ARTICLE IN PRESS

Applied Radiation and Isotopes 67 (2009) 1629–1637

Contents lists available at ScienceDirect

Applied Radiation and Isotopes

journal homepage: www.elsevier.com/locate/apradiso

Treatment planning and delivery of IMRT using 6 and 18 MV photon beams

without flattening filter
Sotirios Stathakis , Carlos Esquivel, Alonso Gutierrez, Courtney R Buckey, Niko Papanikolaou
Department of Radiation Oncology, University of Texas Health Science Center at San Antonio, 7979 Wurzbach Rd, San Antonio, TX 78229, USA

a r t i c l e in fo abstract

Article history: In light of the increasing use of intensity modulated radiation therapy (IMRT) in modern radiotherapy
Received 27 October 2008 practice, the use of a flattening filter may no longer be necessary. Commissioning data have been
Received in revised form measured for a Varian 23EX linear accelerator with 6 and 18 MV photon energies without a flattening
11 March 2009
filter. Measurements collected for the commissioning of the linac included percent depth dose curves
Accepted 11 March 2009
and profiles for field sizes ranging from 2  2 to 40  40 cm2 as defined by the jaws and multileaf
collimator. Machine total scatter factors were measured and calculated. Measurements were used to
Keywords: model the unflattened beams with the Pinnacle3 treatment planning system. IMRT plans for prostate,
Radiation therapy lung, brain and head and neck cancer cases were generated using the flattening filter and flattening
Photons
filter-free beams. From our results, no difference in the quality of the treatment plans between the flat
Intensity modulated radiation therapy
and unflattened photon beams was noted. There was however a significant decrease in the number of
Flattening filter free linac
monitor units required for unflattened beam treatment plans due to the increase in linac out-
put—approximately two times and four times higher for the 6 and 18 MV, respectively.
& 2009 Elsevier Ltd. All rights reserved.

1. Introduction MLC produces inhomogeneous fluence maps of photons beams,

Advances in the technology of radiation therapy planning and
delivery allow for the design and delivery of complex treatments.
With the introduction of intensity modulation radiation therapy
(IMRT), highly conformal treatments are realized in which high
doses are delivered to target volumes while still sparing adjacent
critical structures. The price for such conformal treatments are
increases in treatment time and monitor units (MU). Modern
radiation therapy linear accelerators utilize high energy electron
beams that interact with the linac target to produce high energy
bremsstrahlung X-ray photons. Since the photons created through
this process are forward peaked and the photon fluence is uniform
with angle, flattening filters are introduced in the path of the
beam to preferentially attenuate the photon beam such that the
dose becomes relatively flat and uniform at a given depth in
the patient.
The flattening filter provides a flat beam for a large field area
(40  40 cm2). The photon beam is further modified using
movable collimators (jaws) and multi-leaf collimators (MLC) to
create field projections which are irregular in shape. The
placement of MLC is machine design dependent (above or below
jaws, or replacing jaws), but the end result is similar. The MLC is a
key element in the implementation of an IMRT program. Using the
 Corresponding author. Tel.: +1 210 450 1010; fax: +1 210 616 5682.
E-mail address: stathakis@uthscsa.edu (S. Stathakis).
which allows delivery of very conformal distributions of uniform
dose to the target(s).
The current method of optimizing the photon beam fluence
using MLCs is realized by manipulating the uniform photon beam
fluence by creating areas of high and low intensity according to
the requirements of the plan. The photon beam fluence is divided
into elemental beam areas that are assigned a weight which is
directly proportional to the MU that will be delivered through this
area (segment). The weights of each elemental area are then
manipulated (increased or decreased) with respect to its con-
tribution to the planning dose criteria. The optimized beam
intensity is then delivered using the MLC. Several beam segments
of different MUs are used in each case to deliver the planned
photon fluence.
Since the photon beam fluence is based on the optimization of
elemental areas, it is not conceptually necessary to start with a
uniform beam. The benefits of flattening filter removal have been
reported by several groups (Cashmore, 2008; Fu et al., 2004;
Mesbahi, 2007; Mesbahi et al., 2007; Sharma et al., 2007; Titt
et al., 2006). Initially, unflattened beams have been investigated
for radiosurgical treatment plan deliveries. For such small fields,
the unflattened and flattened beam profiles were identical
however a higher dose rate could be achieved with unflattened
beams (Zefkili et al., 1994). Commercial treatment machines are
currently using flattening filter free linear accelerators. The helical
tomotherapy HiART system (TomoTherapy, Inc., Madison, WI) is
outfitted with a flattening filter free linear accelerator and has

0969-8043/$ - see front matter & 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.apradiso.2009.03.014
 ARTICLE IN PRESS
1630 S. Stathakis et al. / Applied Radiation and Isotopes 67 (2009) 1629–1637
been investigated extensively (Mackie et al., 1993; Jeraj et al.,
2004). Additionally, the CyberKnife system (Accuray Inc, Sunny-
vale CA) does not utilize a flattening filter and is primarily used for
small field stereotactic treatments (Sharma et al., 2007; Adler
et al., 1997).
One of the benefits of flattening filter free linacs is the increase
in output. It has been showed that the removal of the flattening
filter increases the output of a linear accelerator by a factor of 2 or
more (Fu et al., 2004; Mesbahi, 2007; Mesbahi et al., 2007;
Vassiliev et al., 2006a, b, 2007a, b). Although the benefit of the
increased output effects the every day fraction time, this increase
in output has been shown to be of greater advantage during
radiosurgical procedures due to the high doses delivered and
small output factors (Fu et al., 2004). Moreover, flattening filter
free high energy linear accelerators with nominal energies above
15 MeV have included neutron production and hence require
potentially less shielding (Vassiliev et al., 2007b; Kry et al.,
2007a).
In this study, we investigate the ability of commissioning a
flattening filter free photon beam in a commercial treatment
planning system (TPS) and seek to verify the model through
delivery of IMRT treatment plans using a 6 and 18 MV photon
beam. We also investigate the feasibility of using the unflattened
beams for various treatment sites and for radiosurgery.
2. Methods and materials
A Varian 23EX (Varian Medical Systems, Palo Alto, CA) linear
accelerator was used for the measurements undertaken in this
study. The Varian 23EX accelerator has two photon energies: 6
and 18MV. Depending on the energy chosen for operation, the
corresponding flattening filter is normally inserted in the photon
beam path. The flattening filter for each photon beam is housed in
a carousel that rotates. The carousel is located after the primary
collimator. Apart from the flattening filters on the carousel, an
opening exist in which an steel plate of 1 mm thickness sits. The
steel plate was used and was placed in the beam by overriding the
faults. Both photon beams were measured in this manner.
All the commissioning measurements were acquired in water
using a Scanditronix blue phantom (IBA Dosimetry GmbH,
Germany) with Exradin A1SL (Standard Imaging, Inc., WI, USA)
ionization chambers. The source to surface distance (SSD) was set
to 100 cm. Percent depth dose curves and profiles at various
depths were measured for both photon beam energies. Profiles
were measured for a range of field sizes defined by the jaws and
MLC. Output factors (Scp) were measured for various field sizes
since they are needed during the commissioning of the linac in the
treatment planning system. The absolute dose output (cGy/MU)
was measured according to the TG-51 protocol for both energies,
with and without a flattening filter to quantify changes due to the
absence of the flattening filter.
The measured data were used to commission two flattening
filter free machines (one for each photon beam energy) in the
Pinnacle3 (Philips Medical Milpitas, CA) treatment planning
system (TPS).
The commissioned machines were used to produce IMRT
treatment plans for prostate (using 6 and 18 MV photon beams),
head–neck, brain and lung cancer treatments using 6 MV photon
beam only. The plans were created used clinical dose constraints
and the collapsed cone convolutions superposition algorithm was
employed for the final dose calculation. The dose grid was set to
2 mm  2 mm  2 mm for all cases. Corresponding plans using the
exact same dose constraints were generated with a flattening
filter beam. Total number of monitor units, dose distributions,
dose volume histograms were used as comparison metrics
between plans generated with and without a flattening filter. A
prostate IMRT plan was validated for dose delivered in an acrylic
phantom. Point dose measurements were collected using a PTW
0.125 cc ionization chamber (PTW New York) and Kodak EDR2
film (Kodak Inc., Rochester, NY) for planar dose information.
Measurements were used to verify the accuracy of the collapsed
cone convolution superposition (CCCS) algorithm used in Pinna-
cle3 dose computation engine.
3. Results
3.1. Treatment planning commissioning
The physics tool of the Pinnacle3 system has an iterative
algorithm that uses the measured profile data to extract the
convolution model calculation parameters. The model parameters
control different parts of the dose calculation that is separated
into the following regions: in field, out of field, build-up (electron
contamination), and depth dose (photon energy spectrum).
The electron contamination is modeled in Pinnacle using an
approximately exponential dumping filter. Once the dose is
calculated based on the photon beam characteristics (using the
convolution-superposition model) the electron contamination
dose is added to the dose grid. The additional dose attributable
to the electron contamination depends on the energy of the beam,
field size, off axis distance and SSD.
The mean energy for both unflattened beams had to decrease
in order to model accurately the PDD curves measured in water. To
achieve that we had to increase the relative weight of the higher
energy bins in the original energy spectrum of the flattened
beams (Fig. 1).
The relative incidence fluence was specified using the arbitrary
profile editor to account for the conical shape of the unflattened
beam. The altered incidence fluence are shown in Fig. 2. The
profiles of the incidence fluence assume cylindrical symmetry and
they are expressed using a polynomial equation.
The combination of the energy spectrum and the incident
fluence parameters led to two linac models (one for each energy)
that were able to accurately predict the measuremed data. The
accuracy of the models is shown in Fig. 3.
0.35
0.30
relative number of photons
0.25
6MV
6MV no FF
0.20 18MV
18MV no FF
0.15
0.10
0.05
0.00
0 2 4 6 8 10 12 14 16 18 20
energy /MeV
Fig. 1. Relative energy fluence for the 6 and 18 MV photon beams. Comparison
between flatened and flattening filter-free beams.
 ARTICLE IN PRESS
S. Stathakis et al. / Applied Radiation and Isotopes 67 (2009) 1629–1637 1631

3.2. Total scatter factors
The output factors (Scp) of the unflattened beams were
measured and compared against the flattened beams (Fig. 4).
The measurements were all performed along the central axis
using the movable jaws to define the field. The measurements
were performed with SSD ¼ 100 cm at depth of dmax of the
10  10 cm2 field of each energy. All scatter factors were
normalized to the respective 10  10 cm2 field in each case. The
unflattened beam total scatter factors for the 6 MV beam were
lower relative to those of the flattened beam for field sizes larger
than 10  10 cm2 but higher for field sizes smaller than
10  10 cm2.
3.3. Percent depth dose measurements
Percentage depth dose curves were normalized to the depth of
maximum dose for each measured field along the central axis and
compared to its respective value for a flattened beam. The results
are shown in Figs. 5 and 6. It was noted that the 6 MV unflattened
beam PDDs were less penetrating than those of a flattened photon

1.5
1.4 1.2
1.3
1.2
1.1
1.1
1.0
incident fluence

1.0

output factors
0.9
0.8
0.9
0.7
0.6
0.8 6MV no FF
0.5
6MV 18MV no FF
0.4 6MV with FF
6MV FFF 0.7
0.3 18MV with FF
18MV
0.2
18MV FFF
0.1 0.6
0.0
0 2 4 6 8 10 12 14 16 18 20 22 24 0 10 20 30 40
radius /cm field side /cm

Fig. 2. Relative photon incident fluence. Comparison between flattened and Fig. 4. Total scatter factors (Scp) comparison between flattened and unflattened
unflattened beams. photon beams for two given energies.

100 100
90 6MV 10x102 90 18MV 4x42
80 80
70 70
dose /%

dose /%

60 60
50 50
40 40
measured
30 calculated 30 measured
20 20 calculated
10 10
0 0
0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20
depth /cm depth /cm

100 100
90 6MV 10x102 90 18MV 4x42
80 80
70 70
dose / %

60
dose / %

60 measured
50 50 calculated
40 measured 40
calculated
30 30
20 20
10 10
0 0
-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 -7 -6 -5 -4 -3 -2 -1 0
distance /cm distance /cm

Fig. 3. Comparison of measured and modeled by Pinnacle TPS of PDDs and off-axis profiles for a 6 MV 10  10 cm2 and an 18 MV 4  4 cm2 fields.
 ARTICLE IN PRESS
1632 S. Stathakis et al. / Applied Radiation and Isotopes 67 (2009) 1629–1637

100 80
no FF beam Flat beam
90
70
80
60
70

percent dose /%
percent dose /%

50
60

50 40
5x5 no FF
40 5x5 flat
30
10x10 no FF
30 10x10 flat
30x30 no FF 20
20
30x30 flat
10
10

0 0
0 2 4 6 8 10 12 14 16 18 20 22 24 -25 -20 -15 -10 -5 0 5 10 15 20 25
depth /cm distance /cm

Fig. 5. Percent depth dose curve comparison for the 6 MV photon beams with and Fig. 7. Half profiles for the 2  2, 5  5, 10  10, 20  20 and 30  30 cm2 fields of
without flattening filter for field sizes ranging from 5  5 to 30  30 cm2. the 6 MV photon beams. The unflattened profiles are shown with thin lines (left)
and the flat ones with thick lines (right).

100 90
no FF beam Flat beam
90 80
80
70
70
percent dose /%

60
percent depth /%

60
50
50 5x5 no FF
5x5 flat 40
40
10x10 no FF
10x10 flat 30
30
30x30 no FF
20 20
30x30 flat
10 10

0 0
0 2 4 6 8 10 12 14 16 18 20 22 24 -20 -15 -10 -5 0 5 10 15 20
depth /cm distance /cm
Fig. 6. Percent depth dose curve comparison for the 18 MV photon beams with and Fig. 8. Half profiles for the 2  2, 5  5, 10  10, 20  20 and 30  30 cm2 fields of
without flattening filter for field sizes from 5  5 to 30  30 cm2. the 18 MV photon beams. The unflattened profiles are shown with thin lines (left)
and the flat ones with thick lines (right).
beam but the surface dose was higher due to the softer spectrum
of the unflattened beam. unflattened beam profiles to the flattened is shown in Figs. 7
The PDDs for the 18 MV photon beam energy are shown in Fig. and 8.
6. It was observed that the larger field sizes of the unflattened
beam have a shallower dmax and higher surface doses than the
respective ones with the flattening filter. In general, the 3.5. Linac output measurement
differences in the PDD for depth greater than dmax are less
profound than those of 6 MV. The dmax moved towards the surface In order to measure the machine output in cGy/MU, the TG51
by 2–3 mm. protocol was used. The flattened photon beams were calibrated to
deliver 1 cGy/MU at their respective dmax. Apart from the removal
of the flattening filter, no other modification were made to the
3.4. Off-axis measurements linear accelerator. The %ddx were determined according to the
recommendations set forth by the TG51 protocol—i.e. a 1 mm lead
Profiles of various fields were measured using the same sheet was introduced in the 18 MV photon beam for the
setup implemented for the measurement of the PDDs. The measurement of %ddx. The measurements were performed using
absence of a flattening filter caused a cone shape in the profiles. a PTW 31003 0.125 cc ionization chamber. The measured output
The cone shape of the off axis profiles was more profound in the factor for 6 MV was 1.99 and 3.62 cGy/MU for the 18 MV photon
case of the 18 MV photon beam. The comparison of the beam.
 ARTICLE IN PRESS
S. Stathakis et al. / Applied Radiation and Isotopes 67 (2009) 1629–1637 1633

3.6. IMRT delivery
A prostate case was planned using both available energies and
both photon beam options (flat and flattening filter free, see
Fig. 9). The differences between the 6 and 18 MV unflattened
photon beam plans are as expected (higher superficial doses with
6 MV). When comparing the delivery methods with regard to PTV
coverage and critical structure sparing, negligible differences are
noted (see Fig. 10). The IMRT prostate plan generated was also
used for delivery quality assurance (DQA). In our department, the
IMRT QA process consists of point and film dose measurements
using an acrylic phantom. The results of the film QA for the 18 MV
photon beam delivery are shown in Fig. 11. The point dose
measurement was within the 73% tolerance used in our clinic.
Similar results were observed for the 6 MV photon beam.
A comparison of the IMRT plans utilizing the unflattened
photon beams against the those with flat photon beams was
performed. The same dose constraint objectives to the critical

Fig. 9. Isodose distributions comparison of sample prostate IMRT plans using 18 MV (left) and 6 MV (right). Plans with unflattened photon beams are at the top and
flattened photon beams at the bottom.

Dose Volume Histogram Dose Volume Histogram

1.0 1.0
PTV
0.9 0.9
PTV
0.8 0.8
0.7 0.7 bladder
Norm. Volume

Norm. Volume

0.6 bladder 0.6

0.5 0.5
0.4 0.4
rectum
0.3 rectum 0.3
Seminal vesicles
Seminal vesicles
0.2 0.2
0.1 0.1
0.0 0.0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 0 1000 2000 3000 4000 5000 6000 7000 8000 9000
Dose (cGy) Dose (cGy)

Fig. 10. DVH comparison between the flat and unflattened photon beam delivery of prostate IMRT using 6 (left) and 18 MV (right) (solid lines—flattened beam, dotted
line—unflattened beam).
 ARTICLE IN PRESS
1634 S. Stathakis et al. / Applied Radiation and Isotopes 67 (2009) 1629–1637

1 100 100

dose normalized /%
80

dose normalized /%
80

difference /%

difference /%
3
60 60
4
40 calculated 40 calculated
measured measured
5 difference difference
20 20
6
0 0
7
-1 0 1 2 3 4 5 6 7 8 9 10 -1 0 1 2 3 4 5 6 7 8 9 10
1 2 3 4 5 6 7 distance /cm distance /cm
cm

Fig. 11. IMRT QA results for the unflattened 18 MV photon beam delivery. From left to right: gamma index, vertical, and horizontal profiles.

Fig. 12. Isodose distribution comparison for the head–neck, brain, and lung case (top to bottom).
 ARTICLE IN PRESS
S. Stathakis et al. / Applied Radiation and Isotopes 67 (2009) 1629–1637 1635

organs were assigned during the inverse planning optimization Table 1

for both plans. Isodose distributions are shown in Fig. 12. From Monitor units per fraction required for the same prescription.
this figure, it is quite clear that the differences are minimal. This
IMRT site Flattened beam (MU) Unflattened beam (MU)
observation is also reinforced when comparing the respective
DVHs (Fig. 13). There were no substantial differences and based on Prostate 6 MV 501 191
these results no clear dosimetric advantage exists in using Prostate 18 MV 435 132
flattened beams instead of unflattened beams. With regard to Head and neck 590 254
Brain 486 156
treatment parameters, the unflattened photon beam IMRT plans Lung 290 112

Dose Volume Histogram

1.0
required less MUs (Table 1) since the linac output is almost double
0.9
for 6 MV and about three times higher for the 18 MV. It should be
0.8 PTVs
noted that the time required to produce the plans was the same
0.7 Mandible for the unflattened and flattened photon beams.
Norm. Volume

0.6
0.5 4. Discussion
0.4
Our results indicate that the commissioning of an unflattened
0.3
photon beam can be performed in the Pinnacle3 TPS in a similar
0.2 Spinal cord fashion as flattened photon beams. The Pinnacle TPS requires the
0.1 parotids import of beam PDDs and profiles of several field sizes for each
photon beam defined by jaws and MLC. In addition, the total
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 scatter factor for each field size must to be inputted. The
automatic modeling option of the TPS provides a good initial
Normalized Dose (1.0 = 8031.374 cGy)
solution, but the user must fine tune the beam to achieve the best
Dose Volume Histogram agreement. The parameters that needed to be adjusted in the case
1.0 of the unflattened beams were the energy spectrum, the buildup
region of the PDDs and, the inside the field parameters. These
0.9 PTV
adjustments were necessary because the automatic modeling
Optic chiasm
0.8 parameters in the TPS refer to flat beams. After the changes in
0.7 the energy spectrum and the incident fluence were made the
automatic modeling procedures in Pinnacle were employed. The
Norm. Volume

0.6
brainstem models were in very good agreement with the measured data.
0.5 Our measurements show that the output scatter factors for an
0.4 brain unflattened beam differ from those of a flattened beam with
differences up to 3–4% depending on the field size.
0.3 The unflattened photon beam PDDs show that the beams are
0.2 Spinal cord less penetrating relative to flattened beams. This comes to no
surprise since the lower energy photons are not filtered out as is
0.1
the case when the flattening filter is in place—beam hardening
0.0 effect. The spectrum of the unflattened beam usually presents a
0 1000 2000 3000 4000 5000 6000 7000 mean energy of lower value and is more skewed towards lower
Dose (cGy) photon energies in comparison to a flattened one. Comparable
Dose Volume Histogram results on the PDDs and profiles are reported by Vassiliev et al.
1.0 (2006b), also the migration of the dmax to shallower depths in
the absence of flattening filter was observed by the same
0.9
investigators.
0.8 PTV The absence of the flattening filter is more obvious in the
0.7 lateral profiles. The profiles show a cone effect; the profile’s
highest value is in the middle and the shoulders are underdosed
Norm. Volume

0.6 relative to the central axis. This effect is more pronounced for
0.5 carina larger fields, larger than 10  10 cm2. Comparable results have
been previously reported Mesbahi et al. (2007), Vassiliev et al.
0.4
Lt Lung (2006b), showing the ‘‘ice cream’’ cone shape of the profiles. For
0.3 small fields and especially for the small fields sizes (segments)
0.2 Spinal cord used during IMRT delivery, the absence of the flattening filter is
Rt Lung not important. The profile shape for small segments is ‘‘bell’’
0.1
shaped, and it is characterized by lateral disequilibrium behavior.
0.0 The greatest advantage of the removal of the flattening filter is
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
the increase in linac output. We measured a two-fold increase in
Dose (cGy) dose for the 6 MV and more than three-fold increase in dose for
Fig. 13. DVH comparison for plans with flattened beams (solid lines) and
the 18 MV photon beam output. This increase was far greater than
unflattened beams (dashed lines). From top to bottom: head–neck, brain and the decrease of the beam’s penetration. Similar reports of
lung IMRT. increased dose rate along the central axis have been observed
 ARTICLE IN PRESS
1636 S. Stathakis et al. / Applied Radiation and Isotopes 67 (2009) 1629–1637
by other investigators (Cashmore, 2008; Fu et al., 2004;
Mesbahi, 2007; Mesbahi et al., 2007; Vassiliev et al., 2006b).
The dose rate output measured without the flattening filter is
within the range reported in the literature, for example, Vassiliev
et al. (2006b) reported an increase of 2.3 and 5.5 for the 6 and
18 MV beams, respectively. The high output had a great impact in
the number of MUs required to deliver the IMRT plans using
unflattened beams. The decrease of MUs was highest for the
18 MV photon beam which as mentioned above had the largest
output increase. The reduction of the total number of MUs has
many implications: shorter treatment times, less dependence on
patient’s movement due to breathing, and less leakage through
the linac’s head, hence reducing the risk of secondary malig-
nancies (Kry et al., 2005a, b, 2006, 2007b–d; Stathakis et al.,
2007). The fewer MUs required reduce the shielding requirements
of the machine, especially for the higher photon beam energies
where the neutron production needs to be considered (Kry et al.,
2007a).
It should be noted that the quality of the IMRT plans was not
compromised by utilizing the flattening filter free photon beams.
The plans had the same dosimetric characteristics which were
reflected in the isodose distributions and in the DVH comparisons.
Especially in the case of the prostate case, analogous were the
findings in the report by Vassiliev et al. (2007a). The advantage of
the unflattened photon beam IMRT plans was the decrease of MUs
required to deliver the plans. The fewer MUs required, shorten the
beam-on time significantly and that alone can be of great
advantage when targets subject to motion are to be treated such
as lung and liver tumors. Another potential of the high output can
be realized for gated treatments where the beam is ‘‘ON’’ only
during a limited time during the patients breathing cycle
(about 25–33%). With the higher beam output and shorter
treatment times one can also increase patient throughput.
The IMRT QA performed for the prostate case was well within
the acceptance criteria of 3% for the point measurement and more
than 90% of the pixels to have gamma index of 1 or less (3% of dose
difference and 3 mm distance to agreement). The results shown in
Fig. 11, are well within our acceptance criteria. This reveals that
the treatment planning system was able to model the unflattened
beam and predict the dose distributions accurately. The results of
the QA have the typical accuracy that we expected from our
flattened beam IMRT QA. Also, the point dose measurements for
the flattening filter free plans were within the 3% from the
predicted dose by the treatment planning system. Although, there
are some discrepancies between the modeled and the measured
profiles for some of the fields (especially the smaller ones), the
IMRT QAs were within tolerances. Comparing the differences
between the predicted and the measured values for the point
measurements between the plans with and without flattening
filters there was no observed trend to suggest that an over or
under prediction. Overall it appears that the Pinnacle3 TPS could
model the unflattened beam using the measured PDDs, profiles
and total scatter factors.
Other issues that should be addressed and they are out of the
scope of this investigation involve radiobiological consequence
due to the higher dose rate treatment obtained by removing the
flattening filter. Modern radiation therapy linacs have already
being used with dose rate output as high as 1000 MU/min. The
Novalis TX (BrainLab, Feldkirchen, Germany) using the high dose
rate output without flattening filter for stereotactic treatments.
The HiArt Tomotherapy (tomotherapy, Madison, WI) machine has
dose rate that is in the range of 900 MU/min while at the same
time the source axis distance is 80 cm. And the Cyberknife
machine is delivering the dose at 800 MU/min. Although, the high
dose rate treatments are widely used, a closer look of radio-
biological issues might be of great importance.
Finally, it should be pointed out that the delivery of the
treatment at dose rates of 2000 MU/min or higher (as in the case
of the 18 MV) might be compromised by limitations of the MLC.
For example, sliding window delivery might me limited by the
maximum speed the MLC leaves can achieve. IMRT QA is
recommended prior to treatment to ensure that the delivery of
the dose is within tolerances.
5. Conclusions
A flattening filter free linear accelerator with two photon
beam energies was commissioned and modeled for IMRT
planning and delivery. The commissioning process was similar
to that normally undertaken for a flattening filter equipped
linac. Due to the absence of the flattening filter, the PDDs
of the unflattened beams were observed to be less penetrating
than the respective flattened photon beams. The shape of the
profiles was such with a peak at the central axis and an almost
linear decrease at off axis points, resembling the shape of an
upside-down ice cream cone. The profile shape was more
pronounced for larger field sizes. The output of the linac was
increased by a two-fold for the 6 MV energy photon beam and
more than three times for the 18 MV beam. IMRT plans for
various sites created with these beams were clinically
equivalent to the ones created using flat beams. The advantage
of the delivery of the IMRT plans was the lower number of MUs
required per field leading to reduced beam-on times, shorter
overall treatment and potentially reduced incidence of secondary
malignancies.
References
Adler Jr., J.R., Chang, S.D., Murphy, M.J., Doty, J., Geis, P., Hancock, S.L., 1997. The
Cyberknife: a frameless robotic system for radiosurgery. Stereotactic and
Functional Neurosurgery 69, 124–128.
Cashmore, J., 2008. The characterization of unflattened photon beams
from a 6 MV linear accelerator. Physics in Medicine and Biology 53,
1933–1946.
Fu, W., Dai, J., Hu, Y., Han, D., Song, Y., 2004. Delivery time comparison for
intensity-modulated radiation therapy with/without flattening filter: a plan-
ning study. Physics in Medicine and Biology 49, 1535–1547.
Jeraj, R., et al., 2004. Radiation characteristics of helical tomotherapy. Medical
Physics 31, 396–404.
Kry, S.F., Salehpour, M., Followill, D.S., Stovall, M., Kuban, D.A., White, R.A., Rosen II,
2005a. The calculated risk of fatal secondary malignancies from intensity-
modulated radiation therapy. International Journal of Radiation Oncology,
Biology, Physics 62, 1195–1203.
Kry, S.F., Salehpour, M., Followill, D.S., Stovall, M., Kuban, D.A., White, R.A., Rosen II,
2005b. Out-of-field photon and neutron dose equivalents from step-and-shoot
intensity-modulated radiation therapy. International Journal of Radiation
Oncology, Biology, Physics 62, 1204–1216.
Kry, S.F., et al., 2006. A Monte Carlo model for calculating out-of-field dose from a
varian 6 MV beam. Medical Physics 33, 4405–4413.
Kry, S.F., Titt, U., Ponisch, F., Vassiliev, O.N., Salehpour, M., Gillin, M., Mohan, R.,
2007a. Reduced neutron production through use of a flattening-filter-free
accelerator. International Journal of Radiation Oncology, Biology, Physics 68,
1260–1264.
Kry, S.F., Followill, D., White, R.A., Stovall, M., Kuban, D.A., Salehpour, M., 2007b.
Uncertainty of calculated risk estimates for secondary malignancies after
radiotherapy. International Journal of Radiation Oncology, Biology, Physics 68,
1265–1271.
Kry, S.F., Price, M., Followill, D., Mourtada, F., Salehpour, M., 2007c. The use of LiF
(TLD-100) as an out-of-field dosimeter. Journal of Applied Clinical Medical
Physics/American College of Medical Physics 8, 2679.
Kry, S.F., et al., 2007d. A Monte Carlo model for out-of-field dose calculation from
high-energy photon therapy. Medical Physics 34, 3489–3499.
Mackie, T.R., et al., 1993. Tomotherapy: a new concept for the delivery of dynamic
conformal radiotherapy. Medical Physics 20, 1709–1719.
Mesbahi, A., 2007. Dosimetric characteristics of unflattened 6 MV photon beams of
a clinical linear accelerator: a Monte Carlo study. Applied Radiation and
Isotopes 65, 1029–1036.
Mesbahi, A., Mehnati, P., Keshtkar, A., Farajollahi, A., 2007. Dosimetric properties of
a flattening filter-free 6 MV photon beam: a Monte Carlo study. Radiation
Medicine 25, 315–324.
 ARTICLE IN PRESS
S. Stathakis et al. / Applied Radiation and Isotopes 67 (2009) 1629–1637
Sharma, S.C., Ott, J.T., Williams, J.B., Dickow, D., 2007. Commissioning
and acceptance testing of a CyberKnife linear accelerator. Journal of
Applied Clinical Medical Physics/American College of Medical Physics 8, 2473.
Stathakis, S., Li, J., Ma, C.C., 2007. Monte Carlo determination of radiation-induced
cancer risks for prostate patients undergoing intensity-modulated radiation
therapy. Journal of Applied Clinical Medical Physics/American College of
Medical Physics 8, 2685.
Titt, U., Vassiliev, O.N., Ponisch, F., Dong, L., Liu, H., Mohan, R., 2006. A flattening
filter free photon treatment concept evaluation with Monte Carlo. Medical
Physics 33, 1595–1602.
Vassiliev, O.N., Titt, U., Kry, S.F., Ponisch, F., Gillin, M.T., Mohan, R., 2006a. Monte
Carlo study of photon fields from a flattening filter-free clinical accelerator.
Medical Physics 33, 820–827.
1637
Vassiliev, O.N., Titt, U., Ponisch, F., Kry, S.F., Mohan, R., Gillin, M.T., 2006b.
Dosimetric properties of photon beams from a flattening filter free clinical
accelerator. Physics in Medicine and Biology 51, 1907–1917.
Vassiliev, O.N., Kry, S.F., Kuban, D.A., Salehpour, M., Mohan, R., Titt, U.,
2007a. Treatment-planning study of prostate cancer intensity-modulated
radiotherapy with a Varian clinac operated without a flattening
filter. International Journal of Radiation Oncology, Biology, Physics 68,
1567–1571.
Vassiliev, O.N., Titt, U., Kry, S.F., Mohan, R., Gillin, M.T., 2007b. Radiation safety
survey on a flattening filter-free medical accelerator. Radiation Protection
Dosimetry 124, 187–190.
Zefkili, S., Kappas, C., Rosenwald, J.C., 1994. On-axis and off-axis primary dose
component in high energy photon beams. Medical Physics 21, 799–808.