Development of a daily dosimetric control
for radiation therapy using an electronic
portal imaging device (EPID)
Entwicklung einer dosimetrischen Kontrolle der täglichen
Strahlenbehandlung mittels eines MV-Bildgebungssystems (EPID)
Der Medizinischen Fakultät
der Friedrich-Alexander-Universität
Erlangen-Nürnberg
zur
Erlangung des Doktorgrades Dr. rer. biol. hum.
vorgelegt von
Mohammadsaeed Saboori
aus Hamedan (Iran)
Als Dissertation genehmigt von der
Medizinischen Fakultät der Friedrich-Alexander-Universität
Erlangen-Nürnberg
Vorsitzender des Promotionsorgans: Prof. Dr. Dr. h.c. J. Schüttler
Gutachter: Prof. Dr. R. Müller
Gutachter: Prof. Dr. B. Hensel
Gutachter: Prof. Dr. R. Fietkau
Tag der mündlichen Prüfung: 13.Mai 2015
Table&of&content&&
!
Abstract:&.....................................................................................................................&1!
1! Introduction&..........................................................................................................&4!1.1! Facts!about!cancer!........................................................................................................................................!4!1.2! External!Beam!Radiotherapy!(EBRT)!...................................................................................................!6!1.2.1! Intensity+Modulated+Radiation+Therapy+(IMRT)+.........................................................................+8!1.2.2! Biological+point+of+view+of+radiotherapy+.........................................................................................+9!
1.3! Quality!concept!in!radiotherapy!...........................................................................................................!10!1.3.1! In+vivo+dosimetry+in+EBRT+...................................................................................................................+11!1.3.2! Methods+of+IVD+.........................................................................................................................................+12!1.3.3! Electronic+Portal+Imaging+Device+systems+...................................................................................+13!1.3.4! Applications+of+EPID+system+...............................................................................................................+15!
1.4! In!vivo!dosimetry!with!the!EPID!system!...........................................................................................!15!1.5! Purpose!of!study!..........................................................................................................................................!17!
2! Accuracy&of&a&commercial&treatment&planning&system&for&transit&dose&calculations&
at&large&distances&......................................................................................................&19!2.1! INTRODUCTION!...........................................................................................................................................!19!2.2! MATERIALS!AND!METHODS!..................................................................................................................!20!2.2.1! Equipment+..................................................................................................................................................+20!2.2.2! 2D+dose+map+calculation+using+TPS+................................................................................................+20!2.2.3! Reference+transit+dose+measurement+setup+................................................................................+21!2.2.4! Accuracy+of+transit+dose+calculation+by+TPS+...............................................................................+22!
2.3! RESULT!............................................................................................................................................................!24!2.3.1! Field+size+effect+.........................................................................................................................................+24!2.3.2! Phantom+thickness+effect+.....................................................................................................................+25!2.3.3! Air+gap+effect+.............................................................................................................................................+25!2.3.4! IMRT+plan+test+..........................................................................................................................................+26!
2.4! DISCUSSION!...................................................................................................................................................!28!2.5! Conclusion!......................................................................................................................................................!29!
3! 2D&Patient&transit&EPID&dosimetry&for&treatment&fraction&dose&verification&..........&31!3.1! Introduction!...................................................................................................................................................!31!3.2! Materials!and!methods!.............................................................................................................................!34!3.2.1! Equipment+..................................................................................................................................................+34!3.2.2! Transit+dose+distribution+predicted+at+the+EPID+level+.............................................................+35!3.2.3! EPID+dose+measurements+....................................................................................................................+36!3.2.4! Transit+dose+measurement+formula+...............................................................................................+41!3.2.5! ProofRofRprinciple+IMRT+plan+verification+...................................................................................+41!
3.3! Result!................................................................................................................................................................!42!3.3.1! Initial+system+parameters+...................................................................................................................+42!3.3.2! IMRT+plan+verification+..........................................................................................................................+45!
3.4! Discussion!.......................................................................................................................................................!50!3.4.1! Flood+field+correction+............................................................................................................................+51!3.4.2! Open+field+calibration+for+transmission+dosimetry+..................................................................+51!3.4.3! Field+size+correction+...............................................................................................................................+52!3.4.4! Phantom+thickness+correction+...........................................................................................................+52!3.4.5! OverRresponse+correction+....................................................................................................................+53!3.4.6! Transit+measured+data+analysis+.......................................................................................................+53!3.4.7! IMRT+plan+verification+..........................................................................................................................+54!
3.5! Conclusion!......................................................................................................................................................!55!
4! Summary&and&conclusion&.....................................................................................&58!
5! Bibliography&........................................................................................................&61!
6! Acknowledgment&.................................................................................................&67!
7! Curriculum&Vitae&..................................................................................................&70!
Abstract: Electronic Portal Imaging Devices (EPIDs) can be used to perform dose measurements
during radiation therapy treatments if dedicated calibration and correction procedures
are applied. The purpose of this study was to provide a new calibration and correction
model for an amorphous silicon (a-Si) EPID for use in transit dose verification of step-
and-shoot intensity modulated radiation therapy (IMRT). A model was created in a
commercial treatment planning system to calculate the nominal two-dimensional (2D)
dose map of each radiation field at the EPID level. The EPID system was calibrated and
correction factors were determined using a reference set-up, which consisted a patient
phantom and an EPID phantom. The advantage of this method is that for the calibration,
the actual beam spectrum is used to mimic a patient measurement. As proof-of-
principle, the method was tested for the verification of two 7-field IMRT treatment
plans with tumor sites in the head-and-neck and pelvic region. Predicted and measured
EPID responses were successfully compared to the nominal data from treatment
planning using dose difference maps and gamma analyses. Based on our result it can be
concluded that this new method of 2D EPID dosimetry is a potential tool for simple
patient treatment fraction dose verification.
2
Zusammenfassung:
Die an den meisten Beschleunigern der Strahlentherapie angebrachten
Bildempfängersysteme (EPIDs, Electronic Portal Imaging Devices) können bei
entsprechender Kalibrierung auch als Dosimeter eingesetzt werden. Ziel dieser Arbeit
war es, ein System zur Kalibrierung und Korrektur des EPID-Signals (Detektor aus
amorphem Silizium) in Wasserenergiedosis zu entwickeln und bereitzustellen, womit
die transmittierte Dosis während der Strahlenbehandlung zur Verifikation relevanter
Parameter der Therapie herangezogen werden kann. Moderne avancierte
Behandlungstechniken bedienen sich der sogenannten intensitätsmodulierten
Strahlentherapie (IMRT), die sich im technisch einfacheren „step and shoot“-Verfahren
oder in dynamischer Technik durchführen lässt. Es ist bei entsprechender Erweiterung
des „field of view“ der CT-Bildgebung gelungen, die transmittierte Dosis in der Ebene
des EPID im vorhandenen Bestrahlungsplanungssystem vorauszuberechnen. Im
Vergleich der gemessenen 2D-Dosis (EPID dose map) mit der voraus berechneten kann
die Qualität sowohl der integralen Dosis als auch aller einzelnen Subfelder der IMRT
bewertet werden. Mit Hilfe einer praxisrelevanten Anordnung (Patientenphantom und
EPID-Phantom) konnten die wesentlichen Einflussgrößen (Geometrie, Feldgröße,
Streuverhalten und Einfluss auf die spektrale Energieverteilung) erfasst und in die
Kalibrierung zur Wasserenergiedosis übernommen werden. Als „proof-of-principle“
wurde die Methode an zwei typischen Lokalisationen (Kopf-Hals-Region und
Abdomen) getestet. Es wurden dazu typische IMRT-Einstellungen mit jeweils 7
Subfedern herangezogen. Die voraus berechneten und gemessenen 2D „dose maps“
wurden verglichen und die Differenzen (difference maps) dargestellt sowie die
sogenannte Gammaanalyse durchgeführt. Aufgrund der Ergebnisse erweist sich diese
neue Methode der 2D EPID-Dosimetrie als sehr empfindliche und praxistaugliche
Kontrolle und Verifikation der täglichen Patientenbehandlung in der Strahlentherapie.
Mit der Messung der transmittierten Dosis werden sowohl Parameter der Beschleuniger
(MU, Kollimatorwinkel, Lamellenpositionen, Gantrywinkel) als auch
Patientenparameter (Lagerung, Feldposition, Bewegung, Gewichtsverlust,
Organfüllzustand z.B. Blase Rektum) erfaßt.
3
4
1 Introduction
1.1 Facts about cancer
Cancer is a general name for considerable number of complex diseases that could
influence any part of the human body. Cancer is an illness when the abnormal cells
divide without control. Cancer cells have the potential to spread to other parts of body
through the blood and lymph systems (metastasis). According to the WHO report
(2014), there were 14.1 million new cancer cases, 8.2 million cancer deaths and 32.6
million people are living with cancer (within 5 years of diagnosis) in 2012 worldwide.
Figure 1-1 shows the most frequent types of cancer among men and women. It is
estimated that the annual cancer cases will increase from 14 million in 2012 to 22
million in next two decades (WHO, 2014). Cancer treatment needs careful decision of
Figure 1-1.: The most common cancer incidences and mortality in the world in 2012
based on the WHO report. For men, lung cancer and for women, breast cancer were the
most common forms of cancer in the world (WHO, 2014).
5
one or more treatment modalities, such as surgery, chemotherapy and radiotherapy. The
goal of the treatment is either to relieve patients of tumor burden and pain (curative
treatment) or significant life prolongation while improving quality of life of the patients
(palliative treatment).
The choice of cancer treatment depends on many factors such as location, grade and the
stage of tumor. Radiotherapy is one of the cancer treatment modalities, either alone or
in combination with other modalities. Radiotherapy is utilizing ionizing radiation to
control cell growth. There are three main ways of using ionization radiation for patient
treatment (Figure 1-2). They are categorized to: systematic radiation therapy, internal
radiation therapy (Brachytherapy) and external beam radiation therapy (EBRT). In
systematic radiation therapy radioactive drugs, which are called radiopharmaceutical (in
liquid forms), are administrated to the patients to take part in the metabolism in order to
accumulate in the affected organ. Using radioactive iodine for treatment of
hyperthyroidism is a well-established example for that kind of treatment. In
brachytherapy, encapsulated radioactive sources are placed directly into or near the
volume to be treated. The main types of brachytherapy based on the source
implementation are: Intracavitary (sources placed into body cavities close to tumor),
interstitial (sources are implanted within the tumor volume), surface or mould (sources
are places over the tissue to be treated), intraluminal (sources are placed in a lumen)
and intravascular (a single source is placed into arteries) (IAEA, 2005).
6
Figure 1-2.: Different modern radiotherapy modalities.
1.2 External Beam Radiotherapy (EBRT)
The most widely used modality of radiation therapy is EBRT (Figure 1-2). In this
modality the radiation is generated in a machine and guided through different elements
to the patient body. One of the most common types of machine that is widely used in
EBRT is a medical linear particle accelerator (shortened to linac, see Figure 1-3). Linacs
accelerate electrons using power RF microwave fields. For treatment, either the
electrons itself are directed to the patient, e.g., for treatment of superficial tumors or
they are used to generate MV X-rays in a target inside the head of accelerator.
The delivery of the administrated radiation dose to the patient’s body is achieved with
different methods and techniques. By utilizing three dimensional (3D) imaging systems,
a 3D treatment planning system and also 3D dose delivery, it it possible to conform the
dose volume to the prescribed target volume while keeping the dose to healthy organs
bellow their tolerance dose. This technique, which is called 3D conformal radiotherapy
(3D-CRT), is still used in many radiotherapy centers.
Radiotherapy
Brachytherapy
External beam
radiotherapy
Systematic
radiotherapy
EBRT techniques:
- 3D-CRT
- IMRT
- IGRT
- IMAT
- SRS/SRT
Type of implant:
- Intracavitary
- Interstitial
- Surface (mould)
- Intraluminal
- Intravascular
Targeted radiotherapy
7
An important device to achieve such field conformation is the Multi Leaf Collimator
(MLC) which installed as the most down-stream device in the accelerator head. It is
made up of several individual leaves of about 5 mm at isocenter, which can move
individually to conform the field shape to the projection of the target volume. MLCs
were first introduced to market in 1980 to replace conventional field-shielding blocks
(Figure 1-4).
Recently, modern radiotherapy machines are equipped with different imaging devices
that are able to scan the patient anatomy before (or/and during) delivery of a treatment
fraction. This technique of dose delivery is called Image Guided Radiotherapy (IGRT).
It can help to reduce the radiation side effects while ensuring that the relative target
volume and some reference points are in the same geometrical position as in the
reference imaging data set used for treatment parameter optimization (treatment
planning).
By using a 3D localization system (stereotactic apparatus) and multiple narrow beams
there would be possible to have more precise dose delivery system. This type of EBRT
is called Stereotactic Radiation Therapy (SRT). Using this technique used for
Figure 1-2.: A medical linac system (ONCOR) equipped with portal imaging
system (manufactured by SIEMENS).
3
8
intracranial lesions in a single fraction, it is called Stereotactic Radiosurgery (SRS)
(Mayles et al., 2007; Khan, 2010).
1.2.1 Intensity Modulated Radiation Therapy (IMRT)
Intensity Modulated Radiation Therapy (IMRT) is an advanced form of EBRT. IMRT
allows additional optimization of the dose distribution inside the planning target volume
(PTV) as well as in organs at risk (OARs), which are demanding individual restrictions
for dose exposition. In IMRT multiple inhomogeneous field overlaid gives a
homogeneous but much more conformal coverage of the target. A specialized planning
system is indispensable for this kind of optimization considering the potentials of the
individual Linac. IMRT techniques could be categorized into two groups: in the first
method; multiple static MLC shaped fields are utilized. This technique is called Step
and Shoot IMRT. In another IMRT method, dynamic MLC systems are utilized. In this
method leaves are in motion during radiation while on the step and shoot IMRT, leaves
move only when the beam is off. Figure 1-3 shows a conventional and an IMRT plan
created for a patient with oropharynx tumor. Combination of IMRT and arc
radiotherapy, called Intensity Modulated Arc Therapy (IMAT), came to reality with
actual hardware and software development. IMAT allows quicker and more precise
Conventional field
Blocks
Shaped Field at Isocentre Shaped Field at Isocentre
MLC Leaves in Linac Head
Figure 1-3.: Methods of field shaping devices; conventional field shaping
blocks and MLC system (www.medphysfiles.com).
4
5
9
radiation treatment. While in IMRT techniques the degree of freedom in treatment
planning increased during the optimization (including couch, gantry, collimator,
energy,…), it is able to improve dose uniformity inside the target volume and decrease
the delivery time. Immobilization and position verification of patients play a key role in
the promising outcome of IMRT and IMAT treatments (Teoh et al., 2011).
Figure 1-5.: a) Conventional plan for an Oropharynx patient with three open beams, the
yellow arrows represent beam with equal intensities. b) Simplified schematic IMRT
plan with multiple beams, the arrows represent different segments of each beam with
different intensities. The plan is optimized to minimize the dose to parotids and spinal
cord.
1.2.2 Biological point of view of radiotherapy
Radiation biology is not an issue of this thesis, but it is the most important reason for
development of EPID-dosimetry. Ionizing radiation leeds to various alterations in living
mamalian cells. In particular, it causes lesions and alterations at the genetic code
presented in the deoxyribonucleic acid (DNA). That may leed to cell death,
chromosomal alterations or only distinct mutations or activations in the genetic code.
Some of this reactions with the endpoint “cell death” are intended for the tumor cells
inside the CTV (clinical target volume) and are unintended in the surrounding healthy
tissue. The most importand factor for this kind of reactions is the absorbed dose inside
tissue respectively inside singel cells. So treatment planning and application aim at
a) b)
10
realizing proper dose values inside CTV and vice versa dose values in healthy tissue as
small as possible.
Even so that the respond of living cells, tumor cells and healthy tissue, doesn’t diversify
too much, radiation therapy is possibly effective when the optimized dose distributions
are realized in the daily application over treatment periods. This principle demonstrated
in Figure 1-6 by two sigmoid curves, the first one is Tumor Control Probability (TCP)
(curve A) and the second one is the Normal Tissue Complication Probability (NTCP)
(curve B). Any variance of patient’s position and condition (e.g., body weight, pain,
filling conditions of intestine, colon, bladder etc.) may influence and downgrade the
quality of treatment and thus reduce the therapeutic window. Patient movement or only
movement of inner organs especially patient’s breathing have to be checked or
compensated during treatment. Daily checks using EPID dosimetry have the potential
for detecting most of these parameters mentioned above. It’s a valuable control for
getting the right radiation dose at the righ position (IAEA, 2005; Beasley et al., 2005).
Figure 1-6.: The therapeutic window in radiotherapy, the curve A represent the TCP,
curve B the probability of complications (IAEA, 2005).
1.3 Quality concept in radiotherapy
A high level of accuracy is an essential requirement in radiotherapy as a multi-step
complex procedure in order to reach the desired tumor control rates, in combination
with minimized complication rates. The Quality Assurance (QA) program in
radiotherapy diminishes uncertainties and inaccuracies in dosimetry, treatment
Therapeutic window
11
planning, equipment performance, and treatment delivery meanwhile improving
dosimetric and geometric accuracy as well as precision of dose delivery. Uncertainty
should be managed prospectively and dose errors should be kept within acceptable
tolerances. QA programs ensure quality of radiotherapy treatment intrinsically for
patient safety and avoidance of accidental exposure. Consequently, patient safety is
automatically integrated with the QA program. Several QA guidelines have been issued
by a number of organizations such as International Atomic Energy Agency (IAEA),
American Association of Physicist in Medicine (AAPM) and European Society for
Therapeutic Radiology and Oncology (ESTRO) (AAPM, 1994; ESTRO, 1995; IAEA,
1998).
1.3.1 In vivo dosimetry in EBRT
One of the QA methods in EBRT to improve the quality of patient treatment is in vivo
dosimetry (IVD). In vivo dosimetry refers to measuring techniques for recording and
verifying the dose actually received by the patient during treatment. It helps to detect
the major errors and to assess clinically relevant differences between planned and
delivered dose for each individual patient. IVD could help to limit the escalation of a
problem to following treatment session of a particular patient. In addition, it could avoid
systematic errors which could influence other patients. Moreover, if no error is detected,
IVD provides a conformation record of correctly delivered dose within the expected
tolerance. Until now there is no general consensus in radiotherapy organizations for the
cost effectiveness of IVD and its routine implementation is not widespread. While few
errors are detected it has been argued that most treatments are done without any error
and only small number of patient could benefit from rectifying errors. However, recent
series of major accidents in radiotherapy centers would have been ruled out if
appropriate IVD methods had been utilized (IAEA, 2013). The IVD feed back role in
radiotherapy chain is presented in Figure 1-7.
12
Figure 1-7.: The total radiotherapy procedures and role of in vivo dosimetry for the
assessment of out come of treatment.
1.3.2 Methods of IVD
There are different devices and dosimetric methods for patient IVD, which could be
used as real-time and passive dosimeters (methods that need processing after radiation).
Real time detector systems include diode, metal oxide semiconductor field effect
transistor (MOSFETs), and Electronic Portal Imaging Device system (EPID). These
types of noninterventional detectors are able to provide dosimetric information to
evaluate the dose during treatment session. In contrast, the passive detectors such as
Thermoluminescent Dosimeters TLDs, Optically Stimulated Luminescent Dosimeters
(OSLDs) and also Film (radiographic and radiochromic) are not able to provide
immediate measurement. The passive detectors require some finite time from minutes to
Localization -
CT simulation
Delineation of
volumes
Treatment
planning
Verification and
simulation
Patient set-up
Treatment
delivery
Follow-up
Record and
verify
In vivo
dosimetry
Machine QC
Beam data
collection
13
hours for read-out. Table 1-1 summarizes the physical and dosimetric characteristics of
some of available detector system in EBRT (ESTRO, 2006; Mijnheer et al., 2013).
Table 1-1.: Summary of characteristics of various detectors for in vivo dosimetry in
EBRT.
Detector
Parameter
Diode MOSFET TLD OSLD Film EPID
Cables + + - - - -
Bias voltage + + - - - -
Temperature
dependence + +/- - - +/- -
Energy
dependence + + + + + +
Angular
dependence + + - -
Dose + + + + + +
Dose rate + + - - - -
Readout delay no no yes yes yes no
Main
advantage
Good
reproducibility,
immediate
reading
Immediate
reading, minor
fading
No cable, few
correction,
reusable
No cable,
reusable
2D dose
distribution,
high
resolution,
permanent
record
2D and 3D
dose
distribution,
permanent
record
Main
disadvantage
Cumbersome
calibrations,
cable
connection
Limited
lifetime, high
cost
Labor
insensitive,
specific
reading
equipment
Short life,
dependence on
accumulated
dose, specific
reading
equipment
Cost, specific
scanning
device
Cost, limited
availability of
commercial
software
Note: + means no dependence on parameter; - means dependence response to the
parameter; +/- means in some type of detector could be dependent
1.3.3 Electronic Portal Imaging Device systems
In radiotherapy, acquisition of images with therapy beam is called portal imaging
(Langmack, 2001). Modern digital imaging devices were developed as tools to acquire
portal images of the treatment area and to reduce patient setup errors. Utilizing port
films as control systems is time consuming and labor intensive and could decrease the
throughput in a busy radiotherapy clinic. Therefore the need for a fast and accurate
14
portal imaging system to intensify verification of radiation therapy arouse the
development of on-line electronic portal imaging devices (EPIDs) (Herman et al.,
2001). An EPID system consists of a two-dimensional flat radiation detector, which is
mounted on a semi-robotic arm fixed to the Linac chassis (see figure 1-1). The collected
information from the detector system is transferred to the main computer system of the
linac for processing and control. The detected portal images are displayed to the
therapist typically in direct comparison to digitally reconstructed radiographs (DRR)
that are calculated based on the planning computed tomography (CT) scan representing
the nominal position of the patient. Currently there exist different types of detector
systems which are listed bellow:
- Fluoroscopic detectors;
- Ionisation chamber detectors;
- Amorphous silicon detectors (a-Si);
In the fluoroscopic detector system, a scintillator plate converts the high energy X-ray
to light. A camera captures a fraction of this emerging light and transforms it to a video
signal which is digitized and stored by the processing system. The EPID system with
ionisation chamber detectors is working based on a grid of several ionization chambers.
The 2D map of gathered signal will be converted to gray scale images. Currently the
most commonly applied type of EPID is the a-Si EPID. These types of EPIDs are able
to acquire high quality presenting high-resolution images while requiring very small
dose values. The main principle of the a-Si EPID is based on a two-step process. In the
first step, the X-rays are converted into visible light by means of a metal plate and
phosphor screen (typically copper, Cu, and terbium-doped gadolinium oxysulfide, Gd2
O2 S:Tb, respectively). In the second step, the generated photons are absorbed by the
underlying of a-Si photodiodes which create electrical signals (Blake et al., 2013a). A
simplified schematic drawing of the a-Si EPID system in which the MV X-ray converts
to electrical signal is shown in Figure 1-8.
3
15
1.3.4 Applications of EPID system
Since EPID is a fast image acquisition tool, it has different applications in radiotherapy.
It could be utilize for some QC tests at linacs, and also for detecting geometric and
dosimetric errors. It should be mentioned that the geometric accuracy is limited to:
- Uncertainties in a particular patient set-up;
- Uncertainties in the beam set-up;
- Anatomical and physiological changes of patient.
Nowadays many clinics are utilizing EPID systems for verifying patient positioning
during radiotherapy treatment and also measuring patient organ motion during
treatment. In addition, EPID can be used as QA tool to check the functioning of
radiotherapy equipment (e.g. MLCs) and for beam and patient dosimetry (IAEA, 2005).
1.4 In vivo dosimetry with the EPID system
Several groups have developed different methods for checking the planned dose
administrated to the patients. They can be categorized in two groups; pre-treatment
(without patient) verification and in vivo dosimetry (with patient), each approach
having its own benefits and limitations. Pre-treatment verification can detect a large
number of errors in the radiotherapy chain before starting treatment, but if there is any
unexpected change at the time of treatment, it will be missed. It is not easy to translate
the influence of pre-treatment discrepancies on the real dose distribution delivered at the
day of occurrence (McDermott et al., 2007; van Elmpt et al., 2008). Possible sources of
Figure 1-4.: schematic drawing of the EPID system, interaction of X-ray with different layers
generate electric signals.
8
16
errors could be changes in the linac output, incorrect accessory, MLC failure, patient
changes (tumor shrinkage, weight loss, organ motion, anatomical changes in patient
since planning CT) or even changes in patient setup, including obstruction from
immobilization devices (van Elmpt et al., 2008; Chytyk-Praznik et al., 2013). Many of
these errors could be detected if all parameters would be checked during the patient
treatment via patient transmitted dose.
In vivo dosimetry (IVD), as part of a QA program, could detect major errors and
indicate the relevance of the differences between planned and delivered treatments
(Mijnheer et al., 2013). Comprehensive overviews of techniques for IMRT verification
using point detectors, two-dimensional (2D) arrays, films, and also EPIDs, have been
provided by several authors (Low et al., 2011; Mijnheer et al., 2013). The EPID is an
excellent in vivo dosimetry system since it may be used to verify dose in all dimensions
(0, 1, 2, 3 and 4D).
EPIDs can be implemented as a tool for transmission (transit), or non-transmission
(non-transit) dosimetry, i.e. with or without an attenuation object (patient or phantom)
in the beam path. Both methods can be categorized into a forward and back-projection /
dose reconstruction approach. Figure 1-9 shows above mentioned approaches (van
Elmpt et al., 2008). In the forward approach, the verification of fluence or dose is done
at the position of the imager, while in the second approach measurements are made
outside the attenuating medium, and the verification is performed inside the patient or
phantom (Wendling et al., 2006; Wendling et al., 2009).
17
Figure 1-9.: Three different arrangement of EPID dosimetry and verification of dose
distribution, a) non-transmission pre-treatment approach, b) non-transmission during
treatment approach, c) transmission approach for pre-treatment and also during
treatment verification.
1.5 Purpose of study
The purpose of this study was to develop a simple method to evaluate and use EPID
detectors as a dosimetric system distally of the patients treated with step-and-shoot
IMRT. Intense and subtle calibration was necessary for transferring the EPID imaging
information into dose values. By comparing the calibrated and transformed transit dose
images at the EPID level with corresponding dose distributions calculated with a
commercial TPS changes to the nominal conditions can be achieved. For this reason the
calculation matrix of the TPS must be extended to incorporate the complete geometry of
the patient plus the EPID. Based on measurements, the EPID images are calibrated and
transformed into dose ‘portal dose images’. No additional algorithms such as Monte
Carlo or convolution methods were necessary. As proof-of-principle, the feasibility of
this method for 2D transmission dosimetry has been tested on homogenous and
anthropomorphic phantoms. To our knowledge this is the first study using an a-Si EPID
in combination with a commercial TPS to verify transit dose distributions.
EPID
EPID
EPID
Patient Patient or
phantom
18
19
2 Accuracy of a commercial treatment planning system
for transit dose calculations at large distances
2.1 INTRODUCTION A treatment planning systems (TPS) is an essential tool in each radiotherapy
department. Besides their main task, calculating and optimizing a treatment plan, they
have potential for many other applications including the calculation of portal dose
distributions of patients undergoing radiotherapy. Several groups developed in-house
methods for dose prediction at large distances to compare them with measured EPID
data. One of the most common methods is using an independent dose calculation.
Several models have been developed to predict the dose from zero dimension (point
dose) to three dimensional dose distributions (Chytyk and McCurdy, 2009; Fidanzio et
al., 2010; Chytyk-Praznik et al., 2013). Using an independent dose calculation
algorithm instead of a commercial TPS might have several advantages, but on the other
hand needs extra resources to develop and process such a system. There are other
studies, which used existing TPS systems for pre-treatment verification of treatment-
plans (at gantry angle zero). Advanced TPSs are capable of performing are capable of
performing pre-treatment verification using phantoms, as well as of predicting the
transit-dose behind real patients (e.g., in EPID level). This method could be an easy and
fast solution because it does not need extra data collection for a new calculation system.
(Mohammadi et al., 2008; Tyler et al., 2013)
One of the intrinsic problems of a TPS for dose calculation in the extended distances
(e.g. EPID level) is that they are confined to do the calculation within the field of view
(FOV) defined by the CT in use, which usually can’t encompass the geometry of
EPIDs. Reich et al. used a horizontal setup; both phantoms (patient and EPID) were
located on the CT table. They used the TPS to calculate the transmitted dose in the
EPID phantom (Reich et al., 2006). This method is able only to calculate the dose with
gantry and table angle of 90 degree. As substitute to the last method, Mohammadi et al.
extended the acquired CT matrix slices from 512×512 pixels to 1024×1024 pixels.
They set the Hounsfield Units of the extended region in a way that they had density of
air (ρ = 0 g/cm3). In that study, EPID was modeled as virtual phantom (a uniform slab
with water density) and the EPID phantom was in the fixed position and for each beam
the CT images of patient were rotated in the opposite direction of gantry rotation
(Mohammadi et al., 2008).
20
The goal of our study was to check the accuracy of the transit dose calculation by a
commercial TPS in comparison to the ion chamber measurements as gold standard. In
this chapter our method to overcome geometrical limit of the TPS for dose calculation
in the extended distance were explained. Several influential parameters, such as field
size and air gap size, in transmision dose measurement were studied. 14 IMRT fields
were used to test the accuracy of transit dose calculation by the TPS. A calibrated 2-
dimensional ion chamber array (as our reference instrument) was utilized to validate the
calculated transit dose at the EPID level.
2.2 MATERIALS AND METHODS
2.2.1 Equipment
In this study oncor linacs (SIEMENS) with OPTIFOCUS TM 82- leaf MLC systems
using 6 and 15 MV photon beams were used. A calibrated 0.3cm3 air vented ionisation
chamber (IC)(Type 30016, PTW, Freiburg, Germany) and a 2D array detector system
(Matrixx IBA Dosimetry, Schwarzenbruck, Germany) were utilized. The Pinnacle TPS
(V9.4 by Philips Medical Systems, Fitchburg, WI, USA) was used to create treatment
plans and calculate the transit dose distributions. The phantom was scanned in a CT
SOMATOM Sensation open (Siemens Healthcare, Erlangen, Germany).
2.2.2 2D dose map calculation using TPS
To overcome the FOV limit in the Pinnacle planning system, a script has been written to
expand the CT images of patients and copy the treatment plan to the expanded CT
images (Figure 2-1). A second script, which is called planar dose script (PlDo), creates a
CT#expansion
air
EPID%model%
(water)
Figure 2-1.: CT slice expansion with extra surrounding air layer and EPID phantom
21
virtual EPID phantom (as substitute for the real EPID) and calculates the dose maps for
each beam (Figure 2-2). The virtual EPID-phantom is created always at fixed distance
to isocenter and perpendicular to each beam. This virtual phantom with density of water
is positioned in the same position as the real EPID system. The Source to calculation
Plane Distance (SPD) and also the size of this virtual phantom could be defined by user
base on patient treatment condition. The PlDo script creates for each beam a virtual
EPID phantom. Afterwards the EPID dose-map is calculated, the phantom removed and
the procedure repeated for the next beam (Figure 2-3). Analysis of EPID responses to
different parameters is necessary for transit dose measurements but is beyond the scope
of this chapter and will be discussed in following.
Figure 2-3.: Workflow of the two scripts in the TPS to calculate the transit dose maps
for each beam at given angle.
2.2.3 Reference transit dose measurement setup
For studying the influence of different parameters on transit dose measurement, a
reference setup (patient/EPID phantom) was defined by varying distance and/or
dimension. The first phantom, which called patient phantom, is consisting of solid water
slabs (RW3, PTW, Freiburg, Germany) in dimension of 30×30×22cm! (Figure 2-4).
This phantom located at patient table while the isocenter of machine always is located at
Read beam parameters
(gantry angle, isocenter)
Create virtual
phantom for the
beam
CT images expansion (512*512 to
1024*1024)
Calculate 2D
Fluence map for
the beam
Next beam
Remove the virtual
phantom
Figure 2-2.: Representation of virtual phantom for one beam at 288 degree for a pelvic patient
22
the center of this phantom. The second phantom, EPID phantom, is created by PlDo
script at the EPID position. The isocenter distances to measurement plane in the EPID
phantom were chosen variability between 20 and 60 cm. The patient/EPID phantom
configuration was used for all measurements in this study. The ion chamber located in
central axis of the beam in the RW3 EPID phantom in depth of 2 cm. To measure the
transit dose of IMRT fields, a 2D array ionisation chamber system was also positioned
perpendicular to the beam in the position of the EPID phantom.
2.2.4 Accuracy of transit dose calculation by TPS
2.2.4.1 Field size effect
In the current study, checking the accuracy of transit dose calculation in our TPS the
field size has been varied and the dose predicted by the TPS has been compared to ion
chamber measurements. The field size effect was studied in presence of attenuating
media as showed on the patient/EPID phantom setup in Fig. 2b. In this setup the EPID
model was fixed distance of 140 cm(source to measurement plane distance SPD). The
field sizes were ranging from 3×3 to!20×20!"!. The ion chamber measurements were
performed at the center of beam at 2 cm depth inside the RW3 phantom. To find the
best matching depth in the TPS, calculations have been done for fixed field sizes at
different depths 1 to 5 cm for both beam energies.
Beam iso-center = phantom center
Measurement plane
SAD=100cm
Air gap RT table
Patient phantom
EPID phantom
Figure 2-3.: Reference patient/EPID phantom set up 4
media as showed on the patient/EPID phantom setup in Figure 2-4. In this setup the EPID
23
2.2.4.2 Phantom thickness effect
When the patient (phantom) thickness varies, the absorption and scattering pattern and
also the spectrum of the beam changes. Based on detector structure, there might be
increasing or decreasing response to these changes. In the current study, the accuracy of
Collaped Cone algorithm (CC) was tested by changes of phantom thickness. The
patient/EPID phantom (Figure 2-4) was used to study this effect with the field size of
10×10cm!!using 30 MU. The patient phantom thickness ranged from 10 cm to 34 cm
(in the steps of 4 cm). The transit dose calculations have been done by TPS under equal
conditions and were compared with the ion chamber measured values.
2.2.4.3 Air gap effect
The transmitted beam is a combination of primary photons and photons scattered from
the patient (phantom). Modifying the air gap size between patient (phantom) and
detector will change the proportional contribution of the scatter component. Hence the
change of this parameter should be taken into account for transit dose calculation and
measurements. In the current study the defined reference patient/EPID phantom (Figure
2-4) was used. The source to detector distance (SDD) ranged from 130 cm to 160 cm in
the increment of 10 cm.
2.2.4.4 Evaluation of transit dose calculation for IMRT fields
Finally for testing the accuracy of transit dose calculation by the TPS, two IMRT plans
for 6MV and 15MV beams have been created. Each plan consisted of 7 beams and the
field’s areas ranged from 66 to 133 cm2 and the numbers of monitor units ranged from
78.6 to 309.2 MU. The IMRT fields were delivered to the above-mentioned phantom
setup (Figure 2-4). By using PlDo script, the 2D transit dose for each beam was
calculated using the TPS. The 2D array ionization chamber detector system was used to
measure the transit dose of IMRT fields. The detector was placed in the identical
position as the EPID hardware and virtual EPID-phantom. The source to measurement
plane distance was 140 cm (SPD = 140 cm). The internal and external buildup layer for
the detector system was 2cm for both energies. The !-evaluation method was utilized to
compare the calculated and measured dose maps. A dose-difference criterion of 3% and
distance-to-agreement criterion of 3 mm was selected for the evaluation.
24
2.3 RESULT
2.3.1 Field size effect
Figure 2-5 a) and b) show the normalized dose calculation of different field sizes in
depths 1 to 5cm at the virtual phantom in the TPS for 6 and 15MV beams. The
horizontal and vertical axes represent one side (cm) of quadratic field sizes. The results
of dose variation is normalized to a standard field of 10 cm (100 cm2). The calculated
values were compared to the measured dose by the IC at the depth of 2 cm in the patient
phantom. The measurement depth of 1cm is smaller than the !!"# for both beam
energies (1.6 cm for 6 MV and 3 cm for 15 MV), there seems to be more instability in
the calculation of dose at this small depth. On the other measured depths (2 to 5 cm), the
depth of 2cm has the closest accordance to the IC reading.
Figure 2-5.: a) and b) the field size effect for 6 and 15MV beams (calculated output
factor by TPS at different depths in comparison to measured output factor by ion
chamber).
0 5 10 15 200.7
0.8
0.9
1
1.1
1.2
1.3
1.4
field size (cm)
no
rma
lise
d r
esp
on
se
X6MV
IC
d=1cm
d=2cm
d=3cm
d=4cm
d=5cm
0 5 10 15 200.7
0.8
0.9
1
1.1
1.2
1.3
1.4
field size (cm)
no
rma
lise
d r
esp
on
se
X15MV
IC
d=1cm
d=2cm
d=3cm
d=4cm
d=5cm
a) b)
25
2.3.2 Phantom thickness effect
Figure 2-6 shows the phantom thickness variation of calculated and measured EPID-
dose for 6 and 15 MV beams. The horizontal and vertical axes represent the phantom
thickness (in cm) and normalized dose values, respectively. All data were normalized to
the reference phantom thickness of 22cm. As expected, an exponential decrease in the
transmitted dose values was detected for IC measured and TPS calculated values
(Berry et al., 2010; Blake et al., 2014). As shown in Fig. 3.c, the lower energy (6MV)
has greater variation to the changes of phantom thickness than the 15 MV data.
2.3.3 Air gap effect
The IC response by changing the air gap is presented in Figure 2-7. The horizontal axis
represents the air gap and the vertical axis the normalized dose values. All dose
measurements and calculations are normalized to reference phantom setup and field size
(phantom thickness of 22 cm, air gap size of 27 cm and field size of 10×10!cm!).
10 15 20 25 30 350.6
0.8
1
1.2
1.4
1.6
1.8
phantom thickness (cm)
no
rma
lise
d r
esp
on
se
IC X6MV
TPS X6MV
IC X15MV
TPS X15MV
Figure 2-4.: The phantom thickness effect for 6MV and 15MV beams. Measurements
and calculation was done with the reference field size of 10×10 cm2.
2-6
(Berry et al., 2010; Blake et al., 2014). As shown in Figure 2-6, the lower energy (6MV)
26
By Figure 2-7.: the air gap effect for 6 MV and 15 MV beams. The measurements and
calculations were done with the reference field size of 10×10 cm2.
By increasing the air gap less lower-energy photons from the phantom reach the
detector. As shown in fig 3.d. the air gap effect greatly depends on the gap size while it
is less sensitive to the energy of the beam.
2.3.4 IMRT plan test
Figure 2-8. a) and b) show the 2D transit measured and calculated dose map for one of
the tested IMRT fields (field FA7). The Fig.4. c. is the 2D !γ -evaluation map of this
field. In table 1 the detail of tested IMRT field which included the beam energy, number
of MU, field size area and also the number of segments presented. For matching the
calculations and measurements, all beam gantry position set to zero. The result of !γ -
evaluation of measured and calculated dose map of IMRT fields summarized in Table
2-1.
15 20 25 30 35 40 45 500.7
0.8
0.9
1
1.1
1.2
air gap size (cm)
no
rma
lise
d r
esp
on
se
IC X6MV
TPS X6MV
IC X15MV
TPS X15MV
Fig2-7
Fig 2-8
detector. As shown in figure 2-7 the air gap effect greatly depends on the gap size while it
is less sensitive to the energy of the beam.
the tested IMRT fields (field FA7). The Figure 2-8 is the 2D γ - evaluation map of this
27
Figure 2-8.: Comparison of measured and calculated dose map of one IMRT field a)
measured 2D transit dose by Matrixx system, b) calculated transit dose map by the TPS,
c) 2D gamma evaluation result.
a) b)
c)
28
Table 2-1.: Detail of tested IMRT fields and result of gamma evaluation for 14 IMRT
fields.
Field
name
Energy
(MV)
Number
of MU
Field area
(cm2)
Number of
segments
% area
γ>1 γ -mean Std. Dev.
FA1 6 249.9 114 12 2.54 0.22 0.33
FA2 6 173.0 109 8 4.10 0.24 0.36
FA3 6 103.7 69 6 1.33 0.11 0.26
FA4 6 266.7 128 13 3.82 0.26 0.36
FA5 6 114.1 66 7 1.61 0.12 0.27
FA6 6 135.4 109 6 2.55 0.22 0.34
FA7 6 187.3 118 8 2.08 0.23 0.33
FB1 15 309.2 127 17 6.87 0.26 0.41
FB2 15 87.5 68 7 3.11 0.15 0.03
FB3 15 168.1 126 9 6.96 0.28 0.42
FB4 15 153.1 121 9 6.85 0.27 0.42
FB5 15 126.8 133 7 6.69 0.29 0.40
FB6 15 217.0 123 11 7.19 0.26 0.41
FB7 15 78.6 72 5 2.87 0.15 0.32
2.4 DISCUSSION Ideally the dose calculation should match with the measured dose. McCurdy et al.
studied the accuracy of pencil beam algorithm for dose prediction portal imaging
system by comparing with Monte Carlo (MC) simulation (McCurdy and Pistorius,
2000). It was reported that the percentage error was roughly proportional to the field
size in the selected air gaps (0 to 40cm). For their clinical frequent field sizes (smaller
than 20×20!cm!), the maximum deviations between predicted and simulated scatter
pattern was not greater than 0.8%. Dahlgren et al. utilized the collapsed cone (CC)
algorithm to calculate the portal dose and used it as reference for comparison with
measured portal dose images (Dahlgren et al., 2006). Using the EPID with matrix ion
chamber for measuring portal images, on average; 99% of field area passed the gamma
criteria (3%/3mm). Lee et al. studied the dose change by variation of the field size for
the open beam in different depths of water (with out any attenuator in the beam) (Lee et
29
al., 2009). They reported that the measured out put factor in the depth of 5 and 3cm of
water had the same response like the EPID system for 6 and 18MV beams. In the
current study using CC algorithm, the out put factor at different depths of virtual
phantom calculated. The calculated output factor at depth of 2cm has the most similarity
to the measured values by IC for both beam energies.
The impact of varying phantom thickness on transit dose measurement reported by
Blake et al. (Blake et al., 2014). They compared the EPID response to the MC simulated
response. They have shown that by increasing the phantom thickness; the deviation
between EPID measured values and expected values increased. In the current study, it
has been shown that there are slight differences in ionisation chamber measured dose
and calculated values by changes of phantom thickness. The other important issue in
transit dose study is the air gap size. By increasing the air gap, more lower energy
scattered photons don’t reach the detector region (Chytyk-Praznik et al., 2013). The
influence of air gap size on transit dose calculation by the CC algorithm in comparison
to MC simulation was reported by Dahlgren et al (2002). With their setup, for 6MV
beam the deviation was between -3% and 0.3% (at depth of 1.7cm) and for 15MV beam
the deviation was between -2.5% and 0.6% (at depth of 2.9cm). In the current study, the
influence of changing air gap size on the dose response in measured dose with
ionization chamber and collapsed cone algorithm studied. For both energies the air gap
effect is similar.
In our study for the tested IMRT field on the reference setup, the average of γ -value
(γ!"#$) and the mean percent of area that fail the gamma criteria (γ > 1) for the plan
with 6MV beam were 0.20± 0.06 and 2.58± 1.05, respectively. For the IMRT field
with 15MV beam the mean !γ value (γ!"#$) averaged to 0.24± 0.06, and the mean
percent of area that fail the gamma criteria were 5.79± 1.92 .
2.5 Conclusion It could be concluded that the commercial TPS system that was used in this study, is
able to correctly calculate the transit dose in the extended distance. On the other hand
this method could predict the 2D transit dose at EPID level and could be used as
reference to compare with 2D transit dose measured with the EPID system.
30
31
3 2D Patient transit EPID dosimetry for treatment
fraction dose verification
3.1 Introduction The search for better QA and EPID dosimetry methods is ongoing, there are no good,
sufficient commercially available options after more than 20 years of published ideas,
and a new, simple, robust technique is still needed. To improve accurate dose delivery
methods, many modern and complicated radiotherapy techniques were developed, such
as Intensity-Modulated Radiation Therapy (IMRT), Intensity-Modulated Arc Therapy
(IMAT) and stereotactic radiotherapy/radiosurgery (SRT/SRS). The more complicated
the technique is, the higher the demands for verification of the treatment. Independent
of treatment technique, there will always will occur random or systematic errors, which
could affect the outcome of radiation therapy. These errors might be due to anatomical
movement, physiological changes of organ position or dimension, patient mis-
positioning, or change in treatment machine parameters. Failures might accrue during
treatment preparation (inaccuracy in the delineation process) or during the treatment
delivery (van Herk, 2004; Margalit et al., 2011). In external beam radiotherapy, Image-
Guided Radiotherapy (IGRT) techniques have helped to manage changes in patient
anatomy and to check the reliability and reproducibility of radiation field geometry with
respect to the patient anatomy. However pre-treatment imaging techniques cannot
provide direct information about the accuracy of the delivered dose.
Different organizations such as IAEA, AAPM and ESTRO have provided several
comprehensive quality control (QC) and quality assurance (QA) recommendations, to
prevent the errors mentioned above in the radiation therapy chain of procedures (which
are rapidly developing). Most of these guidelines were written to support prevention of
failures prior to treatment. However, regarding the check of dose received by patients
during treatment, there are only a few international publications that make
recommendations. There are, however, various countries that have specific legal
requirements for patient QA, such as France and Sweden (AAPM, 2005; ESTRO, 2006;
IAEA, 2013).
32
Besides the advantages of using an EPID as a dosimetry tool, there are several issues
and problems that should be considered. Some of these are related to the EPID structure
and its design, and should be considered in all forms of dosimetry measurement. First of
all, this device was initially designed as an on-line imaging device to improve patient
positioning. The EPID is constructed from high atomic number materials (far from
water equivalent), which results in a response very different from conventional
dosimetry media (e.g. water). As a consequence, EPIDs have an undesirable over-
response to low energy photons in x-ray beams (Teymurazyan and Pang, 2012).
Another problem related to the EPID structure is that it is attached to the linac by a
holding arm, which causes a non-uniform back-scatter pattern to some types of EPIDs
(Rowshanfarzad et al., 2010). The imaging software of the EPID corrects the individual
pixel sensitivity and the wash out of the beam profile shape to make a flat image (flood
field correction). For dosimetry, the beam shape should be reintroduced to the images;
while the sensitivity correction should be maintained (Greer, 2005). Moreover, the a-Si
detector collects signals in a time frame manner that potentially leads to ghosting and
lag effects (McDermott et al., 2006).
The problems mentioned above should be considered for all forms of a-Si EPID
dosimetry, alike transmission and non-transmission. However, when there is an object
(phantom or patient) in the beam path, additional effects occur, which should be
considered for transit dosimetry. The EPID is an energy dependent system and the
transit beam spectrum varies with size and composition of an attenuating object. The air
gap distance also influence the energy spectrum impinging the EPID because of
increasing loss of scattered low energy photons with increasing distance. These effects
might lead to an over-response in some regions, which should be corrected (Grein et al.,
2002; Nijsten et al., 2007a). Another possible object in the beam path is the treatment
couch, which might attenuate the beam differently, depending on the beam angle and
couch position. Several studies recommend modelling the treatment couch to improve
transmission dosimetry results (Ali et al., 2011; Berry et al., 2014).
Several methods have been developed, and some of them are commercially available, to
perform pre-treatment verification of IMRT and IMAT with an EPID using non-
transmission methods, i.e. without any attenuator (phantom) in the beam. In this way it
is possible to verify, for instance, the 2D absolute fluence of IMRT fields, which can be
done at any planned gantry angle (Nicolini et al., 2008; Bailey et al., 2012; Nicolini et
33
al., 2013). Van Esch et al.(2004) developed an algorithm for predicting 2D portal dose
images at the EPID level in air (non-transmission dosimetry). They used a pencil beam
dose calculation algorithm, which was commissioned entirely based on EPID
measurements. Van Zijtveld (2009) also used the predicted dose images without a
patient in the beam, and then by using the patient’s planning CT scan and an “imaginary
equivalent homogeneous phantom”, included corrections of the primary transmission
and the scattered radiation from the patient.
Several groups have developed methods to check the patient dose by means of EPID-
based transmission dosimetry using point, 2D and 3D approaches. Piermattei et al.
(2012) reported the use of EPID in vivo dosimetry, in which they verified the isocenter
dose of patients treated with 3D conformal radiotherapy techniques with 6, 10 and 15
MV beams. Cilla et al. (2014) expanded this method by considering a new component
(fluence inhomogenity index) for IMRT fields. Nijsten et al. (2007b) performed in vivo
verification of the dose at 5 cm depth on the central beam axis. They correlated the dose
measured with an EPID with dose values in a patient by using a back-projection model
with a semi-empirical relationship. Berry et al. (2012) used the same algorithm as van
Esch et al., along with Monte Carlo simulations to simulate the effect of an attenuating
object in the beam (transmission dosimetry). This method is similar to the 'buildup
factor' calculations used in radiation protection. Nijsten et al.(2007a) developed a model
for dose calibration of a-Si EPIDs for 2D transit dosimetry, in which corrections for the
ghosting effect, image lag, and beam profile were implemented. In addition, a
convolution model was included to correct for the field size dependence and scattering
of the EPID response. Persoon et al. (2012) used this method to analyze the inter-
fractional trend of the patient dose throughout the course of treatment. In The
Netherlands Cancer Institute (NKI), an algorithm was developed that uses a
measurement-based “convolution/correction approach” to reconstruct the 3D dose
distribution in a patient from the EPID-based transmitted dose. They compared the
reconstructed EPID dose with the calculated treatment planning system (TPS) 3D dose
distribution (Wendling et al., 2009; Wendling et al., 2012). This technique is
implemented in NKI clinic as an in vivo treatment verification tool for IMRT/IMAT and
non-IMRT plans (Olaciregui-Ruiz et al., 2013).
McNutt et al.(McNutt et al., 1996a; McNutt et al., 1996b) created a model based on the
collapsed cone superposition (CC) algorithm to predict the portal dose images through
34
patient CT scans in the extended phantom that included the EPID. The authors checked
their algorithm for a single 10cm×10cm field transmitted through different phantoms.
Dahlgren et al. (2002) also used an in-house developed version of the CC algorithm, to
predict the portal dose images of two different types of EPID (liquid-filled matrix ion
chamber and camera-based EPID). They tested the validity of this method for several
rectangular and MLC-shaped fields for several thicknesses of homogenous phantom
(Dahlgren et al., 2006). Reich et al.(2006) modelled the portal dose transmitted through
homogenous phantoms in a commercial TPS (Pinnacle, Philips Medical Systems). They
also used the CC algorithm in combination with the extended phantom and presented
the verification of dose distributions behind phantoms of different thicknesses irradiated
with rectangular fields of a 6 MV beam. The authors found that changes in thickness of
~1 mm corresponded to dose changes of 0.6%. These effects could be detected in the
transmitted dose distributions. Mohammadi et al.(2008) used this technique and
presented the result of pre-treatment verification of an IMRT plan for prostate case on
anthropomorphic phantom with 6 MV beam. In both studies a scanning liquid
ionization chamber (SLIC) EPID was used.
The purpose of this chapter is to develop a simple method to evaluate the dose at the
EPID detector behind patients treated with step-and-shoot IMRT using portal dose
images measured with an amorphous silicon (a-Si) EPID. In this method, transit dose
images at the EPID level are compared with corresponding dose distributions calculated
with a commercial TPS using extended CT scans. Based on measurements, the EPID
images are calibrated and corrected without using additional algorithms such as Monte
Carlo and convolution methods. As proof-of-principle, the feasibility of this method for
2D transmission dosimetry has been tested on homogenous and anthropomorphic
phantoms. To our knowledge this is the first study using an a-Si EPID in combination
with a commercial TPS to verify transit dose distributions.
3.2 Materials and methods
3.2.1 Equipment
Two identically built and commissioned Siemens Oncor AvantGard (Siemens Medical
Solutions, Concord, CA, USA) linacs were used in this study. Both were equipped with
the OPTIFOCUS TM
82-leaf MLC system and have 6 and 15 MV photon beam energies.
35
Portal images were acquired using two OPTIVUE 1000ST Electronic Portal Imaging
Devices (EPIDs). The EPID system consists of an XRD Perkin Elmer a-Si detector
(XRD 1640 AL7), which has a frame rate of 3.5 Hz, an active imaging area of
1024×1024 pixels and a pixel resolution of 0.4 mm. To reduce the noise level, the EPID
images were resampled to a resolution of 2mm. Each stored image is an integral of
several frames in which the number of frames depends on the irradiation time (Greer
and Popescu, 2003).
The TPS used for this study was Pinnacle v9.4 (Philips Radiation Oncology Systems,
Fitchburg, WI, USA). The TPS was used to generate IMRT step-and-shoot plans.
Transit dose images at the EPID level were predicted in the TPS using the collapsed
cone (CC) algorithm.
3.2.2 Transit dose distribution predicted at the EPID level
It has been shown that the dose distribution of the CC algorithm results in acceptable
values of the transit dose at the EPID level. For this reason, the EPID was modelled in
the TPS as an “extended phantom”, whereby the virtual EPID is included in the CT data
set.
Calculation of the 2D transmitted dose in the planning system is limited to the field of
view (FOV) of the CT scanner. TPSs are optimized to perform the calculations within
the image matrix given by the CT scan, which usually makes calculation at EPID level
impossible. In our department we use a SIEMENS CT scanner (SOMATOM Sensation
open). To overcome the limited FOV problem in our TPS, in-house developed software
(Matlab v12, [Mathworks, Natick, Massachusetts, U.S.A.]) was used to expand each of
the CT matrix slices from 512×512 pixels to 1024×1024 pixels, and setting the
Hounsfield Units of the additional pixels outside the patient such that ρ = 0 g/cm3, to
simulate air. For patient treatment the source to EPID distance is usually 140 cm in our
clinic.
To estimate the transit dose at the EPID level in the TPS, the EPID was modelled as a
phantom, defined by a volume of water, 30cm×30cm×5cm, ρ=1.00g/cm3
. This EPID
model needed to be perpendicular to the beam direction and always at a fixed distance
from the radiation source. The source-to-surface distance (SSD) was set to 138cm. The
calculated transit dose map !Ctr!,! was calculated at a depth of 2cm in the phantom,
so the measurement plane was located at 140cm from the source. (Figure 3-1). The
36
depth of 2 cm was chosen as a compromise value of the measured build-up depths
necessary for 6 MV (1.6cm) and 15 MV (3.0cm). The parameters and specify a certain
point in this 2D plane.
Figure 3-1.: Reference combination of patient/EPID setup up for (a) measurements and,
(b) TPS prediction model.
3.2.3 EPID dose measurements
3.2.3.1 Flood field correction
Since EPID systems were primarily invented for imaging, there are several corrections
implemented to improve image quality (Antonuk, 2002). Dark-field (DF), bad pixel
map (defective pixels) and flood field (FF) images are the main components for EPID
image correction. A DF correction is an image acquired without radiation to account for
the offset in response due to background influences. A bad pixel map compensates for
defected pixels by replacing their value with an average response of 'good neighbors'.
The gain of detector pixels is not uniform, they have different sensitivities and the FF
correction is necessary to normalise the pixel response over the detector area. This
correction is essential for imaging but it completely removes the beam profile
information (Greer, 2005). For dosimetry, this correction should be removed, while the
correction of the individual pixel sensitivity across the EPID should remain (Warkentin
et al., 2003). In the present study, FF images (gain files) that are normally used to
automatically correct EPID images were collected. For each energy, a 2D model flood
field correction, !!!"(!,!) , was fitted to the collected images. Since the pixel
37
sensitivity map is incorporated in the flood field correction image in our system, a
model is necessary to remove the flood field without sensitivity correction. This model
was used in all subsequent measurements rather than the original gain image, which
also included noise and small local deviations.
3.2.3.2 Open field calibration for transmission dosimetry
Dose characteristics of a-Si EPIDs have been reported in detail by other groups
(McDermott et al., 2004; Winkler et al., 2005; Chen et al., 2006; McDermott et al.,
2006; Nijsten et al., 2007a). For transmission dose measurements, the recorded signal is
a combination of primary and scattered radiation. Scattering of the primary fluence
occurs in the patient and also in the EPID. The scattered fraction of transmitted dose
depends on many factors such as patient (or phantom) thickness, field size of the beam
and also size of the air gap between the patient and detector (Herman et al., 2001).
McDermott et al. (2004) showed that extra build-up material (2.5 mm copper) could
decrease the sensitivity variation of the EPID system, which depends on the air gap size.
But adding an extra build-up layer is not generally permitted by the manufacturers and
may raise warranty issues if the additions fall outside the original EPID specifications.
As previously stated, the goal of this study was to compare measured transmitted dose
to the calculated planar dose map at the EPID level at any planned gantry angle.
Therefore, in this study, we decided not to use extra build-up material.
A reference set-up was defined to calibrate the EPID for transit dose measurement,
whereby the EPID is calibrated using the transmitted beam spectrum (for measurement
and also in the TPS). The reference patient/EPID set-up in this study consisted of two
phantoms (Figure 3-1). The first phantom (patient phantom) was a 30×30×22cm3 RW3
(PTW Freiburg) homogeneous phantom in which the isocenter of the beam was always
located at the center of the phantom (at 11cm depth). The second phantom (EPID
phantom) was a 30×30×5cm3 RW3 phantom, which was located at the EPID position of
the linac, with 138cm SSD.
The linac was calibrated according to the TRS398 protocol (IAEA, 2000). The protocol
requires the calibrated linac to deliver 100 cGy at a depth of dmax when a beam of 100
MUs is delivered with a field size (FS) of 10×10cm2 (!!!"). To calibrate the EPID
system, the patient/EPID phantom setup (Figure 3-1) was used under the same linac
reference conditions (!!!") for each energy, also with 100 MUs. The patient-phantom
was positioned at an SSD of 89 cm. Under these reference conditions, the isocenter of
38
the machine was located at the midplane of the phantom. A central axis calibration
factor (!") was defined at the center of the EPID phantom (!!,!!) as:
!" =
!ion !!!,!!!", !!, !!
!" !!!,!!!", !!, !!
(1)
With !ion(!!!,!"!", !!,!!) being the transmitted dose behind a 22cm thick phantom,
measured using an ionization chamber in the EPID phantom at a depth of 2cm.
!"(!!!,!"!", !!,!!) is the average central pixel value (in a 1×1cm2 region of interest).
According to some reports, the a-Si EPID response to the integral dose is linear up to
250 MU, which is sufficient for the scope of treatment plans, therefore a single
universal calibration factor was obtained to convert EPID pixel values (!") to dose
(Greer and Popescu, 2003). The CF was calculated for both 6 and 15 MV beam
energies.
3.2.3.3 Field size correction factor
In transit dose measurements, there are several scatter components that affect the EPID
reading. They include head scatter, phantom scatter and also scatter within the structure
of EPID itself. Furthermore, owing to their structure, EPID systems respond differently
to varying scatter conditions, compared to other detector devices such as ionization
chambers (Lee et al., 2009). In currecnt study, a TPS was used to calculate the transit
dose images using an EPID model. The response of the EPID as a function of field size
in the central region (12×12 pixels) was investigated and compared to the calculated
dose at depth of 2 cm of the EPID model in the TPS. The reason for this correction is
that the TPS is designed to calculate the dose-to-water based on ionization chamber
measurements. It is therefore necessary to convert the EPID response to dose-to-water
when comparing its response to TPS dose calculations. The field size correction factor
(!"cor) was defined as the ratio of calculated dose values !"(!!!,!"! , !,!) over the
measured values from the EPID system !"(!!!,!"! , !!,!!), normalized to that of the
reference field (FS=10×10cm2).
!"cor !"! =!"(!22,!!!,!0,!0)!
!"(!22,!!!,!0,!0).×
!!ref!22,!!10,!0,!0
!!ref!22,!!10,!0,!0 .
−1
(2)
Where !"! is the side length of square field sizes from 3×3cm2 to 20×20cm
2.
39
For square fields it is easy to obtain and use !"cor but in IMRT, fields consist of
multiple irregular shaped subfields. The purpose of this study was to verify IMRT
fields, and therefore an estimate of the field size was required. For this purpose the
Sterling formula (Sterling et al., 1964) was used to find the equivalent square field
(EQS) of each IMRT field segment. Unless stated otherwise, all field sizes referred to in
this manuscript are defined at the isocenter. A Matlab code based on an edge detection
algorithm was developed to calculate the area (!!EPID) and perimeter (!!
EPID) of each
segment based on EPID images at the EPID level. The area/perimeter ratio was then
scaled to determine the ESQ at the isocenter level.
ESQisocenter = 0.714×
!×!!
EPID
!!EPID
(3)
The factor 0.714 rescales the ESQ from EPID level (140cm) to isocenter level (100cm).
3.2.3.4 Phantom thickness correction
Changing patient (phantom) thickness leads to changes in the absorption and scatter
pattern of the beam inside the phantom, which results in various beam hardening effects
at the exit side of the attenuating medium. This effect could increase or decrease the
EPID response with respect to the TPS-calculated dose values (Pecharroman-Gallego et
al., 2011; Jung et al., 2012). The reference patient / EPID phantom set-up (Figure 3-1)
was used to investigate this effect. In all measurements, the imager was located at a
fixed distance from the beam source (140 cm). The center of the phantom was always
located on the central axis of beam, and at the isocenter. The phantom thickness
correction factor (!!"# ) was defined as the ratio of the calculated dose value
!"(!! ,!!!", !!,!!) for different phantom thicknesses !! and the corresponding values
determined from EPID images; !"(!! ,!!!", !!,!!). The ratio was normalized to the
corresponding reference values (!" = 10×10cm2, phantom thickness of 22cm).
!!"# !! =!"(!!,!!10,!0,!0)!
!"(!!,!!10,!0,!0).×
!!ref!22,!!10,!0,!0
!!ref!22,!!10,!0,!0 .
−1
(4)
where the phantom thickness (!!) range was 10 cm (for head and neck patients) to 34
cm (for pelvic patients), in steps of 4cm. For IMRT fields, a single thickness correction
is performed per segment, where the thickness is measured along the field axis passing
through the CT scan of the patient.
40
3.2.3.5 Over-response correction
It is well reported that the a-Si EPID detectors have an over-response to lower energy
photons due to their high atomic number structures (McCurdy et al., 2001; Kirkby and
Sloboda, 2005; Blake et al., 2013b). This effect causes deviations in dose measurements
within non-water equivalent media (EPID) and in the calculated dose of the EPID plane
in the TPS. Differences exist inside and also outside the primary beam region, but in the
outer primary beam region there is a higher proportion of low energy photons, which
lead to higher over-response of the EPID system. Several authors, including Nijsten et
al. (2007a) and Greer (2005) investigated this effect. Each IMRT field consists of
several segments, and thus the summation of transmitted dose maps of all segments
could be largely affected by overestimated dose values in the out-of-field regions.
To overcome this problem, an empirical Over-Response Correction (ORC) filter was
defined. Firstly, the filter normalizes each pixel response !!!(!,!) of each segment to
the maximum reading of the entire detector area:
Figure 3-2.: a) Comparison of the flood field corrected EPID image and the TPS profile
for a 10×10 cm2 field for a 6MV beam at the EPID level using the patient/EPID
phantom set-up. Points M and N are used for the over-response correction. b) Linear
over-response correction. The horizontal axis is normalized EPID reading (nPVk !,! )
and the vertical axis is the ORC!.
nPVk !,! =
!!!(!,!)!
max !!! !,! |!∀!!,!
(5)
In the second step, a linear correction was developed based on experimental
observations. Two points (M and N, see Figure 3-2. a)) were determined in both
calculated and measured profiles. Point M, defined as the starting point of this
N
M
41
correction, quantifies the position (in 2D) in which the normalized nPV of measured
and calculated profile is 35%. Point N is defined as the point in which the EPID
response is 50% higher than the TPS value. As shown in Figure 3-2.b), the line that
passes through these points when irradiating a symmetric 10x10 cm2 field is defined as
the linear over-response correction (ORC!) of the EPID:
ORC! !,! =
1 for$nPV! !,! ≥ 0.35
!0×nPV! !,! + !0 for$nPV! !,! < 0.35
(6)
In this formalism bo and co are the linear parameters of the line intersecting N and M,
determined for a symmetric 10×10cm2 field.
3.2.4 Transit dose measurement formula
Finally, by considering all factors mentioned above, the measured transit dose,
!k!"(!,!)for each segment ! is the product !"!(!,!) of the relevant segment with the
universal calibration factor (CF), flood field correction (FFC(x, y)), field size correction
of that segment (!!cor !"! ), ), thickness correction factor for the relevant beam angle
!!"# !! , and over-response correction of the segment ORC! !,! which calculated in
the matlab software.
!k!"!,! = !"!(!,!)×!"×FFC(x,y)×!!cor !"! ×!cor(!!)×ORC! !,! (7)
The total measured transit dose for each field, !tot, is the sum of the transmitted dose of
all relevant individually calculated segments, !tr!!,! , where ! is the total number of
segments in the IMRT field:
!tot = !k!"!,!
!
!!!
(8)
3.2.5 Proof-of-principle IMRT plan verification
To determine the accuracy of the method, verification measurements were carried out
for two treatment sites. In this study, as in clinical practice, 6MV beams were used for
head-and-neck and 15MV beams for prostate IMRT plans. All segments of each field
were individually corrected, summed by equation (8) and compared with calculated
transit dose. The plan verification was first performed on a cubic phantom at zero gantry
angle. To verify the method with respect to changes in geometry, homogeneity and
gantry angle, the verification procedure was also performed on an anthropomorphic
phantom under conditions similar to patient treatments, with the planned gantry angles.
42
Calculated and measured dose were analyzed with global γ-evaluation method (dose-
difference criterion: 4%; distance-to-agreement criterion: 3 mm). For each field, gamma
1% (γ1%), the mean gamma (γmean) and the percentage of gamma values above 1 (γ >1)
were calculated (Stock et al., 2005).
3.3 Result
3.3.1 Initial system parameters
3.3.1.1 Flood field correction
Figure 3-3 shows the model of the fitted flood field correction, based on a Gaussian fit
of the machine flood field image to the specific height of measurements (140 cm).
Figure 3-4 shows the normalized cross-line EPID profiles of an open field (10×10 cm2)
on the patient/EPID phantom.
−200 −100 0 100 2000.95
1
1.05
1.1
1.15
1.2
1.25
off axis distance (mm)
Flo
od F
ield
Cor
rect
ion
fact
or (
FF
C)
FFC X6MV
FFC X15MV
Figure 3-1.: Flood Field correction for X6MV and X15MV at 140cm from the beam
source.
3
43
Figure 3-4.: Normalized EPID cross-line profile of both energies for a field 10×10cm
2
on reference patient-EPID phantom set-up. a) 6MV photon, b) 15MV photon. The
dotted lines are the raw EPID profiles and the solid lines the EPID profiles corrected
with FFC.
3.3.1.2 Open field calibration for transmission dosimetry
The TPS calculations and IC measurements at depth of 2cm in the central EPID
phantom were compared and the average difference was ( 2.41± 0.03 )% and
(2.44± 0.11)% for 6 and 15 MV, respectively. The absolute dose calibration for both
energies was obtained under reference conditions (Figure 3-1) for the Patient/EPID
phantom set-up with a reference open field of 10×10 cm2, SSD 140 cm and 100 MU.
For 6 MV, the calibration factor CF was and for 15 MV.
3.3.1.3 Field size correction
For each energy, a second-degree polynomial curve was fitted to the normalized relative
EPID response:
!"cor !"! = !!×(!"!)
2+ !!× !"! + !! (9)
For the 6MV photon energy, the fit parameters were !! = 1.792×10!!!"
!! , and
!!! = −1.905×10!!!"
!! !!! = 1.172 and for 15MV, !!! = 7.705×10!!!"
!! ,
!! = −3.173×10!!!"
!! and !!! = 1.242. For IMRT fields, equation (3) was used to
calculate EQS for each segment. Every segment was then corrected individually based
on its measured field size.
44
Figure 3-5.: a) The variation of the EPID response and the TPS calculated value with
field size, on the central axis beam (at a depth of 2 cm) for both energies. b) Field size
correction factor for both energies. All data are normalized to the corresponding
10×10 cm2 field.
3.3.1.4 Phantom thickness correction
Figure 3-5 illustrates the difference in EPID reading and TPS calculated values for both
photon energies with changing phantom thickness in the patient/EPID phantom on the
central axis. The phantom thickness variation was determined over the interval 10 to
34cm. Figure 5(c) presents the normalized phantom thickness corrections for both
energies.
10 15 20 25 30 350.8
0.85
0.9
0.95
1
1.05
1.1
1.15
Phantom thickness (cm)
Thic
kness c
orr
ection f
acto
r
TCor X6MV
TCor X15MV
10 15 20 25 30 35
0.6
0.8
1
1.2
1.4
1.6
Phantom thickness (cm)
Norm
aliz
ed r
esponse
Thickness variation response X15MV
TPS
EPID
10 15 20 25 30 35
0.6
0.8
1
1.2
1.4
1.6
Phantom thickness (cm)
Norm
aliz
ed r
esponse
Thickness variation response X6MV
TPS
EPID
Figure 3-2.: Variation of TPS values and EPID response with changing phantom thickness
for (a) 6MV and (b) 15MV photon beam. (c) The normalized phantom thickness correction
factor for both energies. All data are normalized to the value of the corresponding reference
22cm phantom thickness.
field size (cm)
field size correction factor
Thickness correction factors
e
x
6
6
3-6 c)
45
For 6MV the relative difference in response was about ±2%, and for 15MV, there was
greater variation of about ±5%. To simplify use, a first degree polynomial was fitted to
each data set:
!cor !! = !!× !! + !! (10)
For 6MV photon energy, the fit parameters were !!! = −1.0269×10!!!"
!! and
!c! = !1.0242 , and for 15MV, !! = −3.7946×10!!!"
!! and !c! = !1.0851 . This
correction was necessary for measurements of beams transmitted through
anthropomorphic phantoms or patients.
3.3.1.5 Over-response correction
The normalized EPID and TPS profiles for a 10×10cm2 field (6MV beam) in the TPS
EPID model at a depth of 2 cm are shown in Figure 3-2. a). The EPID profile (after
flood field correction) in the irradiated area matches the calculated dose profile.
However, in regions outside the field there are significant differences. This effect
caused large deviations in IMRT field measurements, which consist of multiple
segments. Therefore the ORC! was used to correct each segment individually. The
parameters in equation (6) were obtained from the linear fit in Figure 3-2. b), resulting
in !! = 2 and !!= 0.3.
3.3.1.6 Transit dose measurement
Using above correction methods each segment is treated individually using equation (7).
For each IMRT field, all corrected transmitted measured dose distributions are summed
according to equation (8). Figure 3-7 is an example of the calibration and correction
steps for a segment of an IMRT field. This sample segment belongs to field A1 (Figure
3-8 and Table 3-1), a field from the head and neck plan (6MV).
3.3.2 IMRT plan verification
The first plan verified was a head-and-neck IMRT plan with 7 fields and 6MV beams.
The second plan was a pelvic IMRT plan with 7 fields and 15MV beams. The head-and-
neck plan was copied in the TPS and delivered to a slab phantom (thickness of 22cm)
and all beams were measured with the gantry angle set to zero. To test the effect of non-
flat geometry and a small degree of inhomogeneity, the pelvic plan was copied and
delivered to a pelvic phantom (with fields delivered at their planned gantry angle).
46
Figure 3-7.: Calibration and correction steps of transmitted measured dose for each
single IMRT segment. a) EPID central raw cross-line profile. b) Calibrated EPID profile
and calculated dose profile (TPS). c) Correcting flood field (FFC). d) Field size
correction (!!cor ). e) Over-response correction (!"# ). f) Final 2D transmitted
measured dose.
3.3.2.1 Head-and-neck IMRT plan evaluation with a slab phantom
The number of segments, number of MUs and the field area of each field is presented in
Table 3-1for the head-and-neck case. The plan was delivered to the slab phantom
(thickness of 22cm) and the measured dose was compared with the planned dose for
each field. The results of the measurements and the gamma evaluation for one field
(A1) are presented in Figure 3-8.
DO
SE
(cGy
)
47
Figure 3-8.: IMRT field evaluation for a 6MV head-and-neck IMRT field (A1). a) 2D
measured transit dose, b) 2D calculate transit dose. c) Inline and d) cross line profiles
are as indicated in a) and b); e) 2D gamma evaluation map (global gamma).
Distance (mm)
Dose (
cG
y)
TPS
−170 −100 0 100 170
−170
−100
0
100
170 0
2
4
6
8
10
12
14
16
18EPID
Distance (mm)
Dose (
cG
y)
−170 −100 0 100 170
−170
−100
0
100
170 0
2
4
6
8
10
12
14
16
18
−200 −150 −100 −50 0 50 100 150 2000
2
4
6
8
10
12
Distance (mm)
dose(c
Gy)
middle inline profile
TPS
EPID
−200 −150 −100 −50 0 50 100 150 2000
2
4
6
8
10
12
Distance (mm)
dose(c
Gy)
middle cross line profile
TPS
EPID
Gamma evaluation
Y(4
%/3
mm
)
20 40 60 80 100 120 140 160
20
40
60
80
100
120
140
160
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
a)
e)
d) c)
b)
48
Table 3-1.: Results of the EPID dose verification of an IMRT plan (head-and-neck,
6MV beams, 7 fields) from slab phantom measurements.
field name number of
segments
number of
MUs
field area
(cm2)
γ1% γmean % area γ >1
A1 9 88.2 435.7 1.08 0.44 2.49
A2 7 75 336.1 1.00 0.37 1.02
A3 7 96 359.2 1.12 0.41 2.73
A4 13 148 426.7 1.10 0.44 2.32
A5 16 174 378.3 1.00 0.33 1.13
A6 8 96 335.4 1.08 0.44 2.09
A7 9 98 355.0 1.14 0.43 3.21
3.3.2.2 Pelvic IMRT plan evaluation with an anthropomorphic pelvic phantom
Figure 3-9 shows the pelvic IMRT plan with 15MV beams was verified with the
anthropomorphic pelvic phantom. This plan consisted of 7 beams at various gantry
angles. In this test the effects of varying geometry, couch effect and homogeneity were
included in the measured and calculated planar dose distributions at the EPID level. In
Table 3-2 the beam properties and the results of the gamma evaluations are presented. It
should be noted that all beams passed through the treatment couch (some of them
partially and some of them completely).
49
Figure 3-9.: Evaluation of one field of the 15MV Pelvic IMRT plan (field B4 in Table
3-2). The plan was verified using the anthropomorphic pelvic phantom at the planned
gantry angles. a) 2D measured transit dose, b) 2D calculate transit dose. c) Inline and d)
cross line profiles are as indicated in a) and b); e) 2D gamma evaluation map (global
gamma).
Distance (mm)
Dose (
cG
y)
TPS
−170 −100 0 100 170
−170
−100
0
100
170 0
2
4
6
8
10
12
14
16
18EPID
Dose (
cG
y)
−170 −100 0 100 170
−170
−100
0
100
170 0
2
4
6
8
10
12
14
16
18
−200 −100 0 100 2000
2
4
6
8
10
12
14
16
off axis distance (mm)
dose(c
Gy)
middle inline profile
TPS
EPID
−200 −100 0 100 2000
2
4
6
8
10
12
14
16
off axis distance (mm)
dose(c
Gy)
middle cross line profile
TPS
EPID
Gamma evaluation
Y (
4%
/3m
m)
20 40 60 80 100 120 140 160
20
40
60
80
100
120
140
160
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
e)
a)
d) c)
b)
50
Table 3-2.: Results of the EPID dose verification of an IMRT plan (pelvis, 15MV
beams, 7 fields) from anthropomorphic pelvic phantom measurements.
field
name
number
of
segments
Gantry
angle
phantom
thickness
number of
MU
field area
(cm2)
γ1% γmean % area γ >1
B1 8 30 24.5 116.9 627.0 1.40 0.47 5.49
B2 10 105 31.5 120.8 619.9 1.00 0.38 1.05
B3 8 130 29 87.0 582.4 1.38 0.50 6.44
B4 14 180 20 168.9 625.6 1.30 0.42 4.25
B5 9 230 29 105.1 560.5 1.30 0.38 3.48
B6 9 255 31.8 116.9 592.8 1.28 0.37 3.10
B7 7 330 24 91.9 588.0 1.18 0.42 2.37
3.4 Discussion
This dissertation deals with a method able to control and verify the administrated dose
distribution to patients during single treatment sessions (fractions). In this chapter the
feasibility of a method to verify patient transit dose during step-and-shoot IMRT
treatment was presented. The evaluation of measured dose is performed field-by-field at
the EPID level. The transit dose maps were calculated in the TPS using extended CT
images. This method has been used in previous studies for pre-treatment verification
(Reich et al., 2006; Mohammadi et al., 2008). In these studies, for non-zero gantry
angle measurements, the original CT images were rotated, and for each beam angle a
new rotated CT image set was imported and then transit dose maps of all beams were
calculated at zero gantry angle. This workflow manipulates the original CT data and
increases the risk of errors due to its complexity. In a non-transit study by Tyler et al.,
the XiO and Monaco treatment planning systems were used to calculate the fluence
maps in water at !!"#. In that study, the verification was performed with all gantry,
collimator and couch angles set to zero degrees; i.e it can only be used for pre-treatment
verification (Tyler et al., 2013). In our method, for each gantry angle, the predicted dose
51
map was modeled in its nominal planned position, which makes it possible to use for
dosimetry during patient treatment.
3.4.1 Flood field correction
There are different methods to remove the flood field correction required for non-
dosimetry EPID image processing. A correction method was developed based on the
raw pixel response map and ion- chamber measurements at dmax (Greer, 2005). In
another study by Nijsten et al. (2007a) the dose profile in water was measured and the
data used for a deconvolution of the EPID images, while simultaneously the field size
dependence and beam profile correction were obtained. Since the raw EPID image is
divided by the flood field image in automatic processing, Wendling et al. removed the
flood field correction by multiplying it by each acquired dose image. (Wendling et al.,
2006) This is only possible when the sensitivity correction is performed separately to
the flood (and dark) field correction. In our study, as shown in Figure 3-2and Figure
3-3, the raw flood field images were used and a 2D FFC correction matrix was fitted to
these images. The advantage of this method is that while the original sensitivity
correction of the system is kept, the beam profile is put back to the images. In the
techniques of first two studies mentioned above, the corrections are based on a different
photon beam spectrum (in Perspex or water) in comparison to the clinical conditions
(only the EPID and no extra materials).
3.4.2 Open field calibration for transmission dosimetry
Calibration of the EPID for dosimetry has been done in many different ways. In similar
studies the EPID system has been calibrated without any object in the beam; e.g.
ionization chamber measurements at the EPID level were used as a reference (Wendling
et al., 2006; Nijsten et al., 2007a). In another study the dose at dref in a water phantom
(at a source-to-surface distance of 100cm) calculated by the TPS was used as reference
dose for calibrating the system (Lee et al., 2009). Berry et al. (2012) used the intrinsic
Varian calibration method, which is based on in-air calibration at the SSD of 100cm.
This method is not applicable for all therapy machines because the EPID cannot always
be positioned at 100cm from the beam focus. The problem with methods that calibrate
the EPID with only an open beam (without any attenuator), is that the photon beam
energy spectrum is different from the transmitted spectrum. The methods mentioned
above that calibrate the EPID with multiple correction steps (open and attenuated
3
4
52
beams) have been proven to work, but are extremely complicated. The advantage of our
approach is that the calibration is performed with the transmitted beam from the
beginning (patient-EPID set-up), so while it is simple, the beam spectrum is still
representative of patient measurements. Without back-projecting the dose, there is no
need to isolate/remove the scatter within the EPID.
3.4.3 Field size correction
One of the most common methods for correcting for the field size effect is using a
convolution technique. In one study, ionization chamber measurements were used in a
mini-phantom at the position of the EPID, but without a full-scatter EPID phantom.
This method was designed for back-projection, so it was necessary to obtain a kernel
correction to remove the EPID scatter component, in combination with non-scatter and
full-scatter field size measurements. (Wendling et al., 2006). Nijsten et al. (2007a) also
used a ‘field size dependent kernels’ correction method, which was based on ionization
chamber measurements in water. In an other study, (Lee et al., 2009) determined the
changes in dose with respect to the field size at different water depths, and compared
them with corresponding EPID response. In the present study, an algorithm was
developed for plan verification for the forward transit approach. The effects of square
field size variations were studied for transit dose measurement in a patient / EPID
phantom combination (see figure 1b). One of the advantages of this method compared
to other studies is that the transit beam spectrum was used instead of that of the open
beam (without attenuator in beam path).
3.4.4 Phantom thickness correction
The energy spectrum of the photons that reach the EPID system changes dramatically
with the amount of attenuating material in the beam path to the detector. In one study,
the transmission through slab phantoms of varying thicknesses was studied for two field
sizes (10×10cm2 and 20×20cm
2), with three different beam energies (Pecharroman-
Gallego et al., 2011). According to the findings of this study, the beam hardening with
increasing phantom thickness could be described by adding an extra term in the
exponential attenuation, while there was only a very small deviation in the transmission
for different field sizes. (Jung et al., 2012) used Monte Carlo simulations and reported a
4 to 5% dose-response variation for thickness changes between 20 and 38cm. In our
study, a similar approach was used to correct for the thickness variation effect. This
Figure 3-1
53
correction is relevant during transit dose measurements of patients, where there is a
wide range of patient thicknesses, especially at different gantry angles. For instance, the
radiological path length through the pelvic phantom varied from 21 to 32cm at different
gantry angles, mainly due to varying geometry, but with a small variation due to
inhomogeneity. The current factor corrects for mainly geometrical variation, and further
corrections for more extreme changes in homogeneity is the subject of future work.
3.4.5 Over-response correction
There are different methods to improve the over-response of EPID systems in regions
that receive a dose from lower energy photons. McDermott et al. (2004) used an
additional 2mm copper build-up layer for filtering out the low energy scattered photons.
This solution needs an extra modification of the EPID system and therefore it is not a
reasonable solution for all clinics. (Nijsten et al., 2007a) used an energy spectrum
correction based on a mean transmission correction within the treatment field. To obtain
a model for this energy spectrum correction, they acquired images during a dummy
session without a patient. This method needs extra measurement time and in addition
corrections for the out-of-field regions, based on the energy spectrum inside the field,
which could introduce dose differences in the out-of-field region.
In the present study, a new empirical correction method was introduced that corrects for
the over-response of the EPID in the out-of-field regions. The advantage of our method
is that it is a simple empirical approach, which corrects the pixel values in the out-of-
field region of each IMRT segment individually.
3.4.6 Transit measured data analysis
There are different methods to verify the delivered dose to the patient. Each method has
its advantages and disadvantages. An Italian group developed a method to verify the
isocenter point dose by measuring the transmitted dose with an EPID (Cilla et al.,
2014; Fidanzio et al., 2014). They compared the daily 2D transmitted images (without
calibration) and reported that they could detect changes in patient setup, machine output
and MLC behavior during each treatment. This method could be a fast method but the
limitation of this approach is that the 2D method relies on the first fraction of treatment.
If there would be an error in the first fraction, the reference would be wrong for the rest
of the treatment. A 2D approach was developed by Berry et al. (2012). They simulated
54
the effect of attenuating material (phantom or patient), and modified a method
developed earlier (Van Esch et al., 2004), which is based on in-air verification. They
tested their model for pre-treatment verification using a fixed air gap and two ‘couch-to-
source distances’. In a follow-up study these authors reported the results of applying
this method for the verification of the transit dose of 9 patients treated with IMRT using
a 6MV photon beam (Berry et al., 2014). They reported that on average 95.7% of the
pixels passed the gamma evaluation with 5%/3-mm criteria for the beams that had no
interference with the couch. At the MAASTRO clinic in the Netherlands a 2D
calibration model for transit dose measurements was developed, which included an
energy spectrum correction (Nijsten et al., 2007a). They reported a satisfying gamma
evaluation across the entire EPID plane inside the field, but they found large dose
differences in out-of-field regions. They also determined their corrections for only one
SSD of 150cm. To use this method at other distances it needs an extra energy spectrum
correction while the patient scatter contribution to the EPID may also increase
significantly. In the presented new method, the patient scatter at different positions has
been taken into account by modeling and correcting the segments individually (see eq.
7) In addition, the over-response of the EPID, which could cause a problem in the out-
of-field region, has been taken into account. This method is simple and does not need
elaborate modeling time.
3.4.7 IMRT plan verification
Several studies have developed EPID transit dosimetry methods and similar results have been
reported. In the Netherlands Cancer Institute a back-projection method was developed for
verifying the patient treatment both pre-treatment and in vivo. The accuracy of the approach was
tested for a pre-treatment verification of a five-field IMRT prostate plan. In this test, the
midplane dose for each field was checked separately. The gamma criterion to compare EPID
versus film in this study was 2%/2mm. For all beams, the mean γ value was 0.4 and the mean
maximum γ index 1.2 and the mean % of pixels with γ>1 was 0.05% (Wendling et al., 2006).
This method is now applied routinely for the verification of all IMRT and IMAT treatments in
NKI (Olaciregui-Ruiz et al., 2013). Chung et al.(2011) used a 2D back-projection dose-
reconstruction algorithm to verify six conformal fields and also five IMRT fields in a slab
phantom using a 2D array detector (MapCHECK). With the gamma criterion of 3%/3mm, they
found for the IMRT fields (each with five segments) that three had a 100% gamma index
passing rate, and two had gamma index passing rates of 99.6% and 98.8%. In the present study
the global gamma evaluation of the 6MV head-and-neck IMRT plan resulted in γ1%, γmean and
55
percent of area with γ>1 with mean values of [1.07 ± 0.06, 0.41 ± 0.04, 2.41 ± 0.81%] and
maximum values of 1.14, 0.44 and 3.2%, respectively. The 15 MV pelvic IMRT plan
verification was performed with an anthropomorphic phantom.
A similar approach was reported by (Mohammadi et al., 2008). According to their
results, for fields in the absence of the treatment couch, approximately 95% of the field
area passed the gamma criteria (2.54 mm and 3% of central axis dose). For the field
with a treatment couch in the beam path, the result reduced to 85%. In our study, for the
IMRT fields tested with the pelvic phantom at the planned gantry angles, the mean value
of γ!"#$, the mean value of γ!%, and mean percent of area with γ>1 were 3.74 ± 1.83,
0.42±0.05!!"# 1.26 ± 0.14 respectively. Therefore in this chapter the verification method
presented per field showed significantly better results compared to previously reported
methods that passed through the treatment couch.
Our method has been developed for treatment fraction dose verification and in this
paper the first results of a phantom study are reported. Issues that were investigated
concerned the dependence of different treatment plan parameters, such as field size,
phantom thickness, and gantry rotation on the measured transit doses, and as proof of
principle, the analysis of the results using gamma evaluation. It should be noted that
gamma evaluation alone is not sufficient and could miss clinically relevant dose errors
(Nelms et al., 2013). However, it is a very common tool that could help to detect major
errors. Compared to studies mentioned above, our method is simple and could be used
to detect errors outside pre-defined gamma alert criteria. Further work is in progress to
test the accuracy and sensitivity of our method with respect to other parameters such as
patient mis-positioning, changes in patient anatomy as well as errors in machine
parameters. In addition, the clinical implementation of this method for actual patient
treatments will be investigated.
3.5 Conclusion We developed a method for patient transit fraction dose verification of IMRT. In this
method an EPID was calibrated against an ionization chamber. The EPID system is
constructed from non-water equivalent materials, and is primarily designed as an
imaging device, therefore its response had to be corrected for 2D transit dose
measurement. In the protocol, corrected EPID data are compared to predicted dose
images calculated by the commercial TPS system used in our clinic. The presented
56
method is simple and accurate, and is expected to form part of our QA program for
IMRT. In our study we have used global gamma evaluation criteria of 4% and 3 mm,
and it has shown that it could verify IMRT treatments delivered under clinical
conditions. Future studies including actual patient measurements at different treatment
sites are needed to determine its clinical potential.
57
58
4 Summary and conclusion In practice of radiation therapy, especially using advanced technologies like IMRT
(intensity modulated radiation therapy) and IMAT (intensity modulated arc therapy) a
daily check of series of treatment parameters is absolutely essential. Treatment
parameters include beam information, patient position and movement. This study
sought to show that EPIDs (electronic portal imaging devices), which often are installed
at treatment units, have got the potential for controlling patient position, quality of beam
shaping as well as the dose distribution administered to the PTV (planning target
volume).
Initially, the treatment planning systems (TPS) are designed for calculating and
optimizing the dose distribution in patients. In addition to this aim, they are able to
calculate the transmitted dose through the patient. The transmitted dose which is
impinging the EPID system could be measured separately for each beam direction
(projection) and any subfield. TPSs usually are optimized for using CT-data of patients
having a fixed frame for the field of view. Inside the treatment room the EPID is far out
of this matrix. For calculating the dose with the EPID, an expansion of this calculation
matrix of the TPS was necessary. Transit dose calculation should provide correct
information about both magnitude and position of dose distribution. Therefore
influential parameters in dose calculation and measurement in transit dose were reported
in this study.
Conclusion can be drawn that the commercial TPS system (Pinnacle), used in this
study, is able to calculate correctly the transit dose at extended distances. On the other
hand the pre-calculated transit dose distribution can be used for comparison with the
actually measured one at the EPID level under clinical conditions with a real patient.
The experiment in chapter three was carried out to calibrate the EPID system for transit
dose measurement. The EPID was calibrated against an ionization chamber. While the
EPID system is constructed from non-water equivalent materials, and is primarily
designed as an imaging device, its response had to be corrected for 2D transit dose
measurement. Influences of different treatment parameters such as beam energy, field
size and phantom thickness on the transit dose were investigated. In the protocol,
corrected EPID data are compared to predicted dose images calculated by the Pinnacle
TPS used in our clinic. As proof of principle measured transit dose maps using EPID
were compared with predicted maps from TPS.
59
Since the invention of EPID systems, there were many studies to develop one universal
in vivo dosimetry method for patient treatment verification. Each work has advantages
and disadvantages, which were discussed in chapter two and three, but up to now the
EPID dosimetry methods are not yet gained wide clinical use because of limited
availability of commercial software solution (Mijnheer et al., 2013). The presented
method in this study is simple to calibrate and transform EPID response into transit dose
measurement. The calibrated EPID dose maps are comparable with the calculated dose
map by the planning system. By comparing the measured and calculated data it would
be possible to detect all deviations larger than the intrinsic error margin of the EPID-
TPS combination. For clinical quality management practice also error margins are
defined, which are usually bigger than the intrinsic ones.
Future research is necessary in multiple areas to have fast online EPID dosimetry
solution. They could be listed as bellow;
• The presented method has been tested using a slab phantom and also an
anthropomorphic pelvic phantom; it should be tested for dealing with bigger
inhomogeneity changes such as inside the thorax region.
• In order to evaluate the efficiency of the method it is recommended to simulate
clinically relevant errors (such as missing segment, wrong beam energy, patient
miss positioning,…).
• In the current report the physical principle of EPID transit dosimetry was
studied. It might be recommendable for acceleration and improvement of the
clinical procedure to create a graphical user interface capable of grabbing the
calculated data from the planning system and the measured data from EPID (this
part of work is on progress by a master student).
• The presented method has been tested on Siemens Oncor Linacs, which are
equipped with flat panel detectors manufactured by Perkin Elmer, Inc. The
calibration and correction methods of the EPID system could thus be used in a
similar way for the other machines that have identical detector systems (Elekta,
Vero4DRT). In future this method will also be tested with the Varian and
Vero4DRT machines in our clinic.
• There is a need to test EPID transit dosimetry for several patient treatment sites
to find the clinical applications and limits of this method.
60
5 Bibliography
AAPM!1994!Physical!aspects!of!quality!assurance!in!radiation!therapy.!!
AAPM!2005!Report!of!TG!62!of!the!Radiation!Therapy!Committee,!“Diode!in!vivo!
dosimetry! for! patients! receiving! external! beam! radiation! therapy,”! AAPM!
Report!No.!87!(Medical!Physics!Publishing,!Madison,!WI,!2005).!!
Ali!A!S,!Dirkx!M!L,!Breuers!M!G!and!Heijmen!B!J!2011!Inclusion!of! the!treatment!
couch! in! portal! dose! image! prediction! for! high! precision! EPID! dosimetry!
Med+Phys!38!377`81!
Antonuk! L! E! 2002! Electronic! portal! imaging! devices:! a! review! and! historical!
perspective! of! contemporary! technologies! and! research!Phys+Med+Biol!47!
R31`65!
Bailey!D!W,!Kumaraswamy!L,!Bakhtiari!M,!Malhotra!H!K!and!Podgorsak!M!B!2012!
EPID! dosimetry! for! pretreatment! quality! assurance!with! two! commercial!
systems!J+Appl+Clin+Med+Phys!13!3736!
Beasley!M,!Driver!D!and!Dobbs!H!J!2005!Complications!of!radiotherapy:!improving!
the!therapeutic!index!Cancer+Imaging!5!78`84!
Berry! S! L,! Polvorosa! C,! Cheng! S,! Deutsch! I,! Chao! K! S! and!Wuu! C! S! 2014! Initial!
clinical! experience! performing! patient! treatment! verification! with! an!
electronic! portal! imaging! device! transit! dosimeter! Int+ J+ Radiat+ Oncol+ Biol+
Phys!88!204`9!
Berry! S! L,! Polvorosa! C! S! and! Wuu! C`S! 2010! A! field! size! specific! backscatter!
correction!algorithm!for!accurate!EPID!dosimetry!Medical+Physics!37!2425`
34!
Berry! S! L,! Sheu! R! D,! Polvorosa! C! S! and!Wuu! C! S! 2012! Implementation! of! EPID!
transit!dosimetry!based!on!a!through`air!dosimetry!algorithm!Med+Phys!39!
87`98!
Blake!S!J,!McNamara!A!L,!Deshpande!S,!Holloway!L,!Greer!P!B,!Kuncic!Z!and!Vial!P!
2013a!Characterization!of!a!novel!EPID!designed!for!simultaneous!imaging!
and!dose!verification!in!radiotherapy!Medical+Physics!40!091902`11!
Blake!S!J,!McNamara!A!L,!Vial!P,!Holloway!L,!Greer!P!B!and!Kuncic!Z!2014!Monte!
Carlo! simulation! of! the! transit! dosimetric! response! of! an! a! `Si! electronic!
portal!imaging!device!Journal+of+Physics:+Conference+Series!489!012005!
Blake! S! J,! Vial! P,! Holloway! L,! Greer! P! B,! McNamara! A! L! and! Kuncic! Z! 2013b!
Characterization! of! optical! transport! effects! on! EPID! dosimetry! using!
Geant4!Med+Phys!40!041708!
Chen! J,! Chuang! C! F,! Morin! O,! Aubin! M! and! Pouliot! J! 2006! Calibration! of! an!
amorphous`silicon! flat! panel! portal! imager! for! exit`beam! dosimetry!Med+
Phys!33!584`94!
Chung! H,! Li! J! and! Samant! S! 2011! Feasibility! of! using! two`dimensional! array!
dosimeter! for! in!vivo!dose!reconstruction!via! transit!dosimetry! J+Appl+Clin+
Med+Phys!12!3370!
Chytyk! K! and! McCurdy! B! M! C! 2009! Comprehensive! fluence! model! for! absolute!
portal!dose!image!prediction!Medical+Physics!36!1389`98!
62
Chytyk`Praznik!K,!VanUytven!E,! vanBeek!T!A,!Greer!P!B!and!McCurdy!B!M!2013!
Model`based!prediction!of!portal!dose!images!during!patient!treatment!Med+
Phys!40!031713!
Cilla!S,!Azario!L,!Greco!F,!Fidanzio!A,!Porcelli!A,!Grusio!M,!Macchia!G,!Morganti!A!G,!
Meluccio! D! and! Piermattei! A! 2014! An! in`vivo! dosimetry! procedure! for!
Elekta!step!and!shoot!IMRT!Phys+Med!30!419`26!
Dahlgren!C!V,!Eilertsen!K,! Jørgensen!T!D!and!Ahnesjö!A!2006!Portal!dose! image!
verification:! the! collapsed! cone! superposition! method! applied! with!
different!electronic!portal!imaging!devices!Phys+Med+Biol!51!335`49!
ESTRO! 1995! Recommendations! for! a! Quality! Assurance! Programme! in! External!
Radiotherapy!(Booklet!No.2).!!
ESTRO!2006!Methods!for!In!vivo!Dosimetry!in!external!radiotherapy,!2!ed,!physics!
for!clinical!radiotherapy,!Booklet!no.!1,!ESTRO,!Brussels.!!
Fidanzio!A,!Greco!F,!Mameli!A,!Azario!L,!Balducci!M,!Gambacorta!M!A,!Frascino!V,!
Cilla!S,!Sabatino!D!and!Piermattei!A!2010!Breast!in!vivo!dosimetry!by!EPID!J+
Appl+Clin+Med+Phys!11!3275!
Fidanzio!A,!Porcelli!A,!Azario!L,!Greco!F,!Cilla!S,!Grusio!M,!Balducci!M,!Valentini!V!
and! Piermattei! A! 2014! Quasi! real! time! in! vivo! dosimetry! for! VMAT!Med+
Phys!41!062103!
Greer!P!B!2005!Correction!of!pixel! sensitivity!variation!and!off`axis! response! for!
amorphous!silicon!EPID!dosimetry!Med+Phys!32!3558`68!
Greer!P!B! and!Popescu!C!C!2003!Dosimetric! properties! of! an! amorphous! silicon!
electronic! portal! imaging! device! for! verification! of! dynamic! intensity!
modulated!radiation!therapy!Med+Phys!30!1618`27!
Grein!E!E,!Lee!R!and!Luchka!K!2002!An!investigation!of!a!new!amorphous!silicon!
electronic!portal!imaging!device!for!transit!dosimetry!Med+Phys!29!2262`8!
Herman!M!G,!Balter!J!M,!Jaffray!D!A,!McGee!K!P,!Munro!P,!Shalev!S,!Van!Herk!M!and!
Wong! J!W!2001!Clinical! use! of! electronic! portal! imaging:! report! of!AAPM!
Radiation!Therapy!Committee!Task!Group!58!Med+Phys!28!712`37!
IAEA!1998!Quality!Assurance!in!Radiotherapy,!IAEA!TECDOC!989.!!
IAEA! 2000! Absorbed! dose! determination! in! external! beam! radiotherapy:! an!
international!code!of!practice!for!dosimetry!based!on!standards!of!absorbed!
dose! to! water! (Vienna:! IAEA)!IAEA!TRS`398! (www`
pub.iaea.org/mtcd/publications/pdf/trs398_scr.pdf).!!
IAEA! 2005! Radiation+ Oncology+ Physics:+ A+ Handbook+ for+ Teachers+ and+ Students,+
IAEA,+vienna+
IAEA!2013!Development!of!procedures!for!in+vivo!dosimetry!in!radiotherapy,!IAEA!
Human! Health! Report! No.! 8.! (International! Atomic! Energy! Agency,! IAEA,!
Vienna,!Austria,!2013).!!
Jung!J!W,!Kim!J!O,!Yeo!I!J,!Cho!Y!B,!Kim!S!M!and!DiBiase!S!2012!Fast!transit!portal!
dosimetry! using! density`scaled! layer! modeling! of! aSi`based! electronic!
portal!imaging!device!and!Monte!Carlo!method!Med+Phys!39!7593`602!
Khan!F!M!2010!The+Physics+of+Radiation+Therapy:!Lippincott!Williams!&!Wilkins)!
Kirkby! C! and! Sloboda! R! 2005! Consequences! of! the! spectral! response! of! an! a`Si!
EPID!and!implications!for!dosimetric!calibration!Med+Phys!32!2649`58!
Langmack!K!A!2001!Portal!imaging!Br+J+Radiol!74!789`804!
Lee! C,! Menk! F,! Cadman! P! and! Greer! P! B! 2009! A! simple! approach! to! using! an!
amorphous!silicon!EPID!to!verify!IMRT!planar!dose!maps!Med+Phys!36!984`
92!
63
Low!D!A,!Moran!J!M,!Dempsey!J!F,!Dong!L!and!Oldham!M!2011!Dosimetry!tools!and!
techniques!for!IMRT!Med+Phys!38!1313`38!
Margalit!D!N,!Chen!Y!H,!Catalano!P!J,!Heckman!K,!Vivenzio!T,!Nissen!K,!Wolfsberger!
L!D,!Cormack!R!A,!Mauch!P!and!Ng!A!K!2011!Technological!advancements!
and!error!rates!in!radiation!therapy!delivery!Int+J+Radiat+Oncol+Biol+Phys!81!
12!
Mayles!P,!Mayles!P,!Nahum!A,!Nahum!A!E,!Rosenwald!J!C!and!Rosenwald!J!C!2007!
Handbook+of+Radiotherapy+Physics:+Theory+and+Practice:!Taylor!&!Francis)!
McCurdy!B!M,!Luchka!K!and!Pistorius!S!2001!Dosimetric!investigation!and!portal!
dose!image!prediction!using!an!amorphous!silicon!electronic!portal!imaging!
device!Med+Phys!28!911`24!
McCurdy!B!M!and!Pistorius!S!2000!Photon!scatter!in!portal!images:!accuracy!of!a!
fluence!based!pencil!beam!superposition!algorithm!Med+Phys!27!913`22!
McDermott!L!N,!Louwe!R!J,!Sonke!J! J,!van!Herk!M!B!and!Mijnheer!B!J!2004!Dose`
response! and! ghosting! effects! of! an! amorphous! silicon! electronic! portal!
imaging!device!Med+Phys!31!285`95!
McDermott!L!N,!Nijsten!S!M,!Sonke!J! J,!Partridge!M,!van!Herk!M!and!Mijnheer!B!J!
2006!Comparison!of!ghosting!effects! for! three!commercial!a`Si!EPIDs!Med+
Phys!33!2448`51!
McDermott! L! N,! Wendling! M,! Sonke! J! J,! van! Herk! M! and! Mijnheer! B! J! 2007!
Replacing! pretreatment! verification! with! in! vivo! EPID! dosimetry! for!
prostate!IMRT!Int+J+Radiat+Oncol+Biol+Phys!67!1568`77!
McNutt! T! R,! Mackie! T! R,! Reckwerdt! P! and! Paliwal! B! R! 1996a! Modeling! dose!
distributions!from!portal!dose!images!using!the!convolution/superposition!
method!Med+Phys!23!1381`92!
McNutt! T! R,! Mackie! T! R,! Reckwerdt! P,! Papanikolaou! N! and! Paliwal! B! R! 1996b!
Calculation!of!portal!dose!using!the!convolution/superposition!method!Med+
Phys!23!527`35!
Mijnheer! B,! Beddar! S,! Izewska! J! and! Reft! C! 2013! In! vivo! dosimetry! in! external!
beam!radiotherapy!Med+Phys!40!070903!
Mohammadi! M,! Bezak! E! and! Reich! P! 2008! Verification! of! dose! delivery! for! a!
prostate!sIMRT!treatment!using!a!SLIC`EPID!Appl+Radiat+Isot!66!1930`8!
Nelms!B!E,!Chan!M!F,! Jarry!G,!Lemire!M,!Lowden! J,!Hampton!C!and!Feygelman!V!
2013! Evaluating! IMRT! and! VMAT! dose! accuracy:! practical! examples! of!
failure!to!detect!systematic!errors!when!applying!a!commonly!used!metric!
and!action!levels!Med+Phys!40!111722!
Nicolini!G,!Clivio!A,!Vanetti!E,!Krauss!H,!Fenoglietto!P,!Cozzi!L!and!Fogliata!A!2013!
Evaluation!of!an!aSi`EPID!with!flattening!filter! free!beams:!applicability!to!
the! GLAaS! algorithm! for! portal! dosimetry! and! first! experience! for!
pretreatment!QA!of!RapidArc!Med+Phys!40!111719!
Nicolini!G,!Fogliata!A,!Vanetti!E,!Clivio!A,!Vetterli!D!and!Cozzi!L!2008!Testing!the!
GlaaS! algorithm! for! dose!measurements! on! low`! and! high`energy! photon!
beams!using!an!amorphous!silicon!portal!imager!Med+Phys!35!464`72!
Nijsten! S! M,! van! Elmpt!W! J,! Jacobs!M,! Mijnheer! B! J,! Dekker! A! L,! Lambin! P! and!
Minken!A!W!2007a!A!global!calibration!model!for!a`Si!EPIDs!used!for!transit!
dosimetry!Med+Phys!34!3872`84!
Nijsten!S!M!J!J!G,!Mijnheer!B!J,!Dekker!A!L!A!J,!Lambin!P!and!Minken!A!W!H!2007b!
Routine! individualised! patient! dosimetry! using! electronic! portal! imaging!
devices!Radiotherapy+and+Oncology!83!65`75!
64
Olaciregui`Ruiz! I,! Rozendaal! R,! Mijnheer! B,! van! Herk! M! and! Mans! A! 2013!
Automatic!in!vivo!portal!dosimetry!of!all!treatments!Phys+Med+Biol!58!8253`
64!
Pecharroman`Gallego!R,!Mans!A,!Sonke!J!J,!Stroom!J!C,!Olaciregui`Ruiz!I,!van!Herk!
M!and!Mijnheer!B! J!2011!Simplifying!EPID!dosimetry! for! IMRT! treatment!
verification!Med+Phys!38!983`92!
Persoon!L!C,!Nijsten!S!M,!Wilbrink!F!J,!Podesta!M,!Snaith!J!A,!Lustberg!T,!van!Elmpt!
W!J,!van!Gils!F!and!Verhaegen!F!2012!Interfractional!trend!analysis!of!dose!
differences!based!on!2D!transit!portal!dosimetry!Phys+Med+Biol!57!6445`58!
Piermattei! A,! Greco! F,! Azario! L,! Porcelli! A,! Cilla! S,! Zucca! S,! Russo! A,! Castro! E!D,!
Russo!M,!Caivano!R,!Fusco!V,!Morganti!A!and!Fidanzio!A!2012!A!National!
project! for! in! vivo! dosimetry! procedures! in! radiotherapy:! First! results!
Nuclear+ Instruments+ and+ Methods+ in+ Physics+ Research+ Section+ B:+ Beam+
Interactions+with+Materials+and+Atoms!274!42`50!
Reich! P,! Bezak! E,!Mohammadi!M! and! Fog! L! 2006! The! prediction! of! transmitted!
dose!distributions!using!a!3D!treatment!planning!system!Australas+Phys+Eng+
Sci+Med!29!18`29!
Rowshanfarzad!P,!McCurdy!B!M,!Sabet!M,!Lee!C,!O'Connor!D!J!and!Greer!P!B!2010!
Measurement! and! modeling! of! the! effect! of! support! arm! backscatter! on!
dosimetry!with!a!varian!EPID!Med+Phys!37!2269`78!
Sterling! T! D,! Perry! H! and! Katz! L! 1964! Automation! of! Radiation! Treatment!
Planning.! Iv.! Derivation! of! a! Mathematical! Expression! for! the! Per! Cent!
Depth!Dose!Surface!of!Cobalt!60!Beams!and!Visualisation!of!Multiple!Field!
Dose!Distributions!Br+J+Radiol!37!544`50!
Stock!M,!Kroupa!B!and!Georg!D!2005!Interpretation!and!evaluation!of!the!gamma!
index!and!the!gamma!index!angle!for!the!verification!of!IMRT!hybrid!plans!
Phys+Med+Biol!50!399`411!
Teoh!M,!Clark!C!H,!Wood!K,!Whitaker!S!and!Nisbet!A!2011!Volumetric!modulated!
arc! therapy:!a! review!of!current! literature!and!clinical!use! in!practice!Br+J+
Radiol!84!967`96!
Teymurazyan! A! and! Pang! G! 2012! Monte! Carlo! simulation! of! a! novel! water`
equivalent!electronic!portal!imaging!device!using!plastic!scintillating!fibers!
Med+Phys!39!1518`29!
Tyler!M,!Vial!P,!Metcalfe!P!and!Downes!S!2013!Clinical!validation!of!an! in`house!
EPID!dosimetry!system!for!IMRT!QA!at!the!Prince!of!Wales!Hospital!Journal+
of+Physics:+Conference+Series!444!012043!
van!Elmpt!W,!McDermott!L,!Nijsten!S,!Wendling!M,!Lambin!P!and!Mijnheer!B!2008!
A!literature!review!of!electronic!portal!imaging!for!radiotherapy!dosimetry!
Radiother+Oncol!88!289`309!
Van!Esch!A,!Depuydt!T!and!Huyskens!D!P!2004!The!use!of!an!aSi`based!EPID!for!
routine! absolute! dosimetric! pre`treatment! verification! of! dynamic! IMRT!
fields!Radiother+Oncol!71!223`34!
van!Herk!M!2004!Errors!and!margins!in!radiotherapy!Semin+Radiat+Oncol!14!52`64!
van! Zijtveld!M,!Dirkx!M,! Breuers!M,! de!Boer!H! and!Heijmen!B! 2009!Portal! dose!
image! prediction! for! in! vivo! treatment! verification! completely! based! on!
EPID!measurements!Med+Phys!36!946`52!
Warkentin!B,!Steciw!S,!Rathee!S!and!Fallone!B!G!2003!Dosimetric!IMRT!verification!
with!a!flat`panel!EPID!Med+Phys!30!3143`55!
65
Wendling!M,!Louwe!R! J,!McDermott!L!N,!Sonke! J! J,! van!Herk!M!and!Mijnheer!B! J!2006!Accurate! two`dimensional! IMRT!verification!using!a!back`projection!EPID!dosimetry!method!Med+Phys!33!259`73!
Wendling!M,!McDermott! L!N,!Mans!A,! Olaciregui`Ruiz! Í,! Pecharromán`Gallego!R,!Sonke! J! J,! Stroom! J,! van!Herk!M!and!Mijnheer!B! J!2012! In!aqua!vivo!EPID!dosimetry!Med+Phys!39!367`77!
Wendling!M,!McDermott!L!N,!Mans!A,!Sonke!J!J,!van!Herk!M!and!Mijnheer!B!J!2009!A!simple!backprojection!algorithm!for!3D! in!vivo!EPID!dosimetry!of! IMRT!treatments!Med+Phys!36!3310`21!
WHO!2014!Cancer!Facts!sheet!No297.!World!Health!Organization.!February!2014.!Retrieved!10!June!2014.!!
Winkler! P,! Hefner! A! and! Georg! D! 2005! Dose`response! characteristics! of! an!amorphous!silicon!EPID!Med+Phys!32!3095`105!
66
67
6 Acknowledgment
It is only through the support and encouragement of numerous people that the
completion of a Ph.D program and a doctoral thesis can be accomplished. Foremost, I
would like to express my sincere gratitude to Prof. Dr. Reinhold G. Müller for the
continuous support of my Ph.D study and research, for his patience, motivation,
enthusiasm, and immense knowledge. His guidance helped me in all the time of
research and writing of this thesis.
I will take the privilege to express my sincere gratitude to Prof. Dr. Rainer Fietkau; the
director of radiotherapy department for the pleasant, cooperative and well-equipped
working environment. And many thanks for his permissions to use all equipment and
also for his support for joining professional international meetings.
I would like to thank Prof. Dr. Christoph Bert for his helps and guidance. His advises
assisted me in further development of my writing and communication skill, as well as
the ability to carry out research on a more effective and professional level.
I would like to express my deepest gratitude to Dr. Manfred Schmidt, for his excellent
guidance, caring, patience, and providing me with an excellent atmosphere for doing my
experiments.
I really appreciate Prof. Ben Mijnheer for his great help for my careers and also
research work. He was always available to me with supportive idea since 2004 that I
start my career in radiotherapy field. Although he was in the Netherlands (The
Netherlands Cancer Institute), I felt him beside me.
Thank is due to Dr. Mohammad Mohammadi for initial idea and his support for Matlab
programing.
A warm thank you to Markus Kellermeier, whom I had the pleasure of sharing an office
with him for five years. Thanks for your patience, honesty, approachability.
Furthermore I would also like to acknowledge with much appreciation the crucial role
of the medical physics team of radiotherapy department of Erlangen university clinic,
who supported me to use all required equipment and the necessary materials to
complete the task; Dr. Michael Lotter (the Godfather, who knows everything and kindly
help everyone), Dr. Ulrike Lambrecht, Dr. Nils Achterberg, Dr. Stefan Speer, Stephan
Kreppner and Wilhelm Stillkrieg, Lukas Kober, Dr. Andreas Richter, Alexander Weiss,
Antje Brückner and also my former colleague Dr. Andreas Klein.
68
I would also like to thank for the helps and supports of the research member of medical
physic team specially Dr. Indra Yohannes for his technical support for the planning
system issues, and also Jens Wölfelschneider, Tobias Brandt, Josefin Hartmann, Heru
Prasetio, Felix Böckelmann, Marco Serpa, Barbara Witulla and Dr. Stefan Vasiliniuc.
I would also like to thank my parents, and my brother. They were always supporting me
and encouraging me with their best wishes.
Last but not least I would like to thank my best friend, my wife Rahil. She was always
there cheering me up and stood by me through the good times and bad.
69
70
7 Curriculum Vitae PERSONAL
INFORMATION
Mohammadsaeed Saboori
Sex: Male | Date of birth 1977 |
Nationality: Iranian | Marital status: Married
EDUCATION AND
TRIANING
12/2010–Present Ph.D. (Candidate) Medical Physics
Universitätsklinikum Erlangen, Strahlenklinik, Erlangen
(Germany)
09/2002–11/2005 M.Sc. Medical Physics
Tarbiat Modarres University, Faculty of Medical Science, Tehran
(Iran)
Topic of thesis: Performing diode technical specification and
correction factors for patient dose verification using diode
dosimeter during External Radiotherapy of Prostate cancer.
09/1997–03/2002 B.Sc. Applied Physics (majoring in Nuclear Physics)
Shahid Beheshti University, Faculty of Science, Tehran (Iran)
PROFESSIONAL
CERTIFICATE
24/02/2012 Certified Medical Physic Expert (MPE)
Anerkennung Fachkunde zur Medizinphysik-Experte
WORK
EXPERIENCE
05/2009–Present Medical Physicist
Strahlenklinik; Universität Erlangen, Erlangen (Germany)
11/2003–04/2009 Medical Physicist
Radiotherapy Oncology Department, Shohada Hospital, Shahid
Beheshti University of Medical Science, Tehran (Iran)
07/2006–03/2009 Medical Physicist and Responsible Health Physicist (part time)
Radiotherapy Oncology Department, Payambaran hospital, Tehran
(Iran)
PUBLICATION
AND
PRESENTATION
Publication 2D patient transit EPID dosimetry for treatment fraction dose
verification, M. Saboori, M. Schmidt, C. Bert and R. G. Müller,
(submitted to Phys. Med. Biol. on Sep 2014)
71
Publication Radiation dose distribution under the area protected using a
Cerrobend block during external beam radiotherapy: a film
study,M. Mohammadi, A. Taherkhani, and M. Saboori, Journal of
Radiotherapy in Practice / Volume 13 / Issue 04 / December 2014,
pp 456-464
Publication Evaluation of the physical characteristic of Cerrobend blocks used
for radiation therapy, A. Taherkhani, M. Mohammadi, M. Saboori,
V. Changizi, International Journal of Radiation Research. 2010; 8
(2) :93-101.
Poster presentation QC tests and initial evaluation of transit dose measurement in
Vero4DRT using EPID system, M. Saboori, M. Schmidt, R.G.
Müller, C. Bert, EPI2k14,13th International Conference on
Electronic Patient Imaging, Aarhus, Denmark, 31 Aug - 3 Sep
2014
Poster presentation An experimental method for IMRT transit dose verification using
an a-Si EPID system, M. Saboori, M. Schmidt, M. Mohammadi,
C. Bert, R. Muller, ESTRO 33, Vienna, Austria from 04-08 April
2014
Poster presentation Specific patient exit Dose verification base on treatment planning
system and experimental data, M. Saboori, M Schmidt, R. Mueller,
M. Mohammadi, AAPM 54th annual meeting, Charlotte, North
Carolina, USA, 29 Jul- 2 Aug 2012
Poster presentation Validation of a model for EPID transit dose prediction for two
dimensional IMRT verification,M. Saboori,M. Schmidt, R. Müller,
M. Mohammadi,The 18th congress of radiation oncology german
society, (DEGRO 2012),Wiesbaden, Germany, 7-10 June 2012,
Poster presentation Experimental validation of a simple model for EPID transit dose
prediction by M. Saboori, M. Schmidt, M. Tzivaki, R. Müller, M.
Mohammadi, ESTRO 31 conference, Barcelona, Spain, 09-13 May
2012
Poster presentation Methods for determination of shielding requirements in HDR
Brachytherapy bunkers, M. S. Saboori, M. Abbaci, A. S.
Meigooni, B. H. Malayeri, A. J. Arfaee, ESTRO 25 meeting,
Leipzig, Germany, October 9 - 12, 2006.
Poster presentation Correction factors for in-vivo diode dosimetry in measurement of
Entrance Dose, M.S. Saboori, A. Jabari Arfaei. Sixth Iranian
Congress of Medical Physics, Mashhad, Iran, May 2004
PROFESSIONAL
TRAININGS AND
MEETINGS
EPI2k14, 13th
International Conference on Electronic Patient
72
Imaging, Aarhus, Denmark, 31 Aug - 3 Sep 2014.
ESTRO 33,Annual ESTRO meeting, Vienna, Austria, 4 -8 Apr
2014
IMRT/VMAT Comprehensive course, IBA Dosimetry
International Competence Center (ICC), Schwarzenbruck,
Germany, 9-11 Oct 2013.
VMAT - RapidArc course, The Netherlands Cancer Institute,
Amsterdam, The Netherlands, 12-14 Sep 2013
Dose Modelling and Verification for External Beam
Radiotherapy, (ESTRO teaching course) Florence, Italy, 10 - 14
Mar 2013
World congress of brachytherapy, (ESTRO meeting),
Barcelona, Spain, 10-12 May 201
Image-guided Radiotherapy in Clinical Practice (ESTRO
teaching course), Amsterdam, The Netherlands, 6 – 10 Nov 2011
EPI2kX, 11th International Conference on Electronic Patient
Imaging, Leuven, Belgium, 7-8 Jun 2010
IMRT and other conformal techniques in practice, (ESTRO
teaching course) Ghent, Belgium, 2 - 6 May 2010
Modern brachytherapy techniques, (ESTRO teaching course)
Istanbul, Turkey, 9 - 13 Mar 2008
Joint ICTP-IAEA Activity on Imaging in Advanced
Radiotherapy Techniques, Trieste-Italy, 20-24 Oct 2008.
AAPM TRAINING COURSE "Clinical Dosimetry and
Quality Assurance for Advanced Radiation Therapy
Techniques", Prague, Czech Republic, 16-20 Jun 2008.
Calculation and Verification for External Beam Therapy,
(ESTRO teaching course) Budapest, Hungary, 29 Apr - 3 May
2007
8th
European School of Medical Physics, Archamps, France, 27
Oct - 29 Nov 2005, The school's programs that I participated
were:(3rd week: Medical Computing,4th week: Physics of Modern
Radiotherapy, 5th week: Brachytherapy)
Workshop on Medical Applications of Synchrotron Radiation
MASR2004, Trieste, Italy, 23-25 Sep 2004, (Organized by IAEA
73
and UNESCO in International Center for Theoretical Physics)
College on Medical Physics, Trieste, Italy, 30 Aug – 22 Sep 2004,
(Organized by IAEA and UNESCO in International Center for
Theoretical Physics)