Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Measurement of proton beam dose profiles using a sensitive scintillation screen observed with a CCD camera Ryneveld, Susan C. 1998

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

Item Metadata


831-ubc_1998-0600.pdf [ 5.26MB ]
JSON: 831-1.0085162.json
JSON-LD: 831-1.0085162-ld.json
RDF/XML (Pretty): 831-1.0085162-rdf.xml
RDF/JSON: 831-1.0085162-rdf.json
Turtle: 831-1.0085162-turtle.txt
N-Triples: 831-1.0085162-rdf-ntriples.txt
Original Record: 831-1.0085162-source.json
Full Text

Full Text

MEASUREMENT OF PROTON BEAM DOSE PROFILES USING A SENSITIVE SCINTILLATION SCREEN OBSERVED WITH A CCD CAMERA by SUSAN C. RYNEVELD B.Sc. University of Victoria, 1995 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Physics) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA July 1998 © Susan C. Ryneveld, 1998 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of T r 7 Y c S / The University of British Columbia Vancouver, Canada DE-6 (2/88) A B S T R A C T The Proton Therapy Facility at TRIUMF has been in routine operation since 1995 using 70 MeV protons extracted from the 500 MeV H" cyclotron to treat ocular melanomas. The main rationale behind using proton therapy over conventional therapy is its improved conformity (i.e. manipulating the beam to deliver maximum dose to the tumour volume and minimum dose to the healthy surrounding tissue) and uniformity of dose delivery. Efforts have been made to develop a rapid method of measuring dose profiles using a sensitive scintillating screen observed by a CCD camera. The main advantage of this sort of dosimetry system is its ability to do near real-time observation and analysis of the proton beam; the system offers a practical and timesaving method of quality control. Conventional proton dosimetry methods, which consist of a miniature detector scanned in a water box, can be slow and impractical for some studies. Experiments have been carried out using beam energies and intensities well matched to clinical doses. A scintillating screen enclosed in a light tight box is placed in the path of the proton beam, and is observed by a CCD camera. The CCD camera, which views the screen through a 45-degree mirror has an integrating mode for increased sensitivity and a frame grabber for immediate viewing. Various commercial intensifying screens have been evaluated and one screen Rarex PFG (ZnCdS:Ag) has been found to be 17 times more sensitive than the more commonly used LANEX (Gd202S:Tb) screen. The system has demonstrated to be useful for rapid visualization of lateral dose distributions to look at collimator scattering, beam steering, raw Bragg peaks, and effect of wedges. ii Table of Contents A B S T R A C T ii T A B L E O F C O N T E N T S iii L I S T O F T A B L E S vi L I S T O F F I G U R E S vii A C K N O W L E D G M E N T S ix 2. I N T R O D U C T I O N 1 2.1. HISTORY OF PROTON THERAPY 1 2.2. RATIONALE FOR PROTONS 4 2.2.1. Radiobiological Characteristics 4 2.2.2. Absorption Characteristics 5 2.3. MOTIVATION AND SCOPE OF THESIS 6 3. B A C K G R O U N D I N F O R M A T I O N 10 3.1. INTERACTIONS WITH MATTER 10 3.1.1. X-ray interactions 10 3.1.2. Proton interactions • 14 3.2. DOSIMETRY 17 3.2.1. Ionization Chamber 17 3.2.2. Secondary Emission Monitor 18 3.2.3. Diode 18 3.2.4. Integrators 19 3.3. PROTON THERAPY FACILITY AT T R I U M F 19 3.3.1. Beam Delivery 19 3.3.2. Patient Positioning and Treatment Planning 22 3.3.3. Dose Monitoring and Measurement 23 3.4. SCINTILLATING SCREENS 24 3.4.1. History of Scintillators in Medical Imaging 24 3.4.2. Screen construction 25 3.4.3. Mechanism behind energy conversion 26 3.4.4. Light Quenching 26 3.5. CAMERAS 27 iii 3.5.1. Charge Coupled Devices 28 3.5.2. Mechanism Behind the CCD as an Optical Imaging Device 28 3.5.3. CCD Properties 30 3.5.4. Radiation Damage 31 3.6. FRAME GRABBER AND IMAGE PROCESSING SYSTEM 32 4. E X P E R I M E N T A L D E S I G N C O N S I D E R A T I O N S 34 4.1. FEASIBILITY STUDY 34 4.2. EQUIPMENT DESCRIPTION 36 4.2.1. Screen 37 4.2.2. Camera 38 4.2.3. A cquisition Module 39 4.2.4. Frame Grabber 40 4.2.5. Integration Controller. 41 4.2.6. Software 43 4.2.7. Photo-diode 43 4.3. EXPERIMENTAL SET-UP 44 4.3.1. CCD Readings : 46 4.3.2. Diode Readings 47 5. D I S C U S S I O N O F R E S U L T S 48 5.1. EVALUATION OF SCINTILLATING SCREENS 48 5.7. / . Proton Irradiation 48 5.1.2. X-ray Irradiation 50 5.1.3. Linearity of CCD signal response 52 5.1.4. Background Corrections 52 5.2. B E A M MONITORING 55 5.2.1. Uniformity 55 5.2.2. Beam Steering Effects 56 5.3. DOSIMETRY 58 5.3.1. Bragg Peak 58 5.3.2. Saturation Effects- Loss of Intensity 60 5.4. OTHER PHENOMENA 62 5.4.1. Edge Enhancement Effect 62 5.4.2. Wedges 63 5.4.3. Collimator and Gaussian Distribution 66 5.5. X -RAY FOR REAL TIME PATIENT POSITIONING 67 5.5.1. Feasibility 68 iv 5.6. RADIATION D A M A G E TO CAMERA 71 5.6.1. Hot Pixels 71 5.6.2. Relaxation 72 6. CONCLUSIONS 74 REFERENCES 78 APPENDICES 82 APPENDIX A : CALCULATION OF STOPPING POWER OF H E A V Y CHARGED PARTICLES 82 APPENDIX B : CLASSICAL ELECTRON RADIUS 87 APPENDIX C : SEMI-CONDUCTOR B A N D GAP THEORY 89 v List of Tables Table 1-1 : Clinical Results of Proton Beam Therapy at Massachusetts General Hospital 3 Table 3-1 : Properties of Tested Commercial Intensifying Screens for X-ray Absorption 38 Table 3-2 : Integration Times for Switch Settings of the Integration Controller 42 Table 4-1 : Sensitivities of Various Intensifying Screens Relative to LANEX 54 vi List of Figures Figure 1-1 : Comparison of Depth Dose Distribution for x-ray, y-ray, and electrons with protons 2 Figure 2-1 : Compton Scattering 11 Figure 2-2 : Primary and Secondary Dose Contributions at Point P 12 Figure 2-3 : Depth dose profile of x-rays into medium 13 Figure 2-4 : Proton Dose Profile Showing a Bragg Peak and a Spread Out Bragg Peak 15 Figure 2-5 : H ~ stripping 20 Figure 2-6 : Layout of Proton Therapy Facility Equipment 21 Figure 2-7 : Pattern of charge collected under electrodes is read out as an analoguesignal by the clocking array 29 Figure 3-1 : Geometry of Experimental Apparatus 35 Figure 3-2 : Flow Chart of CCD Signal Processing System 37 Figure 3-3 : Normal Mode Frame Grabber Acquisition 41 Figure 3-4 : Scintillator Screen and CCD Camera Dose Measuring Arrangement 45 Figure 3-5 : Positioning of CCD Camera/Screen Measuring Device in PTF Equipment 46 Figure 4-1 : Response to Proton Irradiation for Less Sensitive Screens 49 Figure 4-2 : Response to Proton Irradiation for More Sensitive Screens 50 Figure 4-3 : Linearity of Screens for X-ray Irradiation 51 Figure 4-4 : Linearity of CCD response for Proton Irradiation 54 Figure 4-5 : Uniformity of Lateral Dose Profile for Various Beam Intensities 56 Figure 4-6 : Effect of Beam Steering on Lateral Dose Profile 57 Figure 4-7 : Bragg Peak Plots 59 Figure 4-8 : Comparison of Bragg Peak Measurements for CCD camera and Ion Chamber 60 Figure 4-9 : Comparison of TI Chamber and Diode Collected Data 61 Figure 4-10 : Edge Effect due to Multiple Scattering Effects at the Edges of Two Materials 62 Figure 4-11 a & b : Cross Section of CCD image of Penny shows clearly the EDGE EFFECT 63 Figure 4-12 : Lateral Dose Profiles from 60 degree wedge for various Range Shifter Positions 65 Figure 4-13 : Beam Profile for Open Collimator - single integration 66 Figures 4-14 a & b : Lateral Cross Section of Beam Profile from Lead Button in Centre of Beam 67 Figure 4-15: CCD Image of Collimator with cross hairs for X-ray Exposure of 80kV, 50mA 69 Figures 4-16a, b, & c : Lateral Profiles of X-ray signal for 0, 9.5, and 23.0cm attenuated beam 70 Figure 4-17 : CCD dark frame readout showing damaged or "hot" pixels 72 vii Figure 4-18 : Graph of Relaxation of 'Hot Pixels' or Radiation Damaged Pixels 73 Figure A - l : Interaction of a Heavy Charged Particle with Matter 81 Figure A-2 : Cylindrical Shell of Randomly Distributed Electrons in Absorbing Material 84 Figure B-l : Free electron at P with Accelerations a; and a2 Due to Field Vectors E\' and E2' 86 Figure C-l : Energy bands of a semiconductor material 88 viii Acknowledgments I would l ike to thank my supervisor, Dr. Ewart Blackmore, for his support and insight throughout this project. Despite his hectic schedule, he was always available and wi l l ing to give direction, guidance, and encouragement. Thanks also to Dr. E d A u l d , for his willingness to be my academic supervisor and for his feedback and comments on this M . S c . thesis. M a n y people at T R I U M F have given freely of their time and expertise towards my M.Sc . work and I believe T R I U M F ' s success as a research facility is largely a result of its team work environment. Specific thanks go to Anna Gelbart and M i n d y Hapke i n the T R I U M F design office for sharing their expertise i n Corel Draw and helping me with my conference poster and thesis figures; M i k e Mouat for his help with the Proton Therapy Facility computer programs; Brian Evans for his advice and useful discussions about the system hardware; and to Dave Morris for his help with the P C and Visual C++ software. I would not have enjoyed my time at U B C half as much without the friendship and support of Sofia Chavez & A l i s o n Clark. I 'm going to miss our workouts, run-bike-swims, coffee times, and of course muffins!! Finally, love and thanks to Brice for his support and cheerleading throughout the course of this project. Y o u r faith i n me warms my heart! (Whew! I 'm done i n time for our wedding!!) ix 1. Introduction After the discovery of x-rays i n 1895 it soon became evident that these radiations could have a profound effect on l iv ing tissues. Since then, man-made sources of radiation have become widely used i n medicine and in particular in the treatment of cancer. However these early radiations were of limited intensity and had a limited penetration depth. Successful treatment of a cancerous tumour involves adequately irradiating the target volume while protecting the surrounding healthy cells. Historically, any advances i n the radiation treatment of cancer were due to improvements in dose delivery and localization through better machine design such as high energy linacs, refined tumour localization techniques, and more accurate treatment planning and dosimetry. Despite significant steps towards improved dose delivery, many cancers are not curable by radiation techniques used i n standard treatments with photons and electrons. Therefore there has been a strong incentive to examine other types of radiation. 1.1. History of Proton Therapy Heavy charged particle beams, and i n particular, proton beams, show the most promise due to their ability to improve both the uniformity and conformity of dose delivery. When x-rays and y-rays are absorbed into the material through which they pass, they do not create chemical and biological damage themselves. Instead they give up their energy to produce fast moving electrons, many of which can ionize other atoms i n the absorber, break vital chemical bonds, and initiate the change of events that ultimately is expressed as biological damage. Charged particles, however, such as protons, are directly ionizing. Protons interact with matter by continuous slowing down as a result of Coulomb collisions with electrons and nuclei of the 1 medium until all their energy is consumed at the end of their range. The details of these interactions w i l l be discussed further i n Chapter 2; however, at this point it w i l l suffice to say that since the relative amount of interaction increases as the particle slows, the dose increases towards the end of the particle range. This creates a sharp dose peak called the Bragg peak. This is i n complete contrast to the depth dose profiles of more coventional radiations such as 6 0 C o y-rays, high energy x-rays, and electrons as indicated in Figure 1-1. 0 10 20 30 Depth in Tissue (cm) Figure 1-1 : Comparison of Depth Dose Distribution for x-ray, y-ray, and electrons with protons It has been demonstrated that improvements i n dose delivery leads to better local control and thus overall better survival rates. [ 2 ] A s early as 1946, R .R . W i l s o n [ 3 ] proposed the use of proton beams 2 for radiation treatment. The clearly advantageous physical characteristics o f the proton distribution were cited as the main rationale for their clinical use. Since 1946, many institutions have participated and contributed to the development of proton therapy. Initiatory studies were conducted at Berkeley, Cal i fornia [ 4 ] in the 1950's to treat pituitary glands with high-energy proton beams. In the 1960's groups at Uppsala, Sweden and Harvard began developing most of the standard beam delivery techniques that are still used today. During the past twenty years ocular melanomas have been treated at the Paul Scherrer Institute (PSI) and by the Massachusetts General Hospital-Harvard Cyclotron Laboratory. These clinical studies have revealed that charged particle irradiation is an effective treatment for eye tumours (uveal melanomas) and other localized tumours as indicated i n Table 1-1. Table 1-1 : Clinical Results of Proton Beam Therapy at Massachusetts General Hospital Tumour Type No. of Patients Total Dose (Gy) Local Control Survival uveal melanoma 1006 46-90 98% 5 year survival 96% survival with vision 69% skull based sarcoma 110 70-75 84% 5 year local control 74% cervial spine tumour 26 70-75 69% 5 year local control 6 1 % meningioma: malignant 2 60-64 50% survival 18/18 100% benign 16 56-72 100% survival 18/18 100% craniopharyngioma 15 53-63 survival 13/15 87% brain tumour 9 67-75 survival 2/9 Consequently, the number of charged particle therapy facilities has been steadily increasing worldwide with a proton patient total of nearly 20,000 as of January 1998 and an overall patient total o f 25,000 for all particles. [ 5 ] One such facility is the proton therapy facility at T R I U M F , Canada. [ 6 ' 7 ] Other facilities include Loma L i n d a [ 8 ] which was the first hospital based facility and 3 the NPTC-Massachusetts General Hospital^ 9 1 which w i l l be the second and is currently being commissioned to treat the first patient i n the fall o f 1998. Heavy ion treatments are done at the Chiba facility i n Japan. 1.2. Rationale for Protons W h y choose protons? The rationale for the use of protons in cancer radiation therapy is based on their clearly superior absorption characteristics. The main strategy in the physics of radiation treatment is based on dose localization. The classical intent of radiation oncology is to deliver ionizing radiation only to the diseased tissue. In practice, however, this ideal is compromised. 1 1 1 1 Thus, it is important to optimize both conformity (ensuring a maximum dose is delivered to the tumour with a tolerable dose to the surrounding healthy tissue) and uniformity (ensuring a uniform dose is delivered to the tumour target volume). Conformity is paramount since patient survival depends on k i l l ing cancerous cells while retaining healthy cells. Uniformity is also important because local recurrence of tumours is usually determined by cold spots in the dose distribution. [ 1 2 1 Protons are also much easier and therefore cheaper to accelerate to treatment energies than heavy ions even though heavy ions do have some advantages such as reduced scattering and a sharper Bragg peak. 1.2.1. Radiobiological Characteristics A s the main objective in radiation treatment is to k i l l cancer tumours, it is also important to consider the effectiveness of the irradiation. Effectiveness can be measured in terms of the amount of induced radiation damage delivered to the cells. The concept of biological damage is 4 one of the least certain aspects about radiation therapy. However, an approximate determination can be obtained using Relative Biological Effectiveness (RBE) which is defined as the ratio of the dose of reference radiation needed to produce a biological effect to the dose of the test radiation to produce the same biological effect. Atoms with higher atomic charge (Z) have a higher loss of energy per unit distance of absorbing material. This process is described as Linear Energy Transfer (LET) . The effect of radiation on a cell may be strongly dependent on the L E T of the ionizing particle since spacing between ionizations decreases with increases in L E T . From radiological studies, it has been shown that for the same dose, radiation of a higher L E T is more effective i n cell k i l l ing compared to a lower L E T [ 1 3 ] , giving rise to a higher R B E . For neutrons, heavy ions, and pions, the R B E has been found to be much higher than that for conventional x-rays; and this property is exploited, in particular with neutrons for specific tumours. However selecting the correct dose is made difficult because the R B E may be different for different cells. One advantage of protons is that the R B E is only slightly higher (RBE=1.1) than for photons [ 1 4 ] . Since the radiobiological effects of photons are so well known, the comparable R B E value makes the photon experience valid for protons. 1 . 2 . 2 . Absorption Characteristics A s shown earlier i n Figure 1-1, the proton depth dose profile has a comparatively lower entrance dose than more conventional radiation and has a sharp cut-off. A s the beam traverses tissue, the dose deposited is approximately constant until near the end of the range where the dose peaks out to a high value followed by a rapid fall-off to zero. This peak is unique to charged particle beams and can be spread out by means of range modulator to uniformly distribute the beam across the tumour volume. Figure 1-1 shows a spread out proton (Bragg) peak. Because of the Bragg peak effect, proton beams provide a much sought after advantage i n radiotherapy - the ability to 5 concentrate dose inside the target volume and minimize dose to the surrounding normal tissue. The level of tolerance of normal tissue i n the irradiated field often determines the dose to the tumour. Sometimes this limited dose is insufficient to control the cancer. A s a result of the proton dose distribution characteristics, the radiation oncologist can increase the dose to the tumour while reducing the dose delivered to the surrounding healthy cells. The advantage to proton therapy is clearly i n its ability to localize dose. In the past, the use of proton therapy has been limited by the precision of tumour localization. W i t h recent advances in imaging technology ( M R I , C T , P E T , etc.) tumours can now be localized with sufficient precision to justify the use of "precision therapy" such as proton beams. 1.3. Motivation and Scope of Thesis The most important quantity measured i n radiotherapy is absorbed dose, which is defined as the energy absorbed per unit mass of material. Protocols i n radiotherapy for different clinical situations place strict requirements on the accuracy of the dose delivered. For charged particle beams, the requirement is that the dose be measured to an accuracy equal to or better than ± 5 % of overall uncertainty and a reproducibility of ± 2 % precision. [ 1 5 1 Ideally this can only be achieved i f the dose distribution is known at all points in the treatment volume and in the surrounding tissues. Thus the main objective of this project was to develop a position sensitive method of measuring dose profiles using a sensitive scintillating screen observed by a C C D camera. The main advantages of this sort of dosimetry system would be its ability to do near real-time observation 6 and analysis of the proton beam and of the distribution of dose into the irradiated volume. Conventional proton dosimetry methods which consist of a miniature detector scanned in a water box can be slow and impractical whereas the C C D technology would allow for immediate acquisition of information which is, as i n any medical application or treatment, to the benefit of the patient. The system could therefore offer a practical and timesaving method of quality control. Possibly, the system could be incorporated into part of the regular quality control protocol for the T R I U M F Proton Therapy Facility. In the past, the use of a C C D camera to monitor the light emitted from a thin plastic scintillator placed i n the beam path has been successfully employed for x-ray beam analysis t l 6 ' 1 7 ] . More recently, a group at the Paul Scherrer Institute (PSI) began development of a two-dimensional position sensitive dosimetry system, based on a scintillating screen mounted on the beam exit side of a water equivalent phantom viewed by a C C D camera. Fol lowing up on their report1-18] i n which they indicated that preliminary studies showed the device could be a useful tool for quick and reliable quality control of proton beams, this project was initiated i n an effort to develop a similar system for beam monitoring and rapid visualization of dose profiles. Chapter 2 outlines the background of the various subject areas included in this study. Firstly a detailed comparison between the characteristics of protons and those of conventional radiation are presented, clearly outlining the physical advantages of protons over conventional radiation. Secondly, a description of the T R I U M F Proton Therapy Facility is given along with explanations of how the equipment components are utilized to control and shape the proton beam into one that is clinically useful. Thirdly, several standard dosimetry techniques are 7 described, leading to the final sections which present the physics of the mechanisms behind scintillating screens and C C D cameras and their appropriateness for use in this dosimetry system. Chapter 3 discusses the design consideration issues relevant to using a scintillating screen/CCD camera combination in the T R I U M F Proton Therapy Facility for beam monitoring and rapid visualization of dose distributions. The details of preliminary feasibility experiments are outlined along with the relevant conclusions about the final design requirements. A detailed description of each component of the final design is offered: the CCD/screen apparatus, the collection of tested scintillators, image acquisition system, and the image processing system. Finally an account of how the experiments were conducted is given. Chapter 4 presents the results of the sensitivity and linearity measurements of the scintillator output for both protons and x-rays along with several examples of use of this technique for rapid visualization of dose distributions. A discussion, along with some examples, is given on the suitability of this dosimetry system as a diagnostic tool in quality control. The results of several other interesting avenues of investigation are also presented. The possibility of using this technique with x-rays for rapid positioning of the patient is discussed; and finally, a summary of some research into radiation induced damage effects of the C C D pixels is also given. The main intent i n undertaking this study was to build on the concept of using a camera to monitor the light output from a scintillating screen and extend it for application as a dosimetry system. Based on the results of this project, future work may be concerned with improving the 8 system to achieve higher sensitivity, make it capable of monitoring a modulated beam, and allow for online image processing to display dose distributions directly on the PC monitor. 9 2. Background Information 2.1. Interactions with Matter There are a number of different mechanisms by which particles interact with the matter through which they pass. Each of the interactions attenuate the primary beam and transfer varying amounts of their energy to the matter giving rise to energy loss and scattering effects. Photons and charged particles are mediated by different processes, and i n turn suffer distinctly different mechanisms of energy loss and scattering, which gives rise to different depth dose profiles as indicated i n Figure 1-1. The unique characteristics o f each depth dose profile can be exploited i n clinical situations. The three main differences are seen in Bragg peak and distal dose fall-off, i n multiple scattering and lateral dose fall-off, and i n dose localization. 2 .1.1 . X-ray interactions The photoelectric effect and pair production are examples of photon/matter interactions; however, the dominant effect that gives rise to dose contributions from scattering is the Compton effect. Compton scattering, shown i n Figure 2-2, is an inelastic interaction between an incident photon and a loosely bound electron at rest. The incident photon donates some of its energy to the electron, resulting in a scattered photon of reduced energy, and an electron set in motion with the energy it received. Thus the incident photon of energy hv is scattered at an angle 6 with energy hv' and a recoil electron is ejected at an angle cp with energy E as shown. Applying conservation of energy and momentum to this situation leads to a relationship between angle and energy for the scattered photon and recoil electron, the derivation of which can be found i n most particle physics texts. 10 Figure 2-1 : Compton Scattering The Compton electron may be scattered at any angle from 0° to 180°. The differential cross section giving the probability of a photon being scattered at an angle 6 into the solid angle dCl per electron is given by the Klein-Nishina cross section formula i n Equation (2 - l ) . [ 1 9 ] 2moC 1 l + c t ( l - c o s 9 ) (. 2 a a 2 ( l - c o s 9 ) 2 "| 1 + cos 0 + — 1 l + a ( l - c o s 0 ) (2-1) hv where a = j and m0 is the rest mass of the electron rtioc Therefore, the total dose contribution to the absorbing medium is a result of both incident beam interactions and scattered photon interactions. Consider the measurement of dose at a point P at depth d as shown i n Figure 2-2. Two different types of dose determine the measurement: 11 primary dose contribution from incident radiation and secondary dose contribution from Compton scattering. Figure 2-2 : Primary and Secondary Dose Contributions at Point P The total dose D(d) at point P can be described by Equation (2-2) D(d) = D . +D „ (2-2) primary scatter A s the depth of point P increases, the contribution from scattering (i.e. forward, lateral and backward) increases to a maximum at depth dmax. The value of dmax increases with energy. Figure 2-3 shows a graph of absorbed dose versus depth for an x-ray beam normally incident on a homogenous material. The region between the surface and dmax is called the build-up region. Eventually the contributions from scattering begin to drop exponentially as the depth is increased beyond the point where the beam penetrates. 12 dmax Dose 1 1 • Depth Figure 2-3 : Depth dose profile of x-rays into medium A s the aim of radiation oncology is to maximize dose to the tumour while preventing damage from being incurred by the surrounding healthy tissue, medical physicists take advantage of the characteristics of the x-ray dose distribution. Delivering a uniform dose over the target volume is largely a question of geometry. For superficial tumours, treatments are planned so that the tumour volume is located at dmax while the build-up region allows for a "skin sparing" effect. For tumours at depths beyond dmax it is necessary to use multiple beams that deliver dose from different directions. Dose from each beam adds up in the tumour volume such that the ratio of tumour dose to normal tissue dose is increased. Mult iple beams can provide good distribution; however, there are some clinical and technical limitations to these methods such as practical feasibility, setup accuracy, and reproducibility. Although the photon dose profile has been manipulated effectively with combination of beams to achieve an acceptable distribution of dose i n the irradiated field, the proton dose distribution is condusive to achieving far superior conformity and uniformity i n dose delivery with single beam treatments. 13 2.1.2. Proton interactions Whereas photons interact with matter v ia photoelectric, Compton, or pair production processes, charged particles interact directly by ionization and excitation. Proton/matter interactions are mediated by the Coulomb force which results i n energy transfer due to Coulomb Scattering. Charged particles, such as protons, lose energy mainly through collisions with bound electrons via coulomb interactions. A charged particle passing by a nucleus of charge Z w i l l experience a Coulomb force between its own electric field and that of the orbital electrons and the nuclei. The deflection or Coulomb Scattering decelerates the particle and energy is transferred from the heavy particle to the bound electrons resulting in ionizations and excitations. These ionizations translate as energy deposited or dose absorbed i n the medium. Often referred to as the stopping power, Equation (2-3), derived i n Appendix A , describes the total amount of energy lost per unit length (dE/dx) along the particle track. Where p is the density and Ne is the electron density of the material, r0 is the classical electron radius as derived in Appendix B , and m0 is the rest mass of the particle. This process was also considered quantum mechanically by Bethe and Bloch to take into account any relativistic effects. The result is called the Beth-Bloch equation 1 2 0 1 , and can be approximated by Equation (2-4). / denotes the average ionization energy of the atom. dE _ 4npNeZ2r0m0c4 rdb ~dx~~ 7 ' T (2-3) dE _ 4nr2pNeZ2m0c4 In 2m0v2 v2 (2-4) dx v2 / ( 1 - v / c 2 ) c2 14 Note that the stopping power is inversely proportional to the square of the particle's velocity. A s the particle slows down, its rate of energy loss increases thereby also increasing the total number of ionizations which translates into a higher absorbed dose to the medium. In other words, as the particle gives up its energy and slows, its ability to ionize increases and reaches a maximum just before the particle completely stops and loses the ability to ionize. A s described i n Section 1.2.2, the region of increased ionization seen near the end of the particle beam path is known as the Bragg peak. It is apparent from Figure 2-4, which shows a measured dose profile for 70 M e V protons, that beyond the Bragg peak the particle has completely lost ability to ionize and deposits no dose. This unique characteristic of heavy charged particle beam dose distributions makes them especially useful for treating tumours close to critical targets. The Bragg peak can be spread out to cover the target volume and the range can be defined by varying the particle energy so that no dose is deposited beyond the tumour depth. 5 4 H T 3 CD Nl Depth in Water (mm) 4 C Figure 2-4 : Proton Dose Profile Showing a Bragg Peak and a Spread Out Bragg Peak 15 The beam of protons experience interactions with not only the bound electrons, but also with the nuclei o f the traversed material. These multiple scattering or Coulomb interactions can be described as small angle deflections, which lead to a divergence of the beam from a lateral spreading of the particles. The angular distribution of the scattered particles can be described by where p is the particle momentum, JC is thickness of the material, and X0 is the radiation length of the material being traversed. Mult iple scattering causes a lateral penumbra which increases with depth. Multiple scattering can be useful however for spreading the proton beam laterally over the treatment volume using an upstream scatterer. Although charged particle interactions are continuous, there are still some statistical fluctuations among the interactions of the individual particles of the beam. These fluctuations result in a range straggling of the particles giving rise to a distal penumbra. The rms range straggling for protons i n the energy range of 100 to 200 M e V is about 1% of the range. The sharpness of the distal edge, also known as distal fall-off, and the sharpness of the lateral edges are the most clinically important characteristics of the dose profile of proton particle beams. The ability to achieve local control without complications is dependent upon how a projected w i d t h [ 2 1 ] , which is Gaussian with rms as given approximately by Equation (2-5) (2-5) 16 closely the distal edge can be placed between the distal edge of the tumour and the critical tissues beyond and upon how well the lateral distribution conforms to the tumour volume. 2.2. Dosimetry Dosimetry systems for radiotherapy have three main functions [ 2 2 ] : (1) measurement of the dose being delivered to the patient i n real time i n order to terminate treatment at the prescribed dose, (2) measurement of the lateral and longitudinal distribution of the radiation delivered in order to ensure that the patient prescription is satisfied, and (3) monitoring radiation field parameters to maintain accuracy of the beam delivery system. The detectors used must be able to measure some quantity proportional to the dose imparted to the patient but at the same time must not perturb the radiation field. The critical parameter i n dosimetry is not the dose incident on the patient but rather the dose absorbed by the patient tissue. The Gray is the unit of absorbed dose and is defined as 1 G y = 6.24 x 10 1 2 M e V / k g . [ 2 3 1 2.2 .1 . Ionization Chamber Ionization chambers are the most commonly used dose measuring devices i n radiotherapy. A transmission ionization chamber is a cavity full o f a known gas surrounded by a known material. Two inner parallel plates with an electric field applied between them function to collect the charges that result from ionization of the gas as radiation traverses the chamber. Ionization of the gas i n the chamber is proportional to the energy loss of the particles which is i n turn proportional to the absorbed dose i n the detector. The absorbed dose i n the detector can then be related to the absorbed dose i n another medium, such as human tissue, by the ratio of the two materials' stopping powers. Bragg-Gray cavity theory gives the relationship between the 17 ionization charge Q and the absorbed dose D as is indicated i n Equation (2-6) where V is the volume. Q = (Coulombs) (2-6) The ionization energy W is the amount of ionization per unit of deposited energy and varies depending upon the type of ionizing radiation. A value of 33.4 e V for protons i n air is a standard W va lue . [ 2 4 ] Ionization chambers are also used for absolute dosimetry. A n ionization chamber is sent to an accredited laboratory and is calibrated against a standard. The calibrated ionization chamber is then regarded as the facility standard and can be used to calibrate other dosimeters. 2.2.2. Secondary Emission Monitor Secondary Emission Monitors ( S E M ) can be used in circumstances where there is a possibility of saturation of the dosimetry equipment as with a pulsed beam for example. The S E M signal is often used by some control systems to stop beam i f the dose rate exceeds a preset limit. A S E M consists of a set of parallel foils, often made of aluminum, in a vacuum enclosure. They are typically separated by a few m m and are made large enough to encompass the entire beam. Alternate foils are used for signal collection and bias. A n absolute calibration of the S E M can be done against a calibrated ionization chamber. 2.2.3. Diode Photodiodes are semiconductor devices usually made from silicon. The usefulness of ionization chambers can be limited by their size, whereas diodes can be successfully used to measure dose distributions with fine spatial resolution i n instances where even small 'thimble' ionization chambers cannot. [ 2 5 ] The small size of the diodes makes them nearly ideal for exploring the depth dose distribution of small well-collimated beams. When light or radiation is absorbed in 18 the active area, electron-hole pairs are formed. Each electron charge generated contributes to the overall photocurrent. Diodes do experience radiation damage with extended exposure; however, they are excellent for relative measurements over short periods of time and give a larger signal than ion chambers. They usually operate i n an unbiased mode. 2.2.4. Integrators Measuring dose from clinical beams originating from an accelerator demands accounting for fluctuations i n the beam current. Total charge from a dose monitor is commonly measured as an integral of the fluctuating current and recorded by an electronic device called a charge integrator. One type of charge integrator issues a pulse for every fixed increment of input charge which can be fed into a scaler for readout of the integral dose. 2.3. Proton Therapy Facility at TRIUMF The T R I U M F Proton Therapy Facility is a joint endeavour of T R I U M F , the British Columbia Cancer Agency, and the University of British Columbia Department of Ophthalmology. This facility has been in routine operation since 1995 using 70 M e V protons extracted from the 500 M e V cyclotron to treat ocular melanomas. [ 2 6 ' 2 7 ] 2.3.1. Beam Delivery The T R I U M F cyclotron accelerates FT" ions to a peak energy of 520 M e V . Protons are easily extracted from the beam by passage through a thin foi l as shown i n Figure 2-5. A thin carbon wire is used to strip the beam of electrons, leaving only protons behind. Because the net charge has been reversed, the protons are deflected i n a direction opposite from their original path. The extracted proton energy and intensity are varied by moving the stripping wire into the path of the 19 circulating protons in the cyclotron with a mechanism which moves the foi l in 3 orthogonal directions: R to select energy, L to center the beam, and Z to vary the beam intensity. This feature allows a low intensity 70 M e V proton beam to be extracted for use in the Proton Therapy Facility while the beam is operating thereby allowing other users simultaneous access to the beam. H " STRIPPING Foil/Wire Figure 2-5 : FT" stripping Once having obtained a source of protons with the desired energy, the next stage in beam delivery involves defining the beam shape laterally and longitudinally. The beam is widened, the depth of penetration is controlled, and a uniform dose distribution is obtained for the target volume. Figure 2-6 shows the layout of the equipment employed i n the PTF to achieve conformity and uniformity of the desired three-dimensional dose distribution. 20 Collimator/ Range Ion Scatterer Modulator Chamber Beam Figure 2-6 : Layout of Proton Therapy Facility Equipment Spreading the beam to a clinically useful size is achieved using a passive scattering system. The beam is first defined by a collimator/scatterer which has a 12mm circular aperture with a 0.8g/cm 2 thick lead scatterer. Protons passing through the single scattering foil are dispersed laterally into a distribution of intensity that is approximately Gaussian. A second collimator stops a significant fraction of the scattered beam, with the central part of the beam passing through to the proton nozzle. Finally a third collimator, tailored especially for each patient, is used to define the proton irradiation field. 21 The Bragg peak is far too narrow for any tumour of significant size to receive a uniform dose. Therefore, the peak is moved over the target volume by varying the beam energy and depth of penetration. The beam can be modulated in this manner to produce a spread-out Bragg peak over a fixed depth using a 'rotating step absorber' called a Range Modulator. A s this wheel of varying thickness spins, the proton beam is absorbed in different amounts as it passes through various different thicknesses of absorber. The net effect is the superpositioning of several Bragg peaks to achieve a spread out peak as seen previously i n Figure 2-4. Dose delivery also requires that the depth of penetration is controlled and a uniform dose is obtained i n the longitudinal direction. A Range Shifter, which is a variable thickness absorber, is employed to shift the range of the protons to the desired depth. This device consists of a spiral plastic wedge degrader that is rotated through one revolution by a micro-stepping motor with an incremental encoder to measure the position. A small compensating wedge is placed in the beam beside the spiral wheel to provide a uniform thickness of material in the way of the beam. The range shifter has a maximum thickness of 40mm of Lucite. 2.3.2. Patient Positioning and Treatment Planning The present method of patient positioning consists of two perpendicular x-ray images that are taken with Polaroid film (intensified by L A N E X screens) to determine the location of the tantalum clips, surgically placed on the eye to indicate tumour size and location, in relation to treatment system landmarks. The treatment plan produces a set of reference x-ray transparencies showing the desired alignment o f the tantalum clips relative to the beam axis for confirming the correct patient alignment. 22 The patient is seated in a treatment chair equipped with a motor that allows for precise positioning i n six directions. Once the head has been immobilized with a facemask and bite-block, the patient is asked to fixate on a blinking red light. During treatment the patient's eye is monitored using a video camera focused on the patient's eye that tracks motion i n directions both laterally and parallel to the beam. The position of the eye is noted on the monitor screen for confirmation of the correct position during x-ray verification and treatment. The tumour size and location information is entered into the treatment planning program, E Y E P L A N [ 2 8 ] , for calculating the optimum set of treatment parameters : location of the fixation light for the best orientation of the eye, the maximum beam range and modulation required, and the profile of the beam aperture through the patient specific collimator. Each patient receives 4 daily fractions for a total dose of 50 proton-Gy. The patient set up time is 15-30 minutes, followed by a 75-100 second treatment time. 2.3.3. Dose Monitoring and Measurement Several different types of beam monitoring devices are used i n the T R I U M F dose delivery system. The patient dose is measured by a transmission ionization chamber located immediately upstream of the proton nozzle. The transmission chamber consists of a series of nickel-copper coated Kapton foils 0.001" thick and spaced 0.25" apart inside a gas enclosure. Two sets of signal and high voltage foils provide independent readings of the dose and are referred to as the Primary and Backup ion chambers. The signal foils are each connected to Ortec 439 current integrators, which give a pulse output for each 10"1 0 coulombs. The chamber also has a set of quadrant plates for beam centering. The incident proton beam current and rate is measured by a 23 S E M located upstream of the first collimator and the beam profile and position are measured with an integrating wire chamber. A diagnostic ionization chamber, which measures the total beam emerging from the first collimator, is used to normalize the beam intensity for range shifter scans and some other measurements. This chamber is located at the treatment position and is used to calibrate the Primary ion chamber. The proton dose distributions are measured by scanning diodes or miniature 'thimble' ionization chambers i n a water phantom. A n avalanche photodiode, type B P W - 3 4 shows good agreement with a 0.05 cc Markus ionization chamber. These measurements are also compared to data collected with radiochromic film. A calibrated Exradin T l ionization chamber connected to a Keithley 616 electrometer is used for absolute dose calibration. 2.4. Scintillating Screens 2.4.1. History of Scintillators in Medical Imaging In the early 1900's, much attention was focussed on the medical application of x-rays. Not only were x-rays capable of producing diagnostic images and useful for treating a variety of ailments, but it was also discovered that x-rays could cause certain materials to fluoresce when irradiated. This discovery launched investigations to examine different substances and their ability to emit light when exposed to ionizing radiation. Research efforts revealed that certain organic compounds, and i n particular phosphors, fluoresce the most intensely. The sensitivity of film to direct x-ray exposure is i n fact quite low; thus a high dose is required i n order to sufficiently expose the film. Scintillating screens, also called intensifying screens, are 24 used to reduce the amount of dose delivered to the patient. A scintillating screen functions to absorb the energy i n the x-ray beam that has penetrated the patient, and convert this energy into a light pattern which has nearly the same information as the original x-ray beam. The light then forms a latent image on the x-ray f i lm. Phosphors powders were first incorporated into screen materials to act as intensifiers as early as 1896. [ 2 9 ] Crystalline calcium tungstate (CaWCu) was the original phosphor used. Later it was discovered that rare earth compounds, especially terbidium activated gadolinium and lanthanum, were even better scintillators. A t present, zinc sulfide activated with silver or copper is one of the most efficient scintillators k n o w n , [ 3 0 ] although there is not much published work on the response of commercial scintillating screens to protons. 2.4.2. Screen construction A scintillating screen has four layers: a base, a reflective layer, a phosphor layer, and a plastic protective coat. The base, or screen support, is usually made of high-grade cardboard or polyester plastic and is ~7mils thick. Because the light produced by the interaction between x-ray photons and phosphor crystals is emitted in all directions, some of the light photons are directed back towards the base layer of the screen and are lost. The reflecting layer, - l m i l , acts to reflect light back towards the front of the screen. The phosphor layer, ~ 1 or 2mils thick, is applied over the reflecting coat. The phosphor crystals are suspended in a plastic polymer. Finally, the protective layer is applied over the phosphor as serves to prevent static electricity, give physical protection to the phosphor layer, and to provide a surface which can be cleaned easily without damaging the phosphor layer. It forms a layer that is 0.7 to 0.8mil thick. 25 2.4.3. Mechanism behind energy conversion Fluorescence is defined as a form of luminescence (emission of light by a substance) produced Q when light is emitted instantaneously, i.e. within 10" seconds of stimulation. Organic compounds, and in particular phosphors, scintillate when exposed to radiation as a result of a photoelectric interaction between the x-rays and the phosphor. [ 3 1 ] The resulting photo-electrons deposit kinetic energy i n the phosphor crystal v ia ionizations and excitations along their path. A small fraction of the excited states decay via fluorescent de-excitations giving rise to luminescent emission. Conversion efficiency or intrinsic efficiency is a measure of the phosphor's ability to convert absorbed x-ray photons into light photons. Calcium tungstate has a conversion efficiency of 5% whereas the newer phosphor materials may have an efficiency as high as 20%. The ability of light emitted by the phosphor to escape from the screen and expose film is called screen efficiency. For typical screens, about half the generated light reaches the film, the rest is absorbed i n the screen and is dissipated. Screen efficiency can also be defined as the product of absorption, conversion, and emission efficiencies. 2.4.4. Light Quenching Some scintillators experience quenching of the light output i f the ionization density, i.e. dE/dx, is very high as it is at the Bragg peak. B i r k s [ 3 2 ] presented a semi-empirical relation to describe this behaviour which he attributed to quenching of the primary excitations by the high density of ionized and excited molecules. It is assumed that a fraction k of these molecules w i l l dissipate energy non-radiatively, which leads to quenching of the light yield. Equation (2-7) describes the light output energy dL per unit path length dx according to Birks ' model where e is the 26 conversion efficiency of the phosphor and B is a constant of proportionality. B and k cannot be determined separately and are usually treated together as the quenching coefficient. Thus, quenching occurs mainly at the most distal part of the depth dose curve, due to the increasing contribution of the large dE/dx values with depth. 2.5. Cameras A s imaging techniques i n radiography have been modified and improved, the cameras which are such an integral part of the system have evolved along with them. Some of the first systems utilized television techniques, which consist of light incident on a television camera receptor. The light is converted into an electrical signal which is then transmitted via a cable to drive the cathode ray television tube and yield a picture. V id icon cameras are an improved version of the original T V camera and are often used in x-ray detection: they have a higher x-ray absorption ability and retain their properties under prolonged exposures. The Reticon camera, which consists of an integrated array of photodiodes, was one of the first solid state cameras. Its main advantage is an intrinsic noise level far below that of the standard vidicon camera. However, it was the Charge-Coupled Device ( C C D ) camera that revolutionized the imaging industry and the C C D image sensor is now the detector of choice i n almost all high performance imaging systems. A new device called the Charge Injection Device (CLD) camera is touted as having increased high-speed readout capabilities, lower noise, and higher resistance to radiation; however, they are also much more costly. (2-7) 27 2.5.1. Charge Coupled Devices The concept of Charge-Coupled Devices ( C C D ' s ) was proposed i n 1970 by Boyle and Smi th [ 3 3 ] and has since been used extensively in imaging applications. A C C D consists of an array of Metal-Oxide-Semiconductor (MOS) transistors. Essentially a C C D operates by storing information i n the form of electrical charge packets in potential wells created in a semiconductor (substrate) by the influence of overlying electrodes separated from the semiconductor by a thin insulating layer (gate oxide). B y applying an external voltage to the electrodes, the potential wells, and thus the accumulated charge packets, can be shifted through the semiconductor. B y passing the charge packet to one point and detecting it an another, a shift register (or 'bucket brigade') has been created. In imaging applications, charges are introduced into the potential wells by optical means: light incident on the semiconductor generates charge carriers which can be collected in the potential wells and subsequently clocked out of the structure. Because the potential wells are capable of storing variable amounts of charge, the shift register is not merely a digital structure, but rather the C C D is capable of providing an analogue signal. It is the C C D analogue signal handling capability for signal processing that makes them appropriate for use in image sensor systems. 2.5.2. Mechanism Behind the CCD as an Optical Imaging Device A full explanation of the mechanism behind C C D ' s involves a discussion on semiconductor physics and is beyond the scope of this thesis, and the reader is directed towards Appendix C as only the basic physics involved is presented here. A s an optical imaging device, the C C D analogue shift register is used to collect and read out optically generated signals. Figure 2-7 shows the mechanism driving a standard C C D structure. A n optical system, such as a camera lens, is used to focus the image on the front face of the C C D . Initially, any one set of electrodes, 28 e.g. <p>i, is held at a positive clock voltage, thus creating a potential wel l under each of the electrodes. <)>i O -<h O -h o -4>i <h2 <t>3 Light spot i i i C C D electrodes m m r-i r-i • n rn rn rn rn m rn o o o Charge packet Potential wel l y integration period • X -readout period Figure 2-7 : Pattern of charge collected under electrodes is read out as an analogue signal by the clocking array. 29 The other electrodes are held at zero volts or a small resting potential. Photons entering the silicon substrate, generate electron-hole pairs v ia the photoelectric effect. The minority carriers generated are collected under each CO electrode. The number of electrons generated under a given electrode per unit time (usually the integration time) is proportional to the local light intensity. Therefore, the pattern of charge collected under the electrodes is an analogue representation of the light intensity across the original image. A t the end of the integration period the charge pattern is read out by a clocking array - taking care that the clocks have been run long enough to clear out every charge packet. Then the device is switched from readout mode to integration mode and the cycle is repeated. 2.5.3. CCD Properties A n ideal C C D sensor would have a high sensitivity and a wide dynamic range making it capable of operating i n environments of both low level illumination and large variations i n the level of illumination. Dynamic range and sensitivity are determined by the maximum charge handling capability of the C C D and by the various different sources of noise present. Sensitivity is a function of noise and responsivity, which is a term given to describe the generation of a signal charge from light energy. It is convenient to express the measured responsivity of a detector i n terms of the quantum efficiency and the charge transfer inefficiency. Quantum efficiency is described as the mean free number of electrons created by one photon incident on the detector. The efficiency of a C C D detector is almost always less than unity. The charge transfer inefficiency describes the average inefficiency i n transferring charge from one pixel to the next during the read-out process. Transfer can be inhibited by crystal defects 30 referred to as 'traps'. A n increase i n charge transfer inefficiency can be caused by radiation damage since the creation of traps is a long-term effect of radiation damage. Charge transfer inefficiency can also degrade spectral resolution, which is defined as the ability to discriminate between closely spaced points i n the image. Good spectral resolution depends upon accurate determination of the total charge liberated from a single photon. This i n turn depends upon the fraction of charge collected i n a single pixel and upon the fraction of charge lost i n the transfer from pixel to pixel down the 'bucket brigade' or shift register during read-out. Sensitivity is also a function of noise; and unfortunately, there are many factors that contribute to C C D noise. These sources of noise can be divided into three categories: noise arising from the injection of charge into the device, the noise attributed to fluctuation in charge transfers, and the noise introduced by the charge sensing circuitry. Dark current however, is the largest source of noise. If electrons and holes gain sufficient energy to make transitions across the band gap from the absorption of thermal energy, there is an undesirable source of charge carriers that are indistinguishable from photo-generated carriers. This current is often referred to as dark current as it is not generated by absorption of radiant (light) energy. C C D cameras can be cooled to low temperatures to reduce the amount of dark current generated. 2.5.4. Radiation Damage Particle radiation can damage C C D s and degrade their performance. Because C C D s are solid state M O S structures, they are vulnerable to ionization damage i n their oxide layer. The main effects are buildup of trapped charge i n the oxide and the generation of traps at the oxide-31 semiconductor interface^3 4 1. This damage degrades the C C D performance by increasing the charge transfer inefficiency, by the generation of 'hot' pixels, and also by increasing the dark current. Transient effects are due to ionization induced generation of charge within the active region of the detector. These effects are not permanent, and spurious charges are swept out during read-out. However, this results i n a significant source of noise i n the video output. Displacement damage i n a semi-conductor occurs when protons pass through the semiconductor material. Nearly all the energy loss goes into ionization, except a fraction which goes into the displacement of atoms from their lattice site and results i n the creation of vacancy-interstitial pairs. More than 90% of these pairs recombine but a remainder of them migrate through the lattice until they form stable complexes. The trapping of signal charge can be seen in persistent "hot pixels" or pixels that appear to be saturated, even during a dark frame read-out. A l s o seen are ' f l ickering' pixels, which show unstable states of dark current: sometimes the mean signal per pixel is comparable with the background level and sometimes it is much greater, appearing as a 'hot' pixel . 2.6. Frame Grabber and Image Processing System In an image processing system, video cameras are linked to computers via a video/computer interface called a frame grabber. The main source of input to the computer is a solid-state camera that picks up patterns of light and dark. This pattern, or image, is scanned as an array of light intensities that are converted into analogue electrical signals by the camera. Images can be 32 scanned at "real-time" speeds, typically 30 frames per second, and fed into the frame grabber. The image-processing computer can then analyze the images and extract the required information. The main role of a frame grabber is to convert the analogue video signal from the camera into a series of binary words. The video signal w i l l contain information about the brightness of each element i n the display as wel l as information about the vertical and horizontal synchronization pulses. To digitize the analogue signal, each video frame is divided into a fixed number of rows, with each containing a fixed number of picture elements called pixels. The frame grabber reads the contents of each pixel defined in the video frame grid, determines the average brightness, and changes the analogue voltage into a binary value. The spatial resolution of the image depends on the number of pixels defined i n the grid. 33 3. Experimental Design Considerations A s mentioned i n the introduction, one of the main goals of this research project was to develop a rapid method of measuring dose profiles using a scintillation screen observed by a C C D camera. A dosimetry system of this nature should be able to do near real-time observation and analysis of the proton beam thereby offering a practical and timesaving method of quality control. The requirements on such a system include: a high sensitivity in order to detect signals from a wide range of beam intensities and delivered doses; a signal response that is linearly proportional to the absorbed dose regardless of the type or energy of the incident radiation; an integration capability over a designated amount of time; immediate acquisition and display of lateral profiles of the dose distribution; and portability to give it versatility for possible use in other applications. 3.1. Feasibility Study A preliminary experiment was carried out i n T R J U M F ' s Proton Therapy Facility using an incident beam energy of 7 0 M e V and intensities up to lOnA, identical to those used for clinical doses. A scintillating screen enclosed in a light tight box is placed in the path of the proton beam, and is observed by a C C D camera. In order to reduce the amount of radiation damage to the C C D chip, a 45° mirror is used so the camera may be placed outside the primary beam, see Figure 3-1. This preliminary investigation into the plausibility of a C C D camera/scintillating screen combination involved testing the system for sensitivity and noise. A n inexpensive camera, a P U L N i X X C - 3 7 , was set-up to observe the light output from the proton beam incident on a 34 L A N E X (Regular) Screen. This screen was selected as it is used as an x-ray intensifying screen for patient alignment. Figure 3-1 : Geometry of Experimental Apparatus The image was observed directly on a T V monitor with no integration. The resulting signal was weak but detectable; however, the image was "contaminated" by bursts of huge noise pulses. In an effort to determine the nature of the noise pulses, the camera was shielded by 2 inches of lead against the upstream side to eliminate any protons that might be generated via scattering in the collimator. The mirror was also replaced by thin tape to avoid proton scatter off the mirror. It was determined that the pulses were indeed related to beam as the pulses stopped when the beam was turned off. The lead shielding eliminated all of the protons scattered from the collimator but not those from the mirror. Removing the mirror reduced even more scatter. The remaining noise pulses were obviously neutrons which cannot be easily shielded. Fortunately, 35 with shielding, the rate was not regarded as a serious problem; i n particular i f the sensitivity of the system could be increased. Based on these encouraging results a new arrangement was made which had a more sensitive camera, a thinner front silvered (actually aluminized) mirror and an integrating frame grabber system. The system was also made compatible with an optical diode, which replaced the C C D for some sensitivity measurements. 3.2. Equipment Description A dosimetry system of the nature described at the beginning of this chapter consists of three main components: a scintillator screen, a sensitive C C D camera, and a data acquisition and display system. The data collection process starts with signal generation within the pixels of the C C D camera. The camera pixels transfer their collected charges that correspond to the patterns of light from the scintillator screen to create an analogue signal. A n acquisition module ( A M ) acts as an interface between the camera and the frame grabber to digitize the analogue signals, also referred to as video signals. The A M interprets the video signals and generates a video frame, which is recognizable to the frame grabber. The image information is captured by the frame grabber and transferred to memory storage i n the format of a video frame. The video frame can be viewed through an image display system, such as within a software run display window. A schematic diagram of the signal processing system is shown i n Figure 3-2. 36 1 120 Hz -586 INTEL PENTIUM BASED P C CAMERA CONFIGURATOR UTILITY 1 t IMAGE DISPLAY PCI VIEW IC-PCI FRAME GRABBER UTILITY AM-VS ACQUISITION MODULE J MISC CLOCK HORIZONTAL SYNCH VERTICAL SYNCH Light From — Scintillator C C D C A M E R A VIDEO INTEGRATION CONTROLLER S C R E E N GATE POWER SUPPLY Figure 3-2 : F low Chart of C C D Signal Processing System 3.2.1. Screen Although the exact mechanism which determines the sensitivity for the conversion of proton energy to light energy i n a phosphor screen is not known, it has been demonstrated that L A N E X [ 3 5 J as wel l as several other organic phosphors are appropriate scintillators for proton beams. [ 3 6 ' 3 7 ' 3 8 ' 3 9 1 Hence L A N E X was chosen as a reference screen for the scintillator analysis portion of this research project. A l l L A N E X screens have phosphors from the lanthanide series and the primary x-ray absorber is gadolinium. They have a (~0.3 mil) clear overcoat to resist surface abrasion, sealed edges to minimize wear, and a backing layer to eliminate curl. Several 37 other scintillators that are commonly used in particle physics detectors were evaluated, such as a plastic scintillator from B i c r o n f 4 0 ] and a zinc-sulfide powder adhered to an aluminum plate used in T R I U M F beam monitors. A s wel l , a series of Rarex [ 4 1 ] medical intensifying screens were also assessed. A s with the L A N E X , the Rarex screens consists of a phosphor compound mounted to a plastic backing layer. Table 3-1 lists some properties o f the screens tested in this research project. Emission efficiency is defined as the efficiency with which the phosphor converts x-ray particles into light photons. Table 3-1: Properties of Tested Commercial Intensifying Screens for X-ray Absorption Screen Name Composition Emission Efficiency Thickness (+/- 0.1 mm) Colour Peak output wavelength (nm) Calcium Tungstate (historical) CaW04 5% blue 425 LANEX Gd202S:Tb 18% 0.38 green 545 Bicron412 unavailable - 1.71 blue 434 Zinc Sulfide ZnS - 1.57 blue 450 Rarex Green Fine Gd202S:Tb 16% 0.34 green 545 Rarex Blue ILT LaOBnTm 11% 0.37 blue 360, 460 Rarex Green Fast Gd202S:Tb 16% 0.53 green 545 Rarex Green Regular Gd202S:Tb 16% 0.39 green 545 Rarex PFG ZnCdS:Ag 19% 0.63 green 530 3.2.2. Camera A P U L N i X T M - 7 4 5 E high-resolution C C D camera was selected due to its low signal to noise ratio of 50dB and its excellent low light sensitivity. Operating at room temperature, it is capable of detecting a signal as low as 0.5 lux at F1.4 due to its full frame integration feature that 38 provides sensitivity for dark environment applications. More sensitive cameras which operate at temperatures of -25° C and below were considered but were determined to be too expensive for this work. The camera has a high resistance to magnetic fields which is an important consideration since all equipment must be able to withstand and function properly in the magnetic field created by the cyclotron magnets. Furthermore, the P U L N i X camera has other desirable features, such as precise image geometry from exact placement of pixel rows and columns that makes it suitable for imaging applications. The C C D pixels are arranged in a 768(H) x 493(V) array and are 11 x 13 u m 2 i n size. RS-170 video signals from the P U L N i X camera used i n this study produce a frame that contains two fields of horizontal image lines: an odd field comprising of all odd numbered lines and an even field comprising of all even numbered lines. Horizontal-synchronization pulses synchronize the start of each line while vertical-synchronization pulses coordinate the start of each field. The line signals and the synchronization pulses are passed to the Acquisition Module ( A M ) which converts them into a valid video frame. 3.2.3. Acquisition Module The Variable Scan Acquisit ion Module ( A M - V S ) manufactured by Imaging Technology Incorporated [ 4 2 ] digitizes analogue video images and generates horizontal blank, vertical blank, and field timing for a host mother board. The A M - V S has two basic modes of operation: standard and variable scan. In standard mode, usually when operating with standard RS-170 cameras, it generates frame timing from the timing generator. In variable scan (VS) mode an external source, typically the camera, supplies the frame timing. The camera supplies a 39 horizontal-line-enable ( L E N ) , a vertical-frame-enable (FEN), and pixel-clock at the camera connector. These signals drive the window generator to create a valid video frame by determining the size of the acquired image; i.e. the number of pixels per line and the number of lines i n the frame. 3.2.4. Frame Grabber Imaging Technology (ITI) manufactures the IC-PCI which is an image capture board that supports the Acquisit ion Module ( A M ) family of plug-in camera interfaces for fast image capture (up to 40 M H z ) [ 4 3 ] . It is a low cost frame grabber with a high-speed P C I bus interface. It is capable of bus mastering data from image memory directly to a destination within the system (i.e. display memory). The IC-PCI receives camera data from the A M after having been passed through a multiplexer that aligns the three A M channels based on whether the pixel format is 8, 16, or 24-bit. The pixel format must be programmed within the software program prior to image acquisition. Two types of image acquisition are available: Normal and External Trigger. The beginning and end of a normal acquisition is based on the framing signal frame enable (FEN). F E N is the digitized version of the vertical blank signal. Three modes of acquisition exist within normal mode: snap, grab, and freeze. Snap performs a single frame acquire at the beginning of the next F E N rising edge. Grab performs continuous acquisition beginning on the next F E N rising edge. Freeze w i l l stop a l ive acquire at the next falling edge of a F E N . Figure 3-3 shows image acquisition (Snap) i n normal mode. When a snap is requested, acquisition of data is delayed until the next F E N rising edge as indicated by the dotted vertical line. The next F E N falling edge is the trigger to stop collecting data at the falling edge of the field definition. 40 F E N Acquire Bits Field Data Grab status ACQMD=0 N < / ACQMD=2 >^<^  ACQMD=0 Snap requested Valid data acquired Status shows snap in progress Figure 3-3 : Normal Mode Frame Grabber Acquisition External trigger allows the image acquisitions to be synchronized to external events. When acquiring an image in external mode, the acquisition w i l l not start until the A M provides a trigger signal to the frame grabber. 3.2.5. Integration Controller The integration controller device is intended to improve the contrast o f an image by causing the camera to integrate the picture over a specified number of frames, then to trigger a frame grabber i n a P C to capture the resulting picture. The P U L N i X T M - 7 4 5 E camera, when receiving a low signal on its integration control input, integrates the image over a whole number of video frames. Fol lowing the next vertical synchronization interval after the integration control switches to high, the camera outputs one frame of video representing the image accumulated during the integration time. During the 41 integration time, the camera produces a black signal. The integration controller samples the camera video signal to extract the vertical synchronization signal to ensure that the integration control pulse is correctly timed. A t the end of the integration time, it generates a trigger signal that is used to make the IC-PCI frame grabber capture the intensified video frame. The unit can be operated i n a "free run" mode, or can be triggered by an external device, as selected by a front panel switch. In the "free-run" mode, the unit repetitively does integrate/frame grab cycles. In the external trigger mode, the camera operates normally until the unit is triggered. When triggered, the camera goes into integrate mode for the specified time, then produces one frame of intensified video before returning to normal continuous operation. The integration time is set using a lever-wheel switch. The integration time is 2 n frames, where n is the switch setting. Table 3-2 : Integration Times for Switch Settings of the Integration Controller S W I T C H S E T T I N G I N T E G R A T I O N T I M E 0 None - normal operation 1 2 frames 2 4 frames 3 8 frames 4 16 frames 5 32 frames 6 64 frames 7 128 frames 42 3.2.6. Software The ITI software is installed in,a 120Hz 586 Intel based Pentium Computer and operates within a Microsoft Windows 95 environment. The ITI Camera Configurator is a software utility that is required to make the video camera compatible with the frame grabber. The software is a library of camera set-up and selection functions, distributed as object code that can be linked to application programs written i n C . The configurator auto-senses which kind of acquisition module and frame grabber board are installed and then allows the camera type to be identified from a pull-down menu. The configurator establishes all o f the manufacturer's recommended parameters for that camera model. The camera settings are then saved i n a binary configuration file for input into an image processing application. Cl icking on the "Grab" button of the Camera Configurator window does an image display. The Pciview! utility starts automatically, and a live image is displayed i n a window on the P C monitor. In grab mode, frames are continuously acquired at the camera's rate from the acquisition module. A grab w i l l continue until "Snap" or "Freeze" is clicked. "Clear" w i l l clear the current active frame buffer and the video display w i l l turn black. A n image i n the active frame buffer can be saved to a file i n TIFF format by clicking on "Save" and a saved image can be redisplayed by clicking " L o a d " which reads a TIFF formatted image from the system disk into the active frame buffer. 3.2.7. Photo-diode Some sensitivity measurements, such as linearity of the screens, are easier done with the use of a photo-diode. A silicon B P W - 3 4 photo-diode was chosen for its favourable signal/noise ratio even at low illuminance. A l s o , the spectral response of the B P W - 3 4 diode was suitable for the wavelength outputs from each type of scintillator. It has a maximum wavelength sensitivity of 850 nm, a quantum efficiency of 0.88, and a sensitive area of 7.6 m m . The diode was operated 43 as unbiased i n order to avoid the large dark current background often associated with reverse biased diodes thereby increasing its sensitivity to a weak signal. Although not as linear as a reverse biased diode, an unbiased diode's signal response is still linear for short exposures. The diode is mounted on the camera lens at the same position as the C C D . 3.3. Experimental Set-up A light tight box was fashioned so that the 16mm F1.4 lens of the camera can fit securely into one end of the box as shown in Figure 3-4. The box also accommodates an optical diode/lens arrangement. The screen is inserted perpendicular to the beam direction with the active side facing the mirror. The light emitted from the scintillator is reflected off the 45° front-silvered mirror towards the optical components. The tube was lined with light absorbing material to reduce the amount of light from reflections off the tube walls. A copper cylinder sits on the tube, to cover the camera and protect it from excess radiation and possible pixel damage. 44 SCINTILLATOR SCREEN - CCD CAMERA Scintillator Front Silvered Screen Mirror (active side facing mirror) Figure 3-4 : Scintillator Screen and C C D Camera Dose Measuring Arrangement (scale 1:3.1) For both the screen evaluation and the beam monitoring experiments, the box was placed in the path of the beam i n the treatment room. Figure 3-5 indicates the exact position of the box, located directly behind the final collimator, relative to the rest of the P T F equipment. 45 Collimator/ Scatterer Profile \ Monitor \ Range Modulator Ion Chamber Scintillator Screen CCD Camera Equipment at Treatment Position Motorized Chair 6 Motions Figure 3-5 : Positioning of C C D Camera/Screen Measuring Device i n P T F Equipment 3.3.1. CCD Readings The camera video signals are connected via coaxial cables to the integration controller located on a portable. experimental equipment/computer trolley in the P T F control room. From the integration controller, the video signal is connected to the A M - V S video input and the camera's Gate In connects to the integration controller's Gate Out. The integration controller lever-wheel switch is set to a certain number of frames of integration to achieve the desired sensitivity or to average over a certain period of time. Care must be taken to ensure that the period of integration does not saturate the C C D pixels. The camera intensity is defined as a gray-scale value between 0 and 255 and thus the integrated signal must not exceed an intensity value of 255. 46 3.3.2. Diode Readings For both the sensitivity and the linearity tests of each scintillation screen, the diode arrangement was used to detect the amount of light emitted from each screen. An Ortec 439 current integrator with a sensitivity of 10"10 Coulombs per pulse was used to read the diode signal. The diagnostic ionization chamber was also used to record the dose delivered and served as a calibration tool. 47 4. Discussion of Results 4.1. Evaluation of Scintillating Screens A n investigation into the light response of several commonly used scintillating materials and various commercially available medical-intensifying screens was done. A reliable dosimetry system requires that its components give a linear dose response. Hence, it is o f paramount importance that the screen light output is invariably proportional to the dose delivered. Thus the linearity of the dose response of each screen was investigated for both x-ray and proton irradiation. The screens were also tested for the relative sensitivity of their dose response. 4.1.1. Proton Irradiation Measuring the output of each screen as detected by a photo-diode tested the relative sensitivity of each screen to proton irradiation. The photo-diode was mounted on the camera lens as shown i n the Figure 3-3 inset. Focussing of the lens was adjusted to an aperture of F 1.4 to give maximum photodiode output when viewing a green L E D mounted at the screen position. The lens settings were kept constant for all measurements. The photo-diode signals were read-out by an Ortec 439 current integrator into a C A M A C scaler. Proton current as measured by the diagnostic ion chamber was also read-out with an Ortec 439 integrator into a scaler. Proton intensities of 1 to l O n A were used to check linearity with intensity. These intensities are converted to dose i n water i n cGy at the screen position using the standard proton therapy calibration technique. A proton current of 5-7nA is the normal intensity used for therapy treatments. 48 For these measurements the diode was biased with a voltage of 9 V . This produced a signal offset current, which was recorded for each screen set-up and subtracted from the diode output during beam measurements. Typically 10-second integration times were used with a number of measurements taken and averaged. The reproducibility of the measurements was within 1% for diode outputs greater than 30. Figures 4-1 and 4-2 shows the diode output as a function of proton dose at the screen position for the selection of screen materials. L A N E X was used for normalization. Diode Reading versus Delivered Dose 120 1 ' ; — ' Dose (cGy) Figure 4-1 : Response to Proton Irradiation for Less Sensitive Screens 49 1800 1600 1400 o 1 2 0 0 o 1000 3 Q. I 800 •D O ' ° 600 400 200 p PGF Fast Regular * Blue LANEX 100 200 300 400 Dose (cGy) 500 600 Figure 4-2 : Response to Proton Irradiation for More Sensitive Screens 4.1.2. X-ray Irradiation The collection of screens was subsequently tested with x-ray irradiation since the commercially available intensifying screens are truly intended for use with diagnostic x-rays. Hence, it was expected that they would give an adequate light output for x-ray energies well matched to diagnostic exposures. The light emitted from the screen was recorded by a photo-diode that was read by a Keithley 616 current integrator. In this case, the diode was operated as unbiased i n order to increase its sensitivity to a weak signal since this eliminates the dark current background previously seen with the reverse biased diode. Because an unbiased diode's signal response is not as linear as a 50 reverse biased diode, the photo-diode was checked for linearity. This was done by varying the current through a green LED and recording the corresponding LED brightness. The photo-diode response was found to be sufficiently linear. The screen/diode arrangement was positioned 1cm forward of the diagnostic x-ray tube used for patient positioning. The x-ray tube is situated amongst the PTF equipment as seen in Figure 3-5. The diagnostic x-ray tube was set to operate at 50mA, 70kV. Thus irradiating the screens with increasing seconds of exposure tested the response of each screen for increases in x-ray intensity. Since these commercial screens are intended for use as image intensifying screens in x-ray imaging, it was not surprising to see the linear response as is evident in Figure 4-3. Signal Response of Screen as a function of X-ray Exposure 2.5 ^ • PFG 2.0 -f * Green Reg Green Fast Blue III LANEX 0.0 0.5 1.0 1.5 2.0 2.5 Seconds of X-ray Exposure at 50 mA, 70 kV Figure 4-3 : Linearity of Screens for X-ray Irradiation 51 4.1.3. Linearity of CCD signal response Although satisfied that the screens themselves showed linear behaviour, it was nevertheless prudent to investigate the linearity of the C C D camera response. Consequently, the next step was to measure the response of the C C D camera/scintillating screen system for proton irradiation. The screen/CCD camera apparatus was set-up as described previously in Chapter 3. The two most sensitive screens plus the L A N E X screen were irradiated and the C C D camera recorded their emitted light. A background image was taken prior to every reading in order to determine the signal contribution from thermal generation, ambient light, and other noise sources. Care was taken to use integration times and beam intensities that kept the C C D count from saturating. Here the read-out of C C D intensity is not immediate and is instead extracted from the recorded images by the following technique. The TIFF formatted image file is transferred to a U N I X machine and viewed through X v i e w image display program. The image file is then resaved in P G M format, which is a format that is convenient for use in analysis. The data file contains the coordinates of each C C D pixel and its corresponding intensity reading. A s mentioned in Chapter 3, the intensity reading is given as a gray-scale value between 0 - 2 5 5 . However, before the C C D intensity values could be used, it was necessary to correct for the background signal. 4.1.4. Background Corrections Although the system was shielded effectively from the signal contributions due to ambient light, there was still a background signal from the C C D dark current. Each image had to be corrected to separate the thermal generated signal from the real photo-generated signal. The background contribution was determined by recording an image with the same integration time as the actual 52 image, except without radiation. During integration, the camera signal process keeps optical black levels as the reference video black signal. This clamps the video levels resulting in the cancelling out of the thermal noise (dark current) during integration period. This means the average background signal is roughly the same for images with identical camera operating conditions regardless of the period of integration. The average size of the background was consistently about 8.18 +/- 0.01% of the C C D dynamic range. Correction for background signal was done easily by subtraction. Fortunately, very few large noise pulses or 'spikes' were seen on the C C D recorded image as a result of the shielding efforts made i n the preliminary feasibility study. However, as the C C D experienced more and more exposure, pixel damage began to occur. The damaged or 'hot' pixels resulted in large signals i n isolated pixels that appeared as spikes i n the C C D image; however, these damaged pixels did not degrade the image enough to cause concern. Radiation damage effects on the C C D camera w i l l be discussed further on i n this chapter. Once corrected for background, the C C D intensity values were plotted as a function of dose. Figure 4-4 shows the system has high sensitivity, low noise, and linear behaviour for beam energies and intensities wel l matched to clinical doses. It is worth noting that the relative sensitivities of each screen as measured by the C C D are consistent with the sensitivities as measured by the photo-diode. The reproducibility of the measurements is not as good as the 1% seen with the diode since the proton dose and C C D measurements are not perfectly synchronized. 53 Scin t i l la to r Outpu t as a Func t ion of Pro ton Dose 140 120 PGF >• 100 cn c CD £ 80 2 d> E co U Q O O 60 40 Green Fast 20 LANEX 100 200 300 P r o t o n D o s e ( c G y ) 400 500 Figure 4-4 : Linearity of C C D response for Proton Irradiation Tabulated below i n Table 4-1 are the sensitivities of each screen for proton and x-ray exposures relative to L A N E X . Table 4-1 : Sensitivities of Various Intensifying Screens Relative to LANEX Scintillator Material Sensitivity to X-rays Sensitivity to Protons LANEX 1.00 1.00 BICRON 412 N / A 0.37 ZnS N / A 0.64 Rarex Green Fine N / A 1.26 Rarex Blue III 1.74 1.89 Rarex Green Regular 3.77 4.88 Rarex Green Fast 5.49 9.50 Rarex PFG 9.16 17.43 54 Evaluation of each type of scintillator for both x-ray and proton exposure revealed that the Rarex P F G screen was by far the most sensitive and displayed a linear dose response. Hence, it became the screen of choice for the remainder of the project investigations. 4.2. Beam Monitoring Since patient survival depends heavily on the conformity and uniformity of the dose delivered, an accurate and efficient method of obtaining the beam profile is necessary. One aspect of quality control i n proton therapy includes monitoring the position and shape of the beam spot. A t present, beam centering is done using the beam profile monitor and the quadrant transmission plates of the transmission ion chamber. Although the quadrant chamber is a useful diagnostic for changes i n beam position, it does not give information about the shape or uniformity of the beam itself. The uniformity of the beam is measured either with f i lm, which is exposed and subsequently digitized, or by scanning a small ionization chamber i n a water box. Both techniques are slow i f two or three-dimensional information is required. Therefore, the CCD/screen technology was tested as a possible quality control system for rapid beam monitoring. 4.2.1. Uniformity A 6nA beam was centered by the quadrant chamber profile monitor to an accuracy of 2%, and the C C D system was set up to observe the beam profile. A 26mm x 15mm rectangular collimator determined the field size. Figure 4-5 shows the beam profile for different beam currents. Once the image had been corrected for background, the uniformity of the beam profile was seen with ease. 55 300 350 400 450 Pixel Position 500 300 350 400 450 Pixel Position 50C 350 400 450 Pixel Position 350 400 450 Pixel Position 50C Figure 4-5 : Uniformity of Lateral Dose Profile for Various Beam Intensities 4.2.2. Beam Steering Effects Since the system had demonstrated rapid beam profile acquisition capabilities, the beam was purposely steered off center to test whether the system could be a useful diagnostic in verifying beam positioning and shape. Was it possible to detect beam misalignment? The profile monitor was used to steer the beam left and right. Figure 4-6 shows the comparison between a centered 56 beam profile and a beam left (115/85) and a beam right (75/125) profile. It is worth noting that there is an observable loss i n intensity of the beam when it is steered off-center. This is easily accounted for since part o f the beam is actually lost in the beam line when it is so severely misaligned. 1 2 0 100 H >, 80 H c 60 Q o 4-0 H o 20 A 200 solid : Beam Right dashed : Beam Centered ~300 400 . 500 lateral position (pixels) 60C 120 1 0 0 A -1-; c 80 A 60 A Q o 4-0 o 2 0 A o-?00 solid : Beam Left dashed : Beam Centered — i 1 1— 300 400 500 lateral position (pixel number) 60C Figure 4-6 : Effect of Beam Steering on Lateral Dose Profile 57 4.3. Dosimetry Although on-line measurements of beam parameters such as intensity and profile are critical, they do not suffice for the quality control of many treatments. Quality control systems require instrumentation that also measures the dose distribution i n phantoms for various field definitions. 4.3.1. Bragg Peak A n experiment was launched to explore how this technique could be used to measure depth-dose distributions. A square of scintillating material was affixed to a wedge of Lucite (near water-equivalent phantom material) to simulate an increase i n depth. The wedge had a 35° angle, which corresponded to a depth of 24.5 m m , and a tissue equivalent depth of 28.3 mm. See inset i n Figure 3-3. The proton beam was incident on one side of the wedge and the C C D camera recorded the light output from the exit side of the screen. The varying intensity of the emitted light should represent the change i n dose deposited with depth. The Figure 4-7 is of two cross sectional plots of the average C C D measured intensity (background subtracted) as a function of depth for beam currents of 2 and 5 n A . The recognizable Bragg peak is clearly visible as is the expected 4 to 1 ratio in intensity for peak to plateau dose. A direct comparison was made between a dose distribution as measured by the conventional ion chamber scan in water and a dose distribution as measured by the C C D method. The C C D intensity values were normalized and graphically compared to data collected and measured by a Markus ionization chamber. 58 Figure 4-7 : Bragg Peak Plots Figure 4-8 shows that there is a good match between the Markus chamber and the CCD measured depth dose profiles. It should be noted that the position of the peaks had to be shifted 2.27 mm in order to correct for the effect of the thickness of the screen versus the thickness of the cap around the ion chamber. Worth mentioning is the apparent loss in intensity at the Bragg peak for the CCD curve as compared to the ion chamber curve. Although this could be a manifestation of many different effects, one possible explanation is saturation of the scintillating screen at high dE/dx. 59 Figure 4-8 : Comparison of Bragg Peak Measurements for C C D camera and Ion Chamber. 4.3.2. Saturation Effects- Loss of Intensity A s mentioned earlier, for dosimetry purposes, it is essential that light output be proportional to energy deposited i n the scintillator. There was some concern however, that there may be a loss i n intensity of the Bragg peak due to saturation of the P F G scintillator; i.e., the screen may have a maximum light output. A group at PSI i n Switzerland, tested a L A N E X screen and a Bicron screen and observed quenching of the light yield as high as 8% i n the most distal part of a normalized Bragg peak. They suggested it could be due to a high dE/dx that causes saturation of the phosphor light emitting centers. [ 4 4 ] A n explanation of this effect was given i n Section 2.4.4. From the graph i n Figure 4-8 it is reasonable to assume that saturation of the scintillating material could be a real effect; so as a follow up investigation, an experiment was done to 60 compare dose measured with an ionization chamber and dose measured with the light output by the photo-diode as a function of proton energy (i.e. dE/dx). A 25mm diameter collimator was set to define the beam and the T l chamber was used to measure the dose delivered. The T l chamber is normalized to the Primary Ion Chamber so as to act as a calibration for the dE/dx values. Readings were taken at various positions of the range shifter, which had the effect of varying the proton energy and therefore the dE/dx at the treatment position. Measurements were repeated with the photo-diode i n the screen/diode arrangement to record the light output of the screen for different range shifter positions. The T l chamber and diode readings were normalized and plotted versus the water equivalent thickness of the range shifter. Figure 4-9 shows the comparison between the two. Proton Energy (MeV) 0.0 H 1 r — 1 1 1 1 h 0 5 10 15 20 25 30 35 Range Shifter Thickness—mm water equivalent Figure 4-9 : Comparison of T l Chamber and Diode Collected Data It seems plausible that the discrepancy seen between the two curves is a function of the proton stopping powers. Because the screen has a finite thickness, the screen protons would have 61 experienced more coulomb interactions, lost more energy and thus have a higher dE/dx than the protons that travelled through the ion chamber cap for the same range shifter thickness. The dE/dx value of the screen travelled protons could be sufficiently high enough to have caused saturation of the light emitting centres of the phosphor screen. 4.4. Other Phenomena The dosimetry system was used to look at the dose profile o f several other non-homogeneous dose distributions. 4.4.1. Edge Enhancement Effect Edge enhancement is due to multiple scattering effects at the edges of two materials. The denser material produces a broader scattering at the edge so that when the two scattering effects are added there is a characteristic peak and valley i n the lateral dose response as depicted i n Figure 4-10. Proton Beam dense material less dense material Dose [ Summed response protons through dense material protons through less dense material Position Figure 4-10 : Edge Effect due to Mult iple Scattering Effects at the Edges of Two Materials 62 This effect is produced i n any non-homogenous field in which radiation scatters through areas of varying densities (or thickness). A penny was placed in the center of a 25mm diameter collimator. Figure 4-11 displays the recorded C C D image and the lateral dose profile, which shows how clearly this effect can be seen with the C C D camera/scintillating screen dosimetry system. A misaligned collimator, which could lead to an increased localized dose, can also produce this effect. See section 4.4.3 for a discussion on this. l a t e r a l p o s i t i o n ( p i x e l s ) Figure 4-11 a & b : Cross Section of C C D image of Penny shows clearly the E D G E E F F E C T 4.4.2. Wedges The use of wedges has been incorporated into the eye treatment procedure to modify the contour of the distal edge of the dose distribution. Aluminum wedges with tissue-equivalent angles from 5° to 60° are mounted a few cm from the eye on a rotating holder attached to the nozzle. Wedged fields are often essential i n reducing the dose to the macula or the optic nerve. Therefore, it was warranted to investigate i f such a non-homogeneous dose distribution also holds a linear dose response. Figures 4-12 (a,b,c) show a 60° partially wedged, 1 9 x 1 9 m m 2 63 field for various positions of the range shifter. O f further interest is the clearly visible edge effect that shows where the wedge was placed - halfway into the irradiated field. The edge of the wedge is seen as a bright line down the C C D image and corresponds to the peak seen in the cross-sectional dose distribution. Also visible is the placement of the Bragg peak for different range shifter thicknesses. A s the protons lose more energy as they pass through thicker material, their range is decreased since the stopping power has increased. Here it is possible to see the peak i n the dose is deposited closer to the entrance of the wedge and the range of the protons becomes shorter as the protons travel through different thickness of the range shifter. It is worth mentioning that these measurements were made with an un-modulated beam. For patient treatments, a modulated beam would be used and the resulting dose response across the wedge would be more linear. 64 0 10 20 30 Distance in mm Distance in mm Distance in mm Figure 4-12 : Lateral Dose Profiles from 60 degree wedge for various Range Shifter Positions 4 . 4 . 3 . Collimator and Gaussian Distribution A s mentioned i n section 2.2.2 protons passing through a single scattering foil are dispersed laterally into a distribution of intensity that is approximately Gaussian. Figure 4-13 shows the intensity distribution of a proton beam for an open collimator as measured with the CCD/screen dosimetry system. 150 A 4 6 8 10 X —Pos i t i on ( c m ) Figure 4-13 : Beam Profile for Open Collimator - single integration A lead 'button' was placed i n the center of the open collimator to observe attenuation through the button and to look at the dose profile of a non-homogeneous distribution. However, o f more interest to note was the peaks profile due to scattering through the edges of the collimator. Remarkably enough, from the rapid visualization of the dose distribution, it appeared as though something i n the equipment components was tilted. See Figure 4-14. It was thought that the dosimetry system had not been set up properly; however, it was later determined that the 66 collimator itself was tilted about 1/8". The characteristic peak and valley associated with edge enhancement can lead to an increase in the localized dose and reduce the accuracy to which the delivered dose is known. This is an example of how such a set up can be used to monitor the beam and setup of the equipment. 2 4 6 8 10 12 X-Posi t ion (cm) Figures 4-14 a & b : Lateral Cross Section of Beam Profile from Lead Button in Centre of Beam 4.5. X-ray for Real Time Patient Positioning From the initial testing of the equipment's capability, another interesting avenue of investigation emerged. The significant increase i n signal output that can be achieved with the combination of the Rarex P F G scintillating screen and the camera's integration capabilities led to an investigation into the possibility of using this technique with x-rays for rapid positioning of the patient. The present method of patient positioning consists of two perpendicular x-ray images that are taken with Polaroid f i lm (intensified by L A N E X screens) to determine the location of the tantalum clips, surgically placed on the eye to indicate tumour margins, in relation to 67 treatment system landmarks. One x-ray tube is inserted onto beam axis to view the clips through the proton nozzle for the axial view and a second tube is mounted for a lateral view through the isocentre. Instead of using the present method of X-raying Polaroid f i lm backed onto an intensifying screen, the f i l m would be replaced by the C C D technique which would allow for near real-time monitoring of the patient position. 4.5.1. Feasibility A n important consideration i n the application of this technology for rapid patient positioning is the system's ability to operate at patient safe exposures. In practice, radiation levels should be kept at the lowest practical level. Initial feasibility studies involved investigating the sensitivity of the system at standard diagnostic x-ray exposures. The P F G screen was exposed to 80kV, 50mA x-rays through a circular collimator with cross hairs through its center. The screen output was recorded with no integration by the C C D camera, which was set up i n the same configuration as for the screen sensitivity tests described in section 4.1.1. Figure 4-15 shows that it is possible to detect an un-attenuated x-ray signal. The cross hairs are clearly visible. 68 Figure 4-15: C C D Image of Collimator with cross hairs for X-ray Exposure of 80kV, 50mA Measurements were repeated for signals from an x-ray beam that had been attenuated by 9.5 cm and 23.0 cm of plastic in order to simulate attenuation through the patient's head. Neither C C D image showed any visible evidence of a detectable signal. Although it was not possible to capture an integrated image due to synchronization problems between the frame grabber system and the integration controller, 2 and 8 frame integrated images showed detectable signals on the P C image display window. A s a further check, the non-integrated captured image files were background subtracted and a cross-sectional plot was made at 'just o f f center to avoid the horizontal cross-hair of the collimator. Figure 4-16a & b show that the x-ray signal is very weak, especially through a 9.5cm thick head/tissue equivalent phantom, but is nonetheless detectable. However, Figure 4-16c shows that the x-ray signal is completely attenuated through a 23cm thick phantom. 69 350 400 450 500 Pixel Position 550 25 20-1 15 ~ 10 H 5-0 -10 9.5 cm attenuation 350 400 450 Pixel Position 500 550 25 20-15-(U 0 -5--10 23.0 cm attenuation 350 4001 450_ Pixel Position 500 550 Figures 4-16a, b, & c : Lateral Profiles of X-ray signal for 0, 9.5, and 23.0cm attenuated beam. 70 In order for the CCD/scinti l lating screen technology to be compatible for use in rapid patient positioning, the sensitivity of the system needs to be increased. A s mentioned earlier, integration was not possible due to synchronization problems between the C C D camera and the frame grabber system. One option would be to resolve these problems to allow for an integrated image to be captured and analyzed and hence, automatically increase the sensitivity of the system. However, from the calculated rate of attenuation, it was determined that for the same x-ray energy, a maximum of 24=16 frames of an un-attenuated signal could be integrated without saturating the C C D , resulting i n a very weak and barely detectable signal through a 20cm thick head equivalent phantom. Thus, a better option would be to purchase a more sensitive camera. 4.6. Radiation Damage to Camera One drawback to this dosimetry system as applied to protons is radiation damage to the C C D camera mainly due to neutrons from beam fragmentation effects. A s mentioned earlier, after only a few experimental runs damaged was noticed i n several pixels of the C C D . This naturally led into an investigation of C C D radiation damage and "damage control". 4.6.1. Hot Pixels A s the C C D experienced more and more exposure, pixel damage began to occur. The damaged or 'hot' pixels resulted i n large signals i n isolated pixels that appeared as spikes in the C C D image. Figure 4-17 is a magnification of the C C D image that shows individual damaged pixels within the image. 71 Figure 4-17 : C C D dark frame readout showing damaged or "hot" pixels. The radiation damage experienced by the C C D was tracked throughout the course of this project. A background image, or dark frame readout, was recorded before, during, immediately following each experiment and usually some duration after the experiment had been completed in order for the C C D to have an opportunity to 'recover'. A running total of the number of monitor counts was kept as a record of the amount of radiation the C C D experienced throughout the course of the project. 4.6.2. Relaxation Hot pixels were identified by measuring their intensity reading relative to the R M S deviation of the background. If the value was in excess of 4 sigma, the pixel was flagged. It should be noted from Figure 4-18 that immediately after irradiation (points A , C , F) there was a high number of 72 hot pixels but that after a period of time at room temperature, some of the radiation damaged pixels experienced a relaxation (points B , E , H) . Thus pixel damage can be separated into two categories: damage that is short term, and damage that is persistent. Figure 4-18 : Graph of Relaxation of 'Hot Pixels ' or Radiation Damaged Pixels It should also be pointed out that for the two very low readings at points D and G , the R M S background was about 19% higher than on average, possibly due to thermal effects. To give some measure of the radiation damage effect, there are 768 x 493 total pixels, so after relaxation, a pixel count (of permanently affected pixels) of approximately 50 represents a very small fraction of the total number of pixels in the image. The total integrated current of 4.5x10 6 monitor counts corresponds to a total patient dose of 500Gy, which is equivalent to approximately 10 treatments. 150 H 1 2 _ 3 m o n i t o r c o u n t s x 10~6 73 5. Conclusions A rapid method of measuring dose profiles using a sensitive scintillating screen observed by a C C D camera has been successfully developed. The system has been demonstrated to be useful for monitoring beam profiles and for measuring lateral dose distributions to look at collimator scattering, beam steering, raw Bragg peaks, and effect of wedges. It is clear that this system could be used to monitor beam uniformity and symmetry and also possibly be adapted for quality control purposes i n the T R I U M F Proton Therapy Facility. The main advantage of this sort of dosimetry system is its ability to do real-time observation and analysis of the proton beam. Conventional proton dosimetry methods can be time consuming whereas the C C D technology allows for immediate acquisition of information. To ensure that this system offered reliable dosimetry, each component was investigated to test that it gave a linear dose response. The light response of several commonly used scintillating materials and various commercially available medical-intensifying screens was investigated for both protons and x-rays. A s wel l , the screens were tested for the relative sensitivity of their dose response. The Rarex P F G (ZnCdS:Ag) screen has been found to be 17 times more sensitive than the more commonly used L A N E X screen (Gd202S:Tb). The C C D camera response to x-ray and proton irradiation was also evaluated. Overall, the system shows high sensitivity, low noise, and linear behaviour for beam energies and intensities wel l matched to clinical doses. B y mounting the screen on a Lucite wedge^ it was possible to measure the raw Bragg peak. The expected 4 to 1 ratio i n intensity from peak to plateau is clearly observable. A direct comparison between a dose distribution as measured by the conventional ion chamber scan i n water and a 74 dose distribution as measured by the C C D method confirmed there is a good match between the two depth dose profiles. The apparent loss i n intensity at the Bragg peak seen i n the C C D curve, although possibly due to a manifestation of many different effects, is most l ikely due to saturation of the scintillation material at high dE/dx since this behaviour has been observed and reported by others. Whi le this effect may preclude the use of this system for absolute dosimetry, the observed reproducibility and linearity of the C C D system is within the +/- 2% precision requirement for beam profile measurements. A l l o f the data collected up to this point was for an unmodulated beam. Due to synchronization problems between the C C D camera and the frame grabber, a maximum of 8 frames or 240 ms could be integrated. The modulator operates at 16 modulations per second or one modulation in 62 ms. W i t h no synchronization between the modulator and the camera integration, readouts taken at different times w i l l give different dose profiles. For correct recording of a spread out Bragg peak the readout must contain an integral number of modulations, or integrate over a sufficiently large number of modulations. It might be possible to use the proximity switch on the modulator to act as an external trigger on the C C D integration and arrange for the modulator rotation speed to be timed to the camera frame speed. Another limitation of the system is radiation-induced damage to the camera due to neutrons produced by the proton beam. A s the C C D camera experiences more and more exposure, pixel damage begins to occur. Although the pixels appear to experience a relaxation i f given enough time to recover, some persistent pixel damage does remain. This damage appears as white pixels 75 i n the image and as high spikes in the lateral profile. However, the total number of permanently damaged pixels represented a very small fraction of the total pixels i n the C C D over the course of the studies. One solution may be to find a more radiation resistant camera, but the benefits of increased resistance may be offset by cost. A conceivable application of the C C D technology is the use of this technique with diagnostic x-rays for rapid positioning of the patient. Instead of using the present method of x-raying f i lm backed onto an intensifying screen, the film would be replaced by the CCD/Screen apparatus which would allow for near real-time monitoring of the patient. A n important consideration is the system's ability to operate at patient safe exposures. Even with the P F G screen, the present system components are not sufficiently sensitive to detect a signal that has been attenuated through a head equivalent thickness. However, preliminary studies are promising and presumably sensitivity limitations could be solved by a more sensitive camera. The ability of the system to measure dose distribution of non-homogenous fields could lead to a variety of different applications. The advantages of such a system's rapid visualization capabilities was recognized by the H C L group who are using a similar screen/CCD dosimetry system for gantry alignment^4 5 1 Using a metal ball , centered in a hollow sphere, they are observing the characteristic peaks and valleys i n the dose profiles associated with the edge effect with a screen/CCD apparatus as a means of gantry alignment. The main intent i n undertaking this study was to build on the concept of using a camera to monitor the light output from a scintillating screen and extend it for application as a dosimetry system i n the T R I U M F Proton Therapy Facility. It investigated the possibility of using a C C D 76 camera/scintillating screen combination for this task. The important parameters of interests were sensitivity of the system for clinically equivalent doses, linearity of the dose response of the system, and ability to process the image data file for immediate viewing. Based on the results of this project, future work may be concerned with improving the system to achieve higher sensitivity, make it capable of monitoring a modulated beam, and allowing for online image processing to display dose distributions directly on the P C monitor. 77 References 1 Andrews, H. L. Radiation Biophysics. Prentice Hall Inc., (1974) 2 Tsujii, H.; Tsuji, H.; Inada, T.; Maruhashi, A.; Hayakawa, Y.; et al. "Clinical Results of Fractionated Proton Therapy." Int. J. Radiat. Oncol. Biol. Phys. 25 (1993) p 49-60. 3 Wilson, R.R. "The Radiological Use of Fast Protons." Radiology 47, (1946) p. 487. 4 Miller, D. " A Review of Proton Beam Radiation Therapy." Medical Physics 22 (11), November (1998)p 1943-1954. 5 Sisterson, J. (ed.) "World Wide Charged Particle Patient Totals." Particles 21 (1998) 6 Blackmore, E.W.; Vincent, J.; Chavez, S.; Gardney, K.; Lam, G.; Oelfke, U.; Pickles, T.; Patton, K. "Commissioning the TRIUMF Proton Therapy Facility." Proceedings of the MRS International Seminar on the Application of Heavy Ion Accelerator to Radiation Therapy of Cancer. Chiba, Japan. 135 (1994). 7 Blackmore, E.W.; Evans, B.; Mouat, M. ; Duzenli, C ; Ma, R.; Pickles, T.; Paton, K. "Operation of TRIUMF Proton Therapy Facility." Proceedings of the 1997 Particle Accelerator Conference, Vancouver. May 1997. 8 Slater, J. M. ; Facr, J.O.; Archambeau, M.D.; Miller, D.W.; Notarus, M.I.; Preston, W.; Slater, J. D. "The Proton Treatment Center at Loma Linda University Medical Center: Rationale for and Description of its Development." Int. J. Radiation Oncology Biol. Phys. Vol. 22 (1992) p 383-389. 9 Flanz, J.; Bradley, S.; Durlacher, S.; Goitein, M. ; Loeffler, J.; Smith, A.; Woods, S. "Status of the Northeast Proton Therapy Center, Boston Construction." Particles 21 (1998) 11 Deeley, T.J.; Hale, B.T. "The Past Seventy-five Years in Radiotherapy." Brit. J. Radiology. 46 (1973)p 906-910. 12 G. Lam, E. El-Khatib. "Heavy Particles for Radiotherapy." Physics in Canada. 51:4 (1995) p 174-177. 13 Skarsgard, L.D.; Kihlman, B.A.; Parker, L.; Pujara, C.C.; Richardson, S. "Survival, Chromosome Abnormalities, and Recovery in Heavy Ion and X-irradiated Mammalian Cells." Radiol. Res. Supple. 7 (1967) p 208-221. 14 Smith, A. " A Report on the Change in Proton Absorbed Dose Measurement Protocol for the Clinical Trials Conducted at the Harvard Cyclotron Laboratory", presented at the Proton Therapy Co-operative Group (PTCOG) Meeting. April 16, 1998. Rancho Mirage, CA. 15 Chu, W. T.; Ludewigt, B.A.; Renner, T.R. "Instrumentation for treatment of cancer using proton and light-ion beams." Rev. Sci. Instrum. 64 (8), August (1993) p 2055-2122. 78 16 Koch, A. "Lens Coupled scintillating screen-CCD X-ray area detector with a high detective quantum efficiency." Nuclear Instruments and Methods in Physics Research A, 348 (1994) p 654-658. 17 Karellas, A.; Harris, L.; Lui, H.; Davis, M. ; D'Orsi, C. " Charge-coupled device detector: Performance considerations and potential for small-field mammographic imaging applications." Med Phys 19, 4. (1992) p 1015-1022. 18 Schippers, J.M.; Boon, S.N.; van Luijk, P.; Pedroni, E.; Coray, A. "Fast 2-D Dosimetry at a Scanning Proton Beam." PSI Life Sciences Report. (1997) p 23. 19 Robinson, D. M. "Megavoltage Photon Dose Modification Resulting from Tissue Compensator Retraction." Masters of Science Thesis, University of Alberta, Department of Physics. (1986) p 3. 20 Frauenfelder, H.; Henley, E.M. Subatomic Physics. Prentice-Hall, INC. (1974) p 36. 21 Review of Particle Properties, Physical Review. (1994) 22 Chu (1993) 23 Johns(1983) 24 Chu (1993) 25 Koehler, A.M. "Dosimetry of Proton Beams using Small Silicon Diodes." Radiation Research Supplement. 7 (1967) p 53-63. 26 Blackmore(1994) 27 Blackmore(1997) 28 Goitein, M. ; Miller, T. "Planning Proton Therapy of the Eye." Medical Physics. Vol. 10, No. 3 May/June (1983) p 274-283. 29 Christensen, E.; Curry, T.; Dowdey, J. "Chapter 9 : Intensifying and Fluoroscopic Screens." An Introduction to the Physics of Diagnostic Radiology. Henry Kimpton Publishers, London. (1978) 30 Birks, J.B. The Theory and Practice of Scintillation Counting. Macmillan Company (1964) p 541. 31 Newell, John D. Jr.; Kelsey, Charles A. "Chapter 6 : Digital Radiography using Storage Phosphors." Digital Imaging in Diagnostic Radiology. Churchill Livingstone, New York. (1990) p 107-132. 32 Birks, p 202. 33 Beynon, J.D.E.; Lamb, D.R. (eds). Charge Couple Devices and their Applications. McGraw-Hill Book Company, London. (1980) 79 34 Hopkinson, G.R.; Dale, C. J.; Marshall, P.W. "Proton Effects in Charge-Coupled Devices." IEEE Transactions on Nuclear Science. Vol 23, No 2. (1996) p 614-627. 35 Available from Eastman Kodak Company, Rochester, NY 36 Schippers, J.M.; Boon, S. N.; van Lujik, P.; Pedroni, E.; Coray, A. "Fast 2D Dosimetry at a Scanning Proton Beam." PSI Report, Life Sciences (1997) p 23. 37 Muller, w.; Hartmaan, FJ.; Eades, J.; Hayano, R.S.; Ketzer, B.; Maas, F.E. "Online Monitoring of Charged-Particle Beams using a CCD Camera with Image Intensifier." Nuclear Instruments and Methods in Physics Research A. 349(1994) p 307-309. 38 Hollerman, W.A.; Fisher, J.H.; Shelby, G.A.; Holland, L.R.; Jenkins, G.M. "Spectroscopic Analysis of Proton Induced Fluorescence from Yttrium and Gadolinium Oxysulfide Phosphors." IEEE Transactions (1993) p. 791-793. 39 Holland, L.R.; Jenkins, G.M.; Fisher, J.H.; Hollerman, W.A.; Shelby, G.A. "Efficiency and Radiation Hardness of Phosphors in a Proton Beam." Nuclear Instruments and Methods in Physics Research B 56/57 (1991) p. 1239-1241. 40 Bicron-412 premium plastic scintillator available from Bicron, Newbury, OH 41 Available from MCI-Optonix, Cedar Knolls, NJ 42 Imaging Technology Incorporated. Bedford, MA. 43 MVC IC_PCI Hardware Reference Manual. Imaging Technology Incorporated, USA. (1995) 44 Boon, S.N.; van Luijk, P.; Schippers, J. M.; Meertens, H.; Denis, J.M.; Vynckier, S.; Medin, J.; GruselLE. Medical Physics 25(4), (1998) p 464-475. 45 Schippers M. et al. " A Device for Isocenter Alignment in the NPTC Gantry." Presentation at PTCOG XXVHJ Conference, Rancho Mirage, CA. April 15-17, 1998. 46 Johns, H.E.; Cunningham, J.R. The Physics of Radiology. Charles C. Thomas (1983) pg 190. 47 Frauenfelder (1976) 48 Johns, p i 86. 49 Dearnaly, G.; Northrop, D.C. Semiconductor Counters for Nuclear Radiations. John Wiley & Sons, Inc. (1966) p 63-65. 50 Muller, R.; Kamins, T.I. Device Electronics for Integrated Circuits. John Wiley & Sons, Inc. (1986) 80 Appendices Appendix A : Calculation of Stopping Power of Heavy Charged Particles. Figure A - l ' - 4 6 - ' shows the configuration of a heavy charged particle as it interacts with the bound electrons of the material through which it is passing. Let m0 be the mass of the electron, and M be the mass of the heavy charged particle such that M» m0 —> Figure A - l : Interaction of a Heavy Charged Particle with Matter A s the heavy charged particle (Ze) travels along path MQ' with velocity v, an electron at Q' w i l l experience a Coulomb Force F as given by Equation ( A - l ) . FJvLjVp„ ( A - i ) r r Since M» m0 , then the force Fy w i l l deflect the particle slightly i n the y-direction. However, an electron at Q w i l l feel a net impulse, and a transfer of energy from the heavy charged particle to the electron w i l l occur. 81 The loss of energy AE felt by a heavy charged particle due to a series of Coulomb interactions, is calculated by first considering the change i n momentum Ap the particle experiences as given in Equation (A-2). Ap=\Fydt (A-2) From Equation ( A - l ) and we see that Equation (A-2) can be rewritten in terms of the y-component of F using cosG as in Equation (A-3) "kZe2 Ap = f — j - c o s Q d t (A-3) W e see from Figure A - l that tan0=x/fe and dt=dx/dv. Taking the derivative of tanG and substituting in for dx, we get an expression for dt as given by Equation (A-4). dt = - s e c 2 QdQ (A-4) v Rewriting Equation (A-3) with the new expression for dt we get Equation (A-5). W e can further simplify by replacing r with b/cosQ to get Equation (A-6). 0 0 h-7 ^ h 1 Ap = cos9 --sec20^e — — (A-5) J r v cos0 82 kZe1 '} A p = ™±_ C c o s Q d Q ( A . 6 ) A D J . bvJ - . J Solving the integral, and making the substitution ke = r0m0c where r0 is the classical electron radius as calculated i n Appendix C , the momentum transferred to the electron from the heavy charged particle can be described by Equation (A-7). bv The energy transfer written in terms of momentum, relativistically, is described by Equation (A-8) and the kinetic energy of the heavy charged particle is given by Equation (A-9). Using these relations we get a final expression, Equation (A-10), for the transfer of energy from the heavy charged particle to the absorbing material. AE = &t (A-8) 2m„ KE = -Mv2 (A-9) 2 b KE It is important to realize that the particle does not interact solely with one electron, but rather w i l l experience a series of energy transfers along its path as it interacts with the many bound electrons scattered throughout the material. B y assuming the electrons of the absorbing material 83 are randomly distributed in space around the path of the incident particle, we can look instead at the total energy lost per unit distance. Consider the number of electrons, given in Equation (A-11), that w i l l be located in a cylindrical shell of length Ax between radii b, and b+db as shown i n Figure A - 2 . A x radius = b+db radius = b path of incident particle Figure A - 2 : Cylindrical Shell o f radii b+db o f randomly distributed electrons in absorbing material. An- Ne- 2np • bdb • Ax (A-l l ) where Ne is the electron density and p is the mass density of the absorbing material. To find the total energy loss AE(b), i n distance Ax we take the product o f AE(b) and An and integrate over all possible b values from bmin to b„ Often referred to as the stopping power, Equation ( A - 1 2 ) , describes the total amount of energy lost per unit length along the particle track. This process was also considered quantum mechanically by Bethe and Bloch to take into account any relativistic effects. The result is 84 called the Beth-Bloch equation[47], and can be approximated by Equation (A-13). I denotes the average ionization energy of the atom. dE _ AnpNeZ r0m0c4 r db ~dx~~ V ' T (A-12) dE _ 4nr*pNeZ2m0c4 dx v 2 In 2m„v 7(1 - v / c 2 ) c2 (A-13) 85 Appendix B : Classical Electron Radius An electron in an electric field Ei will experience a Coulomb force F as described by Equation (B-l) and thus will have an acceleration a as described by Equation (B-2). F = keEl (B-l) a = ^ = ^ (B-2) Classical theory tells us that an accelerated charge will radiate energy in the form of electromagnetic radiation. Electric field vectors, shown in Figure B - l , set the free electron at P into oscillatory motion with accelerations ai and a2.[48] Figure B- l : Free electron at P with Accelerations aj and 02 Due to Field Vectors E\' and E2' 86 The electric vector of the radiated electromagnetic wave at point Q is given by Equation (B-3) E ; = O £ ™ * ( B _ 3 ) c r From Equation (B-2) we see that Equation (B-3) can be written to give Equation (B-4) Ex = - ^ ^ i - s i n Y (B-4) mc r From this expression we get the classical electron radius, r0, as defined by Equation (B-5) Ice2 r0=^r (B-5) mc 87 Appendix C : Semi-Conductor Band Gap Theory The energy bands in a semiconductor arise from the allowed energy levels of the electrons in the individual atoms which make up the c r y s t a l [ 4 9 ] . In a semiconductor, at absolute zero, all the available electrons f i l l the lowest energy levels causing one or more energy bands to be completely ful l . The highest fil led level is separated from the next higher band by an energy interval Eg called the forbidden gap shown in Figure D - l i n which there are no allowable energy levels. The highest fil led level is called the valence band and the lowest empty level is called the conduction band. conduction band allowed states forbidden gap allowed states Eg valence band Figure C - l : Energy bands of a semiconductor material A t absolute zero no conduction can occur because there are no electrons i n the conduction band. However, at any non-zero temperature, the valence band is not entirely filled because a small number of electrons have enough energy to be excited across the forbidden gap into the 88 conduction band. The smaller the gap, and the higher the temperature, the more electrons that can jump between bands. These electrons can easily gain more energy and respond to an electric field to produce a current . [ 5 0 ] When electromagnetic ( E M ) radiation is incident on a semiconductor, some of the radiation may be absorbed and thereby stimulate electronic transitions from one energy state to another. This process, known as the Opto-Electronic interaction, is the basis of many types of radiation detectors. Radiation, as it propagates through an absorbing material, w i l l experience an exponential decay i n its intensity according to Equation ( C - l ) . a (ko) describes the relative rate of decrease i n radiation intensity I(hv) along the path of propagation x. I(x) = I0exp(-ax) ( C - l ) where I0 is the incident (at surface) intensity A thorough description of the absorption process involves solid state physics theory, which is beyond the scope of this thesis. The absorption of radiant energy i n a semiconductor allows for an electron to be raised from the valence band to the conduction band giving rise to the formation of an electron-hole pair. This process is called Intrinsic Absorption. A t absolute zero, the valence band would be full there would be no energy available for an electron to be excited to a higher energy state. The only mode of excitation would be from the 89 absorption of energy from an 'external' source. Sufficient energy would be needed from impinging photons to excite electrons across the band gap and into the conduction band. A t any non-zero temperature, however, it is possible for the electrons and holes to make transitions via the absorption of thermal energy rather than radiant energy. The usefulness of an image sensor is limited by the competition between optical and thermal current generation. In materials with narrow band gaps, the concentration of intrinsic carriers is high and results in an increased probability of generating a dark current. Semiconductor materials, however, have impurity or crystal lattice defect energy levels in their energy band gaps. These defect levels increase the probability that a valence band electron w i l l transition into the conduction band by radiant energy absorption. Furthermore, semiconductors have wide band gaps and therefore have a smaller number of intrinsic carriers available for thermal generated transitions. These two properties combined make semiconductors popular materials for use i n image sensors. 90 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items