Download Patient dose from kilovoltage cone beam computed tomography

Patient dose from kilovoltage cone beam computed tomography imaging
in radiation therapy
Mohammad K. Islama兲
Department of Radiation Physics, Princess Margaret Hospital and Department of Radiation Oncology,
University of Toronto, Toronto, Canada
Thomas G. Purdie
Department of Radiation Physics, Princess Margaret Hospital and Department of Radiation,
University of Toronto, Toronto, Canada
Bernhard D. Norrlinger
Department of Radiation Physics, Princess Margaret Hospital, Toronto, Canada
Hamideh Alasti, Douglas J. Moseley, and Michael B. Sharpe
Department of Radiation Physics, Princess Margaret Hospital and Department of Radiation Oncology,
University of Toronto, Toronto, Canada
Jeffrey H. Siewerdsen
Ontario Cancer Institute Princess Margaret Hospital and Department of Medical Biophysics,
University of Toronto, Toronto, Canada
David A. Jaffray
Department of Radiation Physics and Ontario Cancer Institute, Princess Margaret Hospital
and Department of Radiation Oncology and Department of Medical Biophysics,
University of Toronto, Toronto, Canada
共Received 26 August 2005; revised 21 March 2006; accepted for publication 24 March 2006;
published 10 May 2006兲
Kilovoltage cone-beam computerized tomography 共kV-CBCT兲 systems integrated into the gantry of
linear accelerators can be used to acquire high-resolution volumetric images of the patient in the
treatment position. Using on-line software and hardware, patient position can be determined accurately with a high degree of precision and, subsequently, set-up parameters can be adjusted to
deliver the intended treatment. While the patient dose due to a single volumetric imaging acquisition is small compared to the therapy dose, repeated and daily image guidance procedures can lead
to substantial dose to normal tissue. The dosimetric properties of a clinical CBCT system have been
studied on an Elekta linear accelerator 共Synergy® RP, XVI system兲 and additional measurements
performed on a laboratory system with identical geometry. Dose measurements were performed
with an ion chamber and MOSFET detectors at the center, periphery, and surface of 30 and
16-cm-diam cylindrical shaped water phantoms, as a function of x-ray energy and longitudinal
field-of-view 共FOV兲 settings of 5,10,15, and 26 cm. The measurements were performed for full
360° CBCT acquisition as well as for half-rotation scans for 120 kVp beams using the 30-cm-diam
phantom. The dose at the center and surface of the body phantom were determined to be 1.6 and
2.3 cGy for a typical imaging protocol, using full rotation scan, with a technique setting of 120 kVp
and 660 mAs. The results of our measurements have been presented in terms of a dose conversion
factor f CBCT, expressed in cGy/ R. These factors depend on beam quality and phantom size as well
as on scan geometry and can be utilized to estimate dose for any arbitrary mAs setting and
reference exposure rate of the x-ray tube at standard distance. The results demonstrate the opportunity to manipulate the scanning parameters to reduce the dose to the patient by employing lower
energy 共kVp兲 beams, smaller FOV, or by using half-rotation scan. © 2006 American Association of
Physicists in Medicine. 关DOI: 10.1118/1.2198169兴
Key words: kV cone beam CT, image guided radiation therapy, patient dose
I. INTRODUCTION
With the introduction of advanced imaging technologies into
the radiation therapy process such as: computerized tomography 共CT兲 simulators, magnetic resonance imaging 共MRI兲,
and positron emission tomography, the ability to define target
共GTV and CTV兲 and normal structures, as well as to assess
the magnitude of internal organ motion has improved sub1573
Med. Phys. 33 „6…, June 2006
stantially. Exceptional advancements have also been made in
the past decade in radiation treatment planning and delivery
with the advent of intensity modulated radiation therapy
共IMRT兲. With IMRT the prescribed dose can be delivered to
the target volume with a high degree of conformality, while
restricting dose to normal tissues. The full potential of these
technologies in radiation treatment can be realized only if the
patient can be positioned accurately and reproducibly during
0094-2405/2006/33„6…/1573/10/$23.00
© 2006 Am. Assoc. Phys. Med.
1573
1574
Islam et al.: Patient dose from kilovoltage cone beam CT
every session of the entire course of treatment delivery. Traditionally, the patient’s treatment position is verified by acquiring orthogonal portal images 共with films or electronic
portal imaging devices兲 using megavoltage 共MV兲 photon
beams at the beginning of the treatment course or weekly.
However, the precision of MV portal imaging is limited in
defining the patient’s position accurately due to the inherent
low contrast and two-dimensional nature of the projection
images.1 To overcome these limitations the use of kilovoltage 共kV兲 CT imaging systems in the treatment room are being investigated.2–9 The use of a conventional axial CT scanner in the treatment room has been described by a number of
investigators.2,7–9 The availability of large area flat panel detectors have facilitated the development of integrated cone
beam CT 共CBCT兲 systems on linear accelerators.3–6 A
kV-CBCT system has the potential of data acquisition and
reconstruction of a large volumetric image in 1 to 2 min.10 A
kV-CBCT system can be used before every treatment session
to localize a patient with high precision and, if necessary,
set-up parameters such as patient position and beam geometry can be adapted for the intended treatment delivery.
There is a growing interest in using kV-CBCT for high precision adaptive radiotherapy and several vendors have started
development and commercialization of this technology.
In addition to the CBCT mode, an integrated kV system
on a linear accelerator also provides radiographic and fluoroscopic modes of patient imaging. A kV imaging system,
therefore, would provide the opportunity of imaging the patient position in two-dimensional radiographic mode or
three-dimensional volume imaging mode before, during, and
after the treatment session with a high degree of accuracy.3 It
is expected that these imaging modalities will facilitate advancement of high precision radiation treatments by providing accurate spatial information of patient position and,
therefore, will be potentially used daily and repeatedly as an
online image guidance tool. Although, the dose to the patient
due to a kV image-guided RT session is small compared to
that of the megavoltage treatment, repeated use of this modality for image guidance may contribute significant dose to
normal tissue. Therefore, it is important to quantify the dose
due to kV imaging of patients undergoing radiation therapy.
In this report, the quantification of kV imaging dose for various imaging techniques is described. A method for estimating the dose delivered during imaging is presented to permit
individual centers to document the doses used in their imaging procedures. The quantification of dose is an important
first step toward developing guidelines for appropriate
image-guidance use with respect to the potentially deleterious effects of additional dose.
II. MATERIALS AND METHODS
Two identical kV-CBCT systems were utilized for extensive measurements for this study. A limited set of measurements has also been made with a newly installed commercial
system. The X-Ray Volume Imaging 共XVI®兲 system, a clinical prototype, integrated with a medical linear accelerator,
Medical Physics, Vol. 33, No. 6, June 2006
1574
FIG. 1. The XVI kV-CBCT system; 共a兲 centered geometry 共b兲 off-set geometry. With a setup of X1 = 19.5 cm and X2 = 6.5 cm the effective lateral field
of view 共FOVx兲 is 39.0 cm.
Synergy RP 共Research Platform兲, manufactured by ELEKTA
Oncology Systems, Norcross, GA, has been used to characterize the beams and also to make in vivo patient dose measurements. The XVI system consists of a conventional kV
x-ray tube 共Comet DX-9, Comet AG, Bern, Switzerland兲
mounted on a retractable arm onto the accelerator gantry’s
drum structure, in such a way that the central axes of the kV
and MV beams intersects at the isocenter and are orthogonal
to each other as shown in Fig. 1. The x-ray tube is powered
by a high-frequency generator 共Medstone XHF-340, Fife,
Scotland, UK兲 and operates in the range of 60– 150 kVp.
The image detection unit for the kV imaging is a flat panel
amorphous silicon 共Perkin Elmer Optoelectronics, Wiesbaden, Germany兲 detector with a dimension of 41 cm
⫻ 41 cm mounted on a retractable arm at a distance of
153.6 cm from the source focal spot. In CBCT mode, the
system can acquire projection data for complete rotation and
reconstruct high-resolution volumetric images in about 1 to
2 min.6
To facilitate the detailed measurements in water phantom
a bench top imaging system was utilized. The bench top
system 共as shown in Fig. 2兲 consists of an x-ray tube, collimation system, and flat panel imager, which are identical to
those of the XVI system. With the source at a fixed position,
a turntable with its center coinciding with the central axis of
the phantom can be set at 100 cm from the source and rotated around the vertical axis to simulate imaging geometry
on the linear accelerator. For optimum image quality an
added filtration of 2 mm Al and 0.1 cm Cu is used to harden
the beam for all imaging protocols on the Synergy RP®. The
dosimetric measurements were therefore, performed with the
added filtration both on the bench top and Synergy RP systems.
1575
Islam et al.: Patient dose from kilovoltage cone beam CT
FIG. 2. The bench top kV-CBCT system and dosimetric setup.
The dose measurements were performed in readily available cylindrical shaped water phantoms made of 6-mm-thick
Plexiglass®. A 30-cm-diam phantom was used to simulate an
average body while a 16 cm diameter was used for an average head. As shown in Fig. 3, a specially designed jig allows
accurate and arbitrary positioning of radiation detectors inside the water phantom in all three dimensions: radial distance r 共equivalent depth d兲, angular position ␪, and axis
position z. However, all the data presented in this report have
been measured for z = 0, along a plane coinciding with the
central axis of the beam. A 0.6 cc Farmer type ionization
chamber 共NE-2571A兲 and Keithley 31614 electrometer, with
air kerma calibration factor, Nk, traceable to a standard dosimetry laboratory 共NRCC, Ottawa兲 was used for measurements within the phantom. Dose was determined by methods
recommended by AAPM report TG61.11 A set of high sensitivity micro MOSFETs 共Model 1002 RD, Thompson and
Nielsen Inc., Ottawa, Canada兲, along with AutosenseTM
共Model number TN-RD-60兲 reader was used in high bias
1575
mode to measure dose on the surface of the phantom. The
MOSFET calibration factors, in terms of cGy/ mV, were determined by measuring the response at a depth of 2.0 cm in
water and comparing the corresponding dose measured by
the ion chamber. An average calibration factor of
0.034 cGy/ mV
共with
a
standard
deviation
of
0.0017 cGy/ mV兲 was determined for 100, 120, and 140 kVp
beams.
To check the consistency of the doses measured on the
bench top system, a limited set of measurements have been
performed with a specially designed phantom on the Synergy
RP as well as on a newly installed linear accelerator, Elekta
Synergy®. Synergy is equipped with a commercial XVI system and offers a set of similar, but slightly different scanning
geometry to that used on the bench top and Synergy RP
systems.
A. Characterization of the x-ray beams
The beam qualities were characterized by measuring the
first half-value layers 共HVL兲 of aluminum 共99.99% pure兲 in
narrow beam geometry. Attenuation measurements were performed in a 4 cm⫻ 4 cm field with attenuator placed at a
distance of 50 cm from the source and ion chamber 共without
build up cap兲 at 100 cm, while ensuring that no additional
scattering medium was within 50 cm from the chamber.
The machine output was quantified, in terms of reference
exposure rate 共mR/ mAs兲 in air at the isocenter for 10 cm
⫻ 10 cm field and relative output factors 共ROF兲 for various
field of views 共FOV兲 and kVp settings, where ROF is defined
as the ratio of the outputs for specified FOV and 10 cm
⫻ 10 cm field. The linearity of output was also assessed as a
function of total mAs settings, with various combinations of
current 共mA兲 and exposure time 共ms兲 settings.
B. Single beam dosimetry in fluoroscopic mode
The dose values for a series of square static fields, as a
function of depth along the central axis, were measured for
both the cylindrical phantoms 共i.e., 16 and 30 cm diameter兲
with a source to phantom axis distance of 100 cm. Although,
in typical fluoro mode the beam exposure technique requires
a few mAs, the dose measurements have been performed for
a technique setting of 200 mAs for higher accuracy.
C. Dosimetry in CBCT mode
FIG. 3. Schematic diagram of the water phantom. The jig allows positioning
of the detector at any arbitrary location in the cylindrical coordinate system:
radial distance r 共equivalent depth d兲, angular position ␪, and axial
position z.
Medical Physics, Vol. 33, No. 6, June 2006
In CBCT mode, imaging data are typically acquired in our
department with 330 projections 共the maximum number of
projections allowed on the Electa Synergy RP system兲 in a
360° 共full兲 rotation around the isocenter. As the size of the
flat panel detector is 41 cm⫻ 41 cm and is located at a fixed
distance of 153.6 cm from the x-ray source, the maximum
field of view in both lateral and longitudinal directions
共FOVx, FOVz兲 for data acquisition is restricted to approximately 26 cm, defined at the isocenter. To overcome this
limitation of the lateral FOVx, an offset scanning geometry is
used,12 in which the imager is shifted laterally and a corresponding asymmetric beam, defined by an asymmetric colli-
1576
Islam et al.: Patient dose from kilovoltage cone beam CT
mator, is utilized to scan a larger FOV. As illustrated in Fig.
1共b兲, the effective scan diameter is twice the size of collimator opening, X1, and therefore the largest reconstructed volume may have a diameter of 52 cm 共reconstructed FOVx兲.
The offset geometry that is used most commonly, involves
shifting the flat panel by approximately 10 cm, and utilizes
an asymmetric collimator, defined by X1 = 19.5 cm, X2
= 6.5 cm, and Z = 26 cm at isocenter. In our clinical experience scanning various sites, including thoracic and abdominal regions, this offset geometry provided the optimum image sets, in terms of quality and reconstructed field of view.
Unless otherwise stated, the dose values presented here have
been measured with this asymmetric collimator geometry,
i.e., with a reconstructed FOVx of 39 cm. Although, for the
head phantom, a centered scanning geometry with smaller
FOVx could have been used, for the convenience of keeping
the same collimator settings and reconstruction algorithm the
offset scanning geometry is used for all imaging procedures.
The dose measurements have been made at various depths
共d兲, with 330 projections and 2 mAs/ projection 共as per our
standard imaging protocol兲 for a number of FOV and kVp
settings. To examine the linearity of dose as a function of
total number of projections, while maintaining a constant
mAs/ projections, measurements have been made using 100–
600 projections. The dose values were measured using
120 kV beam and 2 mAs/ projection, at a depth of 2.0 cm
and at the central axis of the 30-cm-diam cylindrical phantom for a FOV of 10 cm⫻ 26 cm.
D. Formalism for dose estimation in CBCT mode
For a fixed scanning geometry and beam quality the dose
at any point in phantom would be proportional to the total
technique settings 共mAs兲 and also to the in-air exposure rate
at the reference point 共e.g., isocenter兲. Since different x-ray
tubes may have different exposure rates 共mR/ mAs兲 and furthermore users may choose to use different numbers of projections and mAs/ projections, it would be more meaningful
to present the measured dose data in terms of a dose conversion factor, e.g., dose per reference exposure rate at the isocenter. The relationship between dose to the phantom in
CBCT mode and in-air reference exposure rate depends upon
several factors such as; beam energy, phantom size, field of
view, and location within the phantom. We introduce a term
f CBCT to relate the dose and reference exposure rate using
DCBCT共E,R,FOV,d兲 = ẊRef共E兲 • ROF共E,FOV兲 • f CBCT
⫻共E,R,FOV,d兲 • T,
where DCBCT is the dose 共cGy兲 in phantom; E the beam
energy 共quality兲, expressed in HVL; R the radius of the phantom; d the depth in phantom; ẊRef the exposure rate 共R / mAs兲
for a reference field size 共e.g., 10 cm⫻ 10 cm兲; ROF the
relative output factor for the FOV; f CBCT the dose conversion
factor 共cGy/ R兲 per complete rotation of CBCT acquisition as
a function of phantom size 共radius R兲, energy 共E兲, field of
view 共FOV兲, and depth in phantom 共d兲; T the total technique
setting 共mAs兲; product of tube current and time settings.
Medical Physics, Vol. 33, No. 6, June 2006
1576
The results of the dose measurements are presented in
terms of f CBCT as a function of depths for different beam
energy, phantom size, and field of view. Given the reference
exposure rate, technique setting, and imaging geometry one
can estimate the dose to phantom using the above-noted
equation and appropriate f CBCT values.
E. Validation of the dose estimation formalism
1. Doses in phantom
To test the performance of the dose estimation formalism,
using the dose conversion factor, identical sets of measurements 共using the same beam quality, technique settings兲 have
been performed with 120 kV beam at a depth of 2.0 and
10 cm, as well as on the surface of a specially designed
20-cm-diam and 30-cm-long cylindrical solid water phantom, on the bench top system as well as on clinical systems
共Synergy RP and Synergy兲. The FOV utlized for the bench
top and Synergy RP was exactly identical, while for the Synergy system it was slightly different. The collimator cassette
labeled M10 of Synergy system was utilized, which has a
FOVx = 27.7 cm and FOVz = 13.5 cm. This collimator
matches with the default setting of the imager arm, for which
the center of the panel is offset by 11.5 cm from the central
axis of the kV x-ray beams. Thus, the lateral field of view 共at
the isocenter兲 is defined by X1 = 20.2 cm and X2 = 7.5 cm and
provides a reconstructed field of view, FOVx = 40.4 cm. The
Synergy unit 共commercial unit兲 was utilized in standard
clinical mode, for which no added filtration was used. For
proper dose consistency evaluation, the beam quality 共HVL兲
as well as the appropriate in-air exposure rate was measured.
The use of the solid water phantom facilitated the identical
setup on the bench top as well as on the linear accelerators,
with the exception that the phantom was supported by a 7.0cm-block of Styrofoam® for the measurements on linear accelerators, while on the bench top system there was no attenuating medium interfering the dose to the phantom. Based
on the measurements from the bench top system, the values
of f CBCT were derived and subsequently, using the appropriate in-air exposure rate, corresponding dose values were estimated for the clinical systems. The measured and estimated
dose values for the clinical systems were then compared.
2. Patient skin dose
As a first-order approximation, a patient may be assumed
to be water equivalent and cylindrical in geometry. To test
the dose estimation formalism with this assumption, measurements have been performed on five patients undergoing
CBCT in the abdominal and pelvic region using microMOSFETS. Four dosimeters were placed on the anterior skin
surface 共patient supine兲 along the iso-centric plane, approximately 40° apart. Three sets of measurements were performed for each patient. The skin dose was compared with
•
共R / mAs兲, total mAs
the estimated dose based upon the XRef
setting for the scan, and appropriate f CBCT for the surface.
1577
Islam et al.: Patient dose from kilovoltage cone beam CT
1577
FIG. 4. Relative output factor 共measured at the isocenter兲 vs longitudinal
collimator setting 共lateral collimator fixed at X1 = 19.5 cm, X2 = 6.5 cm兲 for
different kVp settings. The reference output 共10 cm⫻ 10 cm field at the
isocenter兲 values for the bench top system are 3.9, 6.5, and 9.5 mR/ mAs
with 100, 120, and 140 kVp, respectively.
FIG. 5. Dose/ mAs in fluoroscopic mode as a function of depth in the 16.0cm-diam phantom for various FOV 共a兲 120 kVp and 共b兲 100 kVp beams.
The source to cylinder axis distance is set at 100 cm. The uncertainties of
measurements on the surface and at depths are within ±5% and ±0.5%,
respectively.
III. RESULTS
response of an ion chamber in a phantom with varying tube
currents 共mA兲, as well as exposure time 共ms兲 independently.
The relative output as a function of the product of current
and exposure time settings 共mA s兲 were found to be highly
linear 共R2 = 0.9999兲.
As mentioned earlier, all the measurements performed in
air as well as at depths in phantoms utilized a 0.6 cc ionization chamber and Kiethley electrometer. The surface dose
measurements for phantoms and patients were performed
with MOSFET dosimeters. The uncertainties of the ion
chamber and MOSFET measurements were determined
共standard deviation/ mean value兲 to be within ±0.5% and
±5%, respectively.
A. Beam characteristics
The qualities of 100, 120, and 140 kVp beams with added
filtration of 2 mm Al and 0.1 mm of Cu were determined to
be 6.9, 7.9, and 8.7 mm of HVL in Al, respectively. The
in-air reference exposure rate 共for 10 cm⫻ 10 cm field兲 at
the isocenter was measured to be 3.9, 6.5, and 9.5 mR/ mAs
for 100, 120, and 140 kVp, respectively. The exposure rate
was found to vary rapidly with increasing beam energy. The
exposure rate relates to kVp settings with a second-order
polynomial. The ROFs, measured at the isocenter for various
field of views, with respect to the reference field, is shown in
Fig. 4. The ROF values increase rapidly initially with the
increase of FOV, and however become less sensitive with
field size beyond the FOVx of 10 cm.
The linearity of dose output has been determined, on both
the XVI and bench top systems, by measuring the relative
Medical Physics, Vol. 33, No. 6, June 2006
B. Single beam dosimetry in fluoroscopic mode
The dose per mAs setting is presented as a function of
depth for various square field sizes in Figs. 5 and 6. The
depth dose varies significantly with field size due to increased scatter contribution at larger field sizes. On the other
hand, the change in depth dose is small as a function of kVp
setting. For a 10 cm⫻ 10 cm field size, the dose at the center
of 16 cm phantom is 38% and 42% of the surface doses for
100 and 120 kVp beams, respectively. While for the 30-cmdiam phantom the dose at the center of the phantom for the
same field size is 14%, 16%, and 16% of the surface dose for
100, 120, and 140 kVp, respectively. For typical imaging
protocols, a technique setting of 1 to 2 mA s is used and
therefore the skin dose to phantom ranges from 0.1 to
0.4 mGy, depending upon the beam energy and phantom
size.
C. Dose in CBCT mode
Dose measurements were made with complete rotation for
a range of FOVs; from 5 cm⫻ 26 cm up to 26 cm⫻ 26 cm,
1578
Islam et al.: Patient dose from kilovoltage cone beam CT
1578
FIG. 7. Linearity of dose in CBCT mode, as a function of total number of
projections. The dose values presented here have been measured with
120 kV beam, 2 mAs/ projection and with a FOV of 10 cm⫻ 26 cm, in 30
-cm-diam cylindrical phantom.
FIG. 6. Dose/ mAs in fluoroscopic mode as a function of depth in the 30.0cm-diam phantom for various FOV 共a兲 140 kVp, 共b兲 120 kVp, and 共c兲
100 kVp beams. The source to cylinder axis distance is set at 100 cm.
using 330 projections at 2 mAs/ projection. Depending on
the FOV, the maximum dose for the body phantom was
found to vary from 1.8 to 2.3 cGy for 120 kVp and from 2.8
to 3.5 cGy for 140 kVp beams. For the head phantom, the
maximum dose values were found to vary from 1.5 to
2.0 cGy for 100 kVp and from 2.6 to 3.4 cGy for 120 kVp
beams. The dose data, at various depths and as a function of
FOV, are shown in Tables I and II. It may be noted that the
dose values at the surface are either very similar to or lower
than those at 2.0 cm depth. This is mainly due to the nonoverlapping radiation exposure, resulting from scanning with
asymmetric collimation of the FOV, as illustrated in the
“white” region of Fig. 1共b兲. In this region, the dose is expected to be lower, as compared to that of scanning with
centered geometry.
The relative dose values as a function of total number of
projections is shown in Fig. 7. As shown in the plot the dose
in 30-cm-diam phantom is highly linear 共R2 = 0.995兲 with
respect to the number of projections, in the range of
100– 600 projections. This demonstrates that the dose values
reported here can be scaled with the number of projections
共e.g., for the Synergy system, in which 622 projections are
acquired for full rotation scan兲, provided that the scanning
geometry, phantom size, beam quality, and mAs/ projections
are the same.
D. Dose conversion factor in CBCT mode „fCBCT…
The dose conversion factors, f CBCT expressed in 共cGy/R兲,
are presented for asymmetric collimators in Figs. 8 and 9. As
shown, the values of f CBCT have higher dependence on the
phantom size and FOV, compared to a weak dependence on
the beam energy. For the head phantom, because of its
shorter penetration depths and, consequently, lower attenuation, the values of f CBCT are relatively uniform 共in the range
of 0.5–0.7兲 across the diameter of the phantom. On the other
hand, as expected, the dose conversion factor varies rapidly
共in the range of 0.2–0.55兲 as a function of depth for the larger
TABLE I. Typical dose in cGy for 30-cm-diam phantom 共2 mAs/ projection; 330 projections兲
Field of view 共cm⫻ cm兲
10⫻ 26
15⫻ 26
5 ⫻ 26
26⫻ 26
Depth
共cm兲
120 kVp
140 kVp
120 kVp
140 kVp
120 kVp
140 kVp
120 kVp
140 kVp
0
2
6
9
15
1.8
1.7
1.1
0.9
0.7
2.8
2.4
1.7
1.4
1.1
2.1
2.0
1.5
1.4
1.1
3.1
3.0
2.3
2.1
1.7
2.2
2.2
1.8
1.5
1.3
3.2
3.3
2.7
2.4
2.0
2.3
2.3
1.9
1.8
1.6
3.5
3.5
3.1
2.8
2.4
Medical Physics, Vol. 33, No. 6, June 2006
1579
Islam et al.: Patient dose from kilovoltage cone beam CT
1579
TABLE II. Typical dose in cGy for 16-cm-diam phantom 共2 mAs/ projection; 330 projections兲
Depth
共cm兲
0
2
8
Field of view 共cm⫻ cm兲
10⫻ 26
5 ⫻ 26
15⫻ 26
100 kVp
120 kVp
100 kVp
120 kVp
100 kVp
120 kVp
1.4
1.5
1.2
2.3
2.6
2.0
1.6
1.9
1.6
2.7
3.2
2.8
1.7
2.0
1.8
2.9
3.4
3.0
body phantom. As described earlier, the appropriate values of
•
f CBCT can be multiplied by the reference exposure rate XRef
共R / mAs兲, ROF, and total mAs setting to determine the dose
to phantom.
E. Validation of the dose estimation formalism
1. Doses in phantom
The ratios of the doses measured at various depths of the
20-cm-diam solid water phantom, and to those estimated for
the Synergy RP and Synergy systems are shown in Table III.
The dose was estimated based on the f CBCT values derived
from the corresponding measurements on the bench top system. In the case of Synergy, the HVL was measured to be
7.3 mm in Al and, therefore, the corresponding f CBCT value
was determined by interpolation of the values measured on
the bench top system. As shown, the dose values at depths
agree within ±5%, but on the surface the dose differs by as
much as 10%. The higher disagreement at surface may be
partially due to larger uncertainties in MOSFET response as
well as the dose perturbation 共attenuation and scatter兲 from
the Styrofoam support used during the measurements on linear accelerators.
2. Patient skin dose
The results of in vivo skin dose measurements 共average
and standard deviation兲 and corresponding dose estimation
are shown in Table IV. Also, shown in Table IV are the
patient anterior-posterior and lateral separations. However,
the dose values were estimated based on the f CBCT values of
30-cm-diam water phantom and total mAs used for the imaging scan. The average discrepancies between the measured
and estimated dose was found to be 0.08 cGy, with a maximum difference of 0.23 cGy.
IV. DISCUSSIONS
A. Dose conversion factor and in vivo dose
measurements
The dose conversion factor in CBCT mode, f CBCT, introduced in this report is a useful parameter which can be utilized readily to estimate patient dose, provided that the beam
geometry, phantom size, and beam qualities are similar to
those used here. The values of the factor depend significantly
upon the phantom size due to the source to phantom distance
as well as the attenuation characteristics of low energy photon beams. For a particular phantom size, however, the value
of f CBCT varies slowly with the change in beam quality
共HVL兲, as shown in Fig. 10. For the 30-cm-diam phantom, at
15.0 cm depth, depending upon the FOV the value of f CBCT
changes from 3 to 4% per 0.1 mm change in HVL of Al.
While at a depth of 2.0 cm, the change ranges only from
0.4% to 0.9% for the same variation of HVL.
The average agreement between measured and estimated
skin doses of five patients were found to be within ±5%,
with a maximum deviation of 14%. The high standard deviation of measurements and magnitude of the maximum disagreement can be attributed to several factors, such as approximation of the patient as a homogeneous body phantom
of 30 cm diameter, presence of couch in the beam path, and
also uncertainty in MOSFET measurements. Further work is
necessary, to address these issues for more accurate patient
dose determination. Nevertheless, the present method of dose
estimation may be considered acceptable, as a first-order approximation, in such a low dose environment.
TABLE III. The ratio of dose measured at various depths on clinical systems and dose estimated for 20-cm-diam
cylindrical phantom based on the dose conversion factor, f CBCT, derived from the measurements on the bench
top system.
Dose ratio 共measured/expected兲
Clinical
system
Synergy RP
Synergy RP
Synergy
Nominal
energy
共kVp兲
100
120
120
Medical Physics, Vol. 33, No. 6, June 2006
Added
filtration
Beam
quality
共HVL in
mm of Al兲
d = 10.0 cm
d = 2.0 cm
d = 0 cm
共surface兲
2 mm Al +0.1 mm Cu
2 mm Al +0.1 mm Cu
None
6.9
7.9
7.3
0.97
0.99
0.97
1.04
1.05
0.97
1.01
1.10
0.95
1580
Islam et al.: Patient dose from kilovoltage cone beam CT
FIG. 8. Dose conversion factors in CBCT as a function of depth in 16-cmdiam cylindrical water phantom for various field of views 共a兲 100 kVp
共HVL 6.9 mm Al兲 and 共b兲 120 kVp 共HVL 7.9 mm Al兲. The lateral field of
view 共FOVx兲 is defined with collimator setting of 26.0 cm 共X1 = 19.5 cm and
X2 = 6.5 cm兲.
As pointed out earlier, the values of f CBCT presented here
have been derived based on measurements on a bench top
system. These values may not exactly reflect the dose values
of a recently released commercial Elekta Synergy linear accelerator due to differences in the energy specifications and
collimator geometry.
B. Total dose due to kV image guidance
and comparison with MV imaging
The total dose to the patient due to kV image guidance
can be put into perspective by comparing the dose due to
conventional MV portal imaging. As an example, we consider that a patient receiving radiation therapy in the pelvic
region will have 30 fractions of treatment and daily imaging
will be performed for localization. First, let us consider the
maximum dose per session due to 共a兲 MV portal imaging
with orthogonal pairs using 4 MU for each portal, 共b兲 kV
fluoroscopic imaging with orthogonal pairs using 120 kVp
and 2 mAs/ field, and 共c兲 kV CBCT with full rotation scan
using 120 kVp and total of 660 mAs. The dose due to orthogonal MV portal imaging with a field size of 10 cm
⫻ 10 cm was calculated in a 30-cm-diam phantom using PINNACLE Version 6.2b 共Philips Medical System兲 and the maximum dose was determined to be 7 cGy. The maximum dose
due to orthogonal pair of kV fluoroscopic fields on the same
phantom is determined to be 0.25 mGy. Finally, the maximum dose per session of kV CBCT imaging is determined to
be 2.3 cGy. It is obvious that the maximum doses due to kV
fluoroscopic and CBCT imaging is lower by approximately a
Medical Physics, Vol. 33, No. 6, June 2006
1580
FIG. 9. Dose conversion factors in CBCT as a function of depth in 30-cmdiam cylindrical water phantom for various field of views 共a兲 120 kVp
共HVL 7.9 mm Al兲 and 共b兲 140 kVp 共HVL 8.7 mm Al兲. The lateral field of
view 共FOVx兲 is defined with collimator setting of 26.0 cm 共X1 = 19.5 cm and
X2 = 6.5 cm兲.
factor of 30 and 3, respectively, compared to that of MV
portal imaging. However, for high precision adaptive radiation therapy it is expected that the kV imaging will be used at
least once daily 共may be even twice daily; before and after
patient position adjustment兲 and, therefore, the total accumulated maximum dose may be in excess of 1 Gy. Moreover,
the dose to patient due to CBCT will be distributed throughout a larger volume, as compared to conventional portal imaging. In the case of MV portal imaging, the additional dose
can be easily added tothe overall dose distribution. However,
because of its low energy and very low dose/fraction the
imaging dose due to kV-CBCT may not be radiobiologically
equivalent to MV dose13–15 and therefore cannot be simply
TABLE IV. Comparison of patient skin dose; measured and estimated. The
measured dose for each patient is shown as average and standard deviation
共␴兲 of multiple measurements. The estimation was based on the dose conversion factor derived from 30-cm-diam water phantom measurement on the
bench top system. The patients were scanned in the pelvic or abdominal
region with 120 kVp beam.
Patient skin dose 共cGy兲
Patient
No.
Patient separation
AP 共cm兲
⫻ lateral 共cm兲
1
2
3
4
5
20.0⫻ 29.6
23.3⫻ 38.7
23.0⫻ 39.0
22.0⫻ 35.0
23.2⫻ 36.0
Measured;
average
共␴兲
1.67
1.12
1.46
1.84
1.55
共0.14兲
共0.10兲
共0.15兲
共0.17兲
共0.12兲
Dose difference
Estimated 共measured-estimated兲
1.71
1.15
1.69
1.78
1.71
−0.04
−0.03
−0.23
0.06
−0.16
1581
Islam et al.: Patient dose from kilovoltage cone beam CT
FIG. 10. Variation of dose conversion factor in 30-cm-diam phantom as a
function of beam quality 共HVL in Al兲, with different FOVs: 共a兲 15 cm depth
and 共b兲 2 cm depth.
accounted for in the patient dose distribution. While the issue
of dose accounting due to kV imaging needs further consideration, it is prudent, at a minimum, to accurately document
the applied dose.
C. Dose reduction strategies
Although the dose due to image guidance with CBCT is
small in the context of the uncertainty in therapy dose delivery, based upon ALARA principle, efforts should be made to
minimize any unwanted dose to patients. The dose can be
easily minimized by the following simple methods. These
include 共a兲 minimizing the overall technique settings 共mAs兲
by reducing the number of projections and/ or reducing the
mAs settings per projection, 共b兲 by using lower kVp 共beam
energy兲, 共c兲 minimizing the longitudinal field of view; FOVz
to that necessary for image guidance. In addition to reducing
patient dose, superior image contrast can be obtained due to
the lower scatter component when using smaller FOV,16 共d兲
by scanning the patient with partial rotation. The dose to
patient on the opposite side of the source rotation can be
reduced by a large factor. To examine the magnitude of the
dose reduction due to half rotation scan, measurements have
been made on the surface and at 2.0 cm depth around the
circumference for the 30-cm-diam-phantom with half rotation plus the fan angle 共194° 兲 using 120 kVp beams. These
measurements have been made with a symmetrical collimator setting, i.e., X1 = 13 cm and X2 = 13 cm. As shown in Fig.
11 the dose on the opposite side of the scan is approximately
15% of that on the source side. Further investigation of these
“low dose” scanning methods is under way.
Depending upon the specific mode of image guidance
with volumetric imaging, it may be possible to utilize several
Medical Physics, Vol. 33, No. 6, June 2006
1581
FIG. 11. Polar diagram of the angular variation of doses due to half-rotation
scan at 2.0 cm depth and on the surface of the body phantom. This dose plot
represents a situation where the phantom is fixed and the source rotates over
194° 共180° + fan angle兲, starting at 7°, scanning through 270° and ending at
187°. The technique used are 2 mAs/ projections, 330 projections/ full rotation, and 178 projections/194° rotation.
of the above-noted strategies to minimize patient dose. For
example, if the bony landmark is to be used for patient localization, then a relatively low soft tissue discrimination
image with lower beam energy and mAs settings 共with reduced number of projections兲 can be utilized. On the other
hand, if a high quality image 共high tissue visibility兲 is desirable then a smaller cone angle can be used to minimize the
dose and maximize image quality. The selection of these
techniques will clearly depend on the objective and benefit of
image guidance.
V. CONCLUSIONS
A comprehensive set of dose measurements has been performed for kV beams both in fluoroscopic and CBCT modes,
using homogeneous cylindrical shaped body and head phantoms. The maximum dose to the phantom due to orthogonal
fluoroscopic imaging, using 2 mAs exposure ranges from 0.1
to 0.4 mGy depending upon the beam energy and phantom
size. The dose due to CBCT in a body phantom for a typical
full rotation imaging protocol using 120 kVp beams, with a
field of view of 26 cm⫻ 26 cm and total technique settings
of 660 mAs, ranges from 1.6 cGy 共at the center兲 to 2.3 cGy
共at surface兲. For similar geometry and mAs settings, the
maximum dose to the phantom can be as high as 3.5 cGy
when 140 kVp beams is used in the body phantom or
120 kVp is used for the head phantom. The results of our
CBCT dose measurements using asymmetric collimator, as
required for the offset imaging geometry, have been presented in terms of dose conversion factor, f CBCT 共cGy/R兲.
This factor depends upon beam energy, phantom size, and
scanning geometry and can be utilized to estimate dose to
1582
Islam et al.: Patient dose from kilovoltage cone beam CT
homogeneous cylindrical phantoms for any arbitrary total
mAs setting and reference exposure rate of the kVsource.
The measurements in homogeneous cylindrical water
phantom will not simulate exactly the patient geometry and
effects of anatomy 共heterogeneity兲. However, the method
presented here can be adequate for estimation and recording
of patient dose due to kV fluoroscopic and CBCT imaging. It
is clear from the results presented that there is a genuine
opportunity to minimize the radiation dose penalty associated with image guidance by selecting geometry and exposure parameters that achieve the necessary image quality.
This is an important first step to bringing proper assessment
of the “dose penalties” associated with the benefits of imageguided radiotherapy.
ACKNOWLEDGMENTS
This work is supported in part by NIH/NIA R21/R33
AG198381, NIH/NIBIB R01 EB002470, and Elekta Oncology Systems.
a兲
Electronic mail: [email protected]
L. J. Pisani, D. Lockman, D. A. Jaffray, D. Yan, A. A. Martinez, and J. W.
Wong, “Set-up error in radiotherapy: On-line correction using electronic
kilovoltage and megavoltage radiographs,” Int. J. Radiat. Oncol., Biol.,
Phys. 47共3兲, 825–839 共2000兲.
2
M. Uematsu, T. Fukui, A. Shioda, H. Tokumitsu, K. Takai, T. Kojima, Y.
Asai, and S. Kusano, “A dual computed tomography linear accelerator
unit for stereotactic radiation therapy: A new approach without cranially
fixated stereotactic frames,” Int. J. Radiat. Oncol., Biol., Phys. 35共3兲,
587–592 共1996兲.
3
D. A. Jaffray, D. G. Drake, M. Moreau, A. A. Martinez, and J. W. Wong,
“A radiographic and tomographic imaging system integrated into a medical linear accelerator for localization of bone and soft-tissue targets,” Int.
J. Radiat. Oncol., Biol., Phys. 45共3兲, 773–789 共1999兲.
4
D. A. Jaffray and J. H. Siewerdsen, “Cone-beam computed tomography
with a flat-panel imager: Initial performance characterization,” Med.
Phys. 27共6兲, 1311–1323 共2000兲.
5
J. H. Siewerdsen and D. A. Jaffray, “Optimization of x-ray imaging ge1
Medical Physics, Vol. 33, No. 6, June 2006
1582
ometry 共with specific application to flat-panel cone-beam computed tomography兲,” Med. Phys. 27共8兲, 1903–1914 共2000兲.
6
D. A. Jaffray, J. H. Siewerdsen, J. W. Wong, and A. A. Martinez, “Flatpanel cone-beam computed tomography for image-guided radiation
therapy,” Int. J. Radiat. Oncol., Biol., Phys. 53共5兲, 1337–1349 共2002兲.
7
W. Cheng, J. Wong, L. Grimm, M. Chow, M. Uematsu, and A. Fung,
“Commissioning and clinical implementation of a sliding gantry CT scanner installed in an existing treatment room and early clinical experience
for precise tumor localization,” Am. J. Clin. Oncol. 26共3兲, e28–e36
共2003兲.
8
S. Shiu, E. L. Chang, J. S. Ye, M. F. Lii, L. D. Rhines, E. Mendel, J.
Weinberg, S. Singh, M. H. Maor, R. Mohan, and J. D. Cox, “Near simultaneous computed tomography image-guided stereotactic spinal radiotherapy: An emerging paradigm for achieving true stereotaxy,” Int. J.
Radiat. Oncol., Biol., Phys. 57共3兲, 605–613 共2003兲.
9
K. M. Yenice, D. M. Lovelock, M. A. Hunt, W. R. Lutz, N. FournierBidoz, C-H. Hua, J. Yamada, M. Bilsky, H. Lee, K. Pfaff, S. V. Spirou,
and H. I. Amols, “CT image-guided intensity-modulated therapy for
paraspinal tumors using stereotactic immobilization,” Int. J. Radiat. Oncol., Biol., Phys. 55共3兲, 583–593 共2003兲.
10
M. B. Sharpe, D. J. Moseley, T. G. Purdie, M. K. Islam, J. H. Siewerdsen,
and D. A. Jaffray, “The stability of mechanical calibration for a kV cone
beam computed tomography sytem integrated with linear accelerator,”
Med. Phys. 33共1兲, 136–144 共2006兲.
11
C.-M. Ma, C. W. Coffey, L. A. DeWerd, C. Liu, R. Nath, S. M. Seltzer,
and J. P. Seuntjens, “AAPM protocol for 40-300 kV x-ray beam dosimetry in radiotherapy and radiobiology,” Med. Phys. 28共6兲, 868–893
共2001兲.
12
P. S. Cho, R. H. Johnson, and T. W. Griffin, “Cone-beam CT for radiotherapy applications,” Phys. Med. Biol. 40, 1863–1883 共1995兲.
13
M. C. Joiner, B. Marples, P. Lambin, S. C. Short, and I. Turesson, “Lowdose hypersensitivity: Current status and possible mechanisms,” Int. J.
Radiat. Oncol., Biol., Phys. 49共2兲, 379–389 共2001兲.
14
M. A. Hill, “The variation in biological effectiveness of x-rays and
gamma rays with energy,” Radiat. Prot. Dosim. 112共4兲, 471–481 共2004兲.
15
C. Boreck, E. J. Hall, and M. Zaider, “X rays may be twice as potent as
rays for malignant transformation at low doses,” Nature 共London兲 301,
156–158 共1983兲.
16
J. H. Siewerdsen, D. J. Moseley, B. Bakhtiar, and D. A. Jaffray, “The
influence of antiscatter grids on soft-tissue detectibility in cone beam
tomography with flat panel detectors,” Med. Phys. 31共12兲, 3506–3520
共2004兲.