5TH JUCTS AND THE 5TH S. TAKAHASHI MEMORIAL INTERNATIONAL JOINT SYMPOSIUM
ASSESSMENT OF SPATIAL UNCERTAINTIES IN THE RADIOTHERAPY PROCESS
WITH THE NOVALIS SYSTEM
NAOKI HAYASHI, M.SC.,*yYASUNORI OBATA, M.D.,yYUKIO UCHIYAMA, M.SC.,z
YOSHIMASA MORI, M.D.,* CHISA HASHIZUME, M.D.,* AND TATSUYA KOBAYASHI, M.D.*
*Nagoya Radiosurgery Center, Nagoya Kyoritsu Hospital, Nagoya, Japan;yGraduate school of Medical Science, Nagoya University,
Nagoya, Japan; andzSchool of Health Sciences, Gifu University of Medical Science, Gifu, Japan
Purpose: The purpose of this study was to evaluate the accuracy of a new version of the ExacTrac X-ray (ETX)
tial uncertainties with the Novalis system.
Methods andMaterials: Threefactors of geometrical accuracy related to theETX system wereevaluatedby phan-
tom studies. First, location dependency of the detection ability of the infrared system was evaluated. Second, ac-
curacy of the automated calculation by the image fusion algorithm in the patient registration software was
evaluated. Third, deviation of the coordinate scale between the ETX isocenter and the mechanical isocenter was
evaluated. From the values of these examinations and clinical experiences, the total spatial uncertainty with the
Novalis system was evaluated.
Results: As to the location dependency of the detection ability of the infrared system, the detection errors between
the actual position and the detected position were 1% in translation shift and 0.1?in rotational angle, respectively.
As to the accuracyof patient verification software, the repeatability and the coincidence of the calculationvalue by
image fusion were good when the contrast of the X-ray image was high. The deviation of coordinates between the
ETX isocenter and the mechanical isocenter was 0.313 ± 0.024 mm, in a suitable procedure.
Conclusions: The spatial uncertainty will be less than 2 mm when suitable treatment planning, optimal patient
setup, and daily quality assurance for the Novalis system are achieved in the routine workload.
? 2009 Elsevier
Spatial uncertainty, Image-guided radiotherapy, Geometrical accuracy, Patient setup, Stereotactic radiotherapy.
quality assurance (QA) and quality control (QC) (10–12). The
delivery of radiation at a sufficient dose to a planning target
volume (PTV) is an ideal of external beam radiation therapy.
In the late 1990s, new image-guided patient positioning
devices were developed for a clinical linear accelerator
(LINAC). In addition, BrainLAB (Germany) recently manu-
factured a Novalis radiotherapy system which was dedicated
to highly precise radiotherapy. The Novalis system for ster-
eotaxy consists of a micro-multileaf collimator (m3) and an
ExacTrac X-ray (ETX) positioning system. The characteris-
tics of a micro-multileaf collimator were reported by many
scientists and physicists, and its usefulness for clinical appli-
cation has already been proven. The ETX system of the
Novalis system is a kilovoltage X-ray-based two-dimen-
localization device dedicated to stereotactic radiosurgery
and stereotactic radiotherapy. Two X-ray tubes and two
flat-panel detectors are at a fixed position in the treatment
room, suspended from the clinical LINAC. The ETX system
is one of the commercially available patient positioning
devices produced by BrainLAB. The ETX system combines
based system. Two IR cameras are fixed to the ceiling. Two
kilovoltage X-ray beams are projected from the two X-ray
tubes in oblique directions for the purpose of verifying
patient localization. The procedure for patient setup with
the ETX system consists of three steps: in step one, the initial
patient setup according to the IR body markers is performed.
Reprint requests to: Naoki Hayashi, M.Sc., School of Medical
Science, Fujita Health University, 1-98, Dengakugakubo, Kutsu-
kake, Toyoake, Aichi, Japan 470-1192. Tel: +81-562939424; Fax:
+81-562934595; E-mail: firstname.lastname@example.org
This study was presented in part at the 5th AnnualJapan-US Can-
cer Therapy Symposium in Sendai, Japan. September 7-9, 2007.
Conflict of interest: none.
Acknowledgment—The authors thank Mr. Kengo Kojima and
Mr. Thomas Jelinek for technical support.
Received Oct 28, 2008, and in revised form Feb 11, 2009.
Accepted for publication Feb 16, 2009.
Int. J. Radiation Oncology Biol. Phys., Vol. 75, No. 2, pp. 549–557, 2009
Copyright ? 2009 Elsevier Inc.
Printed in the USA. All rights reserved
0360-3016/09/$–see front matter
Normally, the IR body markers are placed asymmetrically on
the patient’s body surface. The placement of each IR marker
is detected by two IR cameras, and then the IR-based patient
localization is completed by the automatic couch movement.
In step two, two X-ray images are taken and compared with
a digitally reconstructed radiograph (DRR) by the registra-
tion software for the purpose of calculating the setup error.
In step three, X-ray image-based patient localization is com-
pleted by the automatic couch movement.
Clinical experiences with the Novalis system were
reported by several investigators (9, 13–16). Many proce-
dures of these studies were done with a previous type of
ETX system(Qualisys type). The present upgraded ETX sys-
tem is version 5, and the geometrical system has been
changed to a Polaris type, which is smaller and simpler
than the previous type (Fig. 1). The source-to-isocenter dis-
tance and the source-to-detector distance of the Polaris type
ETX are 2.2 m and 3.5 m, respectively. The purpose of this
study was to evaluate the mechanical accuracy of (17–18)
anew ETXsystem andits spatial uncertaintyintheradiother-
apy workload with the Novalis system.
METHODS AND MATERIALS
As a fundamental study of geometrical accuracy with the Novalis
system, three factors of the patient setup with ETX system were
Location dependency of the IR system
Six IR-reflective spherical markers, each with a diameter of 15
mm, were placed asymmetrically on the graph sheet, fixed on the
exact couch top (Fig. 2). A front pointer was put on the gantry
The gantry angle, the couch angle, and the collimator angle were
each set to 0?. A reference position was defined on the origin of
the graph sheet. After the intentional offsets of the couch shift/angle
were set to the exact couch angle, the actual locations were com-
pared to the ETX detective locations under each condition. The
actual location was measured on the enlarged image of the graph
sheet with a highly precise digital camera. The intentional move-
ments of couch shift were 0, 50, and 100 mm in longitudinal and lat-
eral directions. The intentional couch angles were tried at 0?, 45?,
Accuracy of the patient verification software
After five IR-reflective spherical markers were placed asymmet-
izontally on the CT table (Fig. 3). Simulation CT scanning was
performed under suitable conditions. The slice thickness, the tube
voltage, and the field of view of CT were 1.25 mm, 120 kV, and
500 mm, respectively. A CT-based treatment plan was designed
by Brainscan, and then the reference point was placed at the center
of the lumbar phantom. After the coordinate data were transferred to
the ETX system, the lumbar phantom was set up at the isocenter by
the ETX system and kept at its position. Correction and verification
procedures were repeated many times under different X-ray image
conditions. After automatic verification was repeated 50 times via
Fig. 1. Comparison of two types of ETX system.
550 I. J. Radiation Oncology d Biology d Physics Volume 75, Number 2, 2009
one X-ray image, the statistical analysis of deviation was done. The
X-ray images with the same mA conditions were taken at tube volt-
ages of 60, 80, 100, and 120 kV. The reference couch angle was 0?.
The Winston-Lutz test has been used to estimate the geometrical
accuracy of the mechanical isocenter since the late 1980s (19).
However, it is not adequate to define geometrical accuracy when
patient localization is done by the ETX system. Sometimes, there is
deviation between the mechanical isocenter and the ETX isocenter
because the ETX system exists independently of a Novalis linear
accelerator. The range of deviation among the coordinates depends
on the calibration procedure used for the ETX system (Fig. 4).
Therefore, calibration accuracy based on two types of X-ray calibra-
tion phantoms were evaluated as a comparative study. The size and
implant markerplacement ofeach phantomwerebasicallythesame.
A hollow X-ray calibration phantom (Phantom A) has a hole at the
center, and a solid X-ray calibration phantom (Phantom B) does not
have a hole (Fig. 5).
Routine Winston-Lutz tests were performed with a Winston-Lutz
kit (BrainLAB) as daily QA in our hospital. Routine isocenter ver-
ification and X-ray calibration were performed every day and two
times per week, respectively. In addition, a WL-module test, a spe-
cial program used to check the offset distance between the mechan-
ical isocenter and the ETX isocenter, was applied, calculation of
by a pinhole camera model algorithm by the following formula:
u ¼p11x þ p12y þ p13z þ p14
p31x þ p32y þ p33z þ p34;
Where x, y, and z are three-dimensional coordinates of a point in
the imaging object, u and v are two-dimensional coordinates of its
projection on the kV X-ray images, and pijare unknown project
parameters determined by a dedicated phantom.
For this study, the WL-module test was done constantly during
every Winston-Lutz test (Fig. 6). Phantom A and Phantom B were
used for X-ray calibration for the initial 3 weeks and for the next
4 weeks, respectively. The Winston-Lutz test depended on the
mechanical isocenter, which was defined as the cross-line point of
the in-room laser alignment. The deviation between the mechanical
isocenter and the ETX isocenter based on vertical, longitudinal, and
lateral directions was evaluated by the WL-module test (Fig. 7).
Total deviation of coordinates between the mechanical isocenter
and the ETX isocenter was defined by the root sum of the square
of each direction. That is,
where, DTis the total deviation between the mechanical isocenter
and the ETX isocenter; Dxis the observed value in the lateral direc-
tion; Dyis the observed value in the longitudinal direction; and Dzis
the observed value in the vertical direction. The geometrical accu-
racy with each phantom was evaluated by the retrospective statisti-
cal analyses of these data.
v ¼p21x þ p22y þ p23z þ p24
p31x þ p32y þ p33z þ p34
Fig. 2. Schematic image of the first examination.
Fig. 3. Schematic image of the second examination. Basically, the size and the implant marker placement of the phantom
are the same. Phantom A has a hole at the center.
Spatial uncertainties in the radiotherapy process d N. HAYASHI et al.551
Location dependency in the IR system
Results of the examination for location dependency in
the IR system are shown in Table 1 and 2. As the results
of examination with the intentional translation shift are
shown in Table 1, the location dependency of the IR sys-
tem with translation shift was evaluated. All actual loca-
tions after the intentional movement were confirmed by
the enlarged image of the graph sheet with the digital cam-
era. On the other hand, the detected location of the ETX
system, which meant the IR coordinate of the ETX system,
was confirmed by the IR camera. The deviation of each co-
ordinate was calculated by subtraction from the IR coordi-
nate to the actual coordinate. The deviation between the IR
coordinate and the actual coordinate was observed in each
(lateral/longitudinal) direction. The deviation in the longitu-
dinal direction was larger than in the lateral direction; espe-
cially the deviation at the +100 mm shift point in the
longitudinal direction was 1.08 mm (1%) as the maximum
value. On the other hand, the misdetection of the table an-
gle in each intentional movement ranged within ?0.1?.
As to the results of examination with the intentional couch
system with the couch angle movement was evaluated. All
intentional angles, which defined actual angles, were con-
tor with the digital camera. On the other hand, the detected
locations of the ETX system, which meant the IR-based
Fig. 4. Calibration procedure of the ETX system.
Fig. 5. Two types of X-ray calibration phantoms.
552I. J. Radiation Oncology d Biology d Physics Volume 75, Number 2, 2009
angles, were confirmed by the IR camera. The deviation of
each angle was calculated by subtraction from the
IR-detected angle to the actual angle. The deviation of the
front pointer meant the error from rotation center.
An offset of the front pointer was less than 0.3 mm, and
errorsof the detected position of the ETX systemwere within
0.4 mm on the translation shift. The detection error of the
couch angle was less than 0.7?, and the maximum value
was detected at 45?of the couch angle.
Accuracy of the patient verification software
Results of an examination for the purpose of clarifying the
calculation accuracy of the patient verification software are
shown inFig.8.The purposeofthis examinationwas toeval-
in the patient registration as determined by the image fusion.
The image fusion was tried 50 times per each setup verifica-
tion. Each calculated value was analyzed statistically. The
graduation and the section of bar graph mean the average
and the standard deviation of 50 trials, respectively. From
the results, the values of 100 kV are smaller and more stable
than under other conditions. On the other hand, the values of
60 kV and 80 kV are larger and more variable than under the
couch angle settings.
the mechanical isocenter
Results of the WL-module test during the Winston-Lutz
test are shown in Table 3. By using the WL-module test, an
offset distance between the center of a microsphere and the
center of the ETX origin was evaluated. This value means
a deviation of the coordinate between the ETX origin and
the mechanical isocenter. The deviation values calibrated
by Phantom A and Phantom B were 0.455 ? 0.088 mm
and 0.313 ? 0.024 mm, respectively. These values mean
a sphere of 95% confidence level with each calibration phan-
tom. Sample numbers of Phantom A and Phantom B were 14
the ETX system with Phantom B was better than with Phan-
tom A, with an advantage of approximately 0.1 mm.
As to the location dependency in the IR system, results of
the first examination are divided into two categories: (1)
those in which the error of detection is increased in longitu-
dinal translation shift, and (2) those in which the error of de-
tection exists in couch rotation. The stereoscopic angle of the
Polaris type on the isocenter is sharper than the previous sys-
tem because the distance of each IR camera became shorter.
As a result, the detection accuracy of the Polaris type in lon-
gitudinal direction and/or couchrotation was worse thanwith
the previous ETX system.
As to the accuracy of the patient verification software,
results of the second examination are divided into two cate-
gories: (1) those in which calculation accuracy is precise
when the tube voltage is set to 100 kV,and (2) those inwhich
45?and 90?. The verification error is thought to be improved
bony structures or implant markers. In the couch rotation,
however, there were some uncertainties such as marker
placement and mechanical error, etc.
The preliminary study of the previous type of ETX system
was reported by Verellen et al. in 2003 (14). That study used
the previous ETX type and reported that results of preclinical
verification of the system and clinical validation were limited
to the DRR fusion approach based on bony structures. That
ing system, DRR fusion tests, and an overall deviation within
the ETX system. From their data, the authors found that the
average deviations between the IR tracking measurements
of theisocentricposition andtheactual position of thehidden
target with respect to the treatment isocenter were ?0.24 ?
0.33 mm, 0.45 ? 0.55 mm, and ?0.49 ? 0.59 mm in the ver-
tical, longitudinal, and lateral directions, respectively. Their
experiment featured an intrinsic uncertainty from the
Fig. 6. Schematic image of the third examination.
Fig. 7. WL module test program
Spatial uncertainties in the radiotherapy process d N. HAYASHI et al.553
definition of the isocenter in the three-dimensional CT data
of a 2-mm slice thickness and spacing in sequential scanning
mode, resulting in a voxel size of 0.73 ? 0.73 ? 2.00 mm3.
That study concluded that the ETX system was designed to
support automated patient positioning based on anatomical
data such as bony structures and/or implanted radiopaque
markers by the automated registration of X-ray images and
DRRs. Anthropomorphic phantom measurements showed
that submillimeter three-dimensional treatment setup accu-
racy could be achieved within an acceptable time frame.
Our data were acquired under corresponding conditions sim-
ilar to that study. The average deviations obtained from our
data at the second examination, with 0?of couch angle,
were 0.18 ? 0.14 mm, 0.25 ? 0.08 mm, and 0.26 ? 0.25
mm in the vertical, longitudinal, and lateral directions,
respectively. Comparing our results to those of the study by
Verellen et al. (14), the verification accuracy with the Polaris
type ETX system compared to the new version of the regis-
tration software was better than the previous type. However,
the detection uncertainty of the IR marker in the longitudinal
direction with the Polaris type ETX system was larger than
that of the previous type. Therefore, the detection error of
should be decreased for precise patient localization with the
Liu et al. (16) reported optimal marker placement with
a photogrammetry patient positioning system. An actual
target registration error and fiducial registration error in IR-
based patient positioning with the previous type of ETX
system were evaluated with the anthropomorphic phantom.
That study concluded that a reduction of approximately
50% in target registration error had been achieved by using
theoptimal configurationcomparedto the random
configuration. The authors stated that these data demon-
strated that the optimization of a fiducial configuration could
result in improved tumor targeting ability. We carried out
each examination described by that study’s recommenda-
tions, each marker that was put on phantom asymmetrically
for the reduction of registration error depended on IR marker
The ETX system consists of an IR system and two floor-
fixed kilovoltage X-ray tubes on the LINAC side that project
obliquely from lateral to medial, from the floor to the ceiling,
and from the LINAC side to the couch side, as the vertical,
longitudinal, and lateral directions, respectively, onto two
corresponding flat-panel detectors mounted on the ceiling.
Therefore, the Novalis system has three coordinate scales,
the LINAC-based mechanical coordinate scale, the IR-based
coordinate scale, and the X-ray image-based coordinate
scale. It is very important for precise radiotherapy that these
scales was completed by the ETX calibration procedure. In
general, the LINAC-based coordinate scale and IR-based
coordinate scale are coincident with the calibration phantom
at the isocenter. Next, the IR-based coordinate scale and
X-ray image-based coordinate scale are aligned by the
X-ray calibration tool as the weekly routine QA. We suspect
that this procedure has some uncertainty because two differ-
entphantomsare usedinthecalibration procedure. Inthecal-
ibration step between ‘‘LINAC-based scale to IR-based
scale’’ and ‘‘IR-based scale to X-ray image-based scale,’’
the phantom is replaced manually from the calibration phan-
coincidence between the LINAC-based scale and the X-ray
image-based scale is necessary to reduce geometrical uncer-
tainty because patient localization is finally verified by the
fusion of X-ray images and DRRs. A WL module test during
the Winston-Lutz test is a simple and feasible method for
checking the coincidence at the isocenter as routine QA. In
addition, results of a third examination suggest that the solid
X-ray phantom should be used to calibrate the ETX system
for the purpose of reducing the deviation between the
mechanical isocenter and the ETX isocenter.
Radiation therapy committee task group 24 of the Ameri-
can Association of Physicists in Medicine (AAPM) recom-
mended that the tolerance of the spatial uncertainty
resultingfrommachineinaccuracyandpatient motion should
be within 10 mm in conventional radiotherapy (18). In
Table 1. Results of translation shift movement*
Intentional couch shift0-100 -50+50+100-100 -50+50+100
couch Angle (deg)
Abbreviations: Vert = vertical direction; Lng = longitudinal direction; Lat = lateral direction.
* The deviation between IR detected location and the actual location.
Table 2. Results of couch rotation*
Intentional couch angle0 4590 (deg) 270 315
couch Angle (deg)
tion; Lat = lateral direction.
* The deviation between IR detected location and the actual loca-
554I. J. Radiation Oncology d Biology d PhysicsVolume 75, Number 2, 2009
addition, spatial uncertainty is divided into two factors: (1)
displacement of several fields relative to a target volume at
anominaldistance related tothe isocentric accuracy,thelight
field agreement, the jaw alignment, and the focal spot align-
ment, and (2) error in setup and position of target volume due
to patient or organ motion, including setup error and breath-
ing organ motion other than breathing. Figure 9 shows the
tolerance model, which is spotlighted to the ETX system,
of spatial uncertainties with the Novalis system in our hospi-
tal. This model is divided into the same two factors as the
AAPM model: (1) geometrical uncertainty with ETX-guided
patient setup, and (2) error in setup position of target volume
patient organ motion. To satisfy the tolerance of the total spa-
tial uncertainty within 5 mm, two factors of tolerance have to
be within 2.5 mm and 4.0 mm, respectively. Geometrical
ized by four factors: uncertainty due to coordinate deviation
of ETX, uncertainty due to the simulation CT, uncertainty
due to the registration error by the verification system, and
uncertainty due to marker detection. Yan et al. (15) reported
that slice thickness of the simulation CT had a significant
effect on the positioning accuracy of ETX-guided patient
localization (15). The authors investigated positioning errors
ofthe plannedisocenter, using fivedifferent sets ofCT scans.
They concluded that positioning errors increased signifi-
cantly when the slice thickness of the simulation CT was
larger. Reference CT image data are usually used for dose
calculation, marker detection, and construction of DRRs.
Therefore, it is important to keep the image quality high
and the effective slice thickness thin. The image slice thick-
ness test and the positioning accuracy of nonhelical scanning
with our simulation CT were better than with helical scan-
ning. In our hospital, 1.25-mm slice thickness of nonhelical
simulation CT is basically used for the reference CT. In com-
missioning data, the localization uncertainty due to the sim-
ulation CT is less than 0.2 mm. In addition, this value
agrees with the tendency reported by the study by Yan
et al. (15). Based on this value and the present paper’s results
(17, 19–20), we can summarize the total geometrical uncer-
tainty with the Novalis system with the following formula:
¼ 1:36 ? 0:32mm;
where, Uncerttotal is total geometrical uncertainty with
Novalis system; EETXis the uncertainty due to the coordinate
deviation of ETX; ECTis the uncertainty due to the simula-
tion CT; Everifyis the uncertainty due to the registration error
marker detection; and Eotheris the uncertainty due to other
factors (e.g., head structure).
Patientlocalization withthe Novalis systemhasbeen com-
pleted with image-guided technology, and the setup error
(from landmark structures) is able to hold to within 1 mm.
Therefore, the systematic component of spatial uncertainty
is decreased if the internal target volume is defined enough
to cover the patient’s internal motion. We propose two very
Fig. 8. Results of the second examination. Vert = vertical direction; Lng = longitudinal direction; Lat = lateral direction.
Table 3. Results of the third examination
of the X-ray phantom The total deviation
Spatial uncertainties in the radiotherapy process d N. HAYASHI et al. 555
important points for consideration for using the Novalis sys-
tem: (1) to reduce the geometrical uncertainty with ETX-
well as possible at the time of defining the internal tumor vol-
phase CT images at our hospital, for the purpose of including
all internal motions of the target. An evidence-based setup
margin, which means the leeway to enlarge from ITV to
PTV, should be decided by daily QA/QC and the statistical
analysis of the clinical data in the routine workload. The for-
mulas for defining the setup margin have been reported by
several investigators (7, 22, 23). The favorable condition
for these formulas is basically to reduce uncertainty in the
radiotherapy process. If uncertainties in the image-guided
technology are not ensured, confidence in the calculated
values of the setup margin will be lost. From our examina-
tions, the setup margin with the Novalis system will be
able to be reduced compared to that of the conventional sys-
tem through two significant steps of QA/QC for the geomet-
ricalaccuracy andtheuniteddefinitionmethodoftheITV are
necessary and sufficient conditions.
The spatial uncertainty of the Novalis system was evalu-
ated with several examinations related to the ETX system.
Some factors of spatial uncertainties depended on ETX accu-
racy. In this study, these uncertainties were within 2 mm,
which was secured by the submillimeter three-dimensional
patient setup with the ETX system and daily QA/QC related
to the radiotherapy procedure and mechanical stability.
1. International Commission on Radiation Units and Measure-
ments. Prescribing, recording, and reporting photon beam ther-
apy. ICRU report 50 (Bethesda, MD), 1993.
2. Creutzberg CL, Althof VG, de Hoog M, et al. A quality
control study of the accuracy of patient positioning in irradia-
tion of pelvic fields. Int J Radiat Oncol Biol Phys 1996;34:
3. Bel A, van Herk M, Bartelink H, et al. A verification procedure
to improve patient set-up accuracy using portal images. Radio-
ther Oncol 1993;29:253–260.
4. Bel A, Vos PH, Rodrigus PTR, et al. High precision prostate
cancer irradiation by clinical application of an off-line patient
setup verification procedure, using portal imaging. Int J Radiat
Oncol Biol Phys 1996;35:321–332.
5. Balter JM, Sandler HM, Lam K, et al. Measurement of prostate
movement over the course of routine radiotherapy using im-
planted markers. Int J Radiat Oncol Biol Phys 1995;31:
6. Bel A, van Herk M, Lebesque JV. Target margins for random
geometrical treatment uncertainties in conformal radiotherapy.
Med Phys 1996;23:1537–1547.
7. Stroom JC,HansCJ,de Boer,etal.Inclusionofgeometricalun-
certainties in radiotherapy treatment planning by means of
coverage probability. Int J Radiat Oncol Biol Phys 1999;43:
8. Bova FJ, Buatti JM, Friedman WA, et al. The University of
Florida frameless high-precision stereotactic radiotherapy sys-
tem. Int J Radiat Oncol Biol Phys 1997;38:875–882.
9. Ryken TC, Meeks SL, Pennington EC, et al. Initial clinical
experience with frameless stereotactic radiosurgery: analysis
of accuracy and feasibility. Int J Radiat Oncol Biol Phys
10. Shimizu S, Shirato H, Kimura K, et al. Use of an implanted
marker and real-time tracking of the marker for the positioning
of prostate and bladder cancers. Int J Radiat Oncol Biol Phys
11. Manning MA, Wu Q, Cardinale RM, et al. The effect of setup
uncertainty on normal tissue sparing with IMRT for head-
and-neck cancer. Int J Radiat Oncol Biol Phys 2001;51:
12. Oita M, Ohmori K, Obinata K, et al. Uncertainty in treatment of
head-and-neck tumors by use of intraoral mouth piece and em-
bedded fiducials. Int J Radiat Oncol Biol Phys 2006;64:
13. Scarbrough TJ, Golden NM, Ting JY, et al. Comparison of
ultrasound and implanted seed marker prostate localization
Fig. 9. Tolerance model of spatial uncertainties in Nagoya Kyoritsu Hospital. (1) The setup error is defined as the error
556I. J. Radiation Oncology d Biology d Physics Volume 75, Number 2, 2009
methodsforimage-guidedradiotherapy. IntJRadiatOncolBiol Download full-text
14. Verellen D, Soete G, Linthout N, et al. Quality assurance of
a system for improved target localization and patient set-up
that combines real-time infrared tracking and stereoscopic
X-ray imaging. Radiother Oncol 2003;67:129–141.
15. Yan H, Yin FF, Kim JH. A phantom study on the positioning
accuracy of Novalis Body system. Med Phys 2003;30:
16. Liu H, Yu Y, Schell MC, et al. Optimal marker placement in
photogrammetry patient positioning system. Med Phys 2003;
ary patient motion induced by automated couch movement dur-
ing on-line 6 dimensional repositioning in prostate canter
treatment. Radiother Oncol 2007;83:168–174.
18. Radiation Therapy Committee Task Group 24, with contribu-
tion from Task group 22. Physical aspects of quality assurance
in radiation therapy. American Association of Physicists in
Medicine (AAPM) report 13 (American Institute Physics,
New York), 1994.
19. Radiation Therapy Committee Task Group 42. Stereotactic
radiosurgery. American Association of Physicists in Medicine
(AAPM) report 54 (American Institute Physics, New York),
20. Fukugami S. Preliminary report of ExacTrac system. Katarou-
Kai magazine 2006;11:90–93.
21. Stroom JC, Heijmen BJ. Geometrical uncertainties, radiother-
apy planning margins, and the ICRU-62 report. Radiother
22. van Herk M, Remeijer P, Lebesque JV. Inclusion of geometric
uncertainties in treatment plan evaluation. Int J Radiat Oncol
Biol Phys 2002;52:1407–1422.
23. McKenzie A, van Herk M, Mijnheer B. Margins for geometric
uncertainty around organs at risk in radiotherapy. Radiother
Spatial uncertainties in the radiotherapy process d N. HAYASHI et al. 557