patientmostly scatter in patient
–– Use field size determined by skin collimationUse field size determined by skin collimation
•• Output factor:Output factor:–– Mostly scatter in airMostly scatter in air
–– Use field size determined by secondary Use field size determined by secondary collimator (e.g. collimator (e.g. cerrobendcerrobend cutout)cutout)
15
Field equivalence
• For rectangular fields:
• (Note: neglects collimator scatter)
• Equivalent circular fields (field width=2a):
•• For rectangular fields:For rectangular fields:
•• (Note: neglects collimator scatter)(Note: neglects collimator scatter)
•• Equivalent circular fields (field width=2a):Equivalent circular fields (field width=2a):
16
WxWLxLLxW OFOFOF
17
Calculating PDD for non-square fields
• Depth dose can be determined from data from square fields:
• Need to renormalize for 100% at dmax for new PDD
• Can use same formula for output calculations
•• Depth dose can be determined from data Depth dose can be determined from data from square fields:from square fields:
•• Need to renormalize for 100% at Need to renormalize for 100% at dmaxdmax for for new PDDnew PDD
•• Can use same formula for output Can use same formula for output calculationscalculations
18
A possible scenario:
• Skin collimation 2x5
• Cutout in cone: 4x6
• Energy: 6MeV
• What determines output factor?
• What determines PDD?
•• Skin collimation 2x5Skin collimation 2x5
•• Cutout in cone: 4x6Cutout in cone: 4x6
•• Energy: 6MeVEnergy: 6MeV
•• What determines output factor?What determines output factor?
•• What determines PDD?What determines PDD?
19
Raphex Question: T61, 2002
•• An electron beam with a custom insert has a An electron beam with a custom insert has a measured OF measured OF ofof 0.954cGy/MU at 0.954cGy/MU at ddmaxmax. If 200cGy . If 200cGy are prescribed to the 90% are prescribed to the 90% isodoseisodose, the MU setting , the MU setting is ___.is ___.
A.A. 117117B.B. 210210C.C. 212212D.D. 222222E.E. 233233
20
Bonus question
•• The patient is to be treated to 200cGy, prescribed to the 80% The patient is to be treated to 200cGy, prescribed to the 80% isodoseisodose level.level.
•• The calibrated output for the 15 cone is 1cGy/MU at The calibrated output for the 15 cone is 1cGy/MU at ddmaxmax
•• The field size is 4x12cmThe field size is 4x12cm22 (15cone). (15cone). •• The output factor for a 4x4 field in the 15 cone is 0.954The output factor for a 4x4 field in the 15 cone is 0.954•• The output factor for a 12x12 field in the 15 cone is 0.997The output factor for a 12x12 field in the 15 cone is 0.997•• The MU setting is____.The MU setting is____.
A.A. 176176B.B. 184184C.C. 218218D.D. 250250E.E. 256256
21
Another bonus question
• We decide to add 1cm bolus to a 6MeV electron field. Which of the following is not true (if we do not change the MUs)?
a) The position of dmax moves upstream by 1cmb) The dose at 2cm depth in the tissue goes
downc) The skin dose increases to almost dmax
d) The maximum dose increases by 2%e) The maximum dose decreases by 2%
•• We decide to add 1cm bolus to a 6MeV We decide to add 1cm bolus to a 6MeV electron field. Which of the following is electron field. Which of the following is not true (if we do not change the not true (if we do not change the MUsMUs)?)?
a)a) The position of The position of dmaxdmax moves upstream by 1cmmoves upstream by 1cmb)b) The dose at 2cm depth in the tissue goes The dose at 2cm depth in the tissue goes
downdownc)c) The skin dose increases to almost The skin dose increases to almost ddmaxmax
d)d) The maximum dose increases by 2%The maximum dose increases by 2%e)e) The maximum dose decreases by 2%The maximum dose decreases by 2%
22
Raphex Question: T58, 1999
• Output for an electron cones depends on:
1. Cone size.2. Size of cut-out.3. Beam Energy.4. SSD
A. 1,2,3B. 1,3C. 2,4D. 4 onlyE. All of the above.
• Correct Answer: E. Output for electron beam depends on ALL of these factors.
•• Output for an electron cones depends on:Output for an electron cones depends on:
1.1. Cone size.Cone size.2.2. Size of cutSize of cut--out.out.3.3. Beam Energy.Beam Energy.4.4. SSDSSD
A.A. 1,2,31,2,3B.B. 1,31,3C.C. 2,42,4D.D. 4 only4 onlyE.E. All of the above.All of the above.
•• Correct Answer:Correct Answer: E. Output for electron beam depends E. Output for electron beam depends on ALL of these factors.on ALL of these factors.
23
Electron therapyClass 2: Algorithms
Acknowledgements• Many images, data and many slides are from Richard Popple, Ph.D., Department of Radiation Oncology, The University of Alabama at Birmingham, Birmingham, Alabama• Paul Yokoyama (Varian) was also very helpful
AcknowledgementsAcknowledgements•• Many images, data and many slides are from Richard Many images, data and many slides are from Richard PopplePopple, , Ph.D., Department of Radiation Oncology, The University of Ph.D., Department of Radiation Oncology, The University of Alabama at Birmingham, Birmingham, AlabamaAlabama at Birmingham, Birmingham, Alabama•• Paul Yokoyama (Varian) was also very helpfulPaul Yokoyama (Varian) was also very helpful
Laurence Court
24
Pencil beam algorithm
25
Schematic of Hogstrom algorithm
26
Heterogeneity correction in original Hogstrom pencil-beam algorithm
Slab approach to heterogeneity correction
27
Redefinition pencil-beam algorithm
28
Experimental evaluation
• Dose measured using watertank and diode• Dose measured using watertank and diode
Boyd et al, Med Phys 28, 950-958, 2001
29
20-MeV Horizontal Bone Slab
The University of Texas M. D. Anderson Cancer Center Department of Radiation Physics
Varian Clinac 2100, 15x15Varian Clinac 2100, 15x15--cmcm22 open applicator, 100 cm SSDopen applicator, 100 cm SSD
Off-Axis Position (cm)-10 -8 -6 -4 -2 0 2 4 6 8 10
Dep
th (c
m)
0123456789
101112
MeasuredPBRA
10%
100%
Criteria not met
30
Off-Axis Position (cm)-10 -8 -6 -4 -2 0 2 4 6 8 10
Dep
th (c
m)
0123456789
101112
MeasuredPBRA
10%
100%
Criteria not met
20-MeV Horizontal Air Slab
The University of Texas M. D. Anderson Cancer Center Department of Radiation Physics
Varian Clinac 2100, 15x15Varian Clinac 2100, 15x15--cmcm22 open applicator, 100 cm SSDopen applicator, 100 cm SSD
31
Parotid Gland - Transverse View
The University of Texas M. D. Anderson Cancer Center Department of Radiation Physics
Varian 2100, 16 MeV, 15x15Varian 2100, 16 MeV, 15x15--cmcm22 applicator, 100 cm SSDapplicator, 100 cm SSD
32
Parotid Gland - Transverse View
The University of Texas M. D. Anderson Cancer Center Department of Radiation Physics
Varian 2100, 16 MeV, 15x15Varian 2100, 16 MeV, 15x15--cmcm22 applicator, 100 cm SSDapplicator, 100 cm SSD
X Axis
-16 -14 -12 -10 -8 -6 -4 -2 0 2 4 6
Y A
xis
-4
-2
0
2
4
6
8
10
X Axis
-16 -14 -12 -10 -8 -6 -4 -2 0 2 4 6
Y A
xis
-4
-2
0
2
4
6
8
10
Monte CarloPBRA
10%
90%
AIR
LUNGTISSUE
BONE
33
IMC - Transverse Plane
Varian 2100, 16 MeV, 15x15-cm2 applicator, 105 cm SSDVarian 2100, 16 MeV, 15x15Varian 2100, 16 MeV, 15x15--cmcm22 applicator, 105 cm SSDapplicator, 105 cm SSD
The University of Texas M. D. Anderson Cancer Center Department of Radiation Physics
34
IMC - Transverse Plane
Varian 2100, 16 MeV, 15x15-cm2 applicator, 105 cm SSDVarian 2100, 16 MeV, 15x15Varian 2100, 16 MeV, 15x15--cmcm22 applicator, 105 cm SSDapplicator, 105 cm SSD
The University of Texas M. D. Anderson Cancer Center Department of Radiation Physics
AIR
LUNGTISSUE
BONE
X Axis
-4 -2 0 2 4 6 8 10
Y Ax
is
4
6
8
10
12
14
16
X Axis
-4 -2 0 2 4 6 8 10
Y Ax
is
4
6
8
10
12
14
16
Monte CarloPBRA
10%
50%
100%
35
Electron Monte Carlo dose calculations
• Accurate electron dose calculations (distributions) are very difficult
• Monte Carlo has potential to be most accurate method
• MC is very calculation intensive:– 2% SD at peak dose for 10x10cm field, 0.5cm
resolution requires 1,000,000 histories.
• Macro Monte Carlo (MMC) significantly reduces calculation time
•• Accurate electron dose calculations Accurate electron dose calculations (distributions) are very difficult(distributions) are very difficult
•• Monte Carlo has potential to be most Monte Carlo has potential to be most accurate methodaccurate method
•• MC is very calculation intensive:MC is very calculation intensive:–– 2% SD at peak dose for 10x10cm field, 0.5cm 2% SD at peak dose for 10x10cm field, 0.5cm
resolution requires 1,000,000 histories.resolution requires 1,000,000 histories.
•• Macro Monte Carlo (MMC) significantly Macro Monte Carlo (MMC) significantly reduces calculation timereduces calculation time
36
Primary Particle Transport – PDFs
‘Kugel’(!)
Precalculate PDFs for various Kugel sizes, materials
37
Primary Particle Transport
38
Eclipse implementation of MMC
1. Initial phase space model
2. Local simulation
3. Geometric pre-processing
4. Global simulation
1.1. Initial phase space modelInitial phase space model
2.2. Local simulationLocal simulation
3.3. Geometric preGeometric pre--processingprocessing
4.4. Global simulationGlobal simulation
39
(1) Initial Phase Space Model
• Four-source model
• Compute probability distributions of position, energy, direction of electrons and photons for each source, specificed at PSP
•• FourFour--source modelsource model
•• Compute probability Compute probability distributions of position, distributions of position, energy, direction of energy, direction of electrons and photons for electrons and photons for each source, each source, specificedspecificed at at PSPPSP
40
(2) Local Simulation
• Over 200,000 electrons transported through uniform spheres:– 0.05cm, 0.1cm, 0.15cm, 0.2cm, 0.3cm
– Air, lung, water, bone
– 0.2MeV – 25MeV (25 electron energies)
• Gives PDF representing direction and energy of exiting particles, and also secondary particules
•• Over 200,000 electrons transported through Over 200,000 electrons transported through uniform spheres:uniform spheres:–– 0.05cm, 0.1cm, 0.15cm, 0.2cm, 0.3cm0.05cm, 0.1cm, 0.15cm, 0.2cm, 0.3cm
–– Air, lung, water, boneAir, lung, water, bone
–– 0.2MeV 0.2MeV –– 25MeV (25 electron energies)25MeV (25 electron energies)
•• Gives PDF representing direction and Gives PDF representing direction and energy of exiting particles, and also energy of exiting particles, and also secondary secondary particulesparticules
41
(3) Geometric pre-processing
Pre-processing:
• CT# to density and material conversion (user-defined resolution)
• Identify heterogeneities, and assign ‘krugel index’
• Calculate mean ‘krugel density’
PrePre--processing:processing:
•• CT# to density and material CT# to density and material conversion (userconversion (user--defined defined resolution)resolution)
•• Identify heterogeneities, and Identify heterogeneities, and assign assign ‘‘krugelkrugel indexindex’’
•• Calculate mean Calculate mean ‘‘krugelkrugel densitydensity’’
42
43
(4) Global simulation
• Position, direction ,energy of particles exiting sphere (‘kugel’!) determined by random sampling of PDF for appropriate sphere (size, material, energy)• Dose deposited along straight line inside kugel
• Secondary particules forced to interact in sphere• Always transported with 20deg angle relative to incident electron
44
Calculation parameters – user interaction
45
Calculation parameters
46
Evaluation data set (dose distributions, impact of heterogeneities)
• R. A. Boyd, K. R. Hogstrom, J. A. Antolak, and A. S. Shiu, “A measured data set for evaluating electron beam dose algorithms.”Med. Phys. 28, 950–958 (2001).
• Set of measurements developed and published specifically for evaluating electron dose calculation algorithms.
•• R. A. Boyd, K. R. R. A. Boyd, K. R. HogstromHogstrom, J. A. , J. A. AntolakAntolak, , and A. S. and A. S. ShiuShiu, , ““A measured data set for A measured data set for evaluating electron beam dose algorithms.evaluating electron beam dose algorithms.””Med. Phys. 28, 950Med. Phys. 28, 950––958 (2001).958 (2001).
•• Set of measurements developed and Set of measurements developed and published specifically for evaluating electron published specifically for evaluating electron dose calculation algorithms.dose calculation algorithms.
47
Phantom geometries
48
Measurement configurations
49
Evaluation data set
• The measured data were normalized to the maximum dose on central axis in a water phantom (without inhomogeneity) at 100-cm SSD.
• All data were for a 15 x 15 cm2 field and were acquired using a single Varian Clinac 2100C accelerator (Varian Oncology Systems, Milpitas, CA) having a Series III electron foil/applicator set.
• 9MeV, 20MeV
•• The measured data were normalized to the The measured data were normalized to the maximum dose on central axis in a water phantom maximum dose on central axis in a water phantom (without (without inhomogeneityinhomogeneity) at 100) at 100--cm SSD.cm SSD.
•• All data were for a 15 x 15 cm2 field and were All data were for a 15 x 15 cm2 field and were acquired using a single Varian acquired using a single Varian ClinacClinac 2100C 2100C accelerator (Varian Oncology Systems, Milpitas, accelerator (Varian Oncology Systems, Milpitas, CA) having a Series III electron foil/applicator set.CA) having a Series III electron foil/applicator set.
•• 9MeV, 20MeV9MeV, 20MeV
50
Configuration / commissioning
• Data set of Boyd et al. not sufficient to configure Eclipse MMC engine.
• Depth dose data are present (configurations 1 and 3).
• Not present in data set:– CAX fractional depth dose without applicator
– In-air profile without applicator
– Absolute output
•• Data set of Boyd et al. not sufficient to Data set of Boyd et al. not sufficient to configure Eclipse MMC engine.configure Eclipse MMC engine.
•• Depth dose data are present (configurations Depth dose data are present (configurations 1 and 3).1 and 3).
•• Not present in data set:Not present in data set:–– CAX fractional depth dose without applicatorCAX fractional depth dose without applicator
–– InIn--air profile without applicatorair profile without applicator
–– Absolute outputAbsolute output
51
Evaluation method
• For each configuration/calculation parameter combination, the dose difference and distance-to-agreement were computed.
• The dose difference was determined by interpolating the measured dose at the calculation points using bilinear interpolation.
•• For each configuration/calculation For each configuration/calculation parameter combination, the dose difference parameter combination, the dose difference and distanceand distance--toto--agreement were computed.agreement were computed.
•• The dose difference was determined by The dose difference was determined by interpolating the measured dose at the interpolating the measured dose at the calculation points using bilinear calculation points using bilinear interpolation.interpolation.
52
Results - Accuracy
• Calculations performed for all possible values of accuracy parameter (1%, 2%, 3%, 5%, and 8%).
• 1 mm and 2.5 mm grid spacing for the 9 MeV and 20 MeV configurations, respectively.
• No smoothing.
•• Calculations performed for all possible Calculations performed for all possible values of accuracy parameter (1%, 2%, 3%, values of accuracy parameter (1%, 2%, 3%, 5%, and 8%).5%, and 8%).
•• 1 mm and 2.5 mm grid spacing for the 9 1 mm and 2.5 mm grid spacing for the 9 MeVMeV and 20 and 20 MeVMeV configurations, configurations, respectively.respectively.
•• No smoothing.No smoothing.
53
Results - Accuracy
52.6%65.8%8.9%8%33.9%48.8%5.2%5%17.7%29.8%3.2%3%9.9%19.7%2.4%2%3.1%9.8%1.9%1%2039.3%72.4%10.5%8%22.8%54.5%5.9%5%12.4%37.0%3.7%3%6.6%25.6%3.0%2%1.5%16.5%2.4%1%9
Fraction failing 3% difference and 3 mm distance-to-
agreement
Fraction failing 3% difference
RMS differenceAccuracyEnergy (MeV)
54
20 MeV, 100 cm SSD (Configuration 3)1% Accuracy
Isodose lines are 10%, 30%, 70%, 90%, and 95%. Calculation points violating 3% difference and 3 mm distance-to-agreement are shown in gray.
IsodoseIsodose lines are 10%, 30%, 70%, 90%, and 95%. Calculation lines are 10%, 30%, 70%, 90%, and 95%. Calculation points violating 3% difference and 3 mm distancepoints violating 3% difference and 3 mm distance--toto--agreement are shown in gray.agreement are shown in gray.
x (mm)
Dep
th (m
m)
-100 -50 0 50 100
0
20
40
60
80
100
120
55
20 MeV, 100 cm SSD (Configuration 3)3% Accuracy
x (mm)
Dep
th (m
m)
-100 -50 0 50 100
0
20
40
60
80
100
120
56
20 MeV, nose-shape (Configuration 14)3% Accuracy
x (mm)
Dep
th (m
m)
-100 -50 0 50 100
0
20
40
60
80
100
57
Results – Grid spacing
• Calculations performed for all possible values of grid spacing (1 mm, 1.5 mm, 2 mm, 2.5 mm, and 5 mm).
• 1% accuracy.
• No smoothing.
•• Calculations performed for all possible Calculations performed for all possible values of grid spacing (1 mm, 1.5 mm, 2 values of grid spacing (1 mm, 1.5 mm, 2 mm, 2.5 mm, and 5 mm).mm, 2.5 mm, and 5 mm).
•• 1% accuracy.1% accuracy.
•• No smoothing.No smoothing.
58
Results – Grid spacing
3.4%21.8%13.5%5.0
3.1%21.5%9.8%2.5
2.9%21.1%9.3%2.0
1.5%20.2%4.6%1.5
1.6%19.6%4.5%1.0206.6%9.3%47.6%5.0
2.9%6.1%23.7%2.5
1.9%5.8%18.3%2.0
2.2%6.2%19.1%1.5
1.5%4.9%16.5%1.09
Fraction failing 3% difference
and 3 mm distance-to-agreement
Fraction failing 3 mmDistance-to-agreement
Fraction failing 3% difference
Grid spacing (mm)Energy (MeV)
59
9 MeV, 2 cm stepped surface (Configuration 11)1 mm grid spacing
x (mm)
Dep
th (m
m)
-100 -50 0 50 100
0
50
60
9 MeV, 2 cm stepped surface (Configuration 11)5 mm grid spacing
x (mm)
Dep
th (m
m)
-100 -50 0 50 100
0
50
61
Results – smoothing
• Calculations performed for all possible values of 3-D smoothing (Low, Medium, and Strong).
• 1% accuracy.• Grid spacing
– 1 mm, 1.5 mm, 2 mm, and 2.5 mm for the 9 MeVconfigurations.
– 1.5 mm, 2 mm, 2.5 mm, and 5 mm for the 20 MeV configurations.
•• Calculations performed for all possible Calculations performed for all possible values of 3values of 3--D smoothing (Low, Medium, D smoothing (Low, Medium, and Strong).and Strong).
•• 1% accuracy.1% accuracy.•• Grid spacingGrid spacing
–– 1 mm, 1.5 mm, 2 mm, and 2.5 mm for the 9 1 mm, 1.5 mm, 2 mm, and 2.5 mm for the 9 MeVMeVconfigurations.configurations.
–– 1.5 mm, 2 mm, 2.5 mm, and 5 mm for the 20 1.5 mm, 2 mm, 2.5 mm, and 5 mm for the 20 MeVMeV configurations.configurations.
62
Results – smoothing
12.5%3.4%1.1%3.4%5.0
0.6%0.5%1.3%3.1%2.5
0.4%0.3%1.5%2.9%2.0
0.3%0.5%2.0%3.5%1.520
6.6%2.8%2.9%3.7%2.5
1.3%0.5%1.2%1.9%2.0
0.2%0.5%1.5%2.2%1.5
0.1%0.2%0.7%1.5%1.09
StrongMediumLowNoneGrid spacing
(mm)Energy
Fraction failing 3% difference and 3 mm distance-to-agreement
63
20-MeV beam incident on a flat surface with a horizontal bone heterogeneity (configuration 9) No smoothing
x (mm)
Dep
th (m
m)
-100 -50 0 50 100
0
20
40
60
80
100
64
20-MeV beam incident on a flat surface with a horizontal bone heterogeneity (configuration 9) 3D Medium smoothing
x (mm)
Dep
th (m
m)
-100 -50 0 50 100
0
20
40
60
80
100
65
Accuracy in presence of lung
• Experimental data from University of Bern (Switzerland) or Tampere University Hospital (Finland)
• 1cm perspex over lung-equivalent material
• 6MeV Varian beam
•• Experimental data Experimental data from University of Bern from University of Bern (Switzerland) or (Switzerland) or Tampere University Tampere University Hospital (Finland)Hospital (Finland)
•• 1cm 1cm perspexperspex over lungover lung--equivalent materialequivalent material
•• 6MeV Varian beam6MeV Varian beam
66
67
Conclusion (dose distributions)
• For judicious choices of parameters, dose calculations agree with expt. to better than 3% dose difference and 3-mm distance-to-agreement.
• Lung dose is less accurate.
• Note: Varian have also shown good results for oblique incidence
• The Eclipse MMC implementation is limited to– 2.5% and 2.5 mm for 9 MeV
– 3% and 3 mm for 20 MeV
•• For judicious choices of parameters, dose For judicious choices of parameters, dose calculations agree with expt. to better than 3% calculations agree with expt. to better than 3% dose difference and 3dose difference and 3--mm distancemm distance--toto--agreement.agreement.
•• Lung dose is less accurate.Lung dose is less accurate.
•• Note: Varian have also shown good results for Note: Varian have also shown good results for oblique incidenceoblique incidence
•• The Eclipse MMC implementation is limited toThe Eclipse MMC implementation is limited to–– 2.5% and 2.5 mm for 9 2.5% and 2.5 mm for 9 MeVMeV
–– 3% and 3 mm for 20 3% and 3 mm for 20 MeVMeV
68
Evaluation of output factors for range of clinical fields
• Data above only for simple field shapes •Comparison of measured and calculated outputs• (output includes shape and SSD differences)
69
Eclipse MMC Calculation Parameters
• Accuracy (mean statistical error in dose within the high dose volume) : 1%.
• Grid spacing:
• No smoothing
•• Accuracy (mean statistical error in dose within Accuracy (mean statistical error in dose within the high dose volume) : 1%.the high dose volume) : 1%.
•• Grid spacing:Grid spacing:
•• No smoothingNo smoothing2.5 mm2.5 mm18 18 MeVMeV
2.5 mm2.5 mm15 15 MeVMeV
2 mm2 mm12 12 MeVMeV
1.5 mm1.5 mm9 9 MeVMeV
1 mm1 mm6 6 MeVMeV
Grid spacingGrid spacingEnergyEnergy
70
Scatter plot of calculated output factor versus measured output factor
The solid line shows calculation equal to measurement, and the dashed lines show 3% deviation.
The solid line shows calculation equal to measurement, and the dThe solid line shows calculation equal to measurement, and the dashed lines ashed lines show 3% deviation.show 3% deviation.
0.8 0.85 0.9 0.95 1 1.05 1.10.8
0.85
0.9
0.95
1
1.05
1.1
Measured output factor
Ecl
ipse
MM
C o
utpu
t fac
tor
71
Histogram of difference distribution
-5 -4 -3 -2 -1 0 1 2 3 4 50
10
20
30
40Mean difference: -0.23%Std. deviation: 1.09%
Percent difference
Num
ber o
f ins
erts
72
Conclusion (output factors)
• The mean difference between the Eclipse MMC calculation and measurement was 0.2%. The standard deviation of the difference distribution was 1.1%.
• Of the 218 measurements,– 124 (56.9%) were within 1%. – 211 (96.8%) were within 2%. – 215 (98.6%) were within 3%.
• The largest difference was 3.4%.
•• The mean difference between the Eclipse The mean difference between the Eclipse MMC calculation and measurement was 0.2%. MMC calculation and measurement was 0.2%. The standard deviation of the difference The standard deviation of the difference distribution was 1.1%. distribution was 1.1%.
•• Of the 218 measurements,Of the 218 measurements,–– 124 (56.9%) were within 1%. 124 (56.9%) were within 1%. –– 211 (96.8%) were within 2%. 211 (96.8%) were within 2%. –– 215 (98.6%) were within 3%.215 (98.6%) were within 3%.
•• The largest difference was 3.4%.The largest difference was 3.4%.
73
74
Treatment planning algorithms: Summary
• Pencil beam
• Modified Monte Carlo
• Expect MU to be accurate within ~3%
• Relative isodoses should be accurate for most scenarios, but less reliable for lung
• Algorithms do not include skin collimation, customized bolus or electron arc therapy
•• Pencil beamPencil beam
•• Modified Monte CarloModified Monte Carlo
•• Expect MU to be accurate within ~3%Expect MU to be accurate within ~3%
•• Relative Relative isodosesisodoses should be accurate for should be accurate for most scenarios, but less reliable for lungmost scenarios, but less reliable for lung
•• Algorithms do not include skin collimation, Algorithms do not include skin collimation, customized bolus or electron arc therapycustomized bolus or electron arc therapy