UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Evaluation of an amorphous silicon electronic portal imaging device for use in radiation therapy dosimetry Sibbald, Regan Estcourt 2005

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2005-0313.pdf [ 21.41MB ]
Metadata
JSON: 831-1.0092046.json
JSON-LD: 831-1.0092046-ld.json
RDF/XML (Pretty): 831-1.0092046-rdf.xml
RDF/JSON: 831-1.0092046-rdf.json
Turtle: 831-1.0092046-turtle.txt
N-Triples: 831-1.0092046-rdf-ntriples.txt
Original Record: 831-1.0092046-source.json
Full Text
831-1.0092046-fulltext.txt
Citation
831-1.0092046.ris

Full Text

EVALUATION OF AN AMORPHOUS SILICON ELECTRONIC PORTAL IMAGING DEVICE FOR USE IN RADIATION THERAPY DOSIMETRY. by REGAN ESTCOURT SIBBALD B.Sc, Simon Fraser University, 1991 B.Ed., University of British Columbia, 1993  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES PHYSICS  THE UNIVERSITY OF BRITISH COLUMBIA April 2005 © Regan Estcourt Sibbald, 2005  Abstract The objective of external beam radiation therapy is to deliver a high dose of sterilizing radiation to a diseased tumor site while minimizing the dose to as much of the surrounding normal tissue as possible. This is often accomplished by directing one or more megavoltage xray treatment portals at the diseased area from different directions. In theory, this principle maximizes the dose to the target site while minimizing the dose to surrounding healthy tissue. The efficacy of a course of radiation therapy may however be compromised by several factors. Geometric factors related to the accuracy to which the beam delivery system can reproduce the desired treatment set-up may cause deviations from the desired results. Patient set-up uncertainties are another factor and are related to the alignment of skin markings with respect to the underlying target structures within the body. In addition, a presumablyfixedtarget may be geometrically shifted due to its proximity to neighboring organs or the patient may simply move during treatment thereby shifting the treatment portal with respect to the intended target. All of these factors contribute to a sub-optimal treatment outcome and possibly an increase in the normal tissue complication probability (NTCP). On-line electronic portal imaging devices (EPIDs) have been developed to monitor and help correct the inaccuracies encountered in radiation therapy. EPIDs were developed for verification of geometric accuracy, as they are capable of producing near real-time 2 D projection images of the target volume during a radiotherapy treatment. However, it has also become apparent that in addition to geometric verification, EPIDs can provide valuable dosimetric information, indicating points of interest in terms of over and under dosage. In this thesis, the dosimetric characteristics of the amorphous silicon EPID are investigated.  ii  First, the EPIDs ability to accurately and reproducibly measure dose as a function of fluence rate is investigated. EPID pixel intensity values were found to vary by up to ± 3 % from the average value with a standard deviation of 1.3% when measuring the same dose with varying dose rates for both 6 and 18 M V photons. Next the water equivalent depth of the inherent buildup material on the EPID is determined. The inherent build-up of the portal imaging device was found to be 1.4 ± 0.2 cm for 6 M V and 2.2 ± 0.2 cm for 18 M V . A comparison of EPID and film images acquired during enhanced dynamic wedge treatments indicated that the film profiles show an increased response of film compared to the EPID as the distance from the central axis increases. The EPIDS dose response curve was characterized by comparing ion chamber measurements to EPID pixel intensity values. The dose response dependence on incident photon energy and fluence rate was also investigated. The ratio of dose to total pixel value was found to be to be 1.42 x 10" cGy/pix for 18 M V and 1.44 x 10" cGy/pix for 6 M V with confidence 4  4  intervals of 0.05 x 10" cGy/pix. Dose measurements were found to be reproducible with a 4  standard deviation of 1.3% for both 6 and 18 M V photons. A l l calibration equations were verified over the dose range of 10 cGy to 100 cGy.  iii  TABLE O F CONTENTS Abstract Table of Contents List of Figures Glossary of Terms CHAPTER I  General Introduction  1  1.1 An Introduction to Portal Imaging Devices 1.2 The Amorphous Silicon Electronic Portal Imaging Device (aSi-EPID) 1.3 Film Dosimetry 1.4 Ion Chamber 1.5 Dose 1.6 Photon Interactions with Matter 1.7 Off-Axis Spectral Distribution 1.8 Tissue-Maximum Ratio (TMR) 1.9 Dose at extended SDD 1.10 Measuring Dose with the EPID CHAPTER II  2.1 2.2 2.3 2.4  Introduction Method Results Conclusion  CHAPTER III  3.1 3.2 3.3 3.4  Testing the Dose Rate Stability of an Amorphous Silicon Portal Imaging Device  3 8 11 12 14 14 17 20 21 25 32  32 33 35 39  Effective Thickness of the Inherent Build-up inside the PortalVision aS500 Detector  Introduction Method Results Conclusion  40  40 42 44 48  iv  CHAPTER IV  4.1 4.2 4.3 4.4  50  Introduction Method Analysis Conclusion  CHAPTER V 5.1 5.2 5.3 5.4  A Comparison of 6 MV Enhanced Dynamic Wedge Profiles Acquired with Film and the EPID  50 51 53 62  EPID Response versus Thickness of Solid Water  Introduction Method Analysis Conclusion  CHAPTER VI  63  '"  Dosimetric Comparison of an Ion Chamber and EPID  63 65 66 90 92  6.1 Introduction 6.2 Method 6.3 Analysis  92 93 93  6.4 Conclusion  101  CHAPTER VII,  Conclusion and Recommendations for Further Work  Bibliography  103 109  v  List of Figures Figure  Title  Page  1.1.1 1.1.2  Schematic diagram o f a video based E P I D Conversion o f a Megavoltage X-ray Photon to Visible Photon Commercially Available E P I D S Image Detection Unit Sensitometric curve Ion Chamber The Flattening Filter Dose Profiles at Depth Tissue M a x i m u m Ratio Beam Divergence Dose Rate Instability Determination o f Left and Right Field Edges Determination o f Top and Bottom Field Edges Determination o f the Center o f a Non-aligned F i e l d Experimental Positioning Experimental Results for 6 M V Photons Experimental Results for 18 M V Photons ' Dose (cGy) per Total Pixel Value for 6 M V Dose (cGy) per Total Pixel Value for 18 M V Detector Positioning Relative Depths o f M a x i m u m Signal Total Pixel Value vs. Thickness (cm) for 6 M V Ion Chamber vs. Thickness (cm) for 6 M V Total Pixel Value vs. Thickness (cm) for 18 M V Ion Chamber vs. Thickness (cm) for 18 M V Definition o f Wedge Angle Dose (cGy) vs. P i x e l Value for the F i l m Calibration Dose (cGy) vs. Pixel Position for a F i l m Image o f a 15 Degree Enhanced Dynamic Wedge Treatment Experimental Dose vs. Pixel Values for 6 M V Photons 100 M U , 15 Degree Enhanced Dynamic Wedge Treatment 70 M U , 15 Degree Enhanced Dynamic Wedge Treatment 100 M U , 30 Degree Enhanced Dynamic Wedge Treatment 70 M U , 30 Degree Enhanced Dynamic Wedge Treatment 100 M U , 45 Degree Enhanced Dynamic Wedge Treatment 70 M U , 45 Degree Enhanced Dynamic Wedge Treatment 100 M U , 60 Degree Enhanced Dynamic Wedge Treatment 70 M U , 60 Degree Enhanced Dynamic Wedge Treatment T M R Curves as measured by the E P I D and Ion Chamber 18 M V T M R Data from E P I D and Ion Chamber 18 M V T M R Data from E P I D and Ion Chamber  5 6  1.1.3 1.2.1 1.3.1 1.4.1 1.7.1 1.7.2 1.8.1 1.9.1 1.10.1 1.10.2 1.10.3 1.10.4 2.2.1 2.3.1 2.3.2 2.3.3 2.3.4 3.1.1 3.2.1 3.3.1 3.3.2 3.3.3 3.3.4 4.1.1 4.3.1 4.3.2 4.3.3 4.3.4 4.3.5 4.3.6 4.3.7 4.3.8 4.3.9 4.3.10 4.3.11 5.1.1 5.3.1 5.3.2  vi  7 8 12 13 18 19 20 22 26 27 28 29 33 35 35 38 38 41 43 44 45 46 47 51 53 54 56 57 58 58 59 59 60 60 61 64 67 68  5.3.3 5.3.4 5.3.5 5.3.6 5.3.7  5.3.8 5.3.9 5.3.10 5.3.11 5.3.12 5.3.13 5.3.14 5.3.15 5.3.16 5.3.17 5.3.18 5.3.19 6.2.1 6.3.1 6.3.2 6.3.3 6.3.4 6.3.5 6.3.6 6.3.7 6.3.8 7.1  7.2 7.3  Percent Difference of EPID from Ion Chamber vs. Thickness (cm) 18 M V T M R Data Without Inherent Build-up Included 18 MV TMR Data Without Inherent Build-up Included Percent Difference of EPID vs. Ion Chamber (18 MV) Without Adding the effective Buildup TMR Data for EPID and Ion Chamber with EPID correction (Equation 5.3.3) for Equivalent Thickness of the Inherent Buildup of the EPID TMR Data Using Linearly Corrected Thickness for the EPID TMR Curves as Measured by the EPID and Ion Chamber 6 MV EPID and Ion Chamber TMR Data 6 MV EPID and Ion Chamber TMR Data Percent Difference of EPID from Ion Chamber TMR data for 6 MV with Inherent Build-up Included 6 MV EPID and Ion Chamber TMR Data without Inherent Buildup Included 6 MV TMR Data without Inherent Build-up Included Percent Difference of EPID from Ion Chamber TMR data for 6 MV Without Inherent Build-up Included Dose (cGy) vs. Thickness (cm) of Solid Water for 6 MV EPID Total Average Pixel Value vs. Thickness (cm) for 6 MV Dose (cGy) vs. Thickness (cm) of Solid Water for 18 MV EPID Total Pixel Value vs. Thickness (cm) of Solid Water forl8MV Positioning of the EPID and Ion Chamber Ion Chamber Reading (nC) vs. EPID Total Pixel Value P Calibration Curve for 6 MV Photons EPID Total Pixel Value vs. Dose (Monitor Units) Calibration Curve for 6 MV Photons EPID Calibration Curves for 6 MV Percent Difference in Dose from the Two 6 MV Calibration Curves as a Function of EPID Total Pixel Value Ion Chamber Reading (nC) vs. EPID Total Pixel Value P Calibration Curve fori 8 MV Photons EPID Total Pixel Value vs. Dose (Monitor Units) Calibration Curve for 18 MV EPID Calibration Curves for 18 MV Photons Percent Difference in Dose from the Two 6 MV Calibration Curves as a Function of EPID Total Pixel Value Dose (cGy) vs. Pixel Value as Determined by Three Different Methods for 18 MV Photons as a function of EPID total pixel value. Dose (cGy) vs. Pixel Value as Determined by Three Methods for 18 MV Photons Proposed Experiment T  T  vii  69 70 71 72 74  75 76 78 79 80 81 82 83 85 86 88 89 93 94 95 96 97 98 99 100 101 105  106 108  Glossary of Terms  EPID  Electronic Portal Imaging Device  TFT  Thin Film Transistor  TMR  Tissue Maximum Ratio  PDD  Percentage Depth Dose  Linac  Linear Accelerator  Z  Atomic Number  m  0  Rest Mass of Electron  aSi  Amorphous Silicon  v  Frequency of Incident Radiation/Photon  OD  Optical Density  SDD  Source-Detector Distance  SSD  Source-Surface Distance  IMRT  Intensity Modulated Radiation Therapy  IDU  Image Detection Unit  NTCP  Normal Tissue Complication' Probability  ROI  Region of Interest  viii  Chapter 1 General Introduction In a radiotherapy treatment, X-rays are used to destroy cancer cells. The main goal in radiotherapy is to deliver as much radiation to the tumor as necessary while sparing the normal tissue as much as possible. Therefore the consistency and accuracy of external beam radiotherapy treatments is essential to achieve favorable patient outcomes. Patients that have been prescribed a course of therapy may often undergo a series of planning events prior to any radiation being delivered to the diseased site. This planning may involve an initial diagnostic work-up including CT, MR or PET scan, ultrasound and x-ray simulation followed by the generation of a detailed computer calculated dose distribution that represents the clinician's desired treatment prescription: During treatment, deviations from the calculated dose distribution may occur due to patient motion, geometric set-up uncertainties, internal organ motion due to rectal and bladder filling, etc. Therapy imaging equipment such as electronic portal imaging devices (EPIDs) have been developed to address these issues by providing a simple and efficient means of monitoring daily set-up deviations over the entire course of a treatment. EPIDS produce near real-time images using the radiation transmitted through the patient during a radiotherapy treatment allowing unacceptable displacements or gross errors to be corrected before the entire daily dose of radiation is delivered. Many factors may affect the geometric accuracy of radiation treatment delivery " . Patient movement, errors in shielding block 1  9  placement and beam alignment, and shifting of internal anatomy with respect to the external geometric reference marks are the main contributors to error. Therefore, geometric verification is an important factor in the accurate delivery of radiation therapy ' and is incorporated as part of 5 10  1119  the routine clinical quality assurance program * in most radiotherapy clinics today. -1 -  In the past, geometric verification has been difficult to perform in an efficient manner. ' Imaging with kilovoltage photons was well-developed, but imaging techniques for 13 14  megavoltage photons (as used in radiation therapy) produced lower quality images. Some EPID systems produced acceptable contrast resolution, but required large doses to produce an image while other techniques produced images very rapidly, but with poor contrast. ' Conventional 13 14  portal film imaging techniques do not offer an acceptable alternative as, in addition to poor film image quality, they are not developed in real time, and a considerable amount of time and effort is required to acquire, develop, and analyze the film, thereby limiting its effectiveness in the modern radiation therapy department. These problems made geometric verification with EPIDs and film cumbersome and inefficient, with most clinics acquiring images on the first treatment day and perhaps on a weekly basis thereafter. It wasn't until the advent of the amorphous silicon aSi EPIDs that the majority of these problems have been addressed. Commercially available aSi EPIDs can acquire several images a second ' and have a more than adequate level of contrast and spatial resolution. In addition, 15 16  9  on-line registration tools that are available with the commercial systems simplify the task of image analysis and error detection. In the future, EPIDs may also function as 2D area dosimeters. The fluence measured with the EPID can be back projected through and convolved with a volumetric CT (perhaps acquired during treatment) data set to show the dose that is actually delivered to the patient. In order for the dosimetric role of an EPID to be realized, the imaging technology must be very stable, produce images with reasonable contrast and be artifact free. The aSi EPID has shown promise in the field of radiation dosimetry and has proven useful for verifying absolute and  relative dose. ' ' 15  16  17  The work presented in this thesis will investigate the fundamental dosimetric  properties of an aSi EPID.  1.1 A n Introduction to Portal Imaging Devices  In general a portal imaging device can be described as any imaging modality that captures an image of a radiotherapy treatment portal. The images obtained whether they be electronic or film based may be used to verify geometric accuracy of the radiation treatments. Many factors may affect the geometric accuracy of radiation treatment. Patient movement, errors in shielding block placement and beam alignment, and shifting of internal anatomy with respect to the external geometric reference are the main contributors to error. Geometric verification of the radiation treatment is therefore a very important quality assurance requirement. In the early days of radiation therapy, most geometric verification was done with portal film. The use of this technique can greatly reduce set up errors but there are several drawbacks in using this procedure such as high cost, poor image quality, and lack of prompt feedback due to the need for chemical processing. These drawbacks often reduced the frequency of use of films in treatment centres, and geometric verification may not have been performed as often as necessary. Electronic portal imaging devices (EPIDs) offered an exciting alternative to films! They could generate and display electronic images of the treatment portal in near real-time. In difficult treatments (i.e. treatments close to critical organs, prostate cancer treatments, mantle fields where breathing motion is significant, etc.) the intended treatment volume may move significantly between imaging and treatment so there is definitely a need for a portal imaging device that can give -3-  instantaneous feedback about the accuracy of patient set-up. Once a patient is positioned for treatment, the geometric accuracy can easily and quickly be verified with an EPID before the treatment occurs. Time varyingfieldsare now being investigated, and in order to accurately compensate for these motions, these varying fields require near instantaneous geometric feedback. On-line electronic portal imaging was developed in the 1990s and there are three main types of EPIDs in use today. These are the video based systems, scanning liquid ionization chambers, and amorphous silicon systems. Video based EPIDs (or VEPIDs) consist of a metal plate with a phosphor screen placed under the patient, a front-surface mirror to reflect the signal, and a camera used to collect the information (see Figure 1.1.1). The camera is shielded from the x-radiation and is necessarily located outside of the direct beam. The advantages of these systems are that the images can be acquired with a moderate dose. However, these systems typically have low light collection efficiency. The incident megavoltage photons produce kinetic electrons in the metal plate. These electrons travel into the phosphor layer and produce visible photons (see Figure 1.1.2) which radiate from the phosphor in all directions. Only a small fraction of the photons are detected by the camera. Another drawback of VEPIDS is they must incorporate a large optical path to shield the camera from the radiation, making them bulky and difficult to handle. Removable or collapsible models have been developed to minimize this difficulty (see Figure 1.1.3).  Schematic Diagram of a Video Based EPID Shielding Camera  Figure 1.1.1: Schematic diagram of a video based EPID (VEPID). These devices use a metal plate/phosphor screen as the x-ray detector. Light emitted by the phosphor screen is viewed by a CCD or Vidicon camera using one or more front surface mirrors.  Conversion of a Megavoltage X-ray Photon to Visible Photon  Incoming Photons from Treatment  Electrons  Scattered Photon  Acceptance cone of the lens  Fig. 1.1.2: Schematic diagram showing a part of the copper plate and phosphor layer from Figure 1.1.1. The megavoltage photon energy is transferred to an electron in the copper plate and the electron gives up its energy in the phosphor layer producing visible light. The light collection efficiency of camera-based EPIDs is low due to the narrow angle of the acceptance cone of the lens. The narrow cone shows how only a small fraction of the light emitted from the phosphor screen will actually reach the camera. The longer the optical path is, the narrower the acceptance cone of the camera will be.  Commercially Available EPIDS  Figure 1.1.3: Photographs showing the various commercially available EPIDs retracted for patient setup and deployed for image acquisition. They are: (a,b) Siemens Beamview; (c,d) Elekta iView; (e,f) I n rimed Theraview; and, (g,h) Varian Portalvision. The Varian EPID is based on an amorphous silicon array while all the other EPIDs are video based.  1.2 The Amorphous Silicon Electronic Portal Imaging Device (aSi-EPID) The PortalVision aS500 system used in this study is mounted on a dual energy (6 and 18 x  MV) 2100EX linear accelerator. It has replaced the previous scanning liquid matrix ion 1  chamber system used by Varian . The image chain consists of an image detection unit (IDU), a 13  robotic arm to position the IDU, an image acquisition system, and a computer workstation. The image acquisition system is connected to all of the other devices and reads information from the linear accelerator controller and dose rate servo. The IDU of the aS500 EPID consists of an a-Si panel, drive, read-out, and interface 1 ft  electronics, build-up material, and a plastic cover. A schematic of this system is shown below . Image Detection Unit  Interface Electronics  Read-out Electronics  Image Acquisition System  o <'  ft  m  ectronics  aSi Panel  Figure 1.2.1: Schematic Diagram of the image detection unit of the PortalVision aS500 system. 1  Varian Medical Systems, Palo Alto, CA. -8-  In the EPID the material above the aSi panel serves two purposes. It holds the detector together and converts the energy of the incident megavoltage photons through a number of processes, into lower energy photons which can be detected by the aSi layer. In human tissue, the incident high energy photons pass through and sometimes interact (see section 1.5) with electrons giving them kinetic energy. The photon fluence decreases exponentially with depth. The electrons may move a considerable distance before coming to rest, depositing their energy as they go. The electron fluence and therefore absorbed dose increase with depth until a maximum is reached. This region of an absorbing material is called the build-up region, and its depth is called the depth of maximum dose, d  max  . At depths past d  max  , the dose deposited decreases with  depth. The aSi panel consists of a 1024 by 768 array of photodiodes and thin film transistors (TFT) producing an image of 512 by 384 pixels located below several layers of material consisting of metal (1.0 mm copper), gadolinium oxysulfide phosphor, and other materials including epoxy, glass, scintillator, and paper. In the overlaying material high energy photons set electrons in motion and these electrons in turn create visible photons in the phosphor layer which are detected by the aSi panel. A pressure sensitive collision cover (touch guard) encloses the IDU for safety reasons and to keep dust out, etc. The aS500 EPID used in this study is called an indirect detection EPID which functions by first converting incident megavoltage photons into visible light through interaction with a copper plate backed with a phosphor layer of gadolinium. Then the visible photons that escape the phosphor layer are detected by the aSi substrate which integrates the incoming signal until it is triggered to read out stored charge. The drive electronics trigger the transistors row-by-row to release the charges in the photodiodes so the charges can be -9-  detected by the read-out electronics. This information is sent to the image acquisition system via interface electronics. The overlying materials will partially attenuate the beam before the photons are detected by the photodiodes in the aSi array. Since this build-up material is composed of a thin metal sheet and a layer of phosphor, photons will interact much differently with it than with human tissue. High energy photon interaction depends on energy of the photon and the atomic number of the material as described in section 1.5. Human tissue has an effective atomic number very close to that of pure water meaning that photons will interact with it in a similar manner as to how they interact with water. Other materials have also been produced that behave similarly to water when irradiated with photons and in general are referred to as water-equivalent materials or solid water. In Chapter 3 this converting layer will be compared to water-equivalent material for 6 MV and 18 MV photons and the water equivalent depth of the inherent buildup of the EPID will be determined. In Chapter 5 the EPIDs response under varying depths of solid water material will be evaluated and compared to ion chamber measurements. The data from Chapter 5 will also allow the effects of the converting material of the EPID to be compared to that of solid water. The water equivalent depth of the EPIDs inherent buildup will be determined. The aS500 EPID has several modes of image acquisition including single, multiple, and cine acquisition modes. In addition, a special research mode called IMRT verification mode was used to obtain images in this study. In this mode, image frames are rapidly acquired throughout the treatment and summed in a 16 bit memory buffer at the rate of approximately 8.8 frames per second. A single composite image is then produced by averaging the frame buffer contents and it represents the fluence delivery rate for that particular treatment. This mode is very useful for imaging treatments that have dynamically changing field shapes provided that the frame -10-  sampling rate is fast enough to adequately capture leaf motion based on the Nyquist principle . 19  The problem with using these images for dosimetry is that the average pixel value will only depend on dose rate and not on total dose. In order to get a value that consistently represents dose, the pixel values of the final image must be multiplied by the number of frames that are averaged to form the composite image. This is described in detail in Section 1.9.  1.3 Film Dosimetry  Conventionally radiographicfilmhas been used to verify patient setup accuracy in the clinical setting . In addition, film is used as a tool by the Medical Physicist for quality assurance 14  purposes and it has been found that once exposed to a source of ionizing radiation, the level of blackening of the film can be related to dose ' , although the dosimetric accuracy  is only  within ±5%. To use film effectively as a dosimeter, it must be exposed to a known level of radiation to produce a sensitometric curve (see sample in Figure 1.3.1  ) which relates the  blackening of the film (net optical density, OD) to the dose. The.blackening of thefilmor net optical density, is found by subtracting the optical density of an unexposed film from the optical density of each exposed film. As the amount of dose increases, the response of thefilmto changes in dose decreases because the film is becoming saturated, or reaching its maximum blackness (OD). For the dose range used in this study (20 to 100 cGy), the film (Kodac XV) is very responsive. A densitometer is used to measure the optical density of the film which in turn can be converted to digital form with the aid of afilmscanner. Once converted, the digital image file contains a 2D array of pixel intensity values that correspond to the differential blackening of thefilmand hence to dose through a calibration curve. The calibration curve gives  dose D as a function of the film pixel value F , and the calibration determined in this study (for 6 P  MV photons) is shown in Figure 4.3.1.  Sensitometric Curve for Film Dosimetry  4 3.5 3  i i i  JI  J i..  1 L  l  1  *•  J  ' I ' ' '  ,:.o:: •  '  h '  1 1  i  Ii  1  » l l* | | p f < |  T  4  2.5 2  1  At  x  5<V U DP IWp • W2Q0W/p G  rE3 • • ••'  *  XV Co6Q  *  XV 6 MV  * XV 23 MV ^ XV6M&V + XV 21 MeV  1 0.5  30i : :-2Stt ::j  :  MB  .'.".400:  Figure 1.3.1: A sample sensitometric curve for the Kodak-XV film used in this study for various photon energies (100 kVp, 200 kVp, Co , 6 MV, and 23 MV) and two electron energies (6 MeV and 21 MeV). 31  60  1.4 Ion Chamber In this investigation, measurements made with the aS500 EPID will be compared to those made with an ion chamber which is the standard for accurate dose measurement in the 20 21  radiotherapy field ' . The ion chamber wall is composed of a build-up cap which encloses a - 12-  small volume (usually air) that is ionized by incident radiation. The build-up cap serves to convert some of the incident photons into energetic electrons which can then pass through and ionize air within the chamber. The build-up cap is designed to be thick enough to produce an appreciable secondary electron flux yet small enough so that it does not perturb the primary photon flux significantly. In order to ensure proper measurement it is important that the presence of the ion chamber does not alter the radiation field. Small volumes do not perturb the photon flux but require large accumulation times, and large volumes acquire charge quickly but absorb many photons and perturb the photon flux, so there are many different types of ion chambers to serve different applications. A voltage is applied between two electrodes and positively charged particles will be attracted to one electrode and negative particles will be attracted to the other. An electrometer connected to the ion chamber measures either the charge produced or the current and this is directly proportional to the dose using the appropriate multiplicative factors and under controlled measurement conditions according to a protocol . 24  -13 -  Ion Chamber  Figure 1.4.1: Schematic Diagram of a Farmer Ion Chamber . In this case the aluminum rod acts as one electrode and the inside wall of the graphite acts as build-up material and the other electrode. The chamber has cylindrical symmetry about the aluminum electrode. The outer diameter of the graphite is typically around 7.0 mm. 4  Figure 1.4.1 shows a typical Farmer-type ion chamber. Many commercially available Farmer chambers are constructed similarly but may vary with respect to the materials used.  1.5 Dose The main quantity of interest in this thesis is dose which is a measure of the energy imparted to material per unit mass and has SI units of joules per kilogram (J/kg), or the gray (Gy). Dose is a quantity that is used to quantify the amount of energy delivered to matter by photons through various interactions (see section 1.6). Although absorbed dose is a fundamental quantity, it does not fully describe the effects of the radiation especially in tissue, but is nonetheless useful as the standard measure of radiation in radiotherapy. Dose is defined as the expectation value of the energy imparted to matter per unit mass at a point. -14-  \dm/ Where dE is the energy imparted and dm is the element of mass.  1.6 Photon Interactions with Matter In order for photons to deposit dose in material, they must interact with the subatomic particles in some way in order to transfer their energy to kinetic energy of subatomic particles (mostly electrons). The electrons then travel through the material until their kinetic energy has been depleted. For photon energies typical in radiotherapy (i.e. in the range 1 to 25 MV), photons interact mainly through the Photoelectric Effect, the Compton Effect, and Pair Production (each of these interactions is described in the following sub-sections). The probability of each type of interaction depends on the energy of the photon and the type of material. The attribute of an uncharged material that affects the type of interaction at megavoltage energies is the atomic number (Z) for elements or the effective atomic number for a molecular compound. The inherent buildup of the EPID is composed of several layers of material, all of which may have different effective atomic numbers that may be higher than that of human tissue. The Farmer ion chamber used in this study is designed with its sensitive volume enclosed by a carbon envelope having an effective atomic number very close to that of human tissue so that it will interact with radiation in a similar manner. The difference in materials and the Z-dependence of the interaction probability indicate that the relation of the EPID signal to that of the ion chamber may change depending on the energy of the photon. This effect may require the calibration of the EPID to include a distinct correction for every photon energy available. Varying photon energies within a field may occur due to radiation field spectral variation inherent in the linac . This is explained 21  -15 -  in the next section (1.7). Also, when a radiation field passes through a patient it may traverse different thicknesses of tissue in different regions of the field. Therefore the average photon energies in these different regions may also be different and introduce variations in detector 21  sensitivity when interacting with the EPID.  1.6.1 The Photoelectric Effect In the photoelectric effect, a photon with energy hv collides with an inner shell electron in an atom, ejecting the electron and giving it a kinetic energy of T = hv - Eb, where Eb is the binding energy of the electron in the atom. The photon no longer exists as its energy is completely transferred to the electron. The photoelectric effect is most probable for photons with energy less than 500 keV, and the probability of interaction per atom is proportional to approximately Z for high atomic number materials or Z 4  3  *  for low Z materials. The coefficient  4 8  i  3 8  per unit mass varies as Z for high atomic number materials or Z  for low Z materials  1.6.2 The Compton Effect A photon of energy hv undergoes the Compton Effect when it collides with a nearly free outer shell electron removing it from its shell. The electron is given a kinetic energy T = hv hv' in the process, where hv' is the energy of the scattered photon. The Compton Effect dominates for low Z absorbers and for medium energy photons in the range of 0.5 MeV to 4.5 MeV. The probability of interaction per atom is proportional to Z, and the coefficient per unit mass is almost independent of the atomic number of the absorbing material.  -16-  1.6.3  Pair Production  Pair production occurs when an incident photon interacts with the strong fields near the nucleus of an atom. The energy of the incident photon is converted into a positron-electron pair. The photon must have a minimum energy equal to that of the masses of the two particles, 2m c , 2  0  where m is the rest mass of the electron, and for more energetic photons, the rest of the energy 0  is converted to kinetic energy of the electron-positron pair. Triplet production can also occur in the field of an atomic electron in which case the atomic electron also gains significant kinetic energy, so that two electrons and one positron are ejected from the site of interaction. Pair production dominates for high Z absorbers and for photon energies greater than 4.5 MeV. The threshold energy is 1.022 MeV and 2.044 MeV for pair and triplet production respectively. The interaction cross section per atom for pair production is proportional to Z and the coefficient per 2  unit mass varies as Z.  1.7 O f f - A x i s S p e c t r a l D i s t r i b u t i o n  Linear accelerators accelerate electron beams to megavoltage energies. For electron therapy, these electrons may be uniformly scattered and used directly for treatment. For photon therapy, the energy of the electrons are converted into photons by placing a (high Z) tungsten target in the path of the electrons. When the high energy electrons move in the vicinity of a nucleus of the tungsten atom they will decelerate rapidly and the kinetic energy lost is converted to X-ray radiation known as Bremsstrahlung radiation. The radiation produced has a wide distribution of photon energies ' due to the fact that in each interaction, an electron may lose a 20 21  different amount of energy. The photons emanating from the target have a higher fluence in the forward direction than in other directions. In radiotherapy, it is imperative to have a beam that - 17 -  deposits a uniform dose to tissue at depth within the beam envelope. On a plane perpendicular to the central axis (see Figure 1.7.1) the center of the beam must be attenuated to a larger extent than the periphery. To achieve this, the photons are passed through a flattening filter which is constructed from a precisely machined attenuating material that compensates for the higher fluence of photons along the central axis.  The Flattening Filter  Electron Beai  Figure 1.7.1: Shape and position of the flattening filter. Although the flattening filter produces a uniform dose at depth, it creates a spectral 21  variation within the field . Higher energy photons are more penetrating. In the case of a flatteningfilter;thicker portions of thefilterwill attenuate a larger number of low energy photons. Consider measuring a small area element dA within the field. If the area dA is near the edge of thefieldthe average photon energy will be lower than if it were near the central axis. It is said that the beam is 'hardened' at the central axis. It is difficult to produce a uniform beam at all depths due to the greater penetration of higher energy photons. Therefore the flattening filter is usually designed to produce uniformity at a specified depth, for example 10 cm. A megavoltage beam interacting with water is shown in -18-  Figure 1.7.2 below. The flattening filter was designed to create a flat field at the depth db. Since the average photon energy is higher in the center of the beam, the photons at the center penetrate further into the material. Since the photons in the center are higher in energy the photons in the periphery are scattered more easily which also contributes to the higher dose in the center of the 21  field at depth . For these reasons the dose at large depths will decrease as distance from the central axis increases as shown by the profile for d in Figure 1.7.2. At the edge of the field the c  photons do not penetrate material as easily so in order to have a certain dose at depth the intensity must increase as distance from the central axis increases at depths shallower than db as shown by the profile for d in Figure 1.7.2. Dose Profiles at Depth a  I  Off-axis Position  Figure 1.7.2: Sketches of typical dose profiles for a single beam at different depths in tissue equivalent material. The effect is exaggerated for illustrative purposes. The flattening filter was designed to provide aflatfield at the depth db. The profiles are normalized to 100% on central axis.  -19-  1.8 Tissue-Maximum Ratio (TMR) The tissue maximum ratio (TMR) is defined as the ratio of the dose in water at depth d to 21  that at d  max  for a fixed source-detector distance. Tissue Maximum Ratio  Figure 1.8.1: The TMR is defined as the ratio of the dose at depth d to that at depth d  max  .  TMR measurements were performed with the EPID by placing solid water on the detector surface which was held at a constant source-detector distance (SDD) of 130 cm. The solid water was added in small increments in order to determine the depth of maximum dose. Under these conditions the dose to the detector will increase until d will fall off due to the attenuation of the primary photons.  -20-  max  is reached after which it  1.9 Dose at Extended SDD The linac is typically calibrated so that the number of monitor units (MU) is equal to the dose in cGy received at the depth of maximum dose (d ) at a detector distance of 100 cm and a max  reference field size of 10 x 10 cm . Under these conditions the dose D can be written as 2  D=£(MU)  [1.9.1]  where K is 1.00 cGy/MU and MU is the number of monitor units. All experimental measurements are acquired at a source-detector distance (SDD) of 130 cm and a reference field size of 10 x 10 cm . The surface of the EPID is treated as the detector in this investigation. Since 2  the SDD is different from the calibration conditions, many additional factors will be accounted for when calculating the dose from the monitor units such as scattering effects fromfieldsize changes, depth of measurement (TMR), and the inverse square law. Thefieldsize at the detector can be calculated using the geometry of Figure 1.9.1.  -21 -  Beam Divergence d = 130.0 cm 2  Soi  s = 13.0 cm 2  Figure 1.9.1: Since the beam diverges, the field size will increase as the distance from the source increases. There are two similar isosceles triangles in Figure 1.9.1. One triangle has base si and height di and the other has base s and height d . Since the triangles have the same angles the 2  2  sides are proportional and s = d sj/di. The field size s for this experiment at a distance of 130 2  2  2  cm can be calculated as follows. (130cw)(10cw) d  100cm  = \ 3cm  [1.9.2]  The field size at the detector in this experiment is then 13x13 cm . X-rays that interact 2  with the solid water scatter, and as the field size increases, the amount of scattered radiation increases thereby increasing the dose to the center of the field. Since the field size reaching the detector is larger than at the calibration distance, the dose will be increased by some factor. This -22-  factor is called the phantom scatter ratio, S , and is a function of the field size at the point of p  measurement. Beam data acquired during the commissioning of the linac shows the normalized peak scatter factor to be 1.010 and 1.018 for a 6 and 18 MV beam respectively and a 13 x 13 cm  2  field size. For the purposes of this experiment the phantom scatter ratio and the peak scatter factor are essentially equal. Scattered radiation from the head and collimator of the linac also influences the dose received. Much of the radiation from the source is blocked by the collimators and as they change positions, the amount of scattered radiation from them changes. The ratio of the dose received with the collimators at a certain setting to the dose received when the collimators are set at the reference field size (lOx 10 cm )is called the head-scatter factor or the collimator-scatter factor S . S depends only on the field size at isocenter (100 cm). For all measurements in this c  c  investigation the collimator will be set for a 10 x 10 cm field at isocenter so the value of S will 2  c  be 1.00. As the photons in the beam diverge the intensity of the beam decreases according to the inverse square law for point sources of radiation. This states that the intensity of the radiation is inversely proportional to the square of the distance from the source, when there is no absorption. Due to inverse square effects the dose, D2, to a small mass of tissue along the central axis at the distance d , is related to the dose, Di, at the distance di from the source in the following way. 2  d| 2  d  2  2  o o o ^ 1  (130cm)  2  5 1  9 1  ] L  J  The dose will also depend on the depth in tissue. The tissue maximum ratio (see section 1.8) takes this effect into account. If all of these factors are combined into one equation we obtain the following . 21  -23 -  D = K(S XS )(MU)(TMR)^L C  [1.9.4]  P  Where d is the SDD in cm, which is 130 for all measurements in this experiment. Putting the values into equation 1.9.4 gives the following equations for calculating the dose for both 6 and 18 MV photons. D  6 M V  = (1.00cGv/ML/)(1.00)(1.010)(MU)(rMi?)(0.5917) D  D  1 8 M V  6 M V  =(0.5976cGy/MU)(UU)(TMR)  [1.9.5]  = (1.00cGv/Mc/)(1.00)(1.018)(MU)(rMi?)(0.5917) D  1 8 M V  = (0.6024cGv/Mc/)(MU)(rM£)  [1.9.6] s  -24-  1.10 M e a s u r i n g D o s e w i t h t h e E P I D  The first step in the investigation of the dosimetric properties of the EPID involves developing a relationship between the pixel intensity values recorded by the EPID and dose delivered by the accelerator at the same position. The reproducibility and consistency of this relationship must also be tested in order to determine the effectiveness of using an EPID for dosimetry. Linacs can be programmed to deliver a pre-determined dose of x-rays at a specific dose rate; however a sub-optimally tuned linac may deviate from the intended parameters. For example, the dose rate may attempt to lock to the programmed dose rate but over-shoot the intended target. A feedback loop incorporated into the control electronics of an accelerator will try to compensate for this over-shoot. In the process, it may under estimate the desired target at which point it will try to compensate in the opposite direction. In this way the dose rate fluctuates about the intended dose rate for a short period until the feedback loop can adequately compensate. Figure 1.10.1 shows a rough sketch of accelerator output during the first few milliseconds after an irradiation is initiated. Usually the dose rate stabilizes within a fraction of a second but for a sub-optimally tuned linac there may be times when an entire treatment is delivered without the dose rate ever reaching the target value. The output as a function of time may also vary for exact repeated irradiations. During this study, a sub-optimally tuned linac showed a dose rate variation of up to 25 MU/min from the set dose rate. One of the objectives of the experiment was to determine the EPID response at several different pre-programmed dose rates. The use of a sub-optimally tuned linac introduced an added variability commonly encountered in the clinical setting.  -25-  Dose Rate Instability  Set dose rate  Time  Figure 1.10.1: Sketch of dose rate as a function of time with variance exaggerated for illustrative purposes. For the treatment unit used in this study, the time scale was such that t , the time to reach the first peak, varied from a fraction of a second to several seconds. a  During this study it was not uncommon for the dose rate to vary about the intended target (100 MU/min to 600 MU/min) by up to 25 MU/min. In order to accurately determine the total dose, DT, for an irradiation such as that shown in Figure 1.10.1, the dose rate must be integrated over time. This would be equal to the product of the average dose rate, (3D/5t) , and the total ave  time of irradiation, T. D =  [1.10.1]  T  The average rate is determinable from the data stored by the acquisition program. The EPID acquisition mode used in this investigation gives a final image that is representative of the average dose rate. The imager acquires 8.8 frames per second until the irradiation is complete and then the average value for each pixel is computed and a single image is produced. In order to acquire pixel values consistently from the center of the field a program is written to find the center of a squarefieldfor each portal image and the average pixel value for a 5 by 5 region of interest (ROI) in the center of the field. The center of the field is found by starting along the x-direction in the center row of the array (horizontal dotted line in figure 26-  1.10.2), the two adjacent pixels with the largest positive difference in signal are found. The position of the left field edge is then set to be midway between these pixels. The position of the right field edge is set to be in between the two adjacent pixels with the largest negative difference in signal (see figure 1.10.2).  Determination of Left and Right Field Edges High Signal  Low Signal  Figure 1.10.2: The left and right field edges are determined by examining the center row of pixels and finding the locations of the largest signal change. The inset on the right shows that the edge of the field is set to be between the two pixels with the largest signal difference. The average of the edge positions is set to be the x-value of the center of the field. A column of pixels along this x-value (vertical line in Figure 1.10.3) is examined to determine the two sets of adjacent pixels with the largest signal difference. The top and bottom field edges are determined in a similar manner to the left and right field edges. The y-value of the center of the field is then set to the average of the top and bottom edge y-values.  -27-  Determination of Top and Bottom Field Edges  Image of Field  Large Signal Difference  Figure 1.10.3: The top and bottom field edges are determined by examining the vertical column of pixels at the x-value of the center of the field and finding the locations of the largest signal change (or maximum gradient). If the field were perfectly aligned with the detector then the values found above would indeed be the center of the field. However, if the field edges are not perfectly parallel and perpendicular to the EPIDs pixel rows and columns, the values found above may not be in the center of thefield.Therefore this center-finding process is repeated until the algorithm repeatedly returns the same x and y values for the center of thefield(see Figure 1.10.4).  -28-  Determination of the Center of a Non-aligned Field  Figure 1.10.4: After applying the center finding algorithm once (a,b) the center of the field is not necessarily found. Numerous iterations of the center finding process are required to locate the center of the field with accuracy. Once the determined center position is repeated the center is found and the algorithm terminates. Once the central pixel is located, the five by five grid of pixels which it centers is located. The average value of these pixels is determined and called the average pixel value, P, for the remainder of this paper. The average pixel value, P, is representative of the average dose rate, and several studies ' , have found the response of the EPID to be proportional to the dose. 1516  25  Poc  'dD^  [1.10.2]  The EPID acquisition mode used in this investigation captures images into a 16 bit frame buffer at a rate of 8.8 frames per second. The final image displayed to the operator is an average of all the images collected in the frame buffer. The rate of frame acquisition is constant at 8.8  -29-  frames per second, so the number of frames, N , is proportional to the total time, T, of the irradiation. NcxT  [1.10.3]  Combining equations 1.10.2 and 1.10.3 with 1.10.1, we can see that the total dose, D , is T  expected to be proportional to the product of the average pixel value and the number of frames. D oc PN T  [1.10.4]  For this investigation, the product on the right hand side will be called the total pixel value, Pj. D ocP T  r  [1.10.5]  Introducing a constant of proportionality k gives, D =kP T  T  [1.10.6]  In chapter 2 the constant k is measured for 6 and 18 MV photons under conditions of changing fluence rate and in chapter 6 under conditions of changing fluence. In chapter 5 under conditions of changing depth, it is found that the best relationship between dose and pixel value is not linear. Since the total pixel values are obtained by multiplying the average pixel value by the number of frames (an integer) it is expected that the random error could be reduced by using a lower dose rate. A lower dose rate would necessitate a longer treatment time and therefore a larger number of frames would be used in the composite image. Of course all treatments don't require exactly an integer number of frames, so by using an integer number of frames an error is introduced that would be expected to be random. The maximum error due to this should be roughly half a frame and should not depend on the length of the treatment. Therefore the  -30-  percentage error due to this effect is expected to be lower if a lower dose rate setting is used during image acquisition. An error may also be introduced due to the placement of the imager. A change in SDD will result in a detection error that can be approximated by the inverse-square law. In doing this the effect on scattering due to field size change is being neglected. This approximation predicts a 1.5% decrease in the amount of radiation that the imager receives from apositional error of+1.0 cm at an SDD of 130 cm. The imager is easily positioned to within ±0.1 cm, therefore the most significant source of error in taking dose measurements with the EPID will most likely be the variance of the response of the EPID.  -31 -  Chapter 2  Testing the dose rate stability of an amorphous silicon portal imaging device 2.1 Introduction The goal of this investigation is to test the stability of the pixel values of an a-Si EPID at various dose rates. The total pixel value (as described in section 1.10) of a 5 x 5 ROI in the center of the field was calculated on several images acquired at dose rates ranging from 100 MU/min to 600 MU/min in steps of 100 MU/min. The sensitometric response of the EPID (i.e. the relationship between dose and pixel intensity) under these conditions was then calculated and the value of the constant, k, in equation 1.10.6 determined for 6 and 18 MV photon beams. Therapy accelerators can be programmed to deliver a pre-determined dose whereby the operator simply has to key in the desired number of monitor units and select a suitable dose-rate and beam energy. When the beam-on button is activated, the accelerator will deliver the radiation at the selected settings. Although the linac is set to deliver the radiation at a specific dose-rate, it rarely runs at exactly the entered value. This may be due to a sub-optimally tuned linac or due to inherent limitations in the control electronics. During the data collection for this experiment, it was noted that a variation of ±25 MU/min from the set dose rate was not uncommon. The dose rate usually starts low and then varies about the intended value before stabilizing (see section 1.10). It was also noted that repeated exposures under identical linac settings often produced varying beam-on times, an indirect indication that a fluctuation in the dose rate was present.  -32-  2.2 Method The measurements for this study were performed on a dual energy (6 and 18 MV) CL2100EX accelerator equipped with a PortalVision aS500 amorphous silicon EPID. It is 1  1  assumed the EPID will be used with the touch cover in place, and that the easiest way to position it will be with reference to the top surface of the cover. For these reasons, the EPID is treated much like a black box in this study (and the entire thesis), and the top surface is considered to be the detector position. All measurements are made at a SDD of 130 cm, with a field size of 10 x 10 cm at isocenter (100 cm for this linac), as shown in Figure 2.2.1, giving a measured field size of 13 x 13 cm at the level of the imager. The programmed dose in all cases is 100 MU, and 1.5 2  and 3.2 cm of solid water is used for 6 and 18 MV photons respectively, and serves as a build-up layer for the a-Si images.  Experimental Positioning  Solid Water  100 cm j  130 cm  Figure 2.2.1: Sketch ofexperimental set-up for chapter 2.  Varian Medical Systems, Palo Alto, CA. -33 -  Images of the field were acquired at dose rates ranging between 100 and 600 MU/min in increments of 100. A minimum of 9 images were acquired at each dose rate using the IMRT verification mode on the EPID. In order to streamline the analysis process, a MATLAB program was developed to automatically find the average pixel value as described in section 1.10. For each irradiation, the number of frames averaged is multiplied by the average pixel value to get a number representative of the total average pixel value, PT. The constant k in equation 1.10.6 which relates dose to total pixel value can then be determined.  -34-  2.3 Results Tables 2.3.1 and 2.3.2 below show the average, maximum and minimum, and the standard deviation of the total pixel values, PT, measured for each dose rate, and for all dose rates for both energies produced by the linac. Experimental Results for 6 MV Photons Dose Rate (MU/min)  PTave  Max  Min  o~  100 200 300 400 500 600 Average for 6 MV  420416 421225 419308 417262 413518 415392 417510  423650 421968 423000 419808 419540 418582 423650  413996 419889 408884 415944 396831 410269 396831  3581 749 5330 1243 8179 3143 5293  a for 6 MV (as a percent)  2746 (0.66%)  1885 (0.44%)  7282 (1.84%)  Table 2.3.1: Statistics illustrating the reproducibility of the EPID pixel values with varying dose rates for 6 MV. At least nine measurements were made at each dose rate. Experimental Results for 18 MV Photons Dose Rate (MU/min)  PTave  Max  Min  a  100 200 300 400 500 600 Average for 18 MV  430591 425631 426365 426312 417924 425014 425608  430920 428076 426870 426591 426126 425224 430920  430272 417984 425661 426033 401760 424672 401760  212 3958 471 194 10394 182 5408  a for 18 MV (as a percent)  3758 (0.88%)  1829 (0.42%)  9361 (2.33%)  Table 2.3.2: Statistics illustrating the reproducibility of the EPID pixel values with varying dose rates for 18 MV. At least nine measurements were made at each dose rate. -35 -  During data acquisition at 200 MU/min and 500 MU/min at 18 MV, the observed dose rate varied wildly which most likely caused the high variation of values indicated by the large standard deviations for these settings. Even though the standard deviation seems high for 500 MU/min, it is only 2.5% of the average total value, and half of the spread (max-min) is 2.9% of the average. The best results were obtained when the delivered dose rates closely matched the set dose rate. At 300 MU for example, the delivered and set dose rates matched closely, and the standard deviation was only 0.1% of the average and half of the spread was 0.14% of the average. Overall the delivered dose rate varied significantly from the set dose rate and for all 18 MV data, half of the spread was 3.4% of the average, and the standard deviation was only 1.3% of the average. For 6 MV, the worst spread also occurred at the 500 MU/min setting and half of it is 2.7% of the average. The most reliable setting was 200 MU/min where half of the spread was only 0.25% of the average. For all 6 MV data half of the spread is 3.2% of the average, and the standard deviation is 13% of the average. It is interesting that the 500 MU/min setting gave the greatest variation in values for both photon energies. This may indicate an inconsistency with radiation delivery at this setting. Note the difference in the values for the 6 MV and the 18 MV photons. While the standard deviations for the 6 MV and 18 MV photons were approximately 5300 and 5400 respectively, the averages for the two beams differed by more than 8000, i.e. approximately 2%. Since the acquisition mode of the EPID is different for 6 MV and 18 MV this may be due to a difference in gain settings, but since the Linac is calibrated using an electrometer these results  -36-  may also indicate a difference between the responses of the EPID and the electrometer to varying photon energy. This is investigated further in Chapter 5 of this thesis. An inaccuracy in SDD will introduce an error that can be roughly approximated by the inverse-square law. A rigorous error analysis must take into account phantom and head scatter factors since the field size would also change, but this effect is negligible. For example, at an SDD of 130 cm, a positional error of+1.0 cm results in a 1.5% decrease in the amount of radiation that the imager receives according to the inverse-square law. For the +1.0 cm positional error the field size would increase to 13.1 x 13.1 cm and the phantom scatter factor would 2  increase from 1.018 to 1.019 for 18 MV photons and from 1.010 to 1.011 for 6 MV. The EPID is easily positioned with an accuracy of 1.0 mm so these errors are negligible. As mentioned earlier there are also errors in radiation delivery. With all possible sources of error it is encouraging that the standard deviation of the total pixel values for both photon energies was only 1.3%. The dose reaching the surface of the EPID on the central axis can be calculated by using equation 1.9.5 for 6 MV and equation 1.9.6 for 18 MV as follows. As stated previously 100 monitor units is used for all irradiations and the measurements are taken at d  max  (TMR = 1.000).  The inherent build-up of the imager is not considered here as the EPID is being treated as a black box. D D  6 M V  I 8 M V  = (0.5976cGv/Mt/)(100MU)(1.000) = 59.76cG>  [2.3.1]  =(0.6024cGv/ML/)(100MU)(1.000) = 60.24cGv  [2.3.2]  The constant k in equation 1.10.6 can be determined by dividing the doses calculated above by the total pixel values in Tables 2.3.1 and 2.3.2. The constant is calculated for each photon energy and dose rate and shown in Table 2.3.3 and 2.3.4. In the tables the unit pix refers to the total pixel value, PT, as defined in section 1.10. -37-  Dose (cGy) per Total Pixel Value for 6 MV Dose Rate (MU/min)  Average Constant k (x 10")  Minimum Constant k (x 10 )  Maximum Constant k (x 10")  (cGy/pix) 1.421 1.419 1.425 1.432 1.445 1.439 1.431  (cGy/pix) 1.411 1.416 1.413 1.424  (cGy/pix)  4  100 200 300 400 500 600 Average for 6 MV  4  1.424 1.428 1.411  Stated k Value (10" cGy/pix) 4  4  1.44.4  1.42 1.429 1.43 1.43 1.45 1.44 •1.44  1.423 1.462 1.437 1.506 1.457 1.506  ±0.02 ±0.004 ±0.03 ±0.01 ±0.04 ±0.02 ±0.05  Table 2.3.3: Values of the calibration constant, k, for 6 MV photons with confidence intervals for every dose rate. Dose (cGy) per Total Pixel Value for 18 MV Dose Rate (MU/min)  Average Constant k (x 10")  Minimum Constant k (x 10")  Maximum Constant k (x 10")  (cGy/pix) 1.399 1.416 1.413 1.413 1.442 1.418 1.416  (cGy/pix) 1.398 1.407 1.411 1.412 1.414 1.417 1.398  (cGy/pix) 1.400 1.442 1.415 1.414 1.499 1.419 1.499  4  100 200 300 400 500 600 Average for 18 MV  4  Stated k Value (10" cGy/pix) 4  4  1.399 1.42 1.413 1.413 1.44 1.418 1.42  ±0.001 ±0.02 ±0.002 ±0.001 ±0.04 ±0.001 ±0.05  Table 2.3.4: Values of the calibration constant, k, for 18 MV photons with confidence intervals for every dose rate. The k values for 6 MV and 18 MV are equal within experimental error. The constants are within 3% of 1.42 x 10" cGy/pix for 18 MV and 1.44 x 10" cGy/pix for 6 MV where pix refers 4  4  to the total pixel value as defined in section 1.10.  -38-  2.4 Conclusion The amorphous silicon portal imaging device used in this study gives clinically useful dosimetric measurements under conditions of varying dose rates. For 6 MV the spread of total pixel values was only 6.4% for a set of 53 acquired doses. The standard deviation of measured total pixel values for all 6 MV was only 1.3% of the average value. At 18 MV, the spread of total pixel values was 6.8% of the average and the standard deviation was 1.3% for 52 values. With these reliabilities, the amorphous silicon portal imaging device is a promising candidate for taking IMRT treatments to the level of dosimetric verification. The constant in equation 1.9.6 that relates the EPID pixel values to dose was determined to be 1.42 x IO" cGy/pix for 18 MV and 1.44 x 10" cGy/pix for 6 MV with 95 % confidence 4  4  intervals of 0.04 x IO" cGy/pix. Therefore, dose (cGy) may be written as a function of EPID 4  total pixel value as follows. D = 6MV  (1.44x10~*)P  T  D =(L42x\0-*)P imy  r  [2.4.1] [2 A.2]  The 95% confidence interval is approximately 3% and the standard deviation was found to be 1.3%. These values are not sufficiently accurate to be used as the primary monitor, but they are accurate enough that the EPID could be used as a safeguard against gross errors in dose distribution.  -39-  Chapter 3 Effective build-up of the PortalVision aS500 3.1 Introduction The aS500 EPID used in this study is an indirect detection EPID. The construction of indirect detection amorphous silicon (aSi) EPIDS (Section 1.3) is such that an inherent buildup material is always present above the sensitive layer of the detector (see Figure 3.1.1 below).  Build-up Material of the EPID Touch Guard i  Air Gap  ,  y Copper Plate Gadolinium Oxysulfide Phosphor aSi Substrate  Figure 3.1.1: Sketch of the build-up material of the EPID showing relative positions of the aSi substrate, copper plate, gadolinium oxysulfide layer, and the touch cover. The detecting layer in this study is a 512 by 384 array of photodiodes and thin film transistors (TFT) embedded in build-up material which is composed of many layers of different materials including metal (1.0 mm copper), gadolinium phosphor, epoxy, glass, scintillator, and paper. This conglomerate of materials that overlap the aSi substrate converts incident megavoltage photons into visible light through interactions with a copper plate backed with a phosphor layer. Photons of therapeutic (high) energies can penetrate a considerable distance into the material before interacting with the material, and the dose deposited increases with depth -40-  until the attenuation of the photon fluence causes the dose to decrease with depth. The region in which the dose is increasing with depth is called the build-up region. The visible photons that escape the phosphor layer are detected by the aSi substrate. A pressure sensitive collision cover (touch guard) encloses the IDU and provides a small additional source of build-up. As dose measuring instruments can be calibrated for any positioning convention, in this thesis the outer surface of the touch cover of the EPID is considered to be the position of the detector as this is the most practical way to set-up and use the detector. In this way the EPID is treated somewhat as a 'black box'. Figure 3.1.2 below shows the relative position of the EPID and the ion chamber for a measurement at the same depth. Detector Positioning L qin\alcnt Water  Ion ChamberI-.quivaleni \V;itor  Figure 3.1.2: Relative positions of the EPID and Ion Chamber for a same-depth measurement. The EPID and ion chamber will be placed under varying depths of a material commonly referred to as equivalent or solid water, which is designed to interact with photons similarly to water and human tissue. Note that the center of the ion chamber is placed at depth whereas for the EPID, the top surface is placed at depth. The actual detecting layer is below inherent build-up of the EPID. As a step towards understanding the dose characteristics of the EPID, the waterequivalent depth of its inherent build-up is determined for each energy mode (6MV and 18MV) in this chapter.  -41 -  3.2 M e t h o d Images for this study w e r e a c q u i r e d o n an a S 5 0 0 E P I D m o u n t e d o n a d u a l energy (6 a n d 18 M V ) C L 2 1 0 0 E X accelerator. E P I D measurements are made w i t h the outer surface o f the t o u c h guard p o s i t i o n e d at a source-detector distance ( S D D ) o f 130 c m , whereas the i o n c h a m b e r is p l a c e d s u c h that its center is at 130 c m . T h e f i e l d size remains constant at a setting o f 10 x 10 2  2  c m , w h i c h gives a f i e l d s i z e o f 13 x 13 c m at 130 c m . T h e dose setting is 100 M U a n d the dose rate is 300 M U / m i n for a l l measurements. T h e E P I D s p o s i t i o n is h e l d constant a n d water equivalent m a t e r i a l is p o s i t i o n e d o n the surface o f the t o u c h guard. T h e E P I D cannot support m u c h w e i g h t w i t h o u t m o v i n g o r h a v i n g the t o u c h guard activate so the s o l i d water is supported b y the treatment c o u c h . I n the 6 M V m o d e p o r t a l i m a g e s are o b t a i n e d w i t h depths r a n g i n g f r o m 2.0 m m to 10.0 m m a n d i o n c h a m b e r data is a c q u i r e d for depths f r o m 10.0 m m to 30.0 m m . In the 18 M V energy m o d e i m a g e s w e r e r e c o r d e d for depths f r o m 10.0 m m to 30.0 m m o n the E P I D a n d i o n c h a m b e r data w a s a c q u i r e d for depths f r o m 31.0 m m to 55.0 m m . A s expected, the s i g n a l f r o m b o t h instruments increases as m a t e r i a l is p l a c e d above u n t i l a certain depth w h e r e the s i g n a l reaches a m a x i m u m . A f t e r the m a x i m u m is reached, the s i g n a l and dose r e c e i v e d decreases w i t h depth. I o n c h a m b e r measurements are k n o w n to be p r o p o r t i o n a l to dose r e c e i v e d , so the depth w h e r e the m a x i m u m i o n c h a m b e r s i g n a l o c c u r r e d w i l l be the depth w h e r e the m a x i m u m dose is imparted to the m a t e r i a l . T h i s is c a l l e d the depth o f m a x i m u m dose, or d  m a x  . A s the E P I D has some inherent b u i l d - u p , it g i v e s its m a x i m u m signal  w i t h less extra m a t e r i a l p l a c e d o n top (see F i g u r e 3.2.1).  -42-  Relative Depths of Maximum Signal Ion Chamber  !  Figure 3.2.1: Amount of material required for the EPID to give maximum signal, d , is less than that required for the ion chamber, d3. 2  The difference of the two depths of maximum signal is shown in Figure 3.2.1 as di. This can be considered as the amount of water equivalent material that the inherent build-up of the EPID is contributing. In this chapter, the value of di as shown in Figure 3.2.1 will be determined for each energy mode.  -43 -  3.3 Results  Plots of detector signal versus depth are shown below in Figures 3.3.1 to 3.3.4. The plots are similar to TMR (see sectionl .8) plots, but they are not normalized to have a maximum of 1.0. For each case the signal increases with distance to d  max  and then decreases exponentially with  distance. Figure 3.3.1 below shows the EPIDs total average pixel value as a function of depth for 6 MV photons.  Total Pixel Value vs. Depth for 6 MV 557000 - r -  552000 -r 0  0.1  0.2  0.3  0.4  0.5  0.6  0.7  0.8  0.9  1  1.1  Tissue Depth (cm)  Figure 3.3.1: Total pixel value measured by the EPID versus depth of equivalent water (cm) placed on top for 6 M V photons. The best polynomial fit to the data in Figure 3.3.1 is y = 479291.9775x 6  1718687.6953x + 2446033.8857x - 1730830.9939x + 613374.0097x - 97949.8229x + 5  4  3  2  :  561337.6778, where y is the total pixel value and x is the tissue depth. The r-squared value for this fit is 0.9996. By looking at the graph and finding the zeros of the derivative of the -44-  polynomial fit, the maximum value occurs at 0.36 + 0.03 cm. The ion chamber measurements for 6 MV photons are shown below in Figure 3.3.2.  Ion Chamber vs. Depth (cm) for 6 MV. 14.8  - 14.7 U)  c  T3 to 8.  • 14.6 E ra  x: O  * 14.5 4747x + 2 R = 0 9876 2  14.4  i  -  1.4  1.2  1.6  1.8  2  2.2  2.4  2.6  2.8  Depth (cm)  Figure 3.3.2: Ion chamber readings versus depth of equivalent water (cm) placed on top for 6 MV photons.  The polynomial fit to the data in Figure 3.3.2 is y = 0.4397x - 4.9752x + 22.8223x 6  5  4  53.7897x + 67.2599x - 40.4747x + 22.8314, and has an r-squared value of 0.9876. From 3  2  examining the graph, it seems the reading error for depth of maximum dose is ±0.03 cm. By finding the zero of the derivative, d  max  is found to be 1.74 cm. The stated value is 1.74 ± 0.03 cm.  The EPID average pixel values versus depth for 18 MV are shown in Figure 3.3.3 below.  -45-  Total Pixel Value vs. Depth for 18 MV 556000  Depth (cm)  Figure 3.3.3: Total pixel value measured by the EPID versus depth of equivalent water (cm) placed on top for 18 MV photons.  The equation of best fit for figure 3 is y = -1464.7709x + 20598.3248x - 118453.7420x 6  5  + 356943.5677x - 600880.5316x + 547613.3175x + 336357.7008. It has anr-squared value of 3  2  0.9746. The zero of the derivative of this function is 2.28 and the stated depth of maximum detector signal is 2.28 ± 0.05 cm. The ion chamber readings for 18 MV photons are plotted in Figure 3.3.4 below.  -46-  4  Ion Chamber vs. Depth (cm) for 18 MV 15.2  Depth (cm)  Figure 3.3.4: Ion chamber readings versus depth of equivalent water (cm) placed on top for 18 MV. The best fit for the data in Figure 4 is given by y = 0.0604x - 1.4790x + 14.9557x 6  5  4  79.8479x + 237.1616x - 370.8746x + 252.9182 and it has an r-squared value of 0.9886. The 3  2  depth of maximum ion chamber signal is found to be 4.52 ± 0.05 cm. The data in this part of the experiment appears to have some more outliers from the best fit curve than other trials. Because of this the experiment was run four times and the depth of maximum reading was consistent within experimental error even though there were outliers in every trial. By subtracting the depth of maximum signal of the EPID from that of the Ion Chamber the effective water equivalent depth of the inherent build-up material on the EPID is calculated as follows. -47-  For 6 MV  1.74 ± 0.03 cm - 0.36 ± 0.03 cm = 1.38 ± 0.06 cm  For 18 MV  4.52 ± 0.05 cm - 2.28 ± 0.05 cm = 2.24 ± 0.10 cm  The inherent build-up is found to be 1.37 ± 0.06 cm for 6 MV and 2.24 ± 0.10 cm for 18MV. Potential errors in this experiment include variation of the linac output, variation of the EPID response, variation of the ion chamber response, and positioning error as described in Chapter 1.  Conclusion  The inherent build-up of the portal imaging device is found to be 1.37 ± 0.10 cm for 6 MV and 2.24 ± 0.06 cm for 18MV. These values will be added to the depth of solid water when the EPIDs signal versus depth is investigated later in this paper. As expected, the results show that the water equivalent depth of the inherent build-up of the EPID is different for different photon energies. The fact that the water equivalent depth of the inherent build-up is a function of photon energy creates a complex situation for modeling the build-up of the EPID during treatment. During radiotherapy the treatment beams will be passing through various thicknesses of material before they reach the EPID. Furthermore, the linac photon treatment beams are composed of a wide distribution of photon energies, and since low energy photons are attenuated more than high energy photons, the radiotherapy photon beam hardens (becomes higher in average photon energy) as it passes through material. Even in a single field then, the build-up of the EPID may have a different water equivalent thickness in different regions. This justifies the concept of calibrating the EPID for dose in such a way as to treat the surface of the touch guard as the  -48-  /  detecting layer and ignoring the actual position of the detecting a-Si layer and the characteristics of the build-up layers.  -49-  Chapter 4 A Comparison of 6 M V Enhanced Dynamic Wedge Profiles Acquired with Film and the EPID  4.1 Introduction Intensity modulated radiation therapy (IMRT) involves the dynamic adjustment of field size and shape during treatment to produce a complex dose distribution within thefield.Portal film dosimetry is often used to verify the dose distribution in the field for these types of treatments since films produce 2D fluence maps of the treatment field unlike ion chambers which can only produce point measurements. The disadvantages of using film are its poor quality and increased workload required to process the film. EPIDs have the added advantage that they give instantaneous feedback in digital form and can easily be incorporated as part of the patient's plan. As EPID technology continues to improve, it may be possible one day to verify IMRT treatments both geometrically and dosimetrically in real time. In this study, EPID and film images are compared for enhanced dynamic wedge (EDW) 6 MV photonfields.A dynamic wedge treatment necessarily requires one of the collimator jaws to slowly sweep across the field delivering a dose distribution similar to that produced by a physical wedge (see Figure 4.1.1). Wedges are accessories that can be added to a setup that modifies the beam in such a way as to make it more suitable for therapy where a non-uniform beam is desirable (ie. To obtain a uniform dose distribution from beams that intersect at angles). They are designed to change the dose distribution of thefield,thereby angling the isodose curves with respect to the central axis as shown in Figure 4.1.1b. The wedge is defined by the angle by which the isodose lines are rotated (see Figure 4.1.1) at the central axis at a specified depth. There is no general agreement as to the choice of reference depth, but it has been recommended that a depth - 50 -  of 10 cm be used universally . The flattening filter in the linac treatment head is designed such 21  that the isodose distribution for an open field is approximately flat and normal to the central axis as shown in Figure 4.1.1a. Definition of Wedge Angle  Figure 4.1.1: Diagrams taken from Khan, The Physics of Radiation Therapy for Co radiation. The diagram on the left (a) shows isodose lines from an open field. The diagram on the right (b) shows isodose lines of a 45 degree wedged field which are at a 45° angle to the central axis. 21  60  4.2 Method A calibration curve needs to be determined to relate the optical density (OD) of the film (Kodak XV) to dose for the film measurements. Thefilmswere calibrated by placing them on the treatment couch at a source-detector distance (SDD) of 100 cm. Films were irradiated using a -51 -  6 M V p h o t o n b e a m , a c o l l i m a t o r j a w setting o f 10 x 10 c m , a n d doses r a n g i n g f r o m 2 0 c G y to 2  100 c G y d e l i v e r e d at 300 M U / m i n . F o r 6 M V these doses are d e f i n e d i n water at 1.5 c m depth. Therefore 1.5 c m s o l i d water b u i l d u p w a s above the f i l m a n d 10.0 c m o f s o l i d water w a s u s e d as backscatter m a t e r i a l . T h i r t y i r r a d i a t i o n s (and f i l m s ) were used for the c a l i b r a t i o n . A f t e r d e v e l o p i n g the i m a g e s , they w e r e scanned u s i n g a densitometer w h i c h determines the o p t i c a l density ( O D ) at r e g u l a r l y spaced points o n the f i l m g i v i n g a t w o - d i m e n s i o n a l g r i d o f f i l m p i x e l values Fp as a f u n c t i o n o f p o s i t i o n . T h e f i l m p i x e l size is d e t e r m i n e d b y the densitometer ( V I D A R ) used. T h e 5 x 5 p i x e l s centered o n the central a x i s w e r e u s e d as the r e g i o n o f interest (ROI). T h e subsequent ( K o d a k X V ) f i l m a n d E P I D i m a g e s w e r e a c q u i r e d w i t h 15, 30, 4 5 , a n d 60 degree enhanced d y n a m i c w e d g e s u s i n g 6 M V photons at dose settings o f 100 M U a n d 70 M U , S D D o f 130 c m , dose rate o f 300 M U / m i n , a n d f i e l d size setting o f 10 x 10 c m . T h e t w o 2  dose settings w e r e u s e d to observe differences i n response to dose. T h e field size at the p o s i t i o n o f measurement (130 c m ) is 13 x 13 c m . A d d i t i o n a l b u i l d u p w a s p r o v i d e d for the film a n d E P I D 2  i n the f o r m o f 1.5 c m o f s o l i d water (the depth the l i n a c is c a l i b r a t e d w i t h ) to a l l o w m o r e •  accurate film d o s i m e t r y  •  22  •  . W e d g e a n d f i l m profiles (a d i s t r i b u t i o n t h r o u g h the center o f the  field  - see F i g u r e 4.3.2) w e r e extracted f r o m the images at the center o f e a c h f i e l d w i t h the a i d o f the a l g o r i t h m d e s c r i b e d i n s e c t i o n 1.10.  -52-  4.3 Analysis The dose (cGy) versus film grayscale pixel value F calibration data is shown in Figure p  4.3.1. The dose Dg v was obtained from Equation 1.9.4 using k = 1.00 cGy/MU, and S = S = M  p  c  1.00, since the measurement was at a SDD of 100.0 cm using a 10 x 10 cm field size. A TMR 2  value of 1.000 was used since the measurements were taken using a thickness of solid water equal to d  max  as buildup.  Dose (cGy) vs. Pixel Value for the Film Calibration 120 100 cGy)  Hp  > s  to  80 § 1  D  = 4.7534E-08F  6 V V  2 P  + 6.9875E-04F + 6 4967 P  R = 9 9920E-01  s '^  ?  60  y  M  Q o  in  o  40  lillilil liSP 20 ;  0  "  0  I  I  5000  10000  '  ""1  1  15000  20000  25000  Film Pixel Value F  ~  I  •  30000  ~—I  35000  ...  •  —H  40000  P  Figure 4.3.1: Dose (cGy) as a function of film pixel value for the film calibration data. The film was positioned as described in section 4.2. Thirty films were used for the calibration As can be seen in figure 4.3.1, the following relationship between the dose (cGy) and the film pixel value F has good correlation (R = 0.9992) to the data, and was chosen as the 2  p  calibration equation. Equation 4.3.1 shows the best fit to the film calibration data which is a  -53 -  second order polynomial. D v is the dose as calculated by Equation 1.9.4, and F is the film 6 M  P  pixel value. D  =(4.7534x1 (T )F 8  6MV  2 P  + (6.9785x1 ( T ) F +6.4967  [4.3.1]  4  D  Equation 4.3.1 is applied to the pixel data associated with the film profile to obtain dose in cGy versus pixel position. The pixel position FN for each pixel is found by counting the pixels from a certain side of the field. A n example of a profile is shown in Figure 4.3.2.  Dose (cGy) vs. Pixel Position for a Film Image of a 15 Degree Enhanced Dynamic Wedge Treatment 6 MV  >» o o  0)  M O Q  200  400  600  Pixel Position F  800  1000  N  Figure 4.3.2: Dose (cGy) as a function of pixel position for a 70 MU, 15 degree enhanced dynamic wedge treatment. The film was positioned as described in section 4.2. In order to do a quantitative comparison of the film and EPID images it is necessary to geometrically align the images. One way of doing this is to detect the field edges in both images and register them. The field edge is defined as the 50% isodose line. I define the field edges as -54-  the boundary enclosing the beam on a cross-sectional view showing the most abrupt change in signal with position (positions of maximum gradients). The maximum gradient of Figure 4.3.2 occurs at pixel positions 262 and 621, or a width of 359 pixels, which corresponds to 13 cm since the films are taken at an SDD of 130 cm, so there is 13/359 cm per pixel which is predetermined by the settings of the sensitometer. All images were normalized at the central axis as this was the calibration point for the film. The center occurs between pixel number 441 and 442, which means the off-axis position, x (cm), as a function of film pixel position, FN, is given by the following equation. * = j^(F -441.5) N  [4.3.2]  Equations 4.3.1 and 4.3.2 are applied to the film profiles to obtain dose D (cGy) as a function of position (cm). The film had to be replaced for every irradiation so the position of the film was not exactly the same for each trial hence the need for the center reference point. The EPID's position is constant relative to the beam, and the maximum gradients for every image occurred at pixel positions 107 and 272. The maximum gradient was therefore a reliable way (within ±0.35 mm, or a half pixel spacing) to determine the field edge even from wedge treatments of different angles. The off-axis position, x (cm), as a function of EPID pixel position E N is shown below. * = -^(E -189.5) loj N  [4.3.3]  This equation is applied to the EPID data to obtain the EPID grayscale pixel value as a function of off-axis position. The goal of this chapter is to quantify the differences in profiles in an understandable fashion and the profiles will be represented as dose (cGy) as a function of position (cm). The central axis dose for each setting is determined from the film data and -55-  Equation 4.3.1, and then compared to the corresponding EPID central axis total pixel value P . T  The EPID calibration constant k from Equation 1.10.6 is determined (for 6 MV) from the film determined dose and the EPID total pixel value PT for each setting. The results are shown in Table 4.3.3. Experimental Dose vs. Pixel Vales for 6 MV Photons Wedge Angle (degrees)  Monitor Units (MU)  Central Axis Dose (cGy)  15 15 30 30 45 45 60 60 Average  100 70 100 70 100 70 100 70  54.12 37.67 48.56 34.94 44.58 34.70 40.78 26.51  EPID Pixel Value (pix)  Calculated k Value (IO" cGy/pix) 4  371800 266817 344542 245154 313600 225108 . 269290 196683  1.455621 1.411829 1.409407 1.425227 1.421556 1.541482 1.514353 1.347854 1.44  Table 4.3.3: Central axis dose (cGy) from film images, central axis EPID pixel values, and the constant in equation 1.10.6 for each treatment. In Table 4.3.3 the average value of k is shown to be 1.44 x 10" cGy/pix which is equal to 4  the result in Chapter 2 (see Equation 2.4.1), but the values for k vary by up to 0.10 x 10" cGy/pix 4  (or 7%) from the average value with a standard deviation of 0.06 cGy/pix, or 4%. This large variance is due to the complex nature of the comparison performed in this chapter as there are many sources of error including film positioning error (± 1 mm), linac output instability (up to ± 25 MU/min), and variability of film response (± 5% for the film used in this study ). The errors 22  involved with the EPID measurements include positioning error (± 1 mm), linac instability (up to ± 25 MU/min), and variability of the EPID's response (a standard deviation of 1.3% as determined in Chapter 2). Considering the variability of the EPID and film response, using the  -56-  film to predict EPID values will give an error of up to 6.3%, which is close to the maximum deviation from the average value that is seen in table 4.3.3. Using the specific k for each case as determined in Table 4.3.3 the dose D as a function of position for the EPID is determined and the comparison of EPID and film profiles can then be plotted as shown in Figures 4.3.4 through 4.3.11.  100 MU, 15 Degree Enhanced Dynamic Wedge Treatment 70 60 50 | 40 Q  § 30 Q  20 10 $  • EPID  m m  • Film  -5  1  - 2 - 1 0  1  2  Off-Axis Position x (cm)  Figure 4.3.4: Dose D as measured by an EPID and film as a function of position x for a 15 degree enhanced dynamic wedge treatment, 100 MU irradiation.  -57-  70 MU, 15 Degree Enhanced Dynamic Wedge Treatment 6 MV 35 30 i25 S o a  20 •  •  15 10  •m • EPID • Film  -  2  -  1  0  1  2  Off-Axis Position x (cm)  Figure 4.3.5: Dose D as measured by an EPID and film as a function of position for a 15 degree enhanced dynamic wedge treatment, 70 MU irradiation. 100 MU, 30 Degree Enhanced Dynamic Wedge Treatment 6 MV  1  •  1  -  2  -  1  0  1  2  Off-Axis Position x (cm)  Figure 4.3.6: Dose D as measured by an EPID and film as a function of position for a 30 degree enhanced dynamic wedge treatment, 100 MU irradiation. -58-  70 MU, 30 Degree Enhanced Dynamic Wedge Treatment 45  6  40  MV  35  —  30 $25 a 3 20  0^  '  :  "  ""  W  M m  ~  — a  r-r  •  -  <  .  ,  .  .  .  ^  .  M  •  o  a  15 10  •  E P I D  « Film  •  5 0  -  2  -  1  0  1  2  Off-Axis Position x (cm)  Figure 4.3.7: Dose D as measured by an EPID and film as a function of position for a 30 degree enhanced dynamic wedge treatment, 70 M U irradiation. 100 MU, 45 Degree Enhanced Dynamic Wedge Treatment  -  2  -  1  0  1  2  Off-Axis Position x (cm)  Figure 4.3.8: Dose D as measured by an EPID and film as a function of position for a 45 degree enhanced dynamic wedge treatment, 100 M U irradiation. -59-  70 MU, 45 Degree Enhanced Dynamic Wedge Treatment  -7  -6  -5  -4  -3  -2  -1  0  1  2  3  4  5  6  7  Off-Axis Position x (cm)  Figure 4.3.9: Dose D as measured by an EPID and film as a function of position for a 45 degree enhanced dynamic wedge treatment, 70 MU irradiation. 100 MU, 60 Degree Enhanced Dynamic Wedge Treatment  -7  -6  -5  -4  -3  -2  -1  0  1  2  3  4  5  6  7  Off-Axis Position x (cm)  Figure 4.3.10: Dose D as measured by an EPID and film as a function of position for a 60 degree enhanced dynamic wedge treatment, 100 MU irradiation. -60-  70 MU, 60 Degree Enhanced Dynamic Wedge Treatment  -7  -6  -5  -4  |  -3  -2  -1  0  1  2  3  4  5  6  7  Off-Axis Position x (cm)  Figure 4.3.11: Dose D as measured by an EPID and film as a function of position for a 6 0 degree enhanced dynamic wedge treatment, 7 0 M U irradiation. Figures 4.3.4 through 4.3.11 show the comparison of dose profiles measured with the EPID and film. The profiles are normalized in the center of the field and show good correlation considering the expected error in dosimetry of ±5% for film , and of ±1.3% for the EPID (see 22  chapter 2). Outside the field the dose as determined by the film calibration is 20% to 100% higher than the dose as determined by the EPID calibration. A possible explanation for this behavior may be due to thefilmsapparent over-response to low energy scattered photons that are dominant outside the field . Figures 4.3.4 to 4.3.11 show some very interesting features. They show good correlation at the center of the field, but this is where the two signals were calibrated. The signal for the film increases relative to the signal from the EPID as the off-axis distance increases. This is most likely due to the fact that the photons have a higher average energy in the center of thefieldthan -61 -  at the edges. At 6 MV the film seems to over respond to low energy photons compared to the EPID. The EPID images are more linear inside the field edges than the film images. The film images seem to have a sharper change at the field edge than the EPID images. This broadening of thefieldedges may be partly due to scattering in the build-up of the EPID. The 70 MU images correspond better to the EPID images than the 100 MU images for all cases. In all cases the EPID measured dose varies from the film measured dose by approximately 10% near the field edges. This variance could be reduced by fitting the EPID data to the entire field instead of the center of thefieldby finding the best linear fit to the film profile data within the field and adjusting the EPID calibration to correspond to it. Calibrating in this manner would give values off by about 5% in the center and at the edges, but more accurate in the intermediate region.  4.4 Conclusion The constant k in equation 1.10.6 was determined to be 1.44 x 10" cGy/pix with a 4  standard deviation of 4% for 6 MV photons. Althoughfilmdosimetry is only considered to be accurate within ±5%, the average k value found is consistent with the value previously determined in chapter 2. The individually determined values of k varied by up to 7% which is close to what is expected from adding the uncertainties in the film (+5%) and EPID (±1.3%) response. Thefilmover responds to low energy photons compared to the EPID and the EPID gives a more linear dose profile within thefield.The EPID image has broader field edges than the film image due to scattering and imaging processing. The EPIDs response to changing thickness of solid water is investigated in Chapter 5 and is compared to the ion chamber rather thanfilmsince the ion chamber is the standard for measurement of dose. -62-  Chapter 5 EPID Response versus Thickness of Solid Water 5.1 Introduction In section 1.6 it was explained that in an ideal isodose distribution of uniform intensity, the average photon energy decreases as the distance from the central axis increases , but this is 21  only one of the sources of spectral variation within a treatment portal. As a photon beam penetrates through a material (tissue for example) the low energy photons are preferentially removed from the beam before the high energy photons. Therefore the primary beam contains a larger percentage of high energy photons at deeper depths in the material , but this is only one 21  of the sources of spectral variation. As the beam penetrates further into the material, the amount of low-energy scattered radiation increases as well. So there is an increase in the amounts of low and high energy photons relative to average energy photons as the beam penetrates further. In a typical radiotherapy treatment the photons in a given field will travel through a variety of depths of tissue before interacting with the EPID. The EPID will therefore be exposed to a variety of fluences and spectra of radiation when measuring treatment portals. Although the EPID is treated as a black box in this study, it is important to understand that its inherent buildup material interacts with photons differently than solid water or tissue does. The results from chapter 3 show that the effective thickness of the inherent buildup varies with photon energy so it may also vary with the spectral change that occurs as the beam penetrates deeper into material. This is expected to contribute to a varying signal with depth of solid water compared to the ion chamber. The study conducted in this chapter compares the EPID response to ion chamber measurements for various thicknesses of overlying material which is expected to produce - 63 -  spectral variations in the beam. The E P I D is expected to respond differently than the ion chamber to these varying conditions and, if the data is normalized at d  max  then they will show  some difference at greater thicknesses of solid water as shown in Figure 5.1.1.  4  TMR Curves as measured by the EPID and Ion Chamber  Tissue Maximum Ratio  Thickness of Solid Water  Figure 5.1.1: If the TMR curves measured by the EPID and ion chamber are normalized at d , then it is expected they will differ at greater thicknesses of solid water as shown. In this case the EPID can still be used accurately as a dosimeter by correcting for the thickness of solid water (a) or the difference in dose (b). max  Three ways to correct for the difference in response are outlined and compared in this chapter. The first method involves normalizing the curves at d  max  and determining a correction  for the thickness (segment a in Figure 5.1.1) as a function of thickness of solid water. The second method (segment b in Figure 5.1.1) is to determine a correction for the normalized TMR data as a function of thickness of solid water. The third is to use the ion chamber data to determine a direct calibration equation in the form of dose (cGy) as a function of E P I D total pixel value, PT (section 1.10). The direct calibration equation is only valid for thicknesses greater than d -64-  max  .  The EPID will be treated as a black box and the outer surface of the touch guard is set to be the position of the detector. A small error in field size is expected to occur due to the operator's ability to place the surface of the EPID at a known distance. The percent variation of the field size will be small at the SDD of 130 cm used in this study therefore the change in TMR for this small change in field size is not expected to introduce an appreciable error. The true position of the array of photodiodes is actually in a plane with a slightly larger field size than the cover, but this is inconsequential since the EPID is being treated as a black box. The response of the EPID solely to varying fluence is investigated in Chapter 6.  5.2 M e t h o d  Images for this study were acquired on a PortalVision aS500 amorphous silicon EPID x  mounted on a dual energy (6 and 18 MV) CL2100EX accelerator. The EPID was treated as a 1  black box in this study, and the top surface was considered to be the detector position. EPID measurements are compared to ion chamber measurements. The surface of the EPID and the center of the ion chamber are positioned 130 cm from the source for all measurements and the relative positions of the detectors are shown in Figure 3.1.1. The field size is set to 10 x 10 cm at isocenter (100 cm) and is therefore 13 x 13 cm at a distance of 130 cm. The response of the EPID to changing thicknesses of solid water placed on top was investigated using a constant dose (130 MU) and dose rate (300 MU/min) for both 6 MV and 18 MV irradiation. The build-up was varied from 1.0 cm to 56.0 cm to investigate whether spectral hardening will induce different responses between the EPID and the ion chamber.  1  Varian Medical Systems, Palo Alto, CA. -65 -  The ionization data is normalized at d in Figure 5.1.1 are determined. The dose at d  max  max  for both instruments and the corrections shown  , D is calculated at 130 cm from the machine  calibration data for each energy using equations 1.9.5 and 1.9.6, and it is divided by the ion chamber reading Q (nanoCoulombs, nC) at d  max  (at 130 cm) to obtain a calibration constant kj  C  (in units of cGy/nC) for the ion chamber for 6 and 18 MV photons, which relates the charge Q measured to the dose D delivered. D = KQ  [5.2.1]  Multiplying the ion chamber reading Q (nC) as a function of thickness (cm) by the calibration constant kj gives dose (cGy) as a function of thickness (cm). The dose values are C  compared to the E P I D total pixel values PT (section 1.10) as a function of thickness. Three methods of using the E P I D for dosimetry are determined and compared.  5.3 Analysis  The first two methods of comparing the E P I D to the ion chamber involve normalizing the data at d  max  for each detector independently. The effective build-up of the portal imager for 18  MV irradiation, which was found in Chapter 3 (2.24 cm), was added to the thickness of solid water for the E P I D data. This value was found for d d  max  max  and is therefore expected to be accurate at  . The E P I D total pixel value PT and the ion chamber readings were normalized at d  dividing the data from the instruments by their respective values at d E P I D is the total pixel value PT divided by the total pixel value at d  max  max  max  by  . The TMR data from the  and is labeled E P I D in the  following equations. The TMR data from the ion chamber is the total charge Q divided by the total charge at d  max  and is labeled I C in the following equations. This necessarily gives close  -66-  correlation of the data near d  max  , but as the depth increases the two signals differ somewhat. The  data as measured by the EPID and ion chamber for 18 MV is shown in Figure 5.3.1.  18 MV TMR Data from EPID and Ion Chamber 1.1 1 0.9  O  0.8 0.7 _ 0.6 S >- 0.5 0.4 i  0.3  o  •o  0.2 OEPID  0.1  • I.C.  10  15  20  25 30 35 Thickness (cm)  40  45  50  55  60  Figure 5.3.1: Normalized ionization data for 18 MV photons as measured by the EPID (white diamonds) and by the ion chamber (black squares). In this case for the EPID measurements the equivalent thickness of the inherent buildup as determined in Chapter 3 has been added to the thickness of solid water. As expected the curves match closely near d  max  but diverge as the thickness x increases.  In order to examine the difference more closely, a plot of the data in the range of thickness x from 20 cm to 60 cm is shown in Figure 5.3.2. In this graph, lines have been fit to the curves to illustrate the difference. Exponential fits to the data have been used as the primary photon flux is expected to decrease in this way and because they give the highest R values.  -67-  18 MV TMR Data from EPID and Ion Chamber 0.8  EPID = 1 . 2 4 5 0 e - °  OZ54x  R = 0 9999 2  IC = 1 . 1 9 7 5 e - °  02S7,;  R = 1 0000 7  0.2  0.1  • 20  25  30  35  40  Thickness (cm)  45  50  EPID I.C. 55  60  Figure 5.3.2: Normalized (at d ) ionization data for 18 MV photons as measured by the EPID (white diamonds) and by the ion chamber (black squares) for 20 cm to 60 cm only. In this case the equivalent thickness of the inherent buildup of the EPID as determined in Chapter 3 has been added to the thickness of solid water. max  Figure 5.3.2 shows that there is a difference between the signal from the E P I D and ion chamber. The percentage difference o f the E P I D from the ion chamber is found as follows. 1.2450e-  00254A:  -1.1975e-  1.1975e  -0.0257*  00257j;  100% = (1.0397*  This function is shown below i n Figure 5.3.3.  -68-  0.0003*  •1.0000)100%  [5.3.1]  Percent Difference of EPID from Ion Chamber vs. Thickness (cm) 7 6 5 4 3 2 *= Q 0 1  *  -1 -2 -3 -4  -5 -6 -7 10  20  30  40  50  60  70  Thickness (cm)  Figure 5.3.3: The percentage difference of the normalized signal from the EPID compared to the ion chamber for 20 cm to 60 cm overlying material for 18 MV is shown. In this case the equivalent thickness of the inherent buildup of the EPID as determined in Chapter 3 has been added to the thickness of solid water. Figure 5.3.3 shows that the percent difference of the EPID signal from the ion chamber signal increases with increasing thickness of solid water. The curves seem almost parallel all the way and the increase in % Diff is at least in part due to the decreasing dose value. For 18 MV this shows a definite over response of the EPID compared to the ion chamber to high energy photons as the beam becomes harder with thickness. This is plausible since the material of the EPID has a higher effective atomic number than the ion chamber (see chapter 1). Figure 5.3.2 shows that the EPID signal is shifted right compared to the ion chamber signal. Including the equivalent thickness of the inherent build-up of the EPID for TMR data has given good correlation for the two curves near d  max  , but it has produced a large variation at greater -69-  thicknesses of solid water. To investigate this, the data is plotted again without adding the equivalent thickness of the inherent build-up to the thickness of solid water on top of the EPID and the results are shown in Figure 5.3.4.  18 MV TMR Data Without Inherent Build-up Included 1.1 1  c  0.9 0.8  II  0.7  •  Oi  On OI  or 0.6 5 I- 0.5 m 0.4  0.1  Om  4—  O,  •  0.3 0.2  o.  n  OEPID M  0 4— 0  5  10  15  20  25  30  35  40  45  50  I.C.  55  60  Thickness (cm)  Figure 5.3.4: Normalized (at d x) ionization data for 18 MV as measured by the EPID (white diamonds) and by the ion chamber (black squares). In this case the equivalent thickness of the inherent buildup of the EPID as determined in Chapter 3 has not been added to the thickness of solid water. ma  Without the equivalent thickness of the EPID's inherent build-up added, the curves match nicely at depths greater than 30 cm, but differ significantly nearthe depth of maximum response. This again shows a difference in response between the EPID and the ion chamber. Figure 5.3.5 shows this data plotted for depths greater than 20 cm. It can be seen that the two exponential lines coincide very well and fit their respective data very consistently.  -70-  18 MV TMR Data Without Inherent Buildup Included 0.7  0.6  0.5  0.4  EPID= l\J845e"°-  025Sx  FP = 1 0000 0.3  •  EPID I.C.  0.2 20  25  30  35  40  45  50  55  60  Thickness (cm)  Figure 5.3.5: Normalized (at d x ) ionization data for 18 MV as measured by the EPID (white diamonds) and by the ion chamber (black squares) for 20 cm to 60 cm. In this case the equivalent thickness of the inherent buildup of the EPID as determined in Chapter 3 has not been added to the thickness of solid water. ma  The percentage difference between the EPID signal and the ion chamber reading for the TMR data in Figure 5.3.5 is calculated below. 1.1845e  _0 0 2 5 5 x  -1.1994e  1.1994e  _0 0 2 5 7 j c  -0.0257*  0.0002*  -100% = (0.9872e  This function is plotted from 20 cm to 60 cm in Figure 5.3.6.  -71 -  1)100%  [5.3.2]  Percent Difference of Epid vs. Ion Chamber (18 MV) Without Adding the Effective Build-up 6 5 4 3 2 . 1 it:  5 0  -5 -6 -7 10  20  30  40  Thickness (cm)  50  60  70  Figure 5.3.6: The percentage difference of the signal from the EPID compared to the ion chamber for overlying thicknesses of 20 cm to 60 cm for 18 MV. In this case the equivalent thickness of the inherent buildup of the EPID as determined in Chapter 3 has not been added to the thickness of solid water. It can be seen from Figure 5.3.6 that the percentage difference between the EPID and ion chamber signals is much less at thicknesses greater than 20 cm if the equivalent thickness of the EPID's inherent buildup is not added to the thickness of solid water. It can be seen that the difference is greater for more shallow thicknesses if the equivalent thickness of the inherent buildup is included, and certainly does not give good correlation near d  max  . Adding the  equivalent thickness of the EPID's inherent buildup as determined in chapter 3 to the thickness of solid water gives good correlation at shallow thicknesses, and not including it gives good correlation at greater thicknesses. The EPID clearly responds differently to changing thickness of solid water than an ion chamber does. -72-  The first method of accounting for this difference involves correcting for the thickness (segment a in Figure 5.1.1). This leads to the idea of designing a function for the equivalent thickness of the EPID's inherent buildup that varies with the thickness of solid water placed on top of the EPID. The percentage difference in the signals (Figure 5.3.3 and 5.3.6) has varied linearly with distance in both cases. The line of best fit in Figure 5.3.6 is linear and has an xintercept of 62.5 cm, so the function for the equivalent thickness of the EPID's inherent buildup should have a value of zero when the solid water added is 64.5 cm. According to the investigation in chapter 3 the inherent build-up for 18 MV is 2.24 cm with 2.28 cm of solid water on the EPID. The linear function that fits these conditions is as follows where d ff is the e  equivalent thickness of the inherent build up (in cm) of the EPID and d is the thickness of solid water placed above the EPID. d  eff  = -0.0360^ + 2.322  [5.3.3]  Figure 5.3.7 shows a plot of TMR vs. thickness for the ion chamber and EPID data with this function [5.3.3] added to the solid water placed on top of the EPID.  -73 -  TMR Data for EPID and Ion Chamber with EPID Correction (Equation 5.3.3) for Equivalent Thickness of the Inherent Buildup of the EPID  iNPNillli^BllSfil  ... « ^ . o •  |llJl^pi^rtiiSI|M  iti  iiilillliR  5 0.6 i—  •  piSi^pl^Blilll  #il§l|plli| illlllPiiili • illiBllil  lllllll  ijilt  •  • • • EPID  • i.e. 1  0  10  20  30 Thickness (cm)  40  50  60  Figure 5.3.7: Normalized (at d ) ionization data for 18 MV as measured by the EPID (white diamonds) and by the ion chamber (black squares). In this case the equivalent thickness of the inherent buildup of the EPID as determined by Equation [5.3.3] is included in the thickness and is a function of the thickness of solid water placed on the EPID. max  In Figure 5.3.7, the two signals are closer than in Figure 5.3.1 or in Figure 5.3.4 although the linear correction does not appear to satisfy the middle region very well. In order to show the difference more clearly Figure 5.3.8 shows the same data blown-up for depths over 10 cm.  -74-  TMR Data Using Linearly Corrected Thickness for the EPID 1  T l l l i i S ^ ^  0.9 0.8 0.7 -  -I  0.6  on S 0.5 0.4 0.3  0.1  0  •  20  10  30  40  EPID l.C.  50  60  Thickness (cm)  Figure 5.3.8: Normalized (at d ) ionization data for 18 MV as measured by the EPID (white diamonds) and by the ion chamber (black squares). Similar to Figure 5.3.7 except only the data is shown for above 10 cm. In this case the equivalent thickness of the inherent buildup of the EPID as determined by Equation [5.3.3] is included in the thickness and is a function of the thickness of solid water placed on the EPID. max  Figure 5.3.8 shows a large difference in the 10 cm to 25 cm region. This is not a sufficient way to correct for the equivalent thickness of the inherent build-up of the EPID. The equations from Figure 5.3.5 can be used to precisely determine a correction function for thickness. Figure 5.3.9 shows parts of two curves similar to the ones being examined in this study. We need to know how much thickness to add on to the thickness of the solid water in front of the EPID so that the signal will be similar to that of the ion chamber. The thickness correction for the difference in signals is x - xi in the figure. In order to create this correction function, 2  TMR functions are required for the EPID and ion chamber which we have from Figure 5.3.5.  -75-  TMR Curves as measured by the EPID and Ion Chamber  TMR  Figure 5.3.9: Determination of the inherent build up of the imager. The equations of best fit for the ion chamber and the EPID respectively for thicknesses greater than 25 cm and where the equivalent thickness of the EPID's inherent buildup has not been included (Figure 5.3.5) are the following. / C = 1.1994e-  [5.3.4]  00 2 5 7 x  £ ™ = 1.1845e-  [5.3.5]  00255j;  Where IC is the ion chamber signal normalized at d normalized at d  max  max  , and EPID is the EPID signal  without adding any correction for the equivalent thickness of the EPID's  inherent buildup. Using Figure 5.3.9 and equations 5.3.4 and 5.3.5, the following relations can written. '  [5,3.6]  -0.0257*,  [5.3.7]  y =1.1994<r  00257j:  2  y =1.1994e x  76 •  v, =1.1845e-  [5.3.8]  00 2 5 5 X |  Solving equation 5.3.7 for x gives... 2  x = 2 2  ln( * ) 0.0257 1.1994 -  1  [5.3.10]'  y  Then the horizontal shift of the ion chamber compared to the EPID at a thickness of Xj can be found using equations 5.3.10 and 5.3.8 as follows. This is also the correction for the equivalent thickness of the inherent buildup of the EPID. -1 x, - x. = '  , , ln(  0.0257  v, , •—) - x, = 1.1994  -1  , 1.1845 -° ln( /  0.0257  1  0 2 5  g  ^\ ) - x,  [5.3.11]  1.1994  x - x , =0.4864-0.007782x,  [5.3.12]  2  Compared to equation 5.3.3, this relation has a much lower y-intercept and slope. Adding this to the thickness of solid water placed on the EPID should give us very good correlation for the exponential part of the TMR curve. Of course it may not be very accurate near the thickness of maximum signal and very shallow thicknesses. At depths greater than 10 cm for 18 MV, using equation 5.3.12 for the inherent depth of the EPID is a very good approximation and makes the TMR data from the EPID equal to that from the ion chamber. The second method of correction is the vertical correction for the normalized TMR data. The correction for depths greater than 10 cm can easily be calculated to find the ion chamber read TMR (IC) from the EPID signal TMR (EPID) for 18 MV using equations 5.3.4 and 5.3.5 as follows. IC-EPID  = 1.1994e~ '" -1.1845e"  IC = EPID +1.\994e'°  c)my  0251  00255jc  x  -1.1845e-°  -77-  0255j;  [5.3.13] [5.3.14]  Corrections for thicknesses less than 10 cm should be mapped separately or avoided altogether by placing solid water on the EPID when measurements are being taken. The TMR data as measured by the EPID and ion chamber for 6 MV is shown in Figure 5.3.10. The equivalent thickness of the inherent buildup of the portal imager as found in chapter 3 of 1.38 cm was added to the thickness of solid water for the EPID data. This addition seems appropriate for the data near the thickness of maximum dose as the curves match closely in this region as they did for the 18 MV data. As the thickness of solid water increases, the two devices start to give different results.  6 MV EPID and Ion Chamber TMR Data 1.2 o EPID • I.C.  0.8  - 0.6  0.4  0.2  10  20  30 Thickness (cm)  40  50  60  Figure 5.3.10: Normalized (at d ) ionization data for 6 MV as measured by the EPID (white diamonds) and by the ion chamber (black squares). In this case the equivalent thickness of the inherent buildup of the EPID as determined in Chapter 3 has been added to the thickness of solid water. max  -78-  Figure 5.3.10 shows good correlation between the EPID and ion chamber for 6 MV. Figure 5.3.11 below shows the results at depths greater than 10 cm (the exponential region) with corresponding functions. In this graph lines have been fit to the curves to illustrate the difference. Exponential functions are used to fit the data as the intensity of the beam is expected to decrease in this manner. 6 MV EPID and Ion Chamber TMR Data 0.9  EPID = 1 2547e •0.0415X 0.9998  s 0.3 -  IC= 1 2 3 2 5 e  0 0 4 , 0 <  R = 0.9999 z  0.2 0.1 0n 10  EPID I.C. 15  20  25  30  35 40 Thickness (cm)  45  50  55  60  Figure 5.3.11: Normalized (at d ) ionization for 6 MV as measured by the EPID (diamonds) and by the ion chamber (squares). In this case the equivalent thickness of the inherent buildup of the EPID as determined in Chapter 3 has been added to the thickness of solid water. max  The percent difference of the EPID data from the ion chamber data is calculated below. 1 9C47  -0.0415*  1 2325  1  9"50C  0 0 A l x  -0.0410*  100% = (1.0180e-  00005x  -1.0000)100%  This function is plotted from 10 cm to 60 cm in figure 5.3.12 below.  -79-  [5.3.15]  Percent Difference of EPID from Ion Chamber TMR data for 6 MV With Inherent Buildup Included  -7  t  0  •  -  -  i • -  10  t  •  •  20  • - -r 30  i  i  i  40  50  60  -  -i  70  T h i c k n e s s (cm)  Figure 5.3.12: The percent difference of the signal from the EPID compared to the ion chamber for 10 cm to 60 cm for 6 MV. In this case the equivalent thickness of the inherent buildup of the EPID as determined in Chapter 3 has been added to the thickness of solid water.  It can be seen from Figure 5.3.12 that the EPID responds less to the 6 MV beam at large thicknesses compared to the ion chamber. Since the beam has a higher percentage of high energy photons as the thickness increases, this means at 6 MV the EPID is more sensitive to the low energy photons and less sensitive to the high energy photons than the ion chamber. Using the inherent build-up as calculated in Chapter 3 gives a fairly good approximation for depths from 15 cm to 55 cm as the EPID values are within 1% of ion chamber readings in that range. Figure 5.3.13 below shows the signal from the EPID and from the ion chamber as a function of thickness for 6 MV without the equivalent depth of the EPID's inherent buildup added to the thickness of solid water. -80-  6 MV EPID and Ion Chamber TMR Data Without Inherent Buildup Included 1.2 1 ^  0.8 a: 5 0.6  O" h  -  •-•  -  <>•  OEPID • I.C. 6 MV  0.4 - ^ ^ ^ g ^ ^ ^ ^ S  0  .  2-  ^ —  10  ^  ^ 1  :  20  O"  ^  B —  ^  —  o«  i  30  ^  ^ 1  40  ^  ^  ©•  1  50  ^  ^ ~ ~  60  Thickness (cm)  Figure 5.3.13: Normalized (at d ) ionization data for 6 MV as measured by the EPID (diamonds) and by the ion chamber (squares). In this case the equivalent thickness of the inherent buildup of the EPID as determined in Chapter 3 has not been added to the thickness of solid water. max  Figure 5.3.14 shows the same data for 10 cm and beyond only (the exponential region). It can be seen that using the buildup from Chapter 3 gives a better match between EPID and ion chamber data.  -81 -  ^  ^  6 MV TMR Data Without Inherent Buildup Included 0.9 0.8 0.7 IC = 1 2325e  0.6  cc  0M,!  '  R = 0.9999 2  0.5 0.4 EPID = 1 1859e- '' 0M  0.3  5<  R = 0 9998 2  0.2 0.1 -F • 20  10  30 Thickness (cm)  EPID I C 6 MV 50  40  60  Figure 5.3.14: Normalized (at d ) ionization data for 6 MV as measured by the EPID (diamonds) and by the ion chamber (squares). In this case the equivalent thickness of the inherent buildup of the EPID as determined in Chapter 3 has not been added to the thickness of solid water. max  The percentage difference between the two signals is calculated below. 1.1859 e  00415jr  -1.2325 -  1.2325e-°-  e  041j;  00410x  -100% = (0.9622e  -0.0005*  1.0000)100%  [5.3.16]  Equation 5.3.16 is plotted in Figure 5.3.15 for 10 cm to 60 cm for illustrative purposes. It can be see that for 6 MV the percent difference of the EPID signal from the ion chamber signal is much greater when the inherent buildup is not included.  -82-  Percent Difference of EPID from Ion Chamber TMR data for 6 MV Without Inherent Buildup Included 6 5 4 3 2  «  1  5 0 * -1 -2 -3 -4 •••••«  -5 -6 -7  10  20  30  40  50  Thickness (cm)  60  70  Figure 5.3.15: The percent difference of the signal from the EPID compared to the ion chamber for 10 cm to 60 cm for 6 MV. In this case the equivalent thickness of the inherent buildup of the EPID as determined in Chapter 3 has not been added to the thickness of solid water. A similar (but shortened) comparison will be done with 6 MV photons, since they have a different energy spectrum, and may interact differently with the detectors than the 18 MV photons. From an evaluation of looking at Figures 5.3.9 and 5.3.14, and using the same analysis that lead to equation 5.3.11 for 18 MV photons, the thickness correction (the first method) for the EPID as a function of thickness of solid water for 6 MV is found to be: x,-x. = 2  1  0.0410  \n(—^—)-x, 1.2325  1  = ln( 0.0410  x -x, =0.940 + 0.0122x, 2  -83 -  )-x, 1.2325  [5.3.17]  1  [5.3.18]  The second method involving the correction for ion chamber TMR from EPID TMR as a function of depth of solid water on top is calculated straight from the equations in Figure 5.3.14 and is shown below. / C - £ P / D = 1.2325e-  -1.1859e-  [5.3.19]  / C = EP/D + 1.2325e-  -1.1859e-  [5.3.20]  00410j;  00410x  J  00 4 1 5 j :  00 4 1 5 x  Using equation 5.3.18 or 5.3.20 to correct the EPID signal TMR gives very good correlation to ion chamber TMR for 6 MV photons at depths from 10 to 60 cm. At thicknesses less than 10 cm, the correction should be mapped separately. Again, this can be avoided altogether by placing adequate solid water before the EPID and just utilizing the exponential part of the TMR curve. The third method of accurately using the EPID as a dosimeter involves using the ion chamber data to determine a direct calibration equation for dose D as a function of the EPID total pixel value P . In order to obtain the constant k in equation 1.10.6 which relates the EPID total T  pixel value to dose, the dose at d  max  must be determined. For 6 MV photons this can be found  using equation 1.9.5 as shown below. D  6 M V  = 0.5976(MU) = 0.5976(130) = 78.23 cGy  [5.3.21]  Since the ion chamber signal is known to be proportional to the dose the following relation can be written expressing the dose D (cGy) as a proportional function of the ion chamber signal Q (nanoCoulombs, nC), where kj is the proportionality constant of the ion chamber. c  D =k Q  [5.3.22]  ic  The value of the proportionality constant for 6 MV photons kj 6Mv can be found using the C  data at d  max  for 6 MV.  -84-  cQ,  D_T%33cGy Q  \4A29nC  [5.3.23]  nC  Converting the ion chamber data to dose and plotting versus depth of solid water in the exponential region ( 1 0 cm to 6 0 cm) gives Figure  5.3.16.  Dose (cGy) vs. Thickness (cm) of Solid Water for 6 MV 70  60  50  >  Dose = 96.42e"  40  004097><  R = 0 9999 2  0)  (A  o 30 20  10  10  20  30  Thickness (cm)  40  50  60  Figure 5.3.16: Dose (cGy) versus thickness (cm) of solid water as measured by the ion chamber for 10 cm to 60 cm for 6 MV. Figure 5 . 3 . 1 7 below shows the total average pixel value P T plotted versus thickness of solid water x in the exponential region.  -85-  EPID Total Average Pixel Value vs. Thickness (cm) for 6 MV 500000 450000  100000 50000  30  Thickness (cm)  Figure 5.3.17: Total average pixel value versus thickness (cm) as measured by the EPID for 10 cm to 60 cm for 6 MV. From the equation in Figure 5.3.17, the position x can be solved as a function of the total average E P I D pixel value and substituted into the equation in Figure 5.3.16 to obtain the following relationship of dose as a function of the total average pixel values PT from the E P I D for 6 MV. D _ = 96.42e6MV  004100  -0.04100(  *  = 96.42e  D  6MV  D  !  -0.04150  = 96.42( 660800 =(l.7\5x\0- )(P ) 4  6MV  ln(—5—)  09m  T  66osoo' _ .  0.98795 ln(—2_)  96 42^  66osoo  ;  [5324]  0.9879  [5.3.25]  It is apparent that the dose is not exactly linear (as expected from equation 1.10.6 ) to the EPID total pixel value P for changing thicknesses of solid water. The linear relation that best T  - 86 -  approximates equation 5.3.25 in the dose range from 10 cGy to 120 cGy for 6 MV photons is shown below. D  =(1.450x1 Q- )(P ) + 0.6020  [5.3.26]  A  6MV  T  For 18 MV photons the dose at d  max  can be found using equation 1.9.6, and the value of  the proportionality constant in Equation 5.3.22 for 18 MV photons k i vcan be found using the ic  data at  d  m a  v  8M  for 18 MV. D_78.3kG> ic\mv  v  Q  =  5 1 4 7  _cGy  15.19&C  nC  [5.3.27]  Converting the ion chamber data to dose and plotting versus depth in the exponential region 10 cm to 60 cm gives Figure 5.3.18. The exponential line of best fit is shown on the graph.  Dose (cGy) vs. Thickness (cm) of Solid Water for 18 MV  30  Thickness (cm)  -87-  60  Figure 5.3.18: Dose (cGy) versus Thickness (cm) of solid water as measured by the ion chamber for 10 cm to 60 cm for 18 MV. Figure 5.3.19 below shows the total average pixel value plotted in the exponential region with the equation shown on the graph. EPID Total Pixel Value vs. Thickness (cm) of Solid Water for 18 MV 600000  iBllliil  500000 jlBJitfiiiii  SiSlipiliSPi'  o „ 400000 > "a) x  E  o> 300000 O)  2  > < 200000  IB  P , = 6398900e-°  02400x  100000  10  20  30  Thickness (cm)  40  50  60  Figure 5.3.19: Total average pixel value versus thickness (cm) of solid water as measured by the EPID for 10 cm to 60 cm for 18 MV. Using the same analysis as for 6 MV the following relationship of dose as a function of the total average pixel values from the EPID for 18 MV is obtained. ZW=  (1.105x10- )(P .) 1.020 V  4  7  [5.3.28]  This relationship is essentially linear in regions of clinical interest. The linear function that best corresponds to this function in the range from 10 cGy to 120 cGy for 18 MV photons is shown below.  =(1.392xlO- )(P ) + 1.071 4  D  WMV  r  -89-  [5.3.29]  5.4 C o n c l u s i o n  It has been determined that the EPID responds a little different from the ion chamber under conditions of varying thickness of solid water. It was shown that modeling the EPID using a single inherent build-up depth with no other correction will not be accurate for varying depths of solid water. At 6 MV, the EPID over-responded to the low energy photons compared to the ion chamber and at 18MV, the EPID over-responded to the high energy photons compared to the ion chamber. The photoelectric effect dominates for low energy photons and pair production dominates for high energy photons and the probability of interaction for both increases with increasing atomic number. The over-response of the EPID and the dependence of these effects on atomic number indicate that the inherent build-up of the imager has a higher effective atomic number than the ion chamber and solid water. The EPID can be used to give TMR results similar to those from an ion chamber at depths greater than 10 cm using either a correction for thickness (equation 5.3.18 for 6 MV or 5.3.12 for 18 MV), or for signal (equation 5.3.20 for 6 MV or 5.3.14 for 18 MV). The direct calibration for dose from total average pixel value is given by equations 5.3.25 and 5.3.28. Approximate linear relations are given by equations 5.3.26 and 5.3.29. All three methods of using the EPID as a dosimeter proved effective, however the direct calibration will be easier to use in some cases. In the exponential region of the TMR curves (Figures 5.3.2, 5.3.5, 5.3.11, and 5.3.14) the r-squared values of the EPID's line fits are almost as high as the ion chambers, but consistently lower. This shows that the ion chamber consistently fits the expected exponential decrease in dose with thickness better than the EPID.  -90-  These equations are derived from central axis data and may not be useful for the off-axis regions of the field. During treatments the radiation from the linac passes through a flattening filter and a variety of thicknesses of tissue before reaching the EPID, the spectrum of the measured field will vary significantly with position . This means that the EPID is likely to require a calibration equation that is a function of position and thickness of tissue. The effect of the flattening filter will be constant and the depth of tissue could be known by the signal strength. It should be investigated how much the EPID signal varies from the ion chamber reading at point off-axis to determine whether or not one calibration equation is sufficiently accurate for each field or whether the equation should be a function of off-axis position as well.  -91 -  Chapter 6 Dosimetric Comparison of an Ion Chamber and EPID 6.1 Introduction The EPID's response to changing thickness of solid water (Chapter 5) and dose rates (Chapter 2) has been determined. In this chapter the response of the EPID is compared to an ion chamber under conditions of changing dose. The EPID calibrations are determined for both energy modes. The instruments were placed at a SDD of 130 cm with a thickness of solid water equal to d  max  placed above for each energy mode. The dose was varied by changing the number  of monitor units delivered. The linac is calibrated so that the number of monitor units is proportional to the dose delivered. Both the EPID and ion chamber measure ionization which is proportional to incident flux unless the device is saturated. Saturation occurs when the device has the maximum amount of charge that it can hold, but this is not a consideration in this study. The response of the EPID under these conditions is found to be directly proportional to that of the ion chamber. EPID calibration equations in the form of dose (cGy) as a function of total pixel value ?j were determined to be linear for both energy modes.  -92-  6.2 Method Images for this study were acquired on a PortalVision aS500 amorphous silicon EPID 1  mounted on a dual energy (6 and 18 MV) CL2100EX accelerator. The EPID was treated as a 1  black box in this study, and the top surface of the touch guard was considered to be the detector position. The EPID and the center of the ion chamber were positioned at 130 cm from the source for all measurements, and the relative positions of the detectors are shown in Figure 6.2.1. The 9  9  field size is set to 10 x 10 cm at isocenter (13 x 13 cm at a distance of 130 cm). The measurements were done at d  max  , therefore 1.5 cm and 3.2 cm of solid water were  used as buildup for 6 and 18 MV photons respectively for all measurements. The ion chamber was imbedded in the solid water with 10 cm of solid water behind it to provide backscatter as shown in Figure 6.2.1. For 6 and 18 MV photons the EPID and the ion chamber are irradiated with doses varying from 10 to 130 MU using a constant dose rate of 300 MU/min. Positioning of the EPID and Ion Chamber  Solid Water  ^d  Ion Chamber  m a x  A  Solid Water  •  y  Figure 6.2.1: Relative vertical positions of the EPID and Ion Chamber. The ion chamber is embedded in solid water with 10 cm of backscatter.  6.3 Analysis The total pixel value PT (I define the unit to be pix) of the E P I D is directly proportional to the ion chamber reading Q (in nC, 1 nC = 10" C) for 6 MV photons as can be seen in Figure 9  c  Varian Medical Systems, Palo Alto, CA - 93 -  6.3.1. The correlation is excellent, and the r-squared value is 1.000. The relationship is shown on the graph with a proportionality constant of 2.631 x 10" nC/pix and a y-intercept of 6.258 x 10" 5  2  nC.  Ion Chamber Reading (nC) vs. EPID Total Pixel Value P  T  Calibration Curve for 6 MV Photons 1  18  1  1  5  1  ^  —  Figure 6.3.1: Ion chamber reading (nC) versus EPID total pixel value for varying fluence at 6 MV. The response of the E P I D PT as a function of the number of monitor units (MU) used is shown in Figure 6.3.2. The r-squared value for the linearfitis 1.000, which shows excellent linearity for the dosimetric response of the E P I D .  -94-  Dose (Monitor Units) vs. EPID Total Pixel Value Calibration Curve for 6 MV 160 te^::*^  0  100000  200000  300000 400000 EPID Total Pixel Value P  500000  600000  700000  T  Figure 6.3.2: EPID total pixel value versus the number of monitor units used for 6 MV photons. The response of the EPID to total dose is linear. The equation relating the total pixel value and the number of monitor units is the inverse of the equation in Figure 6.3.2. P =(4202)MU-469.6  [6.3.1]  T  The y-intercept is very small compared to typically pixel values measured in this study and contributes an insignificant amount for doses typical used in radiation therapy. Combining this with Equation 1.9.5 and using a TMR of 1.000, a relationship [6.3.2] is obtained-for the EPID total pixel value as a function of the Dose D6MV in cGy at d  max  for 6 MV photons.  /> = ( 4 2 0 2 ) - ^ — 4 6 9 . 6  0.5976  P = (6982)D r  6WK  - 469.6  -95-  [6.3.2]  In order to compare to equations 5.3.25 and 5.3.26, Equation 6.3.2 is written as dose as a i  function of EPID total pixel value PT. D  = (1.430x10~ )i> +0.06717 cGy  [6.3.3]  4  6MV  Equation 5.3.25, the dose (cGy) as a function of EPID total pixel value for varying thickness of solid water, and the above result (Equation 6.3.3) are plotted in Figure 6.3.3 for visual comparison.  EPID Calibration Curves for 6 MV 160  6 MV  140 120  e> u  100 80  o (0  ao  60 40 20 0 100000  200000  300000  400000  500000  600000  EPID Total Pixel Value P  700000  800000 900000  T  Figure 6.3.3: Comparison of the EPID dose calibration curves for 6 MV photons. Equation 5.3.25 is determined by varying the thickness of solid water (diamonds - upper curve) and equation 6.3.3 is determined by varying the fluence (squares - lower line). It is apparent from Figure 6.3.3 that the response of the EPID is similar under the two conditions of varying the dose for 6 MV photons. This is very promising for the prospect of using the EPID as a dosimeter as there is a huge variance in scattering for these two situations. In  -96-  order to quantitatively compare the two curves the percentage difference is determined and shown as Equation 6.3.4. o/.D ff. i  ( 1 =  -  7 1 M 0  " ^ " ' " '- (1.430x10~ )(P ) + 0.06717 )  >  ,  ,  ( 1  1  4 3 t e l 0  , ) ( /  )  0 0 6 7 l 7  ' 00% , 1  4  r  [6.3.4]  This result is plotted in Figure 6.3.4 as a function of total pixel value PT, for visual inspection.  Percent Difference in Dose from the Two 6 MV Calibration Curves as a function of EPID Total Pixel Value 6  MV  * 3  i  •  •  •  2  •  CD 0)  0.  -1 100000 200000 300000 400000  500000 600000 700000  EPID Total Pixel Value P  800000 900000  T  Figure 6.3.4: Percentage difference of the two EPID dose calibration curves for 6 MV photons shown in Figure 6.3.3. It can be observed from Figure 6.3.4 that the response of the EPID under conditions of changing thickness of solid water is different to its response under conditions of changing fluence compared to an ion chamber for 6 MV photons. This implies that for accurate dosimetric verification for radiotherapy treatments, the calibration algorithm for the EPID will require information about the depth of tissue that the beam has gone through. The percent difference -97-  decreases with increasing total pixel value PT (decreasing depth), which is expected since the fluence measurements were taken at shallow depth and had the exact same conditions as the thickness measurements at d  max  . The two calibrations should converge towards the point of  common settings (d ). Since the scattering condition changes with thickness of solid water and max  the EPID has a different effective atomic number than the ion chamber, the EPID calibrations diverge at greater depths. For 18 MV photons the comparison of the EPID signal and the ion chamber reading is illustrated in Figure 6.3.5.  Ion Chamber Reading (nC) vs. EPID Total Pixel Value P  T  Calibration Curve for 18 MV Photons 18 16 a 14 cr  18 MV  °> 12  8 10 Xi  E ra  o °  4  Qi = C  (2 701E-05)P + 7 608E-02 R = 1 000E-01 T  7  2 0 100000  200000  300000  400000  EPID Total Pixel Value P  500000  600000  700000  T  Figure 6.3.5: Ion chamber reading (nC) versus EPID total pixel value for varying fluence at 18 MV.  -98-  As with the 6 MV photons the relationship between the E P I D signal and the ion chamber reading is linear. E P I D total pixel value is plotted versus monitor units used in Figure 6.3.6 in order to determine the associated linear relationship. Dose (Monitor Units) vs. EPID Total Pixel Value Calibration Curve for 18 MV 18 MV  "c L.  O  "E o MU = 2 348E-04(P ) + 4.832E-01 T  R = 1 000E+00 2  0  100000  200000  300000 400000 500000 EPID Total Pixel Value P  600000  700000  T  Figure 6.3.6: EPID total pixel value versus monitor units used for 18 M V photons. As for 6 MV photons, the relationship for 18 MV photons between fluence (MU) and E P I D signal is linear. Combining the equation in Figure 6.3.6 with equation 1.9.6 and using a TMR value of 1.000, a relationship is obtained for the E P I D total pixel value as a function of dose D I 8 M V t d a  max  for 18 MV. P = (4259) T  0.6024  P = (7030)£>, T  -2055  [6.3.5]  2055  [6.3.6]  -99-  Where again, the y-intercepts are insignificant compared to the pixel values measured in this study. Written as dose as a function of EPID signal Equation 6.3.6 becomes the following. D  = (1.423x10" )P +0.2924 cGy  [6.3.7]  4  xmv  r  The calibration for 18 MV photons for varying thickness of solid water was determined in chapter 5 (Equation 5.3.28) and is plotted in Figure 6.3.7 with the above result.  EPID Calibration Curves for 18 MV Photons  0  100000 200000 300000 400000 500000 600000 700000 800000 900000 EPID Total Pixel Value P  Figure  T  6.3.7: E P I D t o t a l p i x e l v a l u e v e r s u s d o s e ( c G y ) u s e d f o r 18 M V p h o t o n s .  Figure 6.3.7 shows that the two calibration curves are very similar but not exactly the same. The percentage difference between the doses calculated from the two calibration functions is shown below. %Diff.=  (  U  0  5  j  d  ° - ^ ' " - 0 - 4 2 3 » 1 0 ^ ) > -0.2924 (1.423x10" )P +0.2924 )  )  r  4  r  This result is plotted in Figure 6.3.8. -100-  Percent Difference in Dose from the Two 18 MV Calibration Curves as a function of EPID Total Pixel Value  0  100000  200000  300000  400000  500000  600000  EPID Total Pixel Value P  700000  800000  900000  T  Figure 6.3.8: Percent Difference of the calibration curves for 18 MV as a function of the EPID total pixel value. Figure 6.3.8 shows the percent difference between the two calibration curves in the range of pixel values of 50,000 to 900,000, corresponding to the range of doses of 7 cGy to 120 cGy. The maximum percentage difference is 2.5%. The percentage difference increases with increasing pixel value (decreasing depth).  6.4 Conclusion The E P I D response is linear to that of the ion chamber as the dose changes for both energies. The calibration equations of dose as a function of total pixel value PT for the two energies are different as shown below. D  =(1.430x10- )P + 0.06717 cGy 4  m  v  r  -101 -  [6.3.3].-  A s w =0 -423x10~*)P +0.2924 cGy  [6.3.7]  T  This means that the calibration equations will depend on the energies of the photons and therefore depth since the spectrum of the radiation changes with depth. In order to account for this, the constants in the above functions will have to be functions of depth themselves. Those functions should be determined so that the imager may be used as a dosimeter more accurately. As mentioned earlier, if points off the central axis are to be measured, then the calibration constants may also have to be functions of position in order to be accurate. Keeping in mind that all of this work has been done at a collimator setting of 10 x 10 cm and at a constant SDD, the 2  constants will be functions of depth, field size, energy, SDD, and off-axis position. Once these calibration factors are determined the EPID should serve as a reliable dosimeter (with a standard deviation of 1.3%, see Chapter 2).  -102-  Chapter 7 Conclusions and Recommendations for Further Work In Chapter 2, the dose rate stability of the aSi EPID was determined. The standard deviation of the dosimetric response of the EPID under these conditions was found to be 1.3%. A constant relating the dose to the total pixel value of the EPID at d  for a 13 x 13 cm field at a 2  max  source to detector distance of 130 cm is determined to be 1.42 x 10" cGy/pix for 18 MV and 4  1.44 x 10" cGy/pix for 6 MV with a standard deviation of 1.3%. The corresponding equations, 4  are 6MV  D  D  = (1 -44x10" )PT  = (1.42xlO" )i 4  xmv  for 6 MV, and  4  J r  for 18 MV.  [2.4.1 ] [2.4.2]  In Chapter 5 the response of the EPID compared to an ion chamber is examined under conditions of changing dose due to changing thickness of solid water placed on top, and the calibration equations for those conditions have been determined to be D  =(1.715x10- )(P )° 4  mv  9879  D  r  = (1.105x10~ )(i>.) 4  nMV  1020  for 6 MV, and  [5.3.25]  forl8MV.  [5.3.28]  The constants in Equations 5.3.25 and 5.3.28 are very different from the constants determined in Chapter 2. However, the exponents in the equations compensate well for this in the range of dose investigated. In Chapter 6 the EPID response was compared to that of an ion chamber under conditions of changing dose due to changing monitor units with a constant thickness. The response was determined to be linear and the calibrations were found to be D  =(1.430x10' )P +0.06717 cGy for 6 MV, and  [6.3.3]  = (1.423x10- )P +0.2924 cGy for 18 MV.  [6.3.7]  4  6MV  D  T  4  imy  r  - 103 -  These equations are equal to the equations of Chapter 2 within experimental error. All three equations for 6 MV photons are shown in Figure 7.1 and for 18 MV photons in Figure 7.2. In both of the figures above the calibration curves are equal within the tolerance found in Chapter 2 (a standard deviation of 1.3%). In Figure 7.1 for 6 MV photons the maximum percentage difference that occurs between any of the curves is about 3% and, as expected it is between the curve for equation 5.3.25 and the other two at doses under 10 cGy. The low doses correspond to greater depths of solid water for equation 5.3.25 and the other two equations were derived using relatively shallow thicknesses of solid water (d ). In Figure 7.2 for 18 MV max  photons the curves are all within 2 percent of each other over 15 cGy and are within 1% of each other over 25 cGy. The maximum deviations for 18 MV are over 5% and they occur for doses less than 10 cGy.  -104-  Dose (cGy) vs. EPID Pixel Value as Determined by Three Different Methods for 6 MV Photons 110 105  • 2.4.1 .L_v_  • 5.3.25 A 6.3.3  100 95 90 85 80 75 70 65 >;  o  6 0  & 55 <D (A O  50 45  g 8  40 35 30 25 20 i  15 10 5 0 0  100000  200000  300000  400000 500000  600000  700000 800000  EPID Pixel Value  Figure 7.1: Comparison of the three 6 M V calibration curves of dose as a function of EPID total pixel value. - 105-  Dose (cGy) vs. Pixel Value as Determined by Three Methods for 18 MV Photons 110  * 2.4.2  105  o 5.3.28  100  A  6.3.7  • N M  95  8  90 85 80 75 70 65 & u ^ 55 6  0  (A  £  50 45 40 35 30 25 20 15 10 5 0 0  100000 200000 300000 400000 500000 600000 700000 800000 EPID Pixel Value  Figure 7.2: Comparison of the three 18 M V calibration curves of dose as a function of EPID total pixel value. -106-  In Chapter 4 the constant in equation 1.10.6 was determined for 6 MV photons using film dosimetry to be 1.44 x 10" cGy/pix with a standard deviation of 4%. This is equal to the value of 4  1.44 x 10" cGy/pix determined in Chapter 2. It is seen (Figures 4.3.3 to 4.3.10) that the drop off 4  in dose at the field edges seems sharper for the film images but this is most likely due to a combination of effects including scattering of the EPID build-up and cover, the EPID imaging algorithm, and over-response of the film compared to the EPID to low energy photons. The first two effects can be corrected for with imaging algorithms. The contribution of dose resulting from scattering from the inherent build-up of the imager should be investigated by comparing fields of different sizes including pencil beams. The third effect implies that we should compare the EPID signal to ion chamber readings for various photon energies. In Chapter 3 the equivalent thickness of the inherent buildup of the EPID was determined to be 1.38 ±0.06 cm for 6 MV and 2.24 ± 0.10 cm for 18MV from d  max  measurements. In  Chapter 5 however it was found that the solid water equivalent depth of the inherent buildup varied with thickness of solid water due to spectral variation. Using the notion of equivalent depth to correct for EPID signal compared to ion chamber signal seems to be a cumbersome method compared to using a direct calibration equation or look-up table. Since many treatments involve rotation of the gantry during irradiation, the EPID will have to rotate with the gantry in order to take images of the treatment. Therefore the most practical use of the EPID will be with no additional build-up, and future work should include calibrating the EPID with no additional build-up. Perhaps two or three practical SDDs could be used, and in order to most accurately calibrate the EPID for dose, the calibration equation should be a function of monitor units, dose rate, field size, energy, SDD, off-axis position, depth of tissue before isocenter, and depth of tissue after isocenter (see Figure 7.3). -107-  Proposed Experiment  Figure 7.3: Suggestion for experimental set-up for calibration of the EPID for dosimetric verification.  With proper calibration in this fashion, the EPID should be able to accurately measure dose to within the tolerances determined in Chapter 2 (a standard deviation of 1.3%).  -108-  Bibliography 1. Siyong Kim, Hilary C. Akpati, Jerrold E. Kielbasa, Jonathan G. Li, Chihray Liu, Robert J. Amdur, and Jatinder R. Palta, "Evaluation of intrafraction patient movement for CNS and head & neck IMRT," Med. Phys. 31, 500 (2004) 2. Tiezhi Zhang, "Respiratory gating and 4-D tomotherapy," Med. Phys. 31, 3529 (2004) 3. Indrin J. Chetty, Mihaela Rosu, Neelam Tyagi, Lon H. Marsh, Daniel L. McShan, James M. Baiter, Benedick A. Fraass, and Randall K. Ten Haken, "A fluence convolution method to account for respiratory motion in three-dimensional dose calculations of the liver: A Monte Carlo study," Med. Phys. 30, 1776 (2003) 4. Chen-Shou Chui, Ellen Yorke, and Linda Hong, "The effects of intra-fraction organ motion on the delivery of intensity-modulated field with a multileaf collimator," Med. Phys. 30, 1736(2003). 5. A. L. Fielding, P. M. Evans, and C. H. Clark, "The use of electronic portal imaging to verify patient position during intensity-modulated radiotherapy delivered by the dynamic MLC technique," Int. J. Radiat. Oncol, Biol, Phys. 54, 1225 (2002). 6. E. C. Ford, G. S. Mageras, E. Yorke, K. E. Rosenzweig, R. Wagman, and C. C. Ling, "Evaluation of respiratory movement during gated radiotherapy using film and electronic portal imaging," Int. J. Radiat. Oncol, Biol, Phys. 52, 522 (2002). 7. S. C. Erridge, Y. Seppenwoolde, S. H. Muller, M. van Herk, K. De Jaeger, J. S. Belderbos, L. J. Boersma, and J. V. Lebesque, "Portal imaging to assess set-up errors, tumor motion and tumor shrinkage during conformal radiotherapy of non-small cell lung cancer," Radiother. Oncol. 66, 75 (2003). 8. R. S. Kaatee, M. J. Olofsen, M. B. Verstraate, S. Quint, and B. J. Heijmen, "Detection of organ movement in cervix cancer patients using a fluoroscopic electronic portal imaging device and radiopaque markers," Int. J. Radiat. Oncol, Biol, Phys. 54, 576 (2002). 9. H. D. Kubo, E. G. Shapiro, and E. J. Seppi, "Potential and role of a prototype amorphous silicon array electronic portal imaging device in breathing synchronized radiotherapy," Med Phys. 26,2410(1999). 10. A. Bakai, F. Paulsen, L. Plasswilm, M. Bamberg, and F. Nusslin, "Positioning accuracy in conformational prostatic irradiation using portal imaging," Strahlenther: Onkol. 178, 84 (2002). 11. Burman C, Chui C, Kutcher G, et al, "Planning, delivery, and quality assurance of intensitymodulated radiotherapy using dynamic multileaf collimator: a strategy for large-scale  -109-  implementation for the treatment of carcinoma of the prostate," Int. J. Radiat. Oncol, Biol, Phys. 39, 863 (1997). 12. M. Woo, P. Charland, B. Kim, and A. Nico, "Commissioning, evaluation, quality assurance and clinical application of a virtual micro MLC technique," Med. Phys. 30, 138 (2003). 13. P. Munro, "Portal Imaging Technology: Past, Present, and Future," Semin. Radiat. Oncol. 5, 115 (1995). 14. L. E. Antonuk, "Electronic portal imaging devices: a review and historical perspective of contemporary technologies and research," Phys. Med. Biol. 47, R31 (2002). 15. Ellen E. Grein, Richard Lee, and Kurt Luchka, "An investigation of a new amorphous silicon electronic portal imaging device for transit dosimetry," Med. Phys. 29, 2262 (2002). 16. Peter B. Greer and Carmen C. Popescu, "Dosimetric properties of an amorphous silicon electronic portal imaging device for verification of dynamic intensity modulated radiation therapy," Med. Phys. 30, 1618 (2003). 17. B. Warkentin, S. Steciw, S. Rathee, and B. G. Fallone, "Dosimetric IMRT verification with a flat-panel EPID," Med. Phys. 30, 3143 (2003) 18. "PortalVision™ aS500 Rel. 6 Reference Manual," Varian Medical Systems Inc., 1997. 19. Rafael C. Gonzalez, Richard E. Woods, "Digital Image Processing," Addison-Wesley Publishing Company, 1992. 20. H. E. Johns and J. R. Cunnungham, "The Physics of Radiology, Fourth Edition," Springfield IL, Charles C. Thomas, 1983. 21. F. M. Khan, "The Physics of Radiation Therapy, 3 Edition," Philadelphia, PA, Lippincott Williams & Wilkins, 2003. rd  22. Yunping Zhu and Jason Inhwan Yeo, "Portal dosimetry using x-ray film: An experimental and computational study," Med. Phys. 26, 2403 (1999). 23. X. R. Zhu, S. Yoo, P. A. Jursinic, D. F. Grimm, F. Lopez, J. J. Rownd, and M. T. Gillin, "Characteristics of sensitometric curves of radiographic films," Med. Phys. 30, 912 (2003). 24. P. R. Almond, P. J. Biggs, B.M. Coursey, W.F. Hanson, M.S. Huq, R. Nath, D.W.O. Rogers, "AAPM's TG-51 Protocol for Clinical Reference Dosimetry of High-Energy Photon and Electron Beams," Med. Phys. 26, 1847 (1999). 25. B. M. McCurdy, K. Luchka, and S. Pistorius, "Dosimetric investigation and portal dose image prediction using an amorphous silicon electronic portal imaging device," Med Phys. 28,911 (2001). -110-  26. P. Munro and D. C. Bouius, "X-ray quantum limited portal imaging using amorphous silicon flat-panel arrays," Med. Phys. 25, 689 (1998). 27. V. N. Hansen, P. M. Evans, and W. Swindell, "The application of transit dosimetry to precision radiotherapy," Med. Phys. 23, 713 (1996). 28. F. H. Attix, "Introduction to Radiological Physics and Radiation Dosimetry," New York, NY, John Wiley & Sons Inc., 1986. 29. J. Leong, "Use of digital fluoroscopy as an on-line verification device in radiation therapy," Phys. Med. Biol. 31,985 (1986). x  30. J. W. Wong, E. D. Slessinger, R. E. Hermes, C. J. Offutt, T. Roy, and M. W. Vannier, "Portal dose images I: quantitative treatment plan verification," Int. J. Radiat. Oncol., Biol., Phys. 18, 1455 (1990). 31. X. Ying, L. Y. Geer, and J. W. Wong, "Portal dose images II: patient dose estimation," Int. J. Radiat. Oncol., Biol., Phys. 18, 1465 (1990) 32. M. C. Kirby and P. C. Williams, "The use of an electronic portal imaging device for exit dosimetry and quality control measurements," Int. J. Radiat. Oncol., Biol., Phys. 31, 593 (1995). 33. M. Kroonwijk, K. L. Pasma, S. Quint, P. C. M. Koper, A. G. Visser, and B. J. M. Heijmen, "In vivo dosimetry for cancer patients using an electronic portal imaging device (EPID); detection of internal organ motion." Radiother. Oncol. 49, 124 (1998). 34. V. Hansen, W. Swindell, and P. Evans, "Extraction of primary signal from EPIDs using only forward convolution," Med. Phys. 24, 1477 (1997). 35. M. C. Kirby and P. C. Williams, "Measurement possibilities using an electronic portal imaging device," Radiother. Oncol. 29, 237 (1993). 36. M. Essers, R. Boellaard, M. Van Herk, J. H. Lanson, and B. J. Mijnheer, "Transmission dosimetry with a liquid filled electronic portal imaging device," Int. J. Radiat. Oncol., Biol., Phys. 34, 931 (1996). 37. C. Fiornio, A. del Vecchio, G. M. Cattaneo, M. Fusca, B. Longobardi, P. Signorotto, and R. Calandrino, "Exit dose measurements by portal film dosimetry," Radiother. Oncol. 29, 336 (1993). 38. C. Fiomio, C. Uleri, G. M. Cattaneo, and R. Calandrino, "On-line exit dose profile measurements by a diode linear array," Phys. Med. Biol. 41, 1291 (1996).  -111 -  39. R. Boellard, M. van Herk, and B.J. Mijnheer, "A convolution model to convert transmission dose images to exit dose distributions" Med. Phys. 24, 189 (1997). 40. R. Boellard, M. van Herk, H. Uiterwaal, and B.J. Mijnheer j "First clinical tests using a liquid-filed electronic portal imaging device and a convolution model for the verification of the midplane dose" Radiother. Oncol. 47, 303 (1998). 41. R. Boellard, M. van Herk, H. Uiterwaal, and B.J. Mijnheer, "Two dimensional exit dosimetry6 using a liquid-filled electronic portal imaging device and a convolution model" Radiother. Oncol. 44, 149 (1997). 42. R. Boellard, M. Essers, M. van Herk, and B.J. Mijnheer, "New method to obtain the midplane dose using portal in vivo dosimetry" Int. J. Radiat. Oncol, Biol., Phys. 41, 465 (1998). 43. B. J. M. Heijmen, K. L. Pasma, M. Kroonwijk, V. G. M. Althof, J. C. J. deBoer, A. G. Visser, and H. Huizenga, "Portal dose measurement in radiotherapy using an electronic portal imaging device (EPID)," Phys. Med. Biol. 40, 1943 (1995). 44. K. L. Pasma, B. J. M. Heijmen, M. Kroonwijk, and A. G. Visser, "Portal dose image (PDI) prediction for dosimetric treatment verification in radiotherapy. I. An algorithm for open beams," Med. Phys. 25, 830 (1998). 45. K. L. Pasma, M. Kroonwijk, J. C. J. deBoer, A. G. Visser, and B. J. M. Heijmen, "Accurate portal dose measurement with a fluoroscopic electronic portal imaging device (EPID) for open and wedged beams and dynamic multileaf collimation," Phys. Med. Biol. 43, 2047 (1998). 46. T. R.. McNutt, T. Mackie, P. Reckwerdt, and N. Papanikolaou, "Calculation of portal dose using the convolution/superposition method," Med. Phys. 23, 527 (1996). 47. Lijun Ma, N. Phaisangittisakul, C. X. Yu, and M. Sarfaraz, "A quality assurance method for analyzing and verifying intensity modulated fields," Med. Phys. 30, 2082 (2003). 48. Jingeng Zhu, Fang-Fang Yin, and Jae Ho Kim, "Point dose verification for intensity modulated radiosurgery using Clarkson's method," Med. Phys. 30, 2218 (2003). 49. Cheng B. Saw, Sicong Li, Komanduri M. Ayyangar, Maung Yoe-Sein, Susha Pillai, Charles A. Enke, and Juan C. Celi, "Dose linearity and uniformity of a linear accelerator designed for implementation of multileaf collimation system-based intensity modulated radiation therapy," Med. Phys. 30, 2253 (2003). 50. Daniel A. Low, Parag Parikh, James F. Dempsey, Sasha Wahab, and Saiful Huq, "Ionization chamber volume averaging effects in dynamic intensity modulated radiation therapy beams," Med. Phys. 30, 1706 (2003). -112-  51. Geetha V. Menon and Ron S. Sloboda, "Compensator quality control with an amorphous silicon EPID," Med. Phys. 30, 1816 (2003). 52. Chester Ramsey, Scott Dube, and William R. Hendee, Moderator, "It is necessary to validate each individual IMRT treatment plan before delivery," Med. Phys. 30, 2271 (2003). 53. Mark Jobse, Jaap Davelaar, Emile Hendriks, Rob Kattevilder, Hans Reiber, and Berend Stoel, "A new algorithm for the registration of portal images to planning images in the verification of radiotherapy, as validated in prostate treatments," Med. Phys. 30, 2274 (2003). 54. D. M. Shepard, Z. Jiang, M. A. Earl, M. C. Ferris, J. Lim, and S. Naqvi, "A toolbox for intensity modulated radiation therapy optimization," Med. Phys. 30, 2320 (2003). 55. M. K. Woo, A. W. Lightstone, G. Shan, L. Kumaraswamy, and Y. Li, "Automatic verification of step-and-shoot IMRT field segments using portal imaging," Med. Phys. 30, 348 (2003). 56. Sanjiv S. Samant, Wei Zheng, Nestor Andres Parra, Jason Chandler, Arun Gopal, Jian Wu, Jinesh Jain, Yunping Zhu, and Marc Sontag, "Verification of multileaf collimator leaf positions using an electronic portal imaging device," Med. Phys. 29, 2900 (2002). 57. E. C. Ford, J. Chang, K. Mueller, K. Sidhu, D. Todor, G. Mageras, E. Yorke, C. C. Ling, and H. Amols, "Cone-beam CT with megavoltage beams and an amorphous silicon electronic portal imaging device: Potential for verification of radiotherapy of lung cancer," Med. Phys. 29, 2913 (2002). 58. A. E. Schach von Wittenau, C. M. Logan, M. B. Aufderheide III, and D. M. Slone, "Blurring artifacts in megavoltage radiography with a flat-panel imaging system: Comparison of Monte Carlo simulations with measurements," Med. Phys. 29, 2559 (2002). 59. G. Pang and J. A. Rowlands, "Development of high quantum efficiency flat panel detectors for portal imaging: Intrinsic spatial resolution," Med. Phys. 29, 2274 (2002). 60. Sandra C. Vieira, Maarten L. P. Dirkx, Kasper L. Pasma, and Ben J. M. Heijmen, "Fast and accurate leaf verification for dynamic multileaf collimation using an electronic portal imaging device, " Med. Phys. 29, 2034 (2002). 61. Lennert S. Ploeger, Monique H. P. Smitsmans, Kenneth G. A. Gilhuijs, and Marcel van Herk, "A method for geometrical verification of dynamic intensity modulated radiotherapy using a scanning electronic portal imaging device," Med. Phys. 29, 1071 (2002). 62. Youcef El-Mohri, Kyung-Wook Jee, Larry E. Antonuk, Manat Maolinbay, and Qihua Zhao, "Determination of the detective quantum efficiency of a prototype, megavoltage indirect detection, active matrix flat-panel imager," Med. Phys. 28, 2538 (2001).  -113-  63. Jenghwa Chang, Gikas S. Mageras, Clifton C. Ling, and Wendell Lutz, "An iterative EPID calibration procedure for dosimetric verification that considers the EPID scattering factor," Med. Phys. 28,2247 (2001). 64. G. Pang, D. L. Lee, and J. A. Rowlands, "Investigation of a direct conversion flat panel imager for portal imaging," Med. Phys. 28, 2121 (2001). 65. M. Lachaine, E. Fourkal, and B. G. Fallone, "Investigation into the physical characteristics of active matrix flat panel imagers for radiotherapy," Med. Phys. 28, 1689 (2001). 66. D. Buck, M. Alber, and F. Nusslin, "Potential and limitations of the automatic detection of fiducial markers using an amorphous silicon flat-panel imager," Phys. Med. Biol. 48, 763 (2003). 67. J. J. Kruse, M. G. Herman, C. R. Hagness, B. J. Davis, Y. I. Garces, M. G. Haddock, K. R. Olivier, S. L. Stafford, and T. M. Pisansky, "Electronic and film portal images: a comparison of landmark visibility and review accuracy," Int. J. Radiat. Oncol., Biol., Phys. 54, 584 (2002). 68. S. C. Vieira, M. L. Dirkx, K. L. Pasma, and B. J. Heijmen, "Dosimetric verification of x-ray fields with steep dose gradients using an electronic portal imaging device," Phys. Med. Biol. 48, 157(2003). 69. J. Pouliot, M. Aubin, K. M. Langen, Y. M. Liu, B. Pickett, K. Shinohara, and M. Roach 3 , "(Non)-migration of radiopaque markers used for on-line localization of the prostate with an electronic portal imaging device,'* Int. J. Radiat. Oncol., Biol., Phys. 56, 862 (2003). rd  70. P. Winkler, H. Bergmann, G. Stuecklschweiger, and H. Guss, "Introducing a system for automated control of rotation axes, collimator and laser adjustment for a medical linear accelerator," Phys. Med. Biol. 48, 1123 (2003). 71. L. Lord, S. May, M. Bailey, and L. Smith, "Is one head and neck immobilization system as good as another? One center's experience," Med. Dosim. 28, 39 (2003). 72. G. Liu, T. van Doom, and E. Bezak, "Assessment of flatness and symmetry of megavoltage x-ray beam with an electronic portal imaging device (EPID)," Australas. Phys. Eng. Sci. Med. 25, 58 (2002). 73. K. K. Brock, D. L. McShan, and J. M. Baiter, "A comparison of computer-controlled versus manual on-line patient setup adjustment," J. Appl. Clin. Med. Phys. 3, 241 (2002). . 74. D. A. Low, E. E. Klein, D. K. Maag, W. E. Umfleet, and J. A. Purdy, "Commissioning and periodic quality assurance of a clinical electronic portal imaging device," Int. J. Radiat. Oncol., Biol, Phys. 34, 117 (1996).  -114-  75. E. Spezi and D. G. Lewis, "Full forward Monte Carlo calculation of portal dose from MLC collimated treatment beams," Phys. Med. Biol. 47, 377 (2002). 76. F. Claus, Y. De Deene, V. Remouchamps, C. De Wagter, K. Van Vaerenbergh, J. Schauvliege, B. Van Duyse, B. Speleers, W. Duthoy, and W. De Neve, "An isocenter position verification device for electronic portal imaging: physical and dosimetrical characteristics," Cancer Radiother. 5, 759 (2001). 77. J. L. Lagrange and F. Lipinski, "The place of conformal radiotherapy in routine practice: national survey of SFRO members," Cancer Radiother. Suppl 1, 44 (2001). 78. J. Chang, G. S. Mageras, C. C. Ling, and W. Lutz, "An iterative EPID calibration procedure for dosimetric verification that considers the EPID scattering factor," Med. Phys. 28, 2247 (2001). 79. V. Casanova Borca, M. Pasquino, M. Bertodatto, S. Tofani, "A quality control test of the image obtained with electronic portal imaging devices," Radiol. Med. (Torino). 102, 266 (2001). 80. M. G. Herman, J. J. Kruse, and C. R. Hagness, "Guide to clinical use of electronic portal imaging," J. Appl. Clin. Med. Phys. 1, 38 (2000). 81. K. A. Langmack, "Portal imaging," Br. J. Radiol. 74, 789 (2001). 82. L. Spies, M. Partridge,B. A. Groh, and T. Bortfeld, "An iterative algorithm for reconstructing incident beam distributions from transmission measurements using electronic portal imaging," Phys. Med. Biol. 46, N203 (2001). 83. O. Pradier, A. Hille, H. Schmidberger, and C. F. Hess, "The role of high-energy imaging in a radiotherapy service and its incorporation in a network," Cancer Radiother. 5(3), 246-54 (2001). 84. A. Van Esch, B. Vanstraelen, J. Verstraete, G. Kutcher, and D. Huyskens, "Pre-treatment dosimetric verification by means of a liquid-filled electronic portal imaging device during dynamic delivery of intensity modulated treatment fields," Radiother Oncol. 60,181 (2001). 85. A. G. Glendinning, S. G. Hunt, and D. E. Bonnett, "Recording accelerator monitor units during electronic portal imaging: application to collimator position verification during IMRT," Phys. Med. Biol. 46, N159 (2001). 86. M. G. Herman, J. M. Baiter, D. A. Jaffray, K. P. McGee, P. Munro, S. Shalev, M. Van Herk, and J. W. Wong, "Clinical use of electronic portal imaging: report of AAPM Radiation Therapy Committee Task Group 58," Med. Phys. 28, 712 (2001).  -115-  87. P. M. Evans, M. Partridge, and J. R. Symonds-Tayler, "Sampling considerations for intensity modulated radiotherapy verification using electronic portal imaging," Med. Phys. 28, 543 (2001). 88. L. S. Ploeger, M. H. Smitsmans, K. G. Gilhuijs, and M. van Herk, "Accurate measurement of the dynamic response of a scanning electronic portal imaging device," Med. Phys. 28, 310 (2001). 89. A. G. Glendinning, S. G. Hunt, and D. E. Bonnett, "A method for controlling image acquisition in electronic portal imaging devices," Phys. Med. Biol. 46, N39 (2001). 90. A. G. Glendinning, S. G. Hunt, and D. E. Bonnett, "Measurement of the response of Gd202S:Tb phosphor to 6 MV x-ray's," Phys. Med. Biol. 46, 517 (2001). 91. G. Liu, T. van Doom, and E. Bezak, "Evaluation of the mechanical alignment of a linear accelerator with an electronic portal imaging device (EPID)," Australas. Phys. Eng. Sci. Med. 23, 74 (2000). 92.1. Barillot, P. Maingon, G. True, B. Steyer, S. Naudy, and J. C. Horiot, "Assessment of the use of electronic portal imaging,"Cancer Radiother. Suppl. 1, 25s (2000). 93. K. E. Hatherly, J. C. Smylie, A. Rodger, M. J. Dally, S. R. Davis, and J. L. Millar, "A double exposed portal image comparison between electronic portal imaging hard copies and port films in radiation therapy treatment setup confirmation to determine its clinical application in a radiotherapy center," Int. J. Radiat. Oncol., Biol, Phys. 49, 191 (2001). 94. M. Partridge, J. R. Symonds-Tayler, and P.M. Evans, "IMRT verification with a camerabased electronic portal imaging system," Phys. Med. Biol. 45, N183 (2000). 95. D. Plattard, M. Soret, J. Troccaz, P. Vassal, J. Y. Giraud, G. Champleboux, X. Artignan, and M. Bolla, "Patient set-up using portal images: 2D/2D image registration using mutual information," Comput. Aided Surg. 5, 246 (2000). 96. M. Partridge, P. M. Evans, M. van Herk, L. S. Ploeger, G. J. Budgell, and H. V. James, "Leaf position verification during dynamic beam delivery: a comparison of three applications using electronic portal imaging," Med. Phys. 27, 1601 (2000). 97. A. G. Glendinning, and D. E. Bonnett, "Dosimetric properties of the Theraview fluoroscopic electronic portal imaging device," Br. J. Radiol. 73, 517 (200Q). 98. J. Chang, G.S. Mageras, C. S. Chui, C. C. Ling, and W. Lutz, "Relative profile and dose verification of intensity-modulated radiation therapy," Int. J. Radiat. Oncol., Biol., Phys. 47, 231 (2000).  -116-  99.1. Barillot, P. Maingon, G. True, A. D'Hombres, B. Steyer, S. Naudy, and J. C. Horiot, "Evolution of the use of the portal imaging device: prospective study over three years," Cancer Radiother. 4, 40 (2000). 100. S. Lavertu, L. M. Girouard, and J. Pouliot, "Observation study of electronic portal images for off-line verification," Radiother. Oncol. 54, 47 (2000). 101. Chiu-Tsao ST, Duckworth T, Zhang C, Patel NS, Hsiung CY, Wang L, Shih JA, Harrison LB, "Dose response characteristics of new models of GAFCHROMIC films: dependence on densitometer light source and radiation energy," Med. Phys. 31, 2501 (2004).  -117-  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0092046/manifest

Comment

Related Items