The neutron dose equivalent evaluation and shielding at the maze entrance of a Varian

Clinac 23EX treatment room

Xudong Wang, Carlos Esquivel, Elena Nes, Chengyu Shi, Nikos Papanikolaou, and Michael Charlton

Citation: Medical Physics 38, 1141 (2011); doi: 10.1118/1.3533713

View online: http://dx.doi.org/10.1118/1.3533713
View Table of Contents: http://scitation.aip.org/content/aapm/journal/medphys/38/3?ver=pdfcov
Published by the American Association of Physicists in Medicine
The neutron dose equivalent evaluation and shielding at the maze entrance
of a Varian Clinac 23EX treatment room
Xudong Wanga兲 and Carlos Esquivel
University of Texas Health Science Center, San Antonio, Texas 78229
Elena Nes
South Texas Accelerated Research Therapeutics (START) Center for Cancer Care, San Antonio,
Texas 78229
Chengyu Shi, Nikos Papanikolaou, and Michael Charlton
University of Texas Health Science Center, San Antonio, Texas 78229
共Received 16 June 2010; revised 11 November 2010; accepted for publication 14 December 2010;
published 10 February 2011兲
Purpose: To evaluate the neutron and photon dose equivalent rate 共Hn,D and HG兲 at the outer maze
entrance and the adjacent treatment console area after the installation of a Varian Clinac 23EX
accelerator with a higher beam energy than its predecessor. The evaluation was based on measure-
ments and comparison with several empirical calculations. The effectiveness of borated polyethyl-
ene 共BPE兲 boards, as a maze wall lining material, on neutron dose and photon dose reduction is also
reported.
Methods: A single energy Varian 6 MV photon linear accelerator 共linac兲 was replaced with a Varian
Clinac 23EX accelerator capable of producing 18 MV photons in a vault originally designed for the
former accelerator. In order to evaluate and redesign the shielding of the vault, the neutron dose
equivalent Hn,D was measured using an Andersson–Braun neutron Rem meter and the photon dose
equivalent HG was measured using a Geiger Müller and an ion chamber ␥-ray survey meter at the
outer maze entrance. The measurement data were compared to semiempirical calculations such as
the Kersey method, the modified Kersey method, and a newly proposed method by Falcão et al.
Additional measurements were taken after BPE boards were installed on the maze walls as a
neutron absorption lining material.
Results: With the gantry head tilted close to the inner maze entrance and with the jaws closed, both
neutron dose equivalent and photon dose equivalent reached their maximum. Compared to the
measurement results, the Kersey method overestimates the neutron dose equivalent Hn,D by about
two to four times 共calculation/measurement ratio⬇ 2.4– 3.8兲. Falcão’s method largely overestimates
the Hn,D 共calculation/measurement ratio⬇ 3.9– 5.5兲. The modified Kersey method has a calculation
to measurement ratio about 0.6–0.9. The photon dose equivalent calculation including McGinley’s
capture gamma dose equivalent equation estimates about 77%–98% of the measurement. After
applying BPE boards as a lining material on the inner corner of the maze wall, the Hn,D and the HG
at maze entrance were decreased by 41% and 59%, respectively.
Conclusions: This work indicates that the Kersey method overestimates the neutron dose equiva-
lent Hn,D for a Varian Clinac 23EX accelerator. The Falcão method overestimates the Hn,D partially
due to the discrepancy in the International Commission on Radiological Protection 共ICRP兲 conver-
sion factors caused by the uncertainties of the estimated average neutron energy. The modified
Kersey method gives the closest estimation of a Varian Clinac 23EX accelerator operated at 18 MV
photon mode in a maze with a similar design as in the authors’ study. However, it should be used
with caution because of its tendency to underestimate the Hn,D. A borated polyethylene lining can
provide a cost effective method to reduce neutron and photon dose equivalent at the maze door for
an existing linac vault, following the installation of a higher energy linac. © 2011 American
Association of Physicists in Medicine. 关DOI: 10.1118/1.3533713兴

Key words: medical accelerator maze, photoneutrons, semiempirical calculations, shielding

I. INTRODUCTION a high energy above this threshold. The neutron dose as well
as the accompanying photon dose at the maze entrance has
When the beam energy of clinical linear accelerators 共linacs兲 been studied for many years by Monte Carlo simulations,4–6
is greater than 10 MeV, the neutron dose and the neutron semiempirical analysis,6–10 and measurements.4,8,10–13 Monte
shielding for operating personnel cannot be neglected.1–3 It is Carlo simulations are capable of providing a variety of dose
important to study the neutron dose at the maze entrance and information in advance. However, an accurate neutron spec-
the corresponding shielding method for the linacs working at trum simulation is required, with detailed information on the

1141 Med. Phys. 38 „3…, March 2011 0094-2405/2011/38„3…/1141/9/$30.00 © 2011 Am. Assoc. Phys. Med. 1141
1142 Wang et al.: Neutron dose evaluation and shielding 1142

geometry and the materials of the linac head and the treat-
ment room. Monte Carlo neutron simulations generally in-
volve elaborated programming and computation. Thus, semi-
empirical calculations are often preferred for neutron dose
evaluation and shielding verification due to its simplicity and
less time consumption. Normally, these calculations relate
the neutron dose equivalent at the maze entrance to the x-ray
absorbed dose at isocenter. A semiempirical calculation of
the dose equivalent at the maze door is easily achievable
when the beam dose rate is known. However, the applicabil-
ity of parameters in semiempirical equations is limited by the
type of machine, the geometry of treatment vault, and the
shielding material used in the studies. Direct measurements
for different linac models are required to examine and sup- FIG. 1. Scheme of the vault for the semiempirical calculations. The distance
port semiempirical calculation methods. d1 is from the isocenter C to the point A on the maze centerline from which
the isocenter is just visible. The distance d2 is from the point A to the outside
The relationship between neutron dose equivalent and the
of the maze door 共point B兲. The inner maze entrance cross-sectional area
absorbed dose at isocenter has been intensely studied by sev- and the cross-sectional area along the maze are denoted by S0 and S1,
eral research groups.6,9,12–17 The parameters of neutron respectively. When the gantry angle is 270°, the beam is directed from point
source strength 共Qn兲, neutron dose equivalent 共H0兲 near the D toward point E.
isocenter, and tenth value distance 共TVD兲 on commonly used
medical accelerators under different vault and maze shield-
ing conditions were provided. The studies include linacs

冉 冊冉 冊
manufactured by Elekta,12 General Electric,6,12–14,17 2
Mitsubishi,14 Philips,9,12,14,16 Siemens,6,12–14 Toshiba,14 and S0 d0
Hn,D = H0 10−共d2/5兲 , 共1兲
Varian.6,9,12–15 However, the relationship between the neutron S1 d1
dose equivalent at the maze door and the absorbed dose from
the photon beam for new linac models, including the Varian where H0 is the total neutron dose equivalent at a distance d0
Clinac 23EX 共Varian Medical Systems, Inc., Palo Alto, CA兲, 共1.41 m兲 from the target per unit absorbed dose of x-ray at
has not been thoroughly evaluated. A study of the neutron the isocenter 共mSv Gy−1兲; d0 is the distance from the target
dose equivalent at the maze entrance will provide informa- to a point, located in a horizontal plane 100 cm from the
tion for neutron dose equivalent evaluation for this fre- target and is 1 m perpendicular away from the axis of a beam
quently encountered linac. In addition, the parameters ex- at 180° gantry angle 关see Followill et al. for Fig. 1 共Ref. 12兲兴;
tracted from the shielding measurement should be applicable d1 is the distance measured from the isocenter C to the point
to a machine working with similar parameters and maze de- A on the maze centerline from which the isocenter is just
sign. visible 共Fig. 1兲; d2 is the distance from the point A to the
This work reports the measurement of neutron dose outside of the maze door 共point B兲; S0 is the inner maze
equivalent at the outer maze entrance and the console area entrance cross-sectional area; and S1 is the cross-sectional
adjacent to the neighboring vault after the installation of a area along the maze. This equation is based on the assump-
Varian Clinac 23EX accelerator with a higher beam energy tion that the maze has a TVD of 5 m for the attenuation of
than the predecessor. The measurement data were compared neutrons in the maze.
to semiempirical calculation results obtained by the Kersey McGinley et al.9 pointed out that the Hn,D could be re-
method,7 the modified Kersey method,10 and a newly pro- solved into the sum of two exponential functions, i.e., neu-
posed method by Falcão el al.6 The measurements were tron dose equivalent decreases exponentially with d2 with
taken before and after lining the maze wall with borated two different TVDs. Wu and McGinley10 further proposed an
polyethylene 共BPE兲 boards. The effectiveness of this neutron equation with consideration of nonstandard surface areas or
absorption material 共BPE兲 on neutron dose and photon dose mazes with exceptional width or length. The neutron dose
reduction was reported. In addition, the photon dose at the equivalent at the outside maze entrance is then given as the
maze entrance was measured and compared with calcula- modified Kersey method by the following equation:
tions.

II. METHODS AND MATERIALS Hn,D = 2.4 ⫻ 10−15␸A 冑 S0

S1
关1.64 ⫻ 10−共d2/1.9兲 + 10−共d2/TVD兲兴,
II.A. Theory 共2兲
Kersey proposed one of the earliest techniques to evaluate
the neutron dose equivalent at a maze entrance.7 According where Hn,D is expressed in Sv Gy−1 and ␸A represents the
to this method, the neutron dose equivalent at the maze en- neutron fluence at the inner maze entrance per unit absorbed
trance per unit absorbed dose due to x-rays at the isocenter dose of photons 共m−2 Gy−1兲 at the isocenter18–20 and can be
共Hn,D兲, expressed in mSv Gy−1, is given by determined according to

Medical Physics, Vol. 38, No. 3, March 2011

1143 Wang et al.: Neutron dose evaluation and shielding 1143

␤Qn 5.4␤Qn 1.26Qn for Varian 1800 linacs19 and by Followill et al. for Varian
␸A = + + , 共3兲 2100C linacs12 共Varian Medical Systems, Inc., Palo Alto,
4␲d21 2␲Sr 2␲Sr
CA兲, which also have a similar gantry head design. These are
where ␤ is the transmission factor for neutrons that penetrate the widely used values for shielding calculations recom-
the head shielding 共1.0 for lead and 0.85 for tungsten head mended in NCRP Report No. 151. The surface area of the
shielding兲, Qn is the neutron source strength in neutrons treatment room 共Sr兲 was estimated by the surface area of the
emitted from the accelerator head per Gray of x-ray absorbed concrete walls including the ceiling and floor. In the Falcão
dose at the isocenter,12,19 and Sr is the surface area of the method calculation, the average neutron energy used to get
treatment room 共m2兲. the neutron fluence to dose equivalent conversion factor
The TVD in Eq. 共2兲 is the tenth value distance of the from ICRP Publication 74 共Ref. 21兲 was taken as 0.1 MeV,
neutron attenuation in the maze expressed in meters. It is as suggested by Falcão et al.
proportional to the square root of the cross-sectional area
along the maze S1 共m2兲
TVD = 2.06冑S1 . 共4兲 II.B. Survey meters and measurement parameters
6
Recently, Falcão et al. proposed another equation, In this study, we used a Rem meter 共AN/PDR-70NRC
NP-2, Nuclear Research Corporation, Southampton, PA兲 as
TVD = 1.7 + 0.55S1 , 共5兲
the neutron dose survey meter. Its design is based on the
to estimate the tenth value distance without resolving the study of Anderson and Braun.23 The NP-2 Rem meter has an
neutron fluence into two components. They suggested con- energy response curve which simulates the ICRP fluence to
verting neutron fluence to dose at the inner entrance of the dose equivalent conversion factor within a factor of 1.0–2.0
maze, using the conversion factor given by the International in the energy range from 0.025 eV 共thermal兲 to 5 MeV 共data
Commission on Radiological Protection 共ICRP兲 Publication from manufacturer兲. McCall and Swanson24 pointed out that
74.21 According to Falcão et al.’s observation, the fluence the photoneutron spectrum outside a concrete shielding re-
should be multiplied by 2 before the conversion. Then the sembles a heavily shielded fission spectrum and the average
Hn,D at the entrance can be calculated by using the TVD energy drop is obvious. According to studies of Kry et al.
value given by Eq. 共5兲. and Howell et al. on a Varian Clinac 21EX accelerator
The dose equivalent due to the neutron capture ␥-rays 共Varian Medical Systems, Inc., Palo Alto, CA兲 operating at
共h␸兲 is a major component of the total photon dose equiva- 15, 18, and 20 MV, the neutron spectrum at 40 cm from the
lent HG at the door. McGinley et al.9 proposed a method to isocenter has a maximum energy below 10 MeV and the
calculate the neutron capture gamma dose equivalent. Ac- average neutron energy was reported between 0.24 and 0.61
cording to the method, the dose equivalent from the neutron MeV.25,26 Liu et al.27 reported that the average neutron en-
capture ␥-rays at the outer maze entrance per unit absorbed ergy could be as low as 0.5 MeV at 1 m away from the
dose of x-rays at the isocenter is given by gantry head of an 18 MV linac. Based on these published
reports, we have assumed that the neutron spectrum at the
h␸ = K␸A10−共d2/TVD兲 , 共6兲
maze entrance has an average energy lower than 0.5 MeV.
where the TVD is 5.4 m for x-ray beams in the range of Hence, the NP-2 Rem meter is suitable for neutron field mea-
18–25 MV and K is the ratio of the neutron capture gamma surement outside of a concrete treatment room.
dose equivalent to the total neutron fluence and is 6.9 A Ludlum 14C survey meter with a Model 44-6 sidewall
⫻ 10−16 Sv m2 as suggested in NCRP Report No. 151.22 Geiger Müller probe 共Ludlum Measurements, Inc., Sweetwa-
Other components of the HG include the dose equivalent due ter,TX兲 and a Victoreen 450P pressurized ion chamber sur-
to the scattering of the primary beam from the room surfaces vey meter 共Victoreen, Inc., Cleveland, OH兲 were used at the
共HS兲, the primary beam scattered from the patient or phan- same time to measure the photon dose equivalent HG. The
tom 共Hps兲, the scattered head leakage photons 共HLS兲, and the results from these two photon dose survey meters were
leakage radiation which is transmitted through the inner crosschecked with each other during the measurement and
maze wall 共HLT兲. These four components 共HS, Hps, HLS, and are in good agreement 共within 8%兲. The dose equivalent
HLT兲 can be calculated using the formulae for maze door readings from the two meters were averaged to get the mean
dose equivalent calculations as described in NCRP Report value.
No. 151. All the measurements were taken at 0.9 m above the floor.
In the Kersey method calculation, as described in Eq. 共1兲, The maze entrance measurements were taken at 0.3 m away
the H0 values were taken as 1.02– 1.6 mSv Gy−1 in this from the maze door 共point B in Fig. 2兲 on the maze center-
study. These are the values suggested by McGinley et al.19 line. The adjacent treatment console area measurements were
for Varian 1800 linacs 共Varian Medical Systems, Inc., Palo taken at 4.0 m away from the maze door and 0.4 m inside of
Alto, CA兲 which have a similar gantry head design with the the console area entrance 共point F in Fig. 2兲, where the maze
Varian Clinac 23EX. In the modified Kersey method and door is just visible. The points A–E in Fig. 2 are the same
neutron capture gamma dose equivalent calculations, the Qn with the points A–E shown in Fig. 1. The lengths of AB 共the
used in the calculation of the ␸A was taken as from 0.87 d2 in Fig. 1兲 and AC 共the d1 in Fig. 1兲 are 5.3 m and 6.1 m,
⫻ 1012 to 1.22⫻ 1012 Gy−1, as suggested by McGinley et al. respectively.

Medical Physics, Vol. 38, No. 3, March 2011

1144 Wang et al.: Neutron dose evaluation and shielding 1144

Neutron dose equivalent rate with closed field

Neutron dose equivalent rate with open field
1200 Photon dose equivalent rate with closed field
Photon dose equivalent rate with open field

1000

Hn,D and HG (μSv/hr)

800

600

400

200

0
90 180 270
Gantry angle (degree)

FIG. 3. The measured Hn,D and HG with different gantry angles and jaw
settings. The closed and open fields correspond to a field size of 0.5⫻ 0.5
and 40⫻ 40 cm2, respectively. Both units of the Hn,D and the HG were
converted to dose equivalent rate 共␮Sv h−1兲.
FIG. 2. Scheme of the point of measurement in this study. The letters from
a to l represent the possible BPE board locations in the maze. The locations
d–i are where the BPE boards were finally installed. Points A–E are the 478 keV ␥-ray will be emitted from 7Li after the neutron
same to those in Fig. 1. When the gantry angle is 270°, the gantry head is absorption, which increases the total photon dose equivalent
close to the west wall and the beam is directed from point D toward point E.
rate.

III. RESULTS
The neutron dose equivalent measurement uncertainty
was estimated to be ⫾15% in this study. The photon dose III.A. Measurements
equivalent measurement uncertainty was estimated to be
A series of measurements was first carried out to find the
⫾11%. The measurement uncertainty is composed of the un-
conditions of the Varian Clinac 23EX accelerator that pro-
certainties from the meter calibration, the measurement ge-
duces the largest dose equivalent rate at the door 共point B in
ometry 共such as the measurement distance and the angular
Fig. 2兲. The linac was operated at 18 MV photon mode with
dependence of the meters兲, and the repeatability of measure-
output dose rate of 600 MU min−1, which is the maximum
ment readings.
output rate of the linac. The beam irradiated a 30⫻ 30
⫻ 20 cm3 acrylic phantom at 90°, 180°, and 270° gantry
II.C. Shielding angles, respectively. The centroid of the phantom was placed
at the isocenter. The field sizes were set at 40⫻ 40 共open jaw
The maze door in this study is made of 0.64 cm 共0.25 in兲
setting兲 and 0.5⫻ 0.5 cm2 共closed jaw setting兲.
thick tempered glass. The major function of the maze door is
In order to make a direct comparison between the results
to act as a safety interlock. Its interaction with the photon
of measurements 共in mR h−1 or mrem h−1兲 and calculations
and neutron field was ignored. After replacing the old linac
共in Sv Gy−1 or mSv Gy−1兲, both units of the Hn,D and the HG
共with 6 MV photon beam only兲 with the new dual energy
were converted to dose equivalent rate, ␮Sv h−1 in this pa-
Varian Clinac 23 EX accelerator 共with 6 MV and 18 MV
per. The measurement unit mR h−1 was taken as equivalent
photon beam兲, the increment of neutron and photon dose
to mrem h−1 共10 ␮Sv h−1兲 in soft tissue. The quality factor is
equivalent required further shielding at the maze door and its
1 for photons. The calculation unit Sv Gy−1 or mSv Gy−1
peripheral region. A new shielding door at the maze entrance
was converted to ␮Sv h−1 using the linac output dose rate
used to reduce the dose equivalent was proposed. However,
共600 MU min−1兲.
the application of using borated polyethylene boards was
found to be more economical and feasible. Six pieces of
borated polyethylene boards 共5% boron element in weight in III.A.1. Neutron dose equivalent
the form of boric oxide, NELCO, Inc., Woburn, MA兲 were The measured neutron dose equivalent Hn,D was shown as
secured to the maze wall using metal screws at positions d–i a function of the gantry angles in Fig. 3. When the jaws are
as seen in Fig. 2. Each board has a size of 2.5⫻ 244 open, the Hn,D increases from 500 to 700 ␮Sv h−1 as the
⫻ 122 cm3 共1 ⫻ 48⫻ 96 in3兲. The high concentration of hy- gantry angle changes from 90° to 180°, where 180° is with
drogen in polyethylene made it an effective material to ther- the beam aiming toward the floor. The neutron dose equiva-
malize fast neutrons. The 10B in the BPE boards has very lent decreases to 400 ␮Sv h−1 when the gantry angle
high thermal neutron absorption cross section 共␴ = 3840 b兲. changes to 270°. With the jaws closed, the Hn,D increases
Therefore, the BPE is a very effective material for neutron with the gantry angle 共when the gantry head gets close to the
shielding. The neutron capture ␥-rays produced by the maze inner maze entrance兲. The Hn,D at 90°, 180°, and 270° are
wall will also decrease with the neutron fluence. However, a 700, 850, and 900 ␮Sv h−1, respectively.

Medical Physics, Vol. 38, No. 3, March 2011

1145 Wang et al.: Neutron dose evaluation and shielding 1145

TABLE I. Comparison of the measured dose equivalent and the calculation results at the maze entrance of the
Varian Clinac 23EX accelerator. The measurement results and the calculation to measurement ratios were
shown with uncertainties 共in parenthesis兲.

Neutron Hn,D
共␮Sv h−1兲
Photon HG
共␮Sv h−1兲 Kersey Modified Kersey Falcão

Measurement 305 共34兲 925 共139兲 925 共139兲 925 共139兲

Calculation 234–298 2251–3532 585–820 3611–5063
Ratio 0.77共8兲–0.98共11兲 2.43共36兲–3.82共57兲 0.63共9兲–0.89共13兲 3.90共59兲–5.47共82兲

III.A.2. Photon dose equivalent approximately 12 ␮Sv h−1; the dose equivalent due to the
The measurement results of the total photon dose equiva- primary beam scattered from the patient or phantom 共Hps兲 is
lent HG were also shown as a function of the gantry angles in approximately 45 ␮Sv h−1; the dose equivalent due to the
Fig. 3. During the measurements, the scattering and leakage scattered head leakage photons 共HLS兲 is approximately
photons as well as the neutron capture ␥-rays were detected 16 ␮Sv h−1; and the dose equivalent due to the leakage ra-
and they were not distinguishable from each other in the diation which is transmitted through the inner maze wall
photon survey meters. With the maximum jaw openings 共HLT兲 is approximately 2 ␮Sv h−1.
共40⫻ 40 cm2兲, the HG is 260 and 145 ␮Sv h−1, respectively, Using the semiempirical calculations described in Sec.
for 90° and 270° gantry angles. The HG reaches its maximum II A, the calculation results of neutron dose equivalent rate
共275 ␮Sv h−1兲 of the largest open jaw settings at the 180° were obtained by the Kersey method, the modified Kersey
gantry angle. When the jaws are closed, the HG at 90°, 180°, method, and the Falcão method. The calculation results are
and 270° are 200, 295, and 300 ␮Sv h−1, respectively. listed in Table I, with their counterparts of the maximum
measurement results. The maximum of measured Hn,D and
III.A.3. Maximum reading setup HG were obtained using the conditions described in Sec.
III A, with the gantry at 270° and the closed jaw setting. The
A similar trend of dose equivalent as a function of gantry ratios of the calculated values to the measured ones are also
angle was observed for both neutrons and photons. Both the listed in Table I.
Hn,D and the HG reach their maximum at 180° gantry angle
when the jaws are open and at 270° when the jaws were
closed. The results show that the Hn,D and the HG at the door III.C. Dose measurement for shielding
are at their overall maximum when the gantry is at 270° and
In order to investigate the shielding effect of lining por-
with the jaws closed 共0.5⫻ 0.5 cm2兲 共Fig. 3兲. At the gantry
tions of the maze wall with borated polyethylene, a series of
angle 270°, the gantry head is close to the west wall and the
measurements were performed with one or multiple 共up to
inner maze entrance and the beam is directed from point D
six兲 BPE boards placed on the maze wall at locations a–l
toward point E 共Fig. 2兲.
共Fig. 2兲. The Hn,D and the HG were measured using the con-
ditions described in Sec. III A, i.e., with the gantry at 270°
III.B. Semiempirical calculations
and the closed jaw setting. According to the single board
The maximum of the calculated total photon dose equiva- shielding effect and comparison of different location combi-
lent 共HG兲 is the sum of its five components: HS, Hps, HLS, nations of multiple boards, we chose d–i as the optimized
HLT, and h␸. Based on Eq. 共6兲 with different values for Qn locations to install the six pieces of the BPE boards. Lining
共0.87⫻ 1012 – 1.22⫻ 1012 Gy−1兲, the calculated neutron cap- the inner corner of the maze at locations d–i with BPE
ture gamma dose equivalent h␸ is 159– 223 ␮Sv h−1. Based boards provides 59% and 41% reductions for the HG and the
on the formulae in NCRP Report No. 151 for maze door dose Hn,D, respectively, at the maze entrance 共Table II兲. Also, a
equivalent calculations, the dose equivalent due to the scat- 56% reduction was achieved for both the dose equivalents at
tering of the primary beam from the room surfaces 共HS兲 is point F in the adjacent console area 共Fig. 2兲.

TABLE II. Final measurements at the maze entrance and the control area of adjacent vault.

Readings and reduction at entrance Readings and reduction at console area of adjacent vault

Photon Neutron Photon Neutron

共␮Sv h−1兲 共␮Sv h−1兲 共␮Sv h−1兲 共␮Sv h−1兲

Reading before shielding 305 925 25 45

Reading after boron shielding 125 550 11 20
Reduction 59% 41% 56% 56%

Medical Physics, Vol. 38, No. 3, March 2011

1146 Wang et al.: Neutron dose evaluation and shielding 1146

IV. DISCUSSION lent with gantry angle is observed through measure-

ments with the minimum field size 共0.5⫻ 0.5 cm2兲.
IV.A. Maximum reading setup in measurements
The trend of the curves of the Hn,D and the HG are
The observations of the maximum reading setup 共270° close to each other. The HG is 50% greater at 270° than
gantry angle and closed jaw setting兲 are consistent with the at the 90° gantry angle, while the Hn,D is 29% greater.
measurements by McGinley and Huffman,28 and the Monte Since the HLS and HLT will not change dramatically
Carlo calculations by Ma et al.29 It is more likely for a scat- with the gantry angle according the calculation results,
tered neutron to escape to the maze when the gantry is at we assume that the large increment of the HG 共50%兲 is
270° as opposed to 90°. Furthermore, when the jaws are majorly caused by the neutron capture ␥-rays. Al-
closed, the average neutron energy will decrease25 and there- though the neutron capture ␥-rays component cannot
fore the neutron absorption cross section may increase. This be quantitatively distinguished from the leakage pho-
will result in a higher neutron dose equivalent and neutron tons, the measurement data indicate that the neutron
capture gamma dose equivalent. capture ␥-rays are a major, possibly a dominant, com-
It is worth mentioning that the major component of the ponent of the photon field at the maze door when the
error bars in Fig. 3 is systematic and does not affect the jaws are closed.
relative relationship between the measurement data 共of either
the neutron or the photon measurement兲. Repeated measure-
ments support the above observation of this trend. IV.B.3. Primary beam scattering photons in
measurements
IV.B. Photon dose equivalent components 共a兲 Open field. All five components of the photon fields,
IV.B.1. Components in the calculations h␸, HS, Hps, HLS, and HLT, contribute to the total pho-
All five components 共HS, Hps, HLS, HLT, and h␸兲 of pho- ton dose equivalent at the maze door. Based on the
ton dose equivalent will contribute to the total photo dose calculation results in Sec. III B, the HLS and the HLT
equivalent 共HG兲 at the maze door when the jaws are open. were assumed to be constant when the gantry rotates.
According to the calculation results in Sec. III B, the HS and The significance of the primary beam scattering pho-
the Hps contribute about 19%–24% of the HG at different tons can be observed in the measurement data. When
gantry angles. When the jaws are closed, the HS and the Hps the gantry is at 90°, the primary beam will irradiate the
can be ignored and the HLS and the HLT contribute about west wall 共Fig. 2兲 so that the primary beam has a
8%–10% of the HG. Changing the gantry angle will cause a higher probability of being scattered into the maze than
variation of 10% of the sum of the HLS and the HLT and only at the 270° gantry angle. The HG increases correspond-
about 1% of the HG with the close jaw setting. According to ingly. When the gantry turns from 90° to 180°, the
the calculations, the neutron capture gamma dose equivalent primary beam has a lower probability to be scattered
is the dominant component of the HG, especially for closed into the maze. However, the HG increases from
jaws 共68%–75% for the open jaw setting and 90%–92% for 260 ␮Sv h−1 at 90° to 275 ␮Sv h−1 at 180°. This may
the closed jaw setting兲. be due to the increase of the neutron capture ␥-rays
with the increasing neutrons fluence in the maze when
the gantry turns from 90° to 180° 共Fig. 3兲.
IV.B.2. Neutron capture ␥-rays in measurements 共b兲 Closed field. The primary beam scatter can be ignored
共a兲 Open field. The neutron dose equivalent is 40% greater for the closed field 共0.5⫻ 0.5 cm2兲. If the primary scat-
at the 180° gantry angle than at 90° where the gantry tering photons are dominant, the HG for the closed field
head is furthest away from the inner maze entrance. should be lower than that for the maximum open field
The neutron dose equivalent is 43% smaller when the 共40⫻ 40 cm2兲. However, the HG only has a lower
gantry is at 270° than at the 180° gantry angle. Corre- value at the 90° gantry angle for the closed field. At the
spondingly, the photon dose equivalent is 5% smaller at 180° gantry angle, the HG for the closed field is 7%
the 90° gantry angle and 47% smaller at 270° when greater than that for the open field. At the 270° gantry
compared to the gantry at 180° 共beam pointing down兲 angle, the HG for the closed field is 107% greater than
for the maximum open field size 共40⫻ 40 cm2兲. The that for the open field. The measurement data suggest
similar trend of the curves suggests that at the maze that the primary beam scattering photons are a major
door, the photon dose is largely affected by the neutron component of the total photon dose equivalent, but are
field and that the neutron capture ␥-rays are a major not dominant.
component of the photon field. However, the much
smaller dose increment of photons than that of neutrons
from 90° to 180° gantry angle indicates that the HG is IV.B.4. Counteractive effect of neutrons capture
also determined by other components of photons, such ␥-rays and primary beam scattering photons
as primary beam scattering photons, when the jaws are From the above discussions, it should be noticed that
at their maximum settings. there is a counteractive effect from the neutron capture
共b兲 Closed field. The increase of the photon dose equiva- ␭-rays and the primary beam scattering photons to the total

Medical Physics, Vol. 38, No. 3, March 2011

1147 Wang et al.: Neutron dose evaluation and shielding 1147

photon dose equivalent. The HG is mainly affected by the ing component did not exist or was negligible during the
following factors: The variation of the primary beam scatter- measurement. However, the calculated HG without the pri-
ing photons with the gantry angle, the variation of the neu- mary beam scattering components 共HS and Hps兲 is only
tron fluence in the maze with the gantry angle, and the varia- 177– 241 ␮Sv h−1 and is lower than the highest possible
tion of the neutron capture gamma production with the dose equivalent in the calculation, which includes all the
change of average neutron energy. components. There is a contradiction of the conditions to
obtain the maximum HG from the measurements and the cal-
共a兲 Open field. The scattering of the primary photon beam
culations. It may partially be caused by the same K value
at the maze door will increase when the gantry head
that was used in both the h␸ calculations with the open jaw
moves away from the inner maze entrance and the pri-
setting and the closed jaw setting. The parameter K is the
mary beam is turning toward the wall near to the inner
ratio of the neutron capture gamma dose equivalent to the
maze entrance. The neutron fluence in the maze is re-
total neutron fluence. With the jaws closed, the average en-
flected by the nonmonotonic Hn,D curve in Fig. 3. The
ergy of the neutrons will decrease.25 The neutron capture
trend of the HG curve with the open jaw setting is
gamma production per neutron fluence will increase due to
majorly affected by these two factors, i.e., the variation
the change of the neutron absorption cross section. There-
of the primary beam scattering photons and the varia-
fore, using the K value for both jaw settings does not reflect
tion of the neutron fluence in the maze.
the changes in the neutron average energy and the corre-
共b兲 Closed field. The HG is mainly affected by the variation
sponding neutron capture gamma production. Before the K
of the neutron fluence in the maze with the position of
values for different jaw settings are available, it is recom-
the gantry head with respect to the inner maze en-
mended to include all the possible components of the pho-
trance. It increases as the gantry head gets closer to the
tons in a treatment room shielding calculation.
inner maze entrance.
共c兲 From closed field to open field. The scattering of the
primary photon beam at the maze door will be higher
with the open field than that with the closed field. Fur-
thermore, the jaws are one of the major sources of neu- IV.C.2. Neutron dose equivalent calculations
tron production.26,30 The average neutron energy will
The results in Table I show that Kersey method overesti-
increase from 0.29 to 0.4 MeV when the jaws are
mates the measured Hn,D by a factor of 2.4–3.8. This result is
open,25 so that the neutron absorption cross section and
consistent with the observations of Carinou et al.4 and
the production of capture ␥-rays may be lower. Due to
McGinley et al.8 The calculated Hn,D by the modified Kersey
the above effects, the HG does not always decrease as
method underestimates the measured dose equivalent by
much as the Hn,D does when using the maximum open
11%–37%. The Falcão method overestimates the Hn,D by a
field versus the closed filed, as shown in Fig. 3. At 180°
factor of 3.9–5.5. With consideration of the measurement
gantry angle, the HG does not decrease significantly
uncertainty, the measurement data are in favor of the modi-
when using the maximum open field versus the closed
fied Kersey method. However, it is worth noting that the
field. The HG even increases by 30% at 90° when jaws
modified Kersey method has the tendency to underestimate
change from the closed jaw setting to the maximum
the Hn,D and should be used with extra caution.
open jaw setting.
In the Falcão method, the Hn,D calculation accuracy de-
pends largely on the knowledge of the average neutron en-
IV.C. Comparison of measurement and semiempirical
ergy at the inner maze entrance. The average neutron energy
calculations could drop from about 0.5 MeV at 1 m distance from the
gantry head27 to about 20 keV in the maze31 for accelerators
IV.C.1. Photon dose equivalent calculations working at 10–18 MV. According to Fig. 31 of the ICRP
The calculated HG underestimates the measured dose Publication 74,21 there is a steep slope of the conversion
equivalent by 2%–23%. The calculated HG is the sum of factors in the neutron energy range from 20 keV to 0.5 MeV.
contributions from the primary beam scattering photons 共HS The value of the conversion factor increases 19 times in this
and Hps兲, the gantry head leakage 共HLS and HLT兲, and the energy range. Determining on the accurate average neutron
neutron capture ␥-rays 共h␸兲. Including all these contributions energy for different accelerators and maze designs will be a
in the calculation will give the highest possible prediction of challenge for those who tend to use the Falcão method to do
the dose equivalent at the maze door. With consideration of the semiempirical shielding calculations. Additional neutron
the measurement uncertainty, the formulae in NCRP Report average energy information should be available for different
No. 151, including Eq. 共6兲 proposed by McGinley et al., give medical accelerators in order to apply this method accurately.
a very close estimation to the HG. However, the tendency to However, the modified Kersey method relies on the neutron
underestimate the HG needs to be carefully considered in source strength Qn, which is available for most commercial
shielding calculations. linacs, and the geometry information of the maze, which is
It should be noted that the maximum photon dose equiva- usually known to the linac users. Therefore, the modified
lent HG obtained in the measurement happened with the Kersey method may still have an important advantage in
closed jaws setting, which means the primary beam scatter- practical shielding calculations.

Medical Physics, Vol. 38, No. 3, March 2011

1148 Wang et al.: Neutron dose evaluation and shielding 1148

IV.D. Dose measurement for shielding ACKNOWLEDGMENTS

The similar trend of dose reduction of both the neutron
The authors are deeply indebted to the staff of Department
and photon field supports the conclusion that neutron capture
of Environmental Health and Safety, students and technicians
␥-rays are a major component of the photon field at maze
at Cancer Therapy Research Center, at University of Texas
entrance. After the shielding, the maze has a 4.39 m TVD
Health Science Center at San Antonio: Christy Shriver, Jen-
共derived by the modified Kersey method兲, which is 15%
nifer Watson, Robert O. Moreno, Dennis Tyll, Nestor Ruiz-
shorter than the TVD before the shielding 共5.15 m兲. The Hn,D
Gonzalez, and Jeffrey Peyton.
reduction after the shielding 共41%兲 is close to the Monte
Carlo simulation results, i.e., 65% reported by Carinou et al.4 a兲
Electronic mail: [email protected]
and 69% by Agosteo et al.32 Agosteo et al. simulated stan- 1
W. P. Swanson, “Improved calculation of photoneutron yield released by
dard concrete wall plus 2.5 cm BPE with 4% boron in incident electrons,” Health Phys. 37, 347–358 共1979兲.
2
C. Ongaro, J. Rodenas, A. Leon, J. Perez, A. Zanini, and K. Burn, “Monte
weight, which is less than the 5% boron in the BPE used by
Carlo simulation and experimental evaluation of photoneutron spectra
Carinou et al. and our group. In the above two Monte Carlo produced in medical linear accelerators,” in Proceedings of the 1999 Par-
simulations, the whole maze wall was covered with BPE ticle Accelerator Conference, New York, 1999 共unpublished兲.
3
boards as opposed to our study where we only lined the inner C. Ongaro, A. Zanini, U. Nastasi, J. Rodenas, G. Ottaviano, and C. Man-
fredotti, “Analysis of photoneutron spectra produced in medical accelera-
corner of the maze with BPE boards. Therefore, the higher tors,” Phys. Med. Biol. 45, L55–L61 共2000兲.
reduction values reported by Carinou et al. and Agosteo et 4
E. Carinou, V. Kamenopoulou, and I. E. Stamatelatos, “Evaluation of
al. are reasonable. A similar shielding method as in our study neutron dose in the maze of medical electron accelerators,” Med. Phys.
reported by Lalonde11 shows that using polyethylene and 26共12兲, 2520–2525 共1999兲.
5
P. H. McGinley, A. H. Dhaba’an, and C. S. Reft, “Evaluation of the
flexboron panels as the shielding materials can reduced the contribution of capture gamma rays, x-rays leakage, and scatter to the
Hn,D by 50% and the HG by 32%, which are close to our photon dose at the maze door for a high energy medical electron accel-
results. erator using a Monte Carlo particle transport code,” Med. Phys. 27共1兲,
225–230 共2000兲.
6
R. C. Falcão, A. Facure, and A. X. Silva, “Neutron dose calculation at the
maze entrance of medical linear accelerator rooms,” Radiat. Prot. Dosim.
123共3兲, 283–287 共2007兲.
7
V. CONCLUSIONS R. W. Kersey, “Estimation of neutron and gamma radiation doses in the
entrance mazes of SL 75-20 linear accelerator treatment rooms,” Medica-
This series of measurements provides neutron and photon mundi 24共3兲, 151–155 共1979兲.
8
dose equivalent data at the maze entrance and at the adjacent P. H. McGinley and E. K. Butker, “Evaluation of neutron dose equivalent
levels at the maze entrance of medical accelerator treatment rooms,” Med.
treatment console area of a Varian Clinac 23EX medical ac- Phys. 18共2兲, 279–281 共1991兲.
celerator operated at 18 MV photon mode. The neutron and 9
P. H. McGinley, M. S. Miner, and M. L. Mitchum, “A method for calcu-
photon dose equivalent measurement data indicate that the lating the dose due to capture gamma rays in accelerator mazes,” Phys.
neutron capture ␥-rays are a major and possibly a dominant Med. Biol. 40, 1467–1473 共1995兲.
10
R. K. Wu and P. H. McGinley, “Neutron and capture gamma along the
component of the photon field at the maze entrance. These mazes of linear accelerator vaults,” J. Appl. Clin. Med. Phys. 4共2兲, 162–
results also compare favorably to the modified Kersey 171 共2003兲.
11
method. The calculations of the modified Kersey method are R. Lalonde, “The effect of neutron-moderating materials in high-energy
linear accelerator mazes,” Phys. Med. Biol. 42, 335–344 共1997兲.
in agreement with the measurements within 11%–37% for 12
D. S. Followill, M. S. Stovall, S. F. Kry, and G. S. Ibbott, “Neutron
our maze design. However, the results show that the Kersey source strength measurements for Varian, Siemens, Elekta, and General
method overestimates the Hn,D by 2.4–3.8 times and the Fal- Electric linear accelerators,” J. Appl. Clin. Med. Phys. 4共3兲, 189–194
cão method overestimates the Hn,D by 3.9–5.5 times for our 共2003兲.
13
J. C. Rivera, R. C. Falcão, and C. E. deAlmeida, “The measurement of
maze design. The large discrepancy of the Falcão method photoneutron dose in the vicinity of clinical linear accelerators,” Radiat.
was partially attributed to the ICRP conversion factors based Prot. Dosim. 130共4兲, 403–409 共2008兲.
14
on estimations of average neutron energy. The widely used R. C. McCall, “Neutron yield of medical electron accelerators,” SLAC
modified Kersey method is recommended for the shielding Publication, No. 4480, 1987.
15
P. H. McGinley and J. C. Landry, “Neutron contamination of x-ray beams
calculation of a Varian Clinac 23EX accelerator in a similar produced by the Varian Clinac 1800,” Phys. Med. Biol. 34, 777–783
maze, with caution to its potential to underestimate the Hn,D. 共1989兲.
16
Borated polyethylene is an efficient material in neutron P. H. McGinley, S. Ghavidel, and J. C. Landry, “A study of photoneutron
dose reduction as well as neutron capture gamma dose re- does levels produced by the Phillips SL medical accelerators,” Radiat.
Prot. Manage. 10, 45–50 共1993兲.
duction. After applying BPE boards as a lining material on 17
J. O. Fenn and P. H. McGinley, “Stray photoneutron fields produced by
the maze wall, the Hn,D at the maze entrance was decreased the GE Saturne accelerator,” Radiat. Prot. Manage. 12, 39–45 共1995兲.
18
by 41% and the HG was decreased by 59% when compared R. C. McCall, T. M. Jenkins, and R. A. Shore, “Transport of acceleration
to the unshielded measurements. The maze in this study has produced neutrons in a concrete room,” IEEE Trans. Nucl. Sci. NS-26,
1593–1602 共1979兲.
a TVD of 5.15 and 4.39 m before and after lining the inner 19
P. H. McGinley, Shielding Techniques for Radiation Oncology Facilities,
corner of the maze wall with 2.5 cm 共1 in兲 BPE boards, 2nd ed. 共Medical Physics, Madison, 2002兲.
20
according to the modified Kersey method. This shielding ap- R. C. McCall, P. H. McGinley, and K. E. Huffman, “Room scattered
neutrons,” Med. Phys. 26共2兲, 206–207 共1999兲.
proach can be applied to retrofitting an existing vault as a 21
ICRP, ICRP Publication 74 共Pergamon, New York, 1996兲.
result of linear accelerator energy upgrades, especially when 22
NCRP, “Structural shielding design and evaluation for megavoltage x-
shielding cost is a concern. and gamma-ray radiotherapy facilities,” NCRP Report No. 151 共National

Medical Physics, Vol. 38, No. 3, March 2011

1149 Wang et al.: Neutron dose evaluation and shielding 1149

28
Council on Radiation Protection and Measurements, Bethesda, MD, P. H. McGinley and K. E. Huffman, “Photon and neutron doses equiva-
2005兲. lent in the maze of a high-energy medical accelerator facility,” Radiat.
23
I. O. Anderson and J. Braun, “A neutron REM counter,” Aktiebolaget Prot. Manage. 17, 43–46共E兲 共2000兲.
Atomenergi Report AE-132, 1964. 29
A. Ma, J. Awotwi-Pratt, A. Alghamdi, A. Alfuraih, and N. M. Spyrou,
24
R. C. McCall and W. P. Swanson, “Neutron sources and their character- “Monte Carlo study of photoneutron production in the Varian Clinac
istics,” SLAC Publication No. 2292共A兲, 1979. 2100C Linac,” J. Radioanal. Nucl. Chem. 276共1兲, 119–123 共2008兲.
25
R. M. Howell, S. F. Kry, E. Burgett, D. Followill, and N. E. Hertel, 30
X. S. Mao, K. R. Kase, J. C. Liu, W. R. Nelson, J. H. Kleck, and S.
“Effects of tertiary MLC configuration on secondary neutron spectra from
Johnsen, “Neutron sources in the Varian Clinac 2100C/2300C medical
18 MV x-ray beams for the Varian 21EX linear accelerator,” Med. Phys.
accelerator calculated by the EGS4 code,” Health Phys. 72共4兲, 524–529
36共9兲, 4039–4046 共2009兲.
26
S. F. Kry, R. M. Howell, U. Titt, M. Salehpour, R. Mohan, and O. N. 共1997兲.
31
Vassiliev, “Energy spectra, sources, and shielding considerations for neu- J.-P. Lin, T.-C. Chu, S.-Y. Lin, and M.-T. Liu, “The measurement of
trons generated by a fattening filter-free Clinac,” Med. Phys. 35共5兲, 1906– photoneutrons in the vicinity of a Siemens Primus linear accelerator,”
1911 共2008兲. Appl. Radiat. Isot. 55, 315–321 共2001兲.
27 32
J. C. Liu, K. R. Kase, X. S. Mao, W. R. Nelson, J. H. Kleck, and S. A. Agosteo, A. F. Para, B. Maggioni, V. Sangiust, S. Terrani, and F.
Johnson, “Calculations of photoneutrons from Varian Clinac accelerators Borasi, “Radiation transport in a radiotherapy room,” Health Phys. 68共1兲,
and their transmissions in materials,” SLAC Publication No. 7404, 1997. 27–34 共1995兲.

Medical Physics, Vol. 38, No. 3, March 2011