The Applicability of Cone-beam CT-based 3D IGRT for Head and Neck Radiotherapy| Zoller
The Applicability of Cone-beam CT-based 3D
IGRT for Head and Neck Radiotherapy
The Development of a Standard Protocol and Methods for
Improvement on Existing Immobilization
BY WESLEY ZOLLER
THE OHIO STATE UNIVERSITY
Abstract: This study explores the applicability of Cone-beam CT (CBCT) image guidance and its ability to detect/adjust translational
and rotational discrepancies in patient set-ups. CBCT has shown merit for alignment use in areas such as the head and neck where
many critical structures lie within close proximity to the PTV. However, this study also looks at the added dose accumulation from
CBCT and the potential harm resulting from this additional dose. In this study, we will measure the set-up discrepancies recognized
by CBCT for 20 patients taken throughout the course of treatment at the Arthur G. James Cancer Hospital & Richard J. Solove
Research Institute as a means of establishing the primary sources of set-up error for the existing immobilization techniques. In
addition, we suggest two immobilization techniques that may reduce this source of error—one involving an Alpha Cradle © foam
headrest built inside an aquaplast mask and another involving the use of a mask-attachable bite block to reduce head tilt. Using
CBCT, the set-up discrepancies will be measured using the same methodology to see if either reduces the set-up and internal
variance throughout treatment. Improvements on immobilization could potentially allow for reduced CBCT exposures, reduced dose,
a smaller PTV margin, and a higher dose to tumor volume.
Introduction
In external beam radiation therapy (EBRT),
the primary goals remain the same. As a standard,
the ideal plan sticks to a three prong premise. The
plan will (1) have high conformity to the target
volume allowing for higher tumor dose for ablative
bio-effects on malignant cells, (2) have sharp dose
fall-off to avoid irradiation to critical structures, and
(3) ultimately produce the best possible outcome for
the patient. To achieve this plan, many innovations
have been made in external beam radiotherapy over
the years. These delivery practices range anywhere
from intensity-modulated radiation therapy with
respiratory gating to linac-based stereotactic
radiotherapy. However, one major player in the goal
of producing a higher and more conformal dose is in
the reliability of the set-up itself. Quite simply,
tumor conformity serves to be useless if the treatment
target varies in location from day to day due to set-up
inconsistencies.
To help alleviate this set-up
The Ohio State University| Radiation Therapy Class of 2012
variability, the use of image guidance (IGRT) has
become a standard method in the treatment process.
Port films and on-board imaging (OBI) digital fluororadiographs provide the ability to align to bony
anatomy each day in accordance with the plan. The
capability to acquire MV or kV conebeam images
from the linac even furthers the ability to look at
internal structures and tumor volumes, in addition to
anatomy of the bone. From this, it is possible to
compare the set-up in the treatment room to the
original planning CT as it pertains to the target of
interest. Many studies have been conducted
documenting the improvements of CBCT and its
imaging capabilities compared to that of orthogonal
imaging. These studies will be outlined in greater
detail later in this paper. Nevertheless, areas of
anatomical variance that possess multiple critical
structures can benefit from additional alignment
practices.
Page
| Abstract-1
The Applicability of Cone-beam CT-based 3D IGRT for Head and Neck Radiotherapy | Zoller
The head and neck region is an anatomical
area that tends to have the greatest range of
variability as it pertains to the clustered presence of
critical structures and day-to-day discrepancies.
Many of these structures such as the spinal cord,
parotid glands, brain stem, esophagus, mandible, and
larynx tend to fall directly in the treatment field for
tumors of this region. In 1995, a group from the New
York University Medical Center detailed the
potential late bio-effects for individuals treated with
EBRT to the head and neck region.1 The group
specifically evaluated the mucosal lining, salivary
glands, teeth, and mandible. Permanent xerostomia,
loss of taste, severe mucositis, radionecrosis of bone,
and perforation and permanent narrowing of the
esophagus were all included in the radio-effects when
the dose became too high. The group evaluated the
tolerance doses for these late effects and concluded
by reiterating the importance of immobilization and
avoidance of these structures whenever possible in
radiotherapy.1
Due to the presence of these structures, as
well as the increased motility of the neck curvature
coupled with head tilt, the set-up of head-and-neck
patients is the primary concentration of this project.
Quite simply, it is an area with high propensity for
variation. To this point, as stated, CBCT has become
a major resource for correcting set-up variances for
this treatment area.
Resulting from conebeam
capabilities, it has even been suggested that PTV
margins could be reduced due to the fact that CBCT
image guidance allows for internal realignment on a
daily basis. However, it is important to keep image
guidance in perspective.
A fact that cannot be forgotten is that daily
image guidance—especially that of CBCT—is excess
dose to the patient. It is important to determine this
level of additional dose before a proper risk-reward
analysis can be performed. Determining the exact
dose from CBCT has been a premise in multiple
existing studies, and these specifically will be
addressed in the Review of Literature. It is also
important to note that many patients with
malignancies have already had multiple diagnostic
imaging procedures performed as a work-up, most of
which involve ionizing radiation. Combined with
this is the fact that follow-up scans will become
regular practice post-completion of treatment. It is
not known exactly what percentage of added bodily
dose can lead to a secondary malignancy, but every
portion of this dose increases the potential for
stochastic effects. In addition to this, shifting daily
based on conebeam images does not necessarily
make reducing PTV margins a reasonable practice.
The Ohio State University| Radiation Therapy Class of 2012
Often times, due to the motility of the anatomy and
set-up variance in the head and neck region, the
alignment from the CBCT image to the planning CT
image has a degree of discrepancy. Due to this, it is
often necessary to utilize every bit of the existing
planned internal tumor margin in order to
accommodate for the avoidance of the critical
structures. “Splitting” between maintenance of the
margin and keeping the spinal cord from lying in the
95% isodose, for example, can be a daily problem.
The internal margin, after all, is there for a reason.
Conebeam is a useful tool in measuring and
adjusting for this variance, but if reducing PTV
margins is the goal, as mentioned previously, the root
of the situation all ties back to the initial set-up
immobilization tactics. A plan can only be as
accurate as its set-up is consistent. To improve the
standard head-and-neck immobilization procedure, it
is necessary to know where the source of the error is
coming from. To do this, MV Conebeam image
guidance will be used in this proposed study to
measure the translational and rotational errors of the
existing set-up—conformal board, custom headrest,
and thermoplastic mask that covers the shoulder.
Once the primary sources have been determined, two
immobilization tactics that are hypothesized to
improve on these weaknesses will be tested—one
involving an Alpha Cradle © foam headrest built
inside the thermoplastic mask and another involving
the use of a mask-attachable bite block to reduce
head tilt. The measurements will be performed using
the same CBCT procedure as for the initial
evaluation. The purpose is to determine if these
methods alone or combined can improve
immobilization reproducibility from day to day. In
turn, this could lessen the need—and thus, the dose—
of CBCT rather than require guidance daily.
Potentially, improved immobilization tactics would
make it possible to achieve reduced PTV margins and
generate a plan that is highly conformal to tumor
volume. With the ability to escalate tumor dose
while still sparing critical structures, it is possible to
achieve the ultimate goal of improving patient
outcomes.
Review of Literature
This proposed study will use conebeam
imaging as a means of measuring set-up error.
Multiple studies have been performed to evaluate the
usefulness of conebeam CT as a method for
evaluating these rotation-based errors in addition to
the translational variety. In addition, the studies
provide rationale behind the selection of conebeam
Page
|1
The Applicability of Cone-beam CT-based 3D IGRT for Head and Neck Radiotherapy | Zoller
imaging as a measuring tool when compared to its
orthogonal counterpart. In order to assess the
positioning accuracy of 3D Conebeam CT IGRT
versus 2D orthogonal kV OBI IGRT, researchers
from the University of Texas M.D. Anderson Cancer
Center studied 21 head-and-neck cancer patients
undergoing radiotherapy.2 The results showed that
when physicians used the IGRT to make couch shifts,
4.1% of the CBCT group had shifts greater than or
equal to 0.5 cm in at least one direction, and 18.7%
were over 0.3 cm. For the kV OBI image guidance,
only 1.7% had shifts greater than 0.5 cm and 11.2%
were greater than or equal to 0.3 cm. Greater shifts
on the 3D image indicated that the ability to visualize
target tissue along with the bone anatomy could
potentially be more accurate than the kV orthogonal
images. These results were deemed statistically
significant, but the team noted that the possibility
remains that the differences could be caused by
immobilization error and potential rotation of the
head and spine due to set-up inconsistencies.2 This
immobilization error is what this further proposed
study will be designed to evaluate and address.
In a similar study, researchers from the
University of Wuerzberg, Germany used CBCT to
determine the effectiveness of conebeam guidance
versus orthogonal imaging in correcting translational
and rotational setup errors.3 In the study, the
researchers assessed the set-up accuracy of 24
patients—8 head and neck, 10 pelvic, and 6
thoracic—using a total of 209 CBCT studies and 148
electronic portal images. To do this, the team
measured translational errors alone, rotational errors
alone, and the effect of the combination of both
translational and rotational error on target coverage
and critical organ sparing. The results of the study
showed that CBCT and portal imagining differed by
less than 1 mm in translational error for 70.7% of the
patients and less than 2 mm for 93.2% of the patients.
However, with the use of CBCT, the researchers
found rotational errors greater than 2° in 3.7% of
pelvic tumors, 26.4% of thoracic tumors, and 12.4%
of head-and-neck tumors. The maximum rotational
errors for each body site were 5° in the pelvis, 8° in
the head and neck, and 6° in the thorax. Based on
comparing this to the initial treatment plans, the
researchers concluded that for patients with longer
target volumes and targets close to critical structures,
the combination of translational and rotational errors
significantly lowered target coverage and highly
increased doses to the organs at risk. The team
suggested that with the feasibility of daily conebeam
guidance, the ability to successfully diminish
significant rotational error is essential for highprecision external beam radiation therapy.3 If the goal
The Ohio State University| Radiation Therapy Class of 2012
is to reduce PTV margins, which is typically around
5 mm, every millimeter counts. Although CBCT and
orthogonal correction only differed by 2mm for
93.2% of patients in the correction of translational
error, this 2mm can be significant to ensuring that the
tumor volume falls within the PTV.
In addition to the Wurzberg study,
researchers at the Institute of Cancer Research in
Surrey, UK evaluated the application of kV
conebeam computed tomography image guidance
when compared to electronic portal imaging (EPI) in
correcting for set-up errors in oesophageal EBRT
patients.4 In the study, the researchers measured
variation in anterior–posterior, right–left, and craniocaudal directions as well as the rotational directions
of pitch, roll, and yaw. In all, 20 patients with
oesophageal tumors were implemented in the study,
and set-ups included skin marks/tattoos, along with
kV-CBCT or EPI prior to treatment which were
compared to planning CT slices or corresponding
DRR’s. A total of 122 EPI orthogonal images and
207 CBCT scans were collected, and the systematic
and random errors were calculated. The authors
found that the systematic and random errors for
CBCT guidance were 1.3 mm, 1.7 mm, 1.4 mm, and
2.6 mm, 3.9 mm, 2.0 mm in right-left, cranio-caudal,
and anterior-posterior directions, respectively. The
results for the EPI were not significantly different.
However, CBCT was able to identify rotational errors
greater than 3o in 44 images, something EPI could not
identify. Based on these results, the research team
concluded that CBCT’s adequate 3D volumetric
image quality improves the accuracy of treatment
delivery, especially by catching rotational errors.4
Expanding the applicability of conebeam CT
image guidance, researchers at the Cancer Research
Center and the University of Heidelberg in
Heidelberg, Germany also included tumors located in
the chest and pelvis in their study.5 The group
selected 6 patients with different tumor diagnoses—
lung cancer, sacral chordoma, head-and-neck,
paraspinal tumor, and two with prostate cancer—in
order to assess the set-up correction applicability of
kV CBCT. The head-and-neck and lung patients
were immobilized using a vacuum pillow in
combination with a head mask, while patients with
prostate cancer or paraspinal tumors were
immobilized using a wrap-around body cast and a
head mask. The image quality of the kV CBCT was
deemed sufficient in all cases, and the bony
landmarks and anatomical target were visualized
clearly, with the exception of poor soft tissue contrast
for an obese prostate patient. The results showed a
maximum set-up deviation of 3 mm for patients
Page
|2
The Applicability of Cone-beam CT-based 3D IGRT for Head and Neck Radiotherapy | Zoller
immobilized with the body frame, and 6 mm for
patients with the vacuum pillow. In 4 of the 6 cases,
the target point was corrected. The authors predicted
that the CBCT took an extra 10-12 minutes
maximum, and was significant in the correction of
set-up deviation—especially in areas with many
critical structures such as the heart, spinal cord, and
esophagus—for high dose stereotactic treatment.5
All of these studies seem to indicate the
usefulness of CBCT as a guidance tool, especially to
correct rotational errors where orthogonal imaging is
inadequate. However, an error due to a patient being
rotated for treatment set-up who was not rotated in
the planning CT cannot be simply remedied by
translational table shifts. This could only be the
result of immobilization error, as the M.D. Anderson
study seemed to indicate. Due to this fact, relying on
CBCT image guidance to correct for inconsistency
errors would only go so far. The immobilization
techniques must improve. CBCT IGRT on a daily
basis cannot fully correct the problem, especially if it
is delivering excess radiation to the patient.
As previously discussed, it is necessary to
get an estimation of this excess radiation dose from
CBCT image guidance before the trade-offs can be
properly examined. Multiple studies have been
performed to get this estimation. In a study from the
Princess Margaret Hospital’s department of oncology
(Hong Kong), a research group sought to determine the
absorbed organ doses and effective tissue doses from
using conebeam CT image guidance.6 Three different
scan sites and 26 organs were specifically outlined
for this study: the head-and-neck, the chest, and the
pelvis.
Using an on-board imager (OBI) kV
conebeam system and numerous thermoluminescent
dosimeters (TLD’s), effective body doses and
absorbed organ doses were measured on a female
anthropomorphic phantom. The study also utilized
both the standard CBCT mode and the low-dose
mode. The results showed single-scan skin doses
from standard CBCT of 6.7 cGy for the head and
neck, 6.4 cGy for the chest, and 5.4 cGy for the
pelvis. The TLD’s showed single-scan effective
body doses of 10.3 mSv for the head and neck, 23.7
mSv for the chest, and 22.7 mSv for the pelvis.
When repeated for low-dose CBCT, the
measurements were approximately 20% of the
standard mode readings. The group estimated that
this could total to over 2 Gy over the course of
treatment for standard conebeam image guidance,
and could potentially increase the risk of secondary
cancer by 2-4%. The research team concluded by
discussing the importance of reducing mAs whenever
possible and using low-dose CBCT in areas where
The Ohio State University| Radiation Therapy Class of 2012
the bony anatomy is the key structural set-up
indicator.6
Along similar lines, a research team from
the department of radiation physics of the University
of Toronto also performed a study to determine the
significance of the dose accumulation to normal
tissue from daily kV-CBCT image guidance.7 The
dose measurements were performed on an Elekta
linear accelerator using a farmer ion chamber and
MOSFET detectors at the surface, periphery (2 cm
depth), and center of 2 cylindrical water phantoms,
16 and 30 cm in diameter. Along with this, the study
utilized both 360° and half-rotation conebeam scans
on the 30 cm diameter water phantom using a
standard setting of 120 kVp and 660 mAs. The
results showed that the maximum dose at the center
measured to be 1.6 cGy and 2.3 cGy at the surface of
the body for full rotation CBCT. The researchers
estimated that this dose could account to over 1 Gy to
normal tissue at the skin surface for patients
receiving the image guidance daily over the course of
external beam treatment. The team also noted that
the dose cannot be accurately added into the
dosimetry treatment plan due to suspicion that kV
energy has different biological effects on cells than
mV treatment. Due to this, the results show that the
scanning parameters should be reduced whenever
possible, and this can be accomplished using less
kVp, smaller fields-of-view in the longitudinal
direction, and by using only half-rotation conebeam
scanning for image guidance.7
The results from these two studies indicate
the significance dose delivery from CBCT imaging,
especially in the terms of kV conebeam. The quality
of kV imaging tends to produce better resolution,
which gives a better comparison to the kV-acquired
planning CT. Avoiding daily guidance would help to
reduce this excess dose accumulation, but only as an
effective trade-off if immobilization can be done
sufficiently enough to merit reducing the number of
CBCT’s.
In 1995, Gunilla Bentel and a research group
from Duke University Medical Center detailed the
use of a customized set-up for head and neck cancer
that included the use of a foaming agent customized
headrest similar to the Alpha Cradle © material used
in this study.8 Though dated, the team documented
the conformity of the headrest to the back of the skull
and stabilization of the neck. The study included a
customized headrest contouring to the posterior head
and extending to the inferior angle of the posterior
neck. This also included a thermoplastic facemask
contouring to the brow, nose, and chin. The custom
Page
|3
The Applicability of Cone-beam CT-based 3D IGRT for Head and Neck Radiotherapy | Zoller
headrest was assembled using a 10 x 18 inch plastic
bag and 100 cc of foaming agent, which the authors
estimated of taking less than 10 minutes to create.
The effectiveness of the set-up was then compared to
conventional set-ups utilizing standard headrests.
The team examined the custom head rests created for
20 head-and-neck patients, 2 being children. The
height under the head ranged from 5.8- 9.8 cm and
under the neck from 3.2 -6.2 cm, while the standard
head supports ranged from 5.0-8.0 cm and 1.0- 5.0
cm, respectively. Also, the superior to inferior
distance between the highest point under the neck
and the lowest point under the head for the
customized supports ranged from 5.0-14.0 cm, while
only 9.0-11.0cm in the standard supports. The study
concluded that the wide span of the measurements
reflected the variability between the patients, and that
the standard head supports do not adequately
conform to the posterior contour. The authors also
suggested that the shape of the customized headrest
provides support to the most inferior aspect of the
neck and the upper thorax, which is crucial for
reproducibility of the cervical spine and is not
addressed by standard head supports.8 This set-up
was as improvement over standard Silverman
headrests due to conformity to the patient’s head and
spine. In the set-up of this proposed study, the Duke
principle will be built upon. By allowing the
foaming agent to expand to the brink of the
thermoplastic shoulder mask, the goal would be to
eliminate gaps, which, as a result, could improve
reproducibility.
In the area of the head and neck, the critical
structures tend to shift with the patient based on the
curvature of the neck and chin—a reason for being
the primary area of interest in the proposed study.
This shift tends to occur if the mask does not fit
correctly or if the head rest does not properly support
the entire neck. In a study performed by the Albert
Einstein College of Medicine in Bronx, NY, the
research team actually hypothesized that some headand-neck cancer patients tend to lose their lordotic
cervical spine curvature throughout the course of
external beam treatment.9 This spinal shifting can
cause for anatomical marks to be varied from the
original marks and cause divergence from the
treatment field originally depicted on the planning
CT. Using a 6 mV linear accelerator, the authors
performed a study detailing the treatment of 50 headand-neck patients over the course of 40-42 elapsed
days. Ensuring reproducibility, the therapists used
shoulder straps, custom thermoplastic face masks,
and multiple tattoos. To measure the change in this
lordotic curvature, the investigation team defined a
cervical spinal angle (CSA) to be calculated by
The Ohio State University| Radiation Therapy Class of 2012
recording the initial lateral port field images and
comparing these to port films in the final week. The
borders of the CSA included a projected line running
parallel to the posterior surface of the C2 vertebral
body and another line lying parallel to the body of the
C6 vertebrae. The results showed that on average,
the CSA of the patients decreased by 2.26° over the
course of the external beam treatment. For patients
with the lowest isocenters, the results were higher
showing a 3.8° change, which translates to a 1 cm
divergence from a distance of 15 cm. Based on the
results, the group suggested that orthogonal films be
used to check spinal cord alignment prior to
treatment as the elapsed days accumulate or perhaps
using a more comfortable means of immobilization to
keep the curvature from decreasing throughout the
duration. 9 This is what the proposed immobilization
tactics hope to account for, as well as expand further
on the study to measure consistency of the curvature
of the spine throughout treatment.
In an effort similar to the proposed study, a
group from the University of Heidelberg and
Department of Oncology of the Mannheim Medical
Center in Mannheim, Germany used 3D/3D
matching, x-ray volume rendering, and conebeam CT
image guidance to assess the accuracy of two
different mask systems—the thermoplastic mask and
the rigid Delta cast.10 In the study, 21 patients—14
had rigid masks and 7 had thermoplastic—with
intracranial or head-and-neck tumors were assessed
using the 3D/3D matching and CBCT, separating the
skull and neck regions. Translational and rotational
errors of the isocenter were recorded as well as an
isolated analysis of the error based on cervical
vertebra. To analyze the results, the researchers
created a displacement vector, v, by taking the square
root of the quantity, (x2 +y2 + z2), where x, y, and z
represented shifts in the lateral, vertical, and
longitudinal directions. This displacement vector
quantity for the rigid masks calculated as 0.312 cm
for intracranial tumors and 0.586 cm for the neck.
For the thermoplastic mask, the researchers measured
a vector of 0.472 cm for intracranial tumors, and
0.725 cm for the neck. In association with this, rigid
masks accompanied with body tattoos had a v length
of 0.35 cm in the neck region. In conclusion, the
research team suggested that the Delta cast rigid
masks were superior for intracranial precision, and
that for the neck region, both thermoplastic and rigid
should be accompanied by CBCT guidance.10
Building on this Heidelberg study, the
proposed study entails the use of the thermoplastic
mask in combination with the Alpha Cradle © foam
custom headrest, along with a mask-attachable bite
Page
|4
The Applicability of Cone-beam CT-based 3D IGRT for Head and Neck Radiotherapy | Zoller
block to reduce head tilt. The effort is to create a
system that diminishes the need for the CBCT
guidance—at least not daily or multiple times per
week. Along similar lines for measurement, the
proposed study will also utilize the calculation of
displacement vector by taking the square root of the
quantity (x2 +y2 + z2).
Methods
In this study, all images will be obtained
from patients who have or will have undergone
IMRT treatment at the Arthur G. James Cancer
Hospital & Richard J. Solove Research Institute for
head and neck tumors of varying stages and
histologies. All conebeam images will be captured
using a MV CBCT image of a 220° arc performed on
a Siemens Oncor model treatment unit complete with
EPID. The number of conebeam studies per patient
will vary from 10-15 as the patients were imaged
throughout the course of treatment 2-3 times per
week.
For the preliminary portion of the study, the
CBCT images will be examined from 20 existing
head and neck cancer patients of varying stages and
histologies. All patients in the study will have
received 35 total fractions—once daily for 5
treatments per week—for a total dose of 70 Gy. In
addition, the patients will be non-boosted and never
have been remarked throughout treatment. The
standard set-up will include a conformal board, an
aquaplastic mask that extends over the shoulders, a
custom head rest that extends to the base of the neck,
and a knee sponge. As a note, the patient will lie
supine with arms placed by his or her sides. The
translational shifts will be measured using the shifts
performed, approved, and documented by the
physician in the Lantis Record and Verify system. X
denotes shifts in the cross-plane or horizontal
direction, Y denotes shifts in the in-plane or vertical
direction, and Z denotes shifts in the longitudinal
position. Once these shifts are obtained, a standard
deviation will be performed by simply taking the
variance from the average of the values (X, Y, and Z
each done independently). In addition, the average
magnitude of the shifts will be obtained by taking the
absolute value of the shifts, and taking an average
based on directional independence. Once these are
determined, two vector quantities will be calculated
to better analyze the set-up error in three dimensions.
To calculate an average shift displacement vector,
Vshift, the square root of the quantity
(Xshift2+Yshift2+Zshift2) will be obtained using the
independent magnitude shift averages in each
direction. Then, a deviation vector will be calculated
The Ohio State University| Radiation Therapy Class of 2012
to determine the average variance of this vector by
using the equation: Vdev = (Xstandard dev2+Ystandard
2
2 (1/2)
.
dev +Zstandard dev )
Once the translational sources of error have
been measured, the next step is to quantify the
rotational errors. This will be done by individual
analysis of all conebeam images for these patients
using the “Image J” program. This program is free
software equipped with angle and distance
measurement tools, complete with screen capture
capabilities. The three rotation tilt planes examined
in this study will be the axial, sagittal, and coronal
directions. To measure the axial rotation, two lines
will be projected on an axial slice. The first will be a
vertically projected line, stopping at the tip of the
posterior tubercle of the atlas vertebrae (C1), and the
second line will start at this same tip of the posterior
tubercle and extend through the lateral tip of the right
transverse process of C1. The consistency of this
angle will be measured throughout treatment based
on the initial angle in the original planning CT. All
measurements will be done using the angle tool.
The measurement of sagittal tilt will be
analyzed on a sagittal slice. An angle will be
obtained between a vertically projected line (ending
at the superior tip of the odontoid process) and a line
starting at the superior tip of the odontoid process and
extending through the tip of the external occipital
protuberance.
This consistency, too, will be
measured from the planning CT as baseline. Using a
coronal image, the coronal tilt determination will be
performed by measuring the angle between a
horizontally projected line and a line running parallel
to the base of the skull at a level anterior to the
cochlea. This, too, will use the planning CT as a
baseline.
Once the primary sources of immobilization
inconsistencies have been quantified, this proposed
study suggests two methods of immobilization
improvement for a board reviewed study. In the first
wing, 10 patients will be set-up using the same set-up
as the preliminary analysis with the addition of a
mask-attachable bite-block. It will be formed on the
day of simulation and will be clipped into the mask
as the aquaplast forms through cooling. This will
allow for it to remain constant and permanent
throughout treatment. The aim of this bite block is to
prevent axial, sagittal, and coronal tilt variance as the
upper maxillary molars and maxillary incisors are
firmly and repetitively placed in the mouthpiece of
the block. The central premise behind this is that the
maxilla is much more stable than the mandible. In
the existing mask system, the primary reliance is on
Page
|5
The Applicability of Cone-beam CT-based 3D IGRT for Head and Neck Radiotherapy | Zoller
the conformity of the mask to the patient’s chin.
Adding one more point of resistance, such as the bite
block, is hypothesized to improve stability.
In the second wing of the proposed study, 10
patients will be treated with a similar set-up as the
preliminary study with the insertion of the Alpha
Cradle © foam custom headrest in the place of the
existing headrest. The headrest will be formed within
the confines of a curved shell mimicking the
maximum expansion of the posterior portion of the
aquaplast mask frame. This study would allow this
foaming agent to expand to its fullest within these
confines, eliminating the gaps between the mask and
the patient’s head as well as areas where the existing
headrests do not conform. The headrest will be
allowed to expand upward behind the ear and
conforming to the superior portion of the patient’s
skull. In addition, the foam will give stability down
the patient’s neck extending to the upper back. The
premise is that reduced gaps would hypothetically
reduce variability. The aquaplast mask will not be
made until after the foaming agent of the custom
head rest has been allowed to completely harden
(approximately 10 minutes).
In the third and final wing, 10 patients will
be set-up utilizing both suggested immobilization
improvements. The set-up would include the Alpha
Cradle © foam custom headrest in addition to the
mask-attachable bite block. Again, the patient would
be supine on a conformal board with arms by his or
her side.
All measurements and quantifications for
these three wings will be determined using CBCT
images and documented shifts, identical to the
evaluation of the preliminary portion of the study.
This will allow for an accurate comparison of set-up
error. All patients will also receive 70 Gy over 35
fractions, in the same lines as the preliminary
examination subjects.
3.
Guckenberger M, Meyer J, Vordermark D, Baier
K, Wilbert J, and Flent M. Magnitude and
clinical relevance of translational and rotational
patient setup errors: a cone-beam CT study. Int.
J. Radiation Oncology Biol. Phys. 2006;
65(3):934–942.
4.
Hawkins M, Aitken A, Hansen V, McNair H, and
Tait D. Set-up errors in radiotherapy for
oesophageal cancers – Is electronic portal
imaging or
conebeam
more
accurate?
Radiotherapy and Oncology. 2011; 98:249–254.
5.
Thilmann C, Nill S, Tücking T, et al. Correction
of patient positioning errors based on in-line cone
beam CTs: clinical implementation and first
experiences. Radiation Oncology. 2006; 1(16):19.
6.
Kan M, Leung L, Wong W, and Lam N.
Radiation dose from cone beam computed
tomography for image-guided radiation therapy.
Int. J. Radiation Oncology Biol. Phys. 2008;
70(1): 272–279.
7.
Islam M, Purdie T, Norrlinger B, et al. Patient
dose from kilovoltage cone beam computed
tomography imaging in radiation therapy. Med.
Phys. 2006; 33(6):1573-1582.
8.
Bentel G, Marks L, Sherouse G, and Spencer D.
A customized head and neck support system. Int.
J. Radiation Oncology Biol. Phys. 1995;
32(1):245-248.
9.
Garg M, Yaparpalvi R, and Beitler J. Loss of
cervical spine curvature during radiotherapy for
head-and-neck cancers: the neck moves, too. Int.
J. Radiation Oncology Biol. Phys. 2004;
58(1):185–188.
References
1.
2.
Cooper J, Fu K, Marks J, and Silverman S. Late
effects of radiation therapy in the head and neck
region. Int. J. Radiation Oncology Biol. Phys.
1995; 31(5):1141-1164.
10. Boda-Heggemann J, Walter C, Rahn A, et al.
Repositioning accuracy
systems—3D revisited:
3D/3D matching with
Radiation Oncology
66(5):1568–1575.
of two different mask
comparison using true
cone-beam CT. Int. J.
Biol. Phys. 2006;
Li H, Zhu XR, Zhang L, et al. Comparison of 2D
radiographic images and 3D cone beam
computed tomography for positioning head-andneck radiotherapy patients. Int. J. Radiation
Oncology Biol. Phys. 2008; 71(3):916–925.
The Ohio State University| Radiation Therapy Class of 2012
Page
|6