<%BANNER%>

Organ Dose Measurements from Multiple-Detector Computed Tomography Using a Commercial Dosimetry System and Tomographic, ...

Permanent Link: http://ufdc.ufl.edu/UFE0024870/00001

Material Information

Title: Organ Dose Measurements from Multiple-Detector Computed Tomography Using a Commercial Dosimetry System and Tomographic, Physical Phantoms
Physical Description: 1 online resource (201 p.)
Language: english
Creator: Lavoie, Lindsey
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: dose, dosimetry, luminescence, mdct, optically, organ, osl, phantom, stimulated
Nuclear and Radiological Engineering -- Dissertations, Academic -- UF
Genre: Nuclear Engineering Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The technology of computed tomography (CT) imaging has soared over the last decade with the use of multi-detector CT (MDCT) scanners that are capable of performing studies in a matter of seconds. While the diagnostic information obtained from MDCT imaging is extremely valuable, it is important to ensure that the radiation doses resulting from these studies are at acceptably safe levels. This research project focused on the measurement of organ doses resulting from modern MDCT scanners. A commercially-available dosimetry system was used to measure organ doses. Small dosimeters made of optically-stimulated luminescent (OSL) material were analyzed with a portable OSL reader. Detailed verification of this system was performed. Characteristics studied include energy, scatter, and angular responses; dose linearity, ability to erase the exposed dose and ability to reuse dosimeters multiple times. The results of this verification process were positive. While small correction factors needed to be applied to the dose reported by the OSL reader, these factors were small and expected. Physical, tomographic pediatric and adult phantoms were used to measure organ doses. These phantoms were developed from CT images and are composed of tissue-equivalent materials. Because the adult phantom is comprised of numerous segments, dosimeters were placed in the phantom at several organ locations, and doses to select organs were measured using three clinical protocols: pediatric craniosynostosis, adult brain perfusion and adult cardiac CT angiography (CTA). A wide-beam, 320-slice, volumetric CT scanner and a 64-slice, MDCT scanner were used for organ dose measurements. Doses ranged from 1 to 26 mGy for the pediatric protocol, 1 to 1241 mGy for the brain perfusion protocol, and 2-100 mGy for the cardiac protocol. In most cases, the doses measured on the 64-slice scanner were higher than those on the 320-slice scanner. A methodology to measure organ doses with OSL dosimeters received from CT imaging has been presented. These measurements are especially important in keeping with the ALARA (as low as reasonably achievable) principle. While diagnostic information from CT imaging is valuable and necessary, the dose to patients is always a consideration. This methodology aids in this important task
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Lindsey Lavoie.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Arreola, Manuel M.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024870:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024870/00001

Material Information

Title: Organ Dose Measurements from Multiple-Detector Computed Tomography Using a Commercial Dosimetry System and Tomographic, Physical Phantoms
Physical Description: 1 online resource (201 p.)
Language: english
Creator: Lavoie, Lindsey
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: dose, dosimetry, luminescence, mdct, optically, organ, osl, phantom, stimulated
Nuclear and Radiological Engineering -- Dissertations, Academic -- UF
Genre: Nuclear Engineering Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The technology of computed tomography (CT) imaging has soared over the last decade with the use of multi-detector CT (MDCT) scanners that are capable of performing studies in a matter of seconds. While the diagnostic information obtained from MDCT imaging is extremely valuable, it is important to ensure that the radiation doses resulting from these studies are at acceptably safe levels. This research project focused on the measurement of organ doses resulting from modern MDCT scanners. A commercially-available dosimetry system was used to measure organ doses. Small dosimeters made of optically-stimulated luminescent (OSL) material were analyzed with a portable OSL reader. Detailed verification of this system was performed. Characteristics studied include energy, scatter, and angular responses; dose linearity, ability to erase the exposed dose and ability to reuse dosimeters multiple times. The results of this verification process were positive. While small correction factors needed to be applied to the dose reported by the OSL reader, these factors were small and expected. Physical, tomographic pediatric and adult phantoms were used to measure organ doses. These phantoms were developed from CT images and are composed of tissue-equivalent materials. Because the adult phantom is comprised of numerous segments, dosimeters were placed in the phantom at several organ locations, and doses to select organs were measured using three clinical protocols: pediatric craniosynostosis, adult brain perfusion and adult cardiac CT angiography (CTA). A wide-beam, 320-slice, volumetric CT scanner and a 64-slice, MDCT scanner were used for organ dose measurements. Doses ranged from 1 to 26 mGy for the pediatric protocol, 1 to 1241 mGy for the brain perfusion protocol, and 2-100 mGy for the cardiac protocol. In most cases, the doses measured on the 64-slice scanner were higher than those on the 320-slice scanner. A methodology to measure organ doses with OSL dosimeters received from CT imaging has been presented. These measurements are especially important in keeping with the ALARA (as low as reasonably achievable) principle. While diagnostic information from CT imaging is valuable and necessary, the dose to patients is always a consideration. This methodology aids in this important task
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Lindsey Lavoie.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Arreola, Manuel M.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024870:00001


This item has the following downloads:


Full Text

PAGE 1

1 ORGAN DOSE MEASUREMENTS FROM MULTIPLE-DETECTOR COMPUTED TOMOGRAPHY USING A COMMERCIAL DOSI METRY SYSTEM AND TOMOGRAPHIC, PHYSICAL PHANTOMS By LINDSEY K LAVOIE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Lindsey K Lavoie

PAGE 3

3 To my Grandma, from whom I got all of my brains

PAGE 4

4 ACKNOWLEDGEMENTS I would like to express my gratitude f irst and foremost to my advisor, Dr. Manuel Arreola. The opportunities he has given me in my five year s of graduate education at the University of Florida and of clinical experience working at Shands at UF are tr uly invaluable and ones that I will take with me after graduation. His en couragement and positive reinforcement throughout this entire process will always be appreciated. To my other two advisors, I am grateful for all they have taught me during my time at UF. Dr. Libby Brateman pushed me to always do my best and constantly challenged me with her essential questions and insightful critiques. He r demand for only high-leve l work is one that I hope to demand of myself in years to come. Dr. Lynn Rill was a source of constant encouragement; she provided a sounding board when I needed a second set of ears. Her door was always open to me for research and clinical questions, as well as those that were a part of life outside Shands, and fo r that I am thankful. There are several others that I would like to acknowledge as helping me along this journey. Dr. David Hintenlang and Dr. Sc ott Banks provided me guidance and counseling throughout this research project. I am grateful to all of the radiologists involved in the RPC. Because of them I have gained a clinical knowledge especially in CT, that complements everything I have learned during these five years of res earch. The CT technologists at Shands answered my countless clinical questions which helped give me a true understanding of CT protocols and how they are used clinically, and they were always gracious when answering my re dundant question, Is CT4 available? For this, I am grateful. I would also like to thank ev eryone in the QC/GA office for the small things they do on a daily basis. Jen Sirera has helped me to unde rstand aspects of a radiology department that go beyond medical physics. Her constantly positive demeanor, despite the everyday stresses of her

PAGE 5

5 job, is one I hope to emulate in the professional world. Carly Williams taught me most of what I have learned about testing x-ray equipment. Bo Hartmanns attention to detail is one that I admire and to which I strive to be better. To m Griglock has assumed most of the responsibility of the GAs, an experience I only hope he can appr eciate when he begins applying for jobs. For all of these people, I am thankf ul. And finally, I would like to express my sincere thanks to Monica Ghita, without whom I could not have comp leted this project. She worked late nights with me, struggling to place dots into the phantom and meticulously aligning the phantom in the scanner. The conversations I had with her are ones that helped sh aped this research into a final product; I could not have done it without her. On a more personal note, I would like to tha nk my Gator friends for their support over the past five years. I would especially like to th ank Duane Ammons for his love and encouragement, particularly in these last few months; he pushed me to keep going and was there to pick me up when I was down. To my colleagues that started at UF with me in 2004, especially Ryan Fisher, Bob Ambrose, and Keelan Seabolt, I would not ha ve survived my first year down here without them. These people, among many others, became my family away from home, and for that I will be forever grateful. To my small family at home, Ed Podesky, Kathy Podesky, and Steve and Connor DeCosta, their support in my lengthy education made the tr ansition to UF as seamless as it could have possibly been and for this I am grateful. Over these past five years, Dr. Rich Maucer i has provided me support in as many ways as he possibly could, and for this I am extremely tha nkful. His excitement for me joining the field of radiology has been contagious and I look forw ard to scholarly discussions with him in the Lounge.

PAGE 6

6 And finally, I want to extend my deepest gratitude to my mom, Dona Lavoie. Without her, I would not be the person I am today. The sacr ifices she has made for me over the years have allowed me the successes I have had in my lif e. She truly is the wind beneath my wings.

PAGE 7

7 TABLE OF CONTENTS page ACKNOWLEDGEMENTS .............................................................................................................4TABLE OF CONTENTS ............................................................................................................. ....7LIST OF TABLES .........................................................................................................................12LIST OF FIGURES .......................................................................................................................14ABSTRACT ...................................................................................................................... .............17 CHAP TER 1 INTRODUCTION .................................................................................................................. 191.1 Increased Use of Computed Tomography .................................................................... 191.2 Effects on Pediat ric Patients ......................................................................................... 201.3 Specific Aims ............................................................................................................. ...222 CLINICAL CT PROTOCOLS ............................................................................................... 242.1 Radiology Practice Committee ..................................................................................... 242.2 Protocol Standardization and Revision .........................................................................252.3 Clinical Indications in Protocols ...................................................................................262.4 Picture Archiving and Communications System ..........................................................262.5 Protocol Development Process .....................................................................................282.5.1 Database Development...................................................................................... 282.5.2 Protocol Template ............................................................................................. 282.5.2.1 Title and illustration ............................................................................ 292.5.2.2 Acquisition parameter description table ............................................. 292.5.2.3 Reconstruction parameter tables ......................................................... 292.5.2.4 Reformation parameter table .............................................................. 302.5.2.5 Additional protocol information ......................................................... 302.5.2.6 Indications ........................................................................................... 312.5.2.7 Using the templates a nd updating the protocols ................................. 312.6 Protocol Website .......................................................................................................... .323 MULTIPLE-DETECTOR COMP UTED TOMOGRAPHY ..................................................363.1 Detector Array Systems ................................................................................................363.2 Modes of Acquisition ....................................................................................................373.2.1 Axial Data Acquisition ...................................................................................... 373.2.2 Helical Data Acquisition ...................................................................................373.2.3 Volumetric Data Acquisition ............................................................................ 38

PAGE 8

8 3.3 Clinical CT Considerations ...........................................................................................393.3.1 Slice Thickness.................................................................................................. 393.3.2 Image Reconstruction ....................................................................................... 393.4 Dose-Reduction Techniques ......................................................................................... 403.4.1 Tube Current (mA) Adjustment ........................................................................ 403.4.2 Tube Current (mA) Modulation ........................................................................ 413.4.2.1 Angular (X-Y) tube current (mA) modulation ................................... 423.4.2.2 Z-axis tube current (mA) modulation ................................................. 423.4.3 Tube Voltage ..................................................................................................... 424 COMPUTED TOMOGRAPHY DOSE DESCRI PTORS ......................................................454.1 Definitions ............................................................................................................... ......454.1.1 Kerma ................................................................................................................ 454.1.2 Absorbed Dose (D) ...........................................................................................454.1.3 Equivalent Dose (HT) ........................................................................................ 464.1.4 Effective Dose (E) .............................................................................................464.2 Effective Dose as a Dose Descriptor .............................................................................464.2.1 Applications of Effective Dose ......................................................................... 474.2.2 Limitations of Effective Dose as a Patient Dose Desc riptor in CT ................... 474.2.2.1 Population-based risk .......................................................................... 474.2.2.2 Tissue-weighting factors .....................................................................484.3 Recommendations ......................................................................................................... 485 COMPUTED TOMOGRAPHY DOSE INDEX ....................................................................515.1 CTDI Definitions .......................................................................................................... 515.1.1 MSAD ............................................................................................................... 515.1.2 CTDI ................................................................................................................. 525.1.3 CTDIFDA ............................................................................................................525.1.4 CTDI100 .............................................................................................................535.1.5 CTDIw ...............................................................................................................545.1.6 CTDIvol ..............................................................................................................545.2 Dose-Length Product .................................................................................................... 555.3 Analysis of the Applicability of the Measurement of CTDI and DLP ......................... 565.3.1 Underestimation of Dose using CTDI in MDCT .............................................. 565.3.2 Proposed Solutions ............................................................................................ 575.3.2.1 Extended phantoms ............................................................................. 575.3.2.2 Small-volume ionization chamber ...................................................... 586 INSTRUMENTATION AND PHANTOMS ......................................................................... 616.1 Ionization Chambers ..................................................................................................... 616.1.1 A 6-cc Chamber ................................................................................................616.1.2 A 3-cc Pencil Chamber ..................................................................................... 626.1.3 A 0.6-cc Chamber .............................................................................................626.2 Solid-State Detector ...................................................................................................... 63

PAGE 9

9 6.3 Tube Voltage Meter ......................................................................................................636.4 Dosimetry Phantoms ..................................................................................................... 636.4.1 Stylized Phantoms ............................................................................................. 636.4.2 Tomographic Phantoms .................................................................................... 646.5 Phantom for Image Quality Evaluations ....................................................................... 657 CHARACTERIZATION OF X-RAY BEAMS OF THE VOLUMETRIC CT SYSTEM ....727.1 Exposure Reproducibility .............................................................................................727.1.1 Service Mode .................................................................................................... 727.1.2 Clinical Mode ....................................................................................................737.2 Beam Quality .............................................................................................................. ..737.2.1 Accuracy in kV ................................................................................................. 747.2.2 Beam Filtration in CT ....................................................................................... 747.2.3 Half-Value Layer .............................................................................................. 767.3 Beam Width ................................................................................................................ ..777.4 Dose Profile .............................................................................................................. .....777.4.1 X-Axis ............................................................................................................... 777.4.2 Z-Axis ............................................................................................................... 778 OPTICALLY-STIMULATED LUMINESCENT DOSIMETERS ........................................828.1 Optically-Stimulated Luminescence ............................................................................. 828.1.1 Disadvantages of TL Dosimeters ...................................................................... 838.1.2 Advantages of OSL Dosimeters ........................................................................ 838.2 Characterization of OSL Dosimetry System .................................................................848.2.1 Other Radiation-Measuring Devices ................................................................. 848.2.2 Tracking of System Standards .......................................................................... 858.2.3 Dosimeter Response .......................................................................................... 868.2.4 Room-Light Erasure .......................................................................................... 878.2.5 Energy Response ............................................................................................... 888.2.6 Scatter Response ...............................................................................................888.2.7 Dosimeter Calibration ....................................................................................... 908.2.8 Linearity Response ............................................................................................ 918.2.9 Angular Response .............................................................................................928.2.9.1 In-air response ....................................................................................928.2.9.2 In-phantom response ........................................................................... 928.2.10 Comparison of OSL Dosimeter in an d out of its Light-Tight Case .................. 939 ORGAN DOSE MEASUREMENTS ..................................................................................... 989.1 Pediatric Head Study .....................................................................................................999.1.1 Volumetric Protocol .......................................................................................... 999.1.2 Helical Protocol ............................................................................................... 1009.1.3 Image Quality Analysis ................................................................................... 1019.2 Adult Brain Perfusion ................................................................................................. 1019.2.1 Volumetric Protocol ........................................................................................ 102

PAGE 10

10 9.2.2 Helical Protocol ............................................................................................... 1049.2.3 Image Quality Evaluation ............................................................................... 1049.3 Adult Cardiac CT Angiography ..................................................................................1059.3.1 Volumetric Protocol ........................................................................................ 1079.3.2 Helical Protocol ............................................................................................... 10810 RESULTS ....................................................................................................................... ......11910.1 Characterization of X-Ray Beam s of Volumetric CT Scanner ................................... 11910.1.1 Exposure Reproducibility ...............................................................................11910.1.1.1 Service mode .................................................................................... 11910.1.1.2 Clinical mode .................................................................................... 11910.1.2 Beam Quality .................................................................................................. 12010.1.2.1 Tube voltage ..................................................................................... 12010.1.2.2 Total filtration ................................................................................... 12110.1.2.3 Half-value layer ................................................................................12210.1.3 Beam Width .................................................................................................... 12210.1.4 Dose Profile ..................................................................................................... 12310.1.4.1 X-axis ................................................................................................ 12310.1.4.2 Z-axis ................................................................................................12310.2 Characterization of OSL Dosimetry System ...............................................................12410.2.1 Tracking of System Standards ........................................................................ 12410.2.2 Dosimeter Response ........................................................................................ 12510.2.3 Room-Light Erasure ........................................................................................ 12510.2.4 Energy Response ............................................................................................. 12610.2.5 Scatter Response .............................................................................................12710.2.6 Dosimeter Calibration ..................................................................................... 12810.2.7 Linearity Response .......................................................................................... 12810.2.8 Angular Response ...........................................................................................12910.2.8.1 In-air response ..................................................................................12910.2.8.2 In-phantom response ......................................................................... 13010.2.9 Comparison of OSL Dosimeter In an d Out of Its Light-Tight Case ............... 13010.3 Organ Dose Measurements ......................................................................................... 13010.3.1 Pediatric Head Study .......................................................................................13110.3.1.1 Organ doses ...................................................................................... 13110.3.1.2 Image quality .................................................................................... 13510.3.2 Adult Brain Perfusion ..................................................................................... 13610.3.2.1 Organ doses ...................................................................................... 13610.3.2.2 Image quality .................................................................................... 13910.3.4 Adult Cardiac CTA ......................................................................................... 14110.3.4.1 Organ doses ...................................................................................... 14110.3.4.2 Image quality .................................................................................... 14410.3.5 Comparisons to Monte Carlo Organ Dose Simulations .................................. 14511 SUMMARY AND CONCLUSIONS ...................................................................................17711.1 Summary of This Research Project .............................................................................177

PAGE 11

11 11.2 Future Work .............................................................................................................. ..17911.3 Final Words .............................................................................................................. ...182 APPENDIX A EXAMPLES OF IMPLEMENTED CT PROTOCOLS ....................................................... 183B DOSIMETRIC PHANTOMS ...............................................................................................192LIST OF REFERENCES .............................................................................................................195BIOGRAPHICAL SKETCH .......................................................................................................201

PAGE 12

12 LIST OF TABLES Table page 4-1 Tissue-weighting factors .................................................................................................. ..508-1 HVL as a function of effective energy ...............................................................................969-1 Scan parameters for volumetr ic pediatric head protocol .................................................1099-2 Locations and number of dosimeter s used in and on pediatric phantom. ........................ 1099-3 Scan parameters for helical pediatric head protocol ........................................................ 1109-4 Details of volumetric br ain perfusion protocols. .............................................................1109-5 Locations and number of dosimeters used in and on adult phantom for brain perfusion protocol. ...........................................................................................................1149-6 Scan parameters for helical adult brain perfusion protocol. ............................................ 1149-7 Scan parameters for ad ult cardiac CTA protocol. ............................................................1159-8 Location and number of dosimeters used in and on adult phantom for cardiac CTA protocol. ..................................................................................................................... ......11510-1 Reproducibility of x-ray tube operated in service mode. ................................................. 14910-2 Reproducibility of volumetric clinical protocol. .............................................................. 14910-3 Reproducibility of helical clinical protocol. .................................................................... 14910-4 Measured beam energies of scanners used for organ dose measurements. ..................... 15010-5 Measured HVLs for each nominal x-ray tube voltage for the Aquilion One .................. 15210-6 Measured HVLs for each nominal x-ray tube voltage for the Aquilion 64 ..................... 15310-7 Measurement standards for microStar reader. ................................................................. 15610-8 Dose response of standard and screened dots. ................................................................. 15910-9 Energy response of the OSL dosimeters. .........................................................................16010-10 Measurements used in flat-field scatter response. ........................................................... 16010-11 Response of the dosimeters to scatter and a flat x-ray field. ........................................... 16110-12 Response of energy and scatter to the CT x-ray beams used for clinical protocols. ....... 162

PAGE 13

13 10-13 The calculated f-factors ....................................................................................................16210-14 Mean doses with a 95% confidence interv al measured by the dosimeters in and out of their plastic cases in two different positions within the CTDI phantom. .................... 16810-15 Mean organ doses for pediat ric craniosynostosis protocol. ............................................. 16810-16 Organ doses measured using the pediatri c phantom and craniosynostosis protocol ....... 16910-17 Number of low-contrast objects viewed using pediatri c craniosynostosis protocol parameters. ................................................................................................................... ....17010-18. Mean organ doses resulting from brain perfusion protocol. ............................................. 17110-19 Mean and maximum organ doses in mG y measured with ad ult brain perfusion protocol. ..................................................................................................................... ......17210-20 Mean organ doses in mGy resul ting from adult cardiac protocol. ................................... 17310-21 Mean and maximum organ doses in mGy resulting from adult cardiac protocol. ........... 17410-22 Organ dose comparison for the 320-slice CT pediatric craniosynostosis protocol between measured doses (mGy) and Monte Carlo (MC) simulate d doses (mGy). ......... 17510-23 Organ dose comparison for the 320-slice CT adult brain protocols between measured doses (mGy) and Monte Carlo (M C) simulated doses (mGy). ........................................ 17610-24 Organ dose comparison for the 320-slice CT adult cardiac CTA protocols between measured doses (mGy) and Monte Carlo (MC) simulated doses (mGy). ........................ 176

PAGE 14

14 LIST OF FIGURES Figure page 2-1 Illustration used for the body protocols .............................................................................342-2 Example of the acquisition table from a body protocol. ....................................................342-3 Two image reconstruction tables are depicted for a body protocol. .................................. 342-4 Reformation section of a neuroradiology protocol. ...........................................................352-5 Portion of CT ENT protocol ..............................................................................................353-1 Illustration of angular tube current modulation ................................................................. 445-1 Standard CTDI phantoms ..................................................................................................606-1 Pediatric tomographic phantom ......................................................................................... 696-2 Adult male tom ographic phantom. .................................................................................... 706-3 Diagram of the low-contrast module of the image quality phantom. ................................ 717-1 Representation of x-ray tube positions used for measurements ......................................... 797-2 Dosimeters suspended perpendicu lar to the anode-cathode axis ....................................... 807-3 Setup to measure beam profile parallel to anode-cathode axis. ......................................... 818-1 nanoDot dosimeters ...........................................................................................................958-2 Energy response, dose rate response and scatter response surface setup .......................... 958-3 Setup of scatter response ....................................................................................................978-4 Diagram of cross-sectional view of ion chamber and dosimeters ..................................... 968-5 Setup of in-phantom angular response ............................................................................... 979-1 Pediatric phantom with surface dosimeters to measure skin, eye and breast doses ........ 1099-2 Placement of skin and lens dosimeters on the forehead and lens of the eyes on the adult phantom for the brain perfusion protocol. ..............................................................1119-3 Dosimeter placement in the esophagus, measur ed in three slices of the adult phantom from the brain perf usion protocol. ...................................................................................1129-4 Locations of dosimeters within the thyroid. .................................................................... 112

PAGE 15

15 9-5 Locations of the dosimeters within the brain. .................................................................. 1139-6 Locations of dosimeters within the br east to capture sca ttered radiation. ....................... 1139-7 Cardiac R-R cycles are show n with exposure conditions ............................................... 1149-8 Placement of dosimeters in the lungs ...............................................................................1169-9 Placement of dosimeters within the stomach ...................................................................1179-10 Placement of the dosimeters to m easure skin dose across the breasts. ............................ 1179-11 Locations of dosimeters for breast do se measurements using the cardiac CTA protocols ..................................................................................................................... ......11810-1 Total filtration of x-ray beam using th e large filter and depicted for three beam qualities. .................................................................................................................... .......15010-2 Total filtration of x-ray beam for small and large filters using a 120 kV x-ray tube voltage. ...................................................................................................................... .......15110-3 Effect of bowtie filter on total filtration measurements. .................................................. 15210-4 Radiographic image of the 16-cm nominal beam width. ................................................. 15310-5 Beam profile of the 16-cm nominal CT x-ray beame. ..................................................... 15410-6 Normalized dose in air across the CT s canner gantry perpendicular to the anodecathode direction. .............................................................................................................15510-7 Normalized dose in air across the CT sca nner gantry in the anode -cathode direction. ... 15610-8 Measurement of DRK counts ...........................................................................................15710-9 Measurement of CAL counts ...........................................................................................15810-10 Measurement of LED counts ...........................................................................................15910-11 Percent decrease from initial dose ................................................................................... 16010-12 Response of the OSL dosimeters to increasing dose rates. .............................................. 16310-13 Linearity of the energy-corrected dosimeter doses .......................................................... 16410-14 Linearity response of the OSL dosimeters to increasing tube curre nts, representative of the range of doses measured. .......................................................................................16510-15 Linearity response of the OSL dosimet ers and ion chamber to increasing tube currents, representative of th e range of doses measured. ................................................. 166

PAGE 16

16 10-16 Dose absorbed by the nanoDot dosimeters, in air, as a function of x-ray tube angle. ..... 16710-17 Dose absorbed by the nanoDots in the h ead phantom, displayed as a function of xray tube angle ................................................................................................................ ...16810-18 Image quality phantom images for pediatric protocol ..................................................... 16910-19 Image quality phantom images for cardiac protocol. ....................................................... 175A-1 Example of a body protocol. ............................................................................................ 185A-2 Example of a cardiac protocol. ........................................................................................ 186A-3 An example of an ENT protocol.. ....................................................................................188A-4 Example of a neur oradiology protocol ............................................................................ 189A-5 Example of an MSK protocol. ......................................................................................... 191

PAGE 17

17 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ORGAN DOSE MEASUREMENTS FROM MULTIPLE-DETECTOR COMPUTED TOMOGRAPHY USING A COMMERCIAL DOSI METRY SYSTEM AND TOMOGRAPHIC, PHYSICAL PHANTOMS By Lindsey K Lavoie August 2009 Chair: Manuel Arreola Major: Medical Physics The technology of computed tomography (CT) imaging has soared over the last decade with the use of multi-detector CT (MDCT) scanne rs that are capable of performing studies in a matter of seconds. While the diagnostic inform ation obtained from MDCT imaging is extremely valuable, it is important to ensure that the ra diation doses resulting from these studies are at acceptably safe levels. This research projec t focused on the measurement of organ doses resulting from modern MDCT scanners. A commercially-available dosimetry system was used to measure organ doses. Small dosimeters made of optically-stimulated lumine scent (OSL) material we re analyzed with a portable OSL reader. Detailed verification of this system was performed. Characteristics studied include energy, scatter, a nd angular responses; dose linearit y, ability to erase the exposed dose and ability to reuse dosimeters multiple times. The results of this verification process were positive. While small correction factors needed to be applied to the dose reported by the OSL reader, these factors were small and expected.

PAGE 18

18 Physical, tomographic pediatri c and adult phantoms were used to measure organ doses. These phantoms were developed from CT imag es and are composed of tissue-equivalent materials. Because the adult phantom is comp rised of numerous segments, dosimeters were placed in the phantom at several organ locations, and doses to select organs were measured using three clinical protocols: pediat ric craniosynostosis, adult brain perfusion and adult cardiac CT angiography (CTA). A wide-beam, 320-slice, vo lumetric CT scanner and a 64-slice, MDCT scanner were used for organ dose measurements Doses ranged from 1 to 26 mGy for the pediatric protocol, 1 to 1241 mGy for the brain perfusion protocol, and 2-100 mGy for the cardiac protocol. In most cases, the doses meas ured on the 64-slice scanner were higher than those on the 320-sl ice scanner. A methodology to measure organ doses with OSL dosimeters received from CT imaging has been presented. These measurements are es pecially important in keeping with the ALARA (as low as reasonably achievable) principle. Wh ile diagnostic information from CT imaging is valuable and necessary, the dose to patients is al ways a consideration. This methodology aids in this important task.

PAGE 19

19 CHAPTER 1 INTRODUCTION The issue of the apparen t increased risk of cancer incidence and mortality from radiation doses resulting from computed tomography (CT) scanning has been widely addressed in the radiology community recently. Among these co ncerns are the documented increase in the number of scans done in the past few years as compared to those a decade ago and especially those from pediatric CT exams.1 Similarly, as the dose from CT exams has become such a widely-discussed and publically debated topic, a brief description of relevant studies is appropriate. 1.1 Increased Use of Computed Tomography A reliable source of data which can be us ed to determ ine the number of medical procedures done in a time period is the reimbur sement payment data. Mettler conducted an institutional study that analyzed medical bill ing data for CT studies from 1990 through 1999.1 Within this nine-year span, there was a 58% increase in the total number of diagnostic procedures. In 1990, CT scans made up 6.1% of all radiology procedures. Almost a decade later, 11.1% of exams were CT studies, which co ntributed to 70% of the total dose to patients from all radiology exams. Furthermore, the study also found that most patients had more than one CT study on the same day. While the ordering of extraneous exams should not be condoned, it must be realized that many CT exams require more than one se ries for proper diagnosis (i.e., a three-phase liver exam for the diagnosis of a liver mass or a pre-transplant evaluation). In the realm of pediatric radiology, Brenner reported a 92% increase from 1996 to 1999 in the number of abdominal and pelvic exams performed on patients under 15 years old.2 In 2009, the National Council on Radiation Protection and Measurements (NCRP) published a report detailing differe nt sources of radiation to wh ich the United States population

PAGE 20

20 was exposed.3 Of all sources of exposure included in the study, they report ed CT for medical purposes contributes 24% of the total collective effective dose due to exposure to ionizing radiation. Within the medical modalities, CT comprises 49% of the total collective dose; according to the same report, this figure was only 3% in the 1980s. In 2000-2001, the Conference of Radiation C ontrol Program Directors, Inc. (CRCPD) conducted a Nationwide Evaluati on of X-Ray Trends (NEXT)4 study and looked at the number and details of CT exams performed at medical ins titutions in the United States. In the year the data were gathered, 45.1 million CT procedures were done nationwide. Head CT exams were the most frequent, making up 33% of the total nu mber of CT exams, while abdomen and pelvis CT exams made up 22% of the total number of exams and second to head exams in frequency. While the statistics illustrated in these studies seem extreme, it is important to note that the increase in the number of CT studies is driven by advancing technology. As CT scanners have become faster and total scan time has been redu ced, there is no longer the need for sedation in many cases, making a CT scan more feasible and efficient. Wide-beam CT has increased the total beam width, allowing entire organs to be im aged in one rotation of the x-ray tube in less than one second. The utility of CT as a diagnos tic tool and the valuable, detailed information provided to physicians has signifi cantly contributed to an increas e in the total number of CT scans performed in the United States. 1.2 Effects on Pediatric Patients The focus of a study by Brenner et al was a pediatric populati on, and his conclusion was that the lifetime risk of radiat ion-induced mortality is larger when the patient undergoes the CT study at an earlier age; and that this risk then increases as age at exposure increases.2 Younger patients have, in principle, a longer period of time in which to express the effect (in this case, cancer), and second, because the organs of pediat ric patients are still de veloping; these rapidly-

PAGE 21

21 dividing, growing cells are more sensitive to radia tion than the same cells of the fully-developed organs of an adult patient. In a more qualitative analysis, Brenner stated that for the same scan parameters, pediatric organ doses were typically hi gher than those in adults.5 Along with increas ed sensitivity to radiation and longer lif etime for expressing radiation-indu ced cancers, pediatric patients, compared to adults, have a smaller body size. Du e to the geometry of CT and its full rotation around isocenter during exposure, a given organ of interest will be found in a position distal to the x-ray tube at some point during the scan. At such a time, there is partial shielding of the organ by the anatomical structures lying in between the xray tube and the organ of interest. In a pediatric patient, this shielding is considerably le ss than in an adult, strictly due to the smaller size of a child, with less shielding between the x-ray tube and the aforementioned organ. When organs are in a closer position to the x-ray tube after a half-rotation, they are exposed in a similar manner for adult and pediatric patients, and thus, th e total dose to the organ is higher in a smaller pediatric patient than in a larger adult. Brenner was careful to point out the caveats of the groups dose cal culations. As with most radiation-effect risk estimates, the data were derived from documentation from atomic bomb survivors who received significantly hi gher doses than those observed in CT. Thus, extrapolation is necessary, and while the accepte d linear-no-threshold model applies, there were very little data in the dose range relevant to a CT exam and its possible effects. Finally, Brenners calculations were overestimates becau se adult protocols were used in the dose calculation which in general implie s a higher tube current value. Brenner cited several studies that have offered suggestions on clinical ways to reduce pediatric doses and admitted that none of these techniques were accounted for in his study.

PAGE 22

22 One of Brenners conclusions concentrates on the fact that, while the reported organ doses seem large, there is a very small percentage increa se in the risk of radiat ion-induced cancer when compared to the natural background. While this translates to a sma ll increase in the risk to an individual, Brenner claims these doses are no t a matter of concern fo r personal health, but instead are of concern for public health. Althoug h the risk per person is small, the number of people in the population who are undergoing CT exams has steadily increased, which has brought up the issue of appropriateness and usage, topics of immense intere st to the radiological and medical communities. One way to address th is issue is through education of physicians on a more careful ordering of CT exams. 1.3 Specific Aims Because of the current relevance of the incr eased number of CT studies in the United States and the apparent higher ra diation doses resulting from CT st udies, it was the goal of this research project to develop a methodology for m easuring organ doses resulting from clinical CT protocols. To this end, the followi ng specific objectiv es were proposed: a) Develop clinical CT protocols. With involvement of experienced radiologists and CT technologists, protocols were to be developed based on clinical indications and by asserting scan parameters necessary to achieve adequate clinical image quality required for each protocol as defined by radiologists. Dose considerations were to be an essential part of the protocol development process. b) Verify commercial dosimetry system. A co mmercially-available optically-stimulated luminescence (OSL) dosimetry system, which can be used in a clinical setting, was to be verified for accuracy and reliability. Among the characteristics of the system to be investigated were energy, angular and scatte r responses, accuracy, reproducibility, ability to erase trapped energy and reusability.

PAGE 23

23 c) Characterize the beams of a 320-slice CT scanner. As these wide-beam scanners were the latest in CT technology, methods were to be developed to char acterize these wide xray beams. The characteristics to be eval uated were reproducibilit y, tube voltage, halfvalue layer, beam width, and beam profiles. d) Measure organ doses. Three clinical CT prot ocols were to be used to measure organ doses resulting from wide-beam CT scanning using a 320-slice CT sy stem. Patients were to be simulated using physical, tomogra phic phantoms developed by the Nuclear and Radiological Engineering group at the University of Fl orida (UF). Small, OSL dosimeters were to be used to measure the doses to selected organs on and in these phantoms. e) Compare organ doses. Organ doses were to also be measured on a 64-slice multipledetector computed tomography (MDCT) system Protocols similar to those of the 320slice scanner were to be used for organ dose measurements using the same physical phantoms. f) Evaluate image quality. A simple low-contrast detect ability test was to be assessed as a feasible image quality indicator for the clinical protocols selected for d) and e) above. It was to be investigated if po ssible changes in acquisition pa rameters could be suggested that could result in a dose reduction while maintaining adequate levels of image quality.

PAGE 24

24 CHAPTER 2 CLINICAL CT PROTOCOLS As described in Chapter 1, the objectives of th is resea rch project were to utilize actual clinical protocols for organ dose measurements rather than measuring doses for arbitrary combinations of scanning parameters, which may be unrepresentative of actual clinical doses. Two adult protocols were chosen: one utilized to assess brain perfusion an d a second that yields images of the heart using CT angiography (CTA). To address some of the concerns raised by Brenner regarding doses in pediatric CT, a pediatric head pr otocol was also selected. Throughout the course of this research project, these protocols, initially developed for the numerous MDCT scanners at Sha nds Hospital at UF, were redesi gned to take full advantage of the newer technology of a 320-slice volumetric CT scanner purchased by the hospital in 2008. The adjustments made to the protocols were a collaborative effort of many members of the radiology department under the leadership of medical physicists, the process of which is detailed in this chapter. 2.1 Radiology Practice Committee In response to the im perative needs of protocol standardization and workflow improvement, and to begin the process of im plementing study appropr iateness criteria, the Radiology Practice Committee (RPC) was formed at Shands Hospital at UF in October of 2005 with the main task of modifying existing CT and magnetic resonance imaging (MRI) protocols, and creating new ones when necessary. The committee was comprised of radiologists, the picture archiving and communications system (PAC S) administrator, medical physicists, CT and MRI technologists, and radiol ogy administrators. While both CT and MRI protocols were developed and edited, the scope of this document is only the CT protocol development process followed by the RPC.

PAGE 25

25 The RPC set various goals in its charter, includ ing the standardization of protocols in form and description and the improvement of workflow The process was iterative in nature, with protocols being evaluated in small groups by the different radiology subspecialty sections under the leadership and guidance of the medical physicists. The following sections describe this complex and lengthy process. 2.2 Protocol Standardization and Revision A standardized set of protocols im proves work flow. As all involved steps are established clinically, the protocols can be used as a teaching tool for residents. In addition, they can further provide clinicians with the information needed to make an appropriate exam order based on clinical indication. When protoc ols are not standardized, they are frequently altered for each specific patient. While sometimes this modifi cation is unavoidable, workflow is slowed down when a technologist needs to make changes to an ex isting protocol to suit the specific needs of a single exam. To further complicate the matter, clarification about the physicians orders is sometimes needed for the technologist to modi fy and accommodate the proper scan. As this process takes time to understand and finalize, pati ent throughput is compromised, with all of the negative ramifications on cost and patient care. The process of standardization was accomplished in a variety of ways. First, radiologists on the committee were responsible for revising and setting anatomical and spatial scan parameters necessary for diagnosis, including slice thickness a nd spacing. Radiologists were also responsible for creating the lis t of clinical indications pertinen t to each protocol. The role of the medical physicists was to make sure dose co nsiderations were kept in mind. Finally, CT technologists were tasked with revising and esta blishing the technical and practical aspects of a scan, including the anatomical star t and end points of the scan, the application of a breath-hold or

PAGE 26

26 a specific breathing technique, oral and/or intravenous (IV) contrast agent administration and the appropriate timing of dela yed scans, among others. 2.3 Clinical Indications in Protocols One of the most im portant aspects of an im aging protocol is the clinical indications associated with that protocol, that is, the cl inical reason for the st udy. While this list of indications may always be a workin-progress due to the vast num ber of clinical indications, it serves as a starting point for the appropriateness of ordering a given CT exam. As all academic hospitals, Shands at UF has a la rge number of residency programs in many specialties. Thus, a properly-designed list of indications may be used by the leastexperienced residents, who may be unsure of which CT exam is needed for a give n patient indication. Allowing access to such a useful list of indications may al so aid in reducing the number of inappropriate CT exams that are ordered every day. In fact, a sec ondary benefit of standardized pr otocols is the realization that CT may not be the most appropria te modality for a given diagnosis. In order to facilitate and take full advantage of the efforts involved in the assignment of clinical indications to each CT protocol, a to ol was developed to allow a search through the entire database of protocols for a given indication. Presented as a simple input text field, the user simply enters the indication, and any associated pr otocols within the databa se are returned. The softcopy version of the selected protocol is displayed, along with its full list of clinical indications. The importance and usefulness of such a tool is extraordinary and without precedent in the field of radiology. 2.4 Picture Archiving and Communications System As digital im aging has become the standard in radiology departments across the country, a PACS provides essential storage and access to di gital images. Network bandwidth, transmission times, and storage capacities are the main concerns of a PACS administra tor and thus, his/her

PAGE 27

27 involvement is indispensible, as the protocol s must contemplate and take into account the number and size of the images generated in a given timeframe. The number of images in a study is generally dictated by the moda lity and the specific exam. The size of each image depends on the number of pixels used during image acquisition.6 The number and size of these pixels are directly re lated to the resolution of an image; a larger number of smaller pixels results in an image with higher resolution than that with fewer and larger pixels, for the same displayed field-of-vie w (FOV). As an example, a typical chest exam involves two radiographs: one in the anteroposterior (AP) orientation and a se cond in the lateral position. The typical pixel matrix size of a di gital radiograph is 2,000 x 2,000. The bit-depth, or number of bits per pixel, is 2 bytes. For CR images, there are 4096 (12-bit) gray scale values available for the image. Thus, roughly sixteen me gabytes (MB) of storage is required for a 2view chest exam. A CT exam, on the other ha nd, could involve anywhere between tens of images up to thousands of images. Despite having a smaller pixel matrix (512 x 512) than a CR radiograph, a standard chest CT study requires roughly 130 MB of space, assuming an average of 250 images. In summary, the larger number of images generated in a given CT study results in larger storage needs than fo r most other imaging modalities. The RPC ensured that estimated study size be considered in the development of each protocol. Determination of slice thickness and spacing are two important parameters in protocols, and both directly affect the size of a CT study. When thinner s lices are reconstructed, a larger number of images is nece ssary to cover the entire scan le ngth; conversely, thicker slices generate fewer images for the same scan length. Important considerations include the level of detail needed for diagnosis, as well as the le ngth of the scan. Radia tion protection guidelines dictate that only the area of in terest be scanned in CT study, thus limiting the number of images

PAGE 28

28 that are generated as well as dose to the patient. As was the case in the entire development process, for each protocol, these tradeoffs we re debated among RPC members, and the most appropriate and efficient values for such parameters were chosen by the committee. 2.5 Protocol Development Process The following is a detailed description of the protocol developm ent process. While all RPC members involved performed specific duties, it was the general leader ship of the medical physicists that kept the committee focused on the task of updating and creating CT protocols. Specifically, RPC meetings were run by one of th e medical physicists (i.e., the author, LKL). Weekly assignments were given to members by this physicist, who was responsible for updating protocol documents throughout the process. When necessary, this medica l physicist coordinated among members so that each protocol was approved by the committee. 2.5.1 Database Development To f acilitate the redesign and organiza tion of CT protocols while beginning implementation of a paperless process, a data base was developed to store protocol-related information. In the database, each protocol ha s its own set of unique descriptors, the most relevant for clinical use being th e Indications field to allow the search for a specific protocol, based on a known clinical indication. It was necessary that this field be dynamic, as the number of indications for a given protocol may be vast and may need to be updated frequently. 2.5.2 Protocol Template Once the protocol datab ase was establis hed, the committee proceeded to develop a softcopy document for use by the CT technologists to use as a reference for each protocol. Templates were created for each anatomical body section, which efficiently condensed information in the database and displayed the protocol in a format most useful for the

PAGE 29

29 technologists. Once completed, these templates were implemented online as the softcopy of each protocol and were made av ailable for physicians to view. 2.5.2.1 Title and illustration Each tem plate is headed by the standardiz ed protocol title, followed by a pictorial anatomical diagram of the body section of interest These diagrams can be presented in axial, coronal or sagittal views, as necessary. For example, the body protocols depict an anterior anatomical view from the top of the neck th rough the femoral heads. The head and neck templates use only a lateral view of the entire sk ull and the neck through th e cervical vertebrae. The protocol then illustrates the start and end poi nts of the scan series by means of lines on the diagram, also indicating a craniocaudal or caudo-cranial direction, wi th an anatomical reference. An arrow describes the direction of the table feed (in the case of helical protocols) with respect to the patient. Figure 2-1 depicts an exam ple of the diagram used for CT body exams. 2.5.2.2 Acquisition parameter description table Adjacent to the ana tomical diagram is the acqui sition table. An image acquisition refers to the actual process of scanning the anatomy of interest to generate raw data. Fields within the acquisition table include the following: patient position, respiration, co ntrast requirements, contrast medium requirements and timing, volum e acquisition specifications, and the start and end point of the scan, as depict ed in Figure 2-2. Multiple imag e acquisitions are labeled with different letters. In this example, there is only one acquisition, thereby de noted with the letter A in Figure 2-2. 2.5.2.3 Reconstruction parameter tables The next set of tables in th e protocol tem plate contains the necessary reconstruction parameters. A reconstruction is defined as an image generated from the raw data of the CT acquisition. These tables include the type of al gorithm used in processing the raw data (e.g.,

PAGE 30

30 bone, lung, soft tissue), the thickness and spacing of the reconstructed images, and finally the FOV. Each reconstruction refers to the corr esponding letter-labeled acq uisition from which the raw data is processed. In the examples shown in Figure 2-3, the first table, labeled A1, is the first reconstruction (1) from the data obtained in acquisition A (s hown in Figure 2.2), and it uses a soft-tissue algorithm. The second reconstruc tion, denoted A2, utilizes the lung algorithm but produces thinner images than the first recons truction (3 mm thickness a nd spacing, as compared in 5 mm in A1). In both tables depicted in Figure 2-3, the FOVs denoted refer to the largest section of the patient within the anatomy to be imaged, and the FOV is increased by an additional 2 cm on each side of the patient; this in crease in FOV is described as patient largest +4 cm in the table. 2.5.2.4 Reformation parameter table The next sec tion of the template is the reformation table. A reformation is an image set that is constructed from processe d data of reconstructed images. The fields within the table are the same as those for the reconstructions, with the addition of the plane in which the images are reformatted (i.e., sagittal, coronal, axial, obli que). Both the reconstruction and reformation sections include a space for a diagram to illustrate the specified plane of the reformatted image. Some protocols have an anatomi cal diagram in the reformation s ection that depict s the FOV, or boundaries, for the reformation. In Figure 2-4, the coronal reformation begins at the tip of the nose but does not continue thr ough the entire head, stopping around the condyle of the mandible. 2.5.2.5 Additional protocol information Following the reform ation table and diagram is a section for additional information related to the protocol labeled Other, intended to include information that may not have a specific place in the template. Special instructions or notes are placed here, as shown in Figure 2-5. Further examples of clinical CT protoc ols developed are found in Appendix A.

PAGE 31

31 2.5.2.6 Indications The final section of the protocol tem plate is perhaps the most important and compiles a list of clinical indications associated with each CT protocol. Changes or updates to the indications are also updated in the protocol database. This section was closely and promptly maintained by the medical physicist, as the information on the pr otocol templates must be the same as that stored in the database. 2.5.2.7 Using the templates and updating the protocols The next step in the protocol develo pm ent and implementation process involved transferring the information from existing CT prot ocols into the database, as well as into the template of the corresponding body section. Initial ly, after the basic information was transferred, members of the committee reviewed the protocols for completeness, appropriateness and accuracy. One of the major changes made to the sets of previously existing protocols was the title. One goal of the committee was to consol idate protocols by separa ting the protocol name from its indication(s). For example, the protocol formerly titled Liver Mass was changed to Three-Phase Liver to allow the protocol to be used for indications other than the specific imaging of a liver mass. A secondary caveat to the specifi c naming of the protocols is an initiative cu rrently being pursued by the Radiological Soci ety of North America (RSNA),7 in conjunction with the American College of Radiology (ACR),8 with the final objective of standardizing words and accepted standardized names used in diagnostic radiology. Termed RadLex, the goal of this RSNA initiative is to implement this standard list as a national standard so that all institutions use the same terminology. The protocol t itles adopted by the RPC follow the RadLex nomenclature whenever available, as the RadLex project is a work-i n-progress and not yet complete.

PAGE 32

32 Efforts were also made to keep all of the pr otocol parameter tables independent of the CT manufacturer and model. For example, the majo r CT manufacturers do not always use the same terminology to describe functions or parameters of the scanner. The localizing tomograph acquired at the start point of each CT scan have manufacturer-specific na mes, such as topogram, scanogram, scout view and pilot scan.9 Similarly, different scanners may follow different scan sequences for a given protocol. One scanner ma y perform a torso acquis ition protocol in one scan, while another scanner will br eak the acquisition into separate sections. The images are still sorted, displayed and viewed in the same order, but the acquisition sequence is different between the two different scanners. 2.6 Protocol Website A website was developed to allow physician access to the CT and MRI protocols online.10 A list of all protocols is availabl e, as well as a fully-functional i ndication search tool. In using the tool, the user types a clinical indication into the search text box. The results of the search are displayed in prioritized order for t hose that best match the term(s) in the text field. A result list may display a statement in red underneath the protocol title for some protocols. This statement is meant to further direct the user to the appropr iate protocol or to alert him/her of the possible inappropriate selection of a protocol. The user is able to view each protocol within the search results by selecting the protocol title. Many revisions to this search tool have been necessary as the process of creating a complete list of indications by clinical specialty occurs in a cumulative manner, and the addition of clinical indications will be continual. Although the RPC is still continuing work to fu rther develop this we bsite, the ground work is complete. The power of this tool is trem endous; it provides physicians with two resources to aid in the process of appropria tely ordering a diagnostic imaging exam. First, the search by indication tool may be used to aid the physic ian in choosing the best exam based on the

PAGE 33

33 particular indication of a given patient. Second, the physician is able to view all details of the indicated protocol. This standardization of protocols based on clinical indication improves the quality of patient care, while optimizi ng workflow and proper use of resources. The implementation of the CT protocols at UF is complete and has been a success. Although small edits are still necessa ry, these are minor compared to the scope of the project. Furthermore, these are dynamic protocols, delib erately designed to allow for the continued development in technology which will ultimately drive major revisions and updates of existing protocols. The RPC committee has put into pl ace a structured process to allow for these advancements.

PAGE 34

34 Figure 2-1. Illustration used fo r the body protocols, with lines showing the start and end points of the scan as the thoracic inlet and pubic symphysis, respectively; the scan direction is cranio-caudal. Figure 2-2. Example of the acquisition table from a body protocol. The letter A in the Begin line notes the first acquisition for the protocol. Figure 2-3. Two image reconstruction tables are depicted for a body protocol. A1 and A2 use different algorithms, slice thickness and slic e spacing. Patient la rgest + 4 cm denotes an increase in the FOV of 2 cm on each side of the patient.

PAGE 35

35 Figure 2-4. Reformation section of a neuror adiology protocol, depicting a diagram for reformation boundaries on the left, and the associated reformation table on the right. Figure 2-5. Portion of CT ENT protocol, depicting the use of the Other section for specific protocol notes.

PAGE 36

36 CHAPTER 3 MULTIPLE-DETECTOR COMP UTED TOMOGRAPHY Invented in 1972 by Godfrey Hounsfield a nd Allan Corm ack independently, CT was designed to solve the problem of the superimposition of anatomi cal projections in clinical laminar tomography. This problem is a result of the process of projection radiography, which generates two-dimensional images of three-dimens ional objects. In addition, poor tissue contrast results from this superimposition of anatomy. The goal of CT is the improvement of ti ssue contrast by generating images of twodimensional nature which do not contain images of overor underlying structures. A true crosssectional view of the scanned anatomy gives the radiologist improved tissue contrast, as well as depth information never before seen on a two-dimensional radiograph. CT has greatly improved since 1972. With comp lete rotation times of less than a halfsecond and single-scan coverage up to 16 cm, it is easy to see how CT has become such a useful clinical tool. A hi story detailing the progression of scanners can be found elsewhere.6,9 Below is a discussion of the relevant CT scanning parameters pertaining to this re search project, including their definitions and clinical impact. Many of the aspects of CT that are outlined in this chapter were considered by the RPC during protocol development. While some dose reduction techniques, such as tube current modulation, were already in place as part of the implemented protocols, other techniques, specifically tube voltage reduction, were investigated during this research project. 3.1 Detector Array Systems In MDCT, sm all, multiple detectors make up the detector array. The minimum slice thickness is determined by the width of the individual detector elements.11 However, the individual slice thickness may be greater than the individual detector widths because multiple

PAGE 37

37 channels can be selected for data acquisition. Th e detector configuration is generally described as the product of the number of data channels and the width of the dete ctor rows within the channel. For example, a detector configuration of 4 x 1.5 mm represents f our slices that are 1.5 mm thick. The total scanned volume coverage per rotation is determined by the physical collimation of the x-ray beam. 3.2 Modes of Acquisition There are tw o modes of acquisiti on in CT: axial and helical. While scanners of today are capable of both, it was only in the 1990s that technology allowed for th e inception of helical scanning. These modes of acquisiti on are described briefly below. 3.2.1 Axial Data Acquisition Early generation scanners were lim ited by elec trical cables that supplied the x-ray tube with the required electrical power and cables that transferred data from the detector array to the data processing unit. This combination allowed one full rotation of the x-ray tube and detector at a time. One slice in the ax ial plane was acquired and the cables wound up, the table was incremented (i.e. advanced) to the next position, and the next slice was acquired as the cables unwound.9 However, in the 1990s, slip-ring technol ogy was developed that allowed the x-ray tube and detector to rotate c ontinuously in the gantry, overcom ing the limitation of needing to stop acquisition for a table increment before the next rotation. This development reduced the time needed in between acquisitions, but data acq uisition was still limited by the time required for the table to advance through the gantry. 3.2.2 Helical Data Acquisition During helical CT acquisition, the table advanc es continually through the gantry during the x-ray exposure. Thus, in the patients fram e of reference, the x-ray tube follows a helical path around the patient. Slip-ring technology also resulted in faster rotation times, all of this helping

PAGE 38

38 to reduce the total scan time as compared to ax ial acquisitions, for which time is needed between exposures for the table to move. 3.2.3 Volumetric Data Acquisition The constant advancem ent of MDCT scanners with the objective of faster scanning and wider coverage has led to the advent of wide-b eam scanners, like the 256slice Philips Brilliance (Philips Medical Systems, Aurora, IL) and the 320-slice Toshiba Aquilion ONE (Toshiba American Medical Systems, Tustin, CA), to addre ss the clinical implications of complete organ coverage in a single scan. As indicated in previous chapte rs, an Aquilion ONE scanner was installed at Shands Hospital at UF in 2008 and t hus, it was one of the focuses of this research project. The detector elements in the Aquilion ONE are 0.5 mm x 0.5 mm,12 and the total cone angle of the x-ray beam is 15.2. A nominal x-ray beam width of 16 cm enables complete coverage of entire organs,13 such as the brain and heart, in one axial rotation of the x-ray tube and detector array. To appreciate the clinical usefulness of this design, consider the case of brain perfusion studies. A contrast agen t is administered to the patient and, at the appropriate time, a scan of the entire brain is captured at once, generating temporally-uniform,14 volumetric data, from which a variety of images can be rec onstructed. A dynamic volum e study is performed by scanning the brain at specific ti me intervals to monitor the prog ress of the contrast agent through the arterial and venous phases. Anatomic info rmation is generated by differences in tissue attenuation, while functional inform ation is calculated using the ti me characteristics of the flow of the administered contrast agent. Perfusion capabilities have been available on 64-slice systems for a few years; however, only a 3.2 cm s ection of the brain is scanned at a time to capture perfusion data. The ability to acquire un iform volumetric informati on of the entire brain is a breakthrough for improving the information av ailable in a clinical setting. However, comparison of doses from such wide beams with those from scans covering the same volume

PAGE 39

39 with narrower beam widths is necessary. This research project assesse d organ doses resulting from clinical scans using both a wide-beam volumetric CT scanner and a 64-slice MDCT scanner to determine if there was significant discrepancy in doses resulting from these scans of different total beam widths. 3.3 Clinical CT Considerations 3.3.1 Slice Thickness Slice thickness is one of the most important parameters of a CT scan, and has significant image quality and file size implications such as transfer times and storage. There are many factors to consider in determin ing the most appropriate scanned and reconstructed slice thickness of a given protocol. Thin slices yield improve d spatial resolution along the Z-axis (assuming all other scan parameters remain constant). Conversely, thin slices suffer from a decrease in the signal-to-noise ratio (SNR), due to the smaller number of photons collected per detector element. For a given volume, thin acquisiti on slices also result in an in creased number of images, putting a potential strain on the networ k and minor impacts on the imag e storage system, namely the PACS archive. Due to the numerous tradeoffs be tween thin and thick sl ices, it should be the clinical need of the CT scan that dictates the scan parameters. 3.3.2 Image Reconstruction Once the raw data are acquired with MDCT, i m ages can be reconstructed in different combinations, depending on the clin ical needs of the study. For example, the data from a 4 x 1.5 mm acquisition may be reconstructed to one 6.0 mm slice, 2 slices that are 3.0 mm thick, or the original acquisition of 4 slices that are 1.5 mm thick. While this allows for flexibility, it is not recommended to reconstruct to a slice thickness that is thinner than the acquisition thickness: data interpolation is necessary to achieve this goal, and the results are not of full integrity.

PAGE 40

40 3.4 Dose-Reduction Techniques As CT technology continues to produce newer and better system s, it is important to keep in mind that the basic principles of x-ray imaging remain true and that dose considerations follow the ALARA (as low as reasonably achievable) prin ciple. While there are numerous techniques to reduce CT dose, the main three are lowering the tube current (mA), dynamic tube current modulation and reduction of tube vol tage. At a given kV, the tube current, along with the total time of exposure, determines the total number of x-ray photons produced by the x-ray tube, sometimes referred to as the mAs. Since the dos e delivered to the patien t is proportional to the number of photons incident to the anatomy, any adjustment in mA (assuming all other parameters remain constant) directly affects the delivered dose. The total number of photons and therefore the tube current also directly impact the level of noise in an image, visually described as the mo ttle in an image. Mathematically, random noise in a digital image is expressed as the square ro ot of the mean number of x-ray photons reaching the imaging detector.6 Because mAs determines the number of x-ray photons in the CT x-ray beam, it directly affects the amount of noise in the image. The acceptable noise content in a clinical image is dictated by the observing radi ologist, as noisy images can negatively affect diagnosis, thus resulting in th e well-known tradeoff between dose to the patient and image noise. 3.4.1 Tube Current (mA) Adjustment CT protocols are pre-programm ed into the scanner by applications personnel of the manufacturer and lead CT technologists. This process is necessary to streamline the performance of CT studies, since the manufacturer-loaded study pa rameters are generic to apply to the average patient. X-ray photon attenuation is affected by the size of the patient; larger patients attenuate a larger number of photons th an a smaller patient, when the same exposure parameters are applied. In many cases, the mA ma y be lowered for a smaller patient and because

PAGE 41

41 of this reduction in size (and therefore attenuating material), an adequate number of photons will still reach the detector. A study on pediatric ches t CT reported that adequate image quality was achieved with a low-dose protocol using 25 mAs and resulted in a dose reduction of 90% compared to the standard protocol that used 250 mAs.15 Manual reduction of mA, depending on patient size, is one method of dose reduction in CT. 3.4.2 Tube Current (mA) Modulation A second option for dose reduction involves using different mAs values at a constant kV throughout a single CT s can series X rays are continuously emitte d from the x-ray tube as it rotates around the patient. Atte nuation profiles are collected at the detector at different angles and reconstructed to produce a cross-sectional image. Tube current modulation refers to adjustments made to the mA during one CT exposure to decrease the dose to the patient while maintaining acceptable noise levels, and therefore, image quality. As the CT x-ray beam rotates around the pati ent, the amount of x-ray transmission through the patient changes. When the tube is in a lateral position, x-rays are absorbed to a greater degree than when the tube is in the AP/PA position because, in general, a patient is thickest in the lateral direction, as depicted in Figure 3-1. Automatic tube current modulation increases the tube current to increase the photon transmission through the thicker part of the patient and decreases it to transmit through th inner sections. Because the noise of an image is dictated by the region of greater attenuation a nd decreased photon fluence at th e individual detector, the tube current is lowered when the x-ray tube is at the AP position to maintain patient dose at lower levels with little degradation in image qua lity. The two techniques of mA modulation are angular, or X-Y modulati on and Z-axis modulation.

PAGE 42

42 3.4.2.1 Angular (X-Y) tube current (mA) modulation Angular m A modulation refers to adjustments to the tube current made by the CT scanner to account for differences in attenuation in the x-y imaging plane of the patient,16 i.e., the scan plane. While different manu facturers have different name s and methods for angular mA modulation techniques, the general approach is the same. One appr oach involves an on-the-fly adjustment in which the scanner utilizes fluen ce information from projections with a 5-degree range to decide if an adjustment to the current is required for the next 5 degrees. Other simple approaches utilize information from the two perpendicular localizing tomographs performed before each CT image acquisition. Based on this attenuation data, the mA is modulated as a function of x-ray tube position with re spect to the patient during the scan. 3.4.2.2 Z-axis tube current (mA) modulation Tube current m ay also be modulated in the Z-axis,16 or along the superio r-inferior axis of the patient. Using this technique, the operato r selects a nominal noise level acceptable for the CT study in question. Informati on is again obtained from a localizing tomograph from which the scanner automatically computes the necessary ap proximate variations in mA values needed throughout the scan to obtain that desired noise level. The Toshib a Real E.C. software uses the AP/PA tomogragh to calculate a thickness of wate r that attenuates an equivalent amount of photons as the patient at different positions along the Z-axis. As the patient and table translate through the CT scanner during imag e acquisition, the mA is modulated to produce the same level of noise in all images, regardless of the size and shape of the anatomy being scanned. 3.4.3 Tube Voltage One technique for dose reduction, used extensively in the case of pedi atric patients, is a decrease in the x-ray tube voltage.17 It has been shown that di agnostic-quality images can be produced using lower tube vo ltages for pediatric patients.18 Because pediatric patients are

PAGE 43

43 smaller in size than adults, there is less tissue at tenuating the x-ray beam as compared to an adult when the same scan parameters are used, thus dose decreases as the tube voltage decreases. Furthermore, a lower tube voltage shifts the x-ray energy spectrum do wnwards in such a way that photoelectric absorptions are more prominen t and away from Compt on scattering, enhancing the tissue contrast of the image.17 One must be careful in this task, however, as more significant beam hardening artifacts may occur in the CT image as a result of a lower tube voltage.18 Beam hardening occurs when lower energy photons are preferentially absorbed by an attenuating material, thus increasing the effective energy of the x-ray beam compared to the initial beam entering the material. This process results in an artificial brightness at th e edges of the material in a CT image and a darkening at the center.6 As described by Cody et al ., the 80 kV tube voltage setting was eliminated from their institu tion because significant beam hardening artifacts were prevalent and affected the diagnostic quality of the images. They were, however, able to use 100 kV in place of 120 kV in patients up to 12 years of age for both chest and abdominal studies. In the case of an abdominal study, this reduction in tube voltage from 120 kV to 100 kV resulted in a 40% reduction in dose. This exam ple highlights the importa nce of recognizing the balance between dose and image quality. Dose re duction techniques fail if a scan needs to be repeated because of poor image quality. While patie nts should be exposed to the lowest doses as possible, images must be of diagnostic quality.

PAGE 44

44 Figure 3-1. Illustration of a ngular tube current modulation, depicting an increase in photon transmission in the AP direction and decrease in the lateral direction. Adapted from M.K. Kalra, M.M. Maher, T.L. Toth, B. Sc hmidt, B.L. Westerman, H.T. Morgan, and S. Saini, Techniques and applications of automatic tube current modulation for CT, Radiology. 233, 649-657.

PAGE 45

45 CHAPTER 4 COMPUTED TOMOGRAPHY DOSE DES CRIPTORS Because one aim of this research project wa s the measurement of organ doses, definitions of appropriate dose quantities ar e appropriate. Discussed in this chapter are the standard definitions of terms relevant to this project, as accepted by the appropriate bodies, mainly the National Council on Radiation Prot ection and Measurements (NCRP)19 and the International Commission on Radiologi cal Protection (ICRP).20 4.1 Definitions 4.1.1 Kerma Kerm a is an acronym for kinetic energy released in matter and describes the events that occur when ionizing radiation passes through matter.6 In the case of x rays, the two-step process begins with indirectly-ionizing photons transferring en ergy to kinetic energy of charged particles via Compton scattering and the photoelectric abso rptions. These charged particles, in turn, deposit energy in matter through ionization and ex citation. Kerma is the amount of kinetic energy transferred from the photons to direct ionization. The Standard In ternational (SI) unit of kerma is J kg-1, termed gray (Gy). From this definiti on, it is clear that kerma does not present information on the amount of energy tr ansferred to the irradiated matter. 4.1.2 Absorbed Dose (D) Absorbed dose is th e mean ener gy imparted by ionizing radiation (d ) to a given mass (dm).20 Absorbed dose is the most fundamental unit of radiation dose. The mathematical expression of absorbed dose is shown in Equation 4-1. The SI unit of abso rbed dose is also the Gy. dm d D (4-1)

PAGE 46

46 4.1.3 Equivalent Dose (HT) Equivalent dose was developed by the ICRP as a radiation protection quantity used for the purpose of defining limits of radiation exposure.20 The product of the average absorbed dose (i.e., the absorbed dose averaged over a vol ume) and a radiation weighting factor (wR) yields the equivalent dose, as shown in Equation 4-2. RTR R TDwH, (4-2) The radiation weighting factor describes the degree of bi ological damage produced by different types of directly and indirectly-ioni zing radiation, and it is based on the relative biological effectiveness (RBE),20 which is defined as the ratio of the absorbed dose from a reference radiation to that of a given radiation, both producin g the same biological effect.21 Photons have a radiation weighting factor of unity. 4.1.4 Effective Dose (E) Effective dose expresses the risk to an indi vidual m ember of a population exposed to a uniform whole-body irradiation that is equivalent to that of a pa rtial-body or organ exposure. Numerically, it is the summation of the individual products of the equivalent organ doses and the individual tissueand organweighting factors. The ICRP20 defines the tissue-weighting factor as the quantity expressing the contribution of a tissue or organ to the total body detriment from stochastic effects. Values of tissue weighting factors are listed in Tabl e 4-1. The expression of effective dose is given by Equation 4-3. RTR R T TDwwE, (4-3) 4.2 Effective Dose as a Dose Descriptor The effective dose has b een historically used to describe the dose associated with a radiation exposure. Thus, a brief discussion of the quantity, its usefulness, its limitations, and possible alternatives for medical applications is pertinent.

PAGE 47

47 4.2.1 Applications of Effective Dose As previously stated, the effective do se is the sum of the average absorbed dose to tissues and organs, weighted to reflect the degree of biological detriment to a given tissue or organ. Generally speaking, effective dose is used wh en comparing the relative biological radiation detriment among different radiological procedur es. Thus, effective dose may be a useful measure in the process of improvement of radi ological procedures that involve partial or nonhomogenous irradiations. 4.2.2 Limitations of Effective Dose as a Patien t Dose Descriptor in CT While effective dose is clearly a useful quant ity in many circumstances, especially when comparing the risk of biological detriment from different diagnostic x -ray procedures, one is tempted to apply its theory to doses for particul ar patients. Below are several arguments against this idea. 4.2.2.1 Population-based risk Because of its risk-bas ed defi nition, effective dose is releva nt when assessing risk to populations, while assessing the risk to individuals is limited by its scope. Such is the case in a diagnostic x-ray exam.22 Similarly, it is relevant only when the exposure conditions are the same as those under which the risk factors have been derived.22 For example, the ICRP formulation of effective dose was derived from considerations of both the working population and the general population which is not intentiona lly exposed. According to the definitions of these populations, a patient does not fall into either of these categories since a patient is intentionally exposed. In addition, not all population groups, including patien ts, can fit into broader definitions. For example, elderly patients receiving diagnostic radiology exams are at a lower risk of radiationinduced effects than the average population simp ly because their age allows less time for any effects to develop.23 On the other hand, pediatric patient s have a much higher risk than the

PAGE 48

48 general population for the opposite reason; their still-growing orga ns are more radiosensitive, and the probability of development of biological detrimental effects is greater because of their potentially longer life span than that of an average member of the public. 4.2.2.2 Tissue-weighting factors The ICRP defined the tissue weighting factors so that the factor associated with a giv en tissue or organ would represent the fraction that such organ or tissue contributes to the total biological effect to the body. These values are displayed in Table 4-1. Thus, the summation of the weighting factors is equal to unity to accoun t for the entire body. In the case of non-uniform, or partial-body irradiation, the definition of the effective dose is such that organs which are not irradiated do not contribute at all to the total body detr iment estimate. 4.3 Recommendations The ICRP states in its 2007 recomm endations20 regarding medical exposure of patients that either the equivalent dose or the absorbed dose to irradiated tissues or organs can be used as good estimates of radiation exposure to patients. Even in the case wher e effective dose is the most commonly used descriptor, th e ICRP states that it is more appropriate to determine organ doses separately, rather than simplify the situation by assuming a whole-body irradiation. Similarly, McCollough and Schueler22 suggest that the best ap proach to a patient dose estimate is to estimate separately and as accurately as possible all pertinent organ doses and apply the most appropriate risk factor for su ch organs. When necessary information is unavailable to conduct such a calculation, they s uggest that a Monte Carlo (MC) simulation be used. This approach is suggested because, if the definition is followed stri ctly, calculation of the effective dose requires direct measurement, or at least an estimate, of the dose to each individual organ and tissue. In practice, it is difficult to acc omplish this task. Dire ct measurement of organ and tissue doses is a practical impossibility since dose measurements in that case would have to

PAGE 49

49 be calculated or estimated, rendering the effective dose an estimate, rather than an assessment. Physical phantoms, as the one used in this rese arch project, typically are not full-body phantoms, or they do not facilitate placemen t of dosimeters in all organs. Thus, they limit the assessment of all organ and tissue doses. MC simulations, as s uggested above, can be util ized in place of actual measurements, but such an a pproach is outside the scope of this resear ch project. Thus, for the purposes of this work and unde r the conditions described, the assessment of effective dose is not only impractical, but it would lead to a large number of assumptions and approximations which would render the number meaningless. Instead, reliable organ dose measurements were made in accordance with the aims of this research project.

PAGE 50

50 Table 4-1. Tissueweighting factors.20 Tissue w T w T Bone-marrow (red), colon, lung, stomach, breast, remainder tissue* 0.12 0.72 Gonads 0.08 0.08 Bladder, esophagus, liver, thyroid 0.04 0.16 Bone surface, brain, salivary glands, skin 0.01 0.04 Total 1.00 Remainder tissues: adrenals, ex trathoracic (ET) region, gall bladde r, heart, kidneys, lymphatic nodes, muscle, oral mucosa, pancreas, prostate, sm all intestine, spleen, thymus, uterus/cervix.

PAGE 51

51 CHAPTER 5 COMPUTED TOMOGRAPHY DOSE INDEX Historically, doses in CT have been expre ssed by and referenced to a quantity known as the computed tomography dose index (CTDI). Therefore, a description of the CTDI and its various definitions, uses and limitations are presented in this chapter. 5.1 CTDI Definitions The CTDI was originally proposed by Shope et a l .24 in 1981 as a method to standardize CT dose estimates. At the time, there were many m easurements and techniques available; Shopes goal was to set a benchmark. There are also ma ny factors that contribu te to the complex dose distributions in CT. Scan motions, filtration, vary ing detector collimations and scan parameters all differ among manufacturers. As a standard method was necessary, the CTDI (without its numerous subscripts) was proposed as a quick and convenient estimate of dose from a CT scan, so that the dose from different scanners could be compared in a consistent manner. 5.1.1 MSAD The f irst quantity specifically designed to represent the radiation dose resulting from exposure of the patient to the rotating x-ray b eam in the geometrical conditions of CT was the multiple scan average dose (MSAD). The MSAD de scribes the limiting value of a dose profile obtained when multiple single-slice axial CT scans are acquired over a given length. As the number of scans increases, the subsequent scatter tails add to the dose on the central axis; the MSAD is approached when the distance between the first and last scans of the series is large enough so that these two scans so not contribute to the dose at the central axis. The MSAD was developed to describe the complex dose distribu tions of clinical scan s; measurement of the MSAD along the length of a scan consisting of multip le rotations of the x-ray tube describes the

PAGE 52

52 dose distribution along the length of the scan. Numerically, Shope24 defines the MSAD in Equation 5-1 as, (5-1) where I is the distance between th e central axis of each scan N is the number of scans. 5.1.2 CTDI The CTDI is defined as the dose as a functi on of position along the Zax is co-ordinate for a single scan dose profile at a given point.24 The numeric integral is inversely proportional to the slice thickness, integrated over infi nity and expressed in Equation 5-2, as dzzD NT CTDI )( 1 (5-2) where N is equal to the number of tomography sl ices in the acquisition, and T is equal to the width of the scanning beam in the Z-direction. Th e product NT is equivalent to the nominal slice thickness. The CTDI is then equal to the av erage dose along the Z-axis at a point over the central scan of a series of scans,24 as long as the distance between each scan is held constant. This definition ensures that if enough scans are acquired in such a way that the contributions from scatter from adjacent slices is included, th e dose profile levels out and remains consistent throughout the scanned length, regardless of the number of scans further added.24 CTDI measurements are much more convenient than the numerous point dose measurements required for an actual calculation of the MSAD. 5.1.3 CTDIFDA Since the time Shope initially introduced the concept of the CTDI, further modifications to the CTDI have been defined, each with a different motivation and purpose. The first modification was made by the Food and Drug Ad ministration (FDA), as it formally adopted 2/ 2/ ,)( 1I I INdzzD I MSAD

PAGE 53

53 CTDI as the parameter to describe doses in CT, realizing that the integral over infinity in the definition of the CTDI was not practical, and also that the dose was dependent on the scan thickness and the sc attering medium.25 In establishing regulatory definitions on how to measure the CTDI, two cylindrical phantoms were introduced: one with diameter of 16 cm to represent an average adult head, and the other 32 cm in diamet er to simulate attenuation and scatter from an average adult torso. Both phantoms are made of polymethyl methacrylat e, better known as PMMA, and 15 cm in length. These phantoms are de picted in Figure 5-1, along with rods used to fill cylindrical sockets in the phantom, created for placement of a dosimeter. At the time, the length of these phantoms guaranteed that all s cattered radiation within the phantom (resulting from a thin axial slice) would be included in the measurement. A standard pencil chamber with a 100-mm active length was used in conjuncti on with these phantoms to measure CTDIFDA, defined in Equation 5-3 below. T T FDANT CTDI7 7D(z)dz 1 (5-3) The limits of integration were assigned as 7T because the maximum beam width used in CT at the time was 10 mm, thus ensuring that all of th e scatter produced with in the phantom from a single 10 mm axial slice was collected by the pencil chamber. However, as CT evolved into helical first, and late MDCT and systems with greater beam widths, these integration limits became obsolete. 5.1.4 CTDI100 With the objective to extend the applicability of the CTDI concept to varied slice thicknesses and scanners, the CTDI100 was defined. Because the pencil chamber specified by the FDA has an active length of 100 mm, the CTDI100 is defined with limits of integration that reflect this length, as shown in Equation 5-4.

PAGE 54

54 mm mmdzzD NT CTDI50 50 100)( 1 (5-4) While this equation was meant to estimate the MSAD (section 5.5.1), CTDI100 underestimates the MSAD when beam widths wider than 10 mm are used because not all of the scatter tails of the scan are collected and meas ured by the pencil chamber. 5.1.5 CTDIw Due to the geometry of a CT acquisition, th e dose distribution is not uniform across a single slice. Except at the center of the cylindrical CTDI phantoms, the attenuation characteristics at any other points on the phant om are different, because of the additional attenuation resulting from the patie nt table and other factors. Clear ly, such is also the case with a patient. To account for these variations a weighted CTDI value was defined26 as indicated in Equation 5-5. edge center wCTDI CTDI CTDI,100 ,1003 2 3 1 (5-5) The first term in the equation refers to the CTDI100 value measured in the center of the PMMA phantom and the latter to the average of the measur ements made at the four peripheral positions. The weighted CTDI is often used as an indi cator of dose across the scanned field-of-view, because the CTDI measured at different positions within the phantom can vary.11 5.1.6 CTDIvol The advent of helical scanning made it clea r that none of the CTDI descriptors would adequately describe the dose resulting from a CT helical acquisition that may use a pitch value other than 1.0. Pitch is defined as the ratio of ta ble travel during one rotati on of the x-ray tube (I) to the total beam width (NT).11 In this way, application of any of the CTDI definitions would yield the same numeric result regardless of wh ether an acquisition resu lted in over-scanning

PAGE 55

55 (pitch less than 1) or under-sca nning (pitch greater than 1). C onsequently, a modified definition called CTDIvol was introduced as stated in Equation 5-6. w volCTDI I NT CTDI (5-6) With the definition of pitch as the table incr ement per axial scan, a further simplification to Equation 5-7 is made. w volCTDI pitch CTDI 1 (5-7) 5.2 Dose-Length Product By def inition, all but the original CTDI definitions are based on measurements on a standard phantom. Thus, the values so obtaine d are not representative of the actual dose deposited in a volume of a different size, shape, or scattering material than the PMMA head and body phantoms, nor can the CTDI values be consider ed accurate when not a ll of the scatter tails are accounted for or measured. As a dose estima te for a 100 mm scan (by virtue of the length of the pencil chamber), such CTDI values are typically applied for a clinical scan, whether its scanned length may be at of 10 mm or 1000 mm. In other words, the length of a clinical scan is not taken into account when using the CTDI as the dose descriptor. To rectify this situation, the dose-length product (DLP) was developed. The DLP, in units of mGy-cm, is described in Equation 5-8, where L is the scan length, in cm and the CTDI100 is measured in mGy. DLP = CTDI100 L (5-8) The evolution of the CTDI, and all of its asso ciated derivations, has spanned over the past three decades. As CT technology has continued to advance, it has been necessary for the imaging community to adapt the definition of CTDI to best describe the actual dose that results from a CT scan. Because of its simple measurement process, CTDI remains an acceptable and

PAGE 56

56 efficient tool in comparing the output of different CT scanners and to track the performance of a scanner over time. However, it presents certai n limitations, as described in the next section. 5.3 Analysis of the Applicability of the Measurement of CTDI and DLP In recent years, there has been m uch disc ussion within the radiology community regarding the measurement of the computed tomography dose index (CTDI) with a standard pencil ionization chamber. The pencil chamber wa s originally proposed by Suzuki and Suzuki27 in 1978 as a convenient way to measure the radiation output of a CT scanner in such a way that geometry and scatter would be included in the measurements. Such measurements would also allow for a relative comparison among scanners, as the measurement is representative of an average dose along the length of the chamber. The chamber was designed to be used the the PMMA CTDI phantoms. While the original con cept behind the pencil chamber was adequate at the time it was proposed, the advancements in CT technology have brought up many limitations to the pencil chamber and the definition of CTDI. 5.3.1 Underestimation of Dose using CTDI in MDCT In the early days of CT technology, the m a ximum 10 mm x-ray beam width was narrow enough that all scatter tails were indeed captured by the pencil cham ber in a CTDI measurement. When the FDA25 adopted the CTDI concept, the limits of integration were modified to T (section 5.1.3), which still encompassed the scatte r tails of such narrow beams. However, CT technology has rapidly advanced: the number of slices per acquisit ion has increased, as have the maximum beam widths, a situation not contem plated in the various CTDI definitions. As MDCT scanners allow wide beams, Di xon has shown that for a given 20-mm total beam width, a single axial measurement using a pencil chamber underestimates the dose in the central region of the body and head CT DI phantoms by 20% and 10%, respectively.28 He further suggests that these discrepancies will only increase as the total beam width increases.

PAGE 57

57 Boone29 addressed the issue of increasing beam widths in a 2007 study that focused on the efficiency of the 100 mm length pencil chamber in capturing and accurately measuring the total dose deposited in a single axial CT scan of beam widths of 10, 20 and 40 mm. He defined the efficiency of the CTDI100 value as the ratio of the dose de posited in a rod 100 mm in length (representing a pencil chamber) to the total dose deposited in a r od of infinite length. Using a Monte Carlo code to simulate energy deposition in the PMMA CTDI head and body phantoms, the CTDI100 efficiencies in the center and peripheral holes of the head phantom were 82% and 90%, respectively, for a 120 kVp x-ray spectrum simulated. The corresponding efficiency values for the body phantom were 63% (center) and 88% (periphery). While these numbers highlight the vast ineffectiveness of CTDI100 as a true dose descriptor, ther e was only a 1% decrease in the efficiency values when a 40 mm be am width was simulated. This result indicates that the dose measured by a pencil chambers for wide-beam MDCT systems is not representative of the actual dose. 5.3.2 Proposed Solutions To address som e of the inadequacies of th e current measurement of CTDI, several groups have suggested different methodologies. The tw o main concepts considered include the design of phantoms of a length greater than 150 mm to provide sufficien t scattering material to better represent a patient more closely and a sm all ionization chamber to measure dose. 5.3.2.1 Extended phantoms The 2003 Dixon28 study suggests that increased phantom lengths appear to be necessary in order to produce equilibrium scatter; however, an act ual length is not suggested in that work. To this end, Mori et al.30 manufactured phantoms of 900 mm le ngth using the same PMMA material as suggested by the FDA.25 By definition, the integration length of CTDI100 is a total of 100 mm (50 mm on each side of the central Z=0 axis). Instead, this study used a 300 mm length pencil

PAGE 58

58 chamber to investigate whether a longer integr ation range would accurately capture all of the dose profile (including scatter ta ils) of CT x-ray beams of increasing widths. The authors concluded that for nominal beam widths of 15 mm or greater, a standard length (150 mm) body phantom estimates the dose profile integral to be 84% of the true value when compared to that measured using a 900 mm length phantom. Be cause a 300 mm length pencil chamber was used in this study, the authors further suggest an integration range of 300 mm for the beam width investigated (20 mm) to estimate bette r an average dose profile integral. 5.3.2.2 Small-volume ionization chamber In order to address this undere stim ation of dose, Dixon and Ballard31 make a recommendation to replace the pencil chamber with a small volume (0.6 cm3) Farmer chamber. The measurement is then performed on the central axis of either the body or head phantom, which are placed on the table of the CT scanne r and allowed to translate through the scanner during the exposure. In this way, the chamber is directly measuring the dose accumulated at the center of the phantom as a result of a CT scan of a given length. Furthermore, the small ion chamber guarantees that the scan length is equal to the integration length. Chapters 4 and 5 have clearly established th e limitations of the vari ous CTDI incarnations to describe, properly and faithfully, the doses re sulting from the wide-beam MDCT scanners, as well as the practical limitations of measuring e ffective doses, which justifies one of the main aims of this research project to measure indivi dual organ doses in physical phantoms as the best alternative to dose characterizati on in MDCT. However, not all of the concepts associated with the CTDI and effective dose are to be discar ded, especially those which focus on scatter and beam width considerations. Fo r example, the concept of the small-volume ion chamber as the most adequate way to perform measurements b ecause it takes into account these issues was applied in this project by approximating the small-volume ion chamber with a commercially-

PAGE 59

59 available, small-volume OSL dosimeter to estimat e organ doses. The dosimeters approximate a small-volume ion chamber and are capable of measuring the dose accumulated in the organ of interest.

PAGE 60

60 Figure 5-1. Standard CTDI phant oms. From left to right: body phantom, cylindrical rods and head phantom.

PAGE 61

61 CHAPTER 6 INSTRUMENTATION AND PHANTOMS This final chapter of background m aterial is dedicated to the description of the dosemeasuring devices and phantoms used throughou t the course of this research project. Specifically included are the details of the various ionization chambers and a solid-state detector used for both the characterization of the CT x-ray beams and the verification of the OSL dosimetry system, as specified in aims (b) and (c ) in Section 1.3. Also, a brief discussion of the different types of dosimetry phantoms, including the tomographic phantoms used in this research project, is given in this chapter. 6.1 Ionization Chambers Ionization cham bers are the standard inst rumentation used by medical physicists in performing all types of measurements with ionizing radiation. The basic details of the principle operation of an ionization chamber are outside th e aims of this project and are found elsewhere.32 Listed in this section are the ionization chambe rs, along with manufacture r-stated specifications and other information pertinent to th eir use in this research project.33 6.1.1 A 6-cc Chamber A 10X6-6 (Radcal, Monrovia, CA ) ionization cham ber with an active volume of 6 cm3 was used to measure air kerma. The usable range for the chamber is 0.01 Gy to 600 Gy, well within the limits of air kerma ranges measured througho ut this research project. The chamber is calibrated by the manufacturer at 60 kVp, 2.8 mm of aluminum (Al) half-value later (HVL) and has an accuracy of %. The nearly-flat en ergy response of the chamber yields correction factors of approximately 0.98-1.00 from 20 keV to 150 keV. While this chamber is most commonly used to measure kerma in the diagno stic energy range, the design and size of the

PAGE 62

62 chamber, when compared with those of the OSL dosimeters, limited its use in this research project. 6.1.2 A 3-cc Pencil Chamber For m easurements involving the PMMA CTDI phantoms, a 3 cm3 pencil chamber was used (10X6-3CT, Radcal, Monrovia, CA). This chamber was calibrated at 150 kVp and 10.2 mm Al HVL and has a manuf acturer-specified accuracy of %. Similar to the 6-cc chamber described in section 6.1.1, the energy dependence is % with a range of correction factors ranging from 0.98 to 1.02, for HVLs of 2 to 20 mm Al. The active length of this chamber is 10 cm and the reported dose is the integral dose ove r the active length. As discussed in Section 5.3, this 10-cm length does not capture all of the scatter tails that result from CT beam widths that are common today. Furthermore, a disadvantage of this chamber is the inability to assess CTDI on modern scanners that do not al low the user to utilize narrow beam widths necessary (i.e., 10 mm) to ensure that all scatter tails are captured by the pencil chamber. 6.1.3 A 0.6-cc Chamber Because of the sm all size of the OSL dosimeters, a 0.6 cm3 chamber (10X6-0.6, Radcal, Monrovia, CA) was purchased for the purpose of comparing and benchmarking measurements made with such dosimeters under similar geometry conditions. This Farmer-type chamber was specifically calibrated by the manufacturer to th e National Institute of Standards and Technology (NIST) beam code M100, which is defined as fo llows: moderately-filter ed 100 kV beam with a 5.25 mm Al HVL.34 This beam code was chosen because it best matched the characteristics of the CT beams used in this research project. The manufacturer-stated energy dependence of this chamber is % from 40 keV to 1.33 MeV, resu lting in correction factors ranging from 0.97 to 1.03 for the 20 keV to 500 keV subrange.

PAGE 63

63 6.2 Solid-State Detector A Barracuda X-ray Multim eter (RTI Electronics Inc., Fairfield, NJ) was used to measure the total filtration of the CT x -ray beam (described in section 7.2.2). This measurement device, in addition to the Ortigo software (RTI Electronics, Inc., Fairfield, NJ) used with it, gives a direct read-out of the total filtration of the x-ray beam, in mm of Al, making such measurements quick and easy. The accuracy of this measurement is %, or 0.3 mm, whichever is larger, in the range of 60-120 kV.35 6.3 Tube Voltage Meter A 40X12W Accu-kV sensor (Radcal, Monrovia, CA) was used to directly measure the accuracy of the nominal kV of the CT x-ray be ams. This sensor has a manufacturer-stated accuracy of kV, or %, whic hever is larger. The range of this meter is 40 kV to 160 kV, adequate to cover the range of tube vo ltages used in this research project. 6.4 Dosimetry Phantoms Over the years, a large variety of phantom s ha s been used for dosimetric purposes. There are two main types of phantoms used for estima ting doses from radiological procedures. Both have advantages and disadvantages. The followi ng sections describe these phantoms and justify the decision for using the anthropomorphic phantoms for this research project. 6.4.1 Stylized Phantoms Stylized phantom s are designed based on simple mathematical shapes. They use threedimensional surface equations to describe the main organs that make up the human body by using simple, general shapes. The organs are represented in generic, often realistic positions and orientations. For example, a leg is represented with a cylinder of a given length, while a sphere is used to represent the head. Because of the lack of detail and generic nature of these shapes, one advantage of these phantoms is that they can be used to sufficiently represent an average

PAGE 64

64 adult. While this type of phantom may be adequate in some equivalent dose estimates, this lack of detail and the inadequate locations and orientat ions of the organ can be a source of error when estimating organ doses.36 6.4.2 Tomographic Phantoms Anthropom orphic tomographic phantoms can be developed using image sets from actual scans of patients from tomographic modalities su ch as CT and MRI. In this way, the actual shapes of organs and structures, and their relati ve positions and orientations, are captured by the image sets. Segmentation methods are generally used to define the contours of the organs from the image set. Throughout the past decade, a multi-disciplinary group at UF as developed a series of tomographic, computational and physics phantoms spanning various age ranges and including both genders.37,38 These unique tissue-equivalent phantom sets were generated from actual patient CT data and we re physically constructed with tissue-equivalent materials.39 Because of these attributes, the phantoms are tom ographic in nature, and they are clearly wellsuited for the measurement of orga n and tissue doses in CT. The re solution of all organs defined in the phantom are limited only by the resolution of the CT or MR image sets from which the organ data were derived.36 Conversely, the smaller voxels that provide good resolution negatively affect the signal-to-noi se ratio (SNR) of the image. When a smaller number of x-ray photons (or hydrogen nuclei in the case of MR) is captured with in a volume element (voxel), there is an increase in the noise of the image. The result is a grainy image which leads to less accuracy during the segmentation process. Furthermore, because tomographic phantoms use patient image sets, the phantoms created are more specific to the few patients from which the CT or MR data were acquired. While one may choo se an image set that closely represents an average patient, the organs may not be as average in size and location as those defined in a

PAGE 65

65 stylized model. However, the detail of the or gans and their relative positions among each other in tomographic phantoms makes them the phantom of choice for organ dose estimates. Two tomographic physical phantoms were used for organ dose measurements in this research project: a pediatric phantom, depicted in Figure 6-1 and an adult male phantom, shown in Figure 6-2. Details regardi ng the construction process of these phantoms, as well as the tissue-equivalent materials used for construction are given in Appendix B. 6.5 Phantom for Image Quality Evaluations There are a num ber of methods used by medi cal physicists to asse ss image quality in a clinical setting. Certai n quantitative measurements, such as signal-to-noise (SNR), can be made using data analysis software available on some im age display and viewing stations. On the other hand, more qualitative measurements can be ma de using phantoms specifically designed for image quality assessment. Since both of these ap proaches are used, a br ief discussion of their advantages and disadvantages is given, as well as justification for the method chosen for this research project to assess low-contract detectability. A quantitative approach to image quality is described by Rill et al.,40 in which acrylic and aluminum phantoms were developed to approxim ate the attenuation characteristics of an adult chest. SNR measurements were made in th ree different locations in these phantoms, corresponding to landmarks of patient anatomy, and using a soft-copy display workstation to assess image quality in the clinical environment. A circular region of in terest (ROI) was drawn at five locations and the mean pixel value was re corded for each of them. Noise was defined as the standard deviation of the pixe l values of all five ROIs. Imaging parameters were assessed by calculating the differences in SNR from images acquired with different techniques. Such a method allows for image quality to be assessed without exposing patien ts to excess radiation

PAGE 66

66 because phantoms are used instead. However, in order to implement dose-reduction techniques suggested by the authors, a follow-up evaluation of clinical images would likely be necessary. In a more qualitative approach, Prasad et al.41 acquired CT chest images which were analyzed by experienced radiologist s. A 5-point scale was used to assess the images for noise, contrast, sharpness and overall quality. A score of 0 was given to images considered to be of the worst quality, while a score of 5 was assigned to im ages of excellent quality. This method yields results that may be put into pr actice immediately because clinical patient images are used for analysis. However, this type of study require s approval by the Institutio nal Review Board (IRB) and, in this specific case, required that patients receive an additional amount of radiation because additional scans were acquired in order to comp are standard CT imaging techniques with the low-dose techniques proposed by the authors. The method chosen to assess image quality fo r two of the protocols in this study was similar to that described by McCollough and Zink.42 A phantom designed and manufactured specifically for image quality analysis in CT was used in their evaluation of the performance of a multi-slice CT system. The phantom was centere d and aligned with the CT scanner gantry, images were acquired using various combinations of scan parameters and resulting images from different modules of the phantom were assessed. While the tomographic phantoms used in this research project in clude many organs and tissues, they are limited from an image quality perspective because the phantoms are made of three tissue-equivalent materials. While this leve l of detail is adequate for the purpose of dose measurements, it was not adequate to evaluate im age quality because of the lack of contrast among different tissues.

PAGE 67

67 To assess image quality of two of the clinical protocols evaluated in th is research project, as described in Chapter 9, the low-contrast detectability (LCD) module of an image quality phantom (Catphan 500, The Phantom Labor atory, Salem, NY) was utilized.43 The phantom contains various sections to assess parameters such as slice thickness and spatial resolution, among others. In particular, the LCD section is 15 cm in diameter and 4 cm thick. Cylindrical rods are embedded in two circular patterns into 6 groups. Each of the outer three groups contains 10 rods of varying diameter: 2, 3, 4, 5, 6, 7, 8, 9, 10 and 15 mm, respectively. Each group has a different contrast level: 1%, 0.5% and 0.3%, and labeled A, B and C, respectively. All rods are 40 mm in length and, thus, these groups are called supra-slice groups because this length is wider than most beam widths in MDCT, excluding, of course, the wide-beam scanners wich as the Aquilion ONE 320-slice scanner. The inner thre e groups, known as the s ub-slice groups, each have 4 rods of varying diameter: 3, 5, 7 and 9 mm. These all have an identical 1% contrast level. Unlike the outer groups, the sub-slice groups cont ain rods of varying le ngth: 3, 5 and 7 mm, and are labeled D, E and F, respectively. MDCT scanners with 16-slice capabilities and up clearly have beam widths larger than these sub-slice groups. As example of a CT image of the LCD section of the Catphan is presented in Figure 6-3. Typical LCD assessments are done in a subjective observation manner. Images of the LCD sector are obtained using kV and mAs values as predetermined by the observer, using the thinnest acquisition slice possi ble and reconstructed using a soft-tissue algorithm and a reconstruction thickness also predetermined by the observer. Sc oring of the phantom is done under dimmed-light conditions and optimal window and level settings, as determined by the user. The number of objects resolved completely (i.e., those for which th e circular margins are

PAGE 68

68 evident) are identified and counte d. The results can be presented in various ways, as described in Chapter 10.

PAGE 69

69 Figure 6-1. Pediatric tomographi c phantom, developed from CT data of a 9 month-old patient.

PAGE 70

70 Figure 6-2. Adult male tomographic phantom.

PAGE 71

71 Figure 6-3. Diagram of the low-contrast module of the image quality phantom. Sections A-C are the supra-slice groups and D-F are th e sub-slice groups. Adapted from Catphan 500 & 600 brochure, The Phantom Laboratory. Available at http://www.phantomlab.com/catphan.html

PAGE 72

72 CHAPTER 7 CHARACTERIZATION OF X-RAY BEAMS OF THE VOLUMETRIC CT SYSTEM Only a handful of 320-slice s canners are currently in use fo r both clinical and research purposes worldwide. No publicatio ns describing the com plete char acteristics of the x-ray beams for these scanners exist to date. Thus, a ch aracterization of the MD CT systems (Aquilion ONE and Aquilion 64, Toshiba America Medical Systems, Tustin, CA) used for this research project was necessary. Though the evaluation of the Aqu ilion 64 was rather limited in scope (HVL and kV accuracy), that for the Aquilion ONE include d exposure reproducibility, kV accuracy, total filtration, HVL, beam width, and dose profile m easurements. For some of these measurements, specifically the beam width and dose profile de terminations, well-known established methods for scanners of narrower beam widths were modifi ed to accommodate the 160 mm full beam width of this scanner. The results of this characterization are pr esented in Section 10.1. 7.1 Exposure Reproducibility 7.1.1 Service Mode In addition to the m easurement of organ doses for selected clinical protocols using the 320slice scanner, as described in Chapter 9, the scanner was also used to verify the dosimetry system (Section 8.2). During the dosimetry system ve rification, the scanner was mostly operated in service mode, which allows the user greater flex ibility in the selection of exposure parameters, and most importantly, the capability of making exposur es with the x-ray tube in a stationary (i.e., non-rotating) mode, a feature un available in clinical mode. To investigate exposure reproduc ibility in this service mode, 0.6 cc ionization chamber, described in Section 6.1.3, was placed on 5.1 cm of acrylic slabs at the ce nter of the CT x-ray beam. Ten exposures were made using identi cal scanning parameters, and the air kerma was measured. The mean and standard deviation of the ten measurements were calculated.

PAGE 73

73 7.1.2 Clinical Mode W hile most of the scan parameters of a clinical protocol are either se lectable or visible to the user, one that remains unknown in present-day scanners is the actual position of the x-ray tube (i.e., the angle in the ga ntry) when the exposure actually begins. The three clinical protocols studied as part of this research project described in Chapter 9, include either a single axial rotation of the x-ray tube to acquire one volume of data, or multiple axial rotations without any table motion to acquire severa l volumes. Some protocols also include helical scans, and so this type of acquisition was also reviewed. Reproducibility in the clinical mode was asse ssed using two types of clinical protocols: volumetric (Craniosynostosis, Section 9.1.1) and helical (Head without Contrast). For both protocols, the CTDI head phantom was placed in the center of the gantry. The 3-cc pencil chamber described in Section 6.1.2 was used to measure the dose at all positions within the phantom. Ten replicate measurements were made at each phantom position for the volumetric protocol, while three were made for the helical protocol. The m ean and coefficient of variation (CV) were calculated for each set of measurements. 7.2 Beam Quality The quality of an x-ray beam affects both th e patient dose and imag e quality, the two main aspects of interest in a diagnostic imaging pr otocol. Beam quality is characterized by tube voltage, total filtration and HVL. X-ray tube vo ltage was measured at the four tube voltage settings available on the sca nner: 80, 100, 120 and 135 kV. Total filtration measurements were made to assess the effects of the different types of bowtie filters used in the CT system. HVLs were measured at all beam energies a nd using three different bowtie filters.

PAGE 74

74 7.2.1 Accuracy in kV Because pho toelectric absorption and Compton scatter are energy-dependent processes, the attenuation characteristics of a tissue are quantita tively represented by the attenuation coefficient of the tissue. This quantity is dependent on th e energy of the x-ray beam that passes through the tissue. In general, when a lower tube voltage is selected for a given exposure, the overall lower energies present in the x-ray spectrum result in more tissue contrast in the image because the relative differences among attenuation coefficients are larger than in th e case of higher energy spectra due to the larger number of photoelectric absorptions wh ich take place at lower energies.6 On the other hand, when lower tube voltage s are used, images suffer from decreased transmission of x-ray photons through the irradi ated material as compared to higher tube voltages, thus, resulting in lower photon transm ission and a smaller amount of photons reaching the detector elements. Because of these reasons, the importance of the accuracy of the voltage selected cannot be overstated. The accuracy of the nominal tube voltages was measured using the kV sensor described in section 6.3 with associated radiation measurement system (9095, Radcal Corporation, Monrovia, CA). The sensor was placed at the scanner isocenter, and the x-ra y tube was positioned stationary at the top of the ga ntry (12 oclock position), as de picted in Figure 7-1. For each voltage available, an exposure was ma de, and the measured kV was recorded. 7.2.2 Beam Filtration in CT Due to the circular geom etry of a CT scanne r, the overall beam attenuation is non-uniform as the x-ray beam passes through th e anatomy of interest during a s can. Thus, the x-ray beam is non-uniformly attenuated and hardened as it rotates around the patient.9 The central rays of the fan beam pass through the middle, thickest section of the body and experience greater attenuation and hardening. The outer rays of the fan beam pa ss through the edges of the body,

PAGE 75

75 where the thinner anatomy resu lts in less attenuation and ha rdening. This difference in attenuation and hardening of the x-ray beam negatively affects the reconstruction of the image. In addition, projection measurem ents through the center of the patient result in a decreased number of transmitted photons as more photons are attenuated through its thick section. The lower number of detected photons for those projecti ons, as compared to th e peripheral sections, results in a decrease in the signal-to-n oise ratio. In order to co mpensate for the non-uniform beam hardening and attenuation resulting from differences in transmitted intensities, a beam compensating filter is placed close to the x-ray tube within the gantry. This bowtie filter,9 named after its particular shape, effec tively attenuates and hardens the x-ra y beam to a greater degree at the edges of the beam, as compared with effect at the center of the beam, where the distance between the x-ray source and the subject is th e shortest possible and the amount of attenuating material is the largest. Several bowtie filters are found in MDCT scanners today; clinicall y, the selection of a particular bowtie filter is depe ndent on the tube voltage and mo re importantly, the field-of-view (FOV) selected. Clearly, larger FOVs require a wider filter to cove r the entire fan beam scanning the anatomy. Correspondingly, a small filte r is used for brain imaging because of the small diameter of the head. A larger filter is used for cardiac imaging because the diameter of the thoracic region is much greater than that of th e head and therefore requires a larger filter to provide uniform beam characteristics. An attempt was made to obtain specific inform ation, such as the materials of these bowtie filters, but such information was deemed as pr oprietary by the manufacturer in spite of a cooperative agreement being in place and a non-di sclosure agreement signed by the author. In order to investigate the characteristics and effects of these filters, the total filtration of the x-ray

PAGE 76

76 beam was measured at points ac ross the gantry using a solid-stat e detector described in Section 6.2. The CT x-ray beam was collimated to approx imate the size of the detector. An initial measurement was made at isocenter; the detector was then moved in 2 cm increments across the gantry out to 24 cm from isocenter along the diamet er of the gantry aperture. Because filters are assumed to be symmetric, only one set of measurem ents was made across one half of the gantry. Measurements were made at three nominal tube voltages: 80, 100 and 120 kV and for two filters: Small-S and Large-L with one exception; due to limitations of the solid-state detector, measurements could not be made using the Sma ll-S filter at 80 kV along the gantry, with the only successful measurement at isocenter. 7.2.3 Half-Value Layer The HVL of an x-ray system is defined as the thickness of a given material needed to attenuate the x-ray beam to one-h alf of its initial intensity. A 6-cc ionization chamber, described in Section 6.1.1, and readout un it (9095, Radcal, Monrovia, CA) wa s used to measure the air kerma in a standard HVL measurement procedure. The x-ray tube was positioned at the bottom of the scanner gantry (6 oclock) and held statio nary during each exposure. A lead shield with a small aperture was placed in the CT gantry betw een the x-ray tube and the ionization chamber to collimate the beam to approximately the size of the chamber in order to attain good geometry conditions. Standard first HVL measurements were performed: the in itial measurement was made without filtration in the beam, followed by exposures made with increasing amounts of 1100 aluminum alloy filters of known thickness placed on top of th e lead collimating device. Measurements were performed at 80, 100, 120 and 135 kV for the Small-S, Medium-M and Large-L filters in place for each of the tube voltage s. HVLs were measured at the center of the field, i.e., corresponding to the center of the bowtie filter.

PAGE 77

77 7.3 Beam Width Actual beam width measurements were conducted for a 160 mm nominal width using a computed radiography (CR) imaging detector (AGFA CR MD4.0, AGFA, Teterboro, NJ) 35 cm x 43 cm in size. A 1 mm copper filter was placed between the x-ray tube and imaging plate to attenuate the x-ray beam and bett er match the exposure expected by the CR imaging plate. The image was processed by a CR digitizer (CR85-X, AGFA, Teterboro, NJ) using a flat-field algorithm and analyzed using the Image J (National Institutes of Health, Bethesda, MD) software analysis program. A beam profile was generate d across the mid-point of the CT x-ray beam. The gray scale value was plotted as a function of distance. The beam width was calculated using the full width at half-maxim um (FWHM) of the profile. 7.4 Dose Profile 7.4.1 X-Axis In order to m easure the intensity of the x-ray beam and create a beam profile, the Landauer OSL nanoDot dosimeters (described in Chapter 8) were positioned suspended at isocenter across the gantry opening of the CT scanner. Dosime ters were placed adjacent to each other and centered vertically in the gantry using the alignm ent lasers, as depicted in Figure 7-2. The x-ray tube was positioned at 90 in the gantry (3 oclock position) and kept stationary during the exposure; the orientation of the x -ray tube with respect to the st rip of dosimeters was intended to create a beam profile (x-ray intensity as a function of distance) in the scanning plane of the fan beam. The dosimeters were exposed and an alyzed by the OSL reader (Chapter 8). 7.4.2 Z-Axis The nanoDot dosim eters were positioned at isoc enter in order to measure the x-ray beam profile in the anode-c athode direction of the x -ray tube and determine if any heel effect could be observed. The heel effect is an intensity grad ient which is the result of the preferential

PAGE 78

78 attenuation of the x-ray beam towards the anode side of the x-ray fi eld by the anode itself.6 The angulation of the target causes photons to travel different distances through the anode towards the image detector and therefore the intensity is lower at the anode side than at the cathode side. Nineteen dosimeters were placed in line to cover approximately 19 cm of the beam which had a nominal width of 160 mm (0.5 mm x 320 detector rows). One row of dosimeters was centered horizontally in the gantr y. Other rows were placed 8 cm above and below the center line. This setup is depicted in Figure 7-3.

PAGE 79

79 Figure 7-1. Representation of x-ray tube positio ns used for measurements. The star-like shape represents the x-ray tube.

PAGE 80

80 Figure 7-2. Dosimeters suspended perpendicula r to the anode-cathode axis, spanning only the top half of the gantry because of symmetry of the fan beam.

PAGE 81

81 Figure 7-3. Setup to measure beam profile parallel to anode-cathode axis.

PAGE 82

82 CHAPTER 8 OPTICALLY-STIMULATED LUMINESCENT DOSIMETERS For the pas t few years, optically-stimula ted luminescent (OSL) dosimetry has been adopted as the technology of c hoice in personnel dosimetry, a nd its use has now extended to clinical applications.44-47 The OSL process is similar to that of thermoluminescence (TL), with the primary difference that an optical stimulus is used in OSL instead of a thermal or heat stimulus (as is the case in the TL process) to i nduce post-exposure lumines cence of the irradiated material. In the typical OSL reader, after i rradiation, light emitting diodes (LEDs) supply the optical energy which releases the OSL signal, wh ich is proportional to th e dose absorbed by the dosimeters. This chapter details the methods of verifying the use of this system in the lowenergy radiation fields used in diagnostic radiology. 8.1 Optically-Stimulated Luminescence OSL has recently been evaluated for its use in do simetry, as it offers great advantages over other conventional dosimetry systems. Only a small amount of OSL material is needed to capture sufficient numbers of photons during ex posure to ionizing radiation, making small dosimeters possible, a useful feature for radiatio n dose estimates. Other positive characteristics include reusability, no observab le dose-rate dependence and good spatial resolution. Jursinic conducted a thorough study in 2007 characterizing OSL dosimeters made by Landauer, Inc. (Glenwood, IL) for the high-energy radiati on fields used in radiation oncology.45 Among the properties he studied were angular and energy res ponse, erasure using diffe rent light sources and signal depletion with repeated exposures. All measurements were condu cted using x-ray beams with energies of 6-15 MV produced by a linear accelerator. Several others have performed studies similar to Jursinic utilizing high beam energies used in radiotherapy.48,49 The results reported by Jursinic and others imply that these OSL dosimeters are good candidates for

PAGE 83

83 radiotherapy dosimetry. However, no similar st udy has been conducted to date in the lower energy range used in diagnostic radiology to de termine whether the same advantages of OSL technology used in radiotherapy ap ply in diagnostic radiology. 8.1.1 Disadvantages of TL Dosimeters While TLDs (therm oluminescence dosimeters ) have been common in dosimetry for decades, there are many disadvantages to their use in clinical radiology.47 It has been shown that the sensitivity of a TLD is dependent on the heat rate at which is it stimulated post-irradiation; as the heat rate increases, the sensitivity decreases. In addition, a lengthy annealing process is associated with each use of a TLD,46 which limits practicality in cl inical situations. Finally, correction factors need to be applied for char acteristics such as a non-linear dose response, angular dependence, time between irradiation an d readout, beam quality characteristics of the irradiation and attenuation of the actual TLD holder. 8.1.2 Advantages of OSL Dosimeters The m aterial used in OSL dosimeters is carbon-doped aluminum oxide (Al2O3:C). Because of its high sensitivity, only a small amount of material needs to be used in making a single dosimeter for adequate dos e absorption, which has several ad vantages. First, the physical size of the dosimeter can be small, therefore in creasing spatial resolution and localization of the dose measurement.46 Second, depending on the method of readout, only a small amount of luminescence is necessary per readout, allowing a la rge amount of charge to be retained within the material which is available for subsequent readouts of the same exposure.50 While a disadvantage of a TLD is its sensitivity to light, this can be turned into an advantage with OSL dosimeters. By encasing Al2O3:C in a light-tight holder, the ex act absorbed dose is held by the material. After readout, exposure to white light effectively erases the remaining signal within the Al2O3:C, and the dosimeter is ready for reuse.

PAGE 84

84 8.2 Characterization of OSL Dosimetry System A comm ercial dosimetry system (microStar Landauer, Glenwood, IL) that employs OSL technology was acquired for its potential use in estim ating doses resulting from CT scans. The microStar system uses small-sized dosimeters to capture and reports the shallow dose equivalent, also known as Hp(0.07) and defined as the shallow dose equivalent at a depth of 0.07 mm in ICRU tissue.51 These nanoDot dosimeters (Landauer, Glenwood, IL), depicted in Figure 8-1, have specifications of 5 mm diameter approximately 0.2 mm thickness and 10x10x2 mm3 holder dimensions. The diameter stated correspo nds to the primary side of the dosimeter; the light-tight holder covers a very small amount of OSL material found on the back, or secondary side of the dosimeter. As the OSL process is stimulated by white light, care must be taken to shield the dot dosimeters from white light after use. Each dosimeter is uniquely identified by a barcode on the underside of the dot dosimeter. T hus, the dosimeter to be read can be simply scanner into the system. Each dosimeter is assigned its own se nsitivity as determined by the manufacturer. The nanoDot dosimeters used in this study had a manufacturer-specified sensitivity of 0.91 and 0.93, respectively. The se nsitivity is included in the readout value reported by the microStar system. 8.2.1 Other Radiation-Measuring Devices Relevant exposures were m ade on a 32-sli ce CT scanner (Aquilion 32, Toshiba America Medical Systems, Tustin, CA) and a 320-sli ce CT scanner (Aquilion ONE, Toshiba America Medical Systems, Tustin, CA). Air kerma measur ements were performed with a 6-cc ionization chamber and a 0.6-cc ionization chamber, both described in Section 6.1. The methods of individual tests are presente d in the following sections.

PAGE 85

85 8.2.2 Tracking of System Standards The m icroStar reader uses two LED beams of di fferent intensities to read the signal on the OSL material after e xposure to radiation.52 The weak LED beam is first used to test the dosimeter to obtain a test count va lue; based on this valu e, the reader automatic ally determines if the weak or strong LED beam is necessary for readout. The signal on the dosimeter is proportional to the dose, and the r eader estimates and reports the Hp(0.07) dose for each dot dosimeter read in the microStar reader. The accuracy of the dose therefore depends on the stability of the LED beams. To assess the functionality and reproducibility of the LED beams, a start-up protocol is used each time the microStar is powered on. Three measurements are made, with the first measurement done with the read dial on the DRK setting, which determines the dark current associated with equipment electronics and is obtained with the LED beam off and the shutter closed. The manufacturer-recommended limit fo r the DRK counts is less than thirty. The second measurement, made with the dial to the CAL setting, assesses the sensitivity and consistency of the photomultiplier tube (PMT). In it, the PMT counts are measured with the shutter open, and a reading from a small amount of 14C radioactive material labeled CAL is made. The manufacturer suggest s that this value be within % of the standard operating average. The third measurement checks the consistency of the intensity of the LED beams and is labeled LED on the reader. In this resear ch project, special at tention was given to reproducibility of the LED measurements because th e same beam intensities must be applied to the OSL material for each reading in order to have reproducible, reliable measurements; fluctuating LED beam intensitie s would lead to a incorrect m easurement of the dose. The manufacturer-recommended tolerance limit for the LED measurement is 10% of the standard operating average.

PAGE 86

86 To establish standard operating values for the DRK, CAL and LED counts, twenty consecutive measurements were made and record ed. The mean values of those measurements were calculated and adopted as ope rating standards for the system. To ensure that the system was operating properly, the DRK, CAL, and LED counts were monitored for two working weeks (ten days) before the system was put into full use. On each day, five measurements were made at each dial position; the average was calculated a nd plotted against the established standards. These measurements were repeated and r ecorded each time the system was used. 8.2.3 Dosimeter Response Two different types of nanoDot dosim eters were used in this research project: standard dosimeters with a manufacturer-specified sens itivity of 93% and screened dosimeters with 91% sensitivity. The manufacturer-specified accuracy of these dosimeters is % and %, respectively. This difference in accuracy defi nes the screened type of dosimeters, which are selected due to their tighter a ccuracy and sold separately. Do simeter precision was investigated to ensure that the dosimeters responded in the same way to an x-ray exposure. The precision was evaluated by the standard deviation of doses measured within each group, and results were analyzed within each group to ensure dosimeters of the same sensitivity had a uniform dose response. It was expected that the screened dots would have a smaller value of the standard deviation due to their higher manufacturer-spec ified accuracy. Since both dosimeter types are made of the same material, it was expected th at standard and screened dots would measure the same dose. To ensure this was the case, the mean dose of each group was compared. Ten dosimeters of each type were used to an alyze dose response. For each measurement, one nanoDot dosimeter was placed in the center of the x-ray field on top of 5.1 cm of rectangular acrylic slabs. A 0.6 cc volume ionization chamber was also placed in the beam as a reference. A total of ten separate exposures was made under the same reference conditions for each dosimeter

PAGE 87

87 type. The mean of each group was calculated, as well as the CV. The outcome of this test, and all others associated with the dosimetry system verification, are li sted in Section 10.2. 8.2.4 Room-Light Erasure During the length of this research project, each nanoDot dosimeter was used in multiple measurements. In most cases, a background readi ng of the dose was acquired first, followed by exposure to an x-ray beam, and then the actu al post-exposure reading; the total dose was calculated as the difference of the post-exposur e reading and the background reading. After usage, the OSL material was taken out of the lig ht-tight casing and exposed to white room light to release the residual trapped energy in th e dosimeter. The length of erasure time was determined to ensure acceptable background dose levels, as described below. Throughout this research project, a background limit of 1.0 mGy was established to decrease statistical uncertainty associated with measurements of the same order of magnitude as the background. Prior to exposure, ten nanoDots were read to determine the background reading. The nanoDots were then exposed to a radiographic xray beam. Dosimeters were exposed one at a time, placed in the center of the x-ray field at a distance of 70 cm from the x-ray source. A 6 cc ion chamber was also placed in the x-ray beam to measure the air kerma for each exposure simultaneously. A post-exposure reading for each nanoDot was obtained, and the difference between the initial and post-exposure readings wa s calculated for all dosim eters. For the purpose of determining the necessary erasur e time, this dose was considered to be the initial dose before white-light exposure, or t=0 reading. Following this procedure, the dosimeters were exposed to room light for thirty minutes. After thirty minutes, one dosimeter was read. Th e displayed dose was recorded and compared to the initial dose reading. If it was greater than 1 mGy, the dosimeter was then returned to white light exposure to continue the erasing procedure. A second dosimeter was read ten minutes later

PAGE 88

88 (total elapsed erasure time of forty minutes), and the process was repeated with all ten dosimeters until an acceptable background level of less than 1 mGy was reached. 8.2.5 Energy Response The m anufacturer provides a se t of dosimeters are that calibrated at 80 kVp to beam code RQR6, defined by an effective energy of 44 keV and HVL of 2.9 mm Al and on PMMA material.53,54 This calibration set is used to calibrate the microStar OSL reader. CT beam energies used clinically ar e higher than 80 kVp, usually 120 kV and occasionally 100 kV, depending on the patient size and the anatomy to be imaged. Similarly, the HVLs found in typical CT systems are approximately 1-5 mm higher than the one us ed for calibration. Therefore, an energy correction procedure was performed to obtain factors to be applied to the dosimetric reading, to account for differe nces in x-ray beam characteristics. To assess such energy response, dosimeters were placed in the center of an x-ray beam of a 320-slice CT scanner. The x-ray tube was held stationary during the exposure and positioned at the top of the gantry (12 ocloc k). Five dosimeters and a 0.6 cc ionization chamber were placed on 5.1 cm of acrylic, as depicted in Figure 8-2. Because the calibration dosimeter set was originally exposed on acrylic, these reference co nditions were matched as best as possible to assess energy response. The five dosimeters we re positioned around the chamber to ensure that both the ion chamber and dosimeters would measur e the same CT x-ray exposure. All other exposure parameters were kept constant, including filter and focal spot so that only the tube voltage was changed. The mean and CV were calculated for each tube voltage. 8.2.6 Scatter Response The nanoDot calibration dosim eters were expos ed on the surface of PMMA material, and the reader reports a shallow dose. While some organ doses measured in this research project were done with the dosimeters placed on the surface, some were also placed in the phantoms.

PAGE 89

89 Because the conditions for non-superficial organs dose measurements do no match those of the dosimeter calibration, scatte r response was evaluated. To do so, dosimeters were exposed under incr easing thicknesses of acrylic, as diagrammed in Figure 8-3. A general x-ray tube was used in order to produ ce a flat x-ray field and avoid a non-uniform field due to the heel effect, and because there is not enough room in the CT scanner gantry to produce such a flat beam. The x-ra y source was placed 100 cm from the surface of the acrylic. For the purpose of asse ssing the scatter respons e of the dosimeters, the three CT beams simulated using a general radiograph x-ray tu be were the 80, 100 and 120 kV beams and the Small-S filter. The HVL of each of these three beams was measur ed (Section 7.2.3) and 1100 aluminum alloy was added to the x-ray beam until the HVL of the beam matched that of the CT beam. To simulate different scatter conditions, five dosimeters were placed around a 0.6 cc ion chamber. The light field of the x-ray tube was used to position the dosimeters and ionization chamber at the center of the field. Surface dose m easurements were made with this setup. For the rest of the measurements, 2.1 cm of acrylic were placed to the right and left of the ion chamber and the dosimeters to support the thicker acryl ic slabs as well as to provide side scatter. More scattering material was placed superior a nd inferior to the chamber and dosimeters to simulate the situation of a dosimeter inside the phantom completely surrounded by scattering material. This setup is depicted in Figure 8-4. The same methodology was followed, and scatte r response was assessed in both the 320slice scanner and the 64-slice sca nner. However, instead of measurements at all beam energies, the tube voltage and bowtie filter combinations used within the cl inical protocols used for organ dose measurements were used for evaluation of scatter response to the CT x-ray beams.

PAGE 90

90 8.2.7 Dosimeter Calibration The nanoDot dosimeters were calibrated for use in this research project. Due to the small size of the 0.6 cc ionization chamber, and the fact that its calibration is traceable to a national standards laboratory, it was the st andard against which the nanoDot dosimeters were compared. First, the dosimeters were calibrated to each en ergy and filter combination by taking the ratio of the dosimeter dose to that of the ion ch amber, as expressed in Equation 8-1, (8-1) where CE,S is the energy and scatter correction factor. For organ doses that were measured on the surface of the phantom, specifically the pediatric doses (excluding th e thyroid), lens of eye doses and skin doses, the surface CE,S correction factor was applied. For all other internal organs, the mean of the energy and scatter CE,S was applied to the raw organ dose measurements. The mean was chosen instead of separate depth-dependent correction factors becau se the uncertainty in which scatter correction f actor to apply to each organ was la rge compared with the uncertainty that resulted in applying the mean. For all ion chamber measurements involved in these calibration factors, the chamber measured dose to air. This value was convert ed to dose to tissue using the ratio of mass attenuation coefficients. These values are a function of effectiv e energy, which is the energy of a polyenergetic x-ray beam that has the same attenuation characteristics of a polyenergetic beam. It was not possible to measure effective energy directly and therefore HVLs were measured on both the 320-slice and 64-s lice scanner and converted to effec tive energies using a table provided by Bushberg6 and reproduced in Table 8-1. Mass attenuation coefficients (en/ ) for soft tissue (ICRU-44) and air were found at each effective en ergy calculated and the dose in air to tissue

PAGE 91

91 was calculated for each of these effective energies as ratio of the two as described by Equation 82: (8-2) The organ doses reported, D, was finally obta ined by dividing the uncorrected dosimeter dose reported by the reader, Draw, by the CE,S correction factor, and then multiplying the result by the f-factor to calculate dose to tissu e, as expressed in Equation 8-3. f C D DSE raw, (8-3) 8.2.8 Linearity Response As with m ost dosimetry systems, it is important to ensure the dosimeters respond uniformly to different amounts of radiation. To assess linearity response, five dosimeters were positioned around a 0.6 cc ionization chamber and placed on 5.1 cm of acrylic, similar to the setup in 8.2.4 and 8.2.5. A tube voltage of 120 kV was used in the 320-slice CT scanner. The xray tube was positioned above the dosimeter and ch amber setup at the top of the scanner gantry (12 oclock position). Each exposure was one second in duration, and the tube current was changed from the minimum to maximum mA as allowed by the scanner operating in service mode (10 to 580 mA). Intermediate tube curren t values were also used to produce a linearity response curve for the OSL dosimeters. These measurements were repeated to assess linearity over the range of doses measured throughout this research project. The same set up was used as previously described, however the second set of measurements were made with a fi xed tube current of 500 mA. Multiple exposures were made to expose the dosimeters and ion chamber to a range of mA from 500 to 8000. air en tissue enf )/( )/(

PAGE 92

92 8.2.9 Angular Response Because the geom etry of a CT scan involves x-ray beam rotation, the angular response of the dot dosimeters was investigated, as the dosimeters receive expos ure at all angles during a CT acquisition, with a variety of experimental setups. To investig ate this angular response in a systematic manner, the CT scanner was operated in service mode, which allo ws the x-ray tube to be parked in a stationa ry position and at any angle. The ze ro-degree position was defined at the top, or 12-oclock position, of the scanner gantry with the x-ray tube rotation in a clockwise direction through 360. Angular response measur ements were made both in-air and using a CTDI phantom, as described below. 8.2.9.1 In-air response For each in -air measurement, a nanoDot dosimeter was suspended within the scanned region of the 320-slice CT scanner within the gantry at isocenter. Exposures were made from 0 to180 at 15 intervals. Five measurements we re made at each x-ray tube position. The mean and CV were calculated for each x-ray tube position. 8.2.9.2 In-phantom response A standard CTDI phantom was used to assess the angular response of the dosimeters under scatter conditions. For each in-air measurement, one dosimeter was placed in the center hole at the Z-axis, midpoint in the phantom. PMMA rods were inserted into th e center hole on each side of the dosimeter to fill voids and provide scatter. All other holes were also filled with PMMA rods. Figure 8-5 depicts the dos imeter setup in the phantom. Exposures were made in the 320-slice scanner. The x-ray tube st arting angular position was 270 (the 9 oclock position within the scan ner gantry) and repositioned clockwise to 90 with the dosimeters face-up in the phantom. The x-ray tube was moved at increments of 10 in order to include the 270 and 90 positions, i.e., wh en the primary x-ray beam is perpendicular to

PAGE 93

93 the active area of the dosimeter. A second set of exposures was made with the dosimeters facedown, using the same tube positions and angular increments. The number of replicate measurements at each x-ray tube position was five. 8.2.10 Comparison of OSL Dosimeter in and out of its Light-Tight Case In order to fit the nanoD ot OSL dosimeters into the tomographic physical phantoms for certain organ dose measurements, the dosimeter materi al needed to be taken out of its light-tight case. The actual size of the ac tive dosimeter is smaller than that of the case, making insertion into the phantom more feasible. To investigate this process and any effect it might have on accuracy and reproducibility of the dosimeters, iden tical CT exposures were made with two sets of dosimeters: one set in the case and one set ou t of the case. Comparison between the two was made to ensure there was no significant deple tion of signal (dose) when the dosimeters were exposed outside of the light-tight plastic case. The nanoDot dosimeters were exposed using a preset clinical protocol (Ba by Chest Volume) using the 320-slice volumetric scanner. The CTDI head phantom was used for these exposures to simulate attenuation and scatter conditions. For each measurement, one dosimeter was placed at the longitudinal midpoint of the phantom. The protocol consists of two localizing tomographs in the AP and lateral orientations directly followed by a 0.4-second volumetric acquisition. This measurement process was repeated ten times with ten different dosimeters. The dosimeters were placed at the 12 ocloc k and center holes of the CTDI phantom. The 12 oclock position was used to simulate organs at a shallow depth, such as the thyroid, while the center position was chosen as representative of deeper organs for example, the heart. Ten separate measurements were made with dosimeters exposed in the plastic case at each position; ten were taken out of the case and exposed i ndividually at the same center and 12 oclock phantom positions. The dimmest room lighting possible was used to minimize the dosimeter

PAGE 94

94 background signal depletion while the active ma terial was out of the case. After each measurement, the dosimeter was placed b ack into the light-tight case and read.

PAGE 95

95 Figure 8-1. From top to bottom, nanoDot dosimet ers with OSL material exposed, secondary side up and primary side up. Figure 8-2. Energy response, dose rate response a nd scatter response surface setup; five nanoDot dosimeters placed around a 0.6 cc ionization chamber and on top of 5.1 cm of acrylic.

PAGE 96

96 Table 8-1. HVL as a func tion of effective energy.6 HVL (mm Al) Effective energy (keV) 1.25 24 1.54 26 1.9 28 2.27 30 3.34 35 4.52 40 5.76 45 6.97 50 9.24 60 11.15 70 Figure 8-3. Diagram of cross-sect ional view of ion chamber and dos imeters, placed on 5 cm of acrylic slabs. Acrylic slabs were added in 2.5 cm intervals on t op of the ion chamber and dosimeters to evaluated the scatte r response of the dosimeters.

PAGE 97

97 Figure 8-4. Setup of scatter response for measur ements made with increasing thicknesses of acrylic, depicting the ionization chambe r surrounded by nanoDot dosimeters. Scattering material was placed around both. Figure 8-5. Setup of in-phantom angular response; nanoDot dosimet er placed at the mid-point of the CTDI head phantom.

PAGE 98

98 CHAPTER 9 ORGAN DOSE MEASUREMENTS This chapter details th e procedure followed in performing organ dose measurements made using the tomographic physical phantoms described in Chapter 6. Three clinical protocols, developed and established by the RPC following the process detailed in Chap ter 2, were used in conjunction with the nanoDot dosimeters to meas ure organ doses. The pediatric phantom was used for with the pediatric craniosynostosis prot ocol, and the adult phantom was used with brain perfusion and cardiac protocols. Localizing tomographs were used to set the boundaries of the scan, and the dosimeters were placed at predet ermined locations on and in the phantoms after these tomographs were acquired. The scanner alignment lasers were used to ensure the positioning of the phantom was as cl ose to that of the clinical st udy case as possible and also to ensure the phantom position on the scanne r table was the same for each scan. The nanoDot OSL dosimeters were used to m easure mean organ doses. The small size of the nanoDots allows a single dosimeter to approxim ate a point organ dose. As much as possible, dosimeters were placed in numerous positions th roughout an organ to cover as much of its volume as possible, and the point doses were averaged to obtain an average organ dose. As a more conservative figure, and to account for the fact that in CT, some organs are only partially exposed, the highest point dose reading was also r ecorded. Dosimeters were placed in organs of interest, as determined by their proximity to the primary CT x-ray beam as well as by their radiosensitivity (as inferred from their corresponding tissue weighti ng factors). Dosimeters in the phantom were put into place by cutting a small sl it into the tissue-equivalent material of the desired slice. In the case of the adult phantom, a vacuum bag wa s used to stabilize the phantom and keep the different sections as compressed toge ther as possible. A more complete description

PAGE 99

99 of dosimeter locations and organ dose measurements is provided in the following sections for each protocol investigated. 9.1 Pediatric Head Study A comm on CT study for pediatric patients is a non-contrast h ead scan to evaluate for craniosynostosis, a deformity of the skull caused by irregular fusi ng of cranial sutures.55 In many cases of craniosynostosis, CT is used to ev aluate the lengths of the patients sutures. While this study has traditionally been performed on a 64-slice scanner, th e ability to scan the entire head in a single rotation with the 320-sl ice scanner makes this study a good candidate for organ dose evaluation and comparison. 9.1.1 Volumetric Protocol The increased coverage com bined with th e 0.35 ms rotation time make the 320-slice scanner conducive to pediatric studies, where pa tient motion during the ac quisition is frequently a problem. The default parameters for the pediat ric head CT study are list ed in Table 9-1. The scan range was chosen, based on the localizing tomographs; a range of 120 mm ensured coverage of the entire brain for the phantom use d. It is important to note that the range on the 320-slice scanner is adjustable in only 20 mm increments. Despite this limitation, the volumetric coverage selection depends on the size of the patients skull and thus affects organ doses. The 9-month-old phantom described in Secti on 6.4.2 was placed on the scanner table and positioned using the alignment lasers. The phant om position was also marked on the table to help with repositioning after dosimeter placement. Two perpendicular localizing tomographs were acquired, and after the scan boundaries were set, dosimeters were placed on and in the phantom. Organ doses resulting from the loca lizing tomographs were not measured. The selection of organs of interest was based on radiosensitive organs that are exposed directly to the

PAGE 100

100 CT x-ray beam or in close proximity to be expos ed to scattered radiation. The following organs were selected: a. Skin. Eight dosimeters were placed equidistant across the forehead to measure a representative average skin dose. b. Breast. Measurements were made on the surface of the pediatric phantom. c. Thyroid. A small hole was drilled into the pha ntom at the location of the thyroid and a dosimeter was taken out of its case to it fit inside the orifice in the phantom. d. Lens of the eye. Two dosimeters were pl aced on the surface of each eye to measure dose to the lenses of the eyes. Table 9-2 summarizes the number of dosimeters us ed for each of these organs of interest. Because the thyroid measurement included only one dosimeter, the measurements with this protocol were repeated five times with five different sets of dosim eters to increase the statistical reliability of the measured thyroid dose. Figur e 9-1 depicts the phantom setup on the CT scanner table. To evaluate the potential for dose reduction, a second series of measurements was performed using the same phantom and protocol but decreasing the tube voltage from 120 kVp to 100 kVp. The same methodology for dosimeter placement and scanning sequence was followed. The results of this methodology, and all ot hers described in this chapter, are presented in Chapter 10. 9.1.2 Helical Protocol Dose m easurements were performed on 64-slice CT scanner using a corresponding pediatric protocol, using a sim ilar non-contrast head study speci fically for the evaluation of craniosynostosis. Dosimeter placements and numbers were the same as those listed in Table 9-2 for the volumetric protocol. Similar scan parameters were chosen, in cluding a range of 120 mm

PAGE 101

101 to cover the entire head; these parameters are listed in Table 93. Positioning of the phantom within the scanner was done us ing the alignment lasers, and extensive efforts were made to position the phantom in the same way as on th e volumetric scanner for best comparison. 9.1.3 Image Quality Analysis In an effort to reduce organ doses resulting fr om the volumetric craniosynostosis protocol, an image quality evaluation was necessary to en sure that any reduction in tube voltage would provide images of diagnostic quality. To do s o, an image quality phantom described in Section 6.5 was scanned on the 320-slice CT scanner used for organ dose measurements. The phantom was scanned using the same acquisition and r econstruction parameters as the volumetric craniosynostosis protocol. Images of the lowcontrast module, acquired at 100 and 120 kV, were compared and scored by three experienced medi cal physicists. This scoring was performed by displaying the image in the center of module with the window and level settings optimized to display the number of objects seen in the image in the best way possible. Dim room lighting was used, and each viewer was seated 2-3 feet from the displayed image. The number of rods fully visible to each viewer was counted, and the total number of objects was reported for both images. 9.2 Adult Brain Perfusion One of the most powerful tools of a 320-slice vo lum etric CT scanner is its ability to image an entire organ in a single rotation of the x-ray tube. As previously described in Section 3.2.3, one clinical application for this type of acquisition is the evalua tion of suspected stroke patients using brain perfusion data. The RPC took on the ta sk of designing this protocol in the clinical setting to maximize image quality and reduce do se, as described in Chapter 2. Organ dose measurements were made for each of the required iterations and then compared to the previous standard of perfusion imaging as preformed on a 64-slice scanner.

PAGE 102

102 9.2.1 Volumetric Protocol The volum etric adult brain perfusion protocol us ed for the evaluation of stroke consists of four general acquisitions. First among these is a non-contrast scan of the head to evaluate for potential bleeding. At Shands at UF, this is performed as a helical acquisition. This non-contrast scan is immediately followed by several dynamic volume scans to acquire perfusion data. The third and fourth acquisitions are a helical head scan with contrast, followed by a helical CT angiogram (CTA) of the head. These acquisi tions are further detailed in Table 9-4. Organ doses were measured for each part of th is protocol. While the scanning parameters used in the non-contrast head, h ead with contrast and CTA of the head remained the same for all iterations, parameters for the dynamic volumes were varied to reflect th e clinical changes the protocol underwent during development. The first volumetric acquisitions assessed were those suggested by the manufacturer a nd are labeled Manufacturer at 80 kV throughout this research project. During the optimization process, volumetric acquisitions were done with the tube voltage increased from 80 kV to 100 and 120 kV labeled Manufacturer at 100 kV and Manufacturer at 120 kV, respectively, and th e organ doses were measured at identical locations. The fourth iteration of the dynami c volumes involved changing the timing of the arterial-phase acquisitions from intermittent to continuous, and is labeled Continuous. Finally, an altogether different acquisiti on protocol was evaluated. In th is protocol, the tube current increased during a certain portion of the arterial phase. This prot ocol is labeled mA Boost in this research project. A total of 27 nanoDot dosimeters was used to measure point organ doses for each of the resulting brain perfusion acquisitions. Similar to those of the pediatric study, organs were chosen based on radiosensitivity and proximity to the CT volume (Section 9.1.1). The following organs were chosen:

PAGE 103

103 a. Skin. 8 dosimeters were placed on the surface of the forehead to measure skin dose, as depicted in Figure 9-2. b. Lens of the eye. Two dosimeters were placed on the surface of each eye for the lens dose, also shown in Figure 9-2. c. Esophagus. There were three possible loca tions for dosimeter placed within the phantom for the esophagus. As was the cas e with all internal organs, the esophagus was outlined in the available phantom slices and was just hollow enough for the dosimeters to be placed in the hollow space. There were three slices with the denotation of the esophagus, and only one dosimet er would fit in each slice due to the small size of the esophagus. Thus, a tota l of three dosimeters was used, and their specific locations are pictured in Figure 9-3. d. Thyroid. The thyroid was contained in onl y one slice of the phantom, and three dosimeters were placed within the thyroid in left, center and right positions as shown in Figure 9-4. e. Brain. While the brain is fairly radioresis tant, the average organ dose was measured, as it is within the primary beam during this protocol. Figure 9-5 shows five dosimeters placed in a slice mid-way through the brain. It was planned to have two additional dosimeters in the frontal portion of the brai n; however, the bone-equivalent material directly below the slice pr evented nanoDot dosimeters fr om being placed at this location. f. Breast. Finally, 2 dosimeters were placed between two slices of each breast, as illustrated in Figure 9-6. Breast tissue rece ives only scatter radiation with these acquisitions and therefore the nanoDots were placed with their primary side

PAGE 104

104 perpendicular to the x-ray beam and flush against the two slices so that the active material of the dosimeters was in the sa me direction as the scattered radiation. A summary of number a nd location of dosimeters is given it Table 9-5. 9.2.2 Helical Protocol For com parison, the protocol used for evaluation of stroke utilized in the 64-slice scanner was also evaluated for organ dose measurements in the same locations and with the same adult phantom. Because of the smaller beam width (3 2 mm) compared to the volumetric scanner (160 mm), the entire brain is not covered during a singl e acquisition of the perfusion data. The full width of the beam is used, and images are acquired consecutively for 1 minute. As was the case with the volumetric scanner, a he lical scan of the head without contrast was acquired before the perfusion acquisition. Helical CTA of the head and a second helical of the head with contrast are also acquired following the perfusion scan. The parameters for each of these acquisitions are detailed in Table 9-6. 9.2.3 Image Quality Evaluation The advent of the 320-slice sca nner and its vo lumetric acquisiti on capabilities led to a need to develop new imaging protocols to utilize the scanner in the mo st effective way possible. As previously explained, the sca nner was equipped with standard protocols developed by the manufacturer, which were used as the starti ng point for the development process. An experienced neuroradiologist was in charge of assessing image quality by evaluating it throughout the development process of the brain pe rfusion protocol, as de scribed in Chapter 2. Before additional changes were made to the prot ocol, the medical physicists in the RPC verified that any proposed change would not result in a significant dose increa se. The CT technologists were given specific instructions from the neur oradiologist detailing which aspect(s) of the protocol to change prior to the study.

PAGE 105

105 9.3 Adult Cardiac CT Angiography A second powerful application of 160 mm volumetric CT scan coverage is cardiac imaging. Like the brain, the entire heart can be imaged with a si ngle gantry rotation; in addition, cardiac CT angiography (CTA) is po ssible. Cardiac CTA is used for the detection of coronary artery disease (CAD).56 While the American College of Cardiology (ACC) considers the 64-slice system as the current standard for cardiac CT st udies, high doses have been seen as a limitation. However, cardiac CTA offers an alternative to the more invasive coronary angiography procedure for detection of CAD. There are several advantages to the full cardiac coverage abil ity of the 320-slice system. Because images are acquired axially, the stair-s tep artifact commonly seen in helicallyacquired images is eliminated.57 As is the case in brain perfusi on studies, the short rotation time allows the contrast bolus to be imaged at a single point in time produc ing temporally-uniform images,57 actually placing the temporal reso lution at one-half the rotation time. There are three general acqui sition protocols for cardiac CTA available on the 320-slice scanner: a. Cardiac functional analysis (CFA). If car diac function needs to be assessed, the CT exposure begins with a half-rotation acquisiti on just before the R-R interval, continuing through the full R-R interval and ending with another half-rotation acquisition just after the R-R interval, as illustrated in Figure 9-7A. Image acquisition during the entire heartbeat allows reconstruction images an any point during the cardiac cycle. This protocol is used in the evaluation of the ejection fraction, st roke volume, cardiac output, end-systolic and end-diastolic volumes and segmental wall motion.58 b. CFA with dose modulation. In an effort to reduce dose to the patient, the CFA protocol may be performed with a dose modulation. While the image acquisition

PAGE 106

106 process is the same, the mA is decreased dur ing the diastole porti on of the heart beat cycle, as depicted in Figure 9-7B. c. Prospective electrocardiogram (ECG) gati ng. When functional information is not necessary, prospective ECG gating is the prefer red protocol. If the patients heart rate is below 65 beats per minute (bpm), images are acquired using a single exposure with one-half rotation of the x-ray tube (0.35 s) at diastole, illustrate d in Figure 9-7C. Because of the shorter exposure time, the dose is generally lower than in the previous protocol. If the patients heart rate is above 65 bpm, several acquisitions occur during different heart beats. In the case of prospective ECG gating, less contrast is administered to the patient because of the decreased time needed for the entire study.58 Details of the cardiac CTA protocols used to measure organ doses are listed in Table 9-7. While there have been several studies comp aring these various imaging protocols for cardiac CTA,58,57,59 all of them report effective doses m easured from pencil chambers and CTDI phantoms. Similar to the methodology and selected criteria described in previous sections, nanoDot dosimeters were placed in or on radiosen sitive organs within the primary x-ray beam and those in close proximity to the primary beam to receive a significant am ount of scatter. As detailed in Table 9-8, five organs were selected to measure average organ doses: a. Thyroid. Three dosimeters were placed in the thyroid at the le ft, center and right positions of the organ (see Fig. 9-4). b. Lungs. A total of 20 dosimeters was used to measure an average dose to the lungs. Because of their large size, a nd the fact that only a portion of the lungs is exposed in the primary beam, doses were measured at five different sections of the phantom. For each slice, two dosimeters were placed in each lung. Two different orientations were

PAGE 107

107 used: dosimeters were aligned in the corona l and sagittal directions. The dosimeter orientation was alternated with each phant om slice, as depicted in Figure 9-7. c. Stomach. The stomach has a relatively high tissue weighting factor, and, while it may not be in the primary beam, the stomach coul d receive a significant am ount of scatter. A total of 8 dosimeters was used to measure the stomach dose as shown in Figure 9-8. Four dosimeters were positioned around the stomach in the two slices that contained the organ. d. Skin. Similar to the brain perfusion meas urements, 8 dosimeters were placed on the surface of the phantom in two rows across the chest to measure skin dose and are pictured in Figure 9-9. e. Breast. Dose to the breast was of most conc ern for this protocol, as breast tissue was in the primary beam and is also highly radiosensitive. The male phantom was used and therefore lacked the anatomy of the female breast. However, dosimeters were placed to estimate female breast dose. Because a 1 cm layer of fatty tissue was assumed to line the breast; dosimeters were placed 1 cm beneath the surface60 and also approximately 1 cm from the chest wall. Two dosimeters were placed at the right and left locations underneath the ar mpit to measure dose to the axillary tail of the breast. This area is where the most radiat ion-induced breast cancers are located.61 Breast tissue was located in two slices of the phantom, shown in Fi gure 9-10, and a total of 24 dosimeters were used to measure breast dose. 9.3.1 Volumetric Protocol Three cardiac CTA protocols were used to m easure organ doses on a 320-slice volumetric scanner: prospectively-gated, functional analysis and functional analysis with dose modulation. The details of these three protoc ols are listed in Table 9-10. Fo r all protocols, the volumetric

PAGE 108

108 scan is a single axial half-rotation. The scan be gins based on the patients heart rate, which was simulated for phantom measurements using an ECG simulator (Model EHS10, Dale Technology, Everett, WA). 9.3.2 Helical Protocol For com parison, a cardiac CTA protocol was utilized on the 64-slice scanner. This protocol was chosen as the best comparison to the three volumetric protocols and is the most commonly-performed cardiac CTA protocol on this scanner. The parameters are detailed in Table 9-7. The scan consists of a single helical scan that covers the entire heart. As with the volumetric protocol, the scan begi ns at the appropriate time in the patients heart cycle.

PAGE 109

109 Table 9-1. Scan parameters for volumetric pedi atric head protocol in Aquilion One 320-slice CT scanner. Scan Parameter Nominal Tube voltage 120 kV Tube Current 200 mA Tube Rotation Time 0.6 sec Effective mAs 121 Scan Range 120 mm Acquisition Thickness 0.5 mm x 320 Focal Spot Small Filter Small S Table 9-2. Locations and number of dosim eters used in and on pediatric phantom. Organ Number of dosimeters Lens of Eye 4 (two on each eye) Skin 8 Thyroid 1 Breast 6 (three on each breast) Figure 9-1. Pediatric phantom with surface dosim eters to measure skin, eye and breast doses; A) anterior view, and B) lateral view.

PAGE 110

110 Table 9-3. Scan parameters for helical pediat ric head protocol in Aquilion 64 64-slice CT scanner. Scan Parameter Tube Voltage 120 kV Tube Current 200 mA Tube Rotation Time 0.5 sec Effective mAs 157 Helical Pitch 0.641 Scan Range 120 mm Acquisition Thickness 0.5 mm x 64 Focal Spot Small Filter Small S Table 9-4. Details of volumetri c brain perfusion protocols. Scan name Scan type Tube voltage (kV) Tube current (mA) Rotation Time (s) No. of volumes Head without Helical 120 300 0.5 CTA head Helical 120 400 0.5 Head with Helical 120 300 0.5 Original at 80 kV Dynamic volume 1 80 310 0.75 5 Dynamic volume 2 80 150 0.75 1 Dynamic volume 3 80 150 0.75 13 Original at 100 kV Dynamic volume 1 100 310 0.75 5 Dynamic volume 2 100 150 0.75 1 Dynamic volume 3 100 150 0.75 13 Original at 120 kV Dynamic volume 1 120 310 0.75 1 Dynamic volume 2 120 150 0.75 13 Dynamic volume 3 120 150 0.75 5 Continuous Dynamic volume 1 80 300 1.0 1 Dynamic volume 2 80 120 1.0 17 Dynamic volume 3 80 120 1.0 7 mA boost Dynamic volume 1 80 310 0.75 1 Dynamic volume 2 80 150 0.75 2 Dynamic volume 3 80 300 0.75 7 Dynamic volume 4 80 150 0.75 4 Dynamic volume 5 80 150 0.75 6

PAGE 111

111 Figure 9-2. Placement of skin and lens dosimeter s on the forehead and lens of the eyes on the adult phantom for the brain perfusion protocol.

PAGE 112

112 Figure 9-3. Dosimeter placement in the esophagus, measured in three slices (A-C) of the adult phantom from the brain perfusion protocol. Figure 9-4. Locations of dos imeters within the thyroid.

PAGE 113

113 Figure 9-5. Locations of the dosimeters within the brain. Figure 9-6. Locations of dosimeters within the breast to capture scattered radiation.

PAGE 114

114 Table 9-5. Locations and number of dosimet ers used in and on adult phantom for brain perfusion protocol. Organ Number of dosimeters Esophagus 3 Thyroid 3 Brain 5 Skin 8 Breast 4 (two on each side) Lens of Eye 4 (two on each eye) Table 9-6. Scan parameters for he lical adult brain perfusion protocol. Scan name Scan type Tube voltage (kV) Tube current (mA) Rotation Time (s) Helical Pitch Detector Array Head without Helical 120 300 0.75 0.641 0.5 x 64 Perfusion Dynamic vol. 120 150 1.0 -8.0 x 4 CTA Head Helical 120 400 0.5 0.641 0.5 x 64 Head w/ delay Helical 120 300 0.75 0.641 0.5 x 64 Figure 9-7. Cardiac R-R cycles are shown with exposure conditions of, A) functional analysis, B) functional analysis with dose modula tion, and C) prospective ECG gating.

PAGE 115

115 Table 9-7. Scan parameters fo r adult cardiac CTA protocol. Protocol Tube voltage (kV) Tube current (mA) Rotation time (s) Slice thickness (mm) Helical pitch Prospectively-gated CTA 120 P400 0.35 0.5 -Functional analysis with dose mod. 120 M500 0.35 0.5 -Functional analysis 120 M500 0.35 0.5 -64-slice 120 490 0.4 0.5 x 64 0.21 Table 9-8. Location and number of dosimeters used in and on adult phantom for cardiac CTA protocol. Organ Number of dosimeters Thyroid 3 Lung 20 Stomach 8 Breast 24 Skin 8

PAGE 116

116 Figure 9-8. Placement of dosimeters in the lungs, A, C,E) with dosimeters aligned right to left in the lung, and B,D) with dosimeters aligned in the anterior-posterior direction, and F) as magnified view.

PAGE 117

117 Figure 9-9. Placement of dosimeters within the stomach as the A) inferior slice, B) magnified view of the superior slice, C) inferior stomach slice and D) magnified view of the inferior slice. Figure 9-10. Placement of the dosimeters to measure skin dose across the breasts.

PAGE 118

118 Figure 9-11. Locations of dosimeters for br east dose measurements using the cardiac CTA protocols with A) specific location of the two dosimeters measuring dose to the axillary tail of the breast and B) the inferior slice of the phantom.

PAGE 119

119 CHAPTER 10 RESULTS The m easurements and results of this research project are presented in this chapter. The characterization of the wide-beam CT scanner was both interesting and informative. A novel method was developed to describe the bowtie filters used by the 320-slice CT system, information which is usually proprietary. A commercially-available dosimetry system was verified for use in diagnostic radiology, a nd a methodology is presented to allow the OSL dosimeters and reader to be used for organ dose measurements. Finally, using this OSL system and tomographic physical phantoms, organ doses were measured fo r three clinical protocols, redesigned by the RPC, using two CT systems: a 320-slice CT scanner and a 64-slice CT scanner for comparison. An image quality analysis was performed for each protocol thereby completing the aims of this research project. 10.1 Characterization of X-Ray Be ams of Volumetric CT Scanner 10.1.1 Exposure Reproducibility 10.1.1.1 Service mode Exposures perform ed in service mode, as de tailed in 7.1.1, were reproducible to 0.03%. These measurements are provided in Table 10-1. Th e importance of this result is the reliability of the x-ray tube to reproduce its output. Wh en multiple exposures were made to analyze the OSL dosimeters (section 8.2), it was important to minimize sources of error, including variability in the x-ray tube exposure. More specifically, th is high degree of reproducibility guarantees very little source of error from the s canner in dosimeter measurements. 10.1.1.2 Clinical mode Reproducibility was assessed for a vo lumetric and a helical protocol (7.1.2). For the volumetric protocol, the smallest CV was 1.2% at the center position, a nd the largest CV of

PAGE 120

120 6.2%, was found at the 9 oclock position. Fu ll details are provided in Table 10-2. The difference among these sets of measurements are mo st likely due to the st arting position of the xray tube for each volumetric acquisition, which is not controllable. When the speed of the rotation of the tube is reached, the acquisition then begins when the technologist manually starts the scan. The measured doses vary by position with in the phantom because of this variability in starting tube position. It follo ws that the center position e xperiences the least amount of fluctuation in the dose measurements, as this position is independent of x-ray tube position; the measurement is made at isocenter, and the mate rial surrounding the center position is uniform in the radial direction. Although mi nimal, this inherent fluctuat ion of dose due to the starting position of the x-ray tube is less than 6.2% and contributed to the uncertainty in organ dose measurements. For the helical protocol, the measured doses were reproducible at al l phantom positions, as shown in Table 10-3. The coefficients of vari ation were 0.1% or less. This high degree of reproducibility is most likely due to the helical rotation of the x-ray tu be around the phantom and pencil chamber. While the starting position of the x-ray may be just as variable as seen in the volumetric protocol, the x-ray tube rotates ar ound the phantom so many times during the full 160 mm scan that any variability in tube starting position is negated. 10.1.2 Beam Quality 10.1.2.1 Tube voltage The differences in tube voltage m easured in the 320-slice scanner (7.2.1) were less than 2.1% of the nominal value. Energies were also measured on the 64-slice scanner and were less than 2.4% of the nominal energy. In both cases the largest inaccuracy of 2.4% occurred at 135 kV. For both scanners, the other three energi es were accurate to less than 1.4%. These measurements are detailed in Table 10-4.

PAGE 121

121 10.1.2.2 Total filtration Measurem ent of the total filtration of the x-ra y beam is described in 7.2.2. The effects of the Large-L filter at three beam energies are sh own in Figure 10-1. The measured total filtration is graphed as a function of measurement position acr oss the CT gantry, with isocenter defined as 0 mm. Despite this instrument having an energy range from 60 to 120 kV, as described in Section 6.2, this graph suggests that the instrument is not accurat e at 80 kV. The shape of the curve at 80 kV should behave in a similar wa y as the 100 and 120 kV curves because these bowtie filters are not dependent on energy, only FOV. Figure 10-2 displays the filtration differences between the Small-S and Large-L filters. In both cases, the increase in filtration measured from 0 cm to 15 cm approximates the shape of the bowtie filter. The decrease in total filtration pa st 15 cm may reflect the edge of the filter, as illustrated in Figure 10-3. The peak at 15 cm corresponds to what may be assumed to be the thickest part of the filter. It is interesting that the Small-S filter has le ss total filtration at the center (X=0 cm) than the Large-L filter. One might assume a Small-S filter would have the highest total filtration to reduce radiation dose to the skin in pediatric st udies, which would use the Small-S filter because of small FOVs. Similarly, one of the most important clinical applications of this scanner is brain perfusion studies. Brain images are acquired wi th a small field-of-view and use this Small-S filter. A thinner filter would allow for a reduc tion in tube current which would lower the overall dose to the patient, as compared to a thicker filt er that would require a higher tube current to compensate for the increase in attenuation of a thicker filter, assuming all other parameters remain the same.

PAGE 122

122 10.1.2.3 Half-value layer Tables 10-5 and 10-6 display the HVLs m easured at the nominal beam energies for the 320-slice and 64-scanners, respectively, in the cent er of the beam, as described in 7.2.3. As expected, for the same filter, the HVL increase s with increasing energy as lower-energy photons are preferentially absorbed, produc ing a harder x-ray beam. For the 320-slice scanner, separate measurements were made using the Medium-M and Large-L filters; the calculated HVLs were the same for both filters; they are listed together in Table 10-5. Similarly, separate HVL measurements were made on the 64-slice scanner using the Small-S filter and Small-M, or medium filter, and these filters were found to be the same. There is a 9% to 17% difference between the HVLs measured on the 320-slice scanner with the Small-S filter in place a nd either the Medium-M or Large-L. The x-ray beam is softer when using the Small-S filter, meaning the eff ective energy of the beam was shifted to lower energy in the spectrum and c ould affect organ doses. There is also a significant difference betw een the HVL measurements between the 320slice and 64-slice scanners. Fo r all measurements, the HVLs of the 64-slice scanner are larger than those of the 320-slice scanner. These differe nces affect both skin doses and image quality. When compared, a lower HVL (for the same CT system) results in higher skin doses than a higher HVL because, as discussed in Section 7.2.1, attenuation coefficients are energy-dependent and more low-energy photons in the polyenergetic beam will be absorbed by the skin in a CT beam with a lower HVL. On the hand, a lower HVL allows for a reduction in tube current because of the increased transmission of photons. 10.1.3 Beam Width The CR i mage of the beam width, acquired as described in 7.3 and displayed in Figure 104, was analyzed by Image J software. Calculatin g the FWHM of the beam profile in Figure 10-5,

PAGE 123

123 the beam width was found to be 17.6 cm across the center of the radiographic image of the nominal 16 cm CT beam. This re sult is in agreement a with b eam width of 18 cm measured by Geleijns et al .62 for the same type of scanner. One possible contributor to this discrepancy between the nominal and measured beam width is th e need for all detector elements to be within the primary x-ray beam to ensure that the photon intensity is same across all detectors thereby avoiding ring artifacts9 caused by inadequate sign al at a single detector. In this case, the actual beam width is wider than the detector elements to avoid the penumbra of the x-ray beam falling on the detectors and known as overbeaming.63 By increasing the width of the beam just slightly beyond the edge of the detector array, the penumbra falls outside the ar ray and ring artifact is avoided. 10.1.4 Dose Profile 10.1.4.1 X-axis As described in 7.4.1, a dose profile was m easured in air in the direction perpendicular to the anode-cathode axis, which is the same direct ion as the fan beam. The normalized dose is displayed as a function of distance from isocenter (X=0 mm) in Figure 10-6. As expected, there is a decrease in dose as the distance from isocenter increases. 10.1.4.2 Z-axis The beam profile in the direction of the Z-ax is was also measured and detailed in 7.4.2. The dose measured by the dosimeter was normalized to the measurement at the center of the beam (X=0 mm) and plotted as a function of distan ce, as displayed in Figure 10-7. As expected, there is a small decrease in dose in the direction parallel to the anode-cathode axis due to the heel effect. This result could be due to the radiati on field being so large in order to accommodate the 160 mm beam width that the heel effect is exaggerated on this 320-slice scanner. However, this profile is inconsistent with the beam profile produced using data acquired with the CR imaging

PAGE 124

124 plate, as described in Section 10.1.3. The beam profile shown in Figure 10-5 does not exhibit the heel effect to the same degree as Figure 10-7. The profile produced using the dosimeters used raw dosimeter dose in air measurements. Because of the shape of the bowtie filters, and the fact that the dosimeters show a non-uniform energy response (see Section 10.2.4), the dosimeters used to measure the beam profile should be correct ed. However, it was outside the scope of this research project to determine correction factors for this measurement. In stead, the beam profile in Figure 10-5, which does show a small decreas e in beam intensity corresponding to the heel effect, is a more accurate asse ssment of the Z-axis profile. 10.2 Characterization of OSL Dosimetry System 10.2.1 Tracking of System Standards As described in 8.2.1, standard operating values were established for the m icroStar OSL reader to ensure proper functioning of the LED beams used to read the dosimeters. The numerical means of these 50 measurements are 1.6 counts for the DRK, 2301.3 counts for CAL and 590.3 counts for the LED. These values, as we ll as the corresponding limits are listed in Table 10-7. After the standard values for the system were established, all three measurements were tracked over the course of two weeks to ensure system stability. Figures 10-8, 10-9 and 1010 display the counts of the DRK, CAL and LED measurements, respectively, graphed at each day measurements were made and shown with the standard value. The DRK is displayed with the manufacturer-recommended upper limit; CAL and LED are displayed with the manufacturer recommended upper and lower limits. After th e DRK, CAL and LED we re tracked for two business weeks and all measurements remained w ithin the established limits, it was decided that the system was functioning properly and ready for use.

PAGE 125

125 10.2.2 Dosimeter Response The results of the m ethodology outlined in 8.2.2 to assess the dose response of the dosimeters are shown in Table 10-8. Exposed unde r the same conditions, the means of the dose measured by the standard dosimeters (93% sensitivity) and screened dosimeters (91% sensitivity) was 47.36 mGy and 48.11 mGy, respectively. The coeffici ents of variation (CV) of the same measurements are 4.2% and 2.3%. These values are in agreement with the manufacturer-specified accuracies of 5% and 2% for the standard and screened dots. The OSL reader takes the sensitiv ity of each dosimeter into account duri ng the calculation of dose, so it is expected that there is no significant difference between the means measured by the two sets of dosimeters. Although there was no difference in the dose re sponse of these two t ypes of dosimeters, the manufacturer-specified accuracy, as well as the difference in the standard deviations of the measured dose response, were the differences be tween two dosimeter types, and this difference was taken into consideration when choosing wh ich type to use for measurements. Because energy and scatter correction factor s were applied to organ doses, the most accurate and precise measurements for energy and scatter corrections were desired. Thus, the screened dosimeters were used for both energy and scatter correcti on factors (10.2.4 and 10.2.5). In the case of the adult brain and cardiac protocols, both dosimeter types were used for organ dose measurements. The screened dots were used for the single exposur e because of their tighter specifications, while the standard dosimeters were exposed five times. 10.2.3 Room-Light Erasure After exposure, dosim eters were exposed to white light and periodically analyzed by the OSL reader, as described in 8.2.3. When all ten dot s were read once (elapsed time 120 minutes), there was still a significant amount of dose trap ped within the OSL mate rial, approximately 73%

PAGE 126

126 of the initial dose. Therefore, the process of reading the dosimeters at 10-minute intervals was repeated. After 320 minutes, almost 92% of the initial dose had been erased by the room light. After 1460 minutes (over night white-light exposure), 98% of the initial dose was depleted. A graph of these results is depict ed in Figure 10-11. A 24-hour er asure period was sufficient to erase the dosimeters to an acceptable background level. 10.2.4 Energy Response Energy response of the dosim eters was measur ed as outlined in 8.2.4. As expected, the dosimeters did not show a uniform response, when compared to the ionization chamber, at CT energies. Due to the small size of the 0.6 cc ionization chamber, it was the standard against which the nanoDot dosimeters were compared. Table 10-9 displays the mean of the five dosimeters used at each tube voltage, the dose measured by the ionization chamber and the ratio of the two. While the dots are calibrated at 80 kV, the HVL used in calibration (2.9 mm Al) is much lower than the one measured for the CT be am (4.65 mm Al, Table 10-5) used for these measurements. Therefore, the dosimeters show a surprisingly good response at this energy. A caveat in this analysis is th e fact that the energy response of the 0.6-cc chamber is not wellestablished at 80 kV, introducing a level of uncertainty in the doses measured at that tube voltage, a matter of future work, as described in the next chapter. However as the tube voltage increased, the response of the dosimeters decrease d when compared to the ionization chamber. For 100, 120 and 135 kV, the difference between the mean dose measured by the dosimeters and the dose measured by the ionization chamber was 5, 11 and 17%, respectively. Because of these significant differences, an energy correction fact or was applied to the dosimeters during organ dose measurements (further detailed in 10.2.5).

PAGE 127

127 10.2.5 Scatter Response The response of the dosim eters to scatter was first measured using the flat beam of a general x-ray tube, as outlined in Section 8.2.5. Fi rst, aluminum was added to the x-ray beam to simulate the HVL of the CT beam at three diffe rent tube voltages. The CT HVL, the simulated HVL and the percent difference between the two is given in Table 10-10. To compare the dosimeter response to the ionization chamber, th e average of the five dosimeters measurements was divided by the ionization chamber dose. Th is quotient was calculated for measurements made at the surface and under increasing thicknesses of acrylic for each tube voltage. Specific values are given in Table 10-11. The larges t difference between the dosimeters and the ionization chamber occurred for measurements made using the 80 kV beam and was 7%, which as explained in the previous section, is most likely due to the unknown energy response of the 0.6-cc ion chamber at 80 kV. This result is not expected, as the dosimeters were calibrated at this energy. A small contribution to this discre pancy could be reproducibi lity of the x-ray tube; however reproducibility was monitored througho ut the measurement process because each exposure was made three times. Another contri bution could be the positioning of the dosimeters and ionization chamber. The chamber was localized in position and an effort was made to keep it fixed in the same location because even though a flat radiation field was used for the exposure, small differences in position coul d result in small differences in the measurements. Similarly, the dosimeters were placed in the same locati ons around the chamber; however, differences in placement would increase the differences among measurements. Smaller differences were calculated for the 100 and 120 kV beam, 2% and 4% both of which are more consistent with expected measurement error with sources of erro r being the manufacturer-s pecified accuracy of the dosimeters.

PAGE 128

128 Because both energy and scatter corrections woul d need to be applied to the organ doses, scatter response was repeated. However this s econd set of measurements utilized the CT x-ray beams used to measure organ doses (Chapter 9). The results were similar to those measured using the flat x-ray field. The small size of the 0.6 cc ionization allowed the five surrounding dots to be close together, thereby minimizing th e heel effect of the CT beam. Measurement details are listed in Table 10-12. The tube volta ge and filter used on th e 64-slice scanner was included with the measurements on the 320-slice scanner to obtain corre ction factors for all clinical protocols used to measure organ doses. 10.2.6 Dosimeter Calibration Effective energies were calcu lated from meas ured HVLs of the beams used for clinical protocols, as described in Section 8.2.6 Table 10-13 lists the e ffective energies, mass attenuation coefficients for tissue and air and calculated calibration factors for each HVL measured during this research project. Valu es of effective energy range from 37.2 keV to 53.1 keV which is consistent with those measured by Mori et al. for a 256-slice scanner.64 The range of the f-factor calculated was from 1.05 to 1.07, consistent with the value of 1.06 published recently by the American Association of Medical Physicists (AAPM) report on the measurement of radiation dose in CT.11 10.2.7 Linearity Response As described in 8.2.6, the linearity response of the OSL dosim eters was analyzed by exposing the dosimeters to changing tube curren ts. The minimum tube current available in service mode was 10 mA and the maximum wa s 580 mA. The mean dose was plotted as a function of the tube current-time product and is displayed in Figure 10-12. The measurements were fitted to a linear line and th e resulting correlation coefficient (R2) was 0.9972, indicating that the dose was linear across the range of mAs values.

PAGE 129

129 If the appropriate energy correction factor is applied to the dosimeters, comparison can be made between the mean dosimeter dose and the ion chamber dose. A correction factor of 0.86, corresponding to the 120 kV and the Medium-L f ilter, was applied to the mean dosimeter doses and plotted as a function of mA in Figure 10-13. The ion chamber dose is also plotted in the same figure as a function of mA As evident by many overlapping measurements, the dosimeter linearity is comparable to that of the ion chamber. In order to assess the linearity of the dosimeters in dose range us ed in this research project, a second set of measurements was made using a tube current range of 500 to 8000 mA. As shown in Figure 10-14, the dosimeters the linear res ponse over the range of tube currents with an R2 value of 0.9998. When the same energy correc tion of 0.86 is applied to the dosimeters, agreement is seen in Figure 10-15 between the re sponse of the dosimeters with that of the ion chamber. 10.2.8 Angular Response 10.2.8.1 In-air response The m ean doses measured by the nanoDot dosimeters in air, as described in 8.2.7.1, are presented as a function of x-ray tube angl e in Figure 10-16 with the corresponding 95% confidence interval (CI). The mean doses at the 0-75 and 105-180 positions were fairly constant: the range of mean doses measured at x-ray tube angles of 0 to 75 and 105 to 180 was 8.0 to 8.5 mGy and 7.9 to 8.4 mGy, respectivel y. The small variation in these doses suggests the dosimeters show very little dependence with angle of exposure when the dosimeters are used in-air. The minimum mean dose for the in-air measurements was 6.45 mGy measured at the 90 x-ray tube position with respect to the dosimeter. This re sult is expected, as it corresponds to the situation when the x-ray tube is perpendicular to the active area of the

PAGE 130

130 dosimeter. The poor response measured at the 90-position is of no concern for clinical purposes, as the dosimeters were not used in air for clinical organ dose estimates. 10.2.8.2 In-phantom response Figure 10-17 depicts the dose response of th e nanoDot dosimeters exposed in the head phantom at different x-ray tube positions. Measurements were made at 15 intervals to include the two positions (90 and 270) when the CT x-ray beam is perpendicular to the active material of the OSL dosimeter. One might expect a drama tic decrease in dose at these two positions, but that was not observed; rather a one-way ANOVA suggests that the m eans of this data set are not significantly different (p=0.680), in dicating that the dosimeters respond similarly to all angles of the x-ray tube. 10.2.9 Comparison of OSL Dosimeter In and Out of Its Light-Tight Case Section 8.2.8 described the m ethod in whic h dosimeters were exposed with the OSL material out of the light-tight plastic case. The mean of these measurements was 10.97 mGy at the 12 oclock position and 8.54 mGy at the center position. For comparison, the same measurements were made with the dosimeters in side their cases. Mean s in the same phantom locations were 10.78 mGy and 8.28 mGy, respectivel y, as listed in Table 10-14. A one-way analysis of variance (ANOVA) suggests that the means of the measurements made with the dosimeters in and out of the plastic case at th e 12 oclock position are no t significantly different (p=0.657). Similar results were found when comparing the means at the center position (ANOVA p=0.116). Therefore, the measured dose is not affected when the OSL material is removed from its plastic casing for organ dose measurements. 10.3 Organ Dose Measurements For each of the three clin ical protocols used for organ dose measurements, small dosimeters were used, and point doses were appr oximated by these dosimeters. For most organs,

PAGE 131

131 multiple dosimeters were placed within the phantom at several locations within the organ. Large organs, such as the lungs, required a larger numbe r of dosimeters to cover the full volume of the organ, compared to small organs, such as the thyr oid, In the case of large organs, there is a possibility that the organ was only partially irra diated by the CT x-ray be am. For example, the lungs extend throughout most of the thoracic cavity. The 160 mm length of the cardiac CTA volumetric scans did not cover th e entire lung volume. Instead, th e volume of the lungs in the primary beam received both primary and scattered radiation, while the rest of the lung volume only received scatter. The averaging of all doses measured in the entire lung volume may not be representative of this non-unifo rm dose distribution when considering the risk of radiationinduced damage to lung tissue. To account for th is uncertainty in the organ doses measured in this research project, two values were reporte d for all organ dose measurements: the mean and the maximum. The mean would mo st likely represent a realistic an alysis of the risk associated with each CT study analyzed in this research project due to the limitations of dosimeter distribution in the phantom. In this case, th e maximum would result in a conservative risk calculation that would also repres ent the worst-case scenario. 10.3.1 Pediatric Head Study 10.3.1.1 Organ doses A com parison of the mean organ doses resu lting from the pediat ric craniosynostosis protocol measured with five sets of dosimeters a nd one set of dosimeters is given in Table 10-15. The mean, maximum and 95% confidence interval (CI) are listed in Tabl e 10-16 for the set of five measurements. Energy correction factors were applied to surface measurements and were acquired as described in 8.2.5 and lis ted in Table 10-10. Skin, lens of eye and breast doses were corrected using the factor calcu lated on the surface of acrylic. The thyroid doses, which were measured in the phantom, were corrected using a scatter correction factor corresponding to 1 of

PAGE 132

132 acrylic. These factors are listed in Table 10-10. The volumetric protocol used the Small-S filter for both 100 and 120 kV; the helical protocol used the Small-S filter and 120 kV. As expected, all organ doses were higher usi ng a tube voltage of 120 kV as compared to 100 kV. For all protocols, the dose to the lens of the eyes was slightly higher (approximately 16% higher) than the dose to the skin. Because these surface measurements were made in close proximity to each other on the phantom, the dos es should be about the same. The small difference between dose to the skin and lens of th e eye could be due to ve rtical height of the dosimeters; those close to the x-ray tube should receive a higher dose simply due to the closer proximity to the x-ray source. The small values of the thyroid dose suggest th at the dosimeter located in the thyroid was not in the primary x-ray beam, as expected. Wh ile an effort was taken to reduce statistical uncertainties by using five different sets of dos imeters and measuring orga n doses five separate times, the 95% CI associated with the thyroid do se is greater than bot h the mean and maximum of the measurements performed on the 320-slice sca nner at 120 kV. The f act that the measured doses are small adds to this statistical error. At the same tube voltage (120 kV), the thyroid dose measured on the 64-slice scanner is roughly 70% higher than the dose measured on the 320-slice scanner. Since the scan lengths were the same for both acquisitions, th e difference in thyroid dose is most likely due to the difference acquisition techniques of the sca nners: volumetric axial acquisition and helical ac quisition. As described in Section 3.2.2, helical CT scans required an additional half-rotation of the x -ray tube at the beginning and end of the scan boundaries to information there is enough data for the necessary interpolation. This over-beaming effect is investigated by Winslow65 and contributes to the measured in organ at the periphery of the scan boundaries. Furthermore, the numerous rotations of the x-ray tube during a helical scan produce

PAGE 133

133 an increasing amount of scattered radiation to tissues and organs inside and outside of the primary beam. This increasing sc atter effect is more obvious for organs outside of the primary beam, as the thyroid was for this protocol. Because of the wider beam coverage of the 320-slice scanner, only one rotation of the x-ray tube is required to cover the same area and less scatter is produced by this single rotation as compared to the multiple rotations of the tube during the helical scan. For all protocols, the breast dose was less th an 2 mGy. The means follow the same trend as the other organ measured: the dose measured at 100 kV was the smallest and of the two scanners, the 320-slice scanner resu lted in a lower dose for the 120 kV set of measurements than the 64-slice scanner. Similar to the thyroid, the dosimeters that m easured dose to the breast were not in the primary x-ray beam during the scan. However, because the breast dose was measured on the surface, the effect of the increased scatter due to the multiple rotations during the helical scan was not observed. Surface measurements are not as susceptible to scatter radiation within the phantom as in-phantom measurements because surface dosimeters receive only backscatter from the phantom while in-phantom dosimeters receive scatter from material surrounding the entire dosimeter. Therefore the breast doses were only slightly higher fo r the 64-slice scanner. While the difference in the thyroid dose betw een the two CT scanners may be explained by an increase in the amount of scatter, the signi ficant difference in surface dose measurements between the two different CT scanners was not expected. The lower HVLs measured using the 320-slice scanner might imply a higher skin dose becau se of the softer beam, as compared to the 64-slice scanner. This trend was not observed an d may be due to a difference in the effective tube current for each scanner. The effective tube current (for this manufacturer) is equal to the produce of the tube current and rotation time fo r the 320-slice scanner an d was equal to 121 for

PAGE 134

134 this protocol. For helical scans, this valu e is divided by the pitch to account for overor underscanning. In this case, a pitch of 0.641 was used on the 64-slice scanner, resulting in an effective mAs of 156. The higher effective mA s used for during acquisition on the 64-slice scanner contributed to the overa ll higher doses than the same acquired on the 320-slice scanner. One parameter that would impact the measured organ doses that coul d vary by patient is the total scan length. For a volumetric acquisitio n, the scan range may be changed by the user from 100 mm to 160 mm in 20 mm increments. The s can length needed to cover the entire brain is determined clinically from the localizing tomograph. This process was used to measure organ doses, and a range of 120 mm was selected. Ho wever an increase in this range would most likely increase organ doses, especially to that of the thyroid, as the scan length extends down towards the neck bringing the primary radiation field closer to the organ. Similarly, a small patient might require a smaller range and receive a slightly smalle r radiation dose. As described in 9.1, the same scan length was used for all pha ntom measurements for the best comparison. One limitation of this pediatric phantom study wa s an inability to assess patient motion. A single volumetric acquisition acquired in 0.6 seconds will most likely result in significantly less patient motion than the helical comparison. While the actual x-ray tube rotation is of shorter duration (0.5 second), the smaller beam width of the 64-slice scanner re quires several rotations of the x-ray tube about the pati ent to cover the entire scan range adequatel y. This longer total scan time allows more time for the patient to move during the scan. Limiting patient motion is of special concern in pediatric cases simply due to the nature of young babies. This phantom study did not allow for investigation into this issue; however, one can conclude the lower dose and decreased total scan time make this protoc ol a more viable option on the 320-slice scanner than the 64-slice scanner.

PAGE 135

135 A second limitation of this pediatric phantom was the inability to measure organ doses other than the thyroid in the phantom. Unlike the adult phantom, the tissue-equivalent material of the pediatric phantom is solid. A hole need ed to be drilled into the phantom for the measurement of the thyroid dose, and the same woul d have had to be done for other organ doses. Since a head protocol was chosen to assess pediatric doses, surface measurements to assess dose to the skin, lens of the eye and breast were appropriate as these are the organs of concern in a head protocol. 10.3.1.2 Image quality To evaluate any possible dose reduction options as outlin ed in 9.1.3, three medical physicists scored images, shown in Figure 10-18, of the LCD section of the Catphan acquired using the same parameters as those of the volumet ric craniosynostosis protocol (Table 9-1). The complied scores are listed in Table 10-17. Viewer 1 observed a difference in LCD between the images acquired at 100 and 120 kV, scoring a tota l of 20 objects for the 120 kV case and only 10 objects for the 100 kV case. Differences in sc ores of the combined supra-slice (A-C) and separately, of the combined s ub-slice (D-F) objects also indi cate that Viewer 1 observed a difference in LCD between the two images. View er 2 did not observe a difference between the two images because the total num ber of objects was scored equall y. Similarly, the number of combined supra-slice and combined sub-slice obj ects was scored the same for the two images. Viewer 3 scored the two images similarly, with small differences between the total number of low-contrast objects detected in the two images and small differences between the combined supra-slice and combined sub-slice objects. Because no definite trend could be inferred fr om the LCD scrores, further analysis was done. The mean of the combined supra-slice object s was calculated from the scores of the three viewers for the 120 kV and 100 kV images and found to be 10 and 9, respectively. Similarly, the

PAGE 136

136 mean of the combined sub-slice scores was 7 and 5 for the 120 kV and 100 kV images, respectively. To determine if these results are significant, a one-way analysis of variance (ANOVA) was used to compare the total scores of all three viewers of the 120 kV and 100 kV images. The difference between the means of the LCD scores was not found to be significant (p=0.4830). The same analysis was used to compare the means of the combined scores for the supra-slice and sub-slice groups; there was no significant differe nce between the 120 kV and 100 kV images for both combined groups (p=0.6784 and p=0.3206, respectivel y), either. While these results appear to indicate that there is not a significant degr adation in image quality when lowering the tube voltage for this protocol, it is clear that future work in this area of th e research project must include a larger sample of viewers in or der to make more definite conclusions. 10.3.2 Adult Brain Perfusion 10.3.2.1 Organ doses Table 10-18 lists two sets of data for each pr otocol and organ resulting from the brain perfusion protocol were measured as described in 9.2. One set of standard nanoDot dosimeters was placed on and in the phantom, and the scan was repeated five times with the same set of dosimeters. The mean of these five scans was calculated. This method was chosen to reduce statistical error in measuring the organ doses once, as well as to minimize error due to reproducibility of the x-ray tube. As seen in 10.1.2, up to a 6.2% standard deviation was measured using the clinical mode of the scanner, and therefore measured organ doses are subject to this variability. Since only one set of dosim eters is used, any def ect in a single dot is exaggerated by scanning five times. To verify these organ doses, screened nanoDot dosimeters were used for the same measurement but scanned only once. As evident by the small differences in the means using these two methods, organ doses are very comparable. The mean organ doses,

PAGE 137

137 as well as the maximum and 95% CI of the measur ements are given in Table 10-19. As with the pediatric phantom measurements, surface measur ements, were corrected for energy and organ doses measured in the phantom were corrected for energy and scatter. Of the three original manufacturer protocols, the acquisitions performed at 120 kV resulted in the highest mean organ doses, as expected. Fo r all three protocols, th e mean dose to the eyes was the largest dose measured and 10-14% higher than the skin dose measurements. This trend was also observed for the same measurements pe rformed with the pediatric protocol, and as described in 10.3.1.1 could be due to the vertic al position of the dosimeters at each organ location and the heel effect. The breast dose was the lowest of all organs measured, as expected because the breast was set outside of the boundaries of the scan. The doses measured in the esophagus provide a good example of reporting the mean and maximum doses. This organ was outlined in three different slices of the phantom. Because the esophagus is not in the direct ra diation field, the amount of scattered received by the dosimeters increases as the slices get clos er to the head. A non-uniform dos e distribution was measured in this organ as a result of different amounts of scattered radiation reaching different parts of the organ. The mA boost and continuous prot ocols resulted in approximately the same mean doses for all organs. Although the acquisiti ons differ between these two prot ocols, the total tube currenttime product is approximately the same for both pr otocols. This total mAs is calculated by multiplying the tube current, rotation time and the number of volumes acquired. The mA boost protocol uses a scan time of 0.75 s, where the continuous protocol volu mes are acquired with a 1.0 s rotation. However the tube current is highe r for the mA boost protocol compared to the continuous protocol, so the mAs is comparable.

PAGE 138

138 Despite the total scan volume being smaller, th e mean skin dose resulting from the 64-slice scanner was higher than all other mean doses. Th e smaller volume spares dose to the eyes, but the mean skin dose is approximately 4.5 times hi gher in the 64-slice scan ner when compared to the mean skin doses resulting from the volumetric protocols. The mean skin dose resulting from this protocol was the highest measured during this research project. Skin erythema effects can occur for doses above 2 Gy.66 While this limit is not appro ached for a single brain perfusion study in the 64-slice scanner, it could be approached if the scan needed to be repeated. With the development of the brain perfusion protocol in the 320-slice scanne r (discussed in 10.3.2.2) complete, the mA boost and continuous protocols are recommended in place of the 64-slice scanner whenever possible. Comparison of the mean doses of the dosimeter s scanned five times to those scanned once shows one of the largest differences for the skin measurements on the 64-slice scanner. For the skin measurements, dosimeters were placed on the surface of the phantom, as shown in Figure 92. Along with the axially-acquired volumetric sca n, the helical protocol also consists of three helical scans, detailed in Table 9-7. While th e nominal beam width used was 32 mm and able to cover the active volume of the OSL dosimeters, the helical nature of the s can could result in the beam not directly passing over one or more of the surface dosimeters. If the x-ray tube is at the bottom of the gantry these dosimeters measure th e radiation transmitted through the entire head. Because the CT scanner table moves during he lical acquisitions, the dosimeter may not be exposed fully to the primary beam. This patter n may result in an inaccurate dose measurement if the primary beam was not measured by the dosim eters. However, this pattern would be representative of the dose distri bution experienced by a single pa tient undergoing the same scan and therefore a reasonable measurement. To report an average dose for this particular scan, the

PAGE 139

139 more representative measurement is that of the do simeters scanned five times; there is a better chance of full exposing the dosimeters during the scan resulting in an average skin dose measurement. 10.3.2.2 Image quality As detailed in 9.2.4, the brain perfusion pr otocol underwent several changes during the developm ent process in an effort to produce hi gh-quality images with the lowest reasonable dose. The initial brain perfusi on protocol (Manufacturer at 80 kV, Table 10-5) resulted in an image set that was of inferior image quality comp ared to the current perf usion studies performed on 64-slice scanner. The timing of the dynamic vol umes generated image sets that was lacking diagnostic information during the arte rial phase of contrast-agent uptak e. In order to ensure that a complete set of diagnostic images were acqui red, the first amendment to the brain perfusion exam was the addition of a helically-acquired CT angiogram (CTA) of the head. Once the timing issue with the dynamic volumes had been resolved, the diagnostic quality of the dynamic volume image sets were still considered poor; the neuroradiologist proposed an increase of the tube voltage used to acquire those volumes from 80 kV to 120 kV. While this is a reasonable approach, the medical physicists s uggested an increase to 100 kV, as dose to the patient increases with increasing tube voltage. The second and thir d iterations of the prot ocol involved modifying the tube voltage to 100 kV and 120 kV, respectively. As expected, the image sets acquired at 120 kV were deemed of higher image quality by the neuroradiologist compared to the same ac quired at 80 and 100 kV. However, due to the considerable dose savings, especially to the skin and eyes, and the fact that the images acquired at 100 kV were deemed by the neuroradiologist to be of diagnostic quality, the RPC accepted the 100 kV modifications to the protocol.

PAGE 140

140 A second protocol, originally designed for evaluation of arteriovenous malformations (AVM), was proposed by the neuroradiologist. This protocol (Continuous, Table 10-5) uses a continuous acquisition of 19 volumes with a one-s econd tube rotation and therefore the beam is on continuously for 19 seconds. The neuroradiologist presented this to the RPC, and, after preliminary dose measurements, the protocol was us ed clinically. The success of this protocol for AVM evaluation led the neuror adiologist to propose the same protocol for brain perfusion. Thus far, this protocol has been a success and offers the lowest organ doses to the patient. Finally, a fifth iteration of th e brain perfusion protocol was developed by the manufacturer in response to the image quality issues reported by the neuroradiologist and the RPC. Instead of increasing the tube voltage of the original pr otocol (Manufacturer at 80kV, Table 10-5), the manufacturers proposed protocol increased the tube curr ent from 100 mA to 310 mA during the initial arterial phase. While the increase in tube current will increase the organ doses compared to those of the original protocol, the increase in dose due to increase mA was less than using a higher tube voltage to achieve images of diagnostic quality, as determined by the neuroradiologist. At the time of the conclusion of this resear ch project, the Continuous protocol was being used clinically more frequently than any of th e other RPC-approved protocols. Considering that this protocol resulted in the lowest measured organ doses, a balance of image quality and dose was achieved. The protocol develo pment process involved several it erations of the protocol, as well as the involvement of several members of th e RPC and radiology department and resulted in successfully providing high-quality care to patients.

PAGE 141

141 10.3.4 Adult Cardiac CTA 10.3.4.1 Organ doses The m ean organ doses resulting from the volum etric and helical protocols are listed in Tables 10-20, detailing the doses measured from five exposures and single exposure. Table 1021 lists the mean doses averaged over the fi ve exposures, including th e maximum and 95% CI associated with the doses. As with all other measurements, organ doses were corrected with energy and scatter corrections appropriate, as ex pressed in Equation 8-3, to the tube voltage and filter combination used for each protocol, as lis ted in Table 10-12. As described in 10.3.2.1, two sets of dosimeters were used to measure organ do ses: one set of standard dosimeters was scanned five times and averaged, and screened dots were scanned one time. For all protocols, the breast and skin doses were the highest among the organ doses measured. Skin doses were measured on the surface across the breast in two longitudinal locations illustrated in Figure 9-9. Breast dose was measured in the phantom, as depicted in Figure 9-10. These results were expected, as both organs are within the radiation field centered at the heart for cardiac imaging. The mean and ma ximum lung doses were less than those of the skin and breast doses. The result may be explaine d by the differences in scatter characteristics of the lung tissue and breast tissues. The lung-tissue equivalent material is similar to an air cavity which provides less scatter than soft-tissue equivalent material. The decrease in the amount of scatter in the lungs compared to the breast resulted in smaller measured doses in the lungs. The same rationale may be applied to the dose differe nces between skin and lung measurements. The skin measurements were made on the surface of the phantom; the dosimeters absorb the primary beam radiation on the surface as well as backsc atter. While the dosimeters in the lung are completely surrounding by lung tissue-equivalent ma terial, minimal scatter is produced and the dose absorbed is less than that on the surface of the skin.

PAGE 142

142 Of the three protocols evaluated on the 320-sli ce scanner, the prospectively-gated protocol resulted in the lowest mean organ doses. Whenev er possible, and clinically indicated, this is the protocol that should be used. The clinical limitation of this protoc ol is the requirement of a low heart rate. The manufacturer-specifies a maximum heart rate of 60 bpm; if the heart rate is too high, the 175 ms scan time is too long to be comple ted while the heart is in diastole. Therefore, patients with heart rates higher th an 60 bpm need to be given medi cation to slow the heart during the scan or receive higher doses because more than one acquisition is required by the system. When complete functional analysis of the h eart is clinically necessary, the prospectivelygated protocol is not an option. Of the two functional analysis protocols, the protocol that includes dose modulation resulted in lower organ doses, as expected. On average, there was a 40% dose reduction in all mean organ doses usin g the dose modulation feature for the functional analysis protocol and is the prefer red protocol for functional analysis. In general, the 64-slice protocol resulted in higher doses than all of the volumetric protocols. Of particular intere st are the mean skin doses measured using a single scan with a mean of 25.37 mGy and those averaged over five scans with a mean of 76.14 mGy. There was not another set of measurements w ith a difference this large between the two sets of dosimeters. One contributor to this large difference is the he lical nature of the sca n. Although measures were taken to reproduce the position of the phantom and the dosimeters within the CT scanner, there was no guarantee that the x-ray t ube exposure would begin at the same location with respect to the phantom. Additionally, the helical scan mean s the table translates through the gantry while the exposure is taking place. The spiral dose prof ile that results could explain the difference in the skin dose measurements, as also described for the skin doses measured for the brain perfusion protocol on the 64-slice scanner (s ection 10.3.2.1). On the surface, the x-ray beam

PAGE 143

143 may not have directly exposed the dosimeters, or some of the dosimeters, causing a non-uniform dose deposition. It is more likely that this was the case of th e single exposure; performing the scan 5 times and taking the average should cause the x-ray tube to travel in a different path with each scan due to the differences in start positions of the tube. The fact that this difference in mean doses is seen only on these skin dose meas urements lends credibility to this argument, despite having no means available for confirmation; skin dose m easurements on the surface of the phantom were subject to both the primary x-ra y beam as well as scattered radiation. Organ doses measured in the phantom result from the attenuated x-ray beam (as it passes through the phantom) and scattered radiation. Similarly, the mean dose to the stomach is si gnificantly larger with the 64-slice protocol than all volumetric protocols. The obvious contri butor to the larger dose could be that the organ was included in the helical and not in the volum etric scan. However, the scan range was 160 mm for all cardiac scans, and the begi nning of the scan was also set to the same slice on the phantom, based on the alignment lasers and localizing tomographs. One possible reason for this discrepancy between mean stomach doses is the added half-rotation necessary for helical CT scanning. Despite data being acquired helically, im ages are reconstructed in the axial plane. There is a half rotation of the x-ray tube at the beginning and end of each helical scan that exposes the patient without data being used for r econstruction purposes. If this half rotation at the end of the scan contained the stomach, a higher dose would be measured than if only scattered radiation contributed to the mean dose. These added half rotations are not necessary in the axial acquisition of the volumetric protocol s, hence the stomach dose is lower in the volumetric protocols. A second contribution involved the increas ed amount of scatter generated by helical scans. As previously described (s ection 10.3.3.1), the small beam width results in an

PAGE 144

144 increased number of helical scans to cover the same total scan le ngth as the volumetric protocol. As the number of x-ray tube rotations increase the amount of scatter produced within the phantom also increases and resulted in higher stomach doses measured on the 64-slice scanner compared to those on the 320-slice scanner. 10.3.4.2 Image quality To evaluate im age quality of these cardiac CTA protocols, the Catphan image quality phantom was used, as described in 9.4.2. The fi rst protocol analyzed was the prospectivelygated protocol. The images are reconstructed at 0.5 mm slices, and Catphan images of the lowcontrast module were also acquire d at that slice thickne ss; the image in the center of the module is displayed in Figure 10-19. During scoring of the image, window width and level were used to adjust the image for best visualization of the lo w-contrast objects. De spite this manipulating, none of the objects were visible in the image. The same situation occurred with the image set reconstructed at 1.0 mm slices. Not being able to score this low-cont rast module with these reconstruction parameters is a lim itation of the Catphan and its abil ity to assess image quality for new volumetric scanners. Thicker slices are n eeded to combine information from neighboring thin slices for proper viewing of objects in th e phantom. Growing technology has allowed for much thinner slices to be reconstructed than ol der generation scanners. Clinical indications for this protocol involve the imaging of small vesse ls which is achieved with images reconstructed with thin slices. Despite the limitations of the Catphan for this analysis, the parameter of most concern in these cardiac protocols is the temporal resolu tion of the acquisition. Because cardiac motion cannot be stopped, it is important to acquire images at the correct point in the cardiac cycle and to do so while the heart is in as motionless of a state as possible. Where this short acquisition time is the parameter that dictates these cardi ac protocols, such that time cannot be changed

PAGE 145

145 without negatively affecting temporal resolution, other parameters may be changed, but doing so will most likely not affect image quality and dose in a positive way. The effective tube current used for prospectively-gated and functional an alysis protocols was 140 and 175, respectively. While these values may appear low, they are the result of the short acquisition time. The actual tube current is 400 and 500, resp ectively; the maximum tube curre nt available on the scanner is 580, so the values used for these protocols are ap proaching the maximum tube current. The tube voltage could be increased to 135 kV, but this increase would result in an increase in organ doses. Since the image quality at 120 kV is su fficient, as evidenced by the number of clinical cardiac CTA studies that have been performed without image quality complaints by the radiologists to the RPC, there is no clinical need to increase th e tube voltage and consequently organ doses. Furthermore, cardiac images are acquired with the use of a contrast agent to image vessels. The clinical use of a contrast agent may be better assessed using the high-contrast resolution module of the Catphan. However, the technology of modern CT scanners has placed the limiting resolution of the images in the detect or resolution, i.e., the si ze of the detectors used to acquire the images defines the limiting resoluti on, thereby limiting the Cat phan in its ability to measure the resolution of a CT system. 10.3.5 Comparisons to Monte Carl o Organ Dose Simulations Work was done by Ghita67 to estimate organ doses using Monte Carol simulations for the same three clinical protocols inve stigated in this research proj ect. These simulations were done for the 320-slice scanner and includ e most of the organs chosen for dose measurements using the OSL dosimeters. Comparisons of the measured and simulate d doses for the pediatric craniosynostosis protocol are shown in Table 1020. The relative differences li sted compare the MC-simulated doses to the mean dose of the OSL dosimeter measurements for the organs selected. The

PAGE 146

146 smallest relative difference (1.72%) among the doses measured is for the lenses of the eyes for the 100 kV case and is due to the fact that the lens of the eye is a small organ completely positioned with the primary CT beam during the scan. Two dosimeters were used for these measurements, and therefore the dose was well-samp led; in addition, the small size of the organ makes the numerical measured average a more r ealistic representation of the dose throughout the organ. The largest relative difference is 11.38% for the thyroid doses at 120 kV; this organ is at the edge of the actual scan boundaries and most lik ely it is only exposed to scattered radiation from the scan. However, one limitation of the dosimeter measurements is highlighted by the results shown for this organ. The fact that the thyroid is assessable only in one slice of the phantom limits the ability to achieve a unifo rm dosimeter placement throughout the organ; measurements therefore correspond only to that slice of the phantom, whereas the Monte Carlo simulations used data for the entire volume of the organ. The relative differences among the measured and simulated skin doses are -4.93% and -8.32%, at 120 and 100 kV, respectively. In fact, the simulated doses fall in between the mean and the maximum measured doses, further justifying the reporting of both th e maximum and mean measured doses as described in Chapter 9. The comparisons between measured and simulated organ doses for the adult brain perfusion protocols are given in Table 10-21. For th e five protocols, doses to the lenses of the eyes are most comparable among all other organs investigated. The relative differences range from 2.01% to 7.15%, due to the adequate sampling because of the small size of the organ and its location in the primary beam during the scan. Of the other four organs, the relative differences among the skin doses are comparable, and range from 12.04% to 33.07%. Similar differences are seen among the thyroid doses, with a range of 13.20% to 21.96%. The thyroid is a small

PAGE 147

147 organ at the edge of the primary beam and, while the relative differences are apparently large, this difference is rather misleading as the doses ar e small thereby inflating the relative difference. For example, a measured difference of only 1.3 mGy between the simulated and measured thyroid doses for the mA boost protocol, which corresponds to a 20% relative difference. The same was found with the comparison among meas ured and simulated esophagus doses; the differences in doses range from approximately 1 to 3 mGy, whereas the relative differences range from 41 to 58%. The largest discrepancy between measured and simulated doses for the brain perfusion protocols corresponds to the br ain doses, ranging from -21.8 to 26.6%. This discrepancy is attributed to dosimeter placement being non-uniform throughout the organ because of limitations in access to it from all s lices in the phantom, as described in Section 9.2.1. Finally, comparisons were made among meas ured and simulated doses for the thyroid, lung, stomach and skin resulting from the adu lt cardiac CTA protocols a nd are found in Table 10-22. Of these four organs, the best comparison was made for the doses to the lungs and the relative differences range from 2.22 to 8.35%. Dosimeters were placed throughout the lung volume in a systematic way, as described in Sect ion 9.3 and depicted in Figure 9-7, and thus are a good representation of the average organ dose because of a better uniformly-distributed dosimeter placement. Good comparison is also seen for the thyroid and stomach doses. Large relative differences among the skin doses for each protocol are the result of using the entire volume of skin in the MC simulation whereas th e dosimeters were placed only on the section of the skin that was in the primary beam and therefore averaged over a smaller area as compared to the simulated skin doses. While limitations existed in both dosimeter measurements and MC simulations, good comparison between the doses obtained from the two methods was achieved. Although it is clear

PAGE 148

148 that the two methods cannot benchmark each othe r, the overall good agreement among all of the results allows for some general conclusions to be drawn. First, dosimeter placement within the phantom has a direct effect on the average organ dose, which is critical especially for larger organs. The lungs were well-sampled by the dos imeters for the cardiac protocol and compared well with the MC simulations of the same protocols. Conversely, due to limitations in the access to all areas of the brain in the phantom, the brain doses resulting from the perfusion protocols was not uniformly-sampled and led to larger di screpancies when compared to the corresponding MC-simulated doses. Doses measured in small organs, such as the lenses of the eyes and the thyroid compared well with si mulated doses. Generally speaking, the dose comparisons support the methodology used to measure organ doses us ing the OSL dosimeters with limitations of accessibility for certain organs in the phantom wh ich, account for the few organ doses that were not as comparable between the measured and simulated results.

PAGE 149

149 Table 10-1. Reproducibility of x-ra y tube operated in service mode. Measurement Measured air kerma (mGy) 1 48.26 2 48.28 3 48.29 4 48.30 5 48.30 6 48.29 7 48.28 8 48.32 9 48.29 10 48.30 Mean (mGy) 48.30 CV 0.03% Table 10-2. Reproducibility of volumetric clinical protocol. Position within phantom Measurement Center 3 oclock6 oclock 12 oclock 9 oclock Measured dose 1 20.10 22.12 24.65 22.74 18.96 (mGy) 2 20.15 21.33 22.98 22.13 20.03 3 20.64 22.00 25.32 21.88 21.12 4 20.35 21.72 24.52 24.08 19.31 5 20.68 21.80 23.00 23.58 19.47 6 20.17 22.60 26.18 24.74 19.62 7 20.65 23.30 23.07 26.29 20.67 8 20.13 21.74 24.10 23.82 19.61 9 20.59 21.63 23.93 24.99 19.54 10 20.59 21.31 26.40 21.88 20.95 Mean (mGy) 20.41 21.96 24.42 23.61 19.93 CV 1.2% 2.8% 5.1% 6.2% 3.7% Table 10-3. Reproducibility of helical clinical protocol. Position within phantom Measurement Center 3 oclock6 oclock 12 oclock 9 oclock Measured dose 1 76.49 78.78 77.08 81.65 80.13 (mGy) 2 76.59 78.83 77.20 81.76 80.23 3 76.60 78.83 77.11 81.80 80.21 Mean(mGy) 76.56 78.81 77.13 81.74 80.19 CV 0.1% 0.0% 0.1% 0.1% 0.1%

PAGE 150

150 Table 10-4. Measured beam energies of s canners used for organ dose measurements. Nominal tube voltage (kV) 80 100 120 135 Measured tube voltage (kV): 320-slice Scanner 80.9 101.3 121.7 137.8 % Error 1.1 1.3 1.4 2.1 Measured tube voltage (kV): 64-slice Scanner 80.8 100.9 121.7 138.3 % Error 1.0 0.9 1.4 2.4 Figure 10-1. Total filtration of x-ray beam usi ng the large filter and depicted for three beam qualities. The increase in total filtration as the distance from isocenter increases is representative of the sh ape of the bowtie filter. 0 2 4 6 8 10 12 14 16 18 04812162024Total Filtration (mm Al)Position from Center (mm) Total Filtration with Large-L Filter 120 kVp Large L 100 kVp Large L 80 kVp Large L

PAGE 151

151 Figure 10-2. Total filtration of x-ray beam for small and large filters using a 120 kV x-ray tube voltage. 0 2 4 6 8 10 12 14 16 048121620Total Filtration (mm Al)Distance from Center (mm) Total Filtration at 120 kV Small-S Large-L

PAGE 152

152 Figure 10-3. Effect of bowtie filter on total filtration measurem ents. While the exact shape of the bowtie filter is unknown, the shape depicted here is to illustrate the shape of the total attenuation curve. Table 10-5. Measured HVLs for each nominal xray tube voltage for the Aquilion One 320-slice scanner. Nominal tube voltage Filter Measured HVL (mm Al) 80 kV Small-S 3.90 80 kV Medium-M/Large-L 4.65 100 kV Small-S 4.98 100 kV Medium-M/Large-L 5.80 120 kV Small-S 6.02 120 kV Medium-M/Large-L 6.85 135 kV Small-S 6.76 135 kV Medium-M/Large-L 7.53 Note: For simplicity of display, the Medium-M and Large-L filters correspond to the 320-slice scanner.

PAGE 153

153 Table 10-6. Measured HVLs for each nominal x-ray tube voltage fo r the Aquilion 64 MDCT scanner. Nominal tube voltage Filter Measured HVL (mm Al) 80 kV Small-S/Small-M 4.84 100 kV Small-S/Small-M 6.09 120 kV Small-S/Small-M 7.16 135 kV Small-S/Small-M 7.92 Figure 10-4. Radiographic image of the 16-cm nominal beam width.

PAGE 154

154 Figure 10-5. Beam profile of the 16-cm nominal CT x-ray beam obtained in air and at scanner isocenter. Gray scale values are plotted as a function of distance. 10000 13000 16000 19000 22000 25000 28000 31000 050100150200250300350CountsDistance (mm) Beam Profile

PAGE 155

155 Figure 10-6. Normalized dose in air across the CT scanner gantry perp endicular to the anodecathode direction. 0 0.2 0.4 0.6 0.8 1 1.2 050100150200250300Normalized DoseDistance (mm) Beam Profile Perpendicular to Anode-Cathode Direction

PAGE 156

156 Figure 10-7. Normalized dose in air across th e CT scanner gantry in the anode-cathode direction. Table 10-7. Measurement standards for microStar reader. Reading Established average (co unts) Acceptable limits (counts) DRK 1.6 DRK<30 CAL 2301.3 2071
PAGE 157

157 Figure 10-8. Measurement of DRK counts over tim e, plotted with the standard value and upper limit. 0 5 10 15 20 25 30 35 12345678910CountsDays Stability of DRK Counts DRK Counts Standard Upper Limit

PAGE 158

158 Figure 10-9. Measurement of CAL counts over tim e, plotted with the standard value and upper and lower limits. 2000 2100 2200 2300 2400 2500 2600 12345678910CountsDays Stability of CAL Counts CAL Counts Standard +10% Limit -10% Limit

PAGE 159

159 Figure 10-10. Measurement of LED counts over ti me, plotted with the standard value and upper and lower limits. Table 10-8. Dose response of standard and screened dots. Standard dots Screened dots Measured dose (mGy) 43.92 49.00 48.03 49.66 50.22 47.78 48.06 47.06 45.79 47.18 49.93 49.67 45.47 46.47 46.46 48.52 48.25 47.82 47.44 48.00 Mean dose (mGy) 47.36 48.11 CV 4.2% 2.3% 500 550 600 650 700 12345678910CountsDays Stability of LED Counts LED Counts Standard +10% Limit -10% Limit

PAGE 160

160 Figure 10-11. Percent decrease from initial dose is displayed as a function of white-light erasure time. Table 10-9. Energy response of the OSL dosimeters. 80 kV100 kV 120 kV 135 kV Mean dose (mGy) measured by the dosimeters 17.07 27.76 43.75 46.42 Dose (mGy) measured by the ion chamber 16.91 29.2 39.09 56.03 Ratio of mean dosimeter dose to ion chamber dose 1.01 0.95 0.89 0.83 Table 10-10. HVLs used in flat-field scatter response. Aluminum filtration was added to a radiographic x-ray beam to match HVLs measured on the 320-slice scanner. Tube voltage Measured CT HVL (mm Al) Measured HVL of radiographic beam (mm Al) Difference 80 kV 3.90 3.82 2.1% 100 kV 4.97 5.04 -1.4% 120 kV 6.02 6.02 < 0.5% 0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% 100.0% 02004006008001000120014001600Percent Decrease from Inital DoseTime of White Light Exposure (min) Dosimeter Erasure

PAGE 161

161 Table 10-11. Response of the dosimeters to scatter and a flat x-ray field. Ratio of dose measured by the dosimeters to that measured by the ionization chamber Nominal tube voltage Thickness of acrylic (in) 80 kV 100 kV 120 kV 0 (surface) 1.05 0.95 0.92 1 1.04 0.96 0.90 2 1.01 0.94 0.91 3 1.01 0.96 0.90 4 0.98 0.94 0.89 5 1.01 0.94 0.91 6 1.01 0.94 0.93 Minimum 0.98 0.94 0.89 Maximum 1.05 0.96 0.93 Mean 1.02 0.95 0.91

PAGE 162

162 Table 10-12. Response of energy and scatter to th e CT x-ray beams used for clinical protocols. Ratio of dose measured by the dosimeters to that measured by the ionization chamber Thickness of acrylic (in) 80 kV M S 100 kV M S 120 kV M L 100 kV S S 120 kV S S 120 kV M S (64-slice) 0 (surface) 1.01 0.95 0.86 0.96 0.90 0.84 1 0.99 0.90 0.86 0.93 0.89 0.83 2 0.96 0.91 0.86 0.91 0.87 0.83 3 0.97 0.89 0.87 0.90 0.87 0.83 4 0.97 0.90 0.85 0.92 0.88 0.82 Minimum 0.96 0.89 0.85 0.90 0.87 0.82 Maximum 1.01 0.95 0.87 0.96 0.90 0.84 Mean 0.97 0.90 0.86 0.91 0.88 0.83 Note: Tube voltage is given at the top of each column, followed by the bowtie filter (M=medium, S=small) and focal spot size (S=small, L=large). Table 10-13. The calculated f-fact ors as a function of HVLs meas ured for all tube voltage and filter combinations of the clinical CT protocols. HVL (mm Al) Effective energy (keV) (en/ )tissue(cm2/g) (en/ )ai r (cm2/g) f-factor 3.87 37.2 9.68E-2 9.18E-2 1.05 4.65 40.5 7.07E-2 6.69E-2 1.06 4.84 41.3 6.85E-2 6.48E-2 1.06 4.88 41.5 6.80E-2 6.44E-2 1.06 5.80 45.2 5.74E-2 5.42E-2 1.06 5.91 45.6 5.61E-2 5.30E-2 1.06 6.09 49.5 4.50E-2 4.23E-2 1.06 6.84 46.4 5.40E-2 5.09E-2 1.06 6.85 49.4 4.52E-2 4.25E-2 1.06 7.16 52.5 4.09E-2 3.84E-2 1.06 7.53 50.8 4.27E-2 4.01E-2 1.07 7.92 53.1 4.02E-2 3.77E-2 1.07

PAGE 163

163 Figure 10-12. Response of the OSL dos imeters to increasing dose rates. R = 0.9972 0 10 20 30 40 50 60 70 80 0100200300400500600700Dose (mGy)Tube Current Time Product (mAs) Dose Rate Response

PAGE 164

164 Figure 10-13. Linearity of the en ergy-corrected dosimeter doses, as depicted with corresponding ion chamber measurements. 0 10 20 30 40 50 60 70 80 90 0200400600800Dose (mGy)Tube Current (mA) Linearity of Dosimeters and Ion Chamber Energy-Corrected Dosimeters 0.6 cc Ion Chamber

PAGE 165

165 Figure 10-14. Linearity res ponse of the OSL dosimeters to increasing tube currents, representative of the range of doses measured. R = 0.99980 200 400 600 800 1000 1200 0100020003000400050006000700080009000Dose (mGy)Tube Current -Time Product (mAs) Linearity Response

PAGE 166

166 Figure 10-15. Linearity response of the OSL dosimeters and ion chamber to increasing tube currents, representative of the range of doses measured. 0 200 400 600 800 1000 1200 1400 0200040006000800010000Dose (mGy)Tube Current (mA) Linearity of Dosimeters and Ion Chamber Energy-Corrected Dosimeters 0.6 cc Ion Chamber

PAGE 167

167 Figure 10-16. Dose absorbed by the nanoDot dosim eters, in air, as a function of x-ray tube angle. Error bars are displaye d as the 95% confidence interval. 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 0153045607590105120135150165180Dose (mGy)Angle (degree) nanoDot Angular Response In Air

PAGE 168

168 Figure 10-17. Dose absorbed by the nanoDots in th e head phantom, displayed as a function of xray tube angle and with 95% confidence intervals. Table 10-14 Mean doses with a 95% confidence in terval measured by the dosimeters in and out of their plastic cases in two different positions within the CTDI phantom. Position Dosimeter in case Dosimeter out of case 12 oclock 10.97 (8.73,13.21) mGy 10.78 (9.79,11.77) mGy Center 8.54 (7.88,9.20) mGy 8.28 (7.64,8.92) mGy Table 10-15. Mean organ doses for pediatric craniosynostosis protocol. Mean organ doses (mGy) Scanner Tube voltage Eff. mAs No. of Scans Skin Lens of eye Thyroid Breast 320-slice 120 kV 121 1 18.98 22.14 1.95 0.81 5 18.95 23.24 1.85 0.84 320-slice 100 kV 121 1 12.02 13.85 1.15 0.48 5 12.85 15.18 1.17 0.50 64-slice 120 kV 157 1 22.97 25.63 6.79 1.55 5 24.07 25.54 7.39 1.56 25.0 30.0 35.0 40.0 45.0 50.0 55.0 60.0 65.0 0306090120150180210240270300330360Average Dose (mGy)Angle (degree) nanoDot Angular Response with Body Phantom

PAGE 169

169 Table 10-16. Organ doses measured using the pe diatric phantom and craniosynostosis protocol; mean and maximum report with 95% confidence interval. Scanner Organ Doses (mGy) Tube voltage Skin Lens of eye Thyroid Breast 320-slice Mean 18.98 22.14 1.95 0.81 120 kV Maximum 22.57 24.08 2.13 1.30 95% CI (15.62,22.33) (19.73,24.55) (1.67,2.23) (0.51,1.11) 320-slice Mean 12.02 13.85 1.15 0.48 100 kV Maximum 14.21 15.33 1.22 1.14 95% CI (10.28,13.76) (12.53,15.17) (1.01,1.30) (0.19,0.78) 64-slice Mean 22.97 25.63 6.79 1.55 120 kV Maximum 28.24 32.40 8.30 2.01 95% CI (18.10,27.84) (13.31,37.95) (4.84,8.73) (1.02,2.08) Figure 10-18. Image quality phantom images for pediatric protocol image using, A) 120 kV and, B) 100 kV. All other parameters remained the same.

PAGE 170

170 Table 10-17. Number of low-c ontrast objects viewed using pediatric craniosynostosis protocol parameters. Number of observed low-contrast objects Viewer Energy (kV) A B C D E F Total 1 120 6 3 3 3 3 2 20 100 5 2 0 2 1 0 10 2 120 6 4 0 2 3 1 16 100 5 4 1 3 3 0 16 3 120 6 2 0 3 3 1 15 100 5 4 1 3 3 0 16 A-C D-F Total Mean 120 10 7 17 100 9 5 14

PAGE 171

171 171 Table 10-18. Mean organ doses result ing from brain pe rfusion protocol. Protocol Tube voltage (kV) No. of scans Mean organ doses (mGy) Esophagus Thyroid Brain Skin Breast Eyes Manufacturer 80 5 2.47 3.90 168.96 225.84 1.09 263.88 1 3.02 3.63 189.72 250.64 1.21 286.18 Manufacturer 100 5 5.22 7.96 325.71 392.39 2.23 459.18 1 6.85 7.81 349.89 407.94 2.50 465.88 Manufacturer 120 5 9.02 13.28 530.80 569.76 3.46 660.09 1 12.30 13.22 588.91 650.98 3.93 723.84 Continuous 80 5 3.46 5.02 193.56 233.32 1.41 263.54 1 3.57 4.20 203.63 250.06 1.54 296.42 mA boost 80 5 3.43 5.14 191.76 241.37 1.42 273.08 1 3.59 4.30 196.81 266.24 1.57 293.98 64-slice 120 5 5.02 7.04 419.59 1175.70 2.24 198.15 1 4.95 6.95 457.15 1240.97 2.22 205.33 Note: Mean organ doses measured with a single scan are reported on the first line. Mean organ doses measured using five consec utive scans are reported on second line.

PAGE 172

172 172 Table 10-19. Mean and maximum organ doses in mG y measured with adult brain perfusion protocol. Protocol Tube voltage Organ Esophagus Thyroid Brain Skin Breast Eyes Manufacturer Mean 2.47 3.90 168.96 225.84 1.09 263.88 80 kV Maximum 3.42 4.09 191.14 253.24 1.24 284.58 95% CI (1.17,3.78) (3.61,4.20) ( 144.70,193,22) (194.49,257.19) (0.83,1.35) (229.64,298.12) Manufacturer Mean 5.22 7.96 325.71 392.39 2.23 459.18 100 kV Maximum 7.25 8.27 351.31 410.54 2.52 464.28 95% CI (2.09,8.35) (7.47,8.46) ( 285.97,365.44) (371.23,413.56) (1.72,2.74) (453.11,465.25) Manufacturer Mean 9.02 13.28 530.80 569.76 3.46 660.09 120 kV Maximum 12.71 13.68 590.33 653.59 3.95 722.25 95% CI (2.90,15.13)(12.62,13.94)(444.43,617.17) (465.12,674.41) (2.67,4.25) (558.78,761.41) Continuous Mean 3.46 5.02 193.56 233.32 1.41 263.54 80 kV Maximum 4.87 5.19 208.05 254.14 1.62 290.46 95% CI (2.18,4.74) (4.79,5.24) ( 180.08,207.03) (218,248.25) (1.19,1.63) (238.31,288.77) mA boost Mean 3.43 5.14 191.76 241.37 1.42 273.08 80 kV Maximum 4.89 5.29 201.23 270.32 1.65 288.02 95% CI (2.13,4.73) (4.98,5.30) ( 1.83,199.87) (217.48,265.27) (1.18,1.66) (256.10,290.06) 64-slice Mean 5.02 7.04 419.59 1175.70 2.24 198.15 120 kV Maximum 6.74 7.27 529.61 1225.80 2.39 203.66 95% CI (2.01,8.04) (6.42,7.65) ( 305.47,533.70) (1112.88,1238.52) (1.92,2.55) (187.61,208.69)

PAGE 173

173 173 Table 10-20. Mean organ doses in mGy resulting from adult cardiac protocol. Protocol No. of Scans Organs Thyroid Lung Stomach Breast Skin Prospectively-gated CTA 5 2.50 13.94 2.08 20.10 21.29 1 2.55 15.22 2.09 25.18 27.09 Functional analysis 5 7.81 43.77 6.25 67.72 69.98 1 8.16 46.09 6.28 68.52 71.49 Functional Analysis with dos e modulation 5 4.61 26.07 3.73 40.03 41.84 1 4.76 27.82 3.79 41.56 43.95 Cardiac CTA (64-slice) 5 5.93 52.03 55.95 100.11 97.24 1 5.81 52.46 62.91 97.60 32.40

PAGE 174

174 174 Table 10-21. Mean and maximum organ doses in mGy resulting from adult cardiac protocol. Protocol Organs Thyroid Lung Stomach Breast Skin Prospectively-gated Mean 2.50 13.94 2.08 20.10 21.29 Maximum 2.61 20.62 2.78 24.75 25.05 95% CV (2.26,2.61) (7.50,20.38) (1.15,3.01) (15.81,24.39) (16.30,26.29) Functional analysis Mean 7.81 43.77 6.25 67.72 69.98 Maximum 8.54 61.76 8.66 80.60 81.34 95% CV (6.38,9.25) (24.07,63.47) (3.29,9.21) (55.53,79.90) (54.99,84.96) Functional Analysis Mean 4.61 26.07 3.73 40.03 41.84 with dose modulation Maximum 4.74 36.06 4.87 52.97 48.25 95% CV (4.23,4.98) (14.31,37.83) (1.90,5.56) (29.79,50.27) (33.52,50.15) 64-slice scanner Mean 5.93 52.03 55.95 100.11 97.24 Maximum 6.16 105.08 86.21 139.37 124.99 95% CV (5.31,6.54) (0,113.81) (18.54,93.36) (58.37,141.85) (58.04,136.44)

PAGE 175

175 Figure 10-19. Image quality phantom images fo r cardiac protocol. A) imaged using 0.5 mm slices, B) using 1.0 mm slices, and C) 3.0 mm slices. All other parameters remained the same. Table 10-22. Organ dose comparison for the 320-slice CT pediatric craniosynostosis protocol between measured doses (mGy) and Monte Carlo (MC) simula ted doses (mGy). 120 kV 100 kV OSL MC Rel. Diff. (%) OSL MC Rel. Diff. (%) max mean max mean Skin 22.57 18.98 19.96 -4.91 14.21 12.02 13.11 -8.31 Lenses 24.08 22.14 21.36 3.54 15.33 13.85 13.62 1.69 Thyroid 2.13 1.95 2.20 -11.36 1.22 1.15 1.28 -10.16

PAGE 176

176 Table 10-23. Organ dose comparison for the 320-slice CT adult brai n protocols between measured doses (mGy) and Monte Carlo (MC) simulated doses (mGy). Organ Protocol EsophagusaThyroid b Brainc Skin d EyeseManufacturer 80 kVp OSL 2.47 3.90 168.96 225.84 263.96 MC 1.56 4.50 221.97 337.40 247.64 Rel. Diff. (%)58.33 -13.33 -23.88 -33.06 6.53 Manufacturer 100 kVp OSL 5.22 7.96 325.71 392.40 459.26 MC 3.56 9.72 420.92 501.27 428.61 Rel. Diff. (%)46.63 -18.11 -22.62 -21.72 7.13 Manufacturer 120 kVp OSL 9.02 13.28 530.80 569.77 697.53 MC 6.39 17.02 678.97 710.98 675.52 Rel. Diff. (%)41.16 -21.97 -21.82 -19.86 3.28 Continuous OSL 3.46 5.01 193.56 233.33 269.74 MC 2.29 6.40 260.18 273.12 264.43 Rel. Diff. (%)51.09 -21.56 -25.61 -14.57 -0.34 mA Boost OSL 3.43 5.14 191.76 241.38 273.24 MC 2.29 6.42 261.28 274.42 265.60 Rel. Diff. (%)49.78 -19.94 -26.61 -12.04 2.82 along organ outside of primary beam; bsmall organ outside primary beam; clarge organ inside primary beam; donly volume of skin inside the primary beam was included in simulation; esmall organ inside primary beam Table 10-24. Organ dose comparison for the 320-s lice CT adult cardiac CTA protocols between measured doses (mGy) and Monte Carlo (MC) simulated doses (mGy). Organ Dose (mGy) Protocol ThyroidaLung b StomachcSkin d Prospectively-gated CTA OSL 2.50 13.94 2.08 21.29 MC 2.76 15.21 2.97 17.34 Rel. diff. (%) -9.42 -8.35 -29.97 22.78 Functional Analysis (CFA) OSL 7.81 43.77 6.25 69.98 MC 8.12 44.76 8.75 51.02 Rel. diff. (%) -3.82 -2.21 -28.57 32.16 CFA with Dose Modulation OSL 4.61 26.07 3.73 41.84 MC 4.88 26.90 5.26 30.66 Rel. diff. (%) -5.53 -3.00 -29.00 36.46 asmall organ outside primary beam; blarge organ inside primary beam; clarge organ at the edge of the primary beam; donly volume of skin inside the primar y beam was included in simulation

PAGE 177

177 CHAPTER 11 SUMMARY AND CONCLUSIONS 11.1 Summary of This Research Project There were several purposes of this research project, each of which was com pleted with a methodical process and in the most efficient way possible. A standardized set of CT protocols was created by the RPC, and volumetric protocol s were developed for the wide-beam technology of the 320-slice scanner. Scan parameters we re chosen by the radiologists on the committee, who also assessed the image quality of these protocols in a clinical setting. Medi cal physicists on the committee were successful in assessing dose considerations of the chosen scan parameters. With the CT prot ocols standardized a nd established at Shands at UF, an OSL dosimetry system was verified for its use in clini cal CT dosimetry. Several characteristics of the system were described, including the ability to erase the dose on the dosimeters using exposure to room light, the ability to re use the dosimeters; energy, scatter, linearity and angular responses; as well as the ability to take the OSL material out of its case for select dose measurements. Because of the advancements in CT technol ogy and the wide beam width of the 320-slice scanner used in this research pr oject, several characteristics of the scanner and x-ray beams were described, including reproducibility, tube voltage accuracy, HVL, beam width and dose profiles. With both the dosimetry system and CT scanne r characterized, organ dose measurements were made for three clinical protocols that were a pproved and implemented by the RPC. Physical, tomographic phantoms were used to measure organ dos es in select organs th at were in the direct radiation field, or close enough to it to receiv e a significant amount of scatter. Organ dose measurements were made on two separate CT systems for comparison: a wide-beam 320-slice scanner and a 64-slice MDCT scanner. Fina lly, image quality was assessed for the three protocols. This analysis was done by a neur oradiologist on the RPC during the clinical

PAGE 178

178 development of the brain perfusion protocol. An image quality phantom was used to assess the low-contrast resolution of the pediatric head a dult cardiac protocols, and analysis was performed by medical physicists. The combination of dosimetry system, tom ographic phantoms and 320-slice volumetric CT scanner makes this research project unique to any other. There has very little work to date published with commercial OSL dosimeters in the diagnostic energy range; a complete evaluation of the system for use in CT was comple ted through the work of this research project and verified for use. A method is described th at details measurements needed to correct for energy and scatter for dose measurements made on other CT systems. The small dosimeters allowed for numerous point organ doses to be made at one time, thus streamlining the measurement process and overcoming some of the challenges to established dose metrics in CT. Because organ doses were measured directly, s cattered radiation was captured by the dosimeters and therefore representative of true organ dose measurements. While physical, tomographic phantoms have been used by others at UF, they have not been used with these OSL dosimeters in a wide-beam MDCT scanner. Shands at UF is one of the few hospitals in the world with the Aquilion ONE 320-slice CT system. Dose measurements resulting from the 320-slice scanne r that have been reported to da te have been in the form of effective dose and the CTDI values displayed by the scanner, and not actual organ doses. Furthermore, comparisons were made between cl inical volumetric and helical protocols, and a dose reduction was reported in most cases us ing the volumetric scanner for the protocols evaluated. Specifically, doses measured on th e Aquilion One were lower than the 64-slice scanner for all organs evaluated with the pediatric craniosynost osis protocol. The skin dose measured with the brain perfusion protocol was at least two times higher on the 64-slice scanner

PAGE 179

179 compared to the Aquilion One, and this measured skin dose approaches the threshold for radiation-induced skin effects. With the ex ception of the thyroid, al l organ doses compared using the adult cardiac CTA prot ocol were higher on the 64-slice scanner than on the Aquilion One. The clinical impact of these findings is tremendous, especially in the case of pediatric patients. A recommendation to scan pediatri c patients with the 320-sl ice scanner will save radiation dose to the patient, in a ccordance with the ALARA principle. 11.2 Future Work The precisio n of the OSL dosimeters has been de scribed in this dissertation; however, their accuracy must be benchmarked against a known and proven radiation dose-measuring instrument. The accepted benchmark is a smallvolume Farmer chamber, or any small-volume ionization chamber that is calibrated in the energy range for which it will be used. To satisfy these conditions to benchmark the dosimeters, the 0.6-cc ion chamber was calibrated at 100 kV for use in this research project and was the init ial approach to quantifyi ng the energy and scatter responses of the dosimeters; future work shoul d include an investiga tion into the energy and scatter characteristics of the 0.6-cc chamber to conduct a better assessment of the accuracy of the dosimeters. If organ doses were to be measured in the future using the methodology presented in this dissertation on a CT scanner other th an the two specific scanners used in this research project, a minimal number of measurements will need to be made, especially the measurements detailed in Sections 8.2.5 through 8.2.6 to correct the non-uniform response of the dosimeters to energy and scatter. While the method recommended by the au thor is the use, for these measurements, the exact scanner to be used for organ dose measurem ents, the flat-field measurements described in Section 8.2.5 could also be used if a service mode in the CT s canner was not available to allow stationary CT exposures. If the flat-field method is followed, the second HVL should be

PAGE 180

180 measured after filtration has been added to th e radiographic tube in order to compare the homogeneity coefficients of the radiographic tube with added filtration to those of the CT x-ray tube, at each tube voltage and bowtie filter combination to be used for organ dose measurements. This comparison would require, of course, the additional measurements of the second HVL in the CT scanner, but would provide a better as sessment and comparison of the two beams. Furthermore, future work should include measuring the HVLs of the CT scanner across the gantry at the same tube voltage and bowtie filter combination to analyze any effect of the shape of the bowtie filter on the dosimeter energy re sponse. Further investigation into the characteristics of the bowtie filters could include analysis using a different method to measure total filtration. The instrument used in this re search project may respond differently to different angle of x-ray beam incidence, meaning that the methodology followed was to move the instrument across the CT scan gantry thereby ma king the initial measurement at isocenter when the beam is perpendicular to the instrument (0); as measurements were made at increasing distance away from isocenter, the angle of incide nce also increased and could have affected the measurements of total filtration. Future work that could be continued as an ex tension of this research project could include measuring organ doses for other clinical protoc ols. The RPC has discussed linking a range of doses to each protocol. This information w ould be available with all other information pertaining to the CT scan. While organ doses should not be the prime reason for ordering a particular diagnostic study, they should be a co nsideration. A dose rang e could help a physician choose between a CT and MR protocol, assuming both provide the necessary information for proper diagnosis. Furthermore, a dose range for each protocol would lay the groundwork for tracking the total dose ac cumulated by a patient resulting from medical exposures. If the dose

PAGE 181

181 information for a CT scan is available, it coul d be put into the patien ts medical record and followed over time. A female tomographic phantom is currently in development at UF. Where the average female size is smaller than that of a man, organ do ses are expected to be different. Similarly, the female phantom has breast tissue and structure, wh ich would allow for more realistic breast dose measurements. Similarly, as the family of phys ical, tomographic phantoms developed by the UF group continues to grow, there will be more phant oms with which to make these types of organ dose measurements. The issue of using small dosimeters to obtain an average organ dose was addressed in 10.3. An issue with all dose measurements is the number of (approximate) point dose measurements needed to calculate an average dose accurately. Theoretically, an infinite number of point doses is needed for a true average dose. However, doi ng so is not realistic in any. Now that a method has been established to reliably measure organ doses with the dosimeter s and physical phantoms, future work in this area could include determin ing the minimum number of dosimeters needed to measure an accurate average dose. Finally, this research project focused on meas uring dose to organs that had a high degree of sensitivity to radiation and were located within or very close to the primary radiation field. However, as evidenced by the high stomach doses measured for the cardiac CTA protocol on the 64-slice scanner, a significant amou nt of radiation may be added to organs adjacent to the scan volume for helical protocols due to scan overranging. Where this is a consequence of the helical rotations of the x-ray tube around the phantom re sulting in additional dose of two half-rotations, there is a possibility to measure even larger reductions in dose to th ese organs outside the

PAGE 182

182 boundaries of the scan by using the wide-beam volumetric scanner that does not require these two extra half-rotations for interpolation. 11.3 Final Words While the medical inform ation acquired from CT scanning is a benefit that normally outweighs the risk associated with radiation, it is important to follow the ALARA (as low as reasonably achievable) principle and keep patient doses as small as possible. The 320-slice volumetric scanner is a perfect example of th e rapid increase in CT technology. Though not addressed in this research project, the comp eting technologies of dua l-source and dual-energy CT scanners are also available and should be included in the discussion. With the concerted effort of the RPC at Shands at UF and the ongoing research of its medical physicists, the safety of patients is a prior ity and one that has been achieved through this research project.

PAGE 183

183 APPENDIX A EXAMPLES OF IMPLEMENTED CT PROTOCOLS CT protocols were com pleted for radiology sub-specialty areas including body, cardiac, neurological, musculoskeletal (MSK )/orthopedic, and ear, nose and throat (ENT). The complete list of protocols can be found on the UF College of Medicine website.10 An example of a body protocol is the Chest/A bdomen/Pelvis protocol displayed in Figure A-1. As with most of the body protocols, it requ ires the use of a contrast agent and thus, the contrast or phase delay is noted in the acquisition table. A phase delay is sometimes necessary to wait for the contrast agent to travel through the va scular system and reach the area of interest of the study at the time of the scan acquisition. A ca rdiac protocol is illust rated in Figure A-2. The cardiac protocols use a diagram that is centered on the heart and have very detailed instructions in the Other section regarding wh en to begin injection of the c ontrast agent, as well as the indicated use of manuf acturer-specific software-based reconstructions which differ by manufacturer, making it necessary to include rec onstruction and reformation instructions for both manufacturer-independent and manu facturer-specific protocols. An ENT protocol is shown in Figure A-3. The Reconstruction section displays an example of the illustrations and FOVs drawn on the protocol to improve visualization of the areas of clinical in terest. Another diagram is used to depict the regions for the coronal reformation image set. Because the area of interest is solely the temporal bones, the coronal reformatted images do not need to go through the entire head; this specification of the reformatted image set is a clear example of implementing limits on the number of reformatted images for storage considerations in the PACS archive. A neuroradiology protocol is depicted in Figure A-4. The need for ve ry thin image slices is typical in neuroradiology studies and indicated in th e B1 reconstruction thickness of 0.5.75mm. A range of values for this parameter is given, ra ther than a single value because of potential

PAGE 184

184 differences among slice thickness ca pabilities of CT scanners. Also of interest are the two separate image sets, one pre-contrast and the ot her post-contrast, generated for comparison. All three image planes are utilized, depi cted in the lines in the illustra tions as well as specified in the tables. Finally, an MSK protocol is shown in Figure A-5. In the case of MSK, there are many instances when the same protocol is applied to the body area of interest. The committee chose to keep the protocols as generic as possible in this case. Therefore, no illustration is used in the Acquisition section for many of these protocols, and the start and end points of the image acquisition are also not specified. Both the Reco nstruction and Reformation tables also indicate the FOV according to the anatomy of interest. By not specifying anatomy, the protocol is flexible in order to avoid gene rating a large number of joint-specific protocols with the same acquisition and recons truction parameters.

PAGE 185

185 Figure A-1. Example of a body protocol. A phase de lay of 50-60 seconds if utilized so that the administered contrast has time to travel through the blood stream to areas of the pelvis.

PAGE 186

186 Figure A-2. Example of a cardiac protocol. These images are ge nerated in different ways on different CT scanners and thus two sets of reconstruction tables are used in this protocol.

PAGE 187

187 Figure A-2. Continued

PAGE 188

188 Figure A-3. An example of an EN T protocol. Two small circles in the locations of the right and left temporal bone denote specific areas for the reconstructed coronal images. Horizontal lines show the plane and speci fic area of reformatted coronal images.

PAGE 189

189 Figure A-4. Example of a neuroradiology prot ocol, depicting the Ac quisition, Reconstruction, Reformation and Indications sections.

PAGE 190

190 Figure A-4. Continued

PAGE 191

191 Figure A-5. Example of an MSK protocol.

PAGE 192

192 APPENDIX B DOSIMETRIC PHANTOMS Phantom Construction While the construction m ethodology and details of the specific phantoms used in this research project are yet to be published, th e methodology was the same as that published by Jones in 2003.39 First, image segmentation was used to de fine the contours of the organs within the data sets; second, tissue-equivalent material s were developed and third, automated methods were used in the physical c onstruction of the phantoms. Image Segmentation For the purp ose of accurate and reproducible se gmentation, and to expedite the process, an automatic segmentation method wa s utilized by the group at UF.36 The goal of this process was to define the contours of each organ displaye d within a CT slice. Various CT number thresholding methods were utilized for this purpose. Once a cont our has been defined, all voxels within that contour are assigne d a tag value. Each organ ha s a unique tag value, thereby differentiating between as ma ny organs as are segmented. Tissue-Equivalent Materials Three tissue-equivalent com positions were de veloped for the purpose of constructing these tomographic phantoms: bone (or skeletal) tissue, soft tissue and lung tissue. The three tissueequivalent materials were matched to three basic physics characteristics: density, mass attenuation coefficient, and ma ss energy absorption coefficient. An epoxy resin-based system was used to manufacture the main matrix of the tissue equivalent materials using phenolic microspheres as needed to adjust the mass density. The reference values used for these tissues of interest were those published by Cristi and Eckerman, in association with the Oak Ridge National Laboratory (ORNL).68

PAGE 193

193 In his paper, Jones argued that there was no need for developing a larger, more specific set of tissue equivalent materials because of the inherent error associated with the dosimetry system he used for organ dose measurements. Thus, in a ddition to the fact that phantom construction is outside the scope of this project, the dosimetry system used for th is research project is different, but sources of error exist in it that are greater than the small differences in tissue composition of all of the segmented organs making use of these phantoms adequate for this research project. Phantom Specifications Adult Male Phantom Two of the UF tom ographic phantom s were used in this research project: an adult male and a pediatric male. The adult male was built using data from work by Lee et al .69 The computational model known as KTMAN-2 was ba sed on whole-body CT images of a 35-year old Korean male whose height and weight closely matched those of the average Korean man. The CT image set was acquired wi th a pixel reso lution of 2x2x5 mm3. The CT scan was performed with contrast to prov ide images with high soft-tissue resolution, and the images were reconstructed at 1-mm thick slices. Work was further done at UF by Ambrose70 to transform the data from KTMAN-2 into a corresponding American counterpart; organ masses were scaled to those of the American reference male (approximately 176 cm tall and a weight of 72.9 kg). Known as gatorman, the data from this tomographic phantom were the basis for the physical phantom used in this research project. The phantom is composed of the same three tissue-equivalent materials as discussed and developed by Jones, and is pictured in Figure 6-1. Pediatric Phantom The pediatric phantom was constructed base d on a 9-month old computational phantom built as part of the UF series.37 Data used for the phantom were obtained from CT scans of a 9-

PAGE 194

194 month old male patient: one of th e chest, abdomen, pelvis and a second of the head. The image segmentation process previously described was followed to assign one of three tissue-equivalent materials to each segmented organ. While the co nstruction process of this 9-month phantom is yet to be published,65 the methodology is similar to that described by Jones and Simon.71 This phantom is depicted in Figure 6-2. While stylized models have been used with success in the past, the UF series of tomographic, physical phantoms is far superior to its predecessors for the purpose of organ dose measurements. The use of CT data sets for image segmentation gives these phantoms the most realistic and detailed organ structure, as well as accurate organ position within the phantom. Tissue-equivalent materials re presenting different human tissu e further substantiate the usefulness of the UF phantoms for organ dose estimates.

PAGE 195

195 LIST OF REFERENCES 1F. A. Mettler, P. W. Wiest, Jr., J. A. Locken, and C. A. Kelsey, "CT scanning: patterns of use and dose," J. Radiol. Prot. 20, 353-359 (2000). 2D. Brenner, C. Elliston, E. Hall, and W. Ber don, "Estimated risks of radiation-induced fatal cancer from pediatric CT," Am. J. Roentgenol. 176 289-296 (2001). 3National Council on Radiation Prot ection and Measurements, "Ion izing Radiation Exposure of the Population of the United States," Report 160 (2009). 4Conference of Radiation Cont rol Program Director, Inc., "N ationwide evaluation of x-ray trends (NEXT) 2000-2001 survey of patient ra diation exposure from computed tomography examinations in the United St ates," Report E-07-2 (2007). 5D. Brenner, "Estimating cancer risks from pediatric CT: going from the qualitative to the quantitative," Pediatr. Radiol. 32, 228-330 (2002). 6J. T. Bushberg, J. A. Seibert, E. M. Leidholdt, Jr., and J. M. Boone, The Essential Physics of Medical Imaging 2nd ed. (Lippencott Williams & Wilkins, Philadelphia, 2002). 7 C. P. Langlotz, "RadLex: A new method for indexing online educational materials," Radiographics. 26, 1595-1597 (2002). 8 American College of Radiology, Index for Radiological Diagnoses 4th ed. (American College of Radiology, Reston, 1992). 9T. M. Buzug, Computed Tomography: From Photon Stat istics to Modern Cone-Beam CT, 1st ed. (Springer, Verlag, 2009). 10Protocols [Internet]. Gainesville: University of Florida, College of Medicine (US); [updated 2009 Jun 24; cited 2009 Jun 24]. Available from: http://xray.ufl.edu/patient-care/protocols 11 C. H. McCollough et al ., AAPM Task Group 23: CT Dosimetry, American Association of Physicists in Medicine. Task Group Report 96, 2008.. 12Aquilion ONE: The Quantum Advantage [Inter net]. Tustin: Toshiba America Medical Systems (US); [cited 2009 Jun 23]. Available from: www.medical.toshiba.com 13E. Siebert, G. Bohner, M. Dewey,F. Masuhr, K. T. Hoffmann, J. Mews, F. Engelken, H. C. Bauknecht, S. Diekmann, and R. Klingebiel, "3 20-slice CT neuroimaging: initial clinical experience and image quality evaluation," Br. J. Radiol. 82, 561-570 (2009).

PAGE 196

196 14Neuro and Acute Stroke Imaging with Dynami c Volume CT [Internet]. Tustin: Toshiba America Medical Systems (US); [c ited 2009 Jun 23]. Available from: http://www.healthcaretechguide.com/index.php?option=com_resource&cont roller=whitepaper &task=inquire&id=207&Itemid=2 15P. Rogalla, B. Stover, I. Scheer, R. Juran, G. Gaedicke, and B. Hamm,"Low-dose spiral CT: applicability to paediatric ch est imaging,"Pediatr. Radiol., 29, 565-569 (1999). 16M. K. Kalra, M. M. Maher, T. L. Toth, B. Schmidt, B. L. Westerman, H. T. Morgan, and S. Saini, "Techniques and applications of automa tic tube current modula tion for CT," Radiology, 233, 649-657 (2004). 17P. Vock, "CT dose reduction in children," Eur. Radiol. 15, 2330-2340 (2005). 18D. D. Cody, D. M. Moxley, K. T. Krugh, J. C. O'Daniel, L. K. Wagner, and F. Eftekhari,"Strategies for formul ating appropriate MDCT techni ques when imaging the chest, abdomen, and pelvis in pediatri c patients," Am. J. Roentgenol. 182, 849-859 (2004). 19National Council on Radiation Protection and Meas urements, "Limits of Exposure to Ionizing Radiation," Report 116, (1993). 20International Commission on Radiological Protection, "The 2007 Recommendations of the International Commission on Radiologi cal Protection," Report 103 (2007). 21International Commission on Radiological Prot ection,"Conversion coefficients for use in radiological protection against external radiation. Adopted by the ICRP and ICRU in September 1995," Annals of the ICRP, 26, 1-205 (1996). 22C. H. McCollough and B. A. Schueler, "C alculation of effective dose," Med. Phys. 27, 828-837 (2008). 23G. Drexler, W. Panzer, N. Petoussi, and M. Zankl, "Effective dose how effective for patients?," Radiat Environ. Biophys. 32, 209-219 (1993). 24T. B. Shope, R. M. Gagne, and G. C. Johnson, "A method for describing the doses delivered by transmission x-ray computed tomography, Med. Phys. 8, 488-495 (1981). 25Food and Drug Administration, "FDA public health notification: reducing radiation risk from computed tomography for pediatric and sm all adult patients," Pediatr. Radiol. 32, 314-316 (2002). 26W. Leitz, B. Axelsson, and G. Szendro, "Computed tomography dose assessment: a practical approach," Radiat Prot. Dosim. 57, 377-380 (1995). 27A. Suzuki and M. N. Suzuki, "Use of a penc il-shaped ionization chamber for measurement of exposure resulting from a computed tomography scan," Med. Phys. 5, 536-539 (1978).

PAGE 197

197 28 R. L. Dixon, "A new look at CT dos e measurement: beyond CTDI," Med. Phys. 30, 1272-1280 (2003). 29J. M. Boone, "The trouble with CTD100," Med. Phys. 34, 1364-1371 (2007). 30S. Mori, M. Endo, K. Nishizawa, T. Tsunoo, T. Aoyama, H. Fujiwara, and K. Murase, "Enlarged longitudinal dose profiles in cone-bea m CT and the need for modified dosimetry," Med. Phys. 32, 1061-1069 (2005). 31R. L. Dixon and A. C. Ballard, "Experimental validation of a versatile system of CT dosimetry using a conventional ion chambe r: beyond CTDI100," Med. Phys. 34, 3399-3413 (2007). 32 G.F.Knoll, Radiation Detection and Measurement 3rd ed. (John Wiley & Son, Hoboken, 2000). 33Radcal Corporation, "Model 9010 Radiati on Monitor Controlle r Manual," (2009). 34National Institute of Standards and Technology, "Ionizing radiation measurements: dosimetry of x ray, gamma rays, and elec trons," (2009). Available from: http://ts.nist.gov/MeasurementSer vices/Calibrations/x-gamma-ray.cfm 35 RTI Electronics, Inc., "Selection guide," (2009). Available at: http://www.rti.se/cuda/index.htm. 36J. C. Nipper, J. L. Williams, and W. E. Bolc h, "Creation of two tomographic voxel models of paediatric patients in the first year of life," Phys. Med. Biol. 47 3143-3164 (2002). 37C. Lee, J. L. Williams, C. Lee, and W. E. Bolc h, "The UF series of tomographic computational phantoms of pediatric patients," Med. Phys. 32, 3537-3548 (2005). 38C. Lee, C. Lee, R.J. Staton, D.E. Hintenlang, M.M. Arreola, J.L. Williams, and W.E. Bolch, "Organ and effective doses in pediatric pa tients undergoing helical multislice computed tomography examination," 34, 1858-1873 (2007). 39A. K. Jones, T. A. Simon, W. E. Bolch, M. M. Holman, and D. E. Hintenlang, "Tomographic physical phantom of the newborn child with read-time dosimetry I. Methods and techniques for construction," Med. Phys. 33, 3274-3282 (2006). 40L.N. Rill, L. Brateman, and M. Arreola,"Evalu ating radiographic parameters for mobile chest computed radiography: phantoms, image quality and effective dose," Med. Phys. 30, 27272735 (2003). 41S.R. Prasad, C. Wittram, J.A. Shepard, T. McLoud, and J. Rhea,"Standard-dose and 50%reduced-dose chest CT: comparing the e ffect on image quality,"Am.J.Roentgenol. 179 461465 (2002). 42C.H. McCollough and F.E. Zink,"Performance ev aluation of a multi-slic e CT system," Med. Phys. 26, 2223-2230 (1999).

PAGE 198

198 43Catphan 500 & 600 brochure [Internet]. Salem: The Phantom Laboratory (US); [cited Jun 24 2009]. Available from: http://www.phantomlab.com/catphan.html 44M. S. Akselrod, L. Botter-Jensen, and S. W. S. McKeever, "Optically st imulated luminescence and its use in medical dosimetry," Radiat. Meas. 41, S78-S99 (2007). 45P. A. Jursinic, "Characterization of optically stimulated luminescent dosimeters, OSLDs, for clinical dosimetric measurements," Med. Phys. 34, 4594-4604 (2007). 46E. G. Yukihara, E. M. Yoshimura, T. D. Lindstrom, S. Ahmad, K. K. Taylor, and G. Mardirossian, "High-precision dosimetry for radiotherapy using the optically stimulated luminescence technique and thin Al 2O3:C dosimeters," Phys. Med. Biol. 50, 5619-5628 (2005). 47S. W. S. McKeever, M. S. Akselrod, L. E. Colyott, N. A. Larsen, J. C. Polf, and V. Whitley, "Characterisation of Al2O3 for use in ther mally and optically stimulated luminscence dosimetry," Radiat. Prot. Dosim. 84, 163-168 (1999). 48A. Viamonte, L.A. da Rosa, L.A. Buckley, A. Cherpak, and J.E. Cygler, "Radiotherapy dosimetry using a commercial OSL system," Med. Phys. 35, 1261-1266 (2008). 49V. Schembri, and B.J. Heijmen, "Optically s timulated luminescence (OSL) of carbon-doped aluminum oxide (Al2O3:C) for film dos imetry in radiotherapy," Med. Phys. 34, 2113-2118 (2007). 50B. G. Markey, L. E. Colyott, and S. W. S. McKeever, "Time-resolved optically stimulated luminescence from alpha-Al2O3:C," Radiat. Meas. 24, 457-463 (1994). 51 American National Standards Institute, Inc. ,"American National Standard for Dosimetry Personnel Dosimetry Performance Criteria for Testing," Report N13.11-2009, (2009). 52Landauer, Inc., "microStar User Manual," (2008). 53International Electrotechnical Commission, "Medic al diagnostic X-ray equipment radiation conditions for use in the determination of characteristics," IEC Report 1267 (1994). 54 International Electrotechnical Commission, "Medical diagnostic X-ray equipment radiation conditions for use in the determination of characteristics," IEC Report 61267 (2005). 55 R. I. Aviv, E. Rodger, and C. M. Hall, "Craniosynostosis," Clin. Radiol. 57 93-102 (2002). 56N. Hirai, J. Horiguchi, C. Fujioka, M. Kiguchi, H. Yamamoto, N. Matsuura, T. Kitagawa, H. Teragawa, N. Kohno, and K. Ito, "Prospective ve rsus retrospective ECG-gated 64-detector coronary CT angiography: assessment of im age quality, stenosis, and radiation dose," Radiology. 248, 424-430 (2008).

PAGE 199

199 57 F. J. Rybicki, H. J. Otero, M. L. Steigner, F. Vorobiof, L. Nallamshetty, D. Mitsouras, H. Ersoy, R. T. Mather, P. F. Judy, T. Cai, K. Coyne r, K. Schultz, A. G. Whitmore, and M. F. Di Carli, "Initial evaluation of coronary images from 320-detector row computed tomography," Int. J. Cardiovasc. Imaging. 24, 535-546 (2008). 58K. Kitagawa, A. C. Lardo, J. A. Lima, and R. T. George, "Prospective ECG-gated 320 row detector computed tomography: implications for CT angiography and perfusion imaging," Int. J. Cariovasc. Imaging. Published online 2-18-2009, available from: http://www.springerlink.com/content/n05324735g457720/ 59M. L. Steigner, H. J. Otero, T. Cai, D. Mits ouras, L. Nallamshetty, A. G. Whitmore, H. Ersoy, N. A. Levit, M. F. Di Carli, and F. J. Rybicki, "Narrowing the phase window width in prospectively ECG-gated single h eart beat 320-detector row corona ry CT angiography," Int. J. Cardiovasc. Imaging. 25, 85-90 (2009). 60L. M. Hurwitz, R. E. Reiman, T. T. Yoshizumi, P. C. Goodman, G. T oncheva, G. Nguyen, and C. Lowry, "Radiation dose from contemporary cardiothoracic multidetector CT protocols with an anthropomorphic female phantom: implications for cancer induction," Radiology. 245, 742750 (2007). 61M. Clemons, L. Loijens, and P. Goss, "Breast cancer risk following i rradiation for Hodgkin's disease," Cancer Treat. Rev. 26, 291-302 (2000). 62J. Geleijns, A.M. Salvado, P.W. de Bruin, R. Mather, Y. Muramatsu, and M.F. Nitt-Gray, "Computed tomography dose assessment for a 160 mm wide, 320 detector row, cone beam CT scanner," Phys. Med. Biol. 54, 3141-3159 (2009). 63M.K. Kalra, M.M. Maher, T.L. Toth, L.M. Hamberg, M.A. Blake, J.A. Shepard, and S. Saini, "Strategies for CT radiati on dose optimization," Radiology. 230, 619-628 (2004). 64S. Mori, K. Nishizawa, M. Ohno, and M. Endo, "C onversion factor to CT dosimetry to assess patient dose using a 256-slice CT scanner," Brit. J. Radiol. 79 888-892 (2006). 65 J.F.Winslow, D.E.Hyer, R.F.Fisher, C.J. Tien, and D.E.Hintenlang,"Construction of anthropomorphic phantoms for use in dosimetry studies," Accepted J.Appl.Clin.Med.Phys. (2009). 66Centers for Disease Control and Prevention, "C utaneous Radiation Injury: Fact Sheet for Physicians," (2005). Available from: http://emergency.cdc.gov/radiat ion/pdf/criphysicianfactsheet.pdf 67 M.Ghita, PhD dissertation, Un iversity of Florida, 2009. 68M. Cristy and K. F. Eckerman, Specific Absorbed Fractions of Energy at Various Ages from Internal Photon Sources ORNL/TM-8381/Volumes I-VII. (Oak Ridge National Laboratory, Oak Ridge, 1987)37C. Lee, C. Lee, S. Park, and J. Lee, "Development of the two Korean adult tomographic computational phantoms for organ dosimetry," Med. Phys. 33, 380-390 (2006).

PAGE 200

200 69C. Lee, C. Lee, S. Park, and J. Lee, "Dev elopment of the two Korean adult tomographic computational phantoms for organ dosimetry," Med. Phys. 33, 380-390 (2006). 70 R. Ambrose, MS thesis, University of Florida, 2006. 71A. K. Jones, T. A. Simon, W. E. Bolch, M. M. Holman, and D. E. Hintenlang, "Tomographic physical phantom of the newborn child with read-time dosimetry I. Methods and techniques for construction," Med. Phys. 33, 3274-3282 (2006).

PAGE 201

201 BIOGRAPHICAL SKETCH Lindsey K Lavoie was born in Boston, Massachusetts on June 28th, 1982 to Charles and Dona Lavoie. She graduated from Bishop St ang High School in 2000. In 2004, she graduated with honors from the College of the Holy Cross, earning a Bachelor of Arts degree in physics. After being accepted into the medical physics program at the University of Florida, she was awarded her masters degree in 2007. In the su mmer of 2009, she graduated from the University of Florida with a Ph.D in medical physics.