UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A dosimetric study of kilovoltage cone beam CT image guidance in gynaecological radiation therapy Qiu, Yue 2012

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

Item Metadata


24-ubc_2013_spring_qiu_yue.pdf [ 5.84MB ]
JSON: 24-1.0073485.json
JSON-LD: 24-1.0073485-ld.json
RDF/XML (Pretty): 24-1.0073485-rdf.xml
RDF/JSON: 24-1.0073485-rdf.json
Turtle: 24-1.0073485-turtle.txt
N-Triples: 24-1.0073485-rdf-ntriples.txt
Original Record: 24-1.0073485-source.json
Full Text

Full Text

A Dosimetric Study of Kilovoltage Cone Beam CT Image Guidance in Gynaecological Radiation Therapy by Yue Qiu  B.Sc., Peking University, 2003 M.Sc., The University of British Columbia, 2006  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in The Faculty of Graduate Studies (Physics)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) December 2012 c Yue Qiu 2012 ⃝  Abstract Cone beam CT (CBCT) is increasingly used in image guided radiation therapy in order to ensure more accurate patient setup and target localization. However, the issue of patient dose due to repeated imaging procedures is an important topic in radiation medicine today. This thesis rigorously explores the organ equivalent doses from kilo-voltage cone beam CT to provide useful data to clinicians for decision making in image protocol development. This work specifically focuses on gynaecological patients undergoing megavoltage radiation therapy using IMRT or RapidArc. A method is established for Monte Carlo simulation of dose distributions from kV CBCT using BEAMnrc/DOSXYZnrc, for an on-board imaging system mounted on a Varian linac. Careful benchmarking was performed to permit further investigation into patient specific dose. In order to account for different biological effects of kV and MV radiation, the investigation focused on equivalent dose, the product of dose and radiation quality factor. Dose mean lineal energy was used to determine beam quality, as opposed to using linear energy transfer, based on microdosimetric principles. Dose mean lineal energy was simulated with a C++ code and the Monte Carlo code NOREC. Quality factor was calculated based on lineal energy by a formula proposed by International Commission ii  on Radiation Units and Measurements (ICRU). Higher relative biological effectiveness of kV CBCT beams compared with MV therapeutic beams was quantified. Finally, the Monte Carlo methods for kV CBCT dose simulation were combined with derived quality factors in a study of organ equivalent dose for gynaecological cancer patients. Using three different dose response models, organ equivalent doses for daily kV CBCT were compared with MV IMRT/RapidArc doses to assess whether patients are at increased risk of radiation induced secondary malignancy due to CBCT. Comprehensive dosimetric information presented in this thesis indicates that patients undergoing IMRT or RapidAcr treatments with daily kV CBCT are not subjected to additional risk due to CBCT imaging dose.  iii  Preface Work related to content in Chapter 3 was presented at the Annual Meeting of Canadian Organization of Medical Physicists 2010: Y Qiu, F Bachand, V Moiseenko, P Lim, C Aquino-Parsons, and C Duzenli. Cinical Implementation of kV CBCT for Gynecological IMRT. Med. Phys. 37, 3891. Dr Aquino-Parsons proposed this study, Dr Bachand did the contours and some data analysis. I worked with Dr Duzenli and Dr Moiseenko on design in the aspect of physics and performed the dose calculations. This study was granted by the BC Cancer Agency Ethics Board (ID: H08-01519). A version of Chapter 4 has been published in Radiation Protection Dosimetry in 2011: Y Qiu, IA Popescu, C Duzenli, V Moiseenko. Megavoltage versus kilo-voltage cone beam CT used in image guided radiation therapy: comparative study of microdosimetric properties. 143(2-4): 47780. In this study, I worked with Dr. Moiseenko and Dr. Duzenli in the design of the method, developed the C++ code and ran the calculations. Additionally, I wrote the manuscript. The work presented in Chapter 6 was partially presented at the AAPM Annual Scientific Meeting in 2010: Y Qiu, V Moiseenko, IA Popescu, and C Duzenli. Equivalent Doses and Secondary Malignancy Risk Estimates for Gynecological Patients Undergoing CBCT and IMRT. Med. Phys. 37, 3462 iv  (2010). This work has been published in Radiotherapy and Oncology: Qiu Y, Moiseenko V, Aquino-Parsons C, Duzenli C. Equivalent doses for gynecological patients undergoing IMRT or RapidArc with kilovoltage cone beam CT. 104(2):257-62 (2012). I performed the Monte Carlo dose calculations, dose measurements, calculation of OEDs and co-wrote the manuscript.  v  Table of Contents Abstract  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ii  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  iv  Table of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . .  vi  List of Tables  x  Preface  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii List of Abbreviations  . . . . . . . . . . . . . . . . . . . . . . . . . xvii  Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . xx 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  1  2 Background  . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  7  Radiation Therapy . . . . . . . . . . . . . . . . . . . . . . . .  7  2.1.1  CT Simulation . . . . . . . . . . . . . . . . . . . . . .  8  2.1.2  Linac . . . . . . . . . . . . . . . . . . . . . . . . . . .  10  2.1.3  Multi-Leaf Collimator (MLC)  . . . . . . . . . . . . .  11  2.1.4  Intensity Modulated Radiation Therapy - IMRT . . .  12  2.1  vi  2.1.5 2.2  2.3  2.4  RapidArc . . . . . . . . . . . . . . . . . . . . . . . . .  Image Guidance in Radiation Therapy  . . . . . . . . . . . .  16  2.2.1  In-room CT Systems  . . . . . . . . . . . . . . . . . .  16  2.2.2  Clinical Implementation of kV CBCT . . . . . . . . .  19  2.2.3  On Board Imager (OBI)  . . . . . . . . . . . . . . . .  20  2.2.4  Scan Modes  . . . . . . . . . . . . . . . . . . . . . . .  22  2.2.5  Image Quality . . . . . . . . . . . . . . . . . . . . . .  23  Radiation Induced Carcinogenesis  . . . . . . . . . . . . . . .  25  2.3.1  Study of Atomic-bomb Survivors . . . . . . . . . . . .  26  2.3.2  Study of Radiation Therapy Patients  . . . . . . . . .  28  2.3.3  Other Studies  . . . . . . . . . . . . . . . . . . . . . .  30  2.3.4  Mechanism of Carcinogenesis . . . . . . . . . . . . . .  31  Risk Assessment Models and Methods  . . . . . . . . . . . .  33  2.4.1  Low Dose Risk Assessment . . . . . . . . . . . . . . .  33  2.4.2  High Dose Risk Assessment: Concept of Organ Equivalent Dose (OED) . . . . . . . . . . . . . . . . . . . .  3 CBCT in Gynaecological Cancer IMRT Treatment 3.1  . . . .  34 36  Introduction to Gynaecological Cancer and Treatment with Radiation Therapy  . . . . . . . . . . . . . . . . . . . . . . .  36  3.1.1  Epidemiology and Treatment . . . . . . . . . . . . . .  36  3.1.2  Radiation Therapy Technique: From Pelvic Radiation Therapy to IMRT . . . . . . . . . . . . . . . . . . . .  3.2  14  CBCT Usage in Gynaecological Radiation Therapy  . . . . .  37 44  vii  4 Cone Beam CT Dose by Monte Carlo Simulation 4.1  . . . . .  49  Introduction to Monte Carlo Simulation . . . . . . . . . . . .  49  4.1.1  Simulation Codes  51  4.1.2  EGSnrc  . . . . . . . . . . . . . . . . . . . .  (Electron Gamma Shower - National Research Coun. . . . . . . . . . . . . . . . . . . . . . . . . . . .  51  4.1.3  BEAMnrc Code . . . . . . . . . . . . . . . . . . . . .  53  4.1.4  DOSXYZnrc Code . . . . . . . . . . . . . . . . . . . .  54  4.1.5  NOREC Code . . . . . . . . . . . . . . . . . . . . . .  55  Calibration of a kV Cone Beam CT system . . . . . . . . . .  56  4.2.1  Building the X-ray Source Model  . . . . . . . . . . .  56  4.2.2  Relative Dose Distribution  . . . . . . . . . . . . . . .  58  4.2.3  Absolute Calibration  . . . . . . . . . . . . . . . . . .  61  Dose Distributions from Cone Beam CT Scans . . . . . . . .  63  4.3.1  . . . . . . . . . . . . . . .  63  cil)  4.2  4.3  Sample Dose Distributions  4.4  Discussion  . . . . . . . . . . . . . . . . . . . . . . . . . . . .  64  4.5  Conclusion  . . . . . . . . . . . . . . . . . . . . . . . . . . . .  67  5 Dose Mean Lineal Energy Based Quality Factor 5.1  Introduction to Beam Quality 5.1.1  . . . . . .  68  . . . . . . . . . . . . . . . . .  68  Relative Biological Effectiveness (RBE) and Quality Factor (Q)  . . . . . . . . . . . . . . . . . . . . . . . .  5.1.2  LET and Quality Factor  5.1.3  Microdosimetry  5.1.4  RBE for Low Energy X-rays  68  . . . . . . . . . . . . . . . .  70  . . . . . . . . . . . . . . . . . . . . .  73  . . . . . . . . . . . . . .  75  viii  5.2  Monte Carlo Simulated Lineal Energy . . . . . . . . . . . . .  77  5.2.1  Materials and Methods . . . . . . . . . . . . . . . . .  77  5.2.2  Results and Conclusion . . . . . . . . . . . . . . . . .  79  5.3  Discussion  . . . . . . . . . . . . . . . . . . . . . . . . . . . .  81  5.4  Conclusion  . . . . . . . . . . . . . . . . . . . . . . . . . . . .  82  6 Organ Equivalent Doses for Gynaecological Patients Undergoing IMRT or RapidArc with kV CBCT  . . . . . . . . . .  83  . . . . . . . . . . . . . . . . . . . . . . . . . . .  83  6.1  Introduction  6.2  Materials and Methods 6.2.1  . . . . . . . . . . . . . . . . . . . . .  Dose Distributions in Gynaecological Patients  86  . . . .  86  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  88  6.3  Results  6.4  Discussion  . . . . . . . . . . . . . . . . . . . . . . . . . . . .  92  6.5  Conclusion  . . . . . . . . . . . . . . . . . . . . . . . . . . . .  97  7 Conclusions and Future Work . . . . . . . . . . . . . . . . . .  99  Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102  ix  List of Tables 2.1  Scan modes and detailed setups for CBCT . . . . . . . . . . .  5.1  Comparison of measured lineal energy and simulated value for monoenergetic photons. . . . . . . . . . . . . . . . . . . .  5.2  24  79  Dose-mean lineal energies and quality factors of kV CBCT for a paediatric abdomen scan (with half fan bowtie).Computational uncertainty is about 2 %. . . . . . . . . . . . . . . . . . . . .  5.3  80  Dose-mean lineal energies and quality factors of kV CBCT for a head and neck scan (with full fan bowtie).Computational uncertainty is about 2 %. . . . . . . . . . . . . . . . . . . . .  6.1  81  Comparison of CBCT doses and IMRT and RapidArc doses for organs in the peripheral region. *Lung and breast are calculated on partial organ volumes due to cutoff at the sup boundary of CT is at the fourth rib and are thus higher than mean dose to the whole organ. . . . . . . . . . . . . . . . . .  6.2  91  Bladder, bowel and rectum OEDs from IMRT and combination of IMRT and CBCT for small (S), medium (M) and large(L) patients. . . . . . . . . . . . . . . . . . . . . . . . . .  93  x  6.3  OED increments with original parameters and parameters adjusted by +/- 50% are listed for patient M. . . . . . . . . . .  94  xi  List of Figures 2.1  Diagram of fan beam geometry with a complete rotation of the x-ray tube around a stationary ring of x-ray detectors.  .  10  asymmetric collimating jaws, (8) multileaf collimator. . . . .  12  2.3  Varian Millennium 120 MLC model. . . . . . . . . . . . . . .  13  2.4  This is the CTVision system from Siemens. It consists of a  2.2  Cutaway view of a modern linear accelerator with multi-leaf collimators for beam shaping. (1) electron gun, (2) standing waveguide, (3) bending magnet, (4) target, (5) carousel of scattering foils and flattering filters, (6) ion chamber, (7)  linear accelerator and a modified diagnostic CT scanner (CTon-rails). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  17  2.5  This is a Varian Trilogy Linac with Cone Beam CT imager. .  18  2.6  This is a Siemens linear accelerator with electronic portal imaging device.  2.7  . . . . . . . . . . . . . . . . . . . . . . . . .  19  Photos of bowtie filters. Left: body bowtie (half fan bowtie) Right: head bowtie (full fan bowtie). . . . . . . . . . . . . . .  21  xii  2.8  CBCT acquisition modes with head fan or body fan. Left: The head fan bowtie is mounted, the detector is centered with x-ray axis and the rotation is 200 degrees. Right: Body fan bowtie is mounted, the detector is shifted and the gantry rotation is 360 degrees. . . . . . . . . . . . . . . . . . . . . . .  3.1  22  Contours of CTV vagina(cyan), CTV vessels(cyan), bladder(red) and rectum(brown) in transverse, coronal and sagittal views. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  3.2  40  Treatment planning CT simulation images and associated contours of PTV vagina(blue), PTV vessels(blue), bladder(red) and rectum(brown) in transverse, coronal and sagittal views. PTVs are defined as CTVs (cyan) plus a 7mm margin.  3.3  . . .  7 field IMRT is demonstrated on a transverse CT slice. The PTVs are covered with 95% isodose line (light green). . . . .  3.4  43  Contours of CTV vagina, rectum and bladder on sagittal planning CT, CBCT and registered CT/CBCT images. . . .  3.6  42  Dose volume histograms for CTV vagina, CTV vessels, rectum, bladder and bowel. . . . . . . . . . . . . . . . . . . . . .  3.5  41  45  Inter-fraction motion of bladder and rectum. Note that the vaginal CTV lies between and abuts both structures. The CTV is not displayed on this image in order not to obscure the bladder and rectal contours. Bladder Day 1 and Rectum Day 1 are structures on the CBCT images and Bladder Day 2 and Rectum Day 2 are the projections from planning CT. .  47  xiii  4.1  Example of linear accelerator constructed in BEAMnrc. Image is from BEAMnrc user’s manual[64].  4.2  . . . . . . . . . . .  54  Accelerator Preview in BEAMnrc. Target (W-Re) , blades (Steel) and halffan bowtie (Aluminum) are shown in the XZ plane. X axis and Z axis show the scales in centimeter.  4.3  . . .  57  Measured (Meas) and Monte Carlo simulated (MC) PDDs for 60, 80, 100,125kVp x-ray beams. The field settings are SSD 100cm, field size 30cmX30cm, without bowtie. . . . . . . . . .  4.4  59  Measured in-plane dose profile (Meas) compared with Monte Carlo simulated dose profile (MC) at depth of 2 cm for 100kVp x-ray beam. The field settings are SSD 100cm, field size 30cmX30cm, without bowtie. Relative dose is normalized to 100% at the center of the field. . . . . . . . . . . . . . . . . .  4.5  60  Measured cross-plane dose profile (Meas) compared with Monte Carlo simulated dose profile (MC) at depth of 2 cm for 100kVp x-ray beam. The field settings are SSD 100cm, field size 30cmX30cm, without bowtie. Relative dose is normalized to 100% at the center of the field. . . . . . . . . . . . . . . . . .  4.6  61  Demonstration of heel effect. Photon intensity at side A is lower than that at side B because of more absorption by the target.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  62  xiv  4.7  Measured dose profile (Meas) compared with Monte Carlo simulated dose profile (MC) at depth of 2 cm for 100kVp xray beam with half-fan bowtie. The field settings are SSD 100cm, field size 30cmX30cm. Relative dose is normalized to 100% at the center of the field. . . . . . . . . . . . . . . . . .  4.8  63  Dose distribution in one scan for gynecological patient. Dose is represented on a color wash scale indicated in the lower right corner.  5.1  . . . . . . . . . . . . . . . . . . . . . . . . . . .  64  Energy deposition for X-ray and particles with different LETs. For x-rays, double-strand breaks can not caused by one track. When LET is around 100keV /µm, the average separation between ionizing events coincides with the diameter of the DNA double helix and the radiation is most efficient in causing double-strand breaks. As LET become higher than 100keV /µm, some energy was “wasted” between the two strands so that RBE decreases.  6.1  . . . . . . . . . . . . . . . . . . . . . . . . .  72  Monte Carlo simulated CBCT dose distribution for a single pelvis scan. . . . . . . . . . . . . . . . . . . . . . . . . . . . .  89  xv  6.2  Comparison of doses for 25 fractions showing cumulative CBCT doses and IMRT and leakage doses along the patient midline (black line) in Fig 1. The X-axis represents the distance from IMRT isocenter from inferior (-12 cm) to superior (+20 cm). The prescribed dose was 4500 cGy. The IMRT field extends from -6 cm to 9.3 cm, and the cone beam CT field extends from -11.8cm to 11.8cm as indicated on the figure. . . . . . .  90  xvi  List of Abbreviations CBCT Cone Beam Computed Tomography CM Component Modules CRT Conformal Radiation Therapy CT Computed Tomography CTV Clinical Target Volume CTDI Computed Tomography Dose Index DNA Deoxyribonucleic Acid DVH Dose Volume Histogram EAR Excess Absolute Risk EGSnrc Electron Gamma Shower - National Research Council EPID Electronic Portal Imaging Device ERR Excess Relative Risk GTV Gross Tumor Volume HPV Human Papillomavirus HU Hounsfield Unit IARC International Agency for Research on Cancer ICRP International Commission on Radiological Protection ICRU International Commission on Radiation Units and Measurements IGRT Image Guided Radiation Therapy xvii  IMRT Intensity Modulated Radiation Therapy kVp Kilovoltage Peak potential LAR Lifetime Attributable Risk LET Linear Energy Transfer LSS Life Span Study MC Monte Carlo MLC Multileaf Collimator MRI Magnetic Resonance Imaging MU Monitor Unit MV Megavoltage accelerating peak potential NIST National Institute of Standards and Technology NOREC New Oak Ridge Electron Transport Code OAR Organ at Risk OBI On-Board Imager OED Organ Equivalent Dose OMEGA Ottawa Madison Electron Gamma Algorithm PDD Percentage Depth Dose PTV Planning Target Volume QUANTEC Quantitative Analyses of Normal Tissue Effects in the Clinic RBE Relative Biological Effectiveness RERF Radiation Effects Research Foundation ROI Region of Interest RR Relative Risk RT Radiation Therapy RTOG Radiation Therapy Oncology Group xviii  SAD Source-to-Axis Distance SBRT Stereotactic Body Radiotherapy SDD Source to Detector Distance SPR Scatter to Primary Ratio SSD Source-to-Surface Distance TPS Treatment Planning System XVI X-ray Volume Imager  xix  Acknowledgements First, much gratitude is owed to Dr. Cheryl Duzenli for her unwavering encouragement, support, and dedication. I cannot thank you and the committee members enough for believing in me and providing every tool I needed for my research. All your encouragement helped me get through the long time filled with frustration and joy. Dr. Vitali Moiseenko, I am truly thankful for you. Your passion in research inspired me. Thank you for sharing the knowledge and thoughts in radiation biophyiscs and always believing in me even when I doubted myself. I would also like to thank my committee members Dr. Anna Celler and Dr. Alex MacKay for their expertise and time. They encouraged me to think outside the boundary of radiation therapy and are always there if I need help. My committee, especially my supervisor Dr. Cheryl Duzenli, made great effort and spent lots of time helping me to improve my scientific writing. As a foreign student, I could not get my papers and dissertation done without their help. They also set a great example on how to be a supervisor! Thanks to Dr. Christina Aquino-Parsons for her help in clinical aspects of the gynecological IMRT/RapidArc study. Much gratitude is owed to Dr. Tony Popescu, Dr. Alanah Bergman and Dr. Tony Teke for sharing with me their knowledge about Monte Carlo simulation. I also would like to xx  acknowledge the team of Medical Physics department at BC Cancer Agency Vancouver Cancer Center for all their help during the past few years. Thanks to my friends for your friendship and support though the long journey. My loving family, you are the source of all of my accomplishments in life. Thank you for all the support for me to be where I am today.  xxi  Chapter 1  Introduction The goal of radiation therapy is to deliver a therapeutic dose of radiation to target tissues while minimizing the risks of normal tissue complications. Modern radiation therapy makes use of advanced treatment delivery technology to deliver highly conformal dose distributions. This results in very high dose gradients around target tissues and necessitates precise localization of anatomical structures and tumor volumes during both the treatment planning and treatment delivery processes. Imaging procedures are now commonly used during the treatment delivery process to ensure that target tissues and sensitive anatomical structures are correctly positioned in an effort to increase the safety and efficacy of treatment. However, some imaging techniques result in additional dose to the patient and increasing concern over radiation dose associated with imaging procedures has captured the attention of radiation medical professionals, patients and the general public. Detailed dosimetric information should be available to oncologists and patients so that potential risks and benefits associated with adopting imaging protocols can be appropriately balanced to optimize the safety and efficacy of radiation therapy treatment. This thesis rigorously explores organ specific doses associated with one  1  of the most frequently used image guidance techniques in radiation therapy today, kilo-voltage cone-beam computed tomography imaging. The specific case of patients being treated for gynaecological cancer using state of the art intensity modulated radiation therapy is examined in detail. Chapter 2 provides background on current technology used in radiation therapy including linear accelerators used to deliver intensity modulated radiation therapy (IMRT) as well as cone beam computed tomography (CBCT) imaging technology found on many modern medical linear accelerators. Chapter 2 also introduces radiation dosimetry and modeling the risk of carcinogenesis associated with radiation dose. Chapter 3 introduces target volume delineation, treatment planning and delivery for gynaecological IMRT. The radiation target for postoperative uterine or ovarian cancer patients is usually a combination of both the vagina and pelvic lymph nodes. The vagina is located between bladder and rectum and its position is affected by variations in bladder and rectum filling. This chapter describes the role of CBCT imaging as a quality assurance tool and discusses the potential benefits of incorporating CBCT imaging into gynaecological IMRT. Chapter 4 describes the methods used to accurately calculate doses associated with kilo-voltage CBCT imaging. Dose distributions were simulated using the Monte Carlo code BEAMnrc/DOSXYZnrc. Monte Carlo simulation is a general stochastic sampling technique and calls on the fundamental physics of radiation interactions to simulate particle transport and dose deposition in different media. This simulation method is widely considered to be the most accurate dose calculation. In the kilo-voltage range, doses 2  to bone and tissues are significantly different due to the photoelectric effect and currently there are no commercially available algorithms that provide accurate modeling of heterogeneous anatomical media for kilovoltage (kV) beams. A description of Monte Carlo theory is followed by a detailed description of the benchmarking and calibration process of the Monte Carlo system used for dose simulations on patient computed tomography image data sets. The effect of radiation on a biological system is not only dependent on dose but also on the type and energy of particles, because the spatial distribution of the ionizing events produced by different particles varies enormously. Dose-equivalent quantities have been defined in terms of a quality factor applied to the absorbed dose at the point of interest in order to take into account the differences in the effects of different types of radiation. The quality factor for photons and electrons is often assigned unity in the field of radiation protection to simplify practical application of radiation protection guidelines. Also, in situations of accidental exposure to radiation, uncertainties in received dose are often larger than effect of radiation quality for kilovoltage vs megavoltage (MV) photon beams. Extensive experimental work shows that kV x-rays have higher relative biological effect compared with MV x-rays when assessed by different endpoints. For example, if two samples of human lymphocytes are irradiated with same dose of kV x-rays and MV x-rays, more chromosome aberrations will be observed in the sample irradiated with kV x-rays. In radiation therapy, doses are very well known, thus it makes sense to account for differences in beam quality when comparing kV CBCT imaging doses with therapeutic doses of megavoltage 3  radiation. This is an important consideration when implementing a CBCT protocol into a radiation therapy treatment delivery protocol. Concepts in microdosimetry have been used successfully to quantify radiation quality. Experimental evidence shows that dose mean lineal energy, which is a stochastic quantity used to describe the energy deposition in a microscopic volume, can be related to quality factor. The International Commission on Radiation Units and Measurements (ICRU) have proposed formulae to describe this relationship and these formulae are used here. Starting with an introduction to microdosimetry, Chapter 5 describes the calculation of a lineal energy based quality factor using a C++ code connecting the results from DOSXYZnrc and another Monte Carlo code NOREC. In order to fully appreciate the impact of additional imaging dose on the health of radiation therapy patients, it is necessary to model radiation induced carcinogenesis. Secondary cancer risk may be estimated from Japanese atomic bomb survivor data and data from studies of patients receiving radiation therapy. Atomic bomb survivor data can be applied only in the dose range lower than 2.5Gy. While higher doses are typically used in radiation therapy, studies of radiation therapy patients are sparse due to the patient numbers required, lack of sufficient follow up data and the wide range of doses and techniques employed in treatment protocols from different institutions. In order to deal with the variation in doses within a particular organ, the concept of organ equivalent dose (OED) has been proposed by Schneider to apply to high dose regions found in radiation therapy studies. For gynaecological patients, critical structures like the rectum, bladder and bowel are in the CBCT imaging field. By design, the imaging field over4  laps with the high dose region of therapeutic megavoltage beam. Kidney, liver and breast are critical structures outside both the CBCT imaging field and the therapeutic MV fields and are subject to dose from scattered radiation. The OED method accounts for dose inhomogeneity typical of clinical dose distributions and can incorporate the effects of cell sterilization and repair. Any dose distribution in an organ is equivalent and corresponds to the same OED if it causes the same radiation-induced cancer incidence. Three different dose response models, the linear model, linear-exponential model and plateau model are used to calculate OEDs. The linear model effectively looks at mean organ dose and overestimates the OEDs in high dose regions. The linear exponential model takes the cell killing effect into account and the plateau model accounts for cellular repair in a fractionated radiotherapy setting. Chapter 6 introduces the OED concept and applies it in the context of gynaecological radiation therapy with daily cone beam CT scans. For organs near the target, the additional organ equivalent doses associated with CBCT imaging volume are compared with the inter-patient variation in treatment dose to these organs and to the current accepted limits for uncertainty in dose delivery. The impact of patient size on imaging dose is also examined. Imaging dose to organs in the low dose regions are compared with IMRT scatter doses and leakage dose resulting from x-rays coming through the shielding in the head of the linear accelerator. In summary, this thesis project describes a new Monte Carlo simulation system for modeling kV CBCT doses and reports patient doses for clinically relevant scanning protocols in the context of a radiation therapy treatment protocol. A gynaecological treatment protocol was studied to demonstrate 5  the potential benefits of using daily CBCT to monitor inter fractional organ motion and the subsequent dosimetric effects of organ volume changes and target displacement. Dose mean lineal energy was simulated by Monte Carlo techniques to derive a quality factor for the imaging beam such that imaging and treatment doses could be compared on the basis of biological effect. Due to the homogeneity in dose across organs, organ equivalent doses both within the treatment field and in peripheral regions were modeled. This dosimetric information should be helpful to radiation oncology professionals when balancing the potential risks and benefits associated with adopting CBCT imaging protocols for quality assurance of radiation treatments using modern IMRT technology.  6  Chapter 2  Background 2.1  Radiation Therapy  Radiotherapy is one of three major treatment modalities for cancer, the others being surgery and chemotherapy. Understanding mechanisms of cell death due to radiation is still an active research area. The general view supports the theory that most cell damage occurs when radiation induces single or double strand breaks or damage to the base molecules in deoxyribonucleic acid (DNA). Photon beams transfer energy to tissue though particle interactions including Rayleigh scattering, the photoelectric effect, the Compton effect, pair production and triplet production. Details of these interactions can be found in the textbook The Physics of Radiology by Johns and Cunningham[1] or other similar works. Energetic particles damage the DNA molecules by direct and indirect actions. Details of these interactions can be found in the textbook Radiobiology for the Radiologist by Hall and Giaccia[2]. Radiation delivery can be achieved in different ways: 1) External Beam Radiation. This technique projects a beam of radiation at the patient from an external source. The medical linear accelerator, or linac is the most commonly used radiotherapy machine; 2) Encapsulated Sources. This technique 7  is often referred to as brachytherapy, which involves the precise placement of radiation sources directly at the site of the cancerous tumor; 3) Unencapsulated sources. This type of radiation delivery involves the ingestion or injection of a soluble radioactive material. A typical external beam radiation therapy process includes simulation, treatment planning and treatment delivery. In simulation, the patient is imaged in the anticipated treatment position using a computed tomography (CT) scanner. Immobilization devices may be used to achieve the reproducibility and rigidity of patient setup. Two purposes for simulation are CT image acquisition for treatment planning and establishment of patient setup that can be reproduced during treatment. After the simulation, tumor and surrounding critical organs will be contoured on the CT image set. Linac parameters (mechanical settings, beam energy, shape and intensity) will then be determined to deliver the prescribed dose to tumor and limit dose to normal structures. That is the treatment planning process. In treatment delivery, the patient will be set up in the planned position on the treatment couch. The medical linear accelerator will deliver radiation to the patient according to the parameters achieved in the treatment planning process.  2.1.1  CT Simulation  In clinical practice, CT simulation of the patient is performed for the purpose of radiotherapy treatment planning. A CT simulator consists of a diagnostic quality CT scanner, laser patient positioning/marking system and virtual simulation/3D treatment planning software. The fourth-generation CT scanner is generally in use today. This device 8  uses a stationary ring of detectors positioned around the patient. Only the x-ray source rotates with a wide fan beam geometry, while the detectors are stationary. A sketch of the beam geometry is shown in Figure 2.1. The detector measures the decrease in x-ray incident particle fluence along a series of linear paths through the patient. The incident particle fluence reduction is a function of x-ray energy, path length, and material linear attenuation coefficient. Fluence from different angles is collected as the xray source rotates around the patient. A specialized algorithm is then used to reconstruct the distribution of x-ray attenuation in the volume being imaged. Details of CT principles, design, image reconstruction can be found in the work of Hsieh[3]. It is worth mentioning that the energy used in diagnostic CT is between 40-150kVp. The dominant physical processes responsible for x-ray attenuation are photoelectric absorption and Compton scattering. The photoelectric effect is the dominant attenuation mechanism at kV xray energies; at around 50kV, photoelectric effect and Compton effect have equal attenuation effect, after that Compton scatter gradually predominates. Differences in attenuation coefficient correspond to differences in material density and atomic number. Because the photoelectric effect cross section is proportional to atomic number Z 3 , whereas Compton scatter is proportional only to Z, low energy (kilo-voltage) x-rays are more sensitive to differences in composition than high energy (Mega-voltage) x-rays. In a CT simulation process, the patient is scanned in the treatment position with radio-opaque markers placed on skin and aligned to the positioning laser system. The same positioning laser system is also installed with the medical linear accelerator. Thus the exact geometry and position of patient 9  Detector  Source  Figure 2.1: Diagram of fan beam geometry with a complete rotation of the x-ray tube around a stationary ring of x-ray detectors. can be reproduced in the treatment room. These radio-opaque markers are visible on the CT scans and provide a reference for tumor and other organ positions in the treatment planning process.  2.1.2  Linac  The medical linear accelerator generates X-ray radiation via the acceleration of electrons. A cutaway view of a Varian linac is shown in Figure 2.2. Low energy electrons are generated by the electron gun and introduced into a waveguide. The waveguide is a vacuum chamber that has an oscillating electromagnetic field and it accelerates electrons close to the speed of light. 10  The accelerated electron beam exits the waveguide and is forced into a 90◦ or 270◦ turn by a bending magnet and then collides with the target. The target composition is usually tungsten, where bremsstrahlung radiation is generated through the rapid deceleration of the electrons. X-rays are projected towards the patient through a primary collimator, a flattening filter (often used to flatten the profile of the forward peaked beam), an ion chamber (used to monitor the dose), secondary tungsten collimating jaws (to create rectangular shapes) and finally a multileaf collimator (MLC, to shape the beam statically or dynamically). For a more in depth description of medical linear accelerators, the reader is directed to the work of Karzmark[4].  2.1.3  Multi-Leaf Collimator (MLC)  A Multi-Leaf Collimator is a device used to fine tune the shape of the photon beams. A typical MLC consists of two banks of leaves with more than 40 pairs of leaves. Each leaf can independently move in and out of the radiation field, thus providing various shapes of the field. A Varian Millennium 120 MLC model is shown in Figure 2.3. It has 60 pairs of leaves made of high density alloy. The central 40 leaf pairs are 0.5cm wide and the remaining leaves (10 pairs on each side) are 1cm wide (widths projected of the isocenter of the linac). In treatment planning, MLC can be shaped to match the contour of tumor in a Beams Eye View thus protecting normal tissues around the tumor. The position of the MLC can also be controlled to create nonuniform beam intensity and deliver a more conformal dose to the tumor target.  11  Figure 2.2: Cutaway view of a modern linear accelerator with multileaf collimators for beam shaping. (1) electron gun, (2) standing waveguide, (3) bending magnet, (4) target, (5) carousel of scattering foils and flattering filters, (6) ion chamber, (7) asymmetric collimating jaws, (8) multileaf collimator. Image is from www.varian.com/us/oncology/radiation oncology/clinac/  2.1.4  Intensity Modulated Radiation Therapy - IMRT  Intensity modulated radiation therapy is a radiation treatment planning and delivery technique. It can generate highly conformal x-ray radiation doses to complex shaped tumors while sparing nearby sensitive tissues[5]. The planning process for IMRT is called inverse planning, as opposed to forward planning used for less complex treatment delivery. For example, conformal radiation therapy, in which the beam is shaped, or conformed, 12  Figure 2.3: Varian Millennium 120 MLC model. Image is from Varian website www.varian.com/us/oncology/radiation oncology/clinac /millennium mlc.html to match the shape of the tumor, is planned with forward planning. In forward planning, the user decides number of beams to use, beam angles that beams will be delivered from, whether wedges will be used to modify the uniform beam intensity or MLC be used to shape the radiation beam and the relative weight of each beam. Once the treatment planner has made an initial plan, the treatment planning system calculates the dose distribution to the patient. The disadvantage of forward planning is that the plan depends on the planner’s subjective decision and is not as conformal as IMRT plans in complex tumor/critical organ geometry. In IMRT the user sets desired dose distribution constraints, such as the dose to the target and dose-volume constraints for surrounding critical structures. In an inverse  13  planning calculation, the beam is divided into beamlets (the beamlet size is the minimum leaf width, often 0.5cm). An algorithm is then used to search for the best intensity of each beamlet to achieve the desired distribution of radiation dose. However, this problem can not be solved analytically since negative fluence pixels are not physically possible. Instead, optimization methods are employed. The goal is to minimize an objective or cost function which describes the difference between the desired dose distribution and the optimized dose distribution. IMRT delivery can be achieved using two methods: step and shoot technique or sliding window technique. In the step and shoot technique, subfields or segments are delivered sequentially. Multileaf collimator leaves are (automatically) positioned while the radiation beam is switched off. After irradiation of one segment the leaves move to the correct positions for the next segment and so on until the total intensity modulated field has been delivered. In the sliding window technique, the beam is always on as the collimator leaves move according to predetermined trajectories designed to give the desired modulation. At the BC Cancer Agency, the sliding window technique is the default for IMRT delivery. IMRT is generally the term used to describe treatment delivery at a series of distinct static beam angles.  2.1.5  RapidArc  Volumetric modulated arc therapy is a technique whereby IMRT is delivered during one or more rotations of the linear accelerator gantry. In the treatment process, rotation speed of the gantry, shape of the treatment R aperture and delivery dose rate can change simultaneously. RapidArc⃝ is  14  a commercial implementation of volumetric modulated arc therapy by Varian. Essentially, RapidArc is also an IMRT technique. However, IMRT and RapidArc are often mentioned together. In this thesis, traditional IMRT is used to refer to fixed gantry angle IMRT. RapidArc can significantly shorten treatment time, two to eight times faster than conventional IMRT, while potentially increasing precision[6, 7]. The key idea of RapidArc planning is progressive optimization, which means that gantry and MLC position sampling is progressively increased throughout the optimization process. Details of the optimization algorithm can be found in the paper by Karl Otto, who developed the prototype RapidArc algorithm[8]. At BCCA, for a typical fraction of gynaecological radiation therapy, treatment time is about 2 minutes using RapidArc while a traditional sliding window IMRT delivery usually requires about 10 minutes. RapidArc is more efficient because the beam is always on as the gantry rotates about the patient and the MLC aperture on average is larger than that of traditional IMRT. In traditional IMRT, the linac gantry delivers at one angle, stops the beam, then moves to the next angle. RapidArc has several advantages over fixed gantry IMRT. First, short treatment time means improved patient comfort and the reduction of the risk of intra-fractional organ motion. Second, less beam on time means less leakage doses from the linac head to other parts of the patient’s body. This potentially reduces the chance of radiation induced secondary malignancy. BC Cancer Agency Vancouver Cancer Center began using IMRT for gynaecological treatment in 2009 and as of 2012 has replaced fixed gantry IMRT with RapidArc to treat this group of patients. 15  2.2 2.2.1  Image Guidance in Radiation Therapy In-room CT Systems  As IMRT and RapidArc are now frequently used in radiation therapy, more accurate and precise target volume localization is required. Image guided radiation therapy (IGRT) is a process using various imaging technologies to adjust for target motion or positional uncertainty and potentially to adapt treatment to tumor response. CT in the treatment room has been integrated with linac-based IMRT dose delivery since the late 1990s[9, 10]. There are three different techniques in use today: kV CT-on-rails, on board kV CBCT and MV CBCT. The CT-on-rails imaging systems consist of a conventional fan beam CT scanner installed in the treatment room. During the image acquisition process, the scanner moves on rails to a position for acquiring helical CT scans of the patient on the treatment couch, shown on Figure 2.4. The image acquisition is slice by slice and the slices are reassembled to form a 3D image. In the kV CBCT imaging system, the x-ray source is mounted on the linac gantry on a retractable arm at 90 degrees to the treatment source and the image is taken with the patient in treatment position on the couch. Commercialized products include Varian’s On-Board Imager (OBI) and Elekta Synergy X-ray Volume Imager (XVI). Figure 2.5 shows Varian’s On-Board Imager. Yellow lines indicate the configuration of the cone beam and unlike the conventional CT scanner, the imager rotates to a maximum of 360o . On-board x-ray CBCT imaging components are described in more detail in sections 2.2.3 to 2.2.5. 16  Linac  CT Scanner  Couch  Rails  Figure 2.4: This is the CTVision system from Siemens. It consists of a linear accelerator and a modified diagnostic CT scanner (CT-on-rails). Image comes from Siemens website: www.medical.siemens.com/webapp/wcs /stores/servlet/P roductDisplay?storeId = 10001&langId= −1catalogId = −1&productId = 4144428&catT ree = 100010, 1008643,12757,1029718, 1031166 An MV CBCT imaging system does not require additional hardware. It uses the treatment beam as the x-ray source and electronic portal imaging device (EPID) positioned distal to the patient as the detector to produce volumetric images as the gantry rotates. Figure 2.6 shows a Siemens linac and the electronic portal imaging device. EPID devices have been in clinical use for more than 15 years. Originally designed for planar x-ray imaging, they have since been adapted for cone beam CT imaging. CT-on-rails imaging systems provide the best image quality because the  17  Figure 2.5: This is a Varian Trilogy Linac with Cone Beam CT imager. Image comes from Varian website: www.varian.com/us/oncology/ radiation oncology/trilogy/treatmenttechniques.html. conventional CT scanner is optimized to produce diagnostic quality images. However, CT-on-rails is more expensive and takes up more space in the treatment room. Thus, CT-on-rails imaging systems are not commonly used in clinical practice. As stated in section 2.1.1, MV is not the optimal energy for CT scan. Although it is the most cost efficient, MV CBCT image quality is compromised. The MV-CBCT imaging procedure has been well integrated into clinical workflow for the patient alignment and image guidance processes. The kV CBCT image quality is better than MV CBCT and is easy to integrate into a linac, thus the usage of kV CBCT image systems has dramatically increased since introduction of this technology. This thesis focuses on the imaging dose associated with kV CBCT only.  18  Linac Head  Treatment Couch  Electronic Portal Imaging Device (EPID)  Figure 2.6: This is a Siemens linear accelerator with electronic portal imaging device. Image comes from website: www.medwrench.com/?equipment.view/equipmentN o/1232/Siemens/ ON COR − Expression/.  2.2.2  Clinical Implementation of kV CBCT  The idea to use a kilo-voltage x-ray source and large-area flat-panel detector on a medical linear accelerator was proposed by Jaffray et al. for cone beam volumetric CT[11–13]. CBCT imaging procedures can be well integrated to the clinical workflow for patient alignment and IGRT , since it can be installed on most linacs, provides good soft tissue contrast and delivers a low dose to patients. The utilization of CBCT has been increasing dramatically since 2000. According to a survey of image guided radiation therapy use in  19  the United States, the percentage of respondents who used volumetric image guided techniques (CBCT and MVCT) in United States has increased from barely zero in 2000 to nearly 70% of in 2009[14]. This survey also shows that CBCT is extensively used in the following treatment sites: prostate, head and neck, lung, CNS(central nervous system), gastrointestinal tract, and gynaecological sites. Prostate tumors can be visualized on kV CBCT without implanting fiducial markers which are commonly used in MV imaging due to poor soft tissue contrast[15]. Stereotactic body radiotherapy (SBRT) used to treat lung cancer patients can be now performed without the need for body frames and fiducial markers, by using kV CBCT for IGRT[16, 17]. In general, CBCT is used as an image guided tool to reduce setup errors and increase precision of radiotherapy.  2.2.3  On Board Imager (OBI)  kV CBCT in this study is performed using the Varian On-Board Imager(OBI). The OBI system can be added as an option to new or currently in use (Varian 21EX or above series) Varian linacs. [18] As shown in Figure 2.5, the OBI system consists of two robotic arms (EXaCTTM arms). The kV source unit is enclosed in a protective plastic cover. The kV source arm carries an X-ray tube and an x-ray collimator. The x-ray tube target is made of Wolfram and Rhenium and two focal spots are 0.4mm and 0.8mm in diameter. The collimator provides the possibility of symmetrical and asymmetrical field setups. The imager support arm carries an amorphous silicon detector with an active rectangular imaging area of 397 mm x 298 mm. Detail principle of flat panel detector can be found in literature and Varian product 20  brochure[19–21]. A bowtie filter made of aluminum is mounted in front of the tube to filter the X-ray beam to improve the quality of the CBCT projection image. A bowtie filter can improve uniformity, CT number accuracy, and contrastto-noise[22]. The bowtie can also reduce patient dose, especially skin dose, because more low energy X-rays are absorbed when passing through the bowtie filter. Figure 2.7 shows the shapes of the two bowties used on the Varian OBI.  Body Bowtie  Head Bowtie  Figure 2.7: Photos of bowtie filters. Left: body bowtie (half fan bowtie) Right: head bowtie (full fan bowtie).  21  Blades  Bowtie Head fan/Body fan  Patient  Detector  Figure 2.8: CBCT acquisition modes with head fan or body fan. Left: The head fan bowtie is mounted, the detector is centered with x-ray axis and the rotation is 200 degrees. Right: Body fan bowtie is mounted, the detector is shifted and the gantry rotation is 360 degrees.  2.2.4  Scan Modes  The On Board Imager can operate in three modes: radiographic, fluoroscopic and CBCT. Fluoroscopic and projection imaging for CBCT are acquired at 15 frames per second. Figure 2.8 shows two CBCT acquisition modes with the two bowties. Under the standard dose head mode, the head fan is mounted. X-ray source of 100kVp rotates 200 degrees around the patient with 360 projections. In theory, half rotation of 180 degrees plus two cone 22  angles are sufficient for image reconstruction[23]. Thus half rotation mode is designed for the purpose of reducing patient dose. Table 2.1.  summarizes acquisition techniques.  For example, under  pelvis mode, in which gynaecological patients are scanned, the body fan is mounted. The x-ray source rotates 360 degrees around the patient with the detector shifted. Because each beam only passes through part of the patient, a full rotation is needed for image reconstruction. A total number of 655 projections are taken during this process using 125kVp beam. The parameters 80 mA and 13ms are the current and beam on time for one projection.  2.2.5  Image Quality  The image quality of CBCT is inferior to fan beam CT used in treatment planning process. The large cone angle causes scatter to contribute undesired signals to the reconstructed image. The primary to scatter ratio at regions of interest (ROI) varies with source to detector distance (SDD) and size of the object being imaged[24]. Scatter to primary ratio of a ROI at the center of the detector for a 32cm diameter cylindrical phantom at 155cm SDD for a 120 kVp beam is about 1.5 for a kV CBCT image. In contrast, for CT body scans on a 4th-generation machine, scatter to primary ratio is approximately 0.05[25]. The large amount of scatter has an adverse effect on image quality. Soft tissue contrast is reduced because of the high number of scatter photons. As a result, kV-CBCT can only resolve objects 7mm in diameter with 1% contrast[26], while the planning CT can resolve objects 3-4.5mm in diameter with 1% contrast [27]. Also, due to inaccurate beam 23  Mode Name  Acquisition Angle (deg)  Fan Type  Techniques mAs (for total acquisition)  Standard dose head  200  head  100kV 20mA 20ms  145  Low dose head  200  head  100kV 10mA 20ms  72  High quality head  200  head  100kV 80mA 25ms  720  Pelvis  360  body  125kV 80mA 13ms  680  Pelvis spot light  200  head  125kV 80mA 25ms  720  Low dose thorax  360  body  110kV 20mA 20ms  262  Table 2.1: Scan modes and detailed setups for CBCT hardening corrections, cupping or capping artifacts, which result in lower or higher Hounsfield unit values towards the center of the phantom, are present in 3D CBCT reconstructions. In addition, relatively long scanning times up to 1 minute cause image blurring in areas with internal organ motion. In such situations, 4D CBCT have been developed to provide respiratory phase resolved volumetric imaging in image guided radiation therapy[28]. CT number accuracy has been evaluated to investigate the feasibility 24  of basing treatment plans on CBCT images instead of the higher quality planning CT images. One study showed that dosimetric accuracy of CBCT based dose calculation was acceptable for prostate dosimetric checks when motion artifacts were absent[29]. Yoo et al. [30] studied the dosimetric feasibility of CBCT-based treatment planning. The discrepancies in dose between CT-based and CBCT-based plans for both forward planned and IMRT cases were clinically acceptable. L´ etourneau et al.[31] has shown that cone beam CT can be used for Online planning for radiotherapy of spinal metastases and the dosimetric accuracy satisfies accepted RT standards. Thus, CBCT-based treatment plans are feasible for some clinical sites.  2.3  Radiation Induced Carcinogenesis  It has been known for a long time that exposure to ionizing radiation can increase the risk of developing cancer over a lifetime. Leukaemia and skin cancer were found in physicists working with x-rays or accelerators[2]. Famous Nobel Laureate Marie Curie was thought to have died of leukemia because of the radiation exposure during her experiments[2]. Radium dial painters and uranium miners were found to develop bone tumors and lung cancers respectively in the early 1900s[2]. The dosimetry in these cases is uncertain, thus the relation between exposure and cancer incidence is difficult to derive. These early problems should never happen again. Nowadays, since harmful effects of radiation exposure are widely understood, radiation protection guidelines are in place to protect workers dealing with radioactive materials or equipment. More accurate information about radiation effects  25  on humans mainly comes from the atomic-bomb survivors and patients undergoing radiation therapy. In order to study the exposure disease relationship, two essential components are required: a measure of exposure and a measure of disease occurrence. The following variables are widely used in quantifying cancer risk. Incidence Rate λ=  d nL  (2.1)  where n is the number of individuals who are disease free, d is the number of new diagnoses during time interval L. Excess Absolute Risk (EAR) EAR(t) = λE (t) − λU (t)  (2.2)  in which λE (t) and λU (t) are incidence rates of the exposed and unexposed groups. Relative Risk (RR) RR(t) =  λE (t) λU (t)  (2.3)  Excess relative risk (ERR) ERR(t) = RR(t) − 1  2.3.1  (2.4)  Study of Atomic-bomb Survivors  Incidence and cancer mortality for solid cancers and leukemia among atomic bomb survivors in Japan have been studied intensively[32, 33]. A series of  26  periodic general reports on cancer mortality in the Life Span Study (LSS) cohort of A-bomb survivors have been published by the Radiation Effects Research Foundation(RERF)[34–36]. 86,572 survivors are available for dose estimates and follow up in the Japanese family registry. 69,308 survivors were exposed to radiation less than 0.1Sv and 679 survivors were exposed to radiation more than 2 Sv with the remainder receiving between 0.1 Sv and 2 Sv . The time between irradiation and the appearance of malignancy is called the latency period. Leukemia has the shortest latency period, with a peak of appearance by 5 to 7 years and mostly within the first 15 years post exposure. Leukemia is the only hematopoietic cancer observed in excess risk in the Japanese life span study data. For solid cancers, the latent period is from 10 to 60 years or more, thus the excess risk is a lifelong elevation of cancer risk. Most types of solid cancers have been observed in excess risk, including cancers of the oesophagus, stomach, colon, liver, lung, non-melanoma skin, breast, bladder, brain, thyroid and central nervous system. Significant excess risk of solid cancer was found among those survivors who were exposed in utero which was comparable to those exposed in early childhood. The dose response for most cancer sites in the life span study can be described by a linear dose response, however, dose response for leukemia is linear-quadratic. There was no evidence of dose threshold in radiation induced solid cancers. In contrast to the solid cancer data, when using a linear quadratic model, the estimated dose threshold for leukemia incidence was 0.11Sv (95% CI, 0.003, 0.27, p=0.04) [37].  27  2.3.2  Study of Radiation Therapy Patients  Cancer risk for patients undergoing radiation therapy is different from that for the population of atomic bomb survivors. The main reason is that patients undergoing radiation therapy are often at high risk of a secondary cancer due to their genes or lifestyle, which are dominant factors compared with radiation risk. With studies of large groups of radiotherapy patients, statistically significant but very small increased risk of secondary malignancies is observed. Credible studies need to meet the following requirements: 1. A sufficiently large group of patients. 2. An appropriate comparison group, i.e., patients with the same cancer treated with other modalities rather than radiation therapy. 3. A sufficiently long follow up. 4. Well characterized dosimetry. Qualified studies are only available for a few cancer sites such as prostate, cervix and Hodgkin’s disease. A number of studies of Hodgkin’s patients have been reported [38–40]. These patients were treated at young ages with curative intent and several organs at risk were included in high dose region. In addition, radiotherapy for Hodgkin’s patients was relatively successful, and the treatment techniques did not vary significantly with time or among institutions. Thus, long follow up times can provide incidence rates for various organs with good precision. The relative risk (RR) of solid tumors increased greatly with younger age at the first treatment of Hodgkin’s disease, especially for breast cancer. The younger the patients were when receiving  28  the treatment, the higher the relative risks were for secondary malignancy. The greatly increased risk of developing solid tumors in patients who were young (< 20 years of age) at first treatment seemed to decrease as these patients grow older(> 40 years of age). Studies of risk of secondary malignancy have been performed for patients receiving radiation therapy for malignancy of the uterine cervix [41–43]. In both radiotherapy and surgical groups, risks for HPV(human papillomavirus)related cancers (of the pharynx, genital sites, and rectum/anus) and smokingrelated cancers (of the pharynx, trachea/bronchus/lung, pancreas, and urinary bladder) were elevated to a statistically significant extent. These risk factors per se are associated with cervical cancers. For patients treated with radiation therapy, secondary cancer sites were classified into three groups according to average dose ranges. Heavily irradiated sites (10-60Gy) were the small intestine, rectum, bladder and other female genital sites . Moderately irradiated sites (1-3 Gy) were the stomach, liver, pancreas, gall bladder and kidneys. Lightly irradiated sites (< 1 Gy) were the lip, tongue, salivary glands, pharynx, esophagus, larynx, lung, breast, eye, brain and central nervous system and thyroid. The radiotherapy group showed increased risk for all secondary cancers and cancers at heavily irradiated sites (colon, rectum/anus, urinary bladder, ovary, and genital sites). Out of 104760 patients, 4703 patients have a secondary cancer at heavily irradiated sites, making the standard standardized incidence ratio 1.50 for this patient group. If comparing secondary cancer risk for patients treated with/withoug radiation therapy, the standardized incidence ratio is 1.59 for patients treated with radiation therapy while it is 0.97 for patients treated without radiation 29  therapy. Statistically significant deficits of breast cancer and melanoma were also observed in both treatment groups. Hysterectomy and ovarian ablation through irradiation may alter hormonal exposure of the breast tissue and thereby reduce subsequent breast cancer risk. In conclusion, excess cancers were associated with radiotherapy, as opposed to surgery. The risks were highest among those who received radiotherapy at relatively young ages.  2.3.3  Other Studies  At low doses of less than 0.2Sv per year, epidemiological studies are not able to detect and quantify statistically significant radiation effects. These studies include occupational, natural background, environmental exposures and medical imaging procedures. The International Agency for Research on Cancer (IARC) carried out a study of 95,000 nuclear industry workers in the United States, the United Kingdom and Canada [44]. For overall solid cancers, there was no evidence of an increased risk associated with radiation in any of the three countries. For leukemia, there was a small statistically significant excess in U.K. workers and a deficit in U.S. workers. The excess disappears when the data from three countries are pooled. Yoshinaga et al.[45] reviewed epidemiologic data on cancer risks from eight cohorts of over 270 000 radiologists and technologists in various countries. Increased mortality due to leukemia was found among early workers employed before 1950, when radiation exposures were high. International Commission on Radiological Protection(ICRP) recommended dose limits are designed to protect workers from these effects. A Nuclear Energy Worker is 30  a person who is required, in the course of the person’s business or occupation in connection with a nuclear substance or nuclear facility, to perform duties in such circumstances that there is a reasonable probability that the person may receive a dose of radiation that is greater than the prescribed limit for the general public. The effective dose limit for Nuclear Energy Workers, is 100 mSv for 5 years for whole-body exposure. Radiation users define individuals who need not be legally classified as Nuclear Energy Workers but who nevertheless require a range of protective measures because they are in regular proximity to radiation sources. For radiation users and members of the public, the corresponding limit is 1 mSv per year. There is no clear evidence of an increased cancer risk in medical radiation workers exposed to current levels of radiation dose. However, given a relatively short period of time for which the most recent workers have been followed up, it is important to continue to monitor the health status of medical radiation workers.  2.3.4  Mechanism of Carcinogenesis  Mechanism of carcinogenesis is largely unknown. It takes multiple steps and a long latent period from the initiating events in chromosomes to tumor development[46, 47]. In order for cells to start dividing uncontrollably, genes that regulate cell growth must be damaged. Proto-oncogenes are genes that promote cell growth and mitosis. Tumor suppressor genes are genes that code for anti-proliferation signals and proteins that suppress mitosis and cell growth. Activation of proto-oncogenes and loss of tumor suppressor genes are considered to contribute to clonal development from an initiated 31  cell. At some point, an evolving clonal population becomes increasingly committed to malignant development. Subsequently the neoplastic cells gain the capacity for invasion of surrounding normal tissue and may spread to distant sites. Overall, only a small fraction of initiating events result in overt malignancy and the whole process may take many years[47]. However, the first step is considered to be initiating events in chromosomes (such as aberrations) or in DNA. When cells are irradiated with xrays, double-strand breaks are produced in DNA. These broken ends may behave in the following ways: 1) The breaks may rejoin in their original configuration. 2) The breaks may fail to rejoin and give rise to an aberration, which is scored as a deletion at the next mitosis. 3) Broken ends rejoin with other broken ends incorrectly to give rise to chromosomes that will be distorted at the following mitosis. For example, a dicentric or a ring can form during this process[2]. Chromosomal aberrations found in lymphocytes circulating in the bloodstream have played an important role as a biomarker of radiation exposure. Within a few days to a few months after total body irradiation, the frequency of unstable chromosomal aberrations, such as dicentrics or centric ring, observed in human lymphocytes reflects the dose received. Advanced methods, such as fluorescence in situ hybridization, which allow us to score stable chromosomal aberrations, such as symmetric translocations, paved the way for estimating doses years after radiation exposure[2]. The frequency of chromosomal aberrations is a linear-quadratic function of dose. The rings and dicentrics result from two chromosome breaks: the linear component is a consequence of the two breaks from a single charged 32  particle while the quadratic component comes from breaks by different particles. In radiation biology, lymphocyte dicentrics and rings are important endpoints representing biological effectiveness of different radiations.  2.4 2.4.1  Risk Assessment Models and Methods Low Dose Risk Assessment  The BEIR(Biologic Effects of Ionizing Radiation) VII committee conducted a comprehensive review of all relevant epidemiologic data related to the risk from exposure to low dose, low LET radiation [47]. Direct estimates of excess absolute risk (EAR) or excess relative risk (ERR) are difficult because small increases in risk associated with low levels of exposure are difficult to detect in the presence of background risk. It is even more difficult, or only feasible in theory, for specific subpopulations defined by stratification on variables such as sex, age and exposure profile. Large groups of individuals are needed to be followed for long periods of time to provide sufficiently precise estimates. Thus, direct estimates of risk are generally not possible for stratified subpopulations. Model based estimation is a feasible alternative to direct estimation.There are two kinds of models: biologically based risk models and empirically based risk models. Because mechanisms of radiation carcinogenesis are not fully understood, data required for biologically based models are not available. Biologically based risk models are not used widely. Empirically based models are subject to statistical limitations imposed by the quantity and quality of data available for model fitting and their 33  validity depends on the appropriateness of the model. Thus, model choice is important. The BEIR VII committee chose five different empirically based models for all solid cancers, site-specific cancers other than breast and thyroid, breast cancer, thyroid cancer, and leukemia. Lifetime attributable risk (LAR) is then calculated and tabulated in the BEIR VII report. These tables list LAR in the format of number of cases per 100,000 persons exposed to a single dose of 0.1Gy for different sites and different ages at exposure.  2.4.2  High Dose Risk Assessment: Concept of Organ Equivalent Dose (OED)  Schneider et al.[48] proposed the concept of organ equivalent dose (OED) for the purpose of estimating organ specific radiation induced cancer incidence rates. For radiation dose greater than 2-4 Gy, the radiation induced cancer incidence rate is not a linear function of the dose due to cell killing effects. The OED takes the cell sterilization effect and cell repair as well as dose inhomogeneity into account. Any dose distribution in an organ is equivalent and corresponds to the same OED if it causes the same radiation induced cancer incidence. To obtain organ specific data about radiation induced cancer, a large population of patients undergoing RT with high dose is needed. Schneider chose to analyze patients with Hodgkin’s disease for the reasons discussed in section 2.3.2. Three dose response models were proposed. The linear dose response model is the conventional way to calculate average organ dose and applies to the low dose region only. For doses less than approximately 2 Gy, organ 34  equivalent doses may be determined using the linear dose response model as follows:  OEDT =  1 ∑ V (Di )Di VT i  (2.5)  in which V (Di ) is the volume that corresponds to the dose Di and the summation is done for all voxels of organ T with volume VT . Cell sterilization effects are present for doses greater than approximately 2-4 Gy[36], so an exponential term is added in the formula to account for the decrease of cell population subjected to potential secondary malignancy. OED for the linear-exponential dose response model is calculated by the following formula  OEDT =  1 ∑ V (Di )Di e−αDi VT i  (2.6)  Finally, OED for the plateau dose response model, which may be more appropriate in a fractionated radiotherapy setting, accounting for effects such as repair , is obtained by  OEDT =  1 ∑ V (Di )(1 − e−δDi )/δ VT i  (2.7)  in which α and δ are 0.044Gy −1 and 0.139Gy −1 respectively. These parameters were derived from data of Hodgkin’s patients and atomic bomb survivors. The OEDs calculated with these models are close to each other at doses up to approximately 4Gy in a fractional protocol and differ widely at high dose depending on the assumption made concerning cell killing.  35  Chapter 3  CBCT in Gynaecological Cancer IMRT Treatment This chapter aims to introduce the methodology used to treat gynaecological cancer using IMRT and to describe the role of CBCT imaging as a quality assurance tool.  3.1  Introduction to Gynaecological Cancer and Treatment with Radiation Therapy  3.1.1  Epidemiology and Treatment  Gynecological cancer is a group of cancers that affect the tissue and organs of the female reproductive system. Gynaecologic cancer is the fourth most prevalent cancer type for women in Canada. According to the Canadian Cancer Registry 8,193 new cases of endometrial, ovarian, and cervical cancer occurred in 2007[49]. The average age for patients with cervical cancer is in the mid- to late 40s and the average age for patients with endometrial and ovarian cancers is about 60. Treatment for gynaecologic cancer depends on the type of cancer and the 36  stage. Common methods of treating gynaecological cancers include surgery, chemotherapy, radiation therapy, and hormonal therapy.  3.1.2  Radiation Therapy Technique: From Pelvic Radiation Therapy to IMRT  Radiation therapy plays a critical role in the management of gynaecological malignancies. From the discovery of radium, RT has been used to treat locally advanced cervical cancer[50]. Over the past several decades, radiation therapy has been used in the treatment of a variety of gynaecological malignancies, including cervical, endometrial, vulvar and vaginal cancers. Many clinical trials have proved that irradiation of the pelvis in the postoperative setting reduces the risk of local recurrence in patients who have disease features associated with a moderate to high risk of recurrence. Standard pelvic radiation therapy is delivered using a 4-field technique, which results in a uniform dose distribution to all organs within the treatment portal. This approach causes irradiation to considerable volumes of normal tissues, causing a variety of treatment-related toxicities. Toxicities to the gastrointestinal tract may result in diarrhea, malabsorption of vitamins, lactose and bile acids. Radiation to a large portion of the bladder may cause genitourinary problems, such as dysuria, urgency or hematuria[51–53]. A QUANTEC (Quantitative Analyses of Normal Tissue Effects in the Clinic) paper reviews radiation dose-volume Effects of the urinary bladder[54]. Because there is a considerable portion of red bone marrow in the pelvic bones, hematologic toxicity is a concern, especially for patients receiving chemotherapy during or after radiotherapy[55]. When extended fields (including para-aortic 37  lymph node) are used for more advanced cancer, larger volumes of normal tissues are involved and toxicities are more prevalent. Concerns regarding normal tissue toxicities limit the radiation dose to the clinical target volume (CTV). Patients with lymph node involvement may benefit from increased doses in the effort of improve tumor control. Patients with high risk of recurrence may also benefit from higher doses. However, only modest dose increases are allowed with the constraints of normal tissue doses. IMRT has been used to treat gynaecological malignancies starting from around 2000. Both 7 field and RapidArc  R ⃝  (RapidArc, Varian Medical  Systems, Palo Alto, CA) pelvic IMRT is currently being evaluated in a prospective clinical trial for patients with gynaecological malignancies at the Vancouver, Cancer Center, BCCA. Eligible patients include women who are to receive post operative adjuvant pelvic radiotherapy for uterine or ovarian malignancies. These patients are treated with 6MV IMRT or RapidArc using a kV CBCT imaging protocol for quality assurance of clinical target volume (CTV) coverage. The IMRT or RapidArc prescribed dose is 45Gy at 1.8Gy per fraction to the planning target volume. For IMRT, seven fields are optimized using the Eclipse treatment planning system (Varian Medical Systems) and delivered by the sliding window technique. Accurate localization of the apex of the vagina is accomplished with placement of interstitial marker seeds introduced just prior to CT simulation. A standard pelvic CT simulation is performed with the slice thickness of 2 mm. Oncologists contour the clinical target volumes (CTV) and critical structures. The clinical target volume includes tumor (when present) and 38  also volumes with suspected (subclinical) tumor considered to need treatment. The CTV for gynaecological cancers has two components. The upper part includes the pelvic nodes by including the soft tissue around the pelvic vessels (CTV vessels). The lower part includes part of the vagina and part of the parametrium (CTV vagina). The CTV vagina is contoured according to the RTOG guidelines [56]. The CTV vagina includes the parametrium with 5 mm of rectal/ bladder fat and laterally to the obturator muscle. The superior extent is 1 cm superior to the marker seeds plus any soft tissue abnormality. The inferior extent is 3 cm inferior to vaginal marker or 1 cm above the inferior obturator foramen, whichever is most inferior. Anteriorly and posteriorly the bladder and rectum are to be excluded except if vaginal CTV would have <1.5 cm AP dimension in midline, in which case, the posterior bladder and anterior rectum will be included in the vaginal CTV. Contours of CTV vagina, CTV vessels, bladder and rectum are shown in Figure 3.1. A 7mm isotropic margin is added to expand clinical target volumes into planning target volumes (PTVs) (shown in Fig.3.2). The margin is used to account for possible uncertainty in beam alignment, patient positioning, organ motion and organ deformation.  39  Figure 3.1: Contours of CTV vagina(cyan), CTV vessels(cyan), bladder(red) and rectum(brown) in transverse, coronal and sagittal views.  40  Figure 3.2: Treatment planning CT simulation images and associated contours of PTV vagina(blue), PTV vessels(blue), bladder(red) and rectum(brown) in transverse, coronal and sagittal views. PTVs are defined as CTVs (cyan) plus a 7mm margin. The planning criteria are that V95, which is the volume of PTV covered by 95% prescribed dose, is more than 98% and the maximum dose is no more than 110% of the prescribed dose (45Gy). There are also dose constraints for critical organs. For example, no more than 60% of bladder (excluding the overlapping part of bladder and PTVs) can receive a dose of 30Gy and no more than 25% of bladder non-overlapping can receive 40Gy. All the constraints for critical organs and PTV coverage must be met for a treatment plan to be delivered. Seven fields are used as shown in Fig 3.3 with the PTVs  41  covered by the 95% isodose line. Dose volume histograms for CTVs, bladder, rectum and bowel are shown in Figure 3.4. The dose volume histogram is very useful to quantitatively evaluate the CTV coverage and critical organ sparing.  Figure 3.3: 7 field IMRT is demonstrated on a transverse CT slice. The PTVs are covered with 95% isodose line (light green).  42  Figure 3.4: Dose volume histograms for CTV vagina, CTV vessels, rectum, bladder and bowel. The CTV vagina location is subject to variable filling of both the bladder and the rectum. In our clinical protocol, patients are asked to have a full bladder and an empty rectum for CT simulation for planning as well as for daily treatments. To maintain reproducibility of an empty rectum, the evening prior to CT simulation and subsequent radiation treatments, the patients are asked to take 2 tablespoons of Milk of Magnesia. Once patients have daily bowel movements prior to treatment, patients do not need to continue to take the Milk of Magnesia.  43  3.2  CBCT Usage in Gynaecological Radiation Therapy  Cone beam CT (CBCT) is used for quality assurance in gynaecological IMRT in our institution. A CBCT of the pelvis is performed immediately before the first three fractions and weekly thereafter during the course of the radiation. The bladder and the rectum contours are evaluated on the images for each CBCT. Immediate feedback to the patients will be given regarding adequacy and consistency of bladder and bowel filling in an attempt to ensure reproducibility of the high precision treatment. If the bladder filling is out of tolerance, the oncologist will make a clinical decision depending on the case. The patient may be required to drink some water and be treated later, or the treatment will be postponed to another day. The CBCT and planning CT image sets may be registered based on bony structures. As shown in Figure 3.5, CTV vagina on planning CT and CBCT are not identical. The difference comes from day to day variations in bladder and rectum filling, the air bubbles in the rectum and internal organ motion. If the position of CTV variation goes beyond the 7mm margin, CTV coverage is compromised. The dosimetric consequences can be evaluated directly by recalculating the dose distribution for the CTV on CBCT images. V95 (the volume covered by 95% prescribed dose) and D98 (the dose received by 98% of the target volume) are chosen for evaluation of CTV coverage. The CTV vagina is surrounded by the posterior wall of the bladder and the anterior wall of the rectum. The shifts of the posterior wall of the bladder and the anterior wall of the rectum on three levels, the upper, middle and 44  Planning CT  CBCT  Registered CT/CBCT  Figure 3.5: Contours of CTV vagina, rectum and bladder on sagittal planning CT, CBCT and registered CT/CBCT images.  45  lower part of the CTV vagina, can be used to quantify the organ motions. From these CBCT images, day to day and patient to patient variation of bladder volume is obvious even when the patients were asked to have a full bladder. The adult bladder can contain up to 1 liter of fluid and the urge to urinate occurs when the bladder contains about 200cc urine[57]. This indicates that protocols for bladder and rectum preparation alone can not guarantee good consistency in organ position and daily image procedures are required for position verification when using highly conformal radiation therapy techniques. This preliminary study also shows that on most CBCT scans, the CTV coverage fits the clinical criteria. However, occasionally large and unpredictable shifts in CTV position may occur, which leads to a reduced CTV coverage. Figure 3.6 shows significant shifts between the planning CT and CBCT. Although our data are based on a limited number of CBCT scans, the observations are consistent with other studies. Haripotepornkul et al. used CBCT before and after each fraction to evaluate intra- and inter- fraction movement of cervix during IMRT[58]. Their results have shown that cervical motion averages approximately 3mm in any given direction. However, maximal movement of the cervix can be up to 18mm. One study on intact cervical cancer by Tyagi et al. is very similar to ours [59]. They contoured CTVs on planning CT and CBCT and projected CTVs from CBCTs to planning CTs to compare the CTVs and dose coverage. Inter-fraction motion caused a high probability of CTV under coverage, however, the mean volume of CTV missed was small (4cc). This study 46  Bladder day 1  rectum day 1  Bladder day 2  rectum day 2  Figure 3.6: Inter-fraction motion of bladder and rectum. Note that the vaginal CTV lies between and abuts both structures. The CTV is not displayed on this image in order not to obscure the bladder and rectal contours. Bladder Day 1 and Rectum Day 1 are structures on the CBCT images and Bladder Day 2 and Rectum Day 2 are the projections from planning CT. supports the need to develop adaptive therapy to improve the radiation delivery in gynaecological patients. Other modalities, such as MRI, are used for IGRT in cervical cancer treatment. T2-weighted MR scans were used before and weekly during the course of treatment to investigate position changes of the vagina after hysterectomy[60]. This study suggested the need to set inhomogeneous PTV margins to accommodate the largest changes in the position of the vaginal CTV in the anterior-posterior direction. On-line MRI guidance has been used for an IMRT planning study [61]. The planning study has shown that healthy tissue involvement can be reduced for patients with cervical  47  cancer. The clinical trial described in this chapter is ongoing and data analysis is incomplete. A comprehensive analysis of the CBCT image data is required to fully appreciate the potential benefit of CBCT beyond the basic QA purpose currently being served. However, based on the imaging data collected on a small number of patients, some conclusions may be drawn. Day to day bladder volume variation is an unavoidable reality. In most cases, displacements of posterior bladder wall and anterior rectum wall are small and thus the CTV coverage is satisfactory. Occasionally, significant organ motion and CTV undercoverage will occur. Further analysis of CBCT image data is beyond the scope of this thesis but remains an important body of work to complete. An increase in CBCT utilization is expected to benefit patients, therefore it is important to fully characterize the CBCT dosimetry, which is the primary focus of the remainder of this thesis.  48  Chapter 4  Cone Beam CT Dose by Monte Carlo Simulation The purpose of this study is to present a Monte Carlo based cone beam CT dose calculation. Due to lack of commercially available dose calculation algorithms and difficulty in modeling heterogeneous anatomical media for kV beams, Monte Carlo simulation is the only accurate method to provide patient specific dose distributions. By careful benchmarking of the Monte Carlo system, accurate patient dose from CBCT and IGRT can be provided to physicians for clinical decision making.  4.1  Introduction to Monte Carlo Simulation  Monte Carlo methods are a class of computational algorithms that rely on repeated random sampling to compute their results. These approaches usually follow a particular pattern: 1.Define a domain of possible inputs; 2.Generate inputs randomly from the domain using a certain specified probability distribution; 3.Perform a deterministic computation using the inputs; 49  4.Aggregate the results of the individual computation into the final result. Monte Carlo methods have become ubiquitous in medical physics over the last 50 years[62]. They are used to simulate radiation beams from medical linear accelerators and dose in target media such as human tissue by modeling electron-photon transport. In this case, the inputs are photons and electrons at certain energies. The specified probability distribution for photons are the probabilities of no interaction, photoelectric effect event, Compton scatter event or pair production event (Triplet production, in which pair production occurs in the field of an electron, is included). Monte Carlo will track the history of every particle and millions of their subsequent interactions can reveal the geographic and spectral distribution of particles as they pass through different media and dose deposited within a patient or a phantom. In practice, condensed history techniques are applied for fast dose distribution calculation. Because electrons undergo a huge number of collisions before being absorbed and the electron range is small (0.141mm for 100keV electrons), it is time-consuming and unnecessary to simulate these collisions for the purpose of dose calculation on a macroscopic scale (2.5mm to 10mm voxel size). A condensed history technique combines effects of many small collisions into a single, large-effect, virtual interaction and thus decreases simulation time without affecting the macroscopic dose distribution. As opposed to condensed history codes, event-by-event or track structure Monte Carlo codes can model the transport of a primary particle, event by event, together with all the secondary electrons. These codes record the 50  location, type of interaction, energy deposited, and other information at each point of interaction.  4.1.1  Simulation Codes  Two Monte Carlo codes are used in this thesis. EGSnrc(Electron Gamma Shower) [63] is the most widely used general purpose Monte Carlo radiation transport package for radiation therapy applications. This code is used to calculate dose distributions from radiation therapy and imaging procedures. This code can simulate the events that generate primary electrons but can not further simulate detailed electron interactions with matter. For the purpose of simulation of microdosimetric quantity, event-by-event simulation for each electron track including its secondaries is required. Thus, another code NOREC is used to provide detailed information regarding how primary electrons and their secondaries interact with liquid water. Also, NOREC code alone can not fulfill the task because it can not simulate the process that electrons are generated by kV x-ray interactions with matter.  4.1.2  EGSnrc  (Electron Gamma Shower - National Research Council) The EGS (Electron-Gamma-Shower) system of computer codes is a general purpose package for the Monte Carlo simulation of the coupled transport of electrons and photons in an arbitrary geometry for particles with energies above a few keV up to several hundreds of GeV. It was first developed at Stanford Linear Accelerator Center (SLAC) for high-energy physics (100 MeV to 100 GeV). Later, NRC and SLAC collaborated to make EGS also 51  work in the energy regime of interest in medical physics, viz 10 keV to 50 MeV. That is the enhanced version called EGSnrc. EGSnrc can simulate radiation transport of electrons (+ or -) or photons in any element, compound, or mixture[63]. It is a condensed history Monte Carlo code. The interactions that can be simulated are: • Bremsstrahlung production using either Bethe-Heitler cross sections or the NIST (National Institute of Standards and Technology) cross sections. • Positron annihilation in flight and at rest. • Multiple scattering of charged particles by Coulomb scattering from nuclei. • Møller (e-e-) and Bhabha (e+e-) scattering. Exact rather than asymptotic formulae are used. • Continuous energy loss applied to charged particle tracks between discrete interactions. • Pair production. • Compton scattering, either Klein-Nishina or bound Compton. • Coherent (Rayleigh) scattering can be included by means of an option. • Photoelectric effect. • Relaxation of excited atoms after vacancies are created and Auger and Coster-Kronig electrons may be produced and tracked if requested. 52  • Electron impact ionization can be modeled.  4.1.3  BEAMnrc Code  In response to a need to model radiation therapy sources for applications in 3D treatment planning, the OMEGA (Ottawa Madison Electron Gamma Algorithm) project was initiated by Canada’s National Research Council and the University of Wisconsin. BEAMnrc was part of this project for radiotherapy source simulation[64]. A feature of BEAMnrc is that the linear accelerator is built by a selection of pre-defined geometries, called component modules (CM). To construct a medical linear accelerator, the following component modules are usually used: • target • primary collimator • flattening filter • mylar mirror • secondary collimating jaws • exit window with reticule (optional) • Multi-Leaf Collimator (MLC, optional) These modules are oriented in a plane perpendicular to the beam axis and can be customized based on the specifications of the particular linear accelerator being modeled.  53  Figure 4.1 shows an example of an accelerator built by BEAMnrc. The input is the incident electron spectrum, with information on density distribution, angular distribution and energy distribution. The BEAMnrc software is capable of tracking the primary, secondary and later generations of photons and electrons emerging from the target. The output is stored in a plane located perpendicular to the beam axis at user specified distance from the radiation source. This plane is called a phase space. The information recorded at each phase space includes the traversing particle’s position, energy, direction, charge, weight and point of origin[64].  Figure 4.1: Example of linear accelerator constructed in BEAMnrc. Image is from BEAMnrc user’s manual[64].  4.1.4  DOSXYZnrc Code  The DOSXYZnrc code is a EGSnrc-based Monte Carlo code to calculate absorbed radiation doses within a 3D voxel-based phantom containing ICRU 54  (International Commission on Radiation Units) defined media and their interaction cross-sections[65]. The phantom can be manually constructed or generated from computed tomography electron density data. Radiation beams are directed at the phantom by defining the distance between the virtual isocentre of the linac and the location of the phase space and the angle of incidence relative to the phantom axis. The same photon interaction probability distributions used in BEAMnrc are applied to the patient/phantom dose deposition simulation. The output is the dose to each medium (water, tissue, air, lung, bone etc) within each voxel.  4.1.5  NOREC Code  NOREC, released in 2003, is a substantially improved version of the Oak Ridge electron transport code (OREC)[66]. It calculates the detailed eventby-event transport of a primary electron and all of its secondaries in liquid water. A database developed at the National Institute of Standards and Technology is used to obtain the differential cross sections for the elastic scattering of electrons by atoms. NOREC is implemented in the form of a C++ class called “Track”. It is designed for Win32 systems. The user needs to write a user code to initiate the transport of an electron. The input parameters for this function are: kinetic energy, starting position and direction of the electron. The results of the electron transport include type of interaction, coordinates of the interaction and energy deposited for each event. This information can be obtained by a function called “getLine”.  55  4.2  Calibration of a kV Cone Beam CT system  For Monte Carlo simulation to be useful in radiation therapy, the entire virtual system must be benchmarked(verified) against measured data from the x-ray source being modeled. The goal is to obtain accurate dose deposition information inside a patient or a patient like material (phantom). Usually, two steps are involved. First, the relative calibration compares the simulated and measured percentage depth doses (PDDs) and profiles for different field sizes. Second, the absolute calibration converts the Monte Carlo result to absorbed dose. The complete benchmarking process of the OBI x-ray source described in this thesis was generated by the author.  4.2.1  Building the X-ray Source Model  The kV beams are produced by an x-ray tube and modified by blades and bowtie, as shown in Figure 4.2. The rotating anode tube with a W-Re target is simulated with a source routine ISOURC10, Parallel Circular Beam Incident from Side. All specifications including the target design, the incident electrons’ angle to the x-ray tube, geometry of blades and bowtie, were obtained from Varian. The focal size of the incident electrons was 0.5mm diameter and the incident electrons were mono-energetic, parallel beam. The blades and bowties were simulated with component modules JAWS and SLAB. The energy thresholds for secondary electron and photon creation (AE, AP) and energy cutoff (ECUT, PCUT) for particle transport were set to AE = ECU T = 0.516M eV for electrons and AP = P CU T = 0.001M eV for photons. Koch-Motz bremsstrahlung angu-  56  Target  x  Blades  z Half Fan Bowtie  Figure 4.2: Accelerator Preview in BEAMnrc. Target (W-Re) , blades (Steel) and halffan bowtie (Aluminum) are shown in the XZ plane. X axis and Z axis show the scales in centimeter.  57  lar sampling and NIST bremsstrahlung cross section data were used. Photoelectron angular sampling(also called Sauter Photoelectric angular sampling, in which Sauter’s formula is used to determine the angle of the photoelectron instead of inheriting the incident photon direction), Rayleigh scattering, atomic relaxations, spin effect and electron impact ionization (EII) were included in the simulation.  4.2.2  Relative Dose Distribution  Dose profiles and percentage depth doses (PDDs) were generated with DOSX -YZnrc using the phase space with and without the bowtie. Corresponding measurements were performed in a motorized 3D water tank. One IC 10 ionization chamber (Scanditronix Wellh¨ofer North America, Bartlett, TN) was used as the reference and the other identical ionization chamber was used to measure PDDs and profiles. In Monte Carlo simulation, the uncertainty is related to the number of particles in each voxel. Small size of voxels is preferred for better representing the profiles and PDDs but requires more histories to be simulated in order to obtain a certain uncertainty. 5mm voxel size and 2% uncertainty were chosen and then the number of histories were adjusted until the uncertainty of simulation was reached. Number of histories was on the order of 108 . Fig 4.3 shows comparison of measured and simulated PDDs for various beam energies. The PDD is normalized to dose at 2cm depth. The discrepancy between measured and simulated doses are mostly within 2%.  58  Percentage Depth Doses 140.00  125MC  120.00  125Meas 100.00  100MC  100Meas percentage  80.00  80MC 80Meas  60.00  60MC 60Meas  40.00 20.00 0.00  0  5  10  15  depth (cm)  Figure 4.3: Measured (Meas) and Monte Carlo simulated (MC) PDDs for 60, 80, 100,125kVp x-ray beams. The field settings are SSD 100cm, field size 30cmX30cm, without bowtie. Fig 4.4 and Fig 4.5 show the in-plane and cross-plane profiles at 2cm depth for 100kVp beam. The profiles are normalized at central axis. In Fig 4.5, the x-ray beam intensity is asymmetrically distributed. This is due to angulation of the target. The distribution of the beam intensity decreases towards the anode due to absorption of the x-ray beam by target and anode material. As shown in Figure 4.6, several photons are given off at a point within the target. Those photons go toward side A stand a greater chance of being absorbed because they have to travel through more target material than those which goes toward side B. Consequently, the intensity of the x-ray beam is greater on side B than on side A. This nonuniformity is called the  59  heel effect. The profile in Figure 4.5 is further modified by bowtie filter as shown in Figure 4.7. As the x-ray source rotates to take volumetric images, the total dose distribution becomes even.  Profile without Bowtie 120  Relative dose  100  80  Meas 60  MC  40  20  0 -30  -20  -10  0  10  20  30  position (cm)  Figure 4.4: Measured in-plane dose profile (Meas) compared with Monte Carlo simulated dose profile (MC) at depth of 2 cm for 100kVp x-ray beam. The field settings are SSD 100cm, field size 30cmX30cm, without bowtie. Relative dose is normalized to 100% at the center of the field.  60  Profile without Bowtie 120  Relative dose  100  80  Meas MC  60  40  20  0 -30  -20  -10  0  10  20  30  position (cm)  Figure 4.5: Measured cross-plane dose profile (Meas) compared with Monte Carlo simulated dose profile (MC) at depth of 2 cm for 100kVp x-ray beam. The field settings are SSD 100cm, field size 30cmX30cm, without bowtie. Relative dose is normalized to 100% at the center of the field.  4.2.3  Absolute Calibration  A calibration factor which converts Monte Carlo result to absolute dose distribution in CT data set was obtained from absolute dose measurement. The measurement Dabs was done in water tank at 2cm depth with 10×10cm field size and 100cm SSD with a Capintec Model C11 ion chamber and Exposure/Exposure Rate Meter model 192 calibrated for kV range according to recommendations by AAPM report TG61[67]. Monte Carlo calculated dose Dxyz was simulated under the same setup. The calibration factor F  61  A  B  Figure 4.6: Demonstration of heel effect. Photon intensity at side A is lower than that at side B because of more absorption by the target. was defined as Dabs = Dxyz × F × mAs  (4.1)  In this equation, Dxyz is the Monte Carlo calculated dose with the unit of Gy/particle and Dabs is the measured dose in cGy. Calibration factor can be calculated based on mAs, Dxyz and Dabs . Using the calibration factor F and total mAs from the image protocol, Monte Carlo results were converted to absorbed dose. After accurate benchmarking work, simulation of CBCT was performed for the gynaecological patients with half fan bowtie.  62  Profile with Half-fan Bowtie 120  100  percentage  80  60  Meas MC  40  20  0 -30  -20  -10  0  10  20  30  position (cm)  Figure 4.7: Measured dose profile (Meas) compared with Monte Carlo simulated dose profile (MC) at depth of 2 cm for 100kVp x-ray beam with half-fan bowtie. The field settings are SSD 100cm, field size 30cmX30cm. Relative dose is normalized to 100% at the center of the field.  4.3  Dose Distributions from Cone Beam CT Scans  4.3.1  Sample Dose Distributions  Figure 4.8 shows a Monte Carlo simulated dose distribution resulting from a pelvic kV CBCT scan for a medium size patient. The dose distribution calculated from EGSnrc was imported to the Eclipse treatment planning system for display. In summary, organ doses are in the 1-5 cGy range and bone doses are in the 3-14 cGy range depending on patient size. Because the photoelectric 63  Figure 4.8: Dose distribution in one scan for gynecological patient. Dose is represented on a color wash scale indicated in the lower right corner. effect is dominant in kV energy range, bone, as a high Z material, absorbs dose 2 to 4 times higher than soft tissues.  4.4  Discussion  The methods developed and validated in this chapter can hence be applied to any Varian OBI CBCT scanning protocol with parameters within the range studied. This includes kVp from 100 to 125, full bowtie, half bowtie and any choice of mAs. Simulations may be performed on any patient CT dataset. Since the CBCT imaging dose is always delivered in combination with treat-  64  ment dose, it is more meaningful to discuss detailed dose information in the context of a specific group of patients with specific radiation therapy dose. In Chapter 6, detailed results for gynaecological patients undergoing IMRT and CBCT image guidance are reported. It should be noted that Varian updated the OBI CBCT image reconstruction algorithm and image protocol from version 1.3 to 1.4 in 2008. Patient dose has been reduced markedly. For example, the previous pelvis scan (OBI 1.3) used 125kV, 80mA and 25ms while the current pelvic scan (OBI 1.4) uses 125kV, 80mA and 13ms. Under the same voltage, patient dose is proportional to total mAs, i.e., the patient dose is reduced by 48%. The CBCT image dose to patient for the TrueBeam system was further reduced by roughly 40% because the additional kV beam hardening filter in the new x-ray source design removes low energy photons in the energy spectrum, which leads to reduction of patient dose[68]. To put the work described in the current chapter into context, it is interesting to discuss other studies on kV CBCT patient dose that have been carried out in recent years. These studies include dose measurements in phantoms, accurate Monte Carlo simulation of CBCT dose, dose estimations, effective dose calculations, and risk estimates of associated secondary malignancy[69–78]. Ding [69–71] is the first author to model Varian CBCT dose by Monte Carlo simulation. His studies include characteristics of kV beams used in CBCT, accurate patient dose associated with CBCT and dose calculation by configuring a kV x-ray source in a commercial treatment planning system. Our Monte Carlo simulation and Ding’s work were essentially done concurrently but independently. Similar parameters were used in both 65  studies. The results were compared directly and excellent agreement validates our work. Downes et al.[72] reported Monte Carlo simulation and patient dosimetry for the x-ray volume imager (XVI), the CBCT mounted on the Elekta Synergy linac. Wen et al. measured skin doses for real patients and surface doses for phantom by TLD from daily pelvic scans [73]. Kan et al. measured an extensive set of organ doses using a female anthropomorphic phantom [74]. Hyer et al. [75] studied the patient organ doses and effective dose for XVI and OBI CBCT systems. This study included a comprehensive set of organ dose measurements using an anthropomorphic phantom and a fiber-optic coupled dosimetry system. Perks et al. [76] reported the peripheral dose of kV CBCT on surface and midline of an anthropomorphic phantom. These measurements would be time consuming and require specialized equipment. Also, the dose information is limited to phantom availability and can not be patient-specific, i.e. patient size can not be accounted for. Hyer et al. [77] used a simple CT dose index (CTDI) phantom to estimate organ doses. An analytical model based calculation was developed by Pawlowski et al.[78]. These calculations are fast and do not require any specialized equipment. However, the accuracy is compromised, particularly for bone . Among these studies, Monte Carlo simulation is the most accurate and patient-specific. Computation time is long compared with model base calculation, however, once the system is calibrated, it takes only a few minutes for physicists to setup the simulation. The dose calculated by Monte Carlo simulation may be imported to Eclipse treatment planning system and can be combined the prescribed dose. The imaging dose can then be accounted 66  for during the optimization of IMRT planning. By including the image dose, the total dose to tumor and surrounding critical organs will be more accurate. This process was done by adapting existing tools specific to BCCA so that they are locally available for future application. Alaei and Ding[71] included the dose from kilovoltage cone beam CT in the radiation therapy treatment plans differently. They configured the Varian kV x-ray source in Pinnacle treatment planning system to calculate the dose to patient and included it in patient treatment plan. Their method is fast but less accurate.  4.5  Conclusion  Careful benchmarking of the Monte Carlo simulation model indicates agreement between measurement and calculation within 2% for simple water tank measurements. Dose calculations on patient CT data sets are consistent with other studies. Doses to soft tissue are in the 1-5 cGy range and doses to bone are in the 3-14 cGy range for a single CBCT scan. Simulation time is around 5 hours on a cluster using five Pentium dual-core CPUs and an uncertainty of 2% can be achieved. Thus, confidence in the Monte Carlo simulation has been established and results may be used for clinical decision making. Results of this chapter will be combined with those of Chapter 5 to determine organ specific biologically effective doses in the patient study described in Chapter 6.  67  Chapter 5  Dose Mean Lineal Energy Based Quality Factor In existing studies regarding effective doses or risks of secondary cancer from CBCT, the quality factor of the kV beam is considered to be unity, which is in conflict with experimental data that show that kV beams have a higher biological effectiveness compared with MV beam[79–81]. In this chapter, we will use a microdosimetric quantity, lineal energy, to quantify the quality factor for kV cone beam CT. This quality factor can be further used in the context of calculating the equivalent doses from kV cone beam CT when making comparisons with MV doses.  5.1 5.1.1  Introduction to Beam Quality Relative Biological Effectiveness (RBE) and Quality Factor (Q)  The effect of ionizing radiation depends on properties of the radiation (type and energy), the way the radiation is delivered (dose, dose rate, fractionation) and endpoint of the assessed cells or tissue. Relative biological ef-  68  fectiveness (RBE) is a very specific concept. It is introduced to account for differences in biological effect (expressed through a particular endpoint) from different types of radiation under specified irradiation conditions. RBE can be defined for cell survival or for incidence of chromosomal aberration or other biological endpoint for a range of doses and its value depends on dose or level of response. Definition of RBE is the ratio of a dose of a standard low linear energy transfer X ray beam (x ray of 250KeV energy or  60 Co  gamma ray)(Dx ) to the dose of the test radiation type (DT ), required to cause the same degree of a particular biological endpoint[82]. It is written as RBE = Dx /DT  (5.1)  Quality factor Q, on the other hand, by its nature applies to low levels of radiation only. It is used to quantify differences in radiation effects related only to the properties of radiation (type and energy). The quality factor is a somewhat arbitrarily chosen conservative value based on a range of RBEs[82]. Thus, the Q factor encompasses RBEs in a very broad sense, independent of the organ or tissue or of the biologic endpoint under consideration. Ideally, to set the quality factor value for cancer risk, an endpoint related to cell transformation from normal to malignant has to be used. However, as was stated above mechanisms of carcinogenesis are not fully understood. Therefore, representative surrogates, such as low dose RBE for yields of chromosomal aberrations, are used to justify setting quality factor to specific values. The dosimetric quantity relevant to radiation protection is the equivalent dose(H), which is used to to place biological effects from  69  exposure to different types of radiation on a common scale. It is defined as: H =D×Q  5.1.2  (5.2)  LET and Quality Factor  Linear energy transfer, or restricted linear collision stopping power, L∆ is a measure of the energy transferred to material as ionizing particles travel through it. The International Commission on Radiological Units in 1962 defined this quantity as follows: The linear energy transfer (L) of charged particles in a medium is the average energy locally imparted to the medium by a charged particle of specified energy in traversing a distance of dl.  L∆ =  dE∆ dl  (5.3)  where dE∆ refers to the energy loss due to electronic collisions minus the kinetic energies of all secondary electrons with energy larger than ∆. L∞ is identical to the linear collision stopping power. Different patterns of energy deposition cause differences in relative biological effectiveness. Figure 5.1 shows why certain ionization densities of radiation cause maximum DNA damage. DNA double-strand breaks have been conclusively shown to lead to formation of chromosomal aberration and thereby cell death or corruption of the genetic code[83]. If radiation events are sparsely distributed, as shown in Track A, probability of a double-strand break being produced by a single particle track is low. As density of energy deposition increases double-strand breaks are effectively produced by a sin-  70  gle track. Track B shows the most efficient pattern for double-strand breaks: the average separation between ionizing events coincides with the diameter of the DNA double helix. In track C, since some energy is deposited between the two strands, the biological effectiveness is lower than that of track B. LET = 100keV /µm corresponds to the peak of RBE. Such high LET value is for charged particles heavier than electrons. Also, spatial distribution of double-strand breaks produced by high LET radiation is different from that from low-LET. In the former case double-strand break are likely to be produced in close proximity and therefore will be prone to exchanges. In the latter case not only likelihood of more than one double-strand break being produced by one particle track is low, even if these breaks are produced they are not proximal to each other due to low ionization density and probability of exchange is diminished. LET for photons and electrons is less than 10keV /µm, for example, LET for 200-kV x-rays is 3.5keV /µm, in contrast, LET for  60 Co  γ-rays (1173kV and 1332kV)is 0.4keV /µm[82]. It should be  noted that photons deposit energy by setting electrons in motion. LET is therefore assigned to photon radiation through electron energy spectra from photon-matter interactions. In the low LET range, RBE increases as LET increases. Thus low energy x-rays have a higher LET compared with high energy x-rays and higher RBE.  71  LET≈100keV/μm  LET ≈ 300keV/μm  X-ray  20Å  A  B  C  Figure 5.1: Energy deposition for X-ray and particles with different LETs. For x-rays, double-strand breaks can not caused by one track. When LET is around 100keV /µm, the average separation between ionizing events coincides with the diameter of the DNA double helix and the radiation is most efficient in causing double-strand breaks. As LET become higher than 100keV /µm, some energy was “wasted” between the two strands so that RBE decreases. LET has been used to quantify quality factor. A simple relationship between Q and L∞ was given in ICRU/ICRP report[84]: Q = 0.8 + k × L∞  (5.4)  in which k = 0.16µm/keV . This equation is usually sufficiently conservative in the range up to 100keV /µm. However, the concept of LET suffers from  72  some deficiencies. LET by definition is an expectation instantaneous value assigned to a particle with a particular energy. LET does not account for stochastic nature of energy depositions in volumes with dimensions of the order of DNA or cell nucleus. Escape of energy via delta rays, changes of LET along a track and track curvature cause complexities in relating the actual energy deposition in small volumes to LET. Thus, ICRU report 40 recommends lineal energy, y, to quantify Q[85].  5.1.3  Microdosimetry  Microdosimetry has been used to provide a better understanding of the mechanism and quantification of biological effectiveness. Linking Q to microdosimetric data could reflect a relation between radiation risk and energy deposited to a small volume of interest. This relationship is more fundamental than that of Q and L∞ . The term microdosimetry was proposed by Rossi and coworkers, describing a conceptual framework and the corresponding experimental method for analysis of the microscopic distribution of energy deposition in irradiated matter[86–89].  Lineal energy The lineal energy, y , is defined as the quotient of ε by ¯l, where ε is the energy imparted to the matter in a volume of interest by an energy deposition event and ¯l is the mean chord length in that volume[90]. y = ε/¯l  (5.5)  73  For a sphere of radius r, ¯l = 4r/3. The lineal energy, y, is a stochastic quantity. The distribution and mean value of y are useful. F (y) is the distribution function, which is the probability that lineal energy is equal to or less than y. Its derivative with respect to y dF (y) dy  f (y) =  (5.6)  is the probability density, also called lineal energy distribution. The frequency mean lineal energy is defined as ∫  ∞  y¯F =  yf (y)dy  (5.7)  0  And dose mean lineal energy is derived as 1 y¯D = y¯F  ∫  ∞  y 2 f (y)dy  (5.8)  0  These two variables are non-stochastic quantities and independent of absorbed dose or dose rate.  Lineal Energy in Radiation Protection As discussed, the pattern of energy deposition is closely related to the biological effectiveness of radiation, and energy deposition in small volumes is complex. Lineal energy describes a property of radiation at just the point of interest in the receptor, not an average of the incident radiation. Lineal energy can be measured in a spherical proportional counter. It is made of tissue equivalent (with the effective atomic number and the same 74  mass energy absorption coefficient for selected range of photon energy) plastic shell filled with tissue equivalent gas. The energy deposited by passage of a charged particle in a tissue equivalent gas volume of density ρ and diameter d is equivalent to the energy that would be deposited in unit density tissue with diameter dρgas /ρtissue . By adjusting the gas density of the proportional counter, very small diameters can be achieved[86]. Ideally, the sphere size should be the same as DNA dimensions, i.e. 20nm. Historically, this dimension was too small and beyond the capability in measurements and 1µm diameter was used in measurement and calculation. An empirical correlation was established between RBE from experimental data and lineal energy for a 1µm diameter sphere. ICRU[91] recommended relationship between Q and yD as Q = 0.8 + k ′ × y D  (5.9)  in which k ′ = 0.14µm/keV . By this equation, Q is empirically fitted to a functional relationship with yD . In this thesis, lineal energy is calculated based on a 1µm diameter sphere and quality factor is calculated by this equation.  5.1.4  RBE for Low Energy X-rays  Extensive experimental work has been carried out measuring different endpoints in many different cell lines, showing that low energy x-rays have a higher RBE compared with higher energy x-rays. These endpoints include cell inactivation, chromosome aberration, cell transformation, micronuclei  75  formation. Chromosome aberration data play a very important role in biological effectiveness comparisons of different radiations and risk estimation. Chromosome aberration in human lymphocytes is a quantitative measure of the amount of damage induced in cells after exposure. In 1960, Moorhead et al.[92]described a method for stimulating human blood lymphocytes to divide in culture and many quantitative studies on human material have followed. Lloyd et al.[79] analyzed how the cell survival parameters in the linear-quadratic model vary with radiation quality. Mestres et al.[80] used fluorescent in situ hybridization to evaluate RBE of 30kVp to 180kVp xray and 60 Co gamma rays. The irradiation dose ranged from 0.05 to 3Gy. Schmid et al. have published papers reporting chromosome aberrations for different radiation qualities[93, 94]. A quality factor of up to 4 for 29kVp x-rays used in mammography screening is under debate[81, 95, 96]. All these data have shown higher RBE of low energy x-ray compared with MV x-rays or γ rays, however the uncertainties and inter-laboratory variations are large. Even though ICRP [97] decided to simply assign Q(L)=1 to all lowLET radiations (L < 10keV /mm) in Report 60, it has been known that low energy x-rays have a higher biological effectiveness compared with MV x-rays. ICRU Report 40 also assigned Q = 1 to photons and electrons, but this report specified that this value should be assigned to photons and electrons with energies in excess of 30keV. In kV cone beam, a great portion of photons are below 30kV. In this thesis, the quality factor for kV cone beams will be explored using dose mean lineal energy.  76  5.2  Monte Carlo Simulated Lineal Energy  5.2.1  Materials and Methods  According to the definition of y D , it can be simulated by Monte Carlo codes EGSnrc and New Oak Ridge Electron Transport Code (NOREC)[66]. The quality factor Q can then be calculated by the equation 5.9 on regions of interest within a patient. This process includes three steps: 1. Generate Electron Spectrum by EGSnrc As discussed in Chapter 4, the kV x-ray source was simulated by BEAMnrc and the energy deposition in patient CT images was simulated by DOSXYZnrc. DOSXYZnrc offers the function to record each interactions in a history. This record includes the types of interactions (photoelectric effect, Compton scattering, Rayleigh scattering etc.), the energies of generated electrons and photons, position and direction of the interaction. Thus, electron information can be extracted from any volume of interest. 2. Generate Electron Tracks by NOREC Input to NOREC is electron energy (in the range of 7.4eV to 1MeV), position and direction. Output is detailed event-by-event transport of a primary electron and all of its secondaries in liquid water. The file format is as follows: • 1st column: 1 for primary or 2 for all secondary and higher order electron; • 2nd column: types of interactions: excitation, ionization, elastic scatting;  77  • 3rd-5th columns: Cartesian coordinates of the interaction; • 6th Column: energy deposited at the point. More than 2000 electrons were recorded for each volume of interest (i.e., volume in water slab, volume in an organ) and were sent to NOREC to generate electron tracks. 3. Lineal Energy Calculation by Track Library The calculation model was set up in this way: a sphere with diameter of 1µm was put at the center of a cube; the dimension of the cube is two times of the maximum range of the electrons (300µm for 125kVp), which means any track that possibly reaches the sphere was initiated in the cube; the electron tracks were distributed in this cube; events in the sphere were recorded and y D was calculated based on its definition. Since the sphere is very small compared with the cube, hundreds of millions of electron tracks need to be simulated to allow enough electron tracks to hit the sphere. It is not realistic to simulate every track. Instead, more than 2000 electron tracks were simulated with NOREC to build a track library and these tracks are recycled many times in the simulation. To determine an appropriate number of tracks to use in this process, testing was performed by comparing results from two sets of electron tracks for the same volume of interest. The number 2000 was chosen based on consistency of results. To establish this method, it was applied to monoenergetic photons and simulation results were compared to ICRU report[91]. After the validation, the method is applied to the CBCT beam dose distributions calculated on 78  patient data sets. This study was initially done on pediatric patients, because this patient group is more sensitive to radiation. Later, it was expanded to adult patients.  5.2.2  Results and Conclusion  In order to validate our simulation, dose mean lineal energies for monoenergetic photons were simulated and compared with experimental results from ICRU. As shown in Table 5.1, discrepancy between simulation and measurements is within 10%. energy (kV)  measured from ICRU  yD  simulated y D  12  5.3  5.3  25  4.1  3.8  36  3.5  3.6  140  3.5  3.7  Table 5.1: Comparison of measured lineal energy and simulated value for monoenergetic photons. Table 5.2 shows the dose-mean lineal energies and quality factors for the paediatric abdominal scan with half fan bowtie. The dose-mean lineal energies vary from 3.6 to 3.9 keV/mm for soft tissue, which corresponds to a quality factor range of 1.31-1.35. Except for superficial locations, the quality factor values vary little. The quality factor for bone is slightly lower than that for soft tissue because in bone the photoelectric effect is the dominant effect, which makes the electron energy spectra different from soft tissue.  79  Organ  yD  Q  Liver  3.6  1.31  Left kidney  3.7  1.32  Spinal cord  3.7  1.32  Spleen  3.7  1.32  Stomach  3.9  1.35  Vertebral bone  3.4  1.27  Table 5.2: Dose-mean lineal energies and quality factors of kV CBCT for a paediatric abdomen scan (with half fan bowtie).Computational uncertainty is about 2 %. Table 5.3 shows the dose-mean lineal energies and quality factors of a head and neck scan with full fan bowtie. The dose-mean lineal energies vary from 3.5 to 4.4 keV/mm, which lead to a quality factor range of 1.29-1.41. Similar to the abdominal scan, the quality factor values for positions other than close to skin vary in a narrow range. The quality factor for eye is slightly higher than other organs because it is superficial. Due to the beam hardening effect, mean photon energy at superficial positions is lower than that at larger depths, which makes the dose-mean lineal energy higher. The lineal energy is also calculated for a pelvic scan on a medium sized woman. Lineal energy in field center is yD = 3.6keV /µm,and lineal energy in peripheral region is yD = 4.0keV /µm. Both of these lineal energy will lead to a quality factor of about 1.35. Photon energy from 60 Co is similar to mean energy of 6MV photon beam, thus is used to be representative of the MV beam. From ICRU report 36,  80  Organ  yD  Q  Right eye  4.4  1.41  Brainstem  3.9  1.35  Cord  3.5  1.29  Left parotid  3.6  1.31  Left temporal lobe  3.8  1.33  Table 5.3: Dose-mean lineal energies and quality factors of kV CBCT for a head and neck scan (with full fan bowtie).Computational uncertainty is about 2 %. the  60 Co  dose-mean lineal energy value yD = 1.8keV /µm, and a quality  factor of 1.05 can be calculated. The quality factor for CBCT is therefore approximately 25 % higher than that of MV beam. Due to the relative homogeneity of the quality factors in different organs, for practical reasons a fixed value 1.3 may be considered instead of organ specific quality factors for the purpose of evaluating risk of secondary malignancy.  5.3  Discussion  Use of a specific imaging modality in IGRT should be subject to thorough dosimetric investigation prior to implementation. This study demonstrates that using a microdosimetric approach to compute quality factors is feasible. Combined with a macroscopic calculation of dose, this approach leads to improved accuracy in the biologically equivalent dose distributions associated with both kV and MV CBCT procedures. This would be particularly important in applications involving paediatric patients who survive for long 81  periods of time following treatment and are at high risk of secondary malignancy. Besides comparing kV and MV CBCT procedures, this study also provides a quality factor which may be used for combining CBCT dose with therapeutic dose. One limitation of this study is that the NOREC code can only simulate tracks for elections with energy lower than 1MeV. Thus, therapeutic x-rays with energy of 6MV are out of the range. However, the reference radiation  60 Co  is a very good approximation of a 6MV beam.  Thus, in Chapter 6, the quality factor for IMRT is chosen as unity.  5.4  Conclusion  When comparing the dose from kV CBCT and dose from an MV treatment beam, the differences between quality factors of the beams should be accounted for. This study demonstrates how Monte Carlo microdosimetric methods can be effectively utilized in this context. Quality factor of 1.3 can be used in CBCT to calculate equivalent doses.  82  Chapter 6  Organ Equivalent Doses for Gynaecological Patients Undergoing IMRT or RapidArc with kV CBCT This chapter combines the Monte Carlo dose simulation methods developed in Chapter 4 and the quality factor data derived in Chapter 5 to obtain organ equivalent does for gynaecological cancer patients. Three different dose response models are used to calculate organ equivalent doses for daily kV CBCT and to compare the CBCT doses with MV IMRT/RapidArc doses to assess whether patients are at increased risk of radiation induced secondary malignancy due to CBCT.  6.1  Introduction  Postoperative pelvic radiation therapy is common in the treatment regimen for many patients with gynaecological malignancies. Intensity modulated  83  radiation therapy (IMRT) is often used to reduce doses to organs at risk (OAR) such as rectum, bladder and bowel. As discussed in previous chapters, organ motion in these patients is a significant concern, encouraging the use of daily imaging. Cone Beam CT (CBCT) may be used to monitor organ motion in order to ensure adequate target volume coverage as well as to assess dose to organs at risk. In order to effectively and safely implement a CBCT protocol it is important to quantitatively explore the CBCT doses compared with the total doses received by the patient. Both patients and staff are sensitive to the issue of imaging dose due to recent literature and media coverage [98, 99]. The need for imaging dose management was discussed in detail in the Report of AAPM Task Group 75 [100]. In this report there were several recommendations that we have considered in the current study. The gynaecological cancer patient population undergoing IMRT is a good cohort in which to demonstrate the relative dosimetric impact of kV CBCT. The imaging dose could be particularly relevant for younger low risk patients who have a good chance of cure and may survive for many years post treatment. Several studies have indicated that gynaecological patients undergoing radiation therapy are at higher risk to develop secondary malignancy compared with patients not receiving radiation therapy [41, 42]. It could be expected that patients in this group would be seeking reassurance that the imaging procedures performed on them are safe. While it is of utmost importance to keep imaging doses as low as possible while providing sufficient image quality to provide benefit to the patient, concerns over imaging dose could lead to under-utilization of CBCT technology resulting 84  in loss of benefit to the patients. It is expected that a detailed and accurate analysis of relative doses, will assist with educating patients and staff in decision making around CBCT utilization. As discussed in Chapter 4, CBCT doses have been reported in literature based on using Monte Carlo simulations, ion chamber, TLD or MOSFET measurements in phantom or at the surface of patients, configuration in a commercial treatment planning system and estimates by cone beam dose index (CTDI). Effective doses have been calculated for different scan protocols [101] and risks of secondary malignancy have been estimated for pediatric patients who are particularly susceptible to radiation induced carcinogenesis [102]. Studies to assess relative risks of second malignancy in gynaecological patients undergoing IMRT have also been carried out [103]. However, some shortcomings exist in these studies. First, the quality factor of the kV beam was simply assumed to be unity while evidence from experimental data suggests that biological effectiveness of kV photon beams is higher than that of MV photon beams. This factor was calculated in Chapter 5 and will be included in our study. Second, the risk associated with CBCT dose was calculated based on CBCT dose only regardless of the fact that the CBCT dose was always combined with high treatment dose. This calculation defaults to the linear no threshold dose model which is not valid when dose is higher than 2.5 Sv [48]. Third, regions peripheral to treatment fields were usually ignored and only phantom measurements were used [76]. Finally, measurements or simulations often did not take patient sizes into account and look at only select points, not using an organ based approach[104]. In this study, in order to combine imaging kV and radiotherapy MV 85  doses for the purpose of risk estimates, we take relative biological effect into account using quality factors. For the risk models, we applied the concept of organ equivalent dose (OED) proposed by Schneider, which was introduced in Chapter 2. Three different dose response models, the linear model, linear-exponential model and plateau model are used to calculate OEDs. Our study of peripheral dose not only reports CBCT doses, but also compares with IMRT scatter dose and linac leakage to demonstrate the relative amount of CBCT dose. CT image data sets from different sizes of gynaecological patients were selected to demonstrate OED variations in this patient group. The aim of this study was to comprehensively evaluate the organ equivalent doses both in-field and in the peripheral region from CBCT for gynaecological cancer patients undergoing IMRT or RapidArc and compare CBCT dose with IMRT scatter and linac leakage doses.  6.2 6.2.1  Materials and Methods Dose Distributions in Gynaecological Patients  The physical doses were obtained by Monte Carlo simulation as stated in Chapter 3. The CT scans for gynaecological patients were intravenous contrast enhanced, causing high Hounsfield units (HU) in kidney and liver (up to 250 HU). In the process of generating phantoms used in DOSXYZnrc, tissue material ICRUTISSUE and density 1.0g/cm3 were assigned to these two organs. 5 × 109 histories were used resulting in 2% uncertainty in doses within the imaging field and up to 15% uncertainty in peripheral regions. 86  The dose mean lineal energy based quality factor was 1.3 as stated in Chapter 5. Equivalent dose is the product of physical dose and quality factor. Small, medium and large patients were chosen from the eligible patients according to the product of the anterior posterior and lateral separation at the isocenter. Patients with a planning CT scan up to the position of the fourth rib were chosen to evaluate mean organ doses in the peripheral region. Peripheral organs are defined as those organs for which the entire organ lies outside of the 5% isodose line and outside the CBCT field. Doses from CBCT were reported as equivalent doses, where equivalent dose is the product of dose from Monte Carlo simulation and Q. Leakage dose per MU from the linac head was measured by Wellh¨ofer IC 10 ionization chamber (IBA dosimetry, Schwarzenbruck, Germany) in solid water (Gammex RMI, Middleton, WI). Measurements of cGy/MU were done with the minimum jaw settings of 0.3 cm by 0.4 cm. The ionization chamber was placed at depth of 1.5 cm at off axis distances of 25cm, 50cm, 75cm from the isocenter along the longitudinal direction (away from the gantry). Leakage dose for each patient was then estimated based on MUs from the Eclipse treatment planning system (TPS). Mean organ doses from IMRT were obtained by Monte Carlo simulation because the Eclipse algorithm is known to underestimate dose in the peripheral region [105]. IMRT dose was calculated using a BEAMnrc/DOSXYZnrc based program. In our model, the 6 MV photon beam was generated using a monoenergetic electron beam of 6.00 MeV incident on the target with a radial Gaussian spatial distribution characterized by a FWHM of 0.75 mm. Dynamic delivery of IMRT was modelled using a phase space modulation code that simulated particle transport through a 87  moving multileaf collimator (MLC) [106, 107].RapidArc delivery was calculated with a new DOSXYZnrc source developed by Lobo and Popescu [108] [109]. The transport parameters AP = PCUT = 0.010 MeV and AE = ECUT = 0.700 MeV were used and the number of histories was 1.0 × 109 in DOSXYZnrc. Dose for 25 CBCTs was calculated to represent the maximum expected usage of CBCT. In gynaecological IMRT treatment, bladder, rectum and bowel are partially in the high dose volume. Mean doses to these organs were about 20-30Gy. The linear dose response model works for doses up to 2.5 Sv and is no longer valid at such high dose levels. Cell killing effect should be taken into account. Additionally, doses are usually inhomogeneous for organs close to the PTV. The concept of organ equivalent dose by Schneider was adopted in this region. Schneider et al. determined the parameters α and δ in these models by fitting a combination of Japanese Atomic bomb data and Hodgkin’s cohort data. These parameters were fitted such that secondary cancer incidence was proportional to OED. Because of the high degree of uncertainty in the values of α and δ, a sensitivity test was done to assess how OED varies with these parameters by adjusting these values by +/- 50% in the linear-exponential model and plateau model.  6.3  Results  Figure 6.1 shows a CBCT dose distribution from a single scan on a coronal plane for a medium size patient. Dose to soft tissue was 1.5-4 cGy and dose  88  to bony structures was up to 10 cGy. Doses of less than 1cGy were delivered to the peripheral region.  Figure 6.1: Monte Carlo simulated CBCT dose distribution for a single pelvis scan. CBCT doses for 25 fractions representing daily imaging compared with IMRT scatter and leakage doses outside the primary beam are shown in Figure 6.2. The CBCT dose is higher toward the inferior end because the depth of mid-plane is shallower here and attenuation from surrounding bone is less compared with the superior end. The dose profile is truncated at the inferior end because this coincides with the inferior extent of the CT data set. The cumulative dose for 25 CBCT images simulating a daily imaging 89  Figure 6.2: Comparison of doses for 25 fractions showing cumulative CBCT doses and IMRT and leakage doses along the patient midline (black line) in Fig 1. The X-axis represents the distance from IMRT isocenter from inferior (-12 cm) to superior (+20 cm). The prescribed dose was 4500 cGy. The IMRT field extends from -6 cm to 9.3 cm, and the cone beam CT field extends from -11.8cm to 11.8cm as indicated on the figure. protocol varies from approximately 15 cGy at the field border to approximately 0.5 cGy, 30 cm away from the treatment field. IMRT leakage dose is approximately 6cGy uniformly distributed throughout the patient while leakage dose from RapidArc is about 3cGy due to the reduced number of monitor units. Scatter doses from IMRT and Rapid Arc are similar in magnitude, ranging from 100’s of cGy near the field border to about 10 cGy 30 cm away. IMRT or RapidArc doses are thus about 10 times higher than  90  Organ  Mean equivalent CBCT dose for 25 fractions  Mean IMRT or RapidArc scatter dose  Leakage dose(IMRT)  Leakage dose(RapidArc)  Lung  0.3 ± 0.1cGy  6 ± 2cGy  6cGy  3cGy  Breast  0.5 ± 0.2cGy  5 ± 2cGy  6cGy  3cGy  Kidney  3 ± 1cGy  40 ± 20cGy  6cGy  3cGy  Liver  1.7 ± 0.5cGy  25 ± 8cGy  6cGy  3cGy  Table 6.1: Comparison of CBCT doses and IMRT and RapidArc doses for organs in the peripheral region. *Lung and breast are calculated on partial organ volumes due to cutoff at the sup boundary of CT is at the fourth rib and are thus higher than mean dose to the whole organ. CBCT doses in peripheral regions. Total equivalent CBCT doses are compared with IMRT, RapidArc, and linac leakage doses for peripheral organs in Table 6.1. The mean number of monitor units (MU) for the RapidArc plans was 583 (8 patients) and mean MU for IMRT was 1192 (15 patients). Thus, the leakage dose from RapidArc was about half of that from IMRT. The estimated variation in CBCT and mean scatter doses is 30 to 50% based on both patient to patient variation and uncertainty from Monte Carlo simulation. Organ equivalent doses for colon, rectum and bladder are listed in Table 6.2 for IMRT alone and IMRT plus daily CBCT. The uncertainties in dose simulations are within 2%. Patient L was the largest patient among the 23 IMRT patients, with separations of 25 cm and 49 cm in the anterior-posterior (AP) and lateral directions, respectively. Patient S was the smallest IMRT patient, with AP separations of 19cm and lateral separation of 35cm, respectively. Patient M was a medium size patient treated with RapidArc 91  technique, with AP separation of 22 cm and lateral separation of 37 cm, respectively. Table 6.3 shows the result for the sensitivity test with α,δ adjusted by +/50%. Although OEDs are highly sensitive to parameters α and δ, percentage ∆OEDs are not significantly affected by the variations of parameters α and δ.  6.4  Discussion  Our results for simulated CBCT doses are consistent with previous studies. The difference is the interpretation of OEDs from kV CBCT. Kim et al. [102] used CBCT dose and cancer incidence data from BEIR VII to estimate secondary cancer incidence, which means applying the linear model in the high dose region. The linear dose response model likely overestimates the risk of secondary malignancy in high dose regions. The exponential component makes the model bell shaped, which means OED will decrease as the dose increases to higher levels, depending on the parameter α. Our results from a linear exponential model or plateau model show lower CBCT OEDs and consequently lower hypothetical incidence of secondary malignancy. When adding CBCT dose to IMRT dose, the OED may actually decrease if the linear exponential model is applied. The plateau model is designed to account for repopulation effects from fractionation and OED for this model continues to increase slowly at high dose levels. The plateau model may be the more appropriate model to use here. Epidemiological data upon which risk estimates are based are limited due to large uncer-  92  Organ  Model  Bladder (Patient L)  Bladder (Patient S)  Bladder (Patient M)  Bowel (Patient L)  Bowel (Patient S)  Bowel (Patient M)  Rectum (Patient L)  Rectum (Patient S)  Rectum (Patient M)  OED(Sv) (IMRT+CBCT) 35.51  ∆OED (Sv)  Linear  OED(Sv) (IMRT) 35.14  Linear-exponential Plateau Linear  7.26 7.07 28.22  7.22 7.08 28.87  -0.04(-0.6%) 0.01 (0.1%) 0.65 (2.3%)  Linear-exponential Plateau Linear  7.43 6.79 35.16  7.44 6.83 35.74  0.01 (0.1%) 0.04 (0.6%) 0.58 (1.6%)  Linear-exponential Plateau Linear  7.27 7.08 15.62  7.22 7.09 15.88  -0.05(-0.7%) 0.01 (0.1%) 0.26 (1.7%)  Linear-exponential Plateau Linear  5.33 4.75 20.75  5.38 4.80 21.23  0.05 (1.0%) 0.05 (1.1%) 0.48 (2.3%)  Linear-exponential Plateau Linear  5.58 5.18 12.58  5.62 5.24 12.90  0.04 (0.7%) 0.06 (1.2%) 0.32 (2.5%)  Linear-exponential Plateau Linear  4.38 3.95 28.34  4.45 4.02 28.78  0.07 (1.6%) 0.07 (1.3%) 0.44 (1.6%)  Linear-exponential Plateau Linear  5.34 5.57 29.10  5.41 5.66 29.74  0.07 (1.3%) 0.09 (1.6%) 0.64 (2.2%)  Linear-exponential Plateau Linear  6.01 6.01 38.44  6.06 6.10 39.05  0.05 (0.8%) 0.09 (1.5%) 0.61 (1.6%)  Linear-exponential Plateau  6.91 7.11  6.85 7.12  -0.06(-0.9%) 0.01 ( 0.1%)  0.37 (1.1%)  Table 6.2: Bladder, bowel and rectum OEDs from IMRT and combination of IMRT and CBCT for small (S), medium (M) and large(L) patients. 93  ∆ OED(Sv)(Linear exponential model)  ∆ OED(Sv)(Plateau model)  organs  α  α + 50%  α − 50%  δ  δ + 50%  δ − 50%  Bladder  −0.05 − 0.7%  −0.06 −1.7%  0.08 0.5%  0.01 0.1%  00  0.08 0.5%  Bowel  0.07 1.6%  0.04 1.6%  0.14 2.0%  0.07 1.8%  0.06 2.0%  0.12 2.0%  Rectum  −0.06 −0.9%  −0.07 −2.2%  0.05 0.3%  0.01 0.1%  00  0.06 0.5%  Table 6.3: OED increments with original parameters and parameters adjusted by +/- 50% are listed for patient M. tainties in dose [47]. The parameters α and δ used in the plateau and linear exponential models are derived from Japanese Atomic bomb and Hodgkin’s cohorts and are therefore approximations for this group of gynaecological cancer patients. However, OED results are relatively insensitive to the values α and δ . Variation in α and δ values of +/-50% alters the ∆OED values by <1.2%. It should be noted that change in OED with the addition of daily CBCT imaging to IMRT or RapidARc treatment is <2% in all cases using the plateau model even with the quality factor considered. This is less than the uncertainty in dose generally accepted in external beam treatment delivery of IMRT or RapidArc. Given that the plateau model likely represents the OED better than the two other models, these results should apply for any CBCT protocol one might practically consider. Within the imaging field of view, CBCT dose varies with patient size. CBCT dose for the largest patient was up to 43% lower than for the smallest patient in these internal organs. Since percentage depth dose for 125kVp x-rays drops off more than 40% in 6cm of tissue, it is not surprising to see a 43% difference in dose for two patients with 14cm difference in lateral  94  separation. In the linear model ∆OED is proportional to CBCT dose and thus varies with patient size. However, with the linear exponential and plateau models, the effect of patient size was less notable. For dose in the peripheral region, we compared the CBCT dose with linac scatter and leakage doses. The low dose regions are at low risk of secondary malignancy and the incremental dose from CBCT is one order of magnitude less than IMRT scatter doses and less than or equal to linac leakage dose. Using RapidArc instead of IMRT can further reduce the doses in this region. It should be pointed out that the number of monitor units required to produce a treatment plan will be impacted by linac calibration conditions. However, the magnitude of linac leakage and scatter doses relative to CBCT doses will not be affected by this. Local differences in planning experience and available optimization tools may also result in differences in MUs to some extent. Likely variations in MUs resulting from such local differences will not impact the conclusions of this study. It is also worth briefly mentioning dose to bone marrow. Dose to red bone marrow is considered to be related to hematologic toxicity [55]. A study by Walters et al. [110] using Monte Carlo simulation and phantoms with modeling of spongiosa based on micro-CT images shows that dose to red bone marrow in a pelvis scan is about 60% of dose to bladder. The shielding effect of trabecular bone by photoelectric effect causes the dose to red bone marrow to be lower than the surrounding soft tissue. With 25 fractions, the extra dose to red bone marrow is about 0.3Gy, which is much smaller than the dose reduction to bone marrow achievable by a bone marrow sparing IMRT technique (IM-WPRT) [111]. Many clinics do not 95  employ bone marrow sparing IMRT, indicating that hematologic toxicity is not a clinical impediment to adopting CBCT for quality assurance in this patient group. The BEIR VII committee has tabulated lifetime attributable risk of cancer incidence; those data can be used to estimate the increased risk due to CBCT for organs in peripheral regions where doses are less than 1 Gy. For organs in high dose regions, the BEIR VII data are not applicable. For a 50 year old woman, the risk of breast cancer is estimated to be 70 cases per 100,000 persons exposed to a single dose of 0.1 Gy. For 25 CBCT scans, dose to breast is 0.5cGy, which will lead to an estimated increase of 35 cased per 100,000 persons undergoing this procedure. (Average incidence of breast cancer for a 50 year old woman is 11.15% [112]). Breast is one of the more radiation sensitive organs so serves as a good example here. Increased cancer risk for liver and lung can be derived using BEIR VII data in a similar way. To put these numbers in perspective it is important to consider that radiation induced cancer has a fairly long latency period for most organs (approximately 15 years) and the fractionated delivery of dose in a daily CBCT regimen are also not accounted for. Additionally, the benefits from CBCT are likely to improve quality of life following gynaecological RT treatment and possibly extend life by ensuring adequate target coverage. Fuller exploration of these issues is beyond the scope of this work, however the data provided here may form the basis for further work in this area. The balance of risk and benefit is a somewhat individualized issue and the goal here was to provide the health care professional with reliable data on organ equivalent doses such that better informed discussions would be possible. 96  A growing number of studies confirm that inter and intra-fractional organ motion is a concern in gynaecological IMRT treatments. A study using weekly magnetic resonance imaging and deformable registration showed that variations in bladder and rectum filling may lead to substantial variation in doses to rectum, sigmoid and bladder compared with planned doses [113]. The reported variation in planned versus delivered dose to these organs is in some cases orders of magnitude greater than CBCT doses reported here. In another study, daily kV imaging has shown the importance of daily monitoring of CTV coverage since occasionally large and unpredictable shifts in CTV position may occur [114]. Another recent study suggests that adaptive radiotherapy strategies based on daily imaging may benefit gynaecological patients [115]. All of these studies, together with the results of the current study suggest that daily-kV CBCT imaging may play an important role in quality assurance of radiotherapy treatment for gynaecological patients and that doses delivered by kV CBCT are inconsequential in relation to improvement in treatment that may result from its appropriate use.  6.5  Conclusion  This study employed Monte Carlo methodology to investigate dose in IMRT and RapidArc treatments, with and without daily CBCT imaging. Peripheral doses from CBCT are, at the most, on the order of linac leakage doses for RapidArc treatments in gynaecological cancer. For organs near the PTV, the additional organ equivalent dose associated with CBCT imaging is less than the inter-patient variation in dose to these organs and within the cur-  97  rent accepted limits for uncertainty in dose delivery. This dosimetric information should be considered when balancing the potential risks and benefits associated with adopting CBCT imaging protocols for quality assurance of IMRT and RapidArc treatments.  98  Chapter 7  Conclusions and Future Work This thesis has explored potential utilization of CBCT and organ equivalent doses associated with daily kV CBCT imaging for gynaecological patients undergoing IMRT. Utilization of CBCT for this group of patients is expected to increase due to unavoidable organ motion observed from day to day during treatment. Occasionally, large variations in bladder volume, the posterior bladder wall position and anterior rectal wall position can be observed, potentially compromising target coverage. A Monte Carlo simulation system for calculating patient specific kV CBCT doses has been built and carefully benchmarked. Discrepancy of less than 2% between simulation and measurement has been achieved. With this system, different scan protocols for different anatomical sites can be simulated and patient-specific dose can be calculated. It has demonstrated that the doses to soft tissue range from 1 to 5cGy per scan and doses to bony structure are 3 to 4 times higher due to the photoelectric effect. Further application to other patient populations such as pediatrics is also feasible. The microdosimetric quantity dose mean lineal energy was investigated  99  to compensate for the differences in relative biological effectiveness between kV and MV radiation. C++ code was developed to compute lineal energy in a sphere of 1 µm diameter by homogenously distributing electron tracks in a cube of dimension equal to twice the largest electron range. Quality factors for different organs based on dose mean lineal energy have been found in the range of of 1.31-1.35. Because the variation of quality factors is small, a quality factor of 1.3 is recommended for the purpose of organ equivalent dose estimates. Experimental validation of this finding is the subject of future work. Combining patient specific doses and quality factors, organ equivalent doses were compared with IMRT/RapidArc scatter and linac leakage dose for gynecological patients. Peripheral doses from CBCT are, at the most, on the order of linac leakage doses for RapidArc treatments in gynecological cancer. For organs near the PTV, the additional organ equivalent dose associated with CBCT imaging is less than the inter-patient variation in dose to these organs and within the current accepted limits for uncertainty in dose delivery. This comprehensive dosimetric information may be used by clinicians to balance the potential risks and benefits associated with adopting CBCT imaging protocols for quality assurance of IMRT and RapidArc treatments. This methodology could be extended to other treatment protocols. With the assurance that daily kV CBCT imaging does not put these patients at elevated risk of radiation induced secondary cancer, clinicians may now be more comfortable applying this technology. Further reduction in the volume of normal tissue being irradiated to high doses may be possible 100  while ensuring that target volumes continue to receive the prescribed dose, if daily CBCT can be applied to adapt daily treatment to patient specific needs. Finally, quantification of absolute risk of secondary cancer induction from imaging and treatment procedures remains a difficult task. For doses above 4Gy, the three OED models explored here differ widely and interpretation of absolute risk is not recommended. This highlights the need to better understand the biological mechanisms involved in cellular response to radiation in order to effectively model OED, particularly for large doses and fractionated treatment. As we develop more understanding of radiobiology in the future, this dose information will become more directly applicable in quantifying risk of radiation-induced second cancers.  101  Bibliography [1] H. E. Johns and J. R. Cunningham. Physics of Radiology. Charles C Thomasm Springfield, IL, 4th edition, 1983. [2] E. J. Hall and A. J. Giaccia. Radiobiology for the Radiologist. Lippincott Williams and Wilkins Publishing, 6th edition, 2006. [3] Jiang Hsieh. Computed Tomography: Principles, Design, Artifacts, and Recent Advances. SPIE Publications, 2nd Edition, 2009. [4] C. J. Karzmark , Craig S. Nunan and Eiji Tanabe. Medical Electron Accelerators. McGraw-Hill, New York, NY. First edition, 1993. [5] Ezzell GA, Galvin JM, Low D, Palta JR, Rosen I, Sharpe MB, Xia P, Xiao Y, Xing L, Yu CX. Guidance document on delivery, treatment planning, and clinical implementation of IMRT: report of the IMRT Subcommittee of the AAPM Radiation Therapy Committee. Med Phys. 30(8):2089-115, 2003. [6] M. Oliver, W. Ansbacher and W. A. Beckham. Comparing planning time, delivery time and plan quality for IMRT, RapidArc and tomotherapy. J Appl Clin Med Phys, 10(4): 3068, 2009.  102  [7] Ong CL, Verbakel WF, Cuijpers JP, Slotman BJ, Lagerwaard FJ, Senan S. Stereotactic radiotherapy for peripheral lung tumors: A comparison of volumetric modulated arc therapy with 3 other delivery techniques. Radiother Oncol, 97(3): 437-442, 2010. [8] Otto K. Volumetric modulated arc therapy: IMRT in a single gantry arc. Med Phys, 35(1): 310-7, 2008. [9] Uematsu M, Fukui T, Shioda A, Tokumitsu H, Takai K, Kojima T, Asai Y and Kusano S. A dual computed tomography linear accelerator unit for stereotactic radiation therapy: A new approach without cranially fixated stereotactic frames. Int J Radiat Oncol Biol Phys, 35(3): 587592, 1996. [10] Kuriyama K, Onishi H, Sano N. A new irradiation unit constructed of self-moving gantry-CT and linac. Int J Radiat Oncol Biol Phys, 55(2): 428-35, 2003. [11] Jaffray D A, Drake D G, Moreau M, Martinez A A and Wong J W. A radiographic and tomographic imaging system integrated into a medical linear accelerator for localization of bone and soft-tissue targets. Int J Radiat Oncol Biol Phys 45: 773-89, 1999. [12] Jaffray D A and Siewerdsen J H. Cone-beam computed tomography with a flat-panel imager: initial performance characterization. Med Phys 27: 1311-23, 2000. [13] Jaffray D A, Siewerdsen J H, Wong J W and Martinez A A Flat-panel  103  cone-beam computed tomography for image-guided radiation therapy. Int J. Radiat Oncol Biol Phys 53: 1337-49, 2002. [14] Simpson DR, Lawson JD, Nath SK, Rose BS, Mundt AJ, Mell LK. A survey of image-guided radiation therapy use in the United States. Cancer, 116(16): 3953-60, 2010. [15] Moseley DJ, White EA, Wiltshire KL, Rosewall T, Sharpe MB, Siewerdsen JH, Bissonnette JP, Gospodarowicz M, Warde P, Catton CN, Jaffray DA.Comparison of localization performance with implanted fiducial markers and cone-beam computed tomography for on-line image-guided radiotherapy of the prostate. Int J Radiat Oncol Biol Phys.,67(3):942-53, 2007. [16] T. G. Purdie, D. J. Moseley, J. P. Bissonnette, M. B. Sharpe, K. Franks, A. Bezjak, and D. A. Jaffray. Respiration correlated cone-beam computed tomography and 4DCT for evaluating target motion in Stereotactic Lung Radiation Therapy. Acta Oncol. 45: 915-22, 2006. [17] J. J. Sonke, M. Rossi, J. Wolthaus, M. van Herk, E. Damen, and J. Belderbos. Frameless stereotactic body radiotherapy for lung cancer using four-dimensional cone beam CT guidance. Int. J. Radiat. Oncol. Biol. Phys., 74: 567-574, 2009. [18] On-Board Imager (OBI) Reference Guide (P/N B501007R01B), Version 1.4, Varian Medical Systems, April 2008. [19] Martin Nikl. Scintillation detectors for x-rays. Meas. Sci. Technol. 17:R37-R54, 2006. 104  [20] Antonuk LE. Electronic portal imaging devices: a review and historical perspective of contemporary technologies and research. Phys Med Biol., 21;47(6):R31-65, 2002. [21] www.varian.com/media/xray/products/pdf/Flat Panel Xray Imaging 2011-11-04.pdf [22] Mail N, Moseley DJ, Siewerdsen JH, Jaffray DA. The influence of bowtie filtration on cone-beam CT image quality. Med Phys, 36(1): 22-32, 2009. [23] Seung Wook Lee and Ge Wang. A Grangeat-type half-scan algorithm for cone-beam CT. Med Phys, 30(4): 689-700 2003. [24] G. Jarry, S.A. Graham, D.J. Moseley , D.J. Jaffray, J.H. Siewerdsen and F. Verhaegen. Characterization of scattered radiation in kV CBCT images using Monte Carlo simulations. Med Phys, 33(11): 4320-9,2006. [25] P.C. Johns and M. Yaffe. Scattered radiation in fan beam imaging systems. Med Phys, 9(2): 231-9,1982. [26] C. B. Saw, Y. Yang, F. Li, N. J. Yue, C. Ding, K. Komanduri, S. Huq, and D. E. Heron. Performance characteristics and quality assurance aspects of kilovoltage cone-beam CT on medical linear accelerator. Med Dosim 32: 80-85, 2000. [27] R.R. Liu, K.L. Prado and D. Cody. Optimal acquisition parameter selection for CT simulators in radiation oncology. J Appl Clin Med Phys, 9(4): 2878, 2008. 105  [28] Sonke JJ, Zijp L, Remeijer P, van Herk M. Respiratory correlated cone beam CT. Med Phys, 32(4):1176-86, 2005. [29] Yang Y, Schreibmann E, Li T, Wang C, Xing L. Evaluation of onboard kV cone beam CT (CBCT)-based dose calculation. Phys Med Biol, 52(3): 685-705, 2007. [30] Yoo, S., and F.-F. Yin. Dosimetric feasibility of cone beam CT-based treatment planning compared to CT-based treatment planning. Int J Radiat Oncol Biol Phys 66(5): 1553-1561, 2007. [31] L´ etourneau D, Wong R, Moseley D, Sharpe MB, Ansell S, Gospodarowicz M, Jaffray DA. Online planning and delivery technique for radiotherapy of spinal metastases using cone-beam CT: image quality and system performance. Int J Radiat Oncol Biol Phys.,67(4):1229-37, 2007. [32] Donald A. Pierce and Dale L. Preston. Radiation-Related Cancer Risks at Low Doses among Atomic Bomb Survivors. Radiat Res, 154: 178-186 (2000) [33] Donald A. Pierce et.al. Effect of recent changes in atomic bomb survivor dosimetry on cancer mortality risk estimates. Radiat Res, 162(4): 37789, 2004. [34] Donald A. Pierce, Yukiko Shimizu, Dale L. Preston, Michael Vaeth, Kiyohiko Mabuchi. Studies of the Mortality of Atomic Bomb Survivors. Report 12, Part I. Cancer: 1950-1990. Radiat Res, 146(1):1-27, 1996. [35] Dale L. Preston, Yukiko Shimizu, Donald A. Pierce et al. Radiation 106  Effects Research Foundation, Life Span Study Report 13: Solid cancer and noncancer disease mortality: 1950-1997. Radiat Res, 160(4): 381407, 2003 [36] Kotaro Ozasa, Yukiko Shimizu, Akihiko Suyama et al. Studies of the Mortality of Atomic Bomb Survivors, Report 14, 1950C2003: An Overview of Cancer and Noncancer Diseases. Radiat Res, 177: 229243, 2012. [37] Little M P and Muirhead C R. Absence of evidence for threshold departures from linearCquadratic curvature inthe Japanese A-bomb cancer incidence and mortality data Int J Low Radiat, 1: 242-55, 2004. [38] Hancock SL, Hoppe RT. Long-term complications of treatment and causes of mortality after Hodgkins disease. Semin Radiat Oncol 6: 225242, 1996. [39] Swerdlow AJ, Barber JA, Hudson GV, et al. Risk of second malignancy after Hodgkins disease in a collaborative British cohort: The relation to age at treatment. J Clin Oncol, 18: 498-509, 2000. [40] van Leeuwen FE, Klokman WJ, Veer MB, et al. Long-term risk of second malignancy in survivors of Hodgkins disease treated during adolescence or young adulthood. J Clin Oncol, 18: 487-497, 2000. [41] Chaturvedi AK , Engels EA , Gilbert ES, et al. Second cancers among 104 760 survivors of cervical cancer: evaluation of long-term risk. J Natl Cancer Inst, 99(21) : 1634-1643, 2007 .  107  [42] Kleinerman RA, Boice JD Jr, Storm HH, Sparen P,et al. Second primary cancer after treatment for cervical cancer. An international cancer registries study. Cancer, 76(3):442-52, 1995. [43] Boice JD Jr, Engholm G, Kleinerman RA,et al. Radiation dose and second cancer risk in patients treated for cancer of the cervix. Radiat Res, 116(1): 3-55, 1988. [44] Cardis E, Gilbert ES, Carpenter L, et al. Effects of low doses and low dose rates of external ionizing radiation: cancer mortality among nuclear industry workers in three countries. Radiat Res, 142(2): 117-32, 1995. [45] Yoshinaga S, Mabuchi K, Sigurdson AJ, Doody MM, Ron E. Cancer Risks among Radiologists and Radiologic Technologists: Review of epidemiologic studies. Radiology, 233(2): 313-21, 2004. [46] John B. Little. Radiation carcinogenesis. Carcinogenesis, 21 (3): 397404, 2000. [47] Committee to Assess Health Risks from Exposure to Low Levels of Ionizing Radiation, National Research Council. Health risks from exposure to low levels of ionizing radiation: BEIR VII phase 2. Washington, DC: National Academies, 2006. [48] Schneider U, Zwahlen D, Ross D, Kaser-Hotz B. Estimation of radiation induced cancer from three-dimensional dose distributions: Concept of organ equivalent dose. Int J Radiat Oncol Biol Phys, 61: 1510-5, 2005.  108  [49] Cancer Incidence in Canada 2007 and 2008, Health Statistics Division, Statistics Canada, Ottawa, Ontario, 2010 [50] Cleaves M. Radium: with a preliminary note on radium rays in the treatment of cancer. Med Rec 64: 1719-23, 1903. [51] Ortin C. G. and Wolf-Rosenblum, S. Dose dependence of complication rates in cervix cancer radiotherapy. Int J Radiat Oncol Biol Phys, 12:3744, 1986. [52] Perez C. A., Breaux S., Bedwinek J. M., Madoc-Jones H., Camel H. M., Purdy J. A. and Walz B. Radiation therapy alone in the treatment of carcinoma of the uterine cervix: II. Analysis of complications. Cancer 54:235-46; 1984. [53] Pourquier H, Delard R, Achille E, Daly NJ, Horiot JC, Keiling R, Pigneux J, Rozan R, Schraub S, Vrousos C. A quantified approach to the analysis and prevention of urinary complications in radiotherapeutic treatment of cancer of the cervix. Int J Radiat Oncol Biol Phys, 13(7):1025-33, 1987. [54] Viswanathan AN, Yorke ED, Marks LB, Eifel PJ, Shipley WU. Radiation dose-volume effects of the urinary bladder. Int J Radiat Oncol Biol Phys,76(3 Suppl):S116-22, 2010. [55] Mell LK, Kochanski JD, Roeske JC, et al. Dosimetric predictors of acute hematologic toxicity in cervical cancer patients treated with concurrent cisplatin and intensity-modulated pelvic radiotherapy. Int J Radiat Oncol Biol Phys, 66: 1356-65, 2006. 109  [56] Small W Jr, Mell LK, Anderson P et al., Consensus Guidelines for the Delineation of the Intensity Modulated Pelvic Radiotherapy CTV in the Postoperatice Treatment of Endometrial and Cervical Cancer. Int J Radiat Oncol Biol Phys 71(2):428-34, 2008. [57] Magill’s Medical Guide Volume III. Englewood Cliffs, NJ: Salem, 1998. [58] Haripotepornkul NH, Nath SK, Scanderbeg D, Saenz C, Yashar CM. Evaluation of intra- and inter-fraction movement of the cervix during intensity modulated radiation therapy. Radioth Oncol, 98: 347-51, 2011. [59] Tyagi N, Lewis JH, Yashar CM, et al., Daily online cone beam computed tomography to assess interfractional motion in patients with intact cervical cancer. Int J Radit Onco Biol Phys 280: 273-80, 2010. [60] J¨ urgenliemk-Schulz IM, Toet-Bosma MZ, de Kort GA et al., Internal motion of the vagina hysterectomy for gynaecological cancer. Radiother Oncol, 98: 244-8, 2011. [61] Kerkhof EM, Raaymakers BW, van der Heide UA, van de Bunt L, Jrgenliemk-Schulz IM, Lagendijk JJ. Online MRI guidance for healthy tissue sparing in patients with cervical cancer: an IMRT planning study. Radiother Oncol, 88(2): 241-9, 2008. [62] D. W. O. Rogers, Fifty years of Monte Carlo simulations for medical physics. Phys Med Biol, 51, R287-R301, 2006. [63] I. Kawrakow and D.W.O. Rogers. The EGSnrc Code System: Monte 110  Carlo Simulation of Electron and Photon Transport. National Research Council of Canada, Ottawa, Canada. PIRS-701 [64] D.W.O. Rogers, B. Walters and I. Kawrakow. BEAMnrc user’s manual. NRCC Report PIRS-0509(A)revL. [65] B. Walters, I. Kawrakow and D.W.O. Rogers. DOSXYZnrc user’s manual. NRCC Report PIRS-794revB. [66] Semenenko, V. A., Turner, J. E. and Borak, T. B. NOREC, a Monte Carlo code for simulating electron tracks in liquid water. Radiat Environ Biophys 42: 213-217, 2003. [67] Ma CM, Coffey CW, DeWerd LA, Liu C, Nath R, Seltzer SM et al. AAPM protocol for 40-300 kV x-ray beam dosimetry in radiotherapy and radiobiology. Med Phys, 28(6): 868-93, 2001. [68] G. Ding and P. Munro. Reduced CBCT imaging dose due to the new X-ray source in TrueBeamTM . Med Phys, 38: 3372-3, 2011. [69] G. X. Ding, D. M. Duggan, and C. W. Coffey, Characteristics of kilovoltage x-ray beams for cone-beam computed tomography in radiation therapy, Phys Med Biol, 52: 1595-1615, 2007. [70] G. X. Ding, D. M. Duggan, and C. W. Coffey. Accurate patient dosimetry of kilovoltage cone-beam CT in radiation therapy. Med Phys 35: 1135-1144, 2008. [71] P. Alaeia, G. Ding, and H. Guan. Inclusion of the dose from kilovoltage  111  cone beam CT in the radiation therapy treatment plans. Med Phys 37: 244-210, 2010. [72] Downes P, Jarvis R, Radu E, Kawrakow I, Spezi E. Monte Carlo simulation and patient dosimetry for a kilovoltage cone-beam CT unit. Med Phys, 36(9): 4156-67, 2009. [73] N. Wen, H. Guan, R. Hammoud, D. Pradhan, T. Nurushev, S. Li, and B. Movsas. Dose delivered from Varians CBCT to patients receiving IMRT for prostate cancer. Phys Med Biol 52: 2267-2276, 2007. [74] W.K. Kan, H.T. Leung, W. Wong, et al, Radiation dose from cone beam computed tomography for image-guided radiation therapy. Int J Radiation Oncology Biol Phys, 70: 272-279, 2008. [75] Hyer DE, Serago CF, Kim S, Li JG, Hintenlang DE. An organ and effective dose study of XVI and OBI cone-beam CT systems. J Appl Clin Med Phys, 11(2): 3183, 2010. [76] Perks JR, Lehmann J, Chen AM, Yang CC, Stern RL, Purdy JA. Comparison of peripheral dose from image-guided radiation therapy (IGRT) using kV cone beam CT to intensity-modulated radiation therapy (IMRT). Radiother Oncol, 89(3): 304-10, 2008. [77] Hyer DE, Hintenlang DE. Estimation of organ doses from kilovoltage cone-beam CT imaging used during radiotherapy patient position verification. Med Phys, 37(9): 4620-6, 2010. [78] Pawlowski JM, Ding GX. A new approach to account for the medium112  dependent effect in model-based dose calculations for kilovoltage x-rays. Phys Med Biol, 56(13): 3919-34, 2011. [79] Lloyd, D.C. and Edwards, A.A. Chromosome aberrations in human lymphocytes: effect of radiation quality,dose, and dose rate. In: Radiation Induced Chromosome Damage in Man. Ishihara, T. and Sasaki, M.S., Eds., (New York: Alan R. Liss) pp. 23-49, 1983. [80] Mestres M, Caball´ın M R, Barrios L, Ribas M and Barquinero J F. RBE of x rays of different energies: a cytogenetic evaluation by FISH Radiat. Res. 170: 93-100, 2008. [81] Schmid E, Regulla D, Kramer HMand Harder D. The effect of 29 kV x rays on the dose response of chromosome aberrations in human lymphocytes Radiat. Res. 158 771-7, 2002. [82] International Commission on Radiological Protection (ICRP). ICRP PUBLICATION 92 Relative Biological Effectiveness (RBE), QualityFactor (Q), and Radiation Weighting Factor (ωR ). Annals of the ICRP 33: 4, 2003. [83] van Gent DC, Hoeijmakers JH, Kanaar R. Chromosomal stability and the DNA double-stranded break connection. Nat Rev Genet. 2(3):196206, 2001. [84] ICRU/ICRP RBE Committee ”Report of the RBE Committee to the International COmmission on Radiological Protection and International Commission on Radiation Units and Measurements,” Helath Physics 9, 357, 1963. 113  [85] The Quality Factor in Radiation Protection, ICRU Report 40. International Commission on Radiation Units and Measurements, Betheda, MD. [86] Rossi,H.H. and Rosenzweig, W. A device for the measurement of dose as a function of specific ionization. Radiology 64: 404, 1955. [87] Rossi,H.H. Specification of radiation quality. Radiat Res 20: 522, 1959. [88] Rossi,H.H. Spatial distribution of energy deposition by ionizing radiation. Radiat Res Suppl.2: 290, 1960. [89] Rossi,H.H. The role of microdosimetry in raidobiology and radiation protection. page 1 in Proceedings of the Third Symposium on Microdosimetry, Report No. EUR 4810, Ebert, H.G. Ed. [90] ICRU: Radiation quantities and units. ICRU Report 33, Washington, 1980. [91] Microdosimetry. ICRU Report 36, 1983. [92] Moorhead PS, Nowell PC, Mellman WJ, Battips DM, Hungerfood DA: Chromosome preparations of leukocytes cultured form huam peripheral blood. Exp Cell Res 20: 613, 1960 [93] Schmid E and Bauchinger M. LET dependence of dicentric yields in human lymphocytes induced by low doses of sparsely ionizing radiations and its implications for risk assessments. Health Phys 74: 719-21, 1998. [94] Schmid E, Krumrey M, Ulm G, Roos H and Regulla D. The maximum low dose RBE of 17,4 and 40 keV monochromatic x rays for the 114  induction of dicentric chromosomes in human peripheral lymphocytes Radiat. Res, 160: 499-504, 2003 [95] Schmid E. Response to the comment on Re-evaluation of the RBE of 29 kV x-rays (mammography x-rays)relative to 220 kV x-rays using neoplastic transformation of human CGL1-hybrid cells, by W G¨oggelmann, C Jacobsen, W Panzer, L Walsh, H Roos, E Schmid (2003) Radiat. Environ. Biophys. 42 175-182 Radiat Environ Biophys, 43:61-2, 2004 [96] W G¨oggelmann, C Jacobsen, W Panzer, L Walsh, H Roos, E SchmidReevaluation of the RBE of 29 kV x-rays (mammography x-rays)relative to 220 kV x-rays using neoplastic transformation of human CGL1-hybrid cells. Radiat Environ Biophys, 42: 175-182, 2003. [97] Recommendations of the International Commission on Radiological Protection. ICRP Publication 60, Ann. ICRP 21(1-3). [98] National Council on Radiation Protection and Measurements Ionizing. Radiation Exposure of the Population of the United States. Bethesda, MD: NCRP Report No. 160; 2009. [99] Brink JA, Amis ES Jr, Image Wisely: A Campaign to increase awareness about adult radiation protection, Radiology, 257: 601-2, 2010. [100] Murphy MJ, Balter J, Balter S, et al. The management of imaging dose during image-guided radiotherapy: Report of the AAPM Task Group 75. Med Phys, 34: 4041-63, 2007. [101] Gu J, Bednarz B, Xu XG, Jiang SB. Assessment of patient organ doses 115  and effective doses using the VIP-Man adult male phantom for selected cone-beam CT imaging procedures during image guided radiation therapy. Radiat Prot Dosimetry, 131(4): 431-43, 2008. [102] Kim S, Yoshizumi TT, Frush DP, Toncheva G, Yin FF. Radiation dose from cone beam CT in a pediatric phantom: risk estimation of cancer incidence. AJR Am J Roentgenol, 194: 186-190, 2010. [103] Zwahlen DR, Ruben JD, Jones P, Gagliardi F, Millar JL, Schneider U. Effect of intensity-modulated pelvic radiotherapy on second cancer risk in the postoperative treatment of endometrial and cervical cancer. Int J Radiat Oncol Biol Phys, 74: 539-45, 2009. [104] Palm A, Nilsson E, Herrnsdorf L. Absorbed dose and dose rate using Varian OBI 1.3 and 1.4 CBCT system. J Appl Clin Med Phys, 11: 229-40, 2010. [105] Howell RM, Scarboro SB, Kry SF, Yaldo Z. Accuracy of out-of-field dose calculations by a commercial treatment planning system. Phys Med Biol 2010;55: 6999-7008. [106] Siebers JV, Keall PJ, Kim JO, Mohan R. A method for photon beam Monte Carlo multileaf collimator particle transport. Phys Med Biol, 47: 3225-49, 2002. [107] Popescu IA, Shaw CP, Zavgorodni SF, Beckham WA. Absolute dose calculations for Monte Carlo simulations of radiotherapy beams. Phys Med Biol, 50: 3375-92, 2005.  116  [108] Teke T, Bergman AM, Kwa W, Gill B, Duzenli C, Popescu IA. Monte Carlo based, patient-specific RapidArc QA using Linac log files. Med Phys, 37(1): 116-23, 2010. [109] Lobo J, Popescu IA. Two new DOSXYZnrc sources for 4D Monte Carlo simulations of continuously variable beam configurations, with applications to RapidArc, VMAT, TomoTherapy and CyberKnife. Phys Med Biol, 55: 4431-43, 2010. [110] Walters BR, Ding GX, Kramer R, Kawrakow I. Skeletal dosimetry in cone beam computed tomography. Med Phys, 36: 2915-22, 2009. [111] Lujan AE, Mundt AJ, Yamada SD, Rotmensch J, Roeske JC. Intensity-modulated radiotherapy as a means of reducing dose to bone marrow in gynecologic patients receiving whole pelvic radiotherapy. Int J Radiat Oncol Biol Phys, 57: 516-21, 2003. [112] SEER Cancer Statistics Review, 1975-2009. Bethesda, MD: National Cancer Institute; 2011. [113] Lim K, Kelly V, Stewart J, et al. Pelvic radiotherapy for cancer of the cervix: is what you plan actually what you deliver? Int J Radiat Oncol Biol Phys, 74: 304-12, 2009. [114] Haripotepornkul NH, Nath SK, Scanderbeg D, Saenz C, Yashar CM. Evaluation of intra- and inter- fraction movement of the cervix during intensity modulated radiation therapy. Radiother Oncol, 98: 347-51, 2011.  117  [115] Ahmad R, Hoogeman MS, Bondar M, et al. Increasing treatment accuracy for cervical cancer patients using correlations between bladderfilling change and cervix-uterus displacements: Proof of principle. Radiother Oncol, 98: 340-6, 2011.  118  


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