04 MRTdosimetryWorkshop Calvert and Tran-Gia SM

Quantitative Imaging and Dose
Calculation Comparison Exercises in

the MRTDosimetry Project
J Tran-Gia1, M Lassmann1, J Tipping2, N Calvert2
1Department of Nuclear Medicine, University of Würzburg
2Nuclear Medicine Group, The Christie NHS Foundation Trust, Manchester
Aims of Comparison Exercises
1. Need for standardization of quantitative imaging

– Problem: Numerous imaging modalities & systems.
No standardization of activity determination.
– Plan: Comparison exercise with different SPECT/CT systems
BUT standardized activities & reconstructions.
– Aim: Protocol for commissioning and QI.
2. Lack of validation between dosimetry software tools

– Problem: Numerous tools available for dose calculation.
No validation across different tools.
– Plan: Calculate doses with available dosimetry tools based
on simulated post-therapy dosimetry dataset.
+ many in-house
– Aim: Protocol for commissioning dosimetry platforms. approaches
2
Participating Systems
SPECT/CT Hardware available at Participating Sites:

AOSP GE Discovery 690 2016 3/8"
ASMN Siemens Symbia T2 2009 3/8"
Christie 1 GE Discovery 670 2015 3/8"
Christie 2 GE Infinia Hawkeye4 2006 3/8"
LUND GE Discovery 670 2012 5/8"
NPL Mediso AnyScan SCP 2017 3/8"
OUHT 1 GE Discovery 670 2011 3/8"
OUHT 2 GE Infinia Hawkeye4 2008 3/8"
RSCH 1 GE Optima NM/CT 640 2014 3/8"
THG (PGNI) GE Optima NM/CT 640 2015 3/8"
UKW 1 Siemens Symbia T2 2006 5/8"
UKW 2 Siemens Symbia Intevo 2016 3/8"
• 5 Countries.
• SPECT/CT systems of 3 large vendors.
• 2 different detector crystal thicknesses.
3
Outline
Quantitative Imaging Comparison Exercise (UKW)

1. Sensitivity Determination (Jaszczak Cylinder)
2. Partial Volume Assessment (6-sphere Phantom)
3. Quantitative Imaging Exercise (3D-printed 2-organ Phantom)
4. Validation of Radioactive Test Sources (Ba-133 Surrogate vs. Liquid I-131)
Dose Comparison Exercise (The Christie)

1. Extend Printed Phantom (4-Organ 3D-printed phantom)
2. Generation of Ground Truth TAC & Dose (PK Model & MC Simulation)
3. Measurement of TAC (Experimental & MC Simulation)
4. Dose Calculation & Comparison (Commercial & in-house)
4
Outline

DosePart
Comparison
1&2 Exercise (The Christie)
Part 3 Part 4
Sensitivity
1. Extend Printed Phantom 3D-printed ICRP (4-Organ 3D-printed
Sealed Ba-133
phantom)
Partial Volume 2-organ phantom vs. liquid I-131
Ba-133 Liquid I-131
5
1.1 Sensitivity Determination
Goal Reconstruction
Counts in
each voxel
Conversion according to
sensitivity (cps/MBq[1])
?
Activity per voxel
Counts → Activity
→ Activity distribution
6
[1] Counts-per-second-per-MBq
Calibration protocol
(produced within the MRTDosimetry project)
7
Approach
• Fill large phantom (to reduce partial volume effects) with
traceable amount of activity.
• SPECT/CT imaging (TACQ) + clinical reconstruction.
• Calculate conversion factor: Filled
Phantom
Total counts Aphantom
SPECT
Decay-corrected Total live time
Imaging
activity ctotal
• Unit: counts-per-s-per-MBq (cps)
8
Protocol:
Suggestions on Carrier Solution
(MRTDosimetry: Only Lu-177 and I-131)
Suggestions on Phantom Selection
Suggestions on > 18 cm
Phantom Filling < 10 L
www.biodex.com
> 20 cm 9
Protocol: Suggestions on SPECT/CT Acquisition

• Positioning in center of rotation.
SPECT/CT Positioning
• Acquisition parameters as in clinical protocol

(e.g. 2×60 projections, 30s-per-projection, non-circular orbit).
• Low-dose CT for attenuation correction.
10
Protocol: Suggestions on
SPECT/CT
SPECT/CT Reconstruction Reconstruction
• Projection-wise decay correction.

• Scatter (e.g. triple-energy-window) & attenuation correction (e.g. CT-
based) as in clinical protocol.
• Filtering (e.g. Gaussian post-filtering) & resolution recovery as in clinical
protocol.
• Iterate until number of
counts has converged
as function of updates
(iterations × subsets).
11
Example from UKW (Lu-177)

SPECT Lu-177 SPECT/CT Fusion
Flash3D, 6i6s0mm
12
Protocol: VOI Analysis

• Define cylinder ROI in central transaxial slice.
• 130% of phantom radius. VOI
• 120% of phantom length.
Phantom
Calculate conversion factor
13
Goal Reconstruction
Counts per
Voxel
Conversion according to
sensitivity (cps/MBq)
?!
Activity per voxel
Counts → Activity
→ Activity distribution
14
Outline

DosePart
Comparison
Part 3 Part 4
Sensitivity
1. Your Choice ☺ 3D-printed ICRP Sealed Ba-133
Ba-133 Liquid I-131
15
1.2 Partial Volume Assessment
Partial Volume Effect
(„Spill-Out“
All spheres should ideally vs. „Spill-In“)
have homogeneous SPECT signal.
Source:
Goal Wikipedia
of this section: Assess the partial volume! Simulation
PSF
(Gaussian)
FWHM
SPECT: 1-2 cm
Image Formation:
Convolution of object and point spread function of imaging system.
(„How the imaging system see the object”).
16
Approach
• Fill sphere phantom with known geometry (known radii) with
traceable amount of activity (Asphere,i).
• SPECT/CT imaging (TACQ) + clinical reconstruction.
• Draw sphere VOIs based on nominal diameters and CT.
• Obtain ratio of spilled out counts:
𝑐𝑡𝑜𝑡𝑎𝑙 ctotal Sphere
SPECT/CT
𝑅𝐶(𝑖) = Phantom
Fusion & VOIs
𝐴𝑠𝑝ℎ𝑒𝑟𝑒,𝑖 ∙ 𝑇𝐴𝐶𝑄 ∙ 𝐶𝐹
From 1.1
SPECT
17
Optional in calibration protocol (Appendix)
http://www.biodex.com
MRTDosimetry comparison exercise

• IEC NEMA Phantom
• 6 sphere inserts (diameters: 10–37 mm, volumes: 0.5–26.5 ml)
18
Phantom Filling Procedure Activity
Carrier
(This procedure assumes that the sphere Solution
volumes Vsphere i were pre-measured)
• Prepare Stock solution (500 ml container).

500 ml Container
𝐴𝑠𝑦𝑟𝑖𝑛𝑔𝑒,𝑓𝑢𝑙𝑙 − 𝐴𝑠𝑦𝑟𝑖𝑛𝑔𝑒,𝑒𝑚𝑝𝑡𝑦
𝑐𝑠𝑡𝑜𝑐𝑘 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛 =
𝑤𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟,𝑓𝑢𝑙𝑙 − 𝑤𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟,𝑒𝑚𝑝𝑡𝑦
• Ensure spheres are empty.
• Fill spheres until liquid “rises up along capillary”.
• Sphere activities: 𝐴𝑠𝑝ℎ𝑒𝑟𝑒 𝑖 = 𝑐𝑠𝑡𝑜𝑐𝑘 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛 ∗ 𝑉𝑠𝑝ℎ𝑒𝑟𝑒 𝑖
19
Acquisition & Reconstruction as for sensitivity.

20
Example from UKW (Lu-177)

Lu-177 SPECT/CT Fusion
Flash3D, 6i6s0mm
SPECT
21
Recovery Curve Fitting

Correct partial volume by volume-based recovery coefficient
• Ratio of spilled outcorrection
(strictly: counts:restricted to sphere geometry). SPECT/CT
ctotal
𝑐𝑡𝑜𝑡𝑎𝑙 0.53 Fusion & VOIs
𝑅𝐶(𝑖) = 0.73
𝐴𝑠𝑝ℎ𝑒𝑟𝑒,𝑖 ∙ 𝑇𝐴𝐶𝑄 ∙ 𝐶𝐹
0.21 SPECT
• Plot RC(i) against sphere volume Vi 0.78
• Obtain volume-dependent recovery coefficient curve by

fitting (α, β, γ: coefficients for fit, V: sphere volume):
𝛼 0.87 0.81 RC(V) curve
𝑅𝐶(𝑉) =
1 + 𝛾 Τ𝑉 𝛽
22
Outline

DosePart
Comparison
Part 3 Part 4
Sensitivity
Ba-133 Liquid I-131

→ Validate QI in quasi-realistic,
anthropomorphic phantom
23
1.3 Quantitative Imaging Exercise
3D Printing Template:
2 Organs (Kidney & Spleen)
of ICRP 110[1] Voxel Phantom
• Average male/female.
• Segmented & modified CT of individual.
• Voxel size (1.8 x 1.8 x 4.8 mm3) places limit on
resolution of organs.
• Accepted as dosimetry standard.
[1] ICRP Publication 110. Ann ICRP 39(2), 2009. 24

Phantom Design
Convert voxels Smoothen Extrude walls, add
to surface surface filling & supports, cut
Voxels Cortex
Medulla
25
Phantom Fabrication
Print Seal
FDM 3D-Printing Printed parts Completed

(Ultimaker 3) (Medulla & Cortex) phantom
26
Phantom Fabrication
Final design Final printed phantom
Jaszczak phantom
Top baseplate
Spleen
Kidney
Support rods
Bottom baseplate
27
Phantom Fabrication
“Bulk production” SPECT/CT quality control
• Designed for the MRTDosimetry project as QI phantom.

• Fits a standard Jaszczak phantom.
• One phantom set will be sent to each participating partner site.
28
Phantom Filling: Setup at UKW

Dose Calibrator
Clamp Stand
2 Stock Solutions:
• High Activity Concentration
(Kidney Medulla)
• Low Activity Concentration
(Spleen & Kidney Cortex)
High
Container
Lab Balance High Low
29
Phantom Filling: Spreadsheet
To be filled out by participant / Automated uncertainty estimation
30
Filling & Mounting: Step-by-Step Instructions
Mounting Instructions
Filling Instructions
31
Phantom-specific
organ-positioning!
Acquisition & Reconstruction as for sensitivity & partial volume.

32
Lu-177 Measurement at UKW:

• Flash3D (Siemens’s OSEM). SPECT/CT
Fusion
• Attenuation correction.
• Scatter correction.
• Resolution recovery.
• 4 iterations, 8 subsets.
• No post-filtering.
SPECT
33
Outline

DosePart
Comparison
Part 3 Part 4
Sensitivity
Ba-133 Liquid I-131

→ Test Ba-133 as surrogate
for liquid I-131
34
1.4 Validation of Radioactive Test Sources
Quality control often based on liquid sources.
→ Disadvantages regarding time effort & radiation protection.
0.53
0.73
0.21
0.78
0.87 0.81
Idea:
Replace liquid I-131 sources by sealed Ba-133 sources.
35
• Design of 4 Ba-133 sealed sources.
Ba-133
Concentration Activity Diameter Length Volume Dimensions & Activities
(kBq/g) (MBq) (mm) (mm) (cm^3)
200 0.34 7.5 38.0 1.7
200 1.34 15.0 38.0 6.7 3D printed
200 5.38 30.0 38.0 26.9 containers
200 21.48 60.0 38.0 107.4
CAD Models.
Liquid
• Production of 4 identical containers I-131
to be filled with liquid I-131.
• Validate sealed Ba-133 against liquid I-131 sources.
36
d = 21.6 cm
7.5mm/40mm
Jaszczak Phantom
Solid Sources
Partial Volume 15mm/50mm
60mm/95mm
30mm/60mm
Attachment holes
Jaszczak cylinder
Estimation of partial volume effects

(resolution high energy collimator: 20 mm)
Baseplate produced by Christie
37
• Production of sealed Ba-133 sources by CMI, Czech Republic

and CEA, France.
• Production of hollow containers for liquid I-131 comparison.
Filling
hole
Ba-133 source production at CEA. Hollow 60-mm container

attached to baseplate.
38
Filling, Mounting, Acquisition & Reconstruction:

Step-by-Step Instructions
Attach base plate (3 screws). Attach Ba-133 sources. Positioning on patient

Attach attachment rods (4 circles). Close Phantom. bed.
Fill with water.
39
Ba-133 Measurement at UKW:

• Flash3D (Siemens’s OSEM).
SPECT/CT Fusion
• Attenuation correction. 4i8s0mm
• Scatter correction.
• Resolution recovery.
• 4 iterations, 8 subsets.
• No post-filtering.
SPECT
40
Outcomes / Deliverables
QI Comparison Exercise:
• Protocol for commissioning and QC of SPECT/CT imaging
systems.
• Multi-center assessment of quantitative accuracy of
different SPECT/CT systems.
• Validation of Ba-133 as surrogate for liquid I-131.
41
Aims of Comparison Exercises
1. Need for standardization of quantitative imaging

– Problem: Numerous imaging modalities & systems.
No standardization of activity determination.
– Plan: Comparison exercise with different SPECT/CT systems
BUT standardized activities & reconstructions.
– Aim: Protocol for commissioning and QI.
2. Lack of validation between dosimetry software tools

– Problem: Numerous tools available for dose calculation.
No validation across different tools.
– Plan: Calculate doses with available dosimetry tools based
on simulated post-therapy dosimetry dataset.
+ many in-house
– Aim: Protocol for commissioning dosimetry platforms. approaches
42
Outline
Part 1 & 2 Part 3 Part 4
4-organ phantom
Quantitative Serial
Imaging Comparison SPECT (UKW)
Exercise Dose Calculations
Ground truth TAC Imaging
x S

43
Outline
4-organ phantom
Quantitative Serial
x S

44
2.1 Extended 3D Printed Phantom
Jaszczak phantom
Top baseplate
Spleen
Support rods
Bottom baseplate
• Current phantom is useful for QI exercise.

• Not fully representative for Dose Calculations.
45
Bespoke
elliptical
phantom
Spleen Tumour
Kidney
Liver
Left Kidney Right Kidney
• 3D Printed phantom extended to include 2 kidneys (each with

separate Cortex & Medulla), Spleen, Liver, and Tumour.
• Phantom based on ICRP110 models.
46
• Imaged in large elliptical phantom.
47
• Preliminary imaging undertaken to help with

develop design of printed phantoms.
48
Outline
4-organ phantom
Quantitative Serial
x S

49
2.2 Ground Truth TAC
Uptake
Liver
Spleen
Time (h)
• Organ Time Activity Curves:

– TAC is required before designing experiments & calculating
ground truth dose.
– Mimic Lu-177 Dotatate Therapy.
50
2.2 Ground Truth TAC
• Organ Time Activity Curves:

– Implemented compartmental Pharmacokinetic model using published data [2], altering the
volumes to match printed phantom
– Cortex & Medulla to have 3:1 activity concentration ratio [3].
– Initial injected activity: 7.4 GBq equivalent.
[2] Brolin, et al. Phys. Med. Biol. 60 (2015) 6131-6149 51

[3] MIRD Pamphlet 20 JNM 49(11) 2008 1884-1899
2.2 S-Factor Computation
Source
Liver LK Cortex LK Medulla RK Cortex RK Medulla Spleen Tumour
Liver 6.57E-02 3.63E-04 2.96E-04 1.08E-03 8.07E-04 2.93E-04 3.27E-03
S (mGy/MBq.hr)
LKCortex 3.64E-04 9.10E-01 1.80E-02 4.21E-04 3.76E-04 1.76E-03 2.85E-04
LK Medulla 2.96E-04 1.79E-02 1.94E+00 4.13E-04 3.87E-04 9.87E-04 2.27E-04
RK Cortex 1.08E-03 4.22E-04 4.18E-04 1.09E+00 2.35E-02 1.85E-04 7.91E-04
RK Medull 6.51E-04 3.08E-04 3.17E-04 1.89E-02 1.87E+00 1.30E-04 5.20E-04
Spleen 2.93E-04 1.75E-03 9.81E-04 1.82E-04 1.59E-04 6.71E-01 2.20E-04
Tumour 3.28E-03 2.91E-04 2.30E-04 7.87E-04 6.48E-04 2.24E-04 5.21E+00
• Simulated in Geant4 & Gate to calculate S-

Factors for the phantom, required for
dosimetry.
52
Outline
4-organ phantom
Quantitative Serial
x S

53
2.3 Post-Therapy Lu-177 SPECT Imaging
• Imaging time point picking:

– Different centres have different guidelines
• Site A typically image PRRT patients at 4 time points: 1 h, 24 h, 72 h, 144 h.
• Site B at 5 time points: 1 h, 4 h, 24 h, 40 h, 70 h.
– We will image at time points 1 h, 4 h, 24 h, 40 h, 72 h, 144 h allowing us to generate number
of datasets of 4, 5, or 6 points.
– Filling & scanning follow SOP defined in QI comparison exercise.
54
2.3 Simulated Measurements
STLs voxels
to to
voxels proj.
• Lund: simulating SPECT with SIMIND and CT with a

stoichiometric calibration method
• INSERM: simulating SPECT with GATE
• SCK·CEN: simulating CT with GATE
55
2.3 Generating device agnostic images
• Experimental data taken on GE Discovery 670.

• Simulated data may be in non-DICOM format.
• SCK·CEN have developed tool to convert simulated images
to DICOM with correct header information to be
reconstructed on clinical workstations.
56
Outline
4-organ phantom
Quantitative Serial
x S

57
2.4 Dosimetry systems available
Commercial In House
LUNDADOSE
RAYDOSE
VoxelMed
MIDAS
58
2.4 Dosimetry systems available
Planar Planar + SPECT MULTI-SPECT Organ Voxel
Voxelmed Voxelmed Voxelmed
Raydose Raydose
Hermes Hybrid Hermes Hybrid Hermes Hybrid Hermes Hybrid Hermes Hybrid
Dosimetry™ Dosimetry™ Dosimetry™ Dosimetry™ Dosimetry™
LundADose LundADose LundADose LundADose LundADose
GE GE GE GE
THG THG THG THG
NUKDOS NUKDOS NUKDOS NUKDOS
MIDAS MIDAS MIDAS MIDAS MIDAS
QDose QDose QDose QDose
DoseFX DoseFX
Planet® Dose Planet® Dose
MIRADA MIRADA
Simplicit90YTM Simplicit90YTM
59
2.4 Patient Dose Comparison
ASMN, CASE 1, Lu-177 DOTATATE Therapy
SPECT/CT Fusion, 24 h post injection SPECT only
60
ASMN, CASE 2, Lu-177 DOTATATE Therapy
SPECT/CT Fusion, 24 h post injection SPECT only
61
Patient 1 WP 4.1.1 of MRTDosimetry
Dose Dose Dose
(Gy)
Liver (Gy)
Total Kidneys (Gy)
Spleen
2.00 2.00 1.50
1.50 1.50
1.00
1.00 1.00
0.50
0.50 0.50
0.00 0.00 0.00
Patient 2
Dose Dose Dose
(Gy)
Liver (Gy)
Total Kidneys (Gy)
Spleen
2.50 2.00 3.00
2.00 1.50
2.00
1.50
1.00
1.00
1.00
0.50 0.50
0.00 0.00 0.00
→ Statistical analysis challenging (only few statistically independent data

Organ-based | Voxel-based
points, e.g. 8 data points from 1 site).
62
2.4 Data collection
• Need to ensure data is collected in same format for easy comparison.

• Developed data recording spreadsheet (based on IAEA RaBiT), that
doubles as a checklist to ensure dose calculations have been performed
correctly.
[4] IAEA Radiotracer Biodistribution Template (RaBiT) , 2016 63

2.4 Data collection
• Important data to collect include TAC fits, etc, but more importantly
the uncertainties associated with these.
• Uncertainties on fit parameters will be used (where possible) to
estimate uncertainties on calculated dose values.
64
2.4 Data collection
• Important data to collect include TAC fits, etc, but more importantly
the uncertainties associated with these.
• Uncertainties on fit parameters will be used (where possible) to
estimate uncertainties on calculated dose values.
65
Data comparison
Voxel-level dosimetry
Whole organ dose
Dose Volume Histogram

1
Fractional Volume
0.8
0.6
0.4
0.2
0
0 5 10 15
Dose (Gy)
• Dose calculations – whole organ & voxel-based - compared to ideal and

variability across different software.
• Effect of using different time points & different number of time points for
the TAC.
• Uncertainties – whether they can be calculated & their value.
66
Outcomes / Deliverables
QI Comparison Exercise:
• Protocol for commissioning and QC of SPECT/CT imaging
systems.
• Multi-center assessment of quantitative accuracy of
different SPECT/CT systems.
• Validation of Ba-133 as surrogate for liquid I-131.
Dose Comparison Exercise:
• Protocol for commissioning dosimetry platforms &
determining overall uncertainty in the absorbed dose
quantification process.
• First ground-truth training set for validation of different
dosimetry software.
67
Thank you very much
for your kind attention!
http://mrtdosimetry-empir.eu/
68

04 MRTdosimetryWorkshop Calvert and Tran-Gia SM

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

04 MRTdosimetryWorkshop Calvert and Tran-Gia SM

Uploaded by

Copyright:

Available Formats

Quantitative Imaging and Dose

Calculation Comparison Exercises in

1. Need for standardization of quantitative imaging

2. Lack of validation between dosimetry software tools

SPECT/CT Hardware available at Participating Sites:

Quantitative Imaging Comparison Exercise (UKW)

Dose Comparison Exercise (The Christie)

Quantitative Imaging Comparison Exercise (UKW)

Ba-133 Liquid I-131

• Unit: counts-per-s-per-MBq (cps)

Suggestions on Phantom Selection

Protocol: Suggestions on SPECT/CT Acquisition

• Acquisition parameters as in clinical protocol

• Projection-wise decay correction.

Example from UKW (Lu-177)

Protocol: VOI Analysis

Calculate conversion factor

Quantitative Imaging Comparison Exercise (UKW)

Ba-133 Liquid I-131

MRTDosimetry comparison exercise

volumes Vsphere i were pre-measured)

• Prepare Stock solution (500 ml container).

• Ensure spheres are empty.

• Fill spheres until liquid “rises up along capillary”.

• Sphere activities: 𝐴𝑠𝑝ℎ𝑒𝑟𝑒 𝑖 = 𝑐𝑠𝑡𝑜𝑐𝑘 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛 ∗ 𝑉𝑠𝑝ℎ𝑒𝑟𝑒 𝑖

Acquisition & Reconstruction as for sensitivity.

Example from UKW (Lu-177)

Recovery Curve Fitting

• Obtain volume-dependent recovery coefficient curve by

Quantitative Imaging Comparison Exercise (UKW)

Ba-133 Liquid I-131

[1] ICRP Publication 110. Ann ICRP 39(2), 2009. 24

FDM 3D-Printing Printed parts Completed

• Designed for the MRTDosimetry project as QI phantom.

Phantom Filling: Setup at UKW

Phantom Filling: Spreadsheet

To be filled out by participant / Automated uncertainty estimation

Filling & Mounting: Step-by-Step Instructions

Acquisition & Reconstruction as for sensitivity & partial volume.

Lu-177 Measurement at UKW:

Quantitative Imaging Comparison Exercise (UKW)

Ba-133 Liquid I-131

Estimation of partial volume effects

Baseplate produced by Christie

• Production of sealed Ba-133 sources by CMI, Czech Republic

Ba-133 source production at CEA. Hollow 60-mm container

Filling, Mounting, Acquisition & Reconstruction:

Attach base plate (3 screws). Attach Ba-133 sources. Positioning on patient

Ba-133 Measurement at UKW:

1. Need for standardization of quantitative imaging

2. Lack of validation between dosimetry software tools

Dose Comparison Exercise (The Christie)

Dose Comparison Exercise (The Christie)

• Current phantom is useful for QI exercise.

Left Kidney Right Kidney

• 3D Printed phantom extended to include 2 kidneys (each with

• Imaged in large elliptical phantom.

• Preliminary imaging undertaken to help with

Dose Comparison Exercise (The Christie)

• Organ Time Activity Curves:

• Organ Time Activity Curves:

[2] Brolin, et al. Phys. Med. Biol. 60 (2015) 6131-6149 51

• Simulated in Geant4 & Gate to calculate S-