JMP_70_18R7

Original Article
1 1
2 2
3
4 Dosimetry Effects Caused by Unilateral and Bilateral Hip 3
4
5
6
Prostheses: A Monte Carlo Case Study in Megavoltage Photon 5
6
7
8
Radiotherapy for Computed Tomography Data without Metal 7
8
9 Artifacts 9
10 10
11 Courage Mahuvava, Frederik Carl Phillipus Du Plessis 11
12 Department of Medical Physics, Faculty of Health Sciences, University of the Free State, Bloemfontein, South Africa 12
13 13
14 14
15 Abstract 15
16 16
17 Background: Hip prostheses (HPs) are routinely used in hip augmentation to replace painful or dysfunctional hip joints. However, 17
high‑density and high‑atomic‑number (Z) inserts may cause dose perturbations in the target volume and interface regions. Aim: To evaluate
18 the dosimetric influence of various HPs during megavoltage conformal radiotherapy (RT) of the prostate using Monte Carlo (MC) simulations.
18
19 Materials and Methods: BEAMnrc and DOSXYZnrc MC user‑codes were respectively used to simulate the linac head and to calculate 3D 19
20 absorbed dose distributions in a computed tomography (CT)‑based phantom. A novel technique was used to synthetically introduce HPs into the 20
21 raw patient CT dataset. The prosthesis materials evaluated were stainless steel (SS316L), titanium (Ti6Al4V), and ultra‑high‑molecular‑weight 21
22 polyethylene (UHMWPE). Four, five, and six conformal photon fields of 6–20 MV were used. Results: The absorbed dose within and beyond 22
metallic prostheses dropped significantly due to beam attenuation. For bilateral HPs, the target dose reduction ranged up to 23% and 17% for
23 SS316L and Ti6Al4V, respectively. For unilateral HP, the respective dose reductions were 19% and 12%. Dose enhancement was always <1%
23
24 for UHMWPE. The 6‑field plan produced the best target coverage. Up to 38% dose increase was found at the bone–SS316L proximal interface. 24
25 Conclusions: The novel technique used enabled the complete exclusion of metal artifacts in the CT dataset. High‑energy plans with more 25
26 oblique beams can help minimize dose attenuation through HPs. Shadowing and interface effects are density dependent and greatest for 26
27 SS316L, while UHMWPE poses negligible dose perturbation. 27
28 28
Keywords: Dose perturbation, hip prostheses, Monte Carlo, radiotherapy, treatment planning
29 29
30 Received on: 14-07-2018 Review completed on: 24-10-2018 Accepted on: 24-10-2018 30
31 31
32 Introduction in kilovoltage CT images.[5‑7] Artifacts degrade the image 32
33 diagnostic quality and impair contour delineation of target and 33
Prostate cancer is one of the most frequently diagnosed
34 critical organs.[3,8] Several metal artifact reduction techniques 34
male malignancies worldwide,[1,2] with the majority of cases
35 have been implemented to partially compensate for CT 35
detected in men above 65 years old. This older population is artifacts.[9‑15]
36 furthermore affected by other age‑related comorbidities such 36
37 as osteoarthritis, for which a common treatment is prosthetic Prostate cancer is often treated with brachytherapy and/or 37
38 hip replacement. The number of hip replacements performed external beam RT (EBRT). EBRT strategies for prostate cancer 38
39 annually has increased from 210 per 100 000 males in include beam arrangements of 4, 5, 6, or more uniform beams, 39
40 1998–1999 to 265 in 2006–2007.[3,4] 40
Address for correspondence: Dr. Frederik Carl Phillipus Du Plessis,
41 Department of Medical Physics, Faculty of Health Sciences, University 41
High‑Z prostheses cause dosimetric calculation errors when
42 setting up a radiotherapy (RT) plan. This can be attributed
of the Free State, P.O. Box 339, Bloemfontein 9300, South Africa. 42
43 E‑mail: duplessisfcp@ufs.ac.za 43
to metal artifacts in the computed tomography (CT) dataset,
44 which are perceived as starburst streaking and blurring 44
45 This is an open access journal, and articles are distributed under the terms of the Creative 45
Commons Attribution‑NonCommercial‑ShareAlike 4.0 License, which allows others to
46 Access this article online remix, tweak, and build upon the work non‑commercially, as long as appropriate credit 46
47 Quick Response Code: is given and the new creations are licensed under the identical terms. 47
48 Website: For reprints contact: reprints@medknow.com 48
www.jmp.org.in
49 49
How to cite this article: Mahuvava C, Du Plessis FC. Dosimetry effects
50 caused by unilateral and bilateral hip prostheses: A Monte Carlo case study
50
DOI:
51 10.4103/jmp.JMP_70_18 in megavoltage photon radiotherapy for computed tomography data without 51
52 metal artifacts. J Med Phys 2018;4:XX-XX. 52

30 © 2018 Journal of Medical Physics | Published by Wolters Kluwer - Medknow

 Mahuvava and Du Plessis: Dosimetry effects of hip prostheses in radiotherapy

1 ranging from 6 to 18 MV.[16,17] To reduce the dosimetric effects of 1

2 hip prostheses (HPs), Eng[18] suggested a 6‑field (6F) conformal 2
3 technique with high‑energy photon beams, increasing the 3
4 anterior–posterior/posterior–anterior (AP/PA) beam weighting 4
5 in a conventional 4‑field  (4F) technique or using bilateral 5
rotational arcs. Efforts have also been made to demonstrate the
6 6
use of advanced RT techniques such as volumetric modulated
7 7
arc therapy (VMAT).[19] Dose conformality may be improved
8 by intensity‑modulated RT (IMRT) techniques.[2,8,20] 8
9 9
10 The American Association of Physicists in Medicine: Radiation 10
11 Therapy Committee Task Group 63 (AAPM TG‑63) report[21] 11
advocates beam arrangements that avoid HPs in the beam’s
12 12
eye view. However, this approach restricts the choice of
13 13
exploitable gantry angles, and the resultant dose distributions
14 suffer from poor dosimetric conformality.[21,22] Oblique beam
14
15 orientation is further limited by its escalation of bladder and Figure 1: The geometry used for modeling the Elekta Precise linac 15
16 head in BEAMnrc. The component modules used to simulate each linac 16
rectal tissue dose.
17 component are shown on the right hand side 17
The concern is tumor dose reduction due to radiation attenuation
18 18
through the HP and induced scattered dose near the prosthesis. distribution in a homogenous water tank (WT) for linac beam
19 19
These dose perturbations (radiation attenuation and scatter) characterization before CT‑based simulations. The dose error
20 may not be accurately predicted in conventional treatment
20
associated with all simulations was kept below 1%.
21 planning systems (TPS), especially near tissue‑bone‑HP 21
22 interfaces.[23] Monte Carlo (MC)‑based systems, on the other Water tank simulations – photon beam characteristics 22
23 hand, can accurately quantify the dosimetric influence of For analysis of photon beam characteristics, a 23
24 HPs. The MC simulation accuracy is, however, limited by the 40 cm × 40 cm × 40 cm homogenous WT was simulated under 24
25 accuracy of the density assignment performed using the CT the linac head at 100 cm source‑to‑surface distance. Voxels of 25
26 dataset and is thus limited by the presence of artifacts. size 0.2 cm × 0.2 cm × 0.2 cm were used. X‑ray square fields 26
27 of 5, 10, 15, and 20 cm per side and energies 6, 10, 15, and 27
This study evaluates the dose perturbation effect of SS316 L, 20 MV incident perpendicularly on the WT were simulated.
28 Ti6Al4V, and UHMWPE HPs during high‑energy conformal 28
Central axis percentage depth dose (PDD) curves and dose
29 X‑ray prostate treatment using MC simulations. The method 29
profiles normalized to the depth of maximum dose (dmax) were
30 used enables the complete exclusion of imaging artifacts by 30
presented.
31 synthetically introducing the prosthesis into the femur of the 31
32 raw patient CT data using a masking method. Computed tomography‑based simulations 32
33 A CT‑based phantom from scanned patient images was used 33
34 Materials and Methods as the patient model. Generating a CT‑based phantom with 34
35 HPs was a two‑step process, contouring the shape of the 35
This study was approved by the University of the Free State
36 prosthesis into a raw pelvis CT dataset in MCSHOW graphical 36
Health Sciences Research Ethics Committee; reference number
user interface (GUI) and then converting the mask media and
37 ECUFS NR 25/2015. 37
density information into the desired HP using an IDL code.
38 The MC technique used is a two‑step process, simulating the 38
This is illustrated in figure 2. The code reads in an *.egs4phant
39 linac head design and then simulating the dose distribution 39
file, the phantom media, the media density data, as well as
40 in a virtual phantom. BEAMnrc[25] was used to build a 3D the mask (delineated HP volume) media data. The code then 40
41 model of an Elekta Precise unit operating at three nominal changes the mask media type into the HP media as well as the 41
42 photon energies of 6, 10, and 15 MV. The linac model was mask physical density into the physical density of the desired 42
43 also modified to create a 20 MV virtual machine. prosthesis material and write the output as a new phantom file. 43
44 The accelerator head [Figure 1] was simulated based on the 44
MCSHOW allows for contouring, slice by slice, the shape
45 manufacturer’s geometric and material specifications. The 45
of the HP, reproduced from a CT dataset of a patient with a
46 phase space data generated from BEAMnrc simulation was 46
HP, onto the actual patient CT dataset  (*.egs4phant file) and
47 scored below the linac head exit window and contains all the substituting bony tissue with the prosthesis material. Unilateral
47
48 information relating to the incident beam, e.g., particle energy, and bilateral HP patient models were generated. The resulting 48
49 position, charge, and direction.[26,27] 49
phantoms were exported to DOSXYZnrc for dose calculations.
50 The second stage of the calculation involved using The dose distribution data were stored in *.3ddose files from 50
51 DOSXYZnrc[17,28] to calculate the dose distribution in a CT where PDD, dose‑volume histogram (DVH), and dose profile 51
52 phantom. DOSXYZnrc was also used to calculate the dose data were extracted. CT‑based profiles were normalized to the 52

Journal of Medical Physics  ¦  Volume 43 ¦ Issue 4 ¦ October-December 2018 31

 Mahuvava and Du Plessis: Dosimetry effects of hip prostheses in radiotherapy

1 1
Table 1: The composition of titanium, stainless steel, and
2 ultra‑high‑molecular‑weight polyethylene used to generate
2
3 implant PEGS4 cross‑sectional data in Monte Carlo simulations 3
4 4
Implant material Ti6Al4V SS316L UHMWPE
5 5
Mass density (ρ) (g/cm3) 4.34 8.10 0.95
6 6
Effective atomic number (Z) 21.4 26.7 5.4
7 Composition (percentage by Ti: 89.947 Fe: 63.70 H:14.29
7
8 weight) Al: 5.80 Cr: 18.50 C:85.71 8
9 V: 3.92 Ni: 12.00 9
10 Fe: 0.18 Mo: 3.00 10
11 O: 0.11 Mn: 2.00 11
12 H: 0.012 Si: 0.75 12
C: 0.01 P: 0.03
13 13
Y: 0.001 S: 0.03
14 Ti6Al4V: Titanium, SS316L: Stainless steel,
14
15 UHMWPE: Ultra‑high‑molecular‑weight polyethylene, PEGS4: Pre- 15
16 processor for Electron Gamma Shower (version 4.0) 16
17 17
18 Results 18
19 Water tank simulations – photon beam characteristics 19
20 For analysis of photon beam characteristics, simulated 20
21 dosimetric parameters from PDD curves and dose profiles 21
22 were compared with measured values from the machine’s 22
23 commissioning data in a WT. 23
24 24
Dose profiles
25 Figure 2: Illustration of the masking method. The top figure shows the
25
Figure 3a‑d present the simulated in‑plane dose profiles
26 transversal delineation of hip prostheses into the original computed (cross‑plane profiles not shown) for the given range of beam
26
27 tomography data using MCSHOW and the bottom figure shows how the energies and field sizes in a WT. An in‑plane profile is defined 27
28 IDL code generates a phantom with hip prostheses and no artifacts by 28
as being along the beam data y‑axis and a cross‑plane profile
29 overriding the mask media data with hip prostheses data 29
is along the x‑axis.
30 30
31 isocenter dose with no HP. Since HPs were embedded into the CT Percentage depth dose curves 31
dataset, the absorbed dose with and without HPs as well as dose Table 2 shows the simulated percentage doses at 10 cm
32 32
perturbation factors were calculated in artifact‑free CT phantoms. depth (PDD 10 cm) as well as dmax values in a WT. The simulated
33 The specific PEGS4 cross‑sectional data were calculated for each
33
values shows good agreement (≤1%) with measured values (in
34 HP material in the EGSnrc Multi‑platform (EGSnrcMP) GUI. 34
brackets) obtained from the Elekta Precise unit commissioning
35 The compositions of the HPs are shown in Table 1.[21,24] data. 35
36 36
37 XiO TPS was used to define treatment beam portals for Computed tomography‑data based treatment planning 37
prostate three‑dimensional conformal RT (3DCRT) to allow and simulations
38 38
translation of the beam setup to DOSXYZnrc for phantom The prostate was delineated as the clinical target volume (CTV).
39 simulations. Beam setup (beam port configuration, beam size,
39
40 OARs (the bladder and rectum) were also delineated to evaluate 40
and isocenter position) translation was performed using a set the dose to these organs which governs the use of fields through
41 of transformation equations derived by Zhan et al.[29] 41
the prosthesis. The PTV was generated by a 1.0 cm isotropic
42 42
The dosimetric impact of HPs was investigated on central expansion of the CTV (excluding seminal vesicles). The dose
43 43
axis PDDs, dose profiles, and DVHs of patient CT‑based was delivered using 4‑, 5‑, and 6 equally‑weighted, isocentric
44 coplanar beams.[21] The isocenter coordinates were positioned at
44
simulations. DVHs were used to compare the dose to the
45 planning target volume (PTV) and to organs at risk (OARs). the PTV centre for a supine, head‑first patient setup. A dose of 45
46 The back‑scattered dose perturbation factor (BSDF) on the HP 75 Gy was prescribed to the PTV with ≥95% of the prescription 46
47 entrance side and the forward dose perturbation factor (FDPF) dose covering the PTV, based on the International Commission 47
48 on the HP exit side were calculated as BSDF (or FDPF) = Di/Dh on Radiation Units and Measurements (ICRU) 50 guideline[31] 48
49 where Di and Dh are the respective doses at a point with and that stipulates that an optimal plan is one wherein the whole PTV 49
50 without the HP. The AAPM TG 105 recommendations[30] were receives between 95% and 107% of the prescription dose. Plans 50
51 followed with respect to normalization, material conversion, were compared for PTV coverage and the influence of increasing 51
52 and dose‑to‑material/dose‑to‑water calculations. the number of beams from 6 to 20 MV. Interface effects were 52

32 Journal of Medical Physics  ¦  Volume 43 ¦ Issue 4 ¦ October-December 2018

 Mahuvava and Du Plessis: Dosimetry effects of hip prostheses in radiotherapy

1 6 MV 10 MV
1
2 100 100 2
FS (cm 2)
3 FS (cm 2) 3
4 80 5x5 80 5x5 4

Relative dose (%)

10 x 10
5 10 x 10 5
Relative dose (%)

60 15 x 15 60 15 x 15
6 6
20 x 20 20 x 20
7 40 40 7
8 8
9 20 20 9
10 10
0 0
11 -20 -15 -10 -5 0 5 10 15 20 -20 -15 -10 -5 0 5 10 15 20
11
12 a Off - axis distance (cm) b Off - axis distance (cm)
12
13 13
14 15 MV
14
20 MV
15 100
15
100
16 FS (cm 2) 16
FS (cm 2)
17 80 20 x 20 80 5x5
17
18 18

Relative dose (%)

5x5
Relative dose (%)

60 10 x 10
19 10 x 10 60
20 x 20
19
20 40
15 x 15
40 15 x 15 20
21 21
22 20 20 22
23 23
0 0
24 -20 -15 -10 -5 0 5 10 15 20 -20 -15 -10 -5 0 5 10 15 20
24
25 Off - axis distance (cm)
25
Off - axis distance (cm)
26 c d 26
27 Figure 3: Simulated in‑plane dose profiles for (a) 6 MV, (b) 10 MV, (c) 15 MV, and (d) 20 MV beams at various field sizes in water at 10 cm depth. 27
28 The 20 MV data were simulated from a virtual linac 28
29 29
30 Maximum PTV dose perturbation occurs for SS316L, followed 30
Table 2: Depth of maximum dose and percentage doses
by Ti6Al4V, and then UHMWPE. The target coverage was
31 at 10 cm depth values for all field sizes and beam 31
quantified by comparing the PTV receiving at least 95%
32 energies used 32
of the prescribed dose (V95%). In the presence of bilateral
33 Energy Dmax PDD10 cm (%) SS316L HPs, up to 89.5% of the PTV is underdosed at 6 MV,
33
34 (MV) (cm) which reduces to 53.5% at 20 MV. This underdosage can be 34
5×5 10×10 15×15 20×20
35 cm2 cm2 cm2 cm2 compensated for using higher AP/PA beam weights.[18,27] The 35
36 6 1.7 (1.7) 63.7 67.3 (67.9) 70 70.2 underdosed volume for a patient with Ti6Al4V ranges from 36
37 10 2.5 (2.5) 71.6 73.8 (73.3) 74.3 75.2 75.8% to 14.2%. No underdosage occurred for UHMWPE. 37
38 15 2.9 (3.0) 75 76.5 (75.9) 77.8 77.9 To evaluate the dose–volume effect to OARs, the rectal 38
39 20 3.5 (‑) 81.9 82.6 (‑) 83.8 83.8 and bladder volume receiving at  ≥80% of the prescription 39
40 Dmax: Depth of maximum dose, PDD10 cm: Percentage doses at 10 cm depth dose (V80%) was calculated. The maximum V80% (or V60Gy) 40
41 value for the rectum and bladder in this 4F plan is 16% and 41
42 also evaluated on lateral dose profiles intersecting the implant. 27%, respectively. 42
43 The beams were forced to intercept the prostheses because the 43
aim was to evaluate the attenuation effect of prostheses during Local dose perturbations (interface effects) also occur in the
44 vicinity of HPs. The “horns” in Figure 6 are due to electron 44
complete HP irradiation. The data are presented in such a way as
45 backscatter from metallic HP surface on the proximal interface, 45
to show under‑ or overdosage if the prosthesis is not accounted for.
46 relative to the lateral beam direction. 46
47 Four–field box plan 47
This effect was absent for UHMWPE, which showed a profile
48 The planar isodose distributions of the 4F plan with opposing 48
close to the one without HPs. The peak local hot spot due to
49 bilateral and AP/PA beams in three different scenarios is 49
backscattered electrons from SS316L was about the same as
50 presented in Figure 4a‑c. 50
the PTV dose. This was for a patient ~36 cm in diameter in
51 The effect of various bilateral HPs on the DVH data for the PTV, the lateral direction. Thus, the interface dose can be as high 51
52 bladder, and rectum using this technique is shown in Figure 5a‑d. as the PTV dose. The BSDF at distance P from the proximal 52

Journal of Medical Physics  ¦  Volume 43 ¦ Issue 4 ¦ October-December 2018 33

 Mahuvava and Du Plessis: Dosimetry effects of hip prostheses in radiotherapy

1 1
Table 3: Back‑scattered dose perturbation factors versus
2 distance from the interface for (a) 6 MV, (b) 10 MV (c), 2
3 15 MV, and (d) 20 MV photon beams 3
4 4
a. 6 MV
5 5
6 P (mm) SS316L Ti6Al4V UHMWPE 6
a b
7 0 1.24 1.18 0.99 7
2 1.08 1.06 0.99
8 8
4 1.04 1.03 1.00
9 6 0.99 1.00 1.00
9
10 10
b. 10 MV
11 11
12 P (mm) SS316L Ti6Al4V UHMWPE 12
0 1.27 1.20 0.99
13 13
c 3 1.10 1.05 0.99
14 6 1.02 1.02 1.00
14
15 9 1.00 1.00 1.00 15
16 c. 15 MV
16
17 17
P (mm) SS316L Ti6Al4V UHMWPE
18 18
0 1.32 1.21 0.99
19 Figure 4: Transverse isodose distributions and isodose levels of a 4‑field 19
5 1.21 1.08 1.00
20 box for a patient with (a) no hip prostheses, (b) a bilateral SS316L,
10 1.07 1.00 1.00
20
21 and (c) unilateral SS316L. The planning target volume is shown as the 21
12 1.00 1.00 1.00
region intersected by the beams
22 d. 20 MV 22
23 23
interface is presented in Table 3a‑d. Backscatter radiation at 6 P (mm) SS316L Ti6Al4V UHMWPE
24 24
MV contributed a small peak extending ~5 mm, which is the 0 1.38 1.25 1.00
25 order of the range of secondary electrons scattered from the 5 1.25 1.13 0.99
25
26 SS316L surface. For 10, 15, and 20 MV beams, backscatter 10 1.08 1.00 0.99 26
27 perturbation extended ~0.9 cm, 1.2 cm, and 1.5 cm from the 15 1.00 1.00 1.00 27
28 HP, respectively. Up to 38% local dose increase was observed SS316L: Stainless steel, Ti6Al4V: Titanium, P: Distance from interface 28
29 UHMWPE: Ultra‑high‑molecular‑weight polyethylene 29
at the bone–SS316L proximal boundary.
30 30
The change in the disequilibrium magnitude at the exit
31 interface was quantified by the FDPF as presented in Table 4a Table 4: Forward dose perturbation factor calculated 31
32 and comparison was made with the FDPF values from Reft versus energy at the distal interface in (a) this study and 32
33 et al.[21] in Table 4b. The FDPF increases above 10 MV due to (b) Reft et al.[21] 33
34 pair production interactions in the metal. 34
a. FDPF at distal interface
35 35
The isocenter dose reduction in the presence of bilateral Energy (MV) SS316L Ti6Al4V UHMWPE
36 36
SS316L HPs ranged from 17 to 23%, and for Ti6Al4V, it 6 0.62 0.84 0.99
37 ranged from 11 to 17%. For a unilateral HP, the respective
37
10 0.75 0.87 1.00
38 dose reductions were 7%–19% and 3%–12%. This is shown 15 0.80 0.94 1.00
38
39 in Table 5. 20 0.93 1.04 1.01 39
40 b. FDPF at distal interface
40
41 Five-field plan 41
42 This plan consisted of five equiangular spaced beams at gantry Energy (MV) Bone Steel Lead
42
angles 54°, 126°, 198°, 270°, and 342°, with the 270° lateral 6 0.94 0.85 0.84
43 43
beam intersecting the HP [Figure 7a]. 18 1.05 1.20 1.41
44 FDPF: Forward dose perturbation factor, SS316L: Stainless steel, 44
45 The effect of various bilateral HPs on the DVH data for the Ti6Al4V: Titanium, UHMWPE: Ultra‑high‑molecular‑weight polyethylene 45
46 PTV, bladder, and rectum at 6–20 MV using a 5F plan is shown 46
47 in Figure 7a‑d. The results are the same for unilateral prostheses 47
This is a double increase in beam energy, though the PTV is still
48 as only one HP blocks the field for the 5F configuration. 48
underdosed. The maximum rectal and bladder volume receiving
49 An interesting observation for this beam configuration is that at least 80% of the prescription dose is 17% and 27%, for 49
50 above 10 MV, increasing the beam energy does not cause a SS316L and Ti6Al4V, respectively. The isocenter dose reduction 50
51 significant increase in the PTV dose, with V95% increasing only in the presence of unilateral SS316  L HPs ranged 2%–14%, 51
52 from 80% to 92% from 10 MV to 20 MV for a SS316L HP. and for Ti6Al4V it ranged 1%–11%. This is shown in Table 6. 52

34 Journal of Medical Physics  ¦  Volume 43 ¦ Issue 4 ¦ October-December 2018

 Mahuvava and Du Plessis: Dosimetry effects of hip prostheses in radiotherapy

1 100 100 1
2 6 MV
10 MV
2
3 80 80 3
4 4
Fractional dose (%)

Fractional volume (%)

5 60 60 5
6 6
7 40 40 7
8 8
9 20 20 9
10 10
11 0 0 11
0 1000 2000 3000 4000 5000 6000 7000 8000 0 1000 2000 3000 4000 5000 6000 7000 8000
12 Dose (cGy) Dose (cGy)
12
13 PTV UHMWPE PTV No HP PTV Ti6Al4V PTV UHMWPE PTV NoP PTV Ti6Al4V 13
PTV SS316L bladder UHMWPE bladder No HP PTV SS316L rectum_NoP bladder_SS316L
14 bladder Ti6Al4V bladder SS316L rectum NoP rectum_316L bladder_Ti6Al4V rectum_Ti6Al4V
14
15 a rectum UHMWPE rectum Ti6Al4V rectum SS316L b bladder_UHMWPE rectum_UHMWPE bladder_Nop 15
16 100 100
16
17 90 15 MV 90 20 MV 17
18 80 80
18
19 70 70
19
Fractional volume (%)

20 60 Fractional dose (%) 60

20
21 50
21
50
22 40
22
40
23 23
30 30
24 24
20 20
25 25
10 10
26 26
0 0
27 0 1000 2000 3000 4000 5000 6000 7000 8000 0 1000 2000 3000 4000 5000 6000 7000 8000
27
Dose (cGy)
28 PTV UHMWPE
Dose (cGy)
PTV NoP PTV Ti6Al4V PTV UHMWPE PTV NoP PTV Ti6Al4V
28
29 PTV SS316L bladder_UHMWPE bladder_NoP bladder_SS316L rectum_UHMWPE bladder_NoP 29
bladder_Ti6Al4V bladder_SS316L rectum_UHMWPE rectum_NoP PTV SS316L rectum_SS316L
30 rectum_NoP rectum_Ti6Al4V rectum_SS316L bladder_Ti6Al4V rectum_Ti6Al4V bladder_UHMWPE 30
c d
31 31
32 Figure 5: Comparison of the cumulative dose‑volume histogram data of the planning target volume, bladder and rectum for various bilateral prostheses 32
using a 4F equiangular plan for (a) 6 MV, (b) 10 MV, (c) 15 MV, and (d) 20 MV, NoP: No prostheses; SS316L: Stainless steel; Ti6Al4V: Titanium;
33 UHMWPE: Ultra-high-molecular-weight polyethylene; PTV: Planning target volume 33
34 34
35 Simulation studies have shown that the dose for the bilateral occurs for UHMWPE. The percentage dose reduction as a 35
36 HP case is within the statistical noise of the data and was not function of beam energy for the 6F plan is shown in Figure 9. 36
37 included. 37
The rate of dose reduction for SS316L and Ti6Al4V alloys is
38 38
Six–field plan more pronounced at lower photon energies. At higher beam
39 energies (>15 MV), the rate of dose reduction diminishes over
39
This plan was composed of 6 fields at gantry angles 45°,
40 the energies studied here. The same trend was observed for the 40
90°, 135°, 225°, 270°, and 315°, i.e., two bilateral fields and
41 4F and 5F plans, although in a 5F plan, the results would be 41
two pairs of opposed oblique fields [Figure 8a]. The largest
42 PTV underdosage in the presence of a unilateral SS316L and the same between unilateral and bilateral prostheses. 42
43 Ti6Al4V HP for this particular 6F plan is 27% and 4.5%, 43
Figure 10a and b shows the relationship between the number
44 respectively. For bilateral HPs, it increases to 30% and 6.9%, 44
of beams and the target coverage for a patient fitted with
45 respectively. Comparison of the cumulative DVH data of the bilateral titanium alloy at 6 and 15 MV. The 6F plan gives the
45
46 PTV, bladder and rectum for various bHPs using a 6F plan for best PTV coverage. 46
47 6-20 MV is shown in Figure 8a-d. 47
48 Discussion 48
From Table 7, the PTV dose reduction in the presence of a
49 49
unilateral SS316L HP for the 6F plan is 10.3%, and for Ti6Al4V, The MC simulation of a full linac head design and beam production
50 50
it is 6.2% at 6 MV. At 20 MV, the dose reduction can be for 6–20 MV X‑ray beams was successfully performed, and the
51 minimized to 1.0% for Ti6Al4V. No significant dose reduction beam parameters were benchmarked against the machine’s 51
52 52

Journal of Medical Physics  ¦  Volume 43 ¦ Issue 4 ¦ October-December 2018 35

 Mahuvava and Du Plessis: Dosimetry effects of hip prostheses in radiotherapy

1 Lateral dose profiles through the isocenter at 6 MV Lateral dose profiles through the isocentre at 10 MV 1
2 100 BONE
2
100 BONE
3 BONE BONE 3
4 80
4
80
5 5
Relative dose (%)

Relative dose (%)

6 6
60 60
7 7
8 8
40 40
9 9
10 10
20 2
20
11 Right Left Right Left 11
HP Tissue HP HP HP
12 Tissue 12
0 0
13 -20 -15 -10 -5 0 5 10 15 20 -20 -15 -10 -5 0 5 10 15 20
13
14 x - axis (cm) x - axis (cm) 14
15 Series1 No prosthesis No prosthesis SS316L Ti6Al4V UHMWPE 15
a b
16 16
17 Figure 6: Comparisons of lateral dose profiles passing through the mid‑prostheses on the central axial slice for a 4‑field box at (a) 6 and (b) 10 MV. 17
Interface regions are marked by broken lines
18 18
19 19
20 100 100 20
21 90 6 MV 90 15 MV 21
22 80 80 22
23 23
Fractional volume (%)

70
Fractional volume (%)

70

24 60 60 24
25 50 50 25
26 40 40 26
27 30 30 27
28 20 20 28
29 10 10 29
30 0 0 30
0 1000 2000 3000 4000 5000 6000 7000 8000
31 0 1000 2000 3000 4000
Dose (cGy)
5000 6000 7000 8000
Dose (cGy) 31
32 prostate_UHMWPE prostate_NoP prostate_Ti6Al4V PTV UHMWPE
PTV SS316L
PTV NoP
bladder_UHMWPE
PTV Ti6Al4V
bladder_NoP
32
prostate_SS316L bladder_UHMWPE bladder_NoP
33 bladder_Ti6Al4V bladder_SS316L rectum_UHMWPE bladder_Ti6Al4V bladder_SS316L rectum_UHMWPE 33
a rectum_NoP rectum_Ti6Al4V rectum_SS316L b rectum_NoP rectum_Ti6Al4V rectum_SS316L
34 34
35 100 100 35
36 90 10 MV 90 20 MV 36
37 80 80 37
38 70 70 38
Fractional volume (%)

Fractional volume (%)

39 60 60 39
40 50 50 40
41 40 40 41
42 30 30 42
43 20 20 43
44 10 10 44
45 0 0 45
46 0 1000 2000 3000 4000 5000 6000 7000 8000 0 1000 2000 3000 4000 5000 6000 7000 8000 46
Dose (cGy) Dose (cGy)
47 PTV UHMWPE PTV NoP PTV Ti6Al4V PTV UHMWPE PTV NoP PTV Ti6Al4V 47
PTV SS316L bladder_UHMWPE bladder_NoP PTV SS316L bladder_UHMWPE bladder_NoP
48 bladder_Ti6Al4V bladder_SS316L rectum_UHMWPE bladder_Ti6Al4V bladder_SS316L rectum_UHMWPE 48
49 c
rectum_NoP rectum_Ti6Al4V rectum_SS316L
d
rectum_NoP rectum_Ti6Al4V rectum_SS316L
49
50 Figure 7: Comparison of the cumulative dose‑volume histogram data of the planning target volume, bladder and rectum for various bilateral prostheses 50
51 using a 5F equiangular plan for (a) 6 MV, (b) 10 MV, (c) 15 MV, and (d) 20 MV; NoP: No prostheses; SS316L: Stainless steel; Ti6Al4V: Titanium; 51
52 UHMWPE: Ultra-high-molecular-weight polyethylene; PTV: Planning target volume 52

36 Journal of Medical Physics  ¦  Volume 43 ¦ Issue 4 ¦ October-December 2018

 Mahuvava and Du Plessis: Dosimetry effects of hip prostheses in radiotherapy

1 100
100 1
2 90 90
10 MV
2
6 MV
3 80 80 3
Fractional volume (%)

4 4

Fractional volume (%)

70 70

5 60 60 5
6 50 50 6
7 40 40 7
8 30 30 8
9 20 20 9
10 10 10 10
11 0
0 1000 2000 3000 4000 5000 6000 7000 8000
0
0 1000 2000 3000 4000 5000 6000 7000 8000
11
12 Dose (cGy) Dose (cGy) 12
PTV UHMWPE PTV NoP PTV Ti6Al4V PTV UHMWPE PTV NoP PTV Ti6Al4V
13 PTV SS316L bladder_UHMWPE bladder_Nop PTV SS316L bladder_UHMWPE bladder_NoP 13
bladder_Ti6Al4V bladder_SS316L rectum_UHMWPE
14 a rectum_NoP rectum_Ti6Al4V rectum_SS316L
b
bladder_Ti6Al4V
rectum_NoP
bladder_SS316L
rectum_Ti6Al4V
rectum_UHMWPE
rectum_SS316L
14
15 15
100 100
16 16
90 90
17 15 MV 20 MV 17
80 80
18 18

Fractional volume (%)

Fractional volume (%)

70 70
19 19
60 60
20 20
50 50
21 21
40 40
22 30
22
30
23 20
23
20
24 10
24
10
25 0
25
0
26 0 1000 2000 3000 4000 5000 6000 7000 8000 0 1000 2000 3000 4000 5000 6000 7000 8000 26
Dose (cGy) Dose (cGy)
27 PTV UHMWPE PTV NoP PTV Ti6Al4V PTV UHMWPE PTV NoP PTV Ti6Al4V 27
PTV SS316L bladder_UHMWPE bladder_NoP PTV SS316L bladder_UHMWPE bladder_NoP
28 bladder_Ti6Al4V bladder_SS316L rectum_UHMWPE bladder_Ti6Al4V bladder_SS316L rectum_UHMWPE 28
rectum_NoP rectum_Ti6Al4V rectum_SS316L rectum_NoP rectum_Ti6Al4V rectum_SS316L
29 c d 29
30 Figure 8: Comparison of the cumulative dose‑volume histogram data of the planning target volume, bladder and rectum for various bilateral prostheses 30
31 using a 6F equiangular plan for (a) 6 MV, (b) 10 MV, (c) 15 MV, and (d) 20 MV; NoP: No prostheses; SS316L: Stainless steel; Ti6Al4V: Titanium; 31
32 UHMWPE: Ultra-high-molecular-weight polyethylene; PTV: Planning target volume 32
33 33
34 14 Table 5: Isocenter dose perturbation versus beam energy 34
35 bilateral SS316L in a 4F box plan for (a) bilateral and (b) unilateral hip 35
12
36 bilateral Ti6Al4V prostheses 36
37 10 37
% dose reduction

unilateral SS316L a. Bilateral HPs

38 8 38
unilateral Ti6Al4V Energy Percentage dose reduction Percentage dose increase
39 6 (MV) 39
SS316L Ti6Al4V UHMWPE
40 40
4 6 22.8 16.8 <1
41 41
10 20.4 14.8 <1
42 2
15 18.5 12.7 <1
42
43 0 20 16.9 10.6 <1 43
5 7 9 11 13 15 17 19
44 b. Unilateral HP 44
Beam energy (MV)
45 45
Energy Percentage dose reduction Percentage dose increase
46 Figure 9: Percentage dose reduction as a function of beam energy in (MV) 46
SS316L Ti6Al4V UHMWPE
47 a 6F plan for a patient with unilateral and bilateral SS316L and Ti6Al4V 47
hip prostheses 6 18.6 12.3 <1
48 48
10 15.1 9.5 <1
49 15 10.3 5.6 <1
49
50 commissioning data. Beam parameter data were within 1% of 50
20 6.8 2.5 <1
51 measured data. For CT‑based simulations, the dosimetric influence SS316L: Stainless steel, Ti6Al4V: Titanium, 51
52 of HPs on the mid axial slice and in some volumes was evaluated UHMWPE: Ultra‑high‑molecular‑weight polyethylene 52

Journal of Medical Physics  ¦  Volume 43 ¦ Issue 4 ¦ October-December 2018 37

 Mahuvava and Du Plessis: Dosimetry effects of hip prostheses in radiotherapy

1 100 100 1
2 90
6 MV
90 15 MV 2
3 3
80 80
4 4
Fractional volume (%)

70 70
5 5

Fractional volume (%)

6 60
4 fields
60
4 fields
6
7 50 50
5 fields
7
5 fields
8 40 40 6 fields
8
6 fields
9 30 30
9
prescribed prescribed
10 10
20 20
11 11
10
12 10 12
13 0
0 1000 2000 3000 4000 5000 6000 7000 8000
0 13
0 1000 2000 3000 4000 5000 6000 7000 8000
14 14
15 Dose (cGy) Dose (cGy) 15
a b
16 Figure 10: Dose‑volume histogram data showing the relationship between the number of beams and the planning target volume coverage for 16
17 (a) 6 MV and (b) 15 MV photon beams for the bilateral Ti6Al4V hip prosthesis case. More beams produce a better dose distribution 17
18 18
19 by quantifying the radiation absorption and scatter in the shadow 19
Table 6: Isocenter dose perturbation versus beam energy
20 of the HP and on the interfaces. The target coverage was quantified 20
in a 5F plan for a patient with unilateral hip prostheses
21 by calculating V95% values for each plan. 21
Unilateral prostheses
22 High‑energy beams have greater penetrating power, hence the 22
23 Energy Percentage dose reduction Percentage dose increase increased V95% with energy. The dose to critical organs was 23
(MV) independent of the beam energy used, or at least no general trend
24 SS316L Ti6Al4V UHMWPE 24
25 6 14.2 11.0 <1 was noticeable. Although OAR dose reduction occurred when 25
26 10 9.6 7.5 <1 high‑Z materials were used, OAR sparing was not the planning 26
27 15 5.3 4.2 <1 intent; therefore little can be said about OAR dose reduction. 27
20 2.3 1.1 <1
28 For a 4F plan, the dependence of dose distribution on beam 28
The dose reduction is exactly the same for the bilateral HP case. The data
29 are within the statistical noise. HP: Hip prostheses, Ti6Al4V: Titanium,
energy at the isocenter was weak, except close to the prosthesis 29
30 SS316L: Stainless steel, UHMWPE: Ultra‑high‑molecular‑weight as also reported by Erlanson et al.[32] in a similar study. They 30
31 polyethylene also found ~25% dose enhancement at the proximal interface, 31
32 extending up to 2 cm from the interface at 50 MV. 32
33 Increasing the number of beam portals generally improves 33
Table 7: Isocenter dose perturbation versus beam energy
34 the plan quality. The 5F plan causes a smaller PTV dose 34
in a 6F plan for bilateral and unilateral hip prostheses
35 reduction compared to a 4F box since only one beam intersects 35
a. Bilateral HPs a HP which is about 20% of the dose delivered by 5 equally
36 36
37 Energy Percentage dose reduction Percentage dose increase weighted fields, while the 4F field box has 2/4 fields (50%) 37
38 (MV)
SS316L Ti6Al4V UHMWPE
intersecting the prostheses. The small impact with regard to 38
beam energy is because five fields of equal weights are used
39 6 15.1 12.6 <1 39
to deliver dose to the PTV and only one passes through the
40 10 10.1 8.9 <1
HP at 270°. The 6F plan produces an even better PTV dose
40
41 15 7.2 4.8 <1 41
20 4.0 2.6 <1 conformity and less dose perturbation compared to the 4F box.
42 This is because the 6F plan has 2/6 fields (33%) passing through 42
43 b. Unilateral HPs 43
the HPs compared to 50% in a 4F plan. The 5F plan has 20%
44 Energy Percentage dose reduction Percentage dose increase of its fields intersecting the HP while the 6F plan has only 44
45 (MV) 1/6 (17%) intersecting the HP; hence, the 6F plan has less dose 45
SS316L Ti6Al4V UHMWPE
46 6 10.3 6.2 <1 perturbation for a unilateral HP. However, for bilateral HPs, 46
47 10 6.9 4.2 <1 the 6F plan has 33% of its beams intersecting HPs compared 47
48 15 3.5 2.4 <1 to 20% in a 5F plan, and therefore the dose perturbation would 48
49 20 2.1 1.0 <1 be worse for a 6F plan. Results highlight the influence of the 49
50 HP: Hip prostheses, Ti6Al4V: Titanium, SS316L: Stainless steel, implant’s composition, the beam energy, and position relative 50
UHMWPE: Ultra-high-molecular-weight polyethylene to the HP on the dose distribution.
51 51
52 52

38 Journal of Medical Physics  ¦  Volume 43 ¦ Issue 4 ¦ October-December 2018

 Mahuvava and Du Plessis: Dosimetry effects of hip prostheses in radiotherapy

1 One approach to correct for the effect of HP is to use 6. Wang G, Snyder DL, O’Sullivan JA, Vannier MW. Iterative deblurring for 1
special beam arrangements (even noncoplanar beam CT metal artifact reduction. IEEE Trans Med Imaging 1996;15:657‑64.
2 7. Robertson DD, Yuan J, Wang G, Vannier MW. Total hip prosthesis
2
3 arrangements) that completely or partially avoid the metal‑artifact suppression using iterative deblurring reconstruction. 3
4 prosthesis. Reducing the weight of beams intersecting the J Comput Assist Tomogr 1997;21:293‑8. 4
5 prosthesis or increasing their monitor units (MUs) to account 8. Rosewall T, Kong V, Vesprini D, Catton C, Chung P, Ménard C, et al.
5
Prostate delineation using CT and MRI for radiotherapy patients with
6 for prosthesis attenuation may provide an acceptable plan. 6
bilateral hip prostheses. Radiother Oncol 2009;90:325‑30.
An understanding of the dose perturbation effect of HPs 9. Brook OR, Gourtsoyianni S, Brook A, Mahadevan A, Wilcox C,
7 7
allows greater exploitation of gantry angles, number of Raptopoulos V, et al. Spectral CT with metal artifacts reduction software
8 for improvement of tumor visibility in the vicinity of gold fiducial 8
beam portals, etc., and potentially improves dosimetric
9 conformality. UHMWPE poses negligible dose perturbation markers. Radiology 2012;263:696‑705. 9
10 10. Boas FE, Fleischmann D. Computed tomography artifacts: Causes and 10
and therefore does not need any MU compensation. It was reduction technique. Imaging Med 2012;4:22940.
11 to be expected that there would be more dispersion in the 11. Axente M, Paidi A, Von Eyben R, Zeng C, Bani‑Hashemi A, 11
12 results for SS316L materials than for the UHMWPE, simply Krauss A, et al. Clinical evaluation of the iterative metal artifact 12
13 reduction algorithm for CT simulation in radiotherapy. Med Phys 13
because of the differences in the atomic number. However, 2015;42:1170‑83.
14 the cause of the inconsistency in the results at this point has 12. Meyer E, Raupach R, Lell M, Schmidt B, Kachelriess M. Normalized 14
15 not yet been understood. metal artifact reduction (NMAR) in computed tomography. Med Phys 15
2010;37:5482‑93.
16 This study confirms the results in similar studies; the advantage 16
13. Buzug T, Oehler M. Statistical image reconstruction for inconsistent CT
17 being the exclusion of imaging artifacts since heterogeneous projection data. Methods Inf Med 2007;46:261‑9. 17
18 media could be introduced into the CT‑phantom in a synthetic 14. Lemmens C, Faul D, Nuyts J. Suppression of metal artifacts in CT 18
using a reconstruction procedure that combines MAP and projection
19 way. This work may also have applicability to patients completion. IEEE Trans Med Imaging 2009;28:250‑60.
19
20 with other pelvic malignancies. However, given the patient 15. Boas FE, Fleischmann D. Evaluation of two iterative techniques 20
21 size variation and the prostate‑to‑prostheses relationship, for reducing metal artifacts in computed tomography. Radiology 21
2011;259:894‑902.
22 no general information can be gleaned, since one patient 22
16. Khan FM, Potish RA. Treatment Planning in Radiation Oncology.
23 anatomy was investigated. The work can nevertheless be Baltimore: William and Wilkins; 1998. 23
24 repeated on a representative number of patient samples. 17. Lin SY, Chu TC, Lin JP, Liu MT. The effect of a metal hip prosthesis on 24
Also, the study is restricted to 3DCRT which might limit the radiation dose in therapeutic photon beam irradiations. Appl Radiat
25 Isot 2002;57:17‑23.
25
26 its relevance to IMRT and VMAT dominated environments. 18. Eng TY. Dose attenuation through a titanium alloy hip prosthesis. Med 26
27 These techniques are, however, difficult to implement in Dosim 2000;25:7‑8. 27
MC simulations. 19. Prabhakar R, Kumar M, Cheruliyil S, Jayakumar S,
28 Balasubramanian S, Cramb J. Volumetric modulated arc therapy 28
29 Financial support and sponsorship for prostate cancer patients with hip prosthesis. Rep Pract Oncol 29
30 This research and the publication thereof is the result of Radiother 2013;18:209‑13. 30
20. Kung JH, Reft H, Jackson W, Abdalla I. Intensity‑modulated
31 funding provided by the South African Medical Research radiotherapy for a prostate patient with a metal prosthesis. Med Dosim 31
32 Council (MRC) in terms of MRC’s Flagship Awards Project 2001;26:305‑8. 32
33 SAMRC‑RFA‑UFSP‑01‑2013/HARD with funds from the 21. Reft C, Alecu R, Das IJ, Gerbi BJ, Keall P, Lief E, et al. Dosimetric 33
National Treasury under its Economic Competitiveness and considerations for patients with HIP prostheses undergoing pelvic
34 irradiation. Report of the AAPM Radiation Therapy Committee Task 34
35 Support Package. Group 63. Med Phys 2003;30:1162‑82. 35
36 22. Carolan M, Dao P, Fox C, Metcalfe P. Effect of hip prostheses on 36
Conflicts of interest radiotherapy dose. Australas Radiol 2000;44:290‑5.
37 There are no conflicts of interest. 23. Buffard E, Gschwind R, Makovicka L, David C. Monte Carlo 37
38 calculations of the impact of a hip prosthesis on the dose distribution. 38
Nucl Instrum Methods Phys Res B 2006;251:9‑18.
39 References 24. Alvarado J, Maldonado R, Marxuach J, Otero R. Biomechanics of hip
39
40 1. Agapito J. Radical radiation therapy for carcinoma of the prostate and knee prostheses. Appl Eng Mech Med 2003:6‑22. 40
41 in patients with a single hip prosthesis: A technique analysis using 25. Rogers DW, Walters B, Kawrakow I. BEAMnrc User’s Manual NRC 41
dose‑volume histograms. Med Dosim 2001;26:243‑50. Report PIRS 509 (a) RevH. Ionizing Radiation Standards; 2004.
42 26. Rogers DW, Faddegon BA, Ding GX, Ma CM, We J, Mackie TR.
42
2. Su A, Reft C, Rash C, Price J, Jani AB. A case study of radiotherapy
43 planning for a bilateral metal hip prosthesis prostate cancer patient. Med BEAM: A Monte Carlo code to simulate radiotherapy treatment units. 43
44 Dosim 2005;30:169‑75. Med Phys 1995;22:503‑24. 44
3. Martin DA, Hruby G, Whitaker MK, Foo KY. Constrained‑beam 27. Ding GX, Yu CW. A study on beams passing through hip prosthesis
45 for pelvic radiation treatment. Int J Radiat Oncol Biol Phys
45
inverse planning for intensity‑modulated radiation therapy of prostate
46 cancer patients with bilateral hip prostheses. J Med Imaging Radiat 2001;51:1167‑75. 46
47 Oncol 2012;56:703‑7. 28. Walters B, Kawrakow I, Rogers DW. DOSXYZnrc User’s Manual. 47
4. Rana SB, Pokharel S. A dosimetric study of volumetric modulated Ionizing Radiation Standards. National Research Council of Canada,
48 Ottawa, NRC Report PIRS-794; 2005. 48
arc therapy planning techniques for treatment of low‑risk prostate
49 cancer in patients with bilateral hip prostheses. South Asian J Cancer 29. Zhan L, Jiang R, Osei EK. Beam coordinate transformations from 49
50 2014;3:18‑21. DICOM to DOSXYZnrc. Phys Med Biol 2012;57:N513‑23. 50
5. Morin RL, Raeside DE. Removal of streaking artifact in computed 30. Chetty IJ, Curran B, Cygler JE, DeMarco JJ, Ezzell G, Faddegon BA,
51 et al. Report of the AAPM Task Group no 105: Issues associated with 51
tomography. J Med Syst 1982;6:387‑97.
52 52

Journal of Medical Physics  ¦  Volume 43 ¦ Issue 4 ¦ October-December 2018 39

 Mahuvava and Du Plessis: Dosimetry effects of hip prostheses in radiotherapy

1 clinical implementation of Monte Carlo‑based photon and electron ICRU Report 50. ICRU, Bethesda, Maryland; 1993. 1
external beam treatment planning. Med Phys 2007;34:4818‑53. 32. Erlanson M, Franzén L, Henriksson R, Littbrand B, Löfroth PO.
2 31. Landberg T, Chavaudra J, Dobbs HJ, Hanks G, Johansson KA, Mller T, Planning of radiotherapy for patients with hip prosthesis. Int J Radiat
2
3 et al. Prescribing Recording and Reporting Photon Beam Therapy. Oncol Biol Phys 1991;20:1093‑8. 3
4 4
5 5
6 6
7 7
8 8
9 9
10 10
11 11
12 12
13 13
14 14
15 15
16 16
17 17
18 18
19 19
20 20
21 21
22 22
23 23
24 24
25 25
26 26
27 27
28 28
29 29
30 30
31 31
32 32
33 33
34 34
35 35
36 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
47 47
48 48
49 49
50 50
51 51
52 52

40 Journal of Medical Physics  ¦  Volume 43 ¦ Issue 4 ¦ October-December 2018