The University of Florida/National Cancer Institute Family of Hybrid Computational Phantoms Representing the Current Uni...

MISSING IMAGE

Material Information

Title:
The University of Florida/National Cancer Institute Family of Hybrid Computational Phantoms Representing the Current United States Population of Male and Female Children and Adolescents—Applications to Computed Tomography Dosimetry
Physical Description:
1 online resource (63 p.)
Language:
english
Creator:
Geyer, Amy, M
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Biomedical Engineering
Committee Chair:
Bolch, Wesley E
Committee Members:
Gilland, David R
Rill, Lynn

Subjects

Subjects / Keywords:
ct -- dosimetry -- obesity -- pediatric -- phantom
Biomedical Engineering -- Dissertations, Academic -- UF
Genre:
Biomedical Engineering thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
According to Report No. 160 by the National Council on Radiation Protection (NCRP), the effective dose per individual in the US due to medical exposures has risen nearly six-fold since 1982. Changes in the frequency of medical exposures and the types of modalities used are largely responsible for this increase and it is now crucial to estimate organs doses to patients for diagnostic medical imaging and radiotherapy procedures. To assist in the rapid reporting of patient organ doses, researchers at the University of Florida and the National Cancer Institute have developed a family of computational hybrid phantoms, constructed from NURBS and polygon mesh surfaces, that fully represent the ICRP 89 reference newborn, 1-year-old, 5-year-old, 10-year-old, 15-year-old male and female, and adult male and female. Coupled with Monte Carlo simulations, these phantoms can be used to estimate patient organ doses. In a study performed by Johnson et al, data from the CDC’s National Center for Health Statistics from 1988 to 1994 were used to create a library of phantoms extending from the 10th to 90th height/weight distributions. These phantoms were created, using modeling software, by scaling the UF/NCI reference phantoms to match targeted values of standing height, sitting height, total body weight, and secondary circumferential values. In total, Johnson et al created 25 adult male and 15 pediatric female patient-dependent phantoms. While suitable for initial studies for patient-phantom matching, substantial increases seen in childhood obesity over the intervening years has prompted us to undergo a major revision to the UF/NCI phantom library. Furthermore, a decision was made to construct the new library in a gridded fashion by height/weight without further reference to time-period-fixed weight/height percentiles. At each height/weight combination, secondary parameters of arm, thigh, waist, and buttocks circumference are also defined and used for phantom construction. All morphometric data for the new library are taken from the CDC NHANES survey data over the time period 1999 to 2006, the most recent reported survey period. A subset of the phantom library was then used in a CT organ dose sensitivity study to examine the degree to which full Monte Carlo simulations would be required to track organ doses for patients that are severely underweight to obese in body size. Through data analysis, it was found that organ dosimetry can be established through data interpolation of a more coarsely defined voxelized phantom library subset. In the future, the UF/NCI phantom library will be used to construct pre-computed dose libraries for individuals undergoing CT examinations. Ultimately, these libraries can be deployed in the clinic for electronic recording of patient organ dosimetry following diagnostic imaging procedures.
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 Amy Geyer.
Thesis:
Thesis (M.S.)--University of Florida, 2012.
Local:
Adviser: Bolch, Wesley E.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-05-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2012
System ID:
UFE0044218:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 THE U NIVERSITY OF FLORIDA /N ATIONAL C ANCER INSTITUTE FAMILY OF HYBRID COMPUTATIONAL PHANTO MS REPRESENTING THE CURRENT UNITED STATES POPULATION OF MALE AND FEMALE CHILDREN AN D ADOLESCENTS APPLICATIONS TO COMPUTED TOMOGRAPHY DOSIMETRY By AMY M. GEYER A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Amy M. Geyer

PAGE 3

3 To my parents, who have believe d in me from the day I was born

PAGE 4

4 ACKNOWLEDGMENTS I thank Dr. Wesley Bolch for the incredible opportunit y he has given me to complete such meaningful research and for his guidance throughout my past academic career and my academic career yet to come. I tha nk Dr. Choonsik Lee for all the knowledge he has imparted to me and for the work he has done making my r esearch possible. I thank Dr. Lynn Rill and Dr. David Gilland for t heir knowledge and contribution I thank my brother, Johnny, for paving the way. I thank my parents for their never ending love, support, and encouragement and for always reminding me

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 6 LIST OF FIGURES ................................ ................................ ................................ .......... 7 LIST OF ABBREVIATIONS ................................ ................................ ............................. 8 ABSTRACT ................................ ................................ ................................ ..................... 9 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 1 1 2 MATERIALS AND METHODS ................................ ................................ ................ 18 Statisti cal Analysis of National Health and Nutrition Examination Survey Data ...... 18 Reshaping of Reference Phantoms ................................ ................................ ........ 19 Body Mass Index Analysis ................................ ................................ ...................... 20 Patient Dependent Phantom Construction ................................ .............................. 21 Computed Tomography Simulated Exams ................................ ............................. 23 3 RESULTS AND DISCUSSION ................................ ................................ ............... 26 Obesity Trends ................................ ................................ ................................ ....... 26 Updated Reference Phantoms ................................ ................................ ................ 26 Phantom Library Grids ................................ ................................ ............................ 27 Patient Dependent Phantom Library Subset for Computed Tomography Dosimetry ................................ ................................ ................................ ............ 28 Computed Tomography Organ Dosimetry ................................ .............................. 28 Organ Dose Interpolation ................................ ................................ ........................ 30 4 CONCLUSION AND FUTURE WORK ................................ ................................ .... 56 APPENDIX COMPUTED TOMOGRAPHY MODELING AND SIMULATION ................................ ... 58 Computed Tomography Scanner Description ................................ ......................... 58 Modeling of the Computed Tomography Scanner X ray Source ............................. 58 Computational Phantom Dose Measurements ................................ ....................... 59 LIST OF REFERENCES ................................ ................................ ............................... 61 BIOGRAPHICAL S KETCH ................................ ................................ ............................ 63

PAGE 6

6 LIST OF TABLES Table page 3 1 Percent difference summary for pediatric males between NHANES databases. ................................ ................................ ................................ .......... 44 3 2 Percent difference summary for pediatric females between NHANES databases. ................................ ................................ ................................ .......... 44 3 3 Updated circumference values for reference phantoms. ................................ .... 45 3 4 Targeted heights, weight, and circumferential data for male pediatric phantoms. ................................ ................................ ................................ ........... 46 3 5 Targeted heights, weight, and circumferential data for female pediatric phantoms. ................................ ................................ ................................ ........... 48 3 6 Summary of organ doses for pediatric female phantoms at a fixed weight of 50 kg, at heights ranging from 135 cm to 175 cm, in 10 cm increments. ............ 50 3 7 Summary of organ doses for pediatric female phantoms at a fixed height of 165 cm, at weights ranging from 40 kg to 115 kg. ................................ .............. 51 3 8 Percent difference between interpolated and Monte Carlo calculated organ doses for 165 cm phantoms, removing every other phantom (10 kg intervals). 52 3 9 Percent difference between interpolated and Monte Carlo calculated organ doses for 165 cm phantoms, at 15 kg intervals ................................ .................. 53 3 10 Percent difference between interpolated and Monte Carlo calculated organ doses for 165 cm phantoms, at 20 kg intervals ................................ .................. 54 3 11 Percent difference between interpolated and Monte Carlo calculated organ doses for 165 cm phantoms, at 25 kg intervals. ................................ ................. 55

PAGE 7

7 LIST OF FIGURES Figure page 2 1 Perspective view of adult female reference phantom demonstrating the coordinate system used during the phantom scaling process. ........................... 25 3 1 Percent of pediatric male individuals that are obese as a function of standing height. ................................ ................................ ................................ ................. 32 3 2 Percent of pediatric female individuals that are obese as a function of standing height. ................................ ................................ ................................ .. 33 3 3 UF/NCI series of ICRP89 compliant hybrid reference phantoms ........................ 34 3 4 Targeted grid for the UF/NCI library o f pediatric male hybrid phantoms. Color code indicates the phantom classification as underweight, healthy, overweight, or obese based upon BMI percentiles and CDC definitions of body morphometry. ................................ ................................ ............................. 35 3 5 Targeted grid for the UF/NCI library of pediatric female hybrid phantoms. Color code indicates the phantom classification as underweight, healthy, overweigh t, or obese based upon BMI percentiles and CDC definitions of body morphometry. ................................ ................................ ............................. 36 3 6 Targeted grid for the UF/NCI library of pedi atric male hybrid phantoms. Comment box opens upon selection of desired cell to display height, weight, average sitting height, and secondary circumferential data. ............................... 37 3 7 Construction methodology for UF/NCI library of pediatric male hybrid phantoms based upon 3D proportional scaling of existing UF/NCI reference phantoms to match targeted values of sitting height. ................................ .......... 38 3 8 Construction methodology for UF/NCI library of pediatric female hybrid phantoms based upon 3D proportional scaling of existing UF/NCI reference ph antoms to match targeted values of sitting height. ................................ .......... 39 3 9 Selection of phantoms at a fixed height of 165 cm at varying weights. ............... 40 3 10 Phantoms at a fixed weight of 50 kg at varying heights. ................................ ..... 41 3 11 Organ doses as a function of height for organs in the beam for a CAP CT scan at 120 kVp, with effective mAs of 100 and pitch of 1.375. .......................... 42 3 12 Organ doses as a function of weight for organs in the beam for a CAP CT scan at 120kVp, with effective mAs of 100 and pitch of 1.0. ............................... 43

PAGE 8

8 LIST OF ABBREVIATION S BMI Body mass index CAESAR Civilian American and European Surface Anthropometry Resource CDC Centers for Disease Control and Protection CT Computed Tomography ICRP International Commission on Radiological Protection MCNPX Monte Carlo n particle extended MR Magnetic Resonance NCI National Cancer Institute NCRP National Council on Radiation Protection NHANES National Health and Nutrition Examination Survey NURBS Non uniform B spline surfaces RPI Rensselaer Polytechnic Institute UF University of Florida

PAGE 9

9 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Mast er of Science THE U NIVERSITY OF FLORIDA /N ATIONAL C ANCER INSTITUTE FAMILY OF HYBRID COMPUTATIONAL PHANTO MS REPRESENTING THE CURRENT UNITED STATES POPULATION OF MALE AND FEMALE CHILDREN AN D ADOLESCENTS APPLICATIONS TO COMPUTED TOMOGRAPHY DOSIMETRY By Amy M. Geyer May 2012 Chair: Wesley Bolch Major: Biomedical Engineering To assist in the rapid reporting of patient organ doses, r esearchers at the University of Florida (UF) and the National Cancer Institute (NCI) have developed a family of computational hybrid phantoms, constructed from non uniform rational B spline ( NURBS ) and polygon mesh surfaces, that fully repres ent the International Commission on Radiological Protection (ICRP) 89 50 th percentile reference newborn, 1 year old, 5 year old, 10 year old, 15 year old male and female, and adult male and female. Coupled with Monte Carlo simulations, these phantoms can be used to estimate patient organ doses Substantial increases seen in childhood obesity in the United States have prompted us to undergo a major revision to the UF/NCI phantom library. Furthermore, a decision was made to construct the new library in a gridded fashion by height/weight without further reference to age dependent weight/height percentiles. At each height/weight combination, secondary circumferential parameters are also defined and used f or phantom construction All morphometric data for the new library are taken from the Centers for Disease Control and Prevention (CDC) National Health and Nutrition Examination Survey data over the time period 1999 to 2006, the most recent reported

PAGE 10

10 surv ey period. A subset of the phantom library was then used in a computed tomography (CT) organ dose sensitivity study to examine the degree to which full Monte Carlo simulations would be required to track organ doses for patients that are severely underweig ht to obese in body size. Through data analysis, it was found that organ dosimetry can be established through data interpolation of a more coarsely defined voxelized phantom library subset. In the future, the UF/NCI phantom library will be used to constru ct pre computed dose librar ies for individuals undergoing CT examinations. Ultimately, these libraries can be deployed in the clinic for electronic recording of patient organ dosimetry following diagnostic imaging procedures.

PAGE 11

11 CHAPTER 1 INTRODUCTION With the rise in medical imaging procedures, particularly from the modalities of computed tomography (CT) and nuclear medicine, the need for accurate and detailed patient dosimetry is becoming increasingly important. 1 As established in a 2009 report by the National Council on Radiation Protection (NCRP) the effective dose per individual in the United States due to medical exposures was estimated to be 3.0 mSv. 2 This value has ris en nearly six fold from a previous per c apita effective dose estimation of 0.54 mSv in 1982. D etailed computational anthropomorphic phantoms, coupled with Monte Carlo simulations, can be employed to estimate organ doses to a patient. Computational anatomic phantoms can be classified into three f ormat types Stylized ( mathematical) p hantoms are characterized by three dimensional geometric surface equations that define the body contour and external anatomy. Although organ re position ing and size scaling can be performed with relative ease, stylized phantoms do not provide high accuracy in dose estimates due to their simplified anatomy. Voxel phantoms represent a phantom type in which organs are defined by groups of voxels from segmented CT or magnetic resonance (MR) images. V oxel (or tomographic) ph antoms provide anatomical accuracy but offer limited flexibility when re scal ing to various patient sizes. Because voxel phantoms are created from segmented CT or MR scans and from grouping voxels to define organs, the se phantoms are limited by the voxel structure. In an effort to combine the flexibility of stylized phantoms with the anatomical accuracy of voxel phantoms, hybrid phantoms were developed by researchers at the University of Florida and later extended at the National Cancer Institute (NCI) Hybrid phantoms employ non uniform rational B spline (NURBS) and

PAGE 12

12 polygon mesh (PM) surfaces that permit one to model organ s with high anatomical realism while allowing eas e in organ and body region deform ation through the m anipulation of control points (NURBS) and polygon vertices (PM) defining the surfaces of individual organ and body regions (head, arms, legs, and torso). For each format type phantoms can be further divided into four morphometric categories reference, patient dependent, patient sculpted, and patient specific. R eference phantoms use 50 th percentile values for a variety of anthropometric parameters such as height, weig ht, organ mass, and total body mass and are designed to represent average individual s in a patient population. They may be matched to individual patient s by age. Reference phantoms are the basis for establishment of reference dose coefficients needed in prospective radiological protection programs, and for establishing means of verifying compliance with radiation protection standards. Nevertheless, they lack anatomical specificity for individuals who diverge from 50 th percentile parameters. By relaxing the restriction on 50 th percentile for the defining body morphometry parameters, one then defines a patient dependent phantom, or more specifically, a phantom library, whereby the patient is matched to a given phantom in that library through not just age, but by both height and weight. Patient dependent phantom libraries thus significantly improve the patient specificity of the dose estimate through closer matching of body shape. Depending upon the differential change in body shape height, weight, body region circumference a close or approximate match to a real patient may be made. The three morphometric category patient sculpted phantoms take this a step closer by adjusting body surface control points of a pat ient dependent phantom to unique and exactly match those of a given

PAGE 13

13 patient. The major disadvantage of this approach is that the phantom is made uniquely for the patient in question, and thus no prior organ dosimetry library could be created, as is possibl e for patien t dependent phantom libraries. P atient specific phantoms provide the most accurate match of both external as w ell as internal organ anatomy. They are created by segmenting whole body CT scans to uniquely model exterior and internal organ morphometry. The requisite s egmenting images is a labor intensive process and whole body scans are o ften not available for patients. Consequently patient specific phantoms are often not used or created for organ dosimetry studies When using computational phantoms for medical dosimetry, the need for accuracy must be balanced with practicality. Patient dependent phantoms provide an acceptable level of accuracy by modifying reference phantoms to match anthropometr ic parameters of patients, but remaining broad enough to represent a diverse population. Because patient dependent phantoms must be created by modifying reference phantoms, it is necessary to work wi th hybrid phantoms due to their ability to be extensively deformed Utilizing hybrid patient dependent phantoms, phantom libraries can be created by matching statistically analyzed anthropometric parameters to represent a population of individuals of varying body sizes and shapes. Phantom libraries are useful in radiation protection studies, as well as pre computed dose libraries and can be used to analyze changes in dose with varying body types. Broggio et a l constructed a phantom library consisting of 25 whole body adult NURBS based Caucasian phantoms with 109 identified organs or tissues 3 Using a sampling strategy, 25 individuals were selected from the European Edition of the

PAGE 14

14 Civilian American and European Surface Anthropometry Resource (CAESAR) database The CAESAR database provides a mesh geometry o f the outer body co ntour for individual volunteers subjected to exterior optical scanning These body contours were then subsequently converted to NURBS surfaces for inclusion in the CAESAR library I nternal organs for the models w ere taken from the commercially available o rgan set s provided by 3DSp ecial ( http://www.3dspecial.com/) and were resized using scaling factors incorporated into each model. 3 The body contour and internal organs and skeletal models were subsequently voxelized and merged together to create the final whole body adult phantoms. In a study by Na et al percentile specific phantoms were created using computer algorithms to deform previously constructed base phantoms (RPI adult male and RPI adult female ) 4 The internal organs with in the phantoms were creat ed using a commercial organ mesh dataset, AnatomiumTM 3D ( http://www.anatomium.com/ ) and were scaled using computer algorithms to match volume and mass percentiles derived from International Commission on Radiologic al Protection ( ICRP ) Publications 23 and 89 To create varying whole body sizes, anthropometric data (height and weight percentil es) for 19 year old males and females were derived from the National Health and Nutrition Examination Survey (NHANES) database (1999 2002) and these parameters were matched by deforming another commercially available skin surface, MakeHuman TM ( http://www.makehuman.org/ ) Cassola et al created a library of 18 anthropometric phantoms (9 adult male phantoms and 9 adult female phantoms) based on 10th, 50th and 90th mass and height percentiles of Caucasian individuals extracted from the commercially available

PAGE 15

15 ergonomic software PeopleSize 5 The 3D modeling software Blender was used to deform pre viously constructed base phantoms, MASH3_sta (male adult mesh standing) and FASH3_sta (female adult mesh standing) to match the extracted height and weight parameters 6 To resize internal anatomy, scaling factors were derived as a function of height based on an autopsy study done by de la Grandmaison et al 7 While the previously discussed studies represent comprehensive libraries of phantoms, the accuracy of commercially available outer body contours and internal organs is questionable. The outer body contours of Caucasian adult males taken from the CAESAR database are comprised of 412 individuals scanned in Italy and 566 individuals scanned in the Netherlands. Due to differing anth ropometric features of individuals of different nationalities, the individuals found in the CAESAR database are not representative of the general Caucasian population, but rather a smaller subset. Like the CAESAR database, 3DSpecial offers a commercially a vailable set of internal organs with little documentation in the study of their origin, thus the degree of their realism was not discussed in great detail. Researchers at the University of Florida (UF) develo ped a family of hybrid phantoms constructed fro m NURBS surfaces, th at contains a reference newborn 1 year old, 5 year old, 10 year old, 15 year old male and female, and adult m ale and female. The phantoms were developed to reflect the reference specifications found in ICRP Publication 89. 8 While mainta ining anatomic al realism, hybrid phantoms are also capable of having their internal and superficial anatomies altered due to the flexibility of the surfaces. In a study performed by Johnson et al a methodology for constructing patient dependent phantoms based on anthropometric percentile distributions from the

PAGE 16

16 NHANES III database was derived to represent a distribution of the United States pediatric and adult populations 9 The p atient dependent p hantoms were then created by from the UF family of hybrid phantoms to match predefined anthropometric parameters. In total, Johnson et al created 25 adult male and 15 pediatric female patient dependent phantoms. Th e NHANES III database included anthropometric data for pediatric and adults collected from examinations from 1988 1994. To provide a more comprehensive library extension to represent the pediatric population and the rise in childhood obesity in the United States this study was performed to provide a vital update and extension to the work done by Johnson et al In 2010, t he Centers for Disease Control and Prevention (CDC) has estimated tha t roughly 12.5 million (17%) children and adolescents ( ages 2 19) are obese in the United States, almost tripling the childhood obesity rate since 1980. 10 With these large increases, it is necessary to update the data from Johnson et al to reflect these changes. It is also necessary to abandon the previous method of age bas ed percentile anthropometric parameter calculations Instead, grids based on height and weight bins that will capture all individuals and allow for varying growth rates of pediatric populations were created The purpose of this study was to create comprehe nsive grids for male and female pediatrics using more recent data from the NHANES database (1999 2006) to reflect obesity trends in the United States among this population These grids will be used in the future to create a library of patient depend ent pha ntoms that represent pediatric populations. By selecting a weight column and height row from the pediatric

PAGE 17

17 female grid, phantoms were constructed to investigate the role of height and weight variation on organ dose sensitivity for CT exams.

PAGE 18

18 CHAPT ER 2 MATERIALS AND METHOD S Statistical Analysis of N ational H ealth and N utrition E xamination S urvey Data Anthropometric pediatric data was obtained from the NHANES database conducted between 1999 and 2006 (which will further be referred to as the NHANES IV database for comparisons) Height and weight were chosen as primary parameters in the construction of a grid for pediatric males and females aged 2 to 20 (with a total of 17,028 individuals) Waist circu mference, thigh circumference and arm circumference were included as secondary parameters. This library will serve as an update to the work performed by Johnson et al which included data on individuals from 1988 to 1994 (NHANES III) The motivation for th is update is the dramatic rise in childhood and adult obesity over the intervening years in the United States Due to the unavailability of average sitting height and buttock s circumference in the database applied in this study, these values were interpol ated from the prior NHANES III database to match the heights and weights of the individuals in the updated grid. University of Florida analyzed male and female pediatric individuals independently and binned by heig ht increments of 5 cm ensuring that the 5th and 95th height percentiles were captured using a MATLAB Gaussian fitting program. Each height bin was further parsed by weight increments of 2.5 kg, with each weight bin containing data for at least ten indiv iduals to assure statistical significance. For each height weight bin, secondary parameters including the average circumferences of the arm, waist, and thigh were determined, along with sitting height and buttock s circumference from the previous NHANES dat abase. For a small number of weight extremes not all secondary parameters contained the necessary ten

PAGE 19

19 individuals for a statistically significant result, so linear interpolation was u sed to determine the proper circumferences. Using primary and secondary p arameters, a grid containing 85 male height weight bins and a grid containing 73 female height weight bins were constructed. These grids will provide the blueprints for later constructing a comprehensive library of patient dependent phantoms containing 158 pediatric phantoms. Reshaping of Reference Phantoms Due to the recent obesity trends, it was necessary to update the UF /NCI family of reference hybrid phantoms to reflect these trends. To begin this update, the primary parameters of height and weight wer e left unchanged to match ICRP reference values, while the secondary circumference parameters were revised. This process included a double linear interpolation of the circumference data. The first interpolation of the circumference data was done to match t he ICRP reference weight, while the second interpolation was to match the height. In cases that lacked sufficient circumferential data to allow for interpolation, extrapolation based on the least squares method was used to obtain the circumference. Updated circumferences were calculated for each phantom in the current UF /NCI reference phantom library, excluding the newborn and 1 year old since this data was not acquired from the NHANES database. After calculating updated circumference data, the phantoms were then reconstructed by Dr. Choonsik Lee, investigator at the National Cancer Institute in the Division of Cancer Epidemiology and G enetics. Circumferences on the phantoms were matched within 3% of targeted values and the total body of the phantom was modified to compensate for circumference adjustments to match ICRP reference masses within 1%.

PAGE 20

20 Body Mass Index Analysis To further aid in the construction of the pediatric phantom library, body mass index (BMI) calculations were performed for each height weight bin within the grids. Because the scaling of phantoms can become quite labor intensive requiring additional steps for overweight individuals (detailed in a later section), it can be useful to know the weight category (underweight, healthy, overweight, or obese) prior to the scaling calculated for pediatrics with relative ease. To calculate the BMI for each pedi atric h eight divided by their height (in meters) squared. For adults, this calculated number would then fall within an acceptable range for each weight category. For pediatric ind ividuals, weight categories are not determined as easily due to body development during the pediatric years. To determine a weight category, an average age of the individual in each bin was estimated using the ages of the individuals that were refined by h eight ( bin height bin height + 0.9 cm) and weight (bin weight bin weight 2.5 kg) Since BMI calculation was used only to predict a rough estimate of body size, it should be noted that average age calculations did not require the minimum of 10 indivi duals per calculation. After obtaining a BMI calculation and average age for each height weight combination on the grids, BMI for age charts provided by the CDC were used to determine a BMI percentile for each individual. BMI percentile ranges then were u sed to determine the weight category of each individual. BMI percentiles below the 5 th percentile are considered underweight individuals, while percentiles ranging between 5 th percentile and less than 85 th percentile represent healthy individuals. Percenti les ranging between 85 th percentile and below 95 th percentile are considered overweight

PAGE 21

21 P ercentiles equal to or greater than the 95 th percentile are considered obese. The pediatric grids were then color coded based on these weight categories. Patient Dep endent Phantom Construction In this study, 20 patient dependent phantoms were constructed to aid in CT dosimetry and dose sensitivity analysis. To aid in the construction of the phantoms, both grids were redesigned to detail which ever reference phantom would serve as the anchor phantom in the scaling process, and determine whether the height would be scaled up or down. Since the effects of height and weight on dose w ere of interest in this study, five female pediatric phantoms with a t otal body mass of 50 kg were made, at 135 cm, 145 cm, 155 cm, 165 cm, and 175 cm standing height s. Additionally, 16 female pediatric phantoms were con structed at a fixed height of 16 5 cm with weights ranging between 40 kg and 11 5 kg, at 5 kg increments (on e phantom was previously constructed at this height when the fixed weight phantoms were created and did not need to be made again) Along with matching the heights and weight detailed within the grid, secondary circumference parameters at each grid point w ere also matched. The scaling process is this study mimics methods previously used by Johnson et al and begins with selecting an appropriate anchor phantom from the phantom mapped grid importing it into the NURBS modeling software Rhinoceros 4.0. 9 Initial ly, the sitting height of the anchor phantom was measured as the distance between the top of the head to the bottom of the ischium. The ratio of the targeted sitting he ight defined in the grid to the measured sitting height produced a scaling factor. The h ead, torso, and arms (including all internal organs and bony anatomy) were scaled uniformly in 3 D by this scaling factor, with the bottom of the ischium acting as the scaling point of origin. To match the target standing height of the phantom the legs we re scaled in the z direction

PAGE 22

22 by a scaling factor equal to the ratio of measured leg length to targeted leg length (coordinate system shown in Figure 2 1) With the proper standing and sitting heights matched, the secondary circumferential parameters were t hen matched. To assure consistency between anchor phantoms for circumferential measurements, a separate layer was added to each anchor phantom with reference planes placed as markers for circumference measurements. These planes were placed at the points of measurement of the circumferences as defined by the CDC. C ontrol points on the outer body contour were modified to simulate a change in adipose tissue thickness to match circumferences For phantoms with waist circumferences smaller than the waist circumf erence of the anchor phantom, a 2D scale was applied to the phantom in the x y plane to accommodate changes to circumferences while preserving previously matched height requirements. For obese phantoms, arms and legs were rotated outward to avoid an overlap of outer body contours due to the increased circumference values. Once circumferences were matched, control points near to the point of circumference measurements were visually anal yzed and fine tuned to preserve correct body shape. To complete the phantom scaling process, total body mass was iteratively matched by adjusting the control points of the outer body contour (adipose tissue) in areas that were not restricted by secondary parameters. These areas of the body include lower arms and legs, upper torso, and adipose tissue of the breasts. Total body mass was estimated in Rhinoceros TM using the volume calculation tool and tissue densities. Five distinct tissue types were defined as adipose tissue, cartilage, homogenous bone, lung, and residual soft tissue and used to calculate the total body mass. Volumes of cartilage, bone (cortical), lungs muscle, and outer body contour were

PAGE 23

23 calculated initially. When calculating the volume the outer body contour body, all objects within this layer were first joined using the Boolean Union tool in Rhinoceros to eliminate any overlap between surfaces. The volume of adipose tissue was estimated by subtracting the volume of the muscle layer from th e outer body contour volume. The volume of the bulk soft tissue was found by subtracting the volumes of the bone, cartilage, and lungs from the muscle volume. Total body mass was found by multiplying the volumes of adipose tissue, bone, lungs, cartilage, a nd residual soft tissue by their respective densities and summing th es e masses C omputed T omography Simulated Exams In collaboration with Mr. Daniel Long, PhD student in medical physics, the effect of body morphometry on organ doses for a chest abdomen pelvis (ranging from the thoracic inlet to the lesser trochanters) CT scan was assessed by simulating a Siemens Sensation 16 CT scanner using the Monte Carlo transport code MCNPX2.6 (using 100 million particle histories) and performing on 20 phantoms. To adequately compile a comprehensive database of varying body morphometries, 5 phantoms were constructed at height increments of 10 cm, starting with 135 cm at a fixed weight of 50 kg to assess the effects of height on organ doses. The effects of weight on organ doses were assessed using 16 phantoms constructed at weight increments of 5 kg, starting with 40 kg, at a fixed height of 165 cm. P hantoms using adult and 15 year old anchor phantoms were v oxelized to resolutions of 3 x 3 x 3 mm 3 and phantoms using the 10 year old anchor phantom were voxelized to a resolution of 2 x 2 x 2 mm 3 The scan parameters for each simulation were beam energy of 120 kVp, body fi lter, pitch of 1.0 (1.3 7 5 for the phantoms at a fixed weight), effective mAs of 10 0, and a 2 .4 c m beam collimation. The

PAGE 24

24 scan ranged from the thoracic inlet to the lesser trochanter on each phantom. Additional information on the CT simulation can be found in Appendix A.

PAGE 25

25 Figure 2 1. Perspective view of adult female reference p hantom demonstrating the coordinate system used during the phantom scaling process.

PAGE 26

26 CHAPTER 3 RESULTS AND DISCUSSI ON Obesity Trends Figures 3 1 and 3 2 show the percentage of obese pediatric individuals as a function of standing height for both NHANES databases. The percentage of obese individuals increases with standing height for both male and female pediatric populations. Consistent with the previously mentioned CDC estimates for obesity, the percentage of obese individuals is higher in the mo re recent NHANES IV database when compared with the NHANES III database. This is true in all cases except males and females with standing heights of 85 cm, and females with a standing height of 125 cm. Tables 3 1 and 3 2 quantify the percent difference be tween pediatric individuals in e ither NHANES database, showing an average increase of 33.71% in the percent of obese individuals for males, and 23.01% for females. Updated Reference Phantoms Using the data collected in the NHANES IV database, reference ph antoms were updated to match secondary circumference data interpolated fr om the database based on ICRP 89 reference heights and weights. Updates to circumferential data also corrected for an anatomical discrepancy in the phantom outer body shape. Previousl y, the buttocks circumference in all phantoms, excluding the 15 year old and adult female phantoms, was found to be less than the waist circumference. This trend was not consistent with circumference values seen in the database, which showed a buttock s cir cumference larger than waist circumference. Table 3 3 shows the measured, suggested updates, and revised circumferential data for all reference phantoms that were updated. For consistency, all circumference measurements were made using

PAGE 27

27 previously discusse d measurement planes. The revised circumferences were matched to the suggested updated dated values within 3 % Once circumferences were matched within 3%, organ masses, total body mass and separable fat mass were matched within 3% of ICRP 89 reference valu es. The updated reference phantoms are shown in Figure 3 3 Phantom Library Grids The final height/weight grid for pediatric males and females are show in Figures 3 4 and 3 5 respectively. The grids are color coded to indicate the phantom classi fication as underweight healthy, overweight, or obese based on BMI percentile analysis CDC definitions of body morphometry. In total, the grids target 85 pediatric males and 73 pediatric females. When viewing the grids in Microsoft Excel, secondary circumferentia l data is displayed in a comment box that opens when selecting a cell of interest, de monstrated in Figure 3 6 Target secondary circumferential data, as well as average sitting height, for male and female pediatric phantoms are shown in Tables 3 4 and 3 5, respectively. To aid in the phantom construction process, a second set of grids were created to incorporate the 3 D up or downscaling methodology of the reference phantoms. These grids are shown in Figures 3 7 and 3 8 for male and female pediatrics, respectively. The color coding of these grids indicates the reference phantom that will be used initially in the scaling process. The direction of the arrow found in each cell indicates either a three dimensional up or d own scaling to the reference phantom torso, arms, and head to match targeted sitting height values.

PAGE 28

28 Patient Dependent Phantom Library Subset for C omputed T omography Dosimetry A subset of patient dependent phantoms was created from the pediatric female gri d to explore the effects of height and weight on organ doses in a simulated CT examination. To investigate the effect of weight on organ doses, 16 patient dependent phantoms were created at a fixed standing height of 165 cm. The weights for these phantoms ranged between 40 kg and 115 kg in 5 kg increments. A selection of t hese phantoms is displayed in Figure 3 9. Additionally, to analyze the effects of patient height on organ dosimetry, 5 patient dependent phantoms were created at a fixed weight of 50 kg, w eight heights ranging from 135 cm to 175 cm, in 10 cm increments. These phantoms are displayed in Figure 3 10. When constructing the 20 patient dependent phantoms, total body mass was matched to targeted values within 1 kg, and all height and secondary ci rcumferential data were matched to targeted values within 1%. After their construction, all phantoms were voxelized using the MATLAB TM code Voxelizer v9.2. Post voxelization total body masses matched the pre voxelization masses within 1% and targeted total body mass values within 1 kg. C omputed T omography Organ Dosimetry To assess the effect of standing height on organ dosimetry, five patient dependent phantoms were created and 120 kVp CT simulations were performed using MXNCPX2.6 for each phantom with a p itch of 1. 3 7 5. A summary of these organ doses can be found in Table 3 6. To better correlate the effects of height on organ d ose, organ doses for organs in the beam were plotted as a function of height in Figure 3 11. As height increases, internal organs a re scaled with the head and torso region to match targeted sitting height values. Due to this scaling, organ volumes are increasing with

PAGE 29

2 9 height. Similarly for phantoms in which the reference waist circumference was significantly larger than targeted circu mferential values, the torso /head regions were scaled in two dimensions to accommodate a smaller waist circumference (165 cm and 175 cm phantoms) This also changes organ volumes. These changes to organ volume result in significant organ variation, making dose correlations very difficult. Figure 3 11 depicts this organ dose variability, most significantly in the kidneys and spleen. Adjusting the scan to a pitch of 1, Monte Carlo CT simulations were performed on 16 patient dependent phantoms to investigate the effect of weight on organ doses. The arms were removed (excluding the femoral head) from each phantom to mimic clinical condit i ons in which the patients would be positioned with their arms raised above their head for the examination. Organ doses for these simulations can be found in Table 3 7. For organs in the beam, organ doses as a function of weight are plotted in Figure 3 12. Figure 3 12 shows the predicted downward trend of in field organ doses with a fixed effective mAs of 100, with increased patient weight due to increased buted to statistical variation. These slight dips and rises in the data as weight increases could also be due to varying distributions of adipose tissue in each phantom due to the phantom scaling process. For example, Figure 3 12 depicts a sharp increase in the lung dose of the 100 kg p hantom, relative to the 95 kg phantom. Upon inspection of the phantoms, the volume of breast tissues in the 95 kg phantoms is significantly larger than that of the 100 kg phantoms, which could contribute to a decrease in lung dose. Similar adipose tissue r edistributions for higher mass phantoms could account for discrepancies in organ doses.

PAGE 30

30 Organ Dose Interpolation Upon completion of the pediatric male and female patient dependent libraries, UF will be working with NCI to construct a CT dose library to lat er be used for patient dose tracking. Though all 158 pediatric phantoms will be available in their original NURBS/polygon mesh format, the po ssibility exists to create a coa rse sampling of this library to be voxelized and imported into MCPNX for organ dose library generation. Using this smaller voxelized phantom library, the extent to which organ doses can be interpolated for phantoms that were not voxelized was investigated. Because organ doses followed predicted trends for the 16 phantoms created at a fix ed height of 165 cm, interpolation was performed using this subset of the patient dependent phantoms created for CT dosimetry. To begin this investigation, the 16 patient dependent library of calculated CT organ doses was further parsed to include onl y 8 phantoms, removing every other phantom (10 kg intervals) Using the organ doses for in field organs, data points were fit with a second degree polynomial trendline using Microsoft Excel (with R 2 values greater than 0.95). Using the equation of each unique trendline for its respective organ, organ doses were calculated for the phantoms that were previously removed from the library. The percent difference between interpolated and Monte Carlo calculated organ doses can be found in Table 3 8. This s ame method was used when interpolating between phantoms at 15 kg intervals, 20 kg intervals and 25 kg intervals and percent diffe rence can be found in Tables 3 9, 3 10, and 3 11 respectively. Generally, organ dose interpolation with each level of coarseness proves efficient and beneficial, with low percent errors. For heavier phantoms (larger than 70 kg), percent differences increase 12.

PAGE 31

31 When interpolating organ doses after removing ever y other phantom, 22 .7 9 % of inter polated doses had percent differences between 5% and 10% of original organ doses values. Interpolated organ values with percent differences greater than 10% were present for phantoms with total body masses 85 kg and greater, accounting for 13.97% of the total interpolated values. After interpolating between phantoms at 10 kg intervals organ dose values with percent differences betwe en 5% and 10% accounted for 21.18 % of total interpolate d organ dose values. Only 5.29 % of all interpolated values had percent differences greater than 10% and were found only in phantoms with total body mass equal to or greater than 90 kg. Interpolated organ doses with percent differences 5% to 10% from original org an doses values accounted for 18.62 % of all interpolated values for phantoms at 15 kg intervals Percent differences greater than 10% accounted for only 16.17 % of all interpolated values and were present in phantoms with total body masses greater than or equal to 85 kg. Lastly, when i nterpolating between phantoms at 20 kg intervals percent differences ranging between 5% and 10% accounted for 20.58 % of all interpolated values. Percent differences greater than 10% for interpolate d organ doses accounted for 12.75 % of all interpolated val ues and were seen in lighter individuals, as well as individuals with larger total body masses. From these calculations, removing every two phantoms proves to be the more efficient and accurate method of interpolation for phantoms at a fixed height with varying weights. This method is most effective for phantoms with total body masses less than 90 kg, with interpolated organ dose percent differences less than 10% from original values.

PAGE 32

32 Figure 3 1. Percent of pediatric male individuals that are obe se as a function of standing height.

PAGE 33

33 Figure 3 2. Percent of pediatric female individuals that are obese as a function of standing height.

PAGE 34

34 Figure 3 3. UF/NCI series of ICRP89 compliant hybrid reference phantoms

PAGE 35

35 Figure 3 4 Targeted grid for the UF/NCI library of pediatric male hybrid phantoms. Color code indicates the phantom classification as underweight, healthy, overweight, or obese based upon BMI percentiles and CDC definitions of body morphometry.

PAGE 36

36 Figure 3 5 Targeted grid for the UF/NCI library of pediatric female hybrid p hantoms. Color code indicates the phantom classification as underweight, healthy, overweight, or obese based upon BMI percentiles and CDC definitions of body morphometry.

PAGE 37

37 Figure 3 6. Targeted grid for the UF/NCI library of pediatric male hyb rid phantoms. Comment box opens upon selection of desired cell to display height, weight, average sitting height, and secondary circumferential data.

PAGE 38

38 Figure 3 7. Construction methodology for UF/NCI library of pediatric male hybrid phantoms based upon 3D proportional scaling of existing UF/NCI reference phantoms to match targeted values of sitting height.

PAGE 39

39 Figure 3 8 Construction methodology for UF/NCI library of pediatric female hybrid phantoms based upon 3D proportional scali ng of existing UF/NCI reference phantoms to match targeted values of sitting height.

PAGE 40

40 Figure 3 9. Selection of phantoms at a fixed height of 165 cm at varying weights.

PAGE 41

41 Figure 3 10. Phantoms at a fixed weight of 50 kg at varying heigh ts.

PAGE 42

42 Figure 3 11. Organ doses as a function of height for organs in the beam for a CAP CT scan at 120 kVp, with effective mAs of 100 and pitch of 1.375.

PAGE 43

43 Figure 3 12. Organ doses as a function of weight for organs in the beam for a CAP CT scan at 120kVp, with effective mAs of 100 and pitch of 1.0.

PAGE 44

44 Table 3 1. Percent difference summary for pediatric males between NHANES databases. Height NHANES III NHANES IV % Difference 85 3.79 2.21 41.66 95 4.49 5.49 22.17 105 6.67 12.26 83.85 115 13.32 14.51 8.92 125 12.45 17.92 44.01 135 17.18 21.83 27.11 145 17.90 23.57 31.65 155 20.52 23.52 14.63 165 15.54 22.13 42.40 175 11.96 22.42 87.44 185 15.10 22.70 50.30 Average 33.71 Table 3 2. Percent difference summary for pediatric females between NHANES databases. Height NHANES III NHANES IV % Difference 85 3.36 3.11 7.46 95 5.56 8.83 58.92 105 9.36 10.49 12.07 115 13.71 13.93 1.58 125 14.97 14.60 2.45 135 17.08 18.86 10.40 145 15.78 21.22 34.50 155 16.86 19.92 18.15 165 16.86 21.58 28.01 175 14.74 26.01 76.42 Average 23.01

PAGE 45

45 Table 3 3. Updated circumference values for reference phantoms. Circumference (cm) Waist Buttock s Arm Thigh Phantom Meas. Updated Target Revised Reference Diff (%) Meas. Updated Target Revised Reference Diff (%) Meas. Updated Target Revised Reference Diff (%) Meas. Updated Target Revised Reference Diff (%) UFH05MF 57.7 52.5 54.2 3% 58 56.2 57.7 3% 14.8 17.5 17.5 0% 26.4 31.2 30.9 1% UFH10MF 70.9 61.2 63.1 3% 72.5 70.4 70 1% 21.4 20.7 21.4 3% 31.6 38.2 37.6 2% UFH15M 89.9 72.2 74.4 3% 85 86.1 86.5 0% 27.2 26.6 27.2 2% 35.6 47 45.8 3% UFH15F 76.9 73.4 75.9 3% 80.5 89.6 88.8 1% 27.7 25.4 25.4 0% 43.5 46.7 46.4 1% UFHADM 97.9 89.3 92 3% 96.3 94.9 96.3 2% 27.3 30.8 31.1 1% 39.9 49.4 48.2 2% UFHADF 84.8 81.8 83.7 2% 93.8 95.9 93.8 2% 29 27.8 28 1% 39.6 47.9 47.6 1%

PAGE 46

46 Table 3 4. Targeted heights, weight, and circumferential data for male pediatric phantoms.

PAGE 47

47 Table 3 4. Continued

PAGE 48

48 Tabl e 3 5. Targeted heights, weight, and circumferential data for female pediatric phantoms.

PAGE 49

49 Table 3 5. Continued

PAGE 50

50 Table 3 6. Summary of organ doses for pediatric female phantoms at a fixed weight of 50 kg, at heights ranging from 135 cm to 175 cm, in 10 cm increments. Dose (mGy) Organ: 135 cm 145 cm 155 cm 165 cm 175 cm Brain 0.22 0.22 0.16 0.15 0.14 Pituitary Gland 0.17 0.16 0.12 0.13 0.10 Lens 0.18 0.13 0.08 0.09 0.08 Eye balls 0.18 0.14 0.10 0.10 0.09 Salivary Glands 0.81 0.74 0.64 0.59 0.55 Oral C avity layer 0.58 0.51 0.72 0.35 0.35 Spinal Cord 4.26 6.34 5.05 5.59 6.68 Thyroid 4.55 3.82 8.22 7.63 7.58 Esophagus 4.27 4.84 5.80 5.81 6.18 Trachea 4.46 5.00 5.98 5.82 6.22 Thymus 4.99 5.28 6.07 5.90 6.27 Lung 5.49 6.11 6.27 6.27 6.71 Breast 4.83 3.69 4.73 4.55 4.28 Heart Wall 5.62 5.42 6.11 6.22 6.10 Stomach Wall 5.20 5.76 6.05 6.86 6.63 Liver 6.06 5.82 6.92 5.80 7.16 Gall Bladder Wall 5.87 5.76 6.22 5.64 7.00 Adrenal 4.77 6.81 6.44 6.40 7.52 Spleen 4.96 6.24 5.55 8.64 7.06 Pancreas 4.81 6.16 5.83 6.96 7.27 Kidney 5.46 5.34 8.43 8.87 10.23 Small Intestine Wall 5.26 5.85 6.31 6.27 6.97 Colon Wall 5.92 6.06 6.57 6.61 7.21 Rectosigmoid Wall 4.74 5.21 5.07 5.27 5.41 Urinary B ladder Wall 5.48 5.01 5.24 5.04 5.66 Uterus 4.44 4.63 4.61 4.38 4.95 Ovaries 4.84 5.13 5.05 5.08 5.41 Skin 3.13 2.85 3.16 3.25 3.35 Residual soft tissue (muscle) 3.45 3.58 3.50 4.03 4.03

PAGE 51

51 Table 3 7 Summary of organ doses for pediatric f emale phantoms at a fixed height of 165 cm at weights ranging from 40 kg to 115 kg

PAGE 52

52 Table 3 8. Percent difference between interpolated and Monte Carlo calculated organ doses for 165 cm phantoms, removing every other phantom (10 kg intervals).

PAGE 53

53 Table 3 9. Percent difference between interpolated and Monte Carlo calculated organ doses for 165 cm phantoms, at 15 kg intervals

PAGE 54

54 Table 3 10. Percent difference between interpolated and Monte Carlo calculated organ doses for 165 cm phantoms, at 20 kg intervals

PAGE 55

55 Table 3 11. Percent difference between interpolated and Monte Carlo calculated organ doses for 165 cm phantoms, at 25 kg intervals.

PAGE 56

56 CHAPTER 3 CONCLUSION AND FUTURE WORK As a result of increasing obesity trends in the United States among pediatrics, an expansion of the current UF/NCI family of reference hybrid phantoms was necessary. The updated grid based library will serve as a comprehensive pediatric library reflecting the most current anthropometric data provided by the CDC. The work presented in this study has been extended to adult male and female populations in the United States, and coupled wit h the p ediatric library, will serve as a useful tool for radiation protection and organ dosimetry purposes. Including the adult male and female library, a total of 358 NURBS phantoms will be created. This is made possible due to the relative ease in which NURBS surfaces can be molded, with the phantom scaling process taking 30 minutes to complete. In the future, these NURBS and polygon mesh based phantoms will be used with MCNPX to create a pre computed dose library for CT examinations that will ultimately be implemented into patient dose tracking programs. Due to large computational run times, interpolation of organ doses using the methods presented in this study will be imperat ive to completing a pre computed dose library for CT examinations. In addition t o CT applications, the extended adult and pediatric libraries can be used in a variety of modalities, including fluoroscopy and nuclear medicine. In fluoroscopy, the reference point air KERMA can be translated to the location of a ent dependent phantoms using skin dose mapping to visually monitor skin dose during interventional fluoroscopy procedures. Implementing skin dose mapping in the future in real time would allow physicians to modify behavior when skin doses approach levels a ssociated with deterministic risks of radiation. Many patients

PAGE 57

57 undergoing interventional fluoroscopic procedures deviate significantly from 50 th percentile reference weight values, making the availability of a comprehensive library of patient dependent com putational models crucial for accurate dose assessments. In addition to their applications in radiation protection and dosimetry studies, the library of patient dependent phantoms can be used in the future to further increase the specificity of phan tom modeling on a patient by patient basis. Through internal organ sculpting, a closest matched patient dependent phantom chosen from the extended library can be altered to create a patient sculpted phantom unique to an individual when increased accuracy i s needed. Though this sculpting process can be done manually, the possibility of automation through the use of Rhinoceros script language can be explored in the future to expedite the sculpting process.

PAGE 58

58 APPENDIX COMPUTED TOMOGRAPHY MODELING AND SIMULAT ION The following write up was provided by Daniel Long (MS Student Medical Physics, ALRADS) from the University of Florida for the CT portion of the simulation. Computed Tomography Scanner Description A SOMATOM Sensation 16 multi detector CT scanner (Siemens Medical Solutions, Erlangen, Germany) was used as the basis for the Monte Carlo source term as well as used for all physical dose measurements. The scanner contained 16 rows of detectors that allowe d beam collimations from 10 24 mm, had the ability to scan in both axial and variable pitch helical modes, and featured two settings of inherent filtration for head and body scanning along with a single bowtie beam shaping filter. The fan beam angle was 52 with a focal spot to isocenter distance of 57 cm. The operator could select tube potentials of 80, 100, 120, or 140 kVp along with varying tube current and gantry rotation speeds. The scanner also allowed for use of tube current modulation during scanni ng, but this feature was not used for the measurements in this study. Modeling of the C omputed T omography Scanner X ray Source The source term of the scanner was created as a custom source file within a general purpose Monte Carlo radiation transport cod e, MCNPX2.6.23 Material and thickness of the two inherent filters for head and body scanning was obtained from the manufacturer. This information, in conjunction with a commercial spectrum generation program, SPEC78 ( Institute of Physics and Engineering in Medicine) was used to create x ray spectra for the head and body filters for all beam energies 11 These spectra were incorporated into the MCNPX source term through an input deck for energy sampling. To account for the effects of the bowtie filter on the shape of the fan beam for all beam

PAGE 59

59 energy/filter combinations, angular dependent weighting factors were applied in the source term based on free in air lateral dose profile measurements made previously by 14 a pencil ion chamber while the x ray tube wa service mode. 12 The effects of overbeaming on the true beam thicknesses for various collimation settings were previously quantified in studies using radiographic film, and were thus taken into a ccount within the source term. 13 The custom source term allowed the user to simulate both axial exams and helical exams with varying pitch. For helical exams, this was accomplished by having the source first sample the location of the x ray focal spot along a mathematically desc ribed helix based on the pitch and scan length selected by the user as well as the previously defined focal spot to isocenter distance of 57 cm. For axial exams, the focal spot location would be sampled along a series of circular rings of radius 57 cm span ning the total specified length of the scan that were spaced apart by the distance of the beam collimation selected. After selecting this starting location of the x ray focal spot, the source sampled an angle within the 52 fan beam as well as within the b eam collimation thickness, therefore selecting a directional path upon which the photon would initially travel. Finally, the photon energy was sampled based on the energy spectrum selected for the exam. This process would be repeated for the total number o f particles to be transported as selected by the user. The final version of the source term allowed the user to select beam energy, head or body filtration, beam collimation, an axial or a helical exam with associated pitch, and starting angle of the beam for helical exams. Computational Phantom Dose Measurements Organ dose calculations using the Monte Carlo CT source term were performed using 20 hybrid computation phantoms. The phantoms were voxelized to resolutions of

PAGE 60

60 3 x 3 x 3 mm 3 (adult and 15 year ol d phantoms) 2 x 2 x 2 mm 3 (10 year old phantom) for input into MCNPX for dose calculations. Within the Monte Carlo simulation setups, fiber patient table. The table was modeled as two concentric cylinders of different radii that wer e truncated to a width of 40 cm corresponding to the width of the patient table, with carbon fiber as its material composition. The source term was then set up to reflect the parameters of each physical scan. For all helical scans, the beam starting angle was assumed to be 0. Organ doses were then calculated using the F6 KERMA approximation tally in MCNPX; therefore, no secondary electrons were transported in the calculation. Considering the energy range of the CT x ray spectra and the subsequent ranges of secondary electrons in tissue, the KERMA approximation offers an acceptably accurate approximation of average organ dose. For all computational calculations, 100 million particles were transported. Since MCNPX provides calculation results in dose p er simulated photon, the number of photons delivered by the scanner per unit mAs, called the Monte Carlo normalization factor, were multiplied to the MCNPX dose results to obtain organ doses in absolute units. The normalization factors were calculated base d on the ratio of pencil ion chamber measurements in free in air (mGy/mAs) to MCNPX simulated free in air ion chamber doses (mGy/pho ton) made in previous studies. 12 Absolute organ doses for each individual scan could then be calculated by multiplying the d ose in mGy/mAs by the total mAs delivered during the exam.

PAGE 61

61 LIST OF REFERENCES 1 H Hricak D. J. Brenner, S. J. Adelstein, D. P. Frush, E. J. Hall, R. W. Howell, C. H. McCollough, F. A. Mettler, M. S. Pearce, O. H. Suleiman, J. H. Thrall, and L. K. Wagner, Manag ing Radiation Use in Medical Imaging: A Multifaceted Challenge Radiology 258, 889 905 (2011) 2 NCRP, NCRP Report No. 160: Ionizing Radiation Exposure of the Population of the United States NCRP Report No. 160. (National Council on Radiation Protect ion and Measurement, Bethesda, MD, 2009) 3 D. Broggio, J. Beurrier, M. Bremaud, A. Desbr e, J. Farah, C. Huet, and D. Franck, 56 7659 7692 (2011) 4 Human Phantoms for Radiation Protection Dosimetry: Anthropometric Data Representing Size Distributions of Adult Worker Populations and Software Medicine and Biology 55 3789 3811 (2010) 5 V. F. Cassola, F. M. Milian, R. Kramer, C. A. B. de Oliveira Lira, and H. J. Khoury, Adult Human Phantoms Based on 10 th 50 th and 90 th Mass and Height Physics in Medicine and Biology 56 3749 3772 (2011) 6 and Male Adult Human Phantoms Based on Polygon Mesh Surfaces: Part I. 55, 4399 4430 (2010) 7 ational 119 149 154 8 Physics in Medicine and Biology 55 339 363 (2010) 9 P. Johnson, S. R. Whalen Patient Dependent Phantoms Covering Statistical Distributions of Body 97 2060 2075 (2009) 10 Centers for Disease Control and Prevention (2011, April 21, 2011). "Overweight and Obesity." Data and Statistics. R etrieved December 12, 2011 from http://www.cdc.gov/obesity/childhood/data.html.

PAGE 62

62 11 K. Cranley, B. Gilmore, diagnostic X Engineering in Medicine (1997). 12 C. Lee, K.P. Kim, D. Long, R. Fisher, C. Tien, S.L. Simon, A. Bouville, W.E. Bolch, rgan doses for reference adult male and female undergoing computer 38 1196 1206 (2011). 13 R. J. Staton, C. Lee, C. Lee, M. D. Williams, D. E. Hintenlang, M. M. Arreola J. L. Biology 51 5151 5166 (2006).

PAGE 63

63 BIOGRAPHICAL SKETCH Amy Marie Geyer was born in Newport Bea ch, California to Bill and Pat Geyer After spending the majority of her childhood living in Idaho, she attended a sports academy in Bradenton, Florida and graduated high school from The Pendleton School in 2004. She graduated with her Bachelor of Science in n uclear engineering s ciences from the University of Florida in May 2010 and graduated with her Master of Science in biomedical engineering with a specialty in medical p hysics from the University of Florida in May 2012. Upon graduation, she began pursuin g her Doctorate of Philosophy in the same field.