Download Simple methods to reduce patient dose in a

The British Journal of Radiology, 82 (2009), 855–859
SHORT COMMUNICATION
Simple methods to reduce patient dose in a Varian cone beam CT
system for delivery verification in pelvic radiotherapy
1
P ROXBY, MSc, 1,2T KRON,
and 1J CRAMB, MSc
PhD,
3
F FOROUDI,
MD,
1,2
A HAWORTH,
PhD,
1
C FOX,
PhD,
1
A MULLEN,
MSc
1
Peter MacCallum Cancer Centre, Department of Physical Sciences, St Andrews Place, East Melbourne, Victoria 3002,
RMIT University, Department of Applied Physics, Melbourne and 3Peter MacCallum Cancer Centre, Department of
Radiation Oncology, St Andrews Place, East Melbourne, Victoria 3002, Australia
2
ABSTRACT. Cone-beam computed tomography (CBCT) is a three-dimensional imaging
modality that has recently become available on linear accelerators for radiotherapy
patient position verification. It was the aim of the present study to implement simple
strategies for reduction of the dose delivered in a commercial CBCT system. The dose
delivered in a CBCT procedure (Varian, half-fan acquisition, 650 projections, 125 kVp) was
assessed using a cylindrical Perspex phantom (diameter, 32 cm) with a calibrated Farmer
type ionisation chamber. A copper filter (thickness, 0.15 mm) was introduced increasing
the half value layer of the beam from 5.5 mm Al to 8 mm Al. Image quality and noise were
assessed using an image quality phantom (CatPhan) while the exposure settings per
projection were varied from 25 ms/80 mA to 2 ms/2 mA per projection. Using the copper
filter reduced the dose to the phantom from approximately 45 mGy to 30 mGy at
standard settings (centre/periphery weighting 1/3 to 2/3). Multiple CBCT images were
acquired for six patients with pelvic malignancies to compare CBCTs with and without a
copper filter. Although the reconstructed image is somewhat noisier with the filter, it
features similar contrast in the centre of the patient and was often preferred by the
radiation oncologist because of greater image uniformity. The X-ray shutters were
adjusted to the minimum size required to obtain the desired image volume for a given
patient diameter. The simple methods described here reduce the effective dose to patients
undergoing daily CBCT and are easy to implement, and initial evidence suggests that they
do not affect the ability to identify soft tissue for the purpose of treatment verification.
Cone beam computed tomography (CBCT) is a threedimensional imaging modality that has recently become
available on linear accelerators used for radiotherapy
treatment [1, 2]. The systems commercially available
include megavoltage imaging [3] as well as kilovoltage
imaging systems [2, 4]. For the latter, linear accelerators
are equipped with a diagnostic imaging set (X-ray tube
and flat panel imaging device) mounted at 90 ˚ to the
treatment beam on the gantry.
CBCTs are reconstructed from projections acquired
during rotation of the gantry by >180˚. The maximum
speed of rotation is 6˚ per second, a restriction imposed by
the International Electrotechnical Commission to reduce
the risk of serious collisions in rotating open gantry systems
[5]. The typical distance between the focal spot and the
detector is 150 cm, as illustrated in Figure 1. Given the high
cost of large flat panel devices, most commercial systems
Address correspondence to: Tomas Kron, PhD, Principal Research
Physicist, Peter MacCallum Cancer Centre, Department of Physical
Sciences, Locked Bag 1, A’Beckett St, Victoria 8006, Australia. E-mail:
[email protected]
The present work was supported in part by a research collaborative
agreement between the Peter MacCallum Cancer Centre and Varian
Medical Systems.
The British Journal of Radiology, October 2009
Received 19 December
2007
Revised 21 October 2008
Accepted 21 October 2008
DOI: 10.1259/bjr/37579222
’ 2009 The British Institute of
Radiology
feature a panel width in the order of 40 cm. With the panel
centred on the axis of the X-ray beam, this would only
allow reconstruction of a cross-sectional image with a
diameter of 25 cm or less. However, if the detector panel is
moved laterally to one side and the X-ray beam collimated
in the same way, it is possible to acquire images in half fan
mode [6, 7]. This requires a 360 ˚ rotation of the gantry,
which will yield at least 180˚ of projections for every voxel
in the image. A small overlap of information in the centre of
the image at the medial aspect of the off-centre flat panel
device can be used for calibration. In order to facilitate the
two different modes of image acquisition (full fan: small
field of view (FOV); half fan: large FOV), the system allows
the mounting of two different bow tie filters.
For radiotherapy treatment delivery, CBCT constitutes
an excellent method to verify that the patient to be
treated is in the same position as they were for treatment
planning [4, 6, 8]. As most radiotherapy treatment
planning makes use of a diagnostic CT scan, the
verification CBCT reproduces the same geometry
imaged with the same modality as the treatment plan.
In addition to external contours, the CBCT can be used to
determine the location of the target and critical normal
structures, which allows optimisation of radiotherapy
855
P Roxby, T Kron, F Foroudi et al
Figure 1. Cone beam CT imaging
geometry for two patients of different sizes. Sup, superior; Inf, inferior.
delivery for every treatment day. Consequently, the most
comprehensive way of utilising this technology is to
acquire an image for each fraction of the treatment.
CT has been identified as a major contributor to the
overall radiation burden to populations in modern society
[9]. While there are clear benefits for the patient from this
image guidance procedure, it also results in a significant
additional dose to the patient, a fact that has led to a recent
report of task group 75 of the American Association of
Physicists in Medicine [10]. In the case of daily CT scanning,
the dose requires careful justification as it may be large, not
only in the region of the target but also in the surrounding
normal structures. Therefore, it is not surprising that dose
delivered in megavoltage [3, 11] and kilovoltage [12–14]
CBCT has received consideration recently.
There are several methods to control and reduce
radiation dose in CT (e.g. compare at www.impactscan.
org). They range from adding filtration to reduced field
sizes, exposure factors and number of projections. The
reduction of the number of projections is complicated in
the present version of the Varian CBCT software (version
2.0; Varian Medical Systems, Palo Alto, CA) and was not
considered. Similarly, a variation in tube potential was
not studied. The system utilises 125 kVp as a default,
which is the accelerating potential used as the standard
factory setting. The selection of a smaller tube potential
may improve the contrast but will generally increase the
radiation dose to patients. Finally, the Varian system
employs a single focused grid to reduce scatter at the
detector system. While the removal of the grid may allow
a reduction of dose by maintaining the photon flux at the
detector, the scatter to primary ratio is already large in
kilovoltage CBCT systems and a further increase would
result in more artefacts and image noise [2, 15].
It was the aim of the present study to determine and
implement simple strategies for reduction of the dose
delivered in CBCT following the general principle of dose
optimisation, i.e. to maintain exposures as low as reasonably practical, included in the recommendations of the
International Commission on Radiological Protection [16].
This was to be performed in the specific context of the
856
commercial Varian CBCT system available at our institution. Several methods of dose reduction were to be
investigated. These were: (i) the use of an additional
0.15 mm thick copper filter on the tube side of the supplied
bow tie filter, (ii) coning down using available collimation
so that only the volume of interest is irradiated and (iii)
reduction of the beam current (mA) per projection.
Methods and materials
CBCT system
All measurements were performed on a CBCTequipped Varian Trilogy linear accelerator with CBCT
2.0 software (Varian Medical Systems). The system uses a
CsI flat panel device with an active area of 30 6 40 cm2.
In our institution, a standard focus to detector distance of
150 cm is employed for all imaging. A single focused
grid is mounted directly above the detector system.
The system acquires 650 images over a 360˚ gantry
rotation with a few degrees added at both the start and the
end of the scan. The factory standard settings (Varian
Medical Systems) are 125 kVp with 80 mA and 25 ms per
projection, which is identical to the parameters used by
Song et al [14] in their recent study. With these settings, the
half value layer (HVL) of the beam is 4.6 mm Al as assessed
with a Unfors XI multimeter (Unfors Instruments,
Sweden). For CBCT acquisition, an aluminium (Al) bow
tie filter is employed that brings the HVL at the thinnest
part of the filter up to 5.5 mm Al. This is comparable with a
HVL of 5.7 mm Al reported in the literature [14]. One full
rotation takes 60 s and the typical frequency of projections
is up to 12 projections per second, resulting in nearly 2
projections per degree.
In our clinic, we use the default Shepp–Logan
convolution with modified Blackman window filter for
CBCT reconstruction. Other filters (Square, Hamming,
Bartlett and Welsh) are available in the software and
were tested but no significant improvement in image
quality was found. Images were routinely reconstructed
The British Journal of Radiology, October 2009
Short communication: Dose reduction in cone beam CT
with a 2.5 mm slice width and spacing, with a pixel
matrix of 512 6 512. The system allows the slice width to
be decreased to 1 mm. This was tested but not routinely
used clinically because of increased reconstruction times.
All experiments and patient images were acquired
using half fan CBCT with a nominal FOV between 35 cm
and 50 cm in diameter.
Dose reduction strategies
Three methods were studied to reduce the dose:
1. A copper filter (thickness, 0.15 mm) was mounted
behind the bow tie filter of the X-ray tube.
2. For accurate CBCT reconstruction, it is required to
have all parts of the patient to be imaged in the FOV
of all projections. Therefore, owing to beam divergence, the width of the cone beam in the superior–
inferior direction depends upon the required scan
length, L, and the thickness of the patient, T. This is
illustrated in Figure 1. From geometric principles, the
X-ray defining blades, b, were set as:
b~ððL=2Þ|100Þ=ð100{ðT=2ÞÞz0:25 cm
ð1Þ
2. The equation applies to a focus to axis distance of
100 cm and the addition of 0.25 cm allows for a small
safety margin.
3. Exposure settings were varied between the factory
setting of 25 ms/80 mA and a minimum setting of
2 ms/2 mA per projection.
region of interest of 100 pixels) was evaluated in three
homogeneous cylindrical water phantoms of 7–30 cm in
diameter and height extending beyond the FOV of the
CBCT image (.20 cm).
Patient studies
Patients
CBCT images were acquired as part of an in-house
clinical trial open for bladder and post-prostatectomy
patients. A total of six patients were enrolled in this
feasibility study. Patients who gave informed consent had
CBCTs taken daily throughout the first week of treatment.
Images were saved and used for off-line evaluation only.
The first three images were acquired with Varian standard
settings (80 mA, 25 ms per projection) and the bow tie
filter only (no copper filter). Image quality was assessed
by asking the clinician whether the CBCT dataset was
adequate for outlining of the bladder, rectum and prostate
(where applicable in bladder patients).
If the image quality of the first three images was
deemed to be acceptable, the copper filter was introduced. An additional two CBCTs were acquired and the
image quality assessed. If the introduction of the copper
filter did not affect the ability of the clinician to outline
structures, weekly images were taken with CBCT using
the copper filter.
Results and discussion
Phantom studies
Phantom studies
Dose assessment
Doses were measured at the centre and periphery
(1.2 cm from the edge) of a 32 cm diameter cylindrical
Perspex phantom with a thickness of 9.5 cm. It should be
noted that the limited thickness of computed tomography dose index (CTDI) phantoms can significantly
underestimate the contribution of scatter [17]. As such,
the dose measurements should be seen as indicative
only; however, the dose comparisons between irradiation with and without the filter, which are the main topic
of the paper, are not affected. A 0.6 cm3 Farmer type
ionisation chamber was used for the dose measurements
in conjunction with a NE 2670 electrometer (NE
Technology, UK). Calibration factors were derived from
values for superficial/orthovoltage radiotherapy treatment beams traceable to the Australian Radiation
Protection and Nuclear Safety Agency (ARPANSA).
Introduction of a copper filter
For the default technique settings, the dose in the CTDI
phantom was 23.7 mGy at the isocentre and 53.3 mGy at
the periphery. Using a weighting of 1/3 to 2/3, this
results in a weighted CTDI of approximately 45 mGy.
This can be compared with a value of 54 mGy reported
for the same procedure in the literature [14].
Using the copper filter of 0.15 mm thickness increased
the HVL of the beam to approximately 8 mm Al.
According to Amer et al [13], the Elekta XVI system
uses a copper filter of 0.127 mm thickness. Therefore, it is
not surprising that the beam quality reported by Song et
al [14] for the Elekta XVI system is 7 mm Al at 120 kVp,
which compares well with the HVL of 8 mm Al for
125 kVp found in our study for a slightly thicker copper
filter. The introduction of the copper filter not only
hardens the beam to correct a potential under filtering
[15], but also reduced the dose to the phantom from
approximately 45 mGy to 30 mGy at standard settings.
As expected for a harder radiation beam, the ratio
between dose in the periphery and the one in the centre
was reduced. The effect of the copper filter on image
quality can be seen in Figure 2, which shows patient
images with and without the additional filtration.
Image quality
Image quality was assessed using a CatPhan (The
Phantom Laboratory, Salem, NY) CT QA phantom with
spatial and contrast resolution objects. In order to mimic
the effect of large patients, additional scatter material
(4 cm of solid water) was added in a triangular structure
around the phantom. This brings the average thickness of
the phantom to approximately 30 cm. The effect of
different materials was studied using a disk phantom
with different materials. In addition to this, noise
(expressed as the standard deviation of CT numbers in a
The British Journal of Radiology, October 2009
Reduction of beam current
The effect of reduction of beam current (mA) and
exposure time (ms) per projection can be seen in Figure 3
for different phantom materials and sizes. The figures
857
P Roxby, T Kron, F Foroudi et al
(a)
(b)
(c)
Figure 2. CT images for the same bladder cancer patient (as Figure 1) taken (a) at planning (Philips Brilliance, 16 slice) and cone
beam CT (CBCT) images at treatment time (c) with and (b) without a copper (Cu) filter. For illustration purposes, all images are
displayed with the same window (2000) and level (0).
show the standard deviation of the CT numbers in small
areas of the image (100 voxels in a volume of approximately 10 6 10 6 2.5 mm3) for different materials and
homogeneous water phantoms of different diameters. As
can be seen in Figure 3b, for a dose reduction of 50% from
the standard settings of 80 mA and 25 ms per projection,
the noise increases significantly only for the largest
phantom size.
The CatPhan phantom experiments showed a small
increase in noise (six rather than seven of the 1% contrast
test objects were visible) with 50% of the default beam
current. Spatial resolution was assessed using the line
pair tool embedded in the Catphan phantom. The results
for two different observers are shown Figure 4. In order
to assess the impact of increased scatter, a separate set of
experiments was performed with the CatPhan phantom
set in a triangular structure of solid water slabs of 4 cm
thickness. This is termed ‘‘large phantom’’ in Figure 4
and it can be seen that the spatial resolution is slightly
reduced in this case.
Patient studies
Introduction of a copper filter
Multiple CBCT images were acquired for six patients
with pelvic malignancies to compare CBCTs with (n53)
and without (n53) a copper filter. Although the
reconstructed image is somewhat noisier with the filter,
it features similar contrast in the centre of the patient and
was scored at least of equal image quality by the
attending clinician. The improved image uniformity
was seen as an advantage. This is demonstrated in
Figure 2. The subjective ability to outline relevant
structures, such as the bladder and rectum, was not
affected by the introduction of the filter.
It has recently been shown by Moore et al [18] that filters
in CBCT can be further optimised using zonal filters, which
result in dose reduction and an improved image quality.
Adjustment of X-ray field size
It is good practice to reduce the irradiated length of the
patient to the area required for reconstruction of images.
Owing to beam divergence, this is affected also by the
858
Figure 3. (a) Cone beam CT (CBCT) number as a function of
exposure settings for the CBCT acquisition for water and
Perspex. As can be seen from the increasing standard
deviation for the p ix el values , a diff erence of
100 Hounsfield units (HU) can clearly be distinguished down
to approximately 200 mA ms per projection. (b) Standard
deviation (SD) of CT numbers in homogeneous water
phantoms of different diameters as a function of beam
current and exposure time per projection. The phantoms
extend over the whole width of the cone beam. In addition
to this, values for lung and bone are shown in a 20 cm
diameter slab phantom.
The British Journal of Radiology, October 2009
Short communication: Dose reduction in cone beam CT
easy to implement and appear not to affect the ability to
identify soft tissue for the purpose of treatment verification
obtained from the images.
References
Figure 4. Spatial resolution assessed by two operators using
a line pair tool in a CatPhan phantom. The results are shown
for the phantom with and without additional scatter
material (4 cm each side of the phantom).
patient diameter, as can be seen in Figure 1. A table of
shutter settings was developed that ensures that the
minimum possible section of the patient is irradiated. The
table is specific to a focus to detector (FDD) distance of
150 cm and includes a geometric ‘‘safety margin’’ of
0.25 cm in both the superior and inferior direction, as can
be seen in Equation 1. The shutter settings are manually
adjusted for each individual patient undergoing CBCT
acquisition in our institution. The dose reduction is
particularly noticeable for thin patients. Similar tables
can easily be developed from geometric principles for
different panel dimensions and FDDs.
Reduction of beam current
Following the phantom study, there appears to be scope
to reduce the beam current for smaller patients. Image
quality was just acceptable with the larger patients and
default beam current settings when improved owing to
increased beam hardening using the copper filter. As such,
we intend to reduce the beam current only for patients
with an average diameter of less than 30 cm as part of a
future study. If image quality is found to be insufficient,
the dose will be increased after two suboptimal images.
It is beyond the scope of the present study to prove
objectively that a reduction of exposure settings (mAs) does
not alter the ability of operators to act upon the images in
adaptive radiotherapy. Elucidating the required receiver
operating characteristics is planned as future work.
Conclusions
It is generally accepted that CT scanning is a ‘‘high dose’’
diagnostic procedure. Diagnostic CT scanners have developed over the past few years to include a large number of
dose reduction schemes (e.g. compare at www.impactscan.org). None of these has as yet been included in CBCT
technology used predominantly in radiotherapy. However,
given the potential for daily imaging in the context of
image-guided radiation therapy, dose optimisation must
be a priority. The simple methods described here reduce
the effective dose to patients undergoing daily CBCT, are
The British Journal of Radiology, October 2009
1. Jaffray DA, Siewerdsen JH, Wong JW, Martinez AA. Flatpanel cone-beam computed tomography for image-guided
radiation therapy. Int J Radiat Oncol Biol Phys 2002;53:
1337–49.
2. Moore CJ, Amer A, Marchant T, Sykes JR, Davies J,
Stratford J, et al. Developments in and experience of
kilovoltage X-ray cone beam image-guided radiotherapy.
Br J Radiol 2006;79:S66–78.
3. Pouliot J, Bani-Hashemi A, Chen J, Svatos M, Ghelmansarai
F, Mitschke M, et al. Low-dose megavoltage cone-beam CT
for radiation therapy. Int J Radiat Oncol Biol Phys
2005;61:552–60.
4. Sorcini B, Tilikidis A. Clinical application of image-guided
radiotherapy, IGRT (on the Varian OBI platform). Cancer
Radiother 2006;10:252–7.
5. International Electrotechnical Commission. Medical electrical equipment Part 2–1: particular requirements for the
safety of electron accelerators in the range of 1MeV to 50
MeV - IEC 60601-2-1. Geneva, Switzerland: IEC, 1998.
6. Smitsmans MH, de Bois J, Sonke JJ, Betgen A, Zijp L, Jaffray
DA, et al. Automatic prostate localization on cone-beam CT
scans for high precision image-guided radiotherapy. Int J
Radiat Oncol Biol Phys 2005;63:975–84.
7. Yang Y, Schreibmann E, Li T, Wang C, Xing L. Evaluation
of on-board kV cone beam CT (CBCT)-based dose calculation. Phys Med Biol 2007;52:685–705.
8. Thilmann C, Nill S, Tucking T, Hoss A, Hesse B, Dietrich L,
et al. Correction of patient positioning errors based on inline cone beam CTs: clinical implementation and first
experiences. Radiat Oncol 2006;1:16.
9. Brenner D, Hall E. Computed tomography — an increasing
source of radiation exposure. N Engl J Med 2007;357:2277–84.
10. Murphy M, Balter J, Balter S, BenComo J, Das I, Jiang S, et
al. The management of imaging dose during image-guided
radiotheraphy: report of the AAPM Task Group 75. Med
Phys 2007;34:4041–63.
11. Peng LC, Yang CC, Sim S, Weiss M, Bielajew A. Dose
comparison of megavoltage cone-beam and orthogonal-pair
portal images. J Appl Clin Med Phys 2007;8:10–20.
12. Islam MK, Purdie TG, Norrlinger BD, Alasti H, Moseley DJ,
Sharpe MB, et al. Patient dose from kilovoltage cone beam
computed tomography imaging in radiation therapy. Med
Phys 2006;33:1573–82.
13. Amer A, Marchant T, Sykes J, Czajka J, Moore C. Imaging
doses from the Elekta Synergy X-ray cone beam CT system.
Br J Radiol 2007;80:476–82.
14. Song W, Kamath S, Ozawa S Ani SA, Chvetsov A, Bhandare
N, et al. A dose comparison study between XVI and OBI
CBCT systems. Med Phys 2008;35:480–6.
15. Ding GX, Duggan DM, Coffey CW. Characteristics of kilovoltage X-ray beams used for cone-beam computed tomography in radiation therapy. Phys Med Biol 2007;52:1595–1615.
16. International Commission on Radiological Protection. The
1990 recommendations if the International Commission on
Radiological Protection, ICRP report 60. Oxford, UK:
Pergamon Press, 1991.
17. Perisinakis K, Damilakis J, Tzedakis A, Papadakis A,
Theocharopoulos N, Gourtsoyiannis N . Determination of
the weighted CT dose index in modern multi-detector CT
scanners. Phys Med Biol 2007;52:6485–95.
18. Moore C, Marchant T, Amer A. Cone beam CT with zonal
filters for simultaneous dose reduction, improved target
contrast and automated setup in radiotherapy. Phys Med
Biol 2006;51:2191–204.
859