Predicted radiation dose and risk associated with pediatric diagnostic x-ray procedures

MISSING IMAGE

Material Information

Title:
Predicted radiation dose and risk associated with pediatric diagnostic x-ray procedures
Physical Description:
xv, 232 leaves : ill. ; 29 cm.
Language:
English
Creator:
Hintenlang, Kathleen M., 1964-
Publication Date:

Subjects

Subjects / Keywords:
Environmental Engineering Sciences thesis, Ph.D   ( lcsh )
Dissertations, Academic -- Environmental Engineering Sciences -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1998.
Bibliography:
Includes bibliographical references (leaves 223-231).
Statement of Responsibility:
by Kathleen M. Hintenlang.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 030019895
oclc - 41093755
System ID:
AA00012983:00001


This item is only available as the following downloads:


Full Text











PREDICTED RADIATION DOSE AND RISK ASSOCIATED
WITH PEDIATRIC DIAGNOSTIC X-RAY PROCEDURES













By


KATHLEEN M. HINTENLANG


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY



UNIVERSITY OF FLORIDA


























Copyright 1998

by

Kathleen M. Hintenlang



























This dissertation is dedicated to my daughter, Lauren Lea Hintenlang, and my son,

Hunter Austin Hintenlang, for bringing the goals of this research that much closer to

home. To my husband David, je t'aime.














ACKNOWLEDGMENTS


The undertaking of this dissertation reminded me of the following koan:
When the mouth wants to speak about it, words fail.

When the mind seeks affinity with it, thought vanishes.

I would like to express my deepest appreciation and gratitutde to the following

people who made the thoughts and words come together: to Drs. Larry Fitzgerald and

Tom Crisman at the University of Florida for their unending support, to Dr. Randy Carter

in Statistics, without whom statistics would be numbers with no physical meaning, and to

Drs. Emmett Bolch and Jon Williams for their continual guidance. Again and always,

immortalis ago gratias vobis agamque. I would also like to thank the Shands at UF

radiology staff for their encouragement and assistance. Additional recognition is due Drs.

Wes Bolch and David Hintenlang; with their assistance, my idea germinated into a seed

and then sprouted many branches. I would like to acknowledge the support I received

from the Children's Miracle Network and the Shands Medical Guild. Sincere appreciation

is extended to my director in the EH&S Division, Dr. Bill Properzio, and to my assistant

director, Don Munroe, not only for the professional development but also for the growth

of a fast friendship. Heartfelt thanks go to my own staff in the Health Center office.

I would like to extend my greatest thanks to my family: Mom and Dad, Tom and

Rachael, Rob and Terry, and Mike and Venessa, for their continued love and support.

Finally, my most sincere and personal thanks go to my husband, David, for his

loving patience, understanding and staunch encouragement.

To quote the words written on a monastery wall in the Middle Ages: "The book is

finished. Let the writer play."















TABLE OF CONTENTS




page


ACKNOWLEDGMENTS ................................. ................................iv

LIST O F TA BLE S ....... .................... ................................................. ........... viii

L IST O F FIG U R E S ............................................................ ............ ............ ....... xi

ABSTRACT................ ............. ............... ........................... xiv

CHAPTERS

1 INTRODUCTION AND BACKGROUND............ ......... ............ ........... 1

G general ...................... ............. .... ............ ......................... 1
Pediatric Radiology as a Subspeciality ........................... ................. ..... ........3
R atio n ale .................. .... ..... .......... .......................................................... ..... .... ..6
Adult Organ Doses in Plain Film Radiography ........................ ........ ...... 8
Pediatric Organ Doses in Plain Film Radiography ........................................... 9
Calculation of Energy Imparted in Diagnostic Radiology............................ .. 11
P h an to m s ........... .... .. .... ........ .................. ............ .................. ....... ... 13
Mathematical ....... ................................... ...................... .... 13
Anthropomorphic............................................................. ... 14
Radiation D election Instrum entation ............................................... ................. 16
G as Filled Ionization C ham bers .............................................. ......... ........ 16
Sem iconductors............................. ................ 16
D term nation of R isk ............................................................ .. ....... 18
Significance and Objectives...................... ............................ ............... ........... 19

2 PATIENT AND EXAMINATION TRENDS OF PEDIATRIC X-RAY
P R A C T IC E S ....................................... .. .. .. .. ...... ..... .. .. .. ...... ................ ..... 2 2

Shands at the U university of Florida................................................ .... ......... ...... 22
Ten Florida Facilities................. ..... ....................................... ............ 26
E xam Sim ulations........................................................................ ..................... 28









A P sk u ll ............................................. ..... ..... 3 0
L ateral skull ......... ....... .. ............... ... .. 31
T ow nes skull ............................ .. ............ .. ......... ...... ................ 32
W aters sinu s .......... ............ .. .. .. .. ... .. .. ... .. ..... .. .. ..... ..... .. 3 3
A P cervical spine .......... ................... ... .. .. ..... .... 33
Lateral cervical spine............. ................ ........................................... 33
AP thoracic spine .................................. ....................... .. ....... .. 34
Lateral thoracic spine ........................................... ................................ .. 34
AP lumbar spine ..... .. .... ... ............ ........................ 35
Lateral lumbar spine ....................................................... ......... .......... .... 35
A P abdom en ............................... .. .. .. .......... ..... .. .. ..... .. ..... .. ... 36
A P p e lv is .................................................................................. ...... ... 3 6
A P bilateral hip.................................. ................... .. ..... .. 36
AP, PA and lateral chest .................................... .............. 37
Survey Data ............... .... ..................... ....... ...40

3 RADIATION MEASUREMENT EQUIPMENT AND TECHNIQUES............ 42

One-Year-Old Anthropomorphic Phantom ......................................... ......... ..42
Metal Oxide Semiconductor Field Effect Transistor (MOSFET) Dosimetry
S y ste m .................................. ...................................................... 4 6
Theory of O peration.................................. ................ ... ..................... 46
Characterization of High Sensitivity MOSFET Dosimeter...................... .......48
M O SFET Placem ent .......................................... ....................................... 50
Measurement of Absorbed Dose........................ .... ............. 55
Sensitivity A nalysis.......................... .................................................................. 56
C alculation of E effective D ose ............................................................................. 58

4 R ISK A SSE SSM E N T ............................................................................................. 60

G e n e ra l ...................... ................. ......................... ....................................... 6 0
Com ponents of a Risk M odel ................................................................... ..... 60
BEIR V M ethodology................................................... ................ 62
Populations for Determining Cancer Risk ...........................................................62
H um an Populations ........................................ ............................... 62
Anim al Populations .................................... .............. 63
Populations for Determining Genetic Risk ..................... ............................. 63
H um an P populations ..................................................................................... 64
Anim al Populations ........................................................... .................... 64
Dose-Response Curves...... .............................. .... ..... 65
D o se-R ate E effects ............................................................................................. 66
Population Transfer Coefficients................................................ ....... ............ 68
BEIR V Cancer Risk Projection M ethod ................................... ....................... 69
Specific C ause of D eath............................................................ ................. 72
A ge of Population ....... .................... ............... .... ..... ......... ...... 72
BEIR V Genetic Risk Projection Method .............................. .......... 73









Calculating the Relative Cancer Risk from Pediatric Diagnostic X-Ray
P ro cedures ........................ ......................... ...... .. ...... .................73

5 RESULTS AND DISCUSSIONS............. .. ...................................... 79

Facility Approach to Handling Special Concerns of Pediatric Patients...................... 79
Patient and Examination Trends........................................ ......... ................. 85
G enerators ............ ...... ................ ..... ...... ........ 93
Exposure Time......................................94
F ocal Spot Size ............................................................. ............... .. 95
A additional Filtration ...................................................... .. ............ 95
G rids ................................. .. .. .. .. .. ....... .......... ............. ... 96
C assettes ........................................ ... ..... .......... ............... 97
Film-Screen Combination ........................ ....... .... ......... 98
Source-Image Distance (SID).................................. ....... ............ ... 98
Automatic Exposure Control......................................................... 98
Field Size and Collim ation ............................................... ......... ..... 99
Radiographic Film Quality ............................................. 100
Repeat Analysis... ...................................................... ................. 100
R radiation Protection................................................................... 102
Organ Doses per Entrance Skin Exposure and Effective Doses............................... 103
Relative Risk..... ......................... .......... ....................... .... ..... ........... 111

6 CON CLU SION S .......................... .... ............ ................ 115

Site Surveys............................................ ................. 115
Phantom M easurements ........................................ ............................ 118
Effective Dose and Risk Predictions .. ......................................... ...................... 120

APPENDICES

A NUMBER OF EXAMS FOR PEDIATRIC PATIENTS UNDER 16 YEARS
O F A G E .. ............................ ............. ... ..... ............ .................... ................. 12 1

B EXAM FREQUENCY FOR PEDIATRIC PATIENTS UNDER 16 YEARS
O F A G E ....................................................... ...................... ...................... 130

C EXAM CHARACTERISTICS DETERMINED FROM FACILITY SITE
SURVEYS ... ................. .......... .................. 136

D AVERAGE MASS ENERGY ABSORPTION COEFFICIENTS........................... 147

E EFFECTIVE DOSE AND RISK CALCULATIONS ............................................ 150

LIST OF REFERENCES............................... ................. 223

BIOGRAPHICAL SKETCH ........................................ 232















LIST OF TABLES


Table page

2-1. Age Group Corresponding to Phantom ................................... ..................... 23

2-2. Annual Number of Plain Film Chest X-Rays by Age Group ..............................24

2-3. Percent Frequency of Plain Film Chest X-Rays ........................................ ........24

2-4. Surveyed Examinations .................................... .. ...................................26

2-5. Defined Locations of Anatomical Landmarks on Phantom Vertical Axis
with the Origin at the Base of the Trunk (cm) ......................................... 30

3-1. Tissue Weighting Factors for Calculation of Effective Dose...............................51

3-2. Position of the MOSFET Dosimeters Within the One-Year-Old Phantom.......... 53

3-3. Percentage Active Bone Marrow and Bone Surface Associated with the
Skeletal M O SFET Locations ................................................... ............... 54

3-4. Sensitivity Calibration Factors (R/mV) at 70 kVp ...................................... 56

3-5. Conversion Factors Converting Exposure to Absorbed Dose............................ 57

4-1. Lifetable Analysis of a Standard Mortality Table..................... ..................... 70

5-1. Effective Doses for Male and Female One-Year-Olds at Shands at UF............. 106

5-2. Comparison of Effective Doses for Male and Female One-Year-Olds
Among Facilities for Chest Exams ............... .................... ................. 107

5-3. Comparison of Shands at UF and FDA Absorbed Dose/ESE........................ 109

5-4. Comparison Among All Facilities and NRPB Effective Dose/ESD.................. 111

5-5. Relative Risk for Shands at UF Examinations ................................................. 112

5-6. Relative Risk for Chest Examinations Among All Facilities............................. 113









C -1. A P Skull .. ................ .. ............ ....... ....................... ........... 137

C -2. T ow nes Skull ................. ........... ........................... .................. 137

C -3 L ateral Sk ull..................................................................................................... 13 8

C -4. A P C ervical Spine .................. ......... .... ........................................... 139

C -5. L ateral C ervical Spine ........................................ ........................ ................ 139

C-6. AP Thoracic Spine ............................................................... 140

C -7. L ateral Thoracic Spine .......................................... .......................... ...... 140

C-8. AP Lumbar Spine............................................ ........... .......... 141

C-9. Lateral Lumbar Spine................................................... 141

C -10. A P A bdom en ...................................................... ....... ............................... 142

C-l A P Pelvis ........................................ 142

C -12 A P H ip ............................................ .. ..... ....... ................... 143

C-13. Waters Sinus... ........................................... 144

C -14. L ateral Sinus..................................... ............ .... ............. 144

C-15. Lateral Chest... .......................... ........... 145

C -16. A P Chest ...................................................... ................ .... ..... ....... 145

C -17. PA C hest........................ .... .. .. ............. ..... .......... .. 146

C-18. Non-Exam Specific Operational Data................................................. ......... 146

E -1. A P Skull ....... .............. ........ .... ........ 151

E-2. A P Skull w ith Thyroid Shield ........................................................................... 153

E -3 P A Skull ............... .. ... ........ .................... ...... .. ............... 155

E-4. PA Skull with Thyroid Shield ............................. ............... ............ 157

E -5 L ateral Skull........... ... ...................... .................................. ................ 159











E-6.

E-7.

E-8.

E-9.

E-10.

E-11.

E-12.

E-13.

E-14.

E-15.

E-16.

E-17.

E-18.

E-19.

E-20.

E-21.

E-22.

E-23.

E-24.

E-25.

E-26.

E-27.


T ow nes Skull ................................... .......................

A P C ervical Spine ...........................................................

Lateral Cervical Spine .......................................................

A P T horacic Spine ................................ .... ...................

Lateral Thoracic Spine ......................................................

AP Lumbar Spine.......................................................

L ateral L um bar Spine........................................................

AP Abdomen Supine........................................................

PA Abdomen Supine...................................

AP Abdomen Upright........................................................

A P P elvis ................ ... ...... .. .. ... .....

A P H ip ........................ ........................................

W aters Sinus ................................ .. .. .............

Lateral Sinus.................

AP Sinus.... ...............................................................

A P C h est................... ...... .. ..... .. .. ..... .. .. .. .. .. .. ..... ..

AP Chest with Thyroid Shield ................................

P A C hest.......... .. ... ..... .. .. .. .. .. .. .. .. .. .. .. ......

L ateral C hest .............. ..... .. .. .. .. .. .. .. ..... .. .. .. .....

L A O C hest.................. ............................

R PO C hest .......................................... ............

R A O C hest ............................................. ................


............................. 161

............................. 163

................ .. 16 6

...................... ....... 169

.................. ......... 172

...................... ........ 17 5

................... ........ 178

.... ........ .............. 181

................. ......... 184

..................... 18 7

........ ....... ....... 19 0

...... ..................... 193

.................... ....... 196

............................... 198

......... .. .. .. ... .. 2 0 0

..... ................ .. 202

.................. ........... 205

....... .. .. ... .. .... .... 2 0 8

S............... ........... 211

.............................. 214

.............................. 217

.. .. .... .. ..... ....... 2 2 0














LIST OF FIGURES


Figure page

2-1. Items frequently used for immobilization. .................................................... 31

2-2. Infant immobilized and supported by the Pigg-o-stat device for an upright
chest film ........ .................... ................ .............38

2-3. Toddler immobilized for an AP chest radiograph on the Tame-Em
im m obilizer................................... .................. ... ................. 39

2-4. Infant immobilized for a left lateral chest radiograph on a Plexiglas
im m obilizer............................... ....... .... .... .......... ................ 39

2-5. Octagon board for immobilization ........................................ ................. 40

3-1. One-year-old pediatric phantom ........................................ ................. 45

3-2. Thomson and Nielsen Electronics Ltd. MOSFET Patient Dose Verification
System ................................. ... .............. ...... ................... 47

3-3. MOSFET dosimeters placed in position in prototype one-year-old phantom...... 52

4-1. Various models used to extrapolate high-dose data on cancer incidence to
the low-dose region, so that risk estimates can be performed ..................... 68

4-2. The risk of leukemia due to low LET radiation as a function of attained age....... 71

4-3. Time dependent risk for exposure at age equal to one year for both male
and female leukemia models................... .......... .......... ................ 75

4-4. Time dependent risk for exposure at age equal to one year for male
respiratory cancer model.......................... .. .. ............... 76

4-5. Time dependent risk for exposure at age equal to one year for female
respiratory cancer model...................... ...... .................... 76

4-6. Time dependent risk for exposure at age equal to one year for male
digestive cancer model................. .. ...........................77









4-7. Time dependent risk for exposure at age equal to one year for female
digestive cancer model......................................................... 77

4-8. Time dependent risk for exposure at age equal to one year for both male
and fem ale other cancer m models ........................................ ...... .. .... ... 77

4-9. Time dependent risk for exposure at age equal to one year for female
breast cancer m odel ....................................... ..................... ....... ......... .. 78

5 -1 A P S ku ll ...................... ............... .. .. .. .. ... .... .. ......... ........ .. .. .. 8 6

5-2 L ateral S kull. ... .. ........ .... ..... .. .. ..... .... .. ..... .. .. ..... ..... ..... .. .. ... 86

5-3. Townes Skull ....... .... ....................... 87

5-4 A P C ervical Spine ................................................................................ ........ 87

5-5. L ateral C ervical Spine .................................................................... ............. 88

5-6 A P T horacic Spine ................................... ........................................................ 88

5-7. L ateral T horacic Spine .................... .............................. ...................... 89

5-8. A P L um bar Spine .................................................................. ........................ 89

5-9. L ateral L um bar Spine................. ................................... ............................... 90

5-10. A P A bdom en ........................................ ........................................ 90

5-11. A P Pelvis .. ..... ........ ... ........ ........................................... 91

5-12 A P L eft H ip ...................................................................................................... 9 1

5-13. W waters Sinus... .............................. 92

5-14. L ateral Sinus ............. ........... ........... ............ ............... 92

5-15. A P C hest................................. .. ....... .. .. .. ........... ... .. .. .. ...... .. .. .. 9 3

5-16. Lateral Chest..... ........ .. ................ 93

5-17. Range of kVp used for the 17 surveyed exams................. ............. ........ 94

5-18. Range of mAs used for the 17 surveyed exams ................. ......................... 95




xii









5-19. R ange of H V L ...................................... ........................................ 96

5-20. Net lung optical densities for AP chest, PA chest and lateral chest films............ 100

5-21. R epeat percentages per facility ............................................ ...................... 101




















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


PREDICTED RADIATION DOSE AND RISK ASSOCIATED
WITH PEDIATRIC DIAGNOSTIC X-RAY PROCEDURES

By

Kathleen M. Hintenlang

December 1998

Chairman: W. Emmett Bolch Jr.
Major Department: Environmental Engineering Sciences


The most effective method of assessing the risks from exposure to radiation is by

calculating the individual tissue absorbed doses and subsequently determining the effective

dose. The absorbed dose to a specific tissue has traditionally been difficult and time

consuming to measure. The effective dose and overall risk are therefore difficult to

calculate directly. The problem is compounded in pediatric radiology, where the patient is

particularly sensitive to radiation effects, and the body and organ sizes differ greatly from

the common adult reference man.

This research developed and applied a new methodology that permitted rapid

measurement of the effective dose from x-ray radiation, and the associated risks, to

pediatric patients for a variety of plain film diagnostic examinations. The methodology










utilizes an anthropomorphic phantom incorporating direct reading MOSFET dosimetry,

allowing the rapid and simple determination of absorbed dose to individual organs. The

research integrated clinical examinations, dosimetry measurements and BEIR V risk

assessment calculations using a prototype, anthropomorphic phantom of a one-year-old to

quantify radiation doses delivered to pediatric patients in clinical radiology exams. Site

surveys were performed in ten facilities using several standard exams to determine

variations in effective dose related to clinical practice. Recommendations resulting from

the site surveys were provided for the optimization of pediatric radiography procedures.

Quantitative studies for a variety of examinations were performed at Shands at UF. The

effective dose for these procedures ranged from 0.1 mrem for an AP skull exam that

utilized thyroid shielding, to 9.2 mrem for an AP upright abdomen exam. The relative

leukemia carcinogenic risk ranged from 1.000041 for the skull exam, to 1.002732 for the

abdomen exam, with corresponding excess risks from 106 exams of 0.005 and 0.3

leukemias per year, respectively. Effective doses and risks for various cancers from all

surveyed exams were quantified. These individual examinations provided a significant

benefit to the patient with minimal risk. This provided the basis for continuing research to

quantify long-term risk assessment to pediatric patients, as well as providing a clinical tool

to evaluate the effective dose delivered to these patients from other current and emerging

radiology imaging modalities.














CHAPTER 1
INTRODUCTION AND BACKGROUND


General

It is well established that diagnostic x-rays constitute the largest and most widely

distributed source of man-made radiation exposure to the general population (NCRP

1987, 1989). While infants and children constitute only about 10% of the total number of

radiographic examinations (NCRP 1989), they are the segment of that population at

higher risk from potential radiation effects. First, their growing tissues are more

susceptible to radiation effects than mature adult tissues (BEIR 1990). Second, their

skeletons encompass a greater fractional distribution of active bone marrow, an organ of

high radiation sensitivity (ICRP 1995). Third, the greater post-exposure lifetimes of

infants and children increase the possibility for any radiation-induced effects to manifest.

Fourth, children are known to have short attention spans and can be non-cooperative; as a

result, they are subject to a greater number and/or longer exposures than adult patients.

Finally, pediatric patients frequently have a larger portion of their anatomies located within

the x-ray beam than adults in similar exams and projections.
In the course of their care, these young patients may be exposed to a wide variety

of x-ray examinations. Conventional radiographs make up three quarters of pediatric

radiological examinations performed (Griscom 1996). These examinations yield medical
benefits and diagnostic information which must be balanced against potential risks from

patient radiation exposure. With ever increasing improvements in clinical care, survival

rates for prematurely born infants have risen over the past decade. As these individuals

age, they frequently experience additional complications requiring a large number of x-ray










examinations from a variety of imaging modalities. Consequently, these patients may

accumulate large radiation doses, the effects of which have not been evaluated since few

survived in earlier years. Technological improvements that permit these newborn infants

and young children to survive increase the need for clinical decision-making based on

4quantitative assessments of radiation risk.

The fundamental quantity of interest from these examinations is the cumulative

absorbed dose to exposed organs. Several studies have focused on average organ doses in

adult patients undergoing radiographic procedures, either through computer simulation

(Rosenstein 1976a, 1976b, 1988; Jones and Wall 1985; Rosenstein et al. 1992; Stern et al.

1995) or physical measurement (Shleien 1973; Ellis et al. 1975; Chen et al. 1978).

Nevertheless, very little current information exists on organ doses from pediatric

radiographic projections, particularly for technique factors common to U.S. practice (Beck

1978, 1979; Rosenstein et al. 1979; Zankl et al. 1989; Hart et al. 1996).

The desire to perform a study of the current patient and examination trends in

pediatric x-ray practices in the State of Florida and incorporate a direct and simple method

to measure organ and effective doses in a pediatric phantom in order to perform a risk

assessment was the basis for this research. An epoxy resin based anthropomorphic

phantom of a one-year-old child constructed in companion research to this dissertation

was utilized (Bower 1997). Site visits were performed at ten hospitals in the State of

Florida. Simulated examinations of a series of plain film radiographs were performed by

pediatric x-ray technologists on the phantom. Real-time Metal Oxide Semiconductor

Field Effect Transistor (MOSFET) dosimeters were inserted into various organ locations

in the phantom to obtain point dose measurements of the organ absorbed dose and

effective dose measurements in real-time. Subsequent risk assessment calculations were

performed utilizing BEIR V methodologies.

In order to facilitate the development and transfer of scientific information for the

improvement in the radiologic care of children, the methods developed in this research can









then be utilized to integrate the computational and advanced experimental components and

determine organ doses and assess risks to various ages of pediatric patients undergoing a

variety of modern diagnostic examinations to include fluoroscopy, computed tomography,

cardiac catheterization, computed radiography and digital radiography.


Pediatric Radiology as a Subspeciality

The history of pediatric radiology is the story of the emergence of pediatrics and

radiology as individual medical specialties, the growth of each of these specialties, and

finally the fusion of these two specialties into the new subspecialty of pediatric radiology,

or Roentgenology, as it was called in the early days.

Roentgen's discovery of x-rays in November and December 1895 was announced

to the world in January 1896. On February 3, 1896, a fourteen-year-old boy who was

thought to have broken his wrist two weeks earlier while skating on the Connecticut

River, was brought to Dartmouth College in Hanover, New Hampshire, because they

supposedly had one of the best collections of vacuum tubes in the western hemisphere at

that time (Spiegel 1995). His wrist was radiographed, with an exposure time of twenty

minutes, and the resulting image on the emulsion-coated glass plate "showed the fracture

in the ulna very distinctly." In addition to what is considered to be the first clinical

radiograph in America, it was also the beginning of pediatric radiology in this hemisphere.

In March of 1896, Dr. E. P. Davis of New York City reported visualizing the trunk

of a living infant and the skull of a dead fetus and also recorded images of a fetus

examined in utero (Davis 1896). Shortly thereafter, the first recorded observation of

disease in children probably belonged to Professor Cox of McGill University, Montreal,

Canada, who used a thirty-minute exposure to find a needle embedded in the wrist of a girl

of unstated age (Cox 1896). William G. Morton, at a meeting of the County of New York

Medical Society held on April 27, 1986, reported that he had already seen the fetal head in










utero and predicted that the time was not far distant when the sex of the fetus in utero

would be disclosed Roentgenographically (Caffey 1956). At this meeting, Morton showed

lantern slides of an entire fetal skeleton and used the terms "inside-seeing or esography"

for Roentgen visualization. In July, a publication listed conditions that the author had

successfully shown by x-ray, including pediatric cases of dactylitis and absence of the radii

(Codmam 1896).

Radiation injury to children apparently was recognized within a few weeks after

the Roentgenology method was instituted, as Dr. W. V. Gage (1896, page 307) of

McCook, Nebraska, reported "erythema and finally sloughing of the region penetrated by

the rays, which at present is the size of the hand" in a child who was being studied for a

foreign body in the stomach. Although this episode was reported in the Medical Record

of August 29, 1896, Caffey stated that it is obvious that the actual Roentgen injury must

have occurred many weeks earlier during the first months of 1896 (Caffey 1956). It is

interesting to note that the spread of Roentgen's discovery was not just limited to large

cities, but was also utilized within months of his publication in a community as medically

remote as McCook, Nebraska, on the prairie which was still "frontier country" at that time

(population 2,346 by the 1890 census). In many of the early reports, there appeared to be

a disproportionate prevalence of pediatric cases, probably because x-rays from the

available tubes at that time could not penetrate larger subjects. In addition, a majority of

the researchers in the initial trial period were not physicians. Sosman (1951, page 552)

observed that "by 1900, the x-ray method was being used by physicists, engineers,

photographers, some charlatans and a few honest physicians". Caffey added bicycle repair

men and the entrepreneurs of circus sideshows to this observation.

After this initial burst of enthusiasm after Roentgen's discovery, Caffey reports

that American pediatric radiology lagged significantly behind the growth of both radiology

and pediatrics and failed to progress until the 1930s, as demonstrated by the limited

number of publications. In the Transactions of the American Roentgen Ray Society,










number of publications. In the Transactions of the American Roentgen Ray Society,

established in 1901, no pediatric subject was discussed until 1904, when Preston M.

Hickey of Detroit reported his Roentgenologic study of skeletal maturation (Hickey 1904,

1906). In addition to the decreased number of studies being reported, the majority of the

case studies published represented superficial subjects. Caffey reported that in 1903,

Volume 20 of the Archives of Pediatrics published a total of only four papers on

Roentgenologic subjects. The most interesting one, Caffey felt, presented Roentgenologic

data by A. C. Cotton purporting to show that wadding of thick diapers between the thighs

produced lateral bowing of the femurs. Caffey also reported that in the first volume of the

American Journal of Diseases of Children in 1911, a study by Long and Caldwell on the

relationship between carpal ossification and mental development was performed with

Roentgen prints of the paired hands of twenty-nine subjects, in which the authors

concluded that "we can find no relation between the carpal development and the quality of

mind" (Caffey 1956, page 440). Caffey noted, with some sarcasm, that this conclusion

has stood the test of time. Griscom indicated that the focus was more on techniques and

equipment at that time than on the medical components of the radiology of children

(Griscom and Jaramillo 1995).

The 1930s, 40s and 50s showed a drastic increase in the number of x-ray studies of

large numbers of healthy children from birth through adolescence. Due to the lack of

normal standards, Dr. Arial George (1908, page 381) explained that "A thorough

knowledge of the normal human anatomy is absolutely necessary, as are its variants. The

whole study of disease in children by the Roentgen method rests on this one point. The

mistakes in diagnosis by the Roentgen ray method will not be the fault of the Roentgen

ray, but the fault of those who are ignorant, and who from lack of training are unable to

interpret its findings ." As a result, Boyd performed studies of the normal thymus (1927).

Lincoln, Hodges, and Josephi investigated cardiac size in healthy infants (Lincoln and

Stillman 1928; Hodges 1933; Josephi 1935). Bouslog and Henderson studied the normal









infantile gastrointestinal tract (Bouslog 1935; Henderson 1942). The normal neonatal

skull was studied Roentgenographically by Henderson and Sherman (1946). Rotch called

attention to the epiphyses, as studied by x-rays, to indicate the state of bodily maturity in

normal standards of growth and development for various age groups (Smith 1951).

Caffey published several norms of the skeleton (Caffey 1956).

All of these early attempts to establish normal values were open to criticism as they

were based on studies of groups of infants and children. It was felt that more accurate

values could be obtained by longitudinal studies in which the same groups of children were

investigated as they progressed from birth to maturity. So, from 1925 through 1940,

Maresh and Washburn performed serial Roentgenographic observations on one hundred

normal infants and children to demonstrate the variation in the size and shape of the

respiratory track (Maresh and Washburn 1940). Griscom indicated that at the second

meeting of the Society of Pediatric Radiology in 1959, sixteen case studies were presented

(Griscom 1995). About half dealt with plain films, the remainder with intravenous

urography, barium studies, nuclear medicine and dosimetry. This was the first instance

where dosimetry for pediatric x-rays is mentioned in the literature. Dr. Donald Darling,

from the Children's Hospital in Pittsburgh, presented a progress report entitled

"Measurement of Gonadal Dose in Routine Pediatric Radiological Practice." Dr. Darling

indicated that results were never published from this study (personal communication,

1998).


Rationale

The organs to be considered in radiologic risk assessment were standardized by the

International Commission on Radiological Protection in 1977 with the introduction of the

effective dose equivalent (EDE) (ICRP 1977). Calculations of the EDE require

knowledge of the absorbed dose (total energy deposited per unit mass) to the following









tissues: breasts, lungs, reproductive tissues (ovaries or testes), thyroid, active bone

marrow, skeletal endosteum and unspecified remainder tissues. These absorbed doses, or

more specifically, dose equivalents,1 are then multiplied by tissue weighting factors (wr)

which represent the fraction of total radiation risk (cancer mortality and severe genetic

damage) attributed to irradiation of that tissue. The EDE is calculated as a summation of

these weighted doses and thus represents a single radiation dose proportional to the total

radiation risk of the exposure, regardless of whether the irradiation is uniformly or

nonuniformly delivered.

This dosimetry concept was further expanded in 1990 with the introduction of the

effective dose (E) in which tissue weighting factors were reassessed based on a more

recent analysis ofradiobiological effects (ICRP 1991). Additional weighting factors were

also given for the esophagus, stomach wall, colon, liver, skin and urinary bladder wall.

While originally defined for radiation protection purposes, these quantities of EDE and E

have been widely reported in the medical literature for both nuclear medicine procedures

and for diagnostic radiology examinations. Their intended use in medicine is to provide

physicians with a unified quantity for risk communication and a quantitive means for

procedure optimization. As defined in ICRP Publications 26 and 60, the tissue weighting

factors are specific only to populations of adults and should not be applied in the

estimation of radiation risk to individual patients (Poston 1993). For use in risk

communication and/or procedure optimization, tissue weighting factors specific for

children have been proposed by Almen and Mattsson (1996).

The determination of organ doses in diagnostic radiology is typically a two-step

process. First, an indicator dosimetry quantity is measured in the clinical setting.

Examples include the entrance skin exposure (ESE), the entrance skin dose (ESD) or the

dose-area product (DAP). Second, these indicator quantities must be multiplied by an

1 Dose equivalent is defined as the product of an absorbed dose and a radiation weighting
factor which for photons is set to unity.









organ dose coefficient to obtain individual organ doses (i.e. mrad per R) (Rosenstein

1988). These conversion factors are obtained either through experimental measurement

within physical phantoms or through computer simulations within mathematical models of

the patient. The specific procedures involved in obtaining the conversion factors utilized

in this research are discussed in Chapter 3, Radiation Measurement Equipment and

Techniques.


Adult Organ Doses in Plain Film Radiography

The first systematic study of organ dose coefficients was published by Rosenstein

in 1976 in Food and Drug Administration (FDA) Reports 76-8030 (Rosenstein 1976b)

and 76-8031 (Rosenstein 1976a). In this study, the Fisher-Snyder mathematical model of

the adult male was modified for use in determining organ doses for a variety of diagnostic

x-ray projections. This model was previously developed for estimates of internal photon

absorbed fractions needed for nuclear medicine dosimetry (Snyder et al. 1978). The

model consisted of geometrical descriptions of eighteen internal organs of the adult and

included three distinct tissue compositions: soft tissue, skeleton (a homogenized mixture

of bone and marrow) and lung (a homogenized mixture of soft tissue and air). While the

model was developed as a description of Reference Man defined in ICRP Publication 23

(1975), it was a hermaphrodite in that it included both testes and a uterus and ovaries.

In order to reduce computational requirements, the following approach was

utilized. Organ doses were first determined for parallel, collimated beams of

monoenergetic photons incident on the surface of the mathematical model. Each photon

beam was situated on one grid of a system of 4 cm x 4 cm grid elements superimposed

upon the front surface of the model (AP projection), back surface of the model (PA

projection), or on a plane just inside the arm bones (lateral projection) to simulate the

actual conditions of exposure in which the arm is generally moved out of the primary









beam. Organ doses for a given examination (e.g. AP chest) were determined by first

summing the individual dose contributions from each grid element contained within the

field-of-view, and then energy weighting these results for the x-ray spectrum of interest.

By using this procedure, beam divergence and air scattering effects were ignored and

oblique views were not accommodated. These data were subsequently used to produce a

handbook providing organ doses per unit entrance skin exposure free-in-air (mrad/R) to

the testes, ovaries, thyroid and active bone marrow for sixteen projections and six beam

qualities (HVLs of 1.5 to 4.0 mm Al). The field size was assumed to equal the film size in

this study. Limited experimental data were collected as part of the study using lithium

fluoride thermoluminescent dosimeter's (LiF TLDs) in anthropomorphic phantoms.

A revision to this handbook was published in 1988 which expanded the number of

projections considered to fifty-four. In addition, Cancer Detriment Indices were also

reported for each exam based upon risk data contained in ICRP Publication 45 (1985). In

the early 1980s, Gesellschaft fir Strahlen und Umweltforschung (GSF) in Germany

(Kramer et al. 1982; Drexler et al. 1984) developed their own male and female adult

phantoms and the British National Radiological Protection Board (NRPB) (Jones and

Wall 1985) further developed the Cristy (1980) modified versions of the original Snyder

Medical Internal Radiation Dosimetry (MIRD) mathematical model (Snyder et al. 1969) to

assess organ doses to adults in diagnostic radiology.


Pediatric Organ Doses in Plain Film Radiography

A comprehensive and systematic investigation of organ dose coefficients for

pediatric radiographic projections was published by Beck and Rosenstein in 1979 as FDA

Reports 79-8078 (Beck 1978, 1979) and 79-8079 (Rosenstein et al. 1979). This study

included a survey of five hospitals in the Baltimore-Washington area to determine the

frequency and types of pediatric radiographic examinations performed and the various










technique factors employed in those exams such as kVp, mAs, field size, SID and beam

quality. The survey data were subsequently used to perform computer simulations of the

more common projections following methods similar to those employed in the adult

report.

Three mathematical models developed at Oak Ridge National Laboratory by

Hwang et al. were utilized representing a newborn, a one-year-old, and a five-year-old

individual (1976a, 1976b, 1976c). Twenty examinations were considered, each

representing at least one percent of the total number of exams in a given age group. For

each exam, organ dose coefficients (mrad/R) were determined for testes, ovaries, thyroid,

active bone marrow, lungs and total body at one representative source-to-image receptor

distance, two field sizes (age dependent) and four beam qualities (HVLs of 2.0, 2.5, 3.0,

and 3.5 mm Al). As with the adult report, limited experimental data were collected as part

of the study using LiF TLDs in anthropomorphic phantoms representing a head and neck,

and two phantoms corresponding to the Monte Carlo mathematical design: a one-year-

old and a five-year-old. Given the experimental uncertainties in the TLD measurements

and the limited number of photon histories transported (60,000), the authors concluded

that good agreement was seen between the two data sets.

To date, no revisions have been made to FDA Reports 79-8078 and 79-8079, and

these data are still widely used for dose estimation in pediatric radiology. Several

arguments can be made for revising and expanding the research of Beck and Rosenstein.

First, while mathematical models did exist for a representative ten-year-old and fifteen-

year-old patient, these two groups were notably absent in the final handbook. Second, the

input photon spectra used in these reports were for single phase generators, while current

radiology practice employs primarily three-phase, constant potential, and high-frequency

x-ray units. Third, while the pediatric models of Hwang et al. included twenty-six internal

organs, dose estimates were given only for the six organs listed above, thus making

calculations of EDE or E impossible. Fourth, substantial revisions and improvements have










been made to the ORNL pediatric model series over the past eighteen years, thus making

the dosimetry of FDA reports 79-3030 and 79-3031 obsolete and somewhat inconsistent

with current practice in radiology and in radiation protection (Cristy 1980, 1987).

In the Hwang et al. 1976 series of pediatric models, a large number of internal

organs were simply scaled downward from the Fisher-Snyder adult model. In 1980, Cristy

reported that this procedure resulted in "shifting and crowding" of various internal organs

in a manner which did not reflect actual anatomical growth trends (Cristy 1980). As a

result, a complete redesign of the ORNL pediatric model series was initiated, resulting in

the publication of new pediatric models by Cristy and Eckerman in 1987. These models

were subsequently adopted for radiation protection dosimetry by the ICRP in 1989 (ICRP

1989) and for nuclear medicine dosimetry in 1988 by the MIRDOSE computer code

maintained by the Radiation Dosimetry Information Center in Oak Ridge, Tennessee

(Stabin 1996). These models were the basis upon which a new MIRD family of

mathematical models are being developed at the University of Florida in companion

research to this dissertation.

Finally, no new pediatric anthropomorphic phantoms have been developed that

correspond to the changes made in the MIRD mathematical models. The only other

comprehensive source of organ doses for pediatric plain film exams is that produced by

the NRPB in 1996 for technique factors specific to the United Kingdom and European

Community utilizing the older 1987 version of the Cristy and Eckerman mathematical

models (Hart et al. 1996).


Calculation of Energy Imparted in Diagnostic Radiolovg

One may also estimate radiation risk for diagnostic procedures by measuring the

total energy imparted to the patient for a given diagnostic exam (the integral dose).

Extensive studies of energy imparted in diagnostic radiology have been performed by










Huda and colleagues (Huda et al. 1989; Huda and Bissessur 1990; Atherton and Huda

1995, 1996; Huda and Atherton 1995; Gkanatsios and Huda 1997; Huda and Gkanatsios

1997). Values of energy imparted, however, in and of themselves do not provide

important information on individual organ doses for a given projection or exam. In his

publications, Huda proposed the use of approximate ratios of effective dose per unit

energy imparted which, in combination with energy imparted on homogeneous phantoms,

can be used to derive an estimate of the effective dose. These ratios must also be obtained

from previous Monte Carlo studies such as those performed by the NRPB (Shrimpton et

al. 1991; Hart et al. 1994, 1996) which use heterogeneous anthropomorphic models of

patients. Huda has recently proposed a simple mass scaling of these ratios to yield

effective doses to children undergoing CT exams (Huda et al. 1997). An important

limitation of this approach is that individual organ doses can never be recovered from a

single value of effective dose and the risk from different diagnostic procedures cannot be

compared when different organs are exposed.

It must be remembered that the effective dose is a weighted average of the

individual organ doses where the tissue weighting factors are chosen based on the current

knowledge of radiation cancer risk and other detriment. This knowledge base changes

with time and with changes in cancer mortality; thus the tissue weighting factors are

influenced by changing success rates in cancer treatment. Consequently, the tissue

weighting factors are always subject to change, whereas the individual organ dose for a

specific diagnostic procedure is not. It is for these very reasons that the MIRD

committee of the Society of Nuclear Medicine has always maintained a focus on individual

organ absorbed dose.









Phantoms


Mathematical

In 1994, companion research to this dissertation was initiated by Dr. Wesley Bolch

to couple the complete series of pediatric mathematical models of Cristy and Eckerman to
the EGS4 radiation transport code (Nelson et al. 1985) for external sources of radiation.
This was additionally performed as a task group effort of the Society of Nuclear

Medicine's MIRD Committee to consider, for the first time, detailed transport of electrons

and beta particles within internal organs for internal sources of radiation. At ORNL, the

Cristy and Eckerman pediatric model series have been used extensively with ALGAM, a

radiation code limited only to the consideration of photon transport (Cristy and Eckerman
1987). EGS4 allows for explicit treatment of coupled photon and electron transport

(including bremsstrahlung production) and has been extensively utilized in both high

energy physics and medical physics research. To verify the coding of the pediatric

mathematical models within the framework of the EGS4 code system, several comparisons

of specific absorbed fractions of energy were performed with the 1987 data published by

Cristy and Eckerman (Kodimer 1995). Excellent agreement was seen over a wide range

of photon energies, confirming the transport of photons in the Cristy and Eckerman
models using EGS4.

A new mathematical model of the adult head and brain has been published

(Bouchet et al. 1996). The model includes a detailed description of the skeletal system

including the maxilla, teeth, mandible, cranium and cervical vertebrae. New regions in the

brain model include the caudate nucleus, cerebellum, cerebral cortex, lateral ventricles,
third ventricle, lentiform nucleus, thalamus and white matter. Other features include eyes

and the thyroid enclosed within an explicit neck region (the original head region was

modeled simply as an elliptical cylinder topped by half an ellipsoid, with no representation










photon transport, its exclusion artificially attenuates external photon sources depositing

dose to the thyroid gland.

Work was initiated in late 1996 to produce a pediatric series of head and brain

models based on modification of the adult model and data published on neural tissue

growth trends. In June of 1997, the MIRD committee also adopted a series of revisions to

the Cristy and Eckerman family of phantoms (Bolch, personal communication, 1997).

These revisions included the addition of an esophagus, a prostate gland, a four-region

kidney (to include a medulla, cortex, papillary, and pelvic region), a mucosal layer within

walled organs (GI tract, gall bladder, and urinary bladder) and the inclusion of a rectum

separate from the sigmoid colon within the current models. It is anticipated that upon

completion of these revisions, future work will include the Monte Carlo-generated doses

being experimentally verified against the MOSFET dosimeter measurements within the

various pediatric anthropomorphic phantoms.


Anthropomorphic


ICRU Report 48 defines a phantom as a "structure that contains one or more

tissue substitutes and is used to simulate radiation interactions in the body" (ICRU 1993,

page 1). They are commonly used to investigate radiographic equipment image quality

and dosimetric evaluations. The complexity of phantoms ranges from very simple

geometries representing singles tissues such as a slab of water simulating soft tissue or a

sheet of copper simulating a chest, to complex geometries with multiple realistic tissue

substitutes for soft tissue, lungs and bones. Such complex phantoms are referred to as

"anthropomorphic" because they simulate the size, shape and composition of a human

body to provide the same attenuation and scattering characteristics.

The majority of anthropomorphic dosimetric phantoms represent an adult man

(Conway et al. 1984; Constantinou et al. 1986; Conway et al. 1990; ICRU 1993). As










previously indicated in the discussion of the mathematical models, a scaled-down adult

phantom is inappropriate to simulate a child, since the organ growth of a child is a

complex process which cannot be described by simply scaling the anatomy of an adult.

The anatomical geometry of a child varies greatly from that of an adult. The weight of

the head with respect to the total body weight is greater for a child than an adult, the trunk

of a child is more cylindrical compared to the elliptical adult trunk, and certain internal

organs such as the thymus gland are larger (Hwang et al. 1976a). The percentage of

extracellular water in children is larger and represents a higher percentage of the total

body weight than in adults (Haschke et al. 1981), and the concentration of certain minerals

is lower in the skeleton of children (Haschke et al. 1981; ICRP 1995). In addition, the

skeleton of a child has more water and less fat than an adult skeleton (Cristy and

Eckerman 1987). Furthermore, the distinction between cortical and trabecular bone and

the percent distribution between yellow and red bone marrow change greatly as children

age (Cristy and Eckerman 1987). The anthropomorphic pediatric phantoms developed by

Chen (Chen et al. 1978) were constructed to represent identically the mathematical models

of the one-year-old and the five-year old, which again, were scaled-down versions of adult

models.

Due to the lack of pediatric anthropomorphic phantoms in the research community

and on the commercial market, research was initiated to construct a prototype one-year-

old phantom at the University of Florida (Bower 1997). The one-year-old representation

was chosen because a review of the examination frequency data provided by Shands at the

University of Florida from the period 1990-present indicated that the majority of pediatric

diagnostic x-ray examinations are of children in their first few years of life; therefore, a

phantom of a one-year-old would be fairly representative of a large class of examinations.

In addition, an anthropomorphic one-year-old phantom has a mass of approximately 10 kg

(21 Ibs), which is easily manageable for construction and its ultimate use in field testing.

The prototype one-year-old anthropomorphic phantom developed by Bower (1997)









utilized in this research is reviewed in detail in Chapter 3, Radiation Measurement

Equipment and Techniques.


Radiation Detection Instrumentation

A variety of detection instrumentation has been used to measure radiation

exposure or absorbed dose. These include gas filled ionization chambers, radiographic

film, semiconductors and thermoluminescent detectors.


Gas Filled Ionization Chambers

Ionization chambers have a response that is proportional to absorbed energy, and,

therefore, they are widely used in performing dosimetry measurements. Most ionization

chambers are cylindrical with an air-equivalent wall and measure exposure. When

radiation interacts with the air in the detector, ion pairs are created and collected

generating a small current. Since the Roentgen is defined as the amount of ionization in

air, measurement of this ionization current will indicate exposure.

The Keithley TRIADM Field Service Kit2 was used in this research to perform

free-in-air exposure measurements in order to calculate entrance skin exposure (ESE) at

the location where the x-rays are expected to enter the patient and characterize beam

quality for the simulated x-ray examinations.


Semiconductors

A semiconductor detector acts as a solid state ionization chamber, whereby the

ionizing particle interacts with atoms in the sensitive volume of the detector to produce

electrons by ionization. The elements in semiconductor materials form crystals that


2 Radiation Measurements Division of Keithley Instruments, Inc., 28775 Aurora Road,
Cleveland, Ohio, 44139, TRIADTM Field Service Kit Model 10100A









consist of a lattice of atoms that are joined together by covalent bonds. Absorption of

energy by the crystal leads to disruption of these bonds which results in a free electron and

a "hole" in the position formerly occupied by the valence electron. This free electron can

move about the crystal with ease. The hole can also move about in the crystal, e.g. an

electron adjacent to the hole can jump into the hole and therefore leave another hole

behind. Connecting the semiconductor in a closed electric circuit results in a current

through the semiconductor as the electrons flow towards the positive terminal and the

holes flow toward the negative terminal. The operation of a semiconductor radiation

detector depends on its having either an excess of electrons or an excess of holes. A

semiconductor with an excess of electrons is called an n-type semiconductor, while one

with an excess of holes is called ap-type semiconductor (Cember 1996).

In order to assess point estimates of organ doses in the one-year-old phantom, the

Patient Dose Verification System designed by Thompson & Nielsen Electronics3 was

utilized. This system is composed of Metal Oxide Semiconductor Field Effect Transistor

(MOSFET) dosimeter arrays. The MOSFET is a common microelectronic device. It is a

layered device consisting of a p-type semiconductor separated from a metal gate by an

insulating oxide layer. Ionizing radiation forms electron-hole pairs in the oxide-insulating

layer of the MOSFET. The applied bias then causes the electrons to travel to the gate,

while the holes migrate to the oxide silicon interface where they are trapped. The trapped

positive charges cause a negative shift in the voltage required to allow conduction through

the MOSFET. The shift in voltage is proportional to the radiation dose deposited, thus

allowing the MOSFET to be used as a dosimeter (Hughes et al. 1988; Gladstone and Chin

1991; Vettese et al. 1996).

This system was originally designed using standard sensitivity dosimeters to

measure skin doses during radiation therapy treatments. Characterization of high

3 Thompson & Nielsen Electronics Ltd., 25E Northside Road, Nepean, Ontario, Canada,
K2H 8S1, Patient Dose Verification System Model TN-RD-50









sensitivity dosimeters by Bower and Hintenlang (1998) for use in diagnostic radiological

applications is detailed in Chapter 3, Radiation Measurement Equipment and Techniques.


Determination of Risk

Radiation risks of diagnostic radiology in the pediatric patient are either

deterministic or stochastic in nature. Deterministic radiation injuries, such as tissue injury

or cataract production, occur when a number of cells are involved and a threshold dose is

required. Above the threshold, the severity of the injury is proportional to the dose.

There is no evidence that deterministic injuries occur as a result of low-dose plain film

procedures. Stochastic radiation injuries, such as genetic effects and carcinogenesis, are

believed to be caused by injury to a single cell and typically a threshold dose is not

required. The probability of the injury is proportional to the dose, but the severity of the

injury is independent of the dose. If one assumes a linear relation without a threshold for

such effects, then any amount of radiation, including low-dose plain film procedures, may

potentially have an effect. A more detailed discussion of the linear, no-threshold (LNT)

theory is presented in Chapter 4, Risk Asssessment.

Risk models were developed to provide a method of predicting the risk of

radiation induced genetic effects and carcinogenesis in relation to the natural incidence in

an unirradiated population. The epidemiological ideal of following several populations

over time was not practical; therefore, statistical models were used to derive risk

estimates. Radiological risk assessments and resulting risk estimates have been developed

by numerous organizations, including the National Academy of Science/National Research
Council's Fifth Committee on the Biological Effects of Ionizing Radiations (BEIR V,

1990), the United Nations Scientific Committee on the Effects of Atomic Radiation
(UNSCEAR, 1988), the International Commission on Radiological Protection (ICRP 60,










1990) and the National Council on Radiation Protection and Measurements (NCRP 115,

1993).

The most recent reassessment of radiation-induced cancer risks from exposure to

low levels of ionizing radiation was performed by the BEIR V committee, and this

methodology was utilized in performing the calculation of the risk estimate from plain film

procedures performed in Chapter 4, Risk Assessment. The BEIR V Report provided

mathematical models to estimate risks for breast cancer, respiratory tract cancer, digestive

tract cancer, leukemia, and other nonleukemia cancers. These equations include terms for

dose, age at exposure, time after exposure and interaction effects. The dose term utilized

in these equations is the effective dose calculated from the MOSFET organ dose

determinations upon applying the dose conversion factors.


Significance and Objectives

Pediatric radiographic examinations are widely thought to yield medical benefits

and/or diagnostic information which greatly exceed any potential risk from patient

radiation exposure. Nevertheless, quantification of that risk is important for a number of

reasons. First, it allows radiographic procedures to be optimized to maximize the medical

benefit while minimizing patient risk. Second, it gives the radiologist a basis for

communicating that risk to concerned patients and/or parents. Third, it allows for

reconstruction of total risk for individuals who have developed radiation-associated

diseases (e.g., leukemia) and who were known to have undergone high-dose or multiple

radiographic procedures at an earlier age (i.e., dose reconstruction). In each case, the

fundamental quantity of interest is the cumulative absorbed dose to the radiosensitive

organs.

As previously discussed, several studies have focused on quantifying average

organ doses in adult patients undergoing radiographic procedures, either through









computer simulation or physical measurement. Nevertheless, very little information exists

on pediatric organ doses from plain film examinations, and no systematic techniques have

been developed to obtain this information for pediatric dynamic procedures such as

fluoroscopy, CT or cardiac catheterization. Consequently, the first objective of this

research project was to establish a baseline and develop a method for assessing pediatric

organ doses for the most frequent plain film procedures used in clinical practice. The

second objective was to verify this method for plain film exams through the use of an

anthropomorphic phantom representing a one-year-old pediatric patient and coupled to a

real-time radiation dosimetry system. The third objective was to utilize the organ doses

determined from the anthropomorphic phantom and calculate an effective dose for each

plain film procedure. The fourth objective was to utilize the effective dose in the BEIR V

methodology to calculate the risk from the procedure. The following specific aims were

designed to accomplish these objectives:

1. To determine, via site survey often facilities, the current examination trends

for one-year-old patients in pediatric x-ray practices in the State of Florida.

2. To determine, utilizing a one-year-old anthropomorphic model providing real-

time radiation dosimetry, organ doses to pediatric patients undergoing the most

common plain film procedures and calculate corresponding effective doses.

3. To estimate, utilizing BEIR V risk model methodology, the risk to pediatric

patients of carcinogenesis from the plain film procedures.

To facilitate specific aim 1, detailed examinations were simulated at Shands at UF

in order to focus the site surveys on the most prevalent examinations currently performed

in pediatric practices. Technique information for these examinations were collected from

all facilities. Due to the time constraint of room downtime, dosimetry data were collected

at the ten facilities for the single most prevalent exam; dosimetry data were collected for

all examinations simulated at Shands at UF. In specific aim 2, point estimates of organ

doses were assessed by selective placement of an array of Metal Oxide Semiconductor









Field Effect Transistors (MOSFET) detectors within the phantom. These measurements

for the x-ray procedures utilized the prototype one-year-old anthropomorphic phantom

developed by Bower (1997). Specific aim 2 also involved the determination of dose

conversions factors (organ absorbed dose per unit entrance skin exposure) for a range of

radiographic projections and radiation beam qualities and the calculation of effective doses

as defined by the ICRP (1977). The task associated with specific aim 3 used equations

developed by BEIR V (1990) to utilize the effective dose calculated in specific aim 2 in

order to predict the risk of radiation induced carcinogenesis from the simulated plain film

procedures.

These tasks not only update and expand obsolete data on pediatric organ doses

from plain film examinations, but establish the method to benchmark experimentally the

mathematical models previously discussed and provide a baseline for the organ dose

estimation computational and experimental techniques necessary to simulate dynamic

pediatric exams. This research provides vital information to the pediatric radiology

community in their efforts to assess organ doses and procedure risk from plain film exams

both prospectively and retrospectively.














CHAPTER 2
PATIENT AND EXAMINATION TRENDS OF PEDIATRIC X-RAY PRACTICES

A general overview of the range of pediatric x-ray practices was obtained through

a site survey of selected facilities within the State of Florida. Ten facilities were selected

to represent their cohorts throughout the rest of the State: six facilities represented

Children's Hospitals and/or Radiology Departments with dedicated pediatric radiologists,

two hospitals represented community patient populations and two hospitals represented

rural patient populations. Selection of the number of facilities was not based on statistical

considerations, as the purpose of this survey was only to identify and characterize the

examinations that are most commonly employed and not establish their precise frequencies

in the U.S. population. Private practices were not included in this survey. In addition,

more specific information was obtained on a single pediatric radiologic practice during in-

depth investigations carried out in the Pediatric section of the Radiology Department at

Shands at the University of Florida.


Shands at the University of Florida


Shands at the University of Florida is a 576-bed not-for-profit tertiary care facility.

The Children's Hospital at Shands at the University of Florida is a 169-bed hospital-

within-a-hospital. The Radiology Department at Shands at UF has a workload of 163,000

examinations per year, approximately six percent of which are for patients under sixteen

years of age. In order to determine which examinations to include in the site surveys,

examination frequency data specific for pediatrics were extracted from the Radiology

Information System (RIS). The RIS provides a computerized patient record from which

physicians can call up all the clinical data that they need about a given patient in










alphanumeric text. The review of the RIS ascertained, retrospectively, the frequency and

types of x-ray procedures performed as a function of age, and simulations were then

performed to determine the important technical parameters of each procedure.

Since the data were to be adapted to five anthropomorphic phantoms and

mathematical models, five age groups were defined for the frequency data. These age

groups were determined from formulae utilized for approximate average height and weight

of normal infants and children that are commonly found in pediatric textbooks (Behrman

and Vaughan 1987). These categories are defined in Table 2-1.

Table 2-1. Age Group Corresponding to Phantom
Phantom Age Group
Newborn 0 to 3 months
One-Year-Old 3 months to 12 months
Five-Year-Old 1 year to 6 years
Ten-Year-Old 7 years to 12 years
Fifteen-Year-Old 13 years to 16 years

The first piece of information extracted from the RIS data was the determination

of the types and frequencies of examinations as a function of age. The RIS data file listed

all the pediatric radiological examinations performed at Shands at the University of Florida

from January 1990 through December 1997. An additional computer program was

written to extract the number of examinations performed in a given year, broken out by

the age group defined in Table 2-1 from the RIS data file. The file encompassed general

radiology (including ultrasound), fluoroscopy, CT, and nuclear medicine examinations.

No exam technique information was contained in this file. An example of the file is

presented in Table 2-2 illustrating the number of pediatric chest x-rays performed in 1997.

The 1997 file in its entirety is located in Appendix A. Prior years were not included as a

trend analysis indicated similar distributions of exams.










Table 2-2. Annual Number of Plain Film Chest X-Rays by Age Group
Year Modality Exam & Projection Newborn 1-Year 5-Year 10-Year 15-Year
97 GENERAL CHEST PA &LATERAL 477 474 1524 665 546
97 GENERAL CHEST 1 VIEW 3763 797 1713 875 973
97 GENERAL BABYORAM (hest & abdomae) 730 30 17 0 0
97 GENERAL CHEST DECUBITUS LEFT 31 3 18 3 9
97 GENERAL CHEST DECUBITUS RIGHT 24 3 11 2 8

The frequency of exams was also determined from this file. The frequency is

defined as the ratio of a specific exam and projection falling within a modality to the total

number of examinations classified for that modality, and is expressed as a percentage. An

example of the data is presented in Table 2-3 demonstrating the percent frequency of

pediatric chest x-rays performed in 1997. The 1997 percent frequency file in its entirety is

located in Appendix B and is tabulated from most to least frequent exam.


Table 2-3. Percent Frequency of Plain Film Chest X-Rays
Year Modality Exam & Projection Newborn 1-Year 5-Year 10-Year 15-Year
97 GENERAL CHEST PA& LATERAL 8 26 26 16 12
97 GENERAL CHEST 1 VIEW 65 44 30 21 22
97 GENERAL BABYGRAM (chet & abdonen) 13 1.7 0.3 0 0
97 GENERAL CHEST DECUBITUS LEFT 0.5 0.2 0.3 007 0.2
97 GENERAL CHEST DECUBITUS RIGHT 0.4 0.2 0.2 0.05 02

The frequency data were further segregated by the principal body region exposed

in the examination: head and neck; thorax and shoulder (excluding humerus); abdomen,

pelvis and hips; and extremities. The most frequent plain film exams per principal body

region were chosen to be included in the dosimetry study. Table 2-4 lists the plain film

exams that were surveyed at each facility based on the data for the one-year-old.

Extremity examinations were not included in the dosimetry survey for the one-year-old

phantom. As demonstrated in Appendix B, the frequency of these exams does not

increase until the older age groups when a child has learned to walk and run, and falls

occur, with subsequent damage to extremities,. Limitations of phantom design with

respect to limb movement also restricted the realism with which such examinations could

be simulated. Radiographic components offluoroscopy examinations were not considered

in this research, as it would be more appropriate to include such dose considerations in a









study of the overall fluoroscopic exam dose. Several additional examinations were

performed at Shands at UF to investigate the effect of shielding on the resultant dose.
The RIS data file also included the projections used in the examinations. The

nomenclature for projections follows that commonly used in radiology (Ballinger 1995).

When referring to a radiograph of a body part as seen from the aspect of the x-ray film,

this is termed a "view," in which case one refers to the side of the body, or body part,

which is closest to the film. For example, in an anterior view of the chest, the patient faces

the film; in a left lateral view, the patient's left side is toward the film. A left posterior

oblique view is performed with the patient turned with his left back toward the film and his

right front away from the film. Generally when the x-ray beam is not incident normally on

the patient's front or back, "view" terminology is used, as opposed to "projection"
terminology. A "projection" defines the path of the x-ray beam through the body or body

part. Thus, in an Anterior-Posterior projection of the chest, the beam enters the front and

exits through the back, producing what could correctly be termed a posterior view. In this

research, projection orientation is used whenever the beam is normally incident on the

front or back ((Anterior-Posterior (AP) or Posterior-Anterior (PA)); view orientation is

used for all others, although the term projection is used for uniformity throughout.
Projections are additionally defined in terms of rotation of the patient or part,

clockwise from the superior aspect beginning with the patient facing the tube. Thus an AP

projection has 0 degree rotation, facing the tube; PA projection has 180 degree rotation,

facing the film; LAT projection (lateral projection) has two possibilities: RLAT (right

lateral) has 90 degree rotation, with the right side towards the film; and LLAT (left lateral)

has 90 degree rotation, with the left side towards the film. The term decubitus is used if

the patient is lying down rather than standing up. Obliques are not included in this data

set, as the majority of oblique projections are associated with fluoroscopy exams.

Additional nomenclature specific for skull and sinus examinations includes "Townes" and

"Waters" projections, respectively. The Townes method for radiographing the skull










involves angling the x-ray tube 20 to 25 degrees caudad (beam penetration through the

top of the head). The Waters method for radiographing the sinuses involves extending the

patient's neck with their chin on the cassette stand.


Table 2-4. Surveyed Examinations
Exam Projection
Skull AP
Skull LAT
Skull Townes
Sinus Waters
Sinus LAT
Cervical spine AP
Cervical spine LAT
Thoracic spine AP
Thoracic spine LAT
Lumbar spine AP
Lumbar spine LAT
Abdomen AP
Pelvis AP
Hip AP
Chest LAT
Chest AP
Chest PA


Ten Florida Facilities


The exposure to pediatric patients during x-ray procedures would be expected to

vary from facility to facility. Factors such as equipment setup and operational

characteristics, patient classification and procedural techniques are of key importance. To

investigate the variation that exists among facilities, a site survey of pediatric x-ray

practices in ten Florida hospitals was conducted. The surveys were conducted in

cooperation with the radiologist, the administrative director and the pediatric x-ray

technologist of the participating facility. One or more of these personnel were present

during each site visit and data collection. In addition to Shands at the University of

Florida, the following facilities participated in this research:









* Tallahassee Memorial HealthCare

Tallahassee Memorial HealthCare is located in Tallahassee, Florida and has 736 beds,

with a radiology workload of 31,200 procedures per year, approximately one and a

half percent of which are estimated to be pediatric.

University Medical Center, Jacksonville

University Medical Center is a 528-bed medical center located in Jacksonville, Florida
with a radiology workload of 190,000 procedures per year, approximately eight

percent of which are estimated to be pediatric.
Shands at AGH

Shands at AGH is located in midtown Gainesville, Florida and has 410 beds, with a

radiology workload of 72,000 procedures per year, approximately four percent of

which are estimated to be pediatric.

Arnold Palmer Hospital for Children & Women

Arnold Palmer Hospital for Children & Women is located on the downtown Orlando

Regional Healthcare System campus in Orlando, Florida and has 267 beds, with a

radiology workload of 33,420 procedures per year, all of which are pediatric.

* All Children's Hospital

All Children's Hospital is located in St. Petersburg, Florida and has 236 beds, with a

radiology workload of 35,930 procedures per year, all of which are pediatric.

* Wolfson Children's Hospital

Wolfson Children's Hospital is part of the Baptist/St. Vincent's Health System located

in Jacksonville, Florida and has 190 beds, with a radiology workload of 41,000

procedures per year, all of which are pediatric.
* Shands at Lake Shore

Shands at Lake Shore is a 128-bed acute care community hospital located in Lake

City, Florida with a radiology workload of 36,000 procedures per year, approximately

eight percent of which are estimated to be pediatric.










* Shands at Starke

Shands at Starke is a 49-bed acute care rural hospital located in Starke, Florida with a

radiology workload of 14,780 procedures per year, approximately six percent of which

are estimated to be pediatric.

* Shands at Live Oak

Shands at Live Oak is a 30-bed acute care rural hospital located in Live Oak, Florida

with a radiology workload of 18,000 procedures per year, approximately three percent

of which are estimated to be pediatric.

The six facilities representing Children's Hospitals and/or Radiology Departments

with dedicated pediatric radiologists were Shands at UF, Tallahassee Memorial

HealthCare, University Medical Center, Arnold Palmer Hospital for Children & Women,

All Children's Hospital and Wolfson Children's Hospital. The two hospitals representing

community patient populations were Shands at AGH and Shands at Lake Shore. The two

hospitals representing rural patient populations were Shands at Starke and Shands at Live

Oak.


Exam Simulations


Per a scheduled appointment approved by the radiologist, initial contact was made

with the director of radiology upon arrival at each facility. The purpose and objectives of

the research were reiterated and the phantom and dosimetry system demonstrated. Copies

of recent news publications about the research were also distributed. The director then

assigned a x-ray room and technologist to the research project for 2.5 hours on average.

The designated time was requested during appointment scheduling and was approximated

based on the time required to carry out the initial characterizations performed at Shands at

UF.









After initial set-up of the phantom to perform real-time dosimetry readings as

described in Chapter 3, the phantom was turned over to the pediatric x-ray technologist to

simulate, as close to clinical practice as possible, the exams listed in Table 2-4 for a one-

year old patient. In order for the pediatric x-ray technologist to simulate the x-ray field

size, shape and location on the phantom, it was necessary to indicate the locations of

certain anatomical landmarks on the phantom that approximated those of an actual one-

year-old patient. The anatomical landmarks, which were defined and placed on the

phantom during the initial simulations at Shands at UF by Dr. J. Williams, a staff pediatric

radiologist, were as follows:

1. Vertex The top or highest part of the phantom head.

2. External acoustic meatus (EAM) Defined to be at the horizontal level of the floor of

the phantom skull.

3. Sternal Notch Defined as the point midway between the anterior medial ends of the

phantom clavicles.

4. Nipples Defined to be located on the level midway between the top of the phantom

trunk and the inferior margin of the twelfth rib; considered to be the midpoint of the

thorax.

5. Xiphoid Process Defined as the inferior margin of the sternum at the level of the

seventeenth rib.

6. Umbilicus Defined as located at the level midway between the phantom diaphragm

and the bottom of the trunk. The umbilicus is considered to be at the midpoint of the

abdomen and at the level of the iliac crests.

7. Symphysis Pubis Defined as located on a level 0.086 times the vertical height of the

trunk above the inferior margin of the trunk, whose relative position is taken from the

vertex. This point is considered to be located at the superior margin of the pubis.

8. Hip Joint Defined as lying on a vertical line, originating at the center of the base of

the leg bone, one-third of the vertical height of the pelvis, above the floor of the trunk.










The distances of these landmarks from the base of the trunk are listed for the one-

year-old phantom in Table 2-5.


Table 2-5. Defined Locations of Anatomical Landmarks on Phantom Vertical Axis with
the Origin at the Base of the Trunk (cm)
Landmark
Vertex 48.5
EAM 39.4
Sternal notch 27.5
Nipples 22.6
Xiphoid process 19.2
Umbilicus 10.6
Hip joint 1.1
Symphysis pubis 0.7

The next section of this chapter describes each exam simulation setup of the

phantom patient in detail. The descriptions are based on technologist interviews and

visual observations obtained of the best practices at each site for actual patient

examinations, supported with the textbook by Kirks and Griscom (1998). Additional

descriptions are provided by Ballinger (1995), Bontrager (1997) and Wilmot and Sharko

(1987). The left lateral projection was chosen for all lateral examinations for ease in

positioning. Images of these examinations are provided in Chapter 5, along with

corresponding phantom images.

AP skull

The patient is positioned supine. The midsaggital plane (divides the body into

right and left halves) of the skull is kept perpendicular to the tabletop. The orbitomeatal

baseline is kept 15 to 20 degrees cephalad from the perpendicular by using a small sponge

wedge or roll of diapers behind the patient's neck to provide some support to maintain this

position. The orbitomeatal line is a frequently used positioning line located between the

outer crease of the eye and the EAM. Several immobilization methods were observed,

ranging from Velcro and tape to foam-backed lead "bookends" placed at both sides of the

head, as shown in Figure 2-1. Typically, one-year olds are also immobilized using the









"bunny" or "mummy" technique, which consists of tightly wrapping the patient in a
blanket to immobilize arms and legs. The thyroid was protected by a shadow shield,
which is a piece of lead attached to the tube head with an adjustable arm. Note that PA
projections are not routinely obtained for this age group, as young children tend to be
frightened of being face down.


Figure 2-1. Items frequently used for immobilization. A: Sandbags, tape, clear Plexiglas
strip. B: Sponges of various shapes and sizes, Velcro strap. C: "Bookends"; patient
"mummified". Figures A and B reproduced with permission from Lippincott-Raven
Publishers. Figure C reproduced with permission from Mosby-Year Book Publishers,
Textbook of Radiographic Positioning and Related Anatomy, K. L. Bontrager (1997).

Lateral skull
The patient is positioned either semiprone or supine. If using a grid, the head is
rotated so that the affected side is closest to the film. If a grid is not used, a cross table
lateral is performed with a film holder. The unaffected side of the body is elevated from
the tabletop using sponges under the neck and shoulders so that the coronal plane (divides


WWO- i
9L









the body into anterior and posterior parts) of the body creates an angle of 45 degrees with

the tabletop. The head is positioned in such a way that the midsaggital plane of the skull is

parallel to the tabletop, and the interorbital line is perpendicular to the film. The

interorbital line is a frequently used positioning line that connects either the pupils or the

outer creases of the patient's eyes. The central ray is perpendicular to the midsaggital

plane and passes midway between the glabella (the smooth raised prominence between the

eyebrows just above the bridge of the nose) and the occipital protuberance (the prominent

"bump" at the inferoposterior portion of the skull). Immobilization and shielding

techniques were similar to those used in the AP projection, except the foam-backed lead

"bookends" were not placed in the field of view.

Townes skull

The patient lies supine on the table so that the orbitomeatal baseline is

perpendicular to the tabletop. Visualization of the foramen magnum (the large opening at

the base of the occipital bone through which the spinal cord passes as it leaves the brain)

may be improved by placing a 20 degree-angled sponge under the skull. For this age

group, an additional angled sponge placed under the lower extremities may also assist in

positioning by depressing the shoulders and encouraging further flexion of the neck. With

the x-ray tube angled 30 to 35 degrees caudal (away from the head end of the body,

toward the feet), the central ray enters the frontal bone in the midsagittal plane at the

hairline passing through a line connecting both external auditory meatuses and exits

through the foramen magnum. An angle of 30 degrees between the orbitomeatal line and

the central ray should be achieved. Immobilization and shielding techniques were similar

to those used in the AP projection. Due to the inability of the phantom's head to

articulate for this method, the pediatric technologist compensated by angling the x-ray

tube to produce the same projection.









Waters sinus

The patient is examined sitting or standing with the chin and the nose in contact

with the film cassette. The patient's mouth is opened during the radiographic procedure.

The neck is extended in such a way so that the orbitomeatal baseline forms a 40 to 45

degree angle with the plane of the film so that the midsaggital plane is perpendicular to the

cassette. The central ray enters at the posterior sagittal suture (the joint connecting the

two parietal bones) just above the occipital protuberance in the midline and runs parallel to

the mentomeatle line and exits at the level where the nose and upper lip meet. The

mentomeatle line is a frequently used positioning line that connects the midpoint of the

chin with the EAM. Again, due to the inability of the phantom's head and jaw to

articulate for this method, the pediatric technologist compensated by angling the x-ray

tube to produce the same projection.

AP cervical spine

As the patient lies supine on the table, the neck is extended to bring the

orbitomeatal baseline to 20 degrees cephalad (toward head end of body) to the vertical.

An angled sponge under the patient's shoulders may be used in order to achieve greater

extension of the neck. The central ray is kept between 10 to 20 degrees cephalad. The

central ray enters at the level ofC-4 or at the thyroid cartilage. The field size is adjusted

so that the view includes the base of the skull as well as the T-l vertebrae. A lead lap

apron was utilized as shielding.

Lateral cervical spine

If using a grid, the patient is positioned supine. The head is in a true lateral

position with the chin raised slightly. The central ray enters the C-4 vertebrae

perpendicular to the film cassette. This procedure may be performed either with extension

or flexion of the neck depending on the kind of functional study. The field is collimated to

include the C-l to T-2 vertebrae. If a grid is not used, a cross table lateral is performed

with a film holder. The patient is placed in a reclined supine position with a sponge placed










under the shoulders. The neck is in a slightly extended position so that the inferior border

of the mandible is vertical. The central ray enters approximately 2.5 cm posterior to the

angle of the mandible. The field is collimated to include C-1 to T-2 vertebrae. Shielding

techniques are similar to those used in the AP projection.

AP thoracic spine

The patient is placed in the supine position. The patient's head rests directly on

the table without the support of a pillow. The patient's chin should be slightly raised. The

hips and knees should be slightly flexed and the knees supported on a pillow so that the

cervical spine will become nearly parallel to the tabletop. Additional restraint is provided

by placing sandbags over the elbows and knees, and by placing Velcro restraining bands

across the chest, hips, and knees. Additional sandbags may be placed along the sides of

the chest and abdomen to further limit movement. The central ray enters at the level ofT-

6 or approximately 10 cm below the sternal notch. The field is collimated to include C-7

to L- vertebrae.

Lateral thoracic spine

The patient is positioned in a true lateral position with the arm closest to the table

flexed to bring the hand up under the head The upper arm also rests on the side of the

head. Note that the arms in the phantom are encased in the trunk and cannot be moved.

When positioning the patient for the lateral view, the head should be placed on a firm

pillow to raise the head and bring the cervical spine into alignment with the thoracic spine.

The hips and knees are flexed to provide stability. In order to reduce the lateral curvature

and ensure that the thoracic spine is parallel to the tabletop, a sponge should be placed

under the patient at the level of the lumbar spine, but above the iliac crests. The arms and

legs should be restrained with sandbags and Velcro restraining bands. A bookend support

should be placed against the back and against the anterior chest wall to discourage arching

of the back. The central ray enters at the level of T-6 just below the inferior angle of the










scapulae. The field is collimated to include C-7 to L-1 vertebrae. The upper 2 or 3

thoracic vertebrae are not well visualized on the standard lateral view.

If upper thoracic vertebrae are of particular interest, a "swimmers" lateral view is

performed. The patient is positioned as described above, but the upper arm is brought to

rest behind the patient's buttocks and the shoulder is rotated slightly posteriorly. This

positioning ensures that neither shoulder will be superimposed on the upper thoracic spine.

The central ray enters at the level ofT-2 and the field is collimated to include C-7 to T-2

vertebrae. Since the arms in the phantom are encased in the trunk and cannot be moved,

this view was not performed.

A "breathing" technique was commonly employed by technologists for lateral

thoracic spine radiographs. Rather than waiting for the time of expiration to take an

exposure, technologists would routinely have the patient breathe normally, so that the

movement would blur the lungs and ribs out.

AP lumbar spine

The patient lies supine with knees and hips flexed and the arms placed to the side.

It was unnecessary for the phantom to perform this flexion, as the posterior aspect of the

phantom lies immediately adjacent to the tabletop. The midsaggital plane of the body is

aligned to the midline of the table. The central ray enters perpendicularly at the level of

the iliac crest or at the L-4 to L-5 interspace. The techniques of immobilization already

described under the discussion for positioning of the thoracic spine are equally applied in

examinations of the lumbar spine.

Lateral lumbar spine

If a cross table lateral is not performed, the patient lies in a lateral recumbent

position with a pillow under the head and the knees and hips flexed with support between

the knees and ankles. The pelvis and the torso are in a true lateral position. The coronal










plane of the patient is aligned to the midline of the table. The central ray enters

perpendicularly at the level of the iliac crest or at the level of the L-4 vertebrae.

AP abdomen

The patient lies supine on the table. The midsaggital plane of the patient is aligned

with the center of the table or the film cassette. Immobilization methods that have been

previously discussed were utilized, such as Velcro or tape, and foam wedges or sandbags

on both sides of the abdomen. The central ray is perpendicular to the film and enters

about one inch above the umbilicus for a one-year-old. The exposure should be made at

the end of expiration. Upright projections of the abdomen may be clinically required and

are performed with immobilization techniques similar to those utilized for upright chest

radiography.

AP pelvis

The patient lies supine with legs extended and separated. The legs are internally

rotated 15 to 20 degrees. Since the phantom's legs are not attached to the trunk of the

body, this positioning does not change the internal anatomy of the phantom. The patient

midsagittal plane is aligned with the central ray which enters a point midway between the

level of the anterior superior iliac spine (prominent anterior border of the iliac crest) and

the superior border of the symphysis pubis. The central ray is perpendicular to the film,

and a gonad shield is utilized. It should be mentioned that a "frog-leg" pelvis view is also

routinely performed for this age group. In a frog-leg view, the knees and hips are flexed

as far as comfortable. Both thighs are abducted as far as possible, and plantar surfaces of

the feet are placed together. This view was not performed with the phantom due to the

inability of the phantom's joints to articulate.

AP bilateral hip

The patient lies supine with their arms immobilized at their sides or crossed at the

chest with the pelvis as symmetrically positioned as possible. The legs are extended and









rotated internally 15 to 20 degrees when a non-trauma patient is being examined. The

central ray enters the midline at a point halfway between the anterior superior iliac spine

and the symphysis pubis. Both hips are always examined in the same projection on the

same film.

A lateral frog-leg hip is routinely performed for this age group. The patient is laid

partially oblique with a support under the affected area. The knee on the affected side is

flexed, and the thigh is drawn up to a 45 degree angle position. The affected femoral neck

is centered to the midpoint of the film cassette. The central ray is perpendicular to the film

and enters at the level of the mid-femoral neck. This view was not performed with the

phantom, due to the inability of the phantom's joints to articulate.

AP. PA and lateral chest

Pediatric chest radiographs are routinely performed upright without a grid. The

challenge of performing upright imaging includes preventing motion and rotation, freeing

the lung fields of superimposition of chin, humeri and scapulae, and obtaining a good

inspiratory radiograph. Various methods of immobilization are used, however, the most

common pediatric positioner and immobilization tool utilized is the Pigg-o-stat. The Pigg-

o-stat is composed of a large support base on wheels, a small adjustable seat and Plexiglas

support sleeves, which come in two sizes. The seat, sleeves, and "turntable" base rotate

as a unit to facilitate quick positioning from the PA to the lateral projection.

As shown in Figure 2-2, the child sits on the little bicycle type seat. The plastic

support sleeves fit snugly around the sides and keeps the arms raised. The child usually

cries with frustration at being confined, but the crying actually helps to obtain a good x-

ray image because, at the end of a cry, the child will take a big gasp of air, and at that

moment the exposure is taken. The orientation of the room, i.e., where the chest stand or

film holder is located relative to the control panel, determines how well the technologist

can see the child's thorax to ensure the exposure is made during inspiration. Inspiration









can also be detected by watching the child's abdomen as it extends on inspiration,

watching the chest wall as the ribs will be outlined on inspiration and watching the rise and

fall of the sternum. The central ray for both PA and lateral projections enters

perpendicular to the film at the level of the midthorax T-6/T-7 vertebrae or mammillary

(nipple) line. The collimation is adjusted so that all of the ribs are included in the
radiograph and should extend from, and include, the mastoid tips to just above the iliac

crests. A lead lap apron was utilized as shielding.




















Figure 2-2. Infant immobilized and supported by the Pigg-o-stat device for an upright
chest film. The device is then rotated 90' for a lateral projection. Note the presence of
shielding for the pelvis. Figure reproduced with permission from Lippincott-Raven
Publishers, Practical Pediatric Imaging: Diagnostic Radiology of Infants and Children
D. R. Kirks and N. T. Griscom (1998).

Radiologists that did not favor the use of the Pigg-o-stat for aesthetic reasons or

because of the potential for sleeve artifacts, frequently used a supine pediatric immobilizer

referred to as a Tam-em board. This one step immobilization device positions the child

with the arms, head and legs immobilized with foam lined Velcro straps to a Plexiglas

frame for both an AP projection (with the film located on the tabletop under the frame as









shown in Figure 2-3) and a cross table lateral projection (with the film in a holder as

shown in Figure 2-4). This immobilization device was also used when upright positioning

was contraindicated.


Figure 2-3. Toddler immobilized for an AP chest radiograph on the Tame-Em
immobilizer. Velcro bands are placed around the arms and legs; a vinyl sheet of lead
covers the lower abdomen and gonads. Figure reproduced with permission from
Lippincott-Raven Publishers, Practical Pediatric Imaging: Diagnostic Radiology of Infants
and Children, D. R. Kirks and N. T. Griscom (1998).


Figure 2-4. Infant immobilized for a left lateral chest radiograph on a Plexiglas
immobilizer. Velcro bands are placed around the arms and head; a vinyl sheet of lead
covers the lower abdomen and gonads. Figure reproduced with permission from Mosby-
Year Book Publishers, Merrill's Atlas of Radiographic Positions and Radiologic
Procedures, P. W. Ballinger (1995).
A variation of this device is an octagonal immobilizer, shown in Figure 2-5, which

is an eight-sided immobilization tool that permits visualization in a variety of positions.























Figure 2-5. Octagon board for immobilization. Figure reproduced with permission from
Lippincott-Raven Publishers, Practical Pediatric Imaging: Diagnostic Radiology of Infants
and Children, D. R. Kirks and N. T. Griscom (1998).


Survey Data

For future comparison of the measured and simulated effective doses, several

parameters from the survey data would need to be correlated. The most important

descriptors collected during the exam simulations at each facility included the type of

generator, e.g. single phase, three phase, constant potential or high frequency; the

collimated size of the x-ray field; the distance from the tube focal spot to the beam

entrance point on the skin or surface of the phantom (SSD); the direction of the x-ray

beam with respect to the long axis of the phantom (projection); and the quality of the x-

ray beam as described by the kVp and half value layer (HVL).

For comparison of technique factors among hospitals, additional parameters were

collected from each facility. These parameters included the mode of operation, e.g.

manual or automatic exposure controlled (AEC, also called phototiming or amplimatting);

the detector configuration of ion chambers that were utilized for AEC-right (R), center

(C) and/or left (L) and the density setting ((-3, -2, -1, Normal (N), +1, +2, or +3)); if

scatter suppression was used; film size; film/screen speed; the distance from the tube focal

spot to the image receptor distance (SID); the output of the x-ray beam as defined by the









milliamperage and time applied (mAs); the focal spot size-large (LFS) or small (SFS); and

shielding used. Repeat and reject rate percentages were also collected.

In addition to recording technique parameters during the simulations, several

measurements were performed. Higher than normal exposures were taken to ensure

reasonable accuracy in the absorbed dose data. After the clinical exam was performed,

exposures from an output of 80 mAs were performed. The accuracy of the kVp was

verified prior to measuring the HVL. The exposure per mAs was measured at a variety of

kVp settings at a given distance in order to compute the entrance skin exposure (ESE),

which was subsequently used in the effective dose calculation. X-ray films were obtained

for all the chest examinations, and the optical density of the middle of the lung fields on

the chest x-rays were quantified with a densitometer and averaged to provide a measure of

image quality. All measurements were performed in accordance with procedures

recommended by the American Association of Physicists in Medicine (Seibert et al. 1994).

The survey data in their entirety are reviewed in Chapter 5 and are tabulated for each

exam and projection in Appendix C.

During the initial survey performed at Shands at UF, x-ray films and dosimetry

measurements were performed for each examination listed in Table 2-4. During the

subsequent site surveys, x-ray films and dosimetry measurements were only performed for

the chest examinations in order to decrease room downtime. The chest examinations were

chosen for comparison, as they were the most frequently performed examination at each

facility.














CHAPTER 3
RADIATION MEASUREMENT EQUIPMENT AND TECHNIQUES


One-Year-Old Anthropomorphic Phantom

Bower successfully constructed a full-scale prototype of a physical heterogeneous

phantom to represent the new MIRD one-year-old mathematical model, which included a

revised head and neck (1997). This prototype was the phantom utilized for the

measurements in the site survey study. The construction entailed several steps. First,

processing techniques were developed to produce tissue equivalent materials

approximating soft, lung and bone tissues which required the addition of a particulate filler

with a moderately high effective atomic number, such as magnesium oxide, to an unfilled

liquid resin with a low effective atomic number. The theoretical interaction coefficients

(mass attenuation and mass energy-absorption coefficients) were matched between these

tissue-equivalent materials and the tissue media used in the 1987 Cristy and Eckerman and

MIRD models over the photon energy range 1.5 keV to 150 keV using the computer

program XCOM written by Berger and Hubbell at the Center for Radiation Research,

National Bureau of Standards (Hubbell 1982). The mass density was adjusted by adding a

small quantity of low-density microballoons (hollow gas-filled spheres). This basic

process followed that published by White (1977) and further developed by White et al.

(1977, 1986) and Herman et al. (1985, 1986).

The soft tissue substitutes were made of materials providing a final elemental

composition of 59.2% carbon, 20.6% oxygen, 10.5% magnesium, 7.5% hydrogen, 2.0%

nitrogen and 0.10% chlorine by weight. The attenuation and mass energy-absorption

coefficients of the tissue-equivalent substitutes were modeled after the atomic










compositions of soft tissue defined in the 1987 models of Cristy and Eckerman and ICRP

Publication 23 (1975). These two sets of photon interaction coefficients were in

agreement within + 3% and were considerably closer at most energies. The soft tissue

substitute's photon interaction coefficients were also favorably compared to those given

for individual tissues given in ICRU Publication 44 (1989). The soft tissue substitute was

modeled to provide the appropriate soft tissue density of 1.04 g/cm3.

Lung-equivalent materials were manufactured using a foaming agent and a

surfactant, similar to the method presented by White et al. (1986). The lung tissue

substitutes were made of materials giving an elemental composition of 59.42% carbon,

17.65% oxygen, 12.06% magnesium, 8.24% hydrogen, 1.64% nitrogen, 0.84% silicon and

0.15% chlorine. The lung tissue substitute density was 0.360 g/cm3, which approximated

the lung tissue density of 0.296 g/cm3 for the mathematical model.

A bone tissue substitute approximating the skeletal tissue of a one-year-old was

also produced. Cristy and Eckerman made the following observations concerning the

skeleton of the newborn: "The skeleton of the newborn contains more water, less fat, and

less mineral than the adult skeleton. Furthermore, the distinction of two bone types,

cortical and trabecular bone, is not evident in the newborn skeleton, and the marrow of the

skeleton is all active. Thus it is clear that the elemental composition of the adult skeleton

cannot be used when evaluating radiation transport in the newborn." [1987, page 42]

Consequently, the authors proposed a different skeletal tissue medium for their newborn

model, but utilized the adult skeletal tissue medium for all other models in the series

including the one-year-old. New data on average skeletal atomic composition as a

function of age available in ICRP Publication 70 (1995) were used to manufacture age-

specific skeletal-equivalent materials incorporated into the one-year-old phantom. The

skeleton composition (in terms of water, protein, mineral and fat by percent weight) was

used to calculate an appropriate elemental composition of the bone-equivalent substitute

for the one-year-old skeleton consisting of 52.98% carbon, 24.54% oxygen, 5.86%









hydrogen, 2.03% nitrogen, 5.79% chlorine and 8.81% calcium. The mineral percent by

weight for a one-year-old skeleton was provided explicitly in ICRP Publication 70. The

water and fat percentages were obtained from the 1995 document's data tables and the

percentage of protein was obtained by subtraction. However, Bower had difficulty in

producing a bone substitute that matched the density for a one-year-old, but could

produce a bone substitute mixture of 1.18 g/cm3 in 1997, which approximated Cristy and

Eckerman's density of 1.22 g/cm3 for the newborn in 1987. Since the new elemental bone

composition for the one-year-old developed by Bower in 1997 was close to the newborn,

and the difference in densities between the newborn and the one-year-old in ICRP

Publication 70 was only 0.01 g/cm3, he decided to use the newborn bone density of 1.18

g/cm3 for the one-year-old.

Molds were then created for the lungs and individual skeleton components

consisting of two leg bones, two arm bones, a pelvis, a spine, twelve ribs, two scapula and

two clavicles. The molds were filled with the respective lung or bone liquid resin tissue

substitutes and allowed to cure. The cured components were milled to final shape. Once

the internal components were completed, molds were created for the external parts

consisting of the trunk, legs and skull. The soft tissue substitute material was then poured

into the external cylindrical trunk mold in a layered method which facilitated the correct

placement of the skeletal and lung components. The legs were frustums of two soft tissue

cones surrounding the leg bones. The head consisted of a skull enclosing the mandible,

teeth, cranium and upper face region, as well as a neck. Guide holes were drilled into the

phantom to allow placement of the dosimeters. Figure 3.1 shows the mathematical model

of the one-year-old, a portion of the internal skeleton with the head model, and an external

view of the completed physical prototype phantom.



































A B C

Figure 3-1. One-year-old pediatric phantom. A: Mathematical model. B: Skeletal model
with 4 ribs. C: External view of physical prototype phantom.

The total mass of the one-year-old prototype is 10.2 kilograms. The length of the

head, trunk and legs, respectively, is 16 cm, 32.5 cm, and 26.5 cm, for a total height of 75

cm. Several aberrations should be noted regarding the prototype phantom. The

mathematical model has soft tissue eyes with a volume of 6.25 cm3, whereas the physical

prototype phantom does not have soft tissue eyes. The space for the eyes is composed of

bone tissue substitute; therefore the mass of the upper face region for the physical

phantom is slightly higher than the mathematical model. The upper portion of the

mandible is slightly thinner in the physical phantom, however, this portion abuts against

the upper face region. Therefore, when averaged together, the two regions are

approximately correct in comparison with the mathematical model. The teeth are also










slightly more massive in the physical phantom, however, the total mass of the teeth is only

22 grams, and the teeth are not a chosen dosimetry monitoring location. Some of the ribs

in the prototype were heavier than the theoretical masses; however, the individual ribs are

thin (volume 7.28 cm3) and are not used as a measurement location.


Metal Oxide Semiconductor Field Effect Transistor (MOSFET) Dosimetrv System



Theory of Operation


The basic structure of a metal oxide semiconductor field effect transistor

(MOSFET) is a sandwich-type device consisting of a p-type channel built on a n-type

silicon semiconductor substrate separated from a metal gate by an insulating oxide layer.

The source and the drain are on top of the positively doped silicon region. When

sufficient negative bias is applied to the gate with reference to the substrate, a significant

number of holes will be attracted to the oxide silicon surface. A sufficient hole population

permits current to flow between the source and the drain. The gate voltage necessary to

allow conduction through the MOSFET is referred to as the threshold voltage (Soubra

1994; Ramani 1997).

When the MOSFET is exposed to ionizing radiation, electron-hole pairs are

formed in the oxide insulation layer. The applied positive potential to the gate causes the

electrons to travel to the gate, while the holes migrate to the oxide silicon interface, where

they are trapped. These trapped positive charges cause a shift in the threshold voltage.

The voltage shift is proportional to the radiation dose deposited in the oxide layer. This is

the basis of the MOSFET as a dosimeter. The Thomson and Nielsen Patient Dose

Verification System used in this research utilizes a dual bias, dual MOSFET device design,

which consists of two MOSFETs fabricated on the same chip. The threshold voltage shift

magnitude is increased when the positive gate bias is raised. Since the MOSFETs are










irradiated simultaneously, the measured difference in the threshold shifts in the two

sensors is representative of the absorbed dose (Soubra 1994).

Each detector of the Patient Dose Verification System consists of the matched pair

of MOSFET dosimeters operated at two different positive gate biases. Each MOSFET

dosimeter has an active area of 0.04 mm2. The pair is mounted under a 1-mm layer of

epoxy to a 20-cm long thin semi-opaque polyamide laminate cable encasing two gold

wires. The extremely thin (0.2 mm), flexible laminate cable is attached to a sturdy 1.4-m

long cable that is connected to a bias supply. The reader can accommodate up to 20

dosimeters. To facilitate monitoring of the lower absorbed doses associated with

diagnostic x-rays, the model TN-RD-19 high-sensitivity bias supply is used. Grouped in

sets of five, each MOSFET set is connected to a bias supply labeled "A" through "D". In

this research, there were two sets of five dosimeters labeled "A" and "D", for a total often

dosimeters. The entire dosimetry system is shown in Figure 3-2 including the reader,

power supply, bias supply, a MOSFET dosimeter and the associated cabling.


Figure 3-2. Thomson and Nielsen Electronics Ltd. MOSFET Patient Dose Verification
System









Characterization of High Sensitivity MOSFET Dosimeter


The commercial Patient Dose Verification System manufactured by Thomson and

Nielsen was originally designed for radiation therapy dose verification. To measure doses

to organs that were outside of the primary therapy beam, Thomson and Nielsen developed

high sensitivity dosimeters to measure the lower doses. Bower and Hintenlang

characterized these high sensitivity dosimeters for diagnostic energies (1998). Bower

measured the sensitivity, linearity, rotational angular response, multiple query response

and post-exposure response of the high sensitivity MOSFETs. He determined that the

high sensitivity MOSFET dosimeters had a nearly uniform angular response, good

linearity, and precise sensitivities at the lower energy ranges that resulted in a standard

error of+ 5% for a typical x-ray exposure. He also determined that the energy

dependence of the sensitivity, the post-exposure drift and the multiple query responses

were not significant limitations. Johnson performed three additional characterization tests,

including a comparison of the MOSFET to TLD, a determination to see if the small air

gap that surrounds the MOSFET when inserted into the phantom influences the dose

measurement, and a determination of the axial angular dependence of the MOSFET

(1998). She determined that the dose reported by the MOSFET is equivalent to the dose

reported by the TLD inside the tissue phantoms and that the small air gap that surrounds

the MOSFET when it is inserted in the holes in the phantom does not affect the reported

dose. Consequently, there is no need to fill the air gap with other tissue equivalent

material. She also determined that angular dependence in the axial (saggital would be a

more appropriate definition) orientation of the MOSFET when placed in the phantom does

not influence the measured absorbed dose. An extrapolation of these determinations for

the rotational and saggital planes is that the angular dependence from a coronal orientation

would also not influence the measured absorbed dose.










The MOSFET is a consumable detector. While the detector is not annealed or

cleared after reading, as is the case for thermoluminescent dosimeters, the system can

provide a reset between desired readings. The reset process consists of measuring the mV

response and using this response as a reference value for the following measurement. The

response is not completely stable over long periods of time post-exposure. Bower

performed a fading study in 1997 to measure the response of the high sensitivity

MOSFETs as a function of time post-exposure out to 500 hours. The dual-bias, dual-

MOSFET system utilizes the stable difference in response from the two detectors for

measuring absorbed dose. This method eliminates temperature and high-dose

dependencies, but does not control post-exposure drift. The drift is a complicated

function attributed to an array of causes. The drift is reproducible, and mathematical

models and deconvolution methods have been proposed and employed (Gladstone and

Chin 1995). In using this dosimeter system to monitor organ doses within the one-year-

old phantom in real-time, delay times between exposure and dosimeter readout are on the

order of minutes, therefore, another fading study was performed to measure the response

of the MOSFET as a function of time post exposure for an immediate read, a one-minute

delay, a two-minute delay and a three-minute delay. The exposure between the four

reading times varied by less than 2%, so a delay time of one minute was utilized in this

research to decrease the room downtime associated with the data collection, and no

correction factors were necessary.

The dosimeters inserted within the physical phantom are required to have good

angular response characteristics since there is no way to adjust the orientation of the

detectors towards the x-ray source when simulating different views of x-ray examinations.

In 1997, Bower demonstrated there is a marked angular dependence of the MOSFET

dosimetry system when it is used free-in-air without an attenuating medium. The

packaging of the MOSFET appears to be the main source of this angular dependence,

providing different responses when exposing the flat surface (60% lower) as opposed to










the epoxy bubble side of the dosimeter. It was initially thought this difference might be

attributed to insufficient material on the flat surface of the dosimeter for the establishment

of electronic equilibrium. It was subsequently demonstrated that this response is not

simply due to scattering from the 1-mm bubble of epoxy, since the angular response

remained after material was added to the flat side of the detector to make it more

symmetrical. The effect is actually due to the inherent asymmetric structure of the

MOSFET itself, specifically resulting from attenuation in the silicon substrate layer. The

angular dependence is much less pronounced when the dosimeter is surrounded by

attenuating media. This was verified by using a tissue equivalent cylinder and measuring

the angular response for the high sensitivity MOSFET dosimeter. The cylinder, with a

detector inserted in the center, was irradiated in 15* increments over 3600 at a tube

potential of 90 kVp, 150 mAs, and a source-to-detector distance of 75 cm. The

procedure was repeated for each MOSFET to obtain an average dose measurement at

each increment. The response was very symmetrical, indicating excellent angular response

when using this dosimeter within a phantom. The nearly uniform angular response and its

small size were primary reasons the MOSFET dosimeter was chosen for the real-time

dosimetry applications within the one-year-old phantom.


MOSFET Placement


The effective dose concept presented in the 1990 recommendations of the ICRP

(1991) identified twelve specific tissues at risk from radiation detriment and supplied

tissue weighting factors for these tissues. Several other tissues were addressed in a

"remainder" term. The specific tissues and their associated weighting factors are shown in

Table 3-1. The MOSFET dosimeters were placed in locations to estimate the absorbed

dose in these tissues for the one-year-old phantom. For localized organs, the MOSFET

was placed at the centroid of the organ to determine the dose. For distributed organs, the










dose was determined from a fractional weighting scheme applied to a specific array of

MOSFET dosimeters. Standard dosimetric allowances were made for measuring the

absorbed dose in both the testes and the ovaries in the hermaphrodite phantom since an

actual patient would not have both types of gonads. The effective dose estimates

performed in Chapter 5 are shown with the testes absorbed dose being used as the gonadal

dose component for the male estimate, and the ovaries absorbed dose being used as the

gonadal dose component for the female estimate. All other tissues and their

corresponding tissue weighting factors were the same. A skin dose estimate was obtained

by measuring the entrance and exit absorbed doses and using an average value of the two

measurements. The organs defined in the remainder term were as follows: muscle,

stomach wall, small intestine wall, upper large intestine wall, lower large intestine wall,

kidneys, pancreas, spleen, thymus, uterus, adrenals, and bladder wall. A remainder dose

estimate was calculated by averaging the measurements in the arms and the trunk of the

body as representative of the overall location of the remainder organs.

Table 3-1. Tissue Weighting Factors for Calculation of Effective Dose (ICRP, 1991)

Tissue or organ Tissue Weighting Factor (wT)
Gonads 0.20
Bone Marrow (red) 0.12
Colon 0.12
Lung 0.12
Stomach 0.12
Bladder 0.05
Breast 0.05
Liver 0.05
Esophagus 0.05
Thyroid 0.05
Skin 0.01
Bone surface 0.01
Remainder 0.05


The MOSFET locations in the one-year-old phantom are shown in Figure 3-2 and

detailed in Table 3-2. The MOSFET locations for the soft tissue organs, including the









lungs, were generally placed at the centroid of the organ. A single MOSFET location was

chosen for the testes and breasts for the following reasons. In the mathematical model

representing a one-year-old, the testes are outside of the trunk and are very small (1.16

cm3). The testes are close together, and great differences were not expected between the

individual testes, therefore, a single monitoring location was chosen at the point where the

two legs and the trunk join. Additional soft tissue material was not added to the phantom

at this representative location. The mathematical model has breasts that are again very

small for the one-year-old (1.06 cm3 including skin), so breasts were not explicitly created

for the anthropomorphic phantom. The dose to the breasts is estimated with the

anthropomorphic phantom by measuring the absorbed dose to the surface midway

between the two breasts. The centroids of the soft tissue organs in the anthropomorphic

phantom were adopted from those supplied by Cristy and Eckerman (1987; Eckerman et

al. 1996) for their mathematical model with the exception of the colon and the esophagus.

Representative monitoring locations were chosen for the colon and the esophagus based

on the equations provided for the colon (ICRP 1995) and the esophagus (Eckerman et al.

1996). The skeletal locations were chosen using three criteria: 1) the bones selected were

larger bones able to accept the MOSFET without loss of a significant percentage of the

bone material, 2) the bones selected represented high concentrations of active bone

marrow and 3) locations were selected to cover a major portion of the skeleton.













Figure 3-2. MOSFET dosimeters placed in position in prototype one-year-old phantom.











Table 3-2. Position of the MOSFET Dosimeters Within the One-Year-Old Phantom
gonads: testes-general (left & right)
ovaries-left
ovaries-right
colon
lung-left
lung-right
stomach
bladder
breasts-general (left & right)
esophagus
liver
thyroid
skeleton: skull & facial
spine (in head)
spine (middle of back)
pelvis
leg-lower left
leg-lower right
arm-left
arm-right

The MOSFETs were placed in the phantom with the epoxy bubble facing toward

the x-ray tube, and the leads were taped down to prevent any movement. Initially, ten

dosimeters were available for measurements. Two of the original compliment often failed

early in the measurement series. One failed mechanically; the reason for the other failure

could not be determined by the manufacturer. The simulations had to be repeated two and

a half times since there are 20 different organ sites and only eight dosimeters were

available. In each successive simulation, the dosimeters were placed in the 10

measurement sites that could be accessed on the top half of the phantom, and the

simulation was repeated in order to measure the 10 remaining sites that were accessed on

the bottom half of the patient. A full complement of dosimeters were not purchased until

data collection was close to completion.

ICRP Publication 70 (1995) provides tables of the active bone marrow found in

various bones as a function of age. These tables are based on work presented by Cristy










(1981) who relates the active marrow in individual bones, parts of bones, or bone groups

expressed as a percentage of active marrow in the body. Weightings of the active marrow

for the different phantom skeletal regions were assigned to the skeletal MOSFET locations

given in Table 3-2. New active marrow percentages associated with the MOSFET

location were assigned by Bower and are presented in Table 3-3.

Table 3-3 has been updated to include new bone surface percentages introduced in

this research. Cristy (1981) also provides the skeletal volume for each region, from which

a percentage of the total bone was calculated. Using an analagous technique, weightings

of the total bone for the different skeletal regions were assigned to the same skeletal

MOSFET locations, and new bone surface percentages associated with the MOSFET

location were assigned.


Table 3-3. Percentage Active Bone Marrow and Bone Surface Associated with the
Skeletal MOSFET Locations
Phantom Skeletal Region % % MOSFET % Active % Bone
Active Bone Location Marrow Surface
Marrow Assigned Assigned
to to
MOSFET MOSFET
Skull (cranium & facial skeleton) 24.74 19.9 Skull & facial 14.7 11.6
Spine (upper portion) 1.88 3.2 Spine head 14.7 11.6
Arm bones upper portion 2.41 5.0 Arm bone left 7.79 9.8
Scapulae 2.73 3.2 Arm bone right 7.79 9.8
Clavicles 0.83 0.8
Ribs 9.61 10.7
Spine (middle portion) 9.27 8.5 Spine middle 15.88 18.5
Ann bones middle portion 2.25 5.0
Arm bones lower portion 4.36 5.0
Spine (lower portion) 3.37 4.1 Pelvis 21.91 21.9
Pelvis 16.47 9.3
Leg bones upper portion 2.07 8.5
Leg bones middle portion 3.88 8.5 Leg lower left 8.64 8.5
Leg bones lower portion 13.4 8.5 Leg lower right 8.64 8.5








Measurement of Absorbed Dose

Sensitivity has been defined as the MOSFET response per unit of exposure (mV

per C kg1 or mV/R) at a given tube potential; consequently the absorbed dose (rad) in the
tissue may be derived as follows. The ion chamber measures exposure in air in units of
roentgen (R); therefore, this unit was used in measuring the sensitivity. The exposure in
air, as measured by the MOSFET dosimetry system is therefore the MOSFET dose
measurement (mV) divided by the sensitivity (mV/R). Expressed symbolically in Equation
1, where X is the exposure, M is the MOSFET dose measurement and S is the sensitivity.
X=M*() (Eq. 1)

One roentgen produces an absorbed dose of 0.876 rad in air; therefore,
D,, =X*0.876 (Eq 2)

The absorbed dose in tissue may be then be calculated with the following equation:

D,_ = D.,Z _1)_, (Eq. 3)

where p,, / p is the average mass energy-absorption coefficient associated with the mean

energy of the x-ray spectrum. Equations 1-3 may be combined into one useful equation as

follows.


D, 0.876* M (Eq. 4)


The value of p, / p for the oxide could be removed from the equation without

error; however, its inclusion indicates that the absorbed dose in the MOSFET device is an
integral part of the corresponding dose value in tissue. The average mass energy-
absorption coefficient ratios are effectively 1.0 for soft tissue, but increase to 3.5 for
skeletal tissue at 30 keV.










The mean energy of the x-ray spectrum was estimated by utilizing the computer

program XCOM5R (Nowotny and Hifer 1985). The program is a spectrum generating

program which accounts for parameters such as the tube potential, the inherent filtration,

and the target materials. The program provides detailed information about the x-ray

spectrum, including the mean photon energy of the spectrum. The determination of the

exact mean spectrum energy is not critical for soft tissue or lungs, since the ratio of mass

energy absorption coefficients is relatively constant and nearly 1.0 over the diagnostic

energy range. The determination is more critical for the bones since that ratio is a rapidly

changing with photon energy and peaks around 30 keV.


Sensitivity Analysis


MOSFET sensitivity, defined as the mV response per R exposure, is shown in

Table 3-4 below for the ten high sensitivity dosimeters initially used in this research. Just

prior to the completion of data collection, an additional ten dosimeters were added to the

system, so during the calibration of the new dosimeters, the old dosimeters were

recalibrated, as shown in column three. Exposure measurements were simultaneously

made with a Keithley calibrated ion chamber. The data points were fit to obtain entrance

skin exposures as a linear function of tube potential; only the most commonly used tube

potential determined from the site surveys is reported.


Table 3-4. Sensitivity Calibration Factors (R/mV) at 70 kVp
MOSFET # Calibration Factor (R/mV)
1 0.028
2 0.031
3 0.029
4 0.029
5 0.030
6 0.030
7 0.031
8 0.031
9 0.029
10 0026


Calibration Factor (R/mV)
0.030
0.035
0.031
0.030
0.029
0.037
0.034
0.036
0.032
0.033









Table 3-4, continued.
11 0.031
12 0.029
13 0.029
14 0.030
15 0.030
16 0.030
17 0.033
18 0.030
19 0.032
20 0.033

Calibration factors needed for subsequent determination of tissue absorbed dose

(R/mV) were determined from these sensitivity data and through the application of the

Bragg-Gray cavity theory, as previously described. The conversion factors given in Table

3-5 contain the ratio of average mass energy absorption coefficients detailed in Appendix

D and a constant for converting exposure to the absorbed dose in air, as described in

Equation 4. Utilization of these conversion factors convert the exposure, obtained by

multiplying the MOSFET measurement by the appropriate value in Table 3-4 to absorbed

dose.


Table 3-5. Conversion Factors Converting Exposure to Absorbed Dose
Tube Potential Soft Tissue Lung Tissue Bone Tissue
70 1.009 1.046 3.316

The following technique was used to assign a dose to the active bone marrow

using a weighted fraction of bone marrow assigned to the MOSFET locations in the

skeleton reported in Table 3-3. The active bone marrow dose was calculated by utilizing

the conversion factor for soft tissue in Table 3-5 to calculate the absorbed dose of each of

the skeletal components, multiplying by the fraction of active bone marrow assigned to the

MOSFET (Fi) reported in Table 3-3, and then summed to get the total absorbed dose to

the active marrow. Equation 5 illustrates how the dose to the active marrow is calculated.









D =0.876*.* *- I *M,* ~F, (Eq. 5)


A similar technique was used to assign a dose to the bone surface using a weighted
fraction of bone surface assigned to the MOSFET locations in the skeleton reported in
Table 3-3. The bone surface dose was calculated by utilizing the conversion factor for
bone in Table 3-5 to calculate the absorbed dose of each of the skeletal, multiplying by the
fraction of bone surface assigned to the MOSFET (Fi) reported in Table 3-3, and then
summed to get the total absorbed dose to the bone surface. Equation 6 illustrates how the
dose to the bone surface is calculated.


D =0.876* *L b *M* F (Eq.6)




Calculation of Effective Dose

Effective dose (E) as presented by the ICRP in 1991 is the most current and
comprehensive measure of radiation detriment. Effective dose equivalent (HE) was a
measure presented earlier by the ICRP in 1976. The effective dose was chosen to be
calculated in this research as it evaluates the absorbed dose from more tissues than the
effective dose equivalent and applies updated tissue weighting factors which are
numerically different from those used with the effective dose equivalent. Various sets of
tissue weighting factors were derived for children ages 0 to 9 and 10 to 19 years by Almen
et al. and tested for use in the calculation of effective dose (1996). They concluded that
the ICRP tissue weighting factors were applicable to children and adolescents; therefore,
the ICRP tissue weighting factors were utilized in this research. The effective dose was
calculated by multiplying the absorbed dose for each organ described in the previous
section by its respective tissue weighting factor reported in Table 3-1. To summarize,






59


organ doses were determined with the MOSFET at the centroid of the organ, the active

marrow and bone surface doses were determined via fractional weightings assigned to

individual MOSFET locations, and the remainder dose was determined from an averaged

measurement over MOSFETs located in the torso region. Since the effective dose

calculation is a weighted average of the MOSFET measurement, the uncertainty in the

effective dose caclulation is determined to be a maximum of+ 5%. Appendix E presents

the measured organ doses and calculated effective doses utilizing the MOSFET dosimetry

system and the one-year-old physical prototype phantom. The results are reviewed in

Chapter 5.














CHAPTER 4
RISK ASSESSMENT


General

Risk assessment has been defined by the National Research Council (NRC) as "the

characterization of the potential adverse health effects of human exposures to

environmental hazards"(NRC 1983, page 18). An assessment of the risks from all types of

hazards, including radiological hazards, requires all or some of the following components:

i) hazard identification, which is investigated to determine whether a particular hazard has

a corresponding health effect; ii) dose-response assessment, in which the relation between

the magnitude of the dose and the probability that the health effect will occur is

determined; iii) exposure or dose assessment, which is the determination of the extent to

which humans will be exposed to the hazard; and iv) risk characterization, which describes

the nature and magnitude of the human risk, including uncertainties surrounding that risk

[NRC, 1983 #2254]. It is the last component of risk characterization that integrates the

results of the previous three components into a risk model that includes one or more

quantitative estimates (Cohrssen and Covello 1989).


Components of a Risk Model

The model used to perform a risk assessment is a function of several parameters,

including populations from which data are obtained, relationships between doses and

effects and methods used to project risks into the future. Due to the limited knowledge of

radiation effects, there is diverse discussion about the interpretation of dose-response data

and the choice of values for the other parameters










Populations used in epidemiological studies vary among risk assessment models.

Some studies provide better data than others do because some populations better

represent the population of concern. In addition, differences in population characteristics

such as lifestyles, cancer rates and types of exposures must be considered in transferring

the results obtained from one population to another population, which is defined in a risk

model as the population transfer coefficient.

The dose-response curve and dose-rate effect applied can differ. Among the

possible risk models, the risk analysis techniques most commonly applied are a linear-

quadratic or a linear relationship.

Risk projection models used to predict the mortality from, or incidence of, cancer

and genetic disease that will occur within a given population can be of several types.

Cancer risk projections use additive or multiplicative methods; genetic risk projections

may use direct or indirect methods. Each method requires choices from among several

additional parameters including death rates, specific causes of death, population statistics,

and dose calculations.

Radiological risk assessments and resulting risk estimates have been developed by

numerous organizations. Organizations such as the National Research Council's fifth

committee on the Biological Effects of Ionizing Radiations (BEIR V), the United Nations

Scientific Committee on the Effects of Atomic Radiation (UNSCEAR), and the

International Commission on Radiological Protection (ICRP) have summarized the

complex data previously discussed that are available on the effects of radiation in a form

that is easily applicable. The BEIR V methodology is considered the most recent consent

document for risk analysis and was utilized in this research.









BEIR V Methodolovg



Populations for Determining Cancer Risk

In the risk assessment by the BEIR V Committee in 1987, cancer risks were based

mainly on epidemiological studies and, in large part, on the population of residents from

Hiroshima and Nagasaki in 1950 (the Life Span Study). The data from this population

were based on dose estimates made in 1986 (the Dosimetry System 1986 {DS86})

(Shimizu et al. 1987). To validate these assessments, additional human and animal

populations were reviewed. For example, thirty-four additional studies were cited in

support of the risk assessment for breast cancer.

Human Populations

The BEIR V Report estimated risks for breast cancer, respiratory tract cancer,

digestive tract cancer, leukemia and other nonleukemia cancers. The human populations

used in the risk assessment were chosen based on the type of cancer. For breast cancer,

four populations were used: residents of Hiroshima and Nagaski in 1950 (the Life Span

Study), most of whom were exposed to the atomic bomb; women examined by

fluoroscopy in Canada for tuberculosis from 1930 to 1952; women examined by

fluoroscopy in Massachusetts from 1930 to 1956; and women treated with radiotherapy

for postpartum mastitis in New York during the 1940s and 1950s. For respiratory and

digestive tract cancers, the population comprised residents of Hiroshima and Nagaski in

1950 (the Life Span Study). For leukemia, two populations were used: residents of

Hiroshima and Nagaski in 1950 (the Life Span Study); and people treated with

radiotherapy for ankylosing spondylitis in the United Kingdom from 1935 to 1954. A

third population, women treated with radiotherapy in several countries for cervical cancer,

may have been used in the leukemia risk calculations, however, the BEIR V report is










unclear on this point. For all other cancers (except leukemia), the study population was

comprised of residents of Hiroshima and Nagaski in 1950 (the Life Span Study).

Animal Populations

In general, the BEIR V report cancer risk assessments considered animal

populations only when necessary to validate or understand data from human populations.

Animal studies raise numerous problems, including extrapolation of results obtained under

experimental conditions to the conditions relevant to population exposure, such as dose

rates, fractionation, and other variables; and extrapolation from an experimental organism

in which radiation effects may be estimated with some confidence, to humans, because

organisms differ in radiation sensitivity.


Populations for Determining Genetic Risk


The assessment for genetic risk relied on animal (especially mouse) data to

supplement human data. To determine human genetic risk by extrapolating from animal

data, three assumptions were made: the amount of genetic damage induced by a given

type of radiation under given conditions in the animal species was the same in human germ

cells; biological factors such as sex, germ cell stage, and age, and physical factors such as

quality of radiation and dose rate, similarly affected genetic damage in the animal species

and in humans; and the relationship between low- linear energy transfer (LET) radiation

and frequency of genetic effects was linear at low doses and low dose rates. Genetic

effects, by definition, require that at least one generation pass before they are expressed,

and their expression is affected by the exposed population's sex, age, and probability of

having children. Genetic risks determined from human data are based on the assumption

that all of the exposed population was of reproductive age and wanted to have children.

In the irradiation of an entire population, the reproducing people are a fraction of the total,

and damage to the germ cells of the nonreproducing people in the population does not









pose a genetic risk. When the irradiation is neither random nor uniform, as is true of the

available human populations utilized in the risk assessment, a reduction factor is required

to adjust the genetic risk estimates.

Human Populations

The BEIR V report used several human populations, divided into three sets, to

assess genetic risk. The first set consisted of people with genetic disorders resulting from

spontaneous mutations. These disorders included dominant autosomal disorders, sex-

linked recessive disorders, recessive autosomal disorders, chromosomal abnormalities,

congenital abnormalities and other multifactorial traits. Multifactorial traits are a group of

disorders about which the exact mode of inheritance is unknown and include such diseases

as diabetes mellitus, gout, schizophrenia, affective psychoses, epilepsy, glaucoma,

hypertension, varicose veins, asthma, psoriasis, ankylosing spondylitis and juvenile

osteochondrosis of the spine. The second set consisted of people with genetic disorders

resulting from several specific spontaneous mutations that were used to calculate mutation

rates. The disorders included dominant autosomal disorders and recessive sex-linked

disorders. The third set was the Hiroshima and Nagasaki survivors and their children,

including members of a pregnancy termination study, a cytogenetic study of the children of

exposed parents, an investigation of rare electrophoretic variants in children of exposed

parents, and doubling dose studies.

Animal Populations

In general, mouse populations were used to assess genetic risk when human data

were unavailable or inappropriate. A few data points were taken from monkey and

marmoset studies. The animals in these studies displayed a specific endpoint, such as

dominant and recessive lethal and visible mutations, reciprocal and heritable

translocations, congenital malformations and aneuploidy.









Dose-Response Curves

A dose-response relationship, or curve, is determined by the change in effect

(response, the dependent variable) with increasing amounts of radiation (dose, the

independent variable). The relationship between dose and response is the primary problem

in predicting the health risks of radiation (Cohrssen and Covello 1989). Determining the

radiation dose has been detailed in Chapter 3. Determining the response to radiation is

difficult, and the details of the process are beyond the scope of this research. Although

somatic effects (such as cancer) and genetic effects are the major responses to radiation,

many intermediate conditions or specific endpoints may be used as measurements. Among

the many endpoints that are considered in nonhuman studies are clonogenic survival of

cultured cells and formation of tumors and death in test animals. In human studies, effects

such as chromosome abnormalities, enzyme aberrations, tumorigenesis and leukemia are

considered.

The effects of cancer and genetic damage are presumed to be stochastic; that is,

any level of irradiation increases the likelihood of inducing genetic damage or cancer (no

threshold relationship). Thus, the dose-response relationship of interest presumably begins

at the level of background radiation, which varies from place to place. This presumption

is currently under great controversy, as a threshold relationship is becoming more

frequently used. As these methodologies evolve, the quantitative risk predictions of this

research may need to be revisited.

Most experimental studies are conducted using high doses, because the effects of

low doses of experimentally applied radiation are indistinguishable from those of natural

background radiation in small experimental populations (UNSCEAR 1986)1. These high

doses are plotted against the responses to obtain a piece of the dose-response curve, and a

mathematical relationship is used to determine the shape of the curve in the low-dose


1 The UNSCEAR defines low dose as less than 0.2 Gy (20 rad).










region. Mathematical relationships explain the two differently shaped dose-response

curves most commonly applied to radiation data: linear and linear-quadratic., as shown in

Equations 1 and 2, respectively, where R is response, D is dose and a and (i are constants.
LINEAR: R=aD (Eq. 1)

LINEAR-QUADRATIC: R = a D + / D (Eq. 2)

The linear relationship extrapolates high-dose data along a straight line into the

low-dose region where response is directly proportional to dose. The linear-quadratic

relationship is effectively linear at low doses but begins to rise more steeply with

increasing dose because of a squared term in the mathematical expression. Thus, at high

doses, response is proportional to the square of the dose. At low doses, the linear

relationship predicts a greater response than does the linear-quadratic relationship and is

often considered as an upper limit. Thus, the linear relationship provides a conservative

estimate of dose useful for radiation protection purposes.

The BEIR V committee chose a linear dose-response curve for quantifying genetic

risk and most cancer risks from low doses of radiation because the epidemiological data

they reviewed were best described by this model. For leukemia, however, the BEIR V

committee used a linear-quadratic dose-response curve, which better fit the DS86

dosimetry data from the residents of Hiroshima and Nagasaki.


Dose-Rate Effects

Since most of the data on radiation effects come from high doses, they must be

extrapolated to low doses, which describe most human irradiation, and the plain film

examinations reviewed in this research. Evidence from medicine and biology suggests that

as dose and dose rate decrease, the effect per unit dose also decreases for low-LET

radiation (NCRP 1980). Thus, high doses and dose rates are more effective at causing










damage than low doses and dose rates. This difference is quantified by the dose-rate

effectiveness factor (DREF) or dose and dose-rate effectiveness factor (DDREF).

The effect of dose rate has been studied for numerous years. For low-LET

radiation, the effect depends on several factors, including repair of sublethal damage,

redistribution of cells in the mitotic cycle, and compensatory proliferation of cells during

protracted irradiation. For high-LET radiation, the dose-rate effect is much reduced. The

effect seems to apply both to cancer and genetic endpoints, although the effect is not

necessarily equal for either endpoint (BEIR 1990)

The dose rate effectiveness factor is estimated using mouse or human data. To

reiterate, with a linear-quadratic relationship, the dose response curve is approximately

linear at low doses at the low end of the curve. With a linear relationship, the dose-

response curve continues to be linear at low doses. Since it is extrapolated from high-

dose data along a straight line, it predicts greater response at a given dose than does the

linear-quadratic relationship. The ratio of these two responses at low doses, which is in

effect the linear extrapolation overestimation factor, approximates the dose-rate

effectiveness factor (Fabrikant 1990). These two models are shown in Figure 4-1. The

solid circles represent hypothetical data for an excess incidence of cancer observed at

relatively high doses. Curve A represents a linear extrapolation; curve B represents a

quadratic relationship between incidence and dose; curve C illustrates a threshold type of

response, which was previously discussed. While all three models fit the high-dose data

equally well, the risk estimates pertinent to this research in the low-dose region are quite

different according to which model is chosen for the extrapolation.











t
z


C-)
z


U,



X-RAY DOSE ~
LOW-DOSE HIGH-DOSE
REGION REGION


Figure 4-1. Various models used to extrapolate high-dose data on cancer incidence to the
low-dose region, so that risk estimates can be performed. Redrawn from: Radiation
Carcinogenesis in Man, United Nations Scientific Committee on the Effects of Atomic
Radiation (1977).

The most commonly cited range for dose-rate effectiveness factor is two to ten;

that is, radiation at high doses and dose rates is from two to ten times more effective at

causing damage than radiation at low doses and dose rates (NCRP 1980). Dose rate

effectiveness factors are applied with care. For example, as previously discussed, BEIR V

used a linear-quadratic relationship to model leukemia, which would already predict

reduced numbers of cases of leukemia at low doses and dose rates (BEIR 1990). The

BEIR V committee also chose to fit separate dose-response curves for each cancer

(Fabrikant 1990) and therefore did not use a dose-rate effectiveness factor. The factor

used in assessing genetic risk, which was based on the ratio of genetic risk from high to

low dose-rate irradiation of mice, was approximately three.


Population Transfer Coefficients


A number of uncertainties are involved in a risk assessment. Some sources of

uncertainty can be evaluated using conventional statistical theory and are incorporated into

risk assessment parameters such as dose-response relationships and risk projection










methods. Other sources of uncertainties cannot be captured by the usual statistical

techniques. One important uncertainty addressed by the BEIR V report was that inherent

in applying risks determined in one population to another population. Referenced in a risk

model as the population transfer coefficient, this issue was of major concern because the

BEIR V methodology's risk assessment was primarily based on results from the Hiroshima

and Nagasaki survivors applied to U.S. populations. The extrapolation required

assumptions about diets, industrial exposures, cancer rates and lifestyles in general. The

BEIR V committee recognized the population transfer issue to be a fundamental problem

and considerable source of uncertainty in estimating risks. The report chose to evaluate

this issue by "a consensus of expert opinion as to the uncertainty, expressed in a number

on a scale commensurate with ordinary statistical measures of variability" (BEIR 1990,

page 220).

The committee obtained this number by judging the range within which it was

believed to lie with 95 percent confidence. In this way, all types of uncertainty, both

statistically calculated and estimated by consensus, were evaluated together to obtain

combined measurements of standard error and confidence intervals. The committee thus

estimated the uncertainty in population transfer to be 20%.


BEIR V Cancer Risk Projection Method


Lifetable analyses using standard mortality tables that were modified to include an

additional incremental risk from radiation were used by the BEIR V committee to

calculate lifetime cancer risks from specific irradiation. Input data for the lifetable analyses

included radiation doses and parameters used in the assumed dose-response relationship

(Bunger et al. 1981).

To comprehend a lifetable analysis of a standard mortality table, consider a lifetime

irradiation at a constant annual rate. A lifetable analysis starts with a hypothetical










population of one million newborns, and the columns of the table provide the following

information shown in Table 4-1.


Table 4-1. Lifetable Analysis of a Standard Mortality Table
Number of Cancer Number of Number of deaths
surviving infants death rate cancer cases from other causes

The first column gives the number of infants expected to survive each age; the

second column gives the cancer death rate predicted by the dose-response curve; the third

column gives the number of cases of cancer deaths (which is the product of the first and

second columns); and the fourth column gives the number of deaths from causes other

than radiation based on mortality rates. The number of infants surviving to each age (first

column) less the number of radiogenic and nonradiogenic cancer deaths (sum of third and

fourth columns) is the number of survivors at the beginning of the next age interval. This

process continues with increasing age until the entire population is dead or until age one

hundred (BEIR 1990).

Quantifying excess cancer deaths from radiation is the key to projecting risks.

When the risk to exposed people exceeds the risk to unexposed people by the same

amount at all ages, the effect of the radiation is additive, and the mathematical expression

for this risk is thus an additive one. This is often called the absolute risk because at all

ages the excess risk is constant. When the risk to exposed people exceeds the risk to

unexposed people by a constant fraction, the effect of radiation is multiplicative. Also

known as relative risk because at all ages after irradiation, the relative risk or risk ratio is

constant, the multiplicative relationship is the mathematical expression of this risk

(Muirhead and Darby 1987). These two models are overly simplistic in projecting risk.

The BEIR V Committee permits the risk to vary as complex functions of time after

exposure throughout the individual's lifetime. This is demonstrated in Figure 4-2, which

illustrates the Committee's preferred risk model for leukemia.










t









ATTAINED AGE -+


Figure 4-2. The risk of leukemia due to low LET radiation as a function of attained age.

Cancer risk projection requires knowledge of several parameters, including the

relationship between excess cancer risk and relative risk, the latency period or the time

from irradiation to the first expression of excess cancer risk, the plateau period or the time

from the first expression of excess risk until the excess cancer risk disappears, the age

distribution of the exposed population and the baseline pattern of age-specific mortality

rates from all causes and from the cancers under consideration, the age at irradiation and

the dose-response function. These parameters are brought together in the cancer risk

projection model used by the BEIR V Committee. The report used cancer risk projection

to provide several measures of risk, i.e., estimating the probability of radiation-induced

cancer death expressed as a percentage, the number of projected cancer deaths expressed

as deaths per thousand or million exposed people per unit dose, and the number of years

of life lost in an exposed population because of radiation-induced cancers. All of these

estimates, however, derive from a single measure of excess lifetime cancer risk chosen by

BEIR V, as discussed in the next section.
The BEIR V Committee chose to use only multiplicative methods to project

cancer risk. The measure chosen by this committee was excess lifetime cancer risk, which

is the increase in the lifetime probability that a person will die from a specific cancer as a










result of a specific irradiation. The method of lifetime excess risk estimation used in the

BEIR V report differs slightly from those previously used. In the BEIR V Report,

separate lifetime risks are estimated for an exposed population and for an unexposed

population. The unexposed population is assumed to be exposed to the same background

radiation as the exposed population. The excess cancer risk is simply the difference

between the two lifetime risks (BEIR 1990). The parameter estimates used for the BEIR

V cancer risk projections were obtained using AMFIT, a program developed for the

analysis of survival data. Further discussion of some of these parameters provides more

insight into the assumptions made by the BEIR V report and are reviewed in the next

sections.

Specific Cause of Death

As previously noted in the discussion on populations, the BEIR V report projected

risks separately for leukemia, breast cancer, respiratory tract cancer, digestive tract cancer

and other nonleukemia cancers. For leukemia deaths, the BEIR V report assumed a two-

year latency period. For deaths from other cancers, a ten-year latency period was

assumed.

Age of Population

The age distribution of a population at the time of irradiation must be specified in

order to predict the effect of that irradiation because some of the cancers used as

endpoints depend strongly on age at irradiation. For example, leukemia risks for people

exposed before age 20 are much greater than leukemia risks for people exposed later in

life (Vaeth and Pierce 1990) and risk varies as a complex function of age and time since

irradiation (Thomas, Darby et al. 1992). For the cancers with the longer latency periods,

there is more uncertainty in the application of the BEIR V risk models. Since there is

nothing available that is more advanced than BEIR V, the BEIR V methodology was










utilized, recognizing that these results may need to be revised if the BEIR V models are

revised.


BEIR V Genetic Risk Projection Method

The BEIR V report did not provide a direct estimate of the risk of total genetic

damage, stating that these estimates were highly uncertain because they did not include

allowance for genetic diseases of complex etiology that may be caused by multiple factors.

The report suggested that these diseases make up the largest category of genetically

related diseases and that further research was required before these probabilities could be

estimated accurately. The report did provide estimates of genetic effects for other types

of genetic disorders, however, the performance of these calculations are beyond the scope

of this research. This research will only utilize the BEIR V methodology to estimate the

risk of radiation induced cancer utilizing the equations detailed in the following section.


Calculating the Relative Cancer Risk from Pediatric Diagnostic X-Ray Procedures

The BEIR V report's general expression for calculating the total cancer risk,

including natural and radiation-induced, to a population is given in Equation 3 (BEIR

1990).


A (d)= [ + f(d)g(fl)] or A (d) = + f(d)g(fl)A (Eq. 3)


where X0 is the individual's age and gender-specific mortality rate for a given type of

cancer in the absence of a radiation exposure other than natural background (or absolute

age specific cancer risk to the unexposed population);f(d) is a function of dose equivalent

in Sievert; g(P) is the excess risk function for a specific cancer that depends upon gender,

age of the individual, age at exposure, and time since exposure; and X(d) represents total









fatal cancer risk. The second term on the right of Equation 3, [1 +Ad) g(f)], therefore,

represents the radiation-induced fatal cancer risk.

The dose-response function,fd), depends upon the type of cancer involved. The

equations forAd), taken from BEIR V (BEIR 1990), are:


Leukemia:


Ad) = 0.243d+ 0.271d'


Respiratory cancer: Ad) = 0.636d

Digestive cancer: Ad) = 0.809d

Other cancer: fd) = 1.22d

Female Breast Cancer: Ad) = 1.22d

Similarly, the radiation -induced cancer risk depends upon a number of factors that

are incorporated into the excess risk function that has been discussed previously. In the

following equations for the excess risk function g(f), E represents the age at exposure,

and T represents the number of years following exposure. Only the equations applicable

to a one-year-old patient simulated in this research are presented.

For males:

A. Leukemia (latent period = two years)

E < 20; T < 15: g(f) = exp(4.885) = 132.3

B. Respiratory cancer (latent period = ten years)

g(3)= exp[-1.437 ln(T/20)]
C. Digestive cancers (latent period = ten years)

E <25: g() = exp(0) = 1.0

D. Other cancers (latent period = ten years)

E < 10: g() = 1.0

For females:

A. Leukemia (latent period = two years)

E < 20; T < 15: g() = exp(4.885) = 132.3










B. Respiratory cancer (latent period = ten years)

g(P) = exp[-1.437 ln(T/20) + 0.711]
C. Digestive cancers (latent period = ten years)

E < 25: g(p) = exp(0.553)= 1.74

D. Other cancers (latent period = ten years)

E < 10: g() = 1.0

E. Breast cancers (latent period = ten years)

E < 15: g() = exp[1.385 0.104 In(T/20)- 2.21 n2(T/20)]

The values ofg(P) were determined for each gender for an exposure that occurs at

an age of one year and the time post-exposure was chosen to be five years past the latency

period for each different type of cancer. The five year post-latent time period was chosen

by plotting g(B) as a function of time for each cancer and gender, as shown in Figures 4-3

through 4-9. These illustrations only extend to 25 years, but the elevated risk continues

further into the future. The five year post-latent time period was chosen as representative

of all the cancer models but was initially based on the leukemia cancer model because that

is approximately in the middle of the time interval of maximum risk for leukemia, as shown

in Figure 4-3.



150

100

50


0 5 10 15 20 25 30
Time (years)

Figure 4-3. Time dependent risk for exposure at age equal to one year for both male and
female leukemia models.













2.5
2
1.5


0.5
0
0 5 10 15 20 25 30
Time (years)

Figure 4-4. Time dependent risk for exposure at age equal to one year for male respiratory
cancer model.


0.8

0.6

| 0.4

0.2
0
0 10 20 30
Time (years)


Figure 4-5. Time dependent risk for exposure at age equal to one year for female
respiratory cancer model.











1.2

0.8
0.6
0.4
0.2
0
0 5 10 15 20 25 30
Time (years)

Figure 4-6. Time dependent risk for exposure at age equal to one year for male digestive
cancer model.


1.5


0.5
0
0 5 10 15 20 25 30
Time (years)


Figure 4-7. Time dependent risk for exposure at age equal to one year for female digestive
cancer model.





1.2

0.8
9 0.6
S0.4
0.2
0
0 5 10 15 20 25 30
Time (years)

Figure 4-8. Time dependent risk for exposure at age equal to one year for both male and
female other cancer models.












3
2.5
2
1.5

0.5
0
0 5 10 15 20 25 30
Time (years)

Figure 4-9. Time dependent risk for exposure at age equal to one year for female breast
cancer model.

The values offtd) were determined utilizing the effective doses calculated in

Chapter 3, rather than dose equivalent, since effective dose is the most current and

comprehensive measure. The values ofg(P) were then multiplied by the appropriate value

offd) and the relative risk predicted by the [1 +f(d) g(p)] term in Equation 1 was

determined using these variables for each procedure simulated. The calculation of the

relative risk for specific pediatric exams are performed in Chapter 5.














CHAPTER 5
RESULTS AND DISCUSSIONS


Facility Approach to Handling Special Concerns of Pediatric Patients

The child is not a small adult. This aphorism is a mantra for any discussion of

pediatric radiology; it is particularly pertinent to the specifics of technique. Most general

diagnostic radiologists and radiologic technologists are uncomfortable examining an infant

or young child. Unfortunately, this uneasiness may be transmitted to the pediatric patient

and frequently to the parent. The result is often an inadequate examination that is either

confusing or uninterpretable and requires retakes, ultimately yielding an increase in

radiation dose to the patient (Kirks and Griscom 1998). Pediatric imaging is no better

than its techniques and technologists. Thus, special approaches are required to handle

pediatric patients.
Children are frequently apprehensive in the hospital setting. These feelings are

compounded by fear of the unknown and the pain so often associated with doctors and

hospitals. As a result, not all technologists enjoy working with children; many become

frustrated by the lack of cooperation of infants and by the time required for pediatric

examinations, and there are, therefore, dedicated pediatric technologists. All of the

pediatric technologists involved in the site surveys appeared to be conscientious and

dedicated individuals who enjoyed working with children and had sufficient patience but
firmness to develop the rapport to handle pediatric patients. Reducing the child's fright

and obtaining his or her trust is of great benefit to the patient psyche and adds to the
technologist's ability to carry out high-quality examinations.









There are many different ways of successfully approaching a child, and certain

basic guidelines were observed during the site surveys that appeared to be adapted to

individual technologist styles and personalities. Of greatest importance is obtaining the

pediatric patient's trust. Obtaining trust can be difficult with any child but is especially

difficult with a child who is ill and suddenly finds herself, or himself, in a strange

environment. Creating an environment in which a child feels unthreatened and

comfortable is a challenge to hospital designers. Children enjoy stimulating surroundings

and require something to occupy their time while waiting for a procedure to be performed.

The environment must accommodate the parent's needs in addition to those of the child.

Since the first area encountered in a radiology department is usually the waiting room,

furniture appropriate to both parents and children should be available. For the

departments in the site survey that had dedicated pediatric facilities, an area equipped with

tables and chairs to accommodate children's activities was commonly found. Unbreakable

toys, puzzles, games, coloring books, crayons and storybooks were observed in all of the

surveyed facilities. In the majority of surveyed facilities, the room commonly used for

pediatric examinations had a mural painted on the wall or a myriad of stuffed animals on

the shelves to provide a soothing, entertaining atmosphere for the child. One of the

facilities visited had the x-ray equipment itself painted in a jungle motif and the table and

lift decorated to resemble a boat; patients were told to relax and enjoy the jungle boat ride.

Most pediatric rooms also carefully kept all medical paraphernalia out of sight but had

accessory equipment such as sponges, sandbags, and restraining devices readily accessible

from the tableside.

All of the pediatric technologists interviewed indicated that the key to establishing

rapport with a child is truth and honesty. As children are not only inquisitive, but very

perceptive, they should never be lied to. All of the technologists attempted to establish a

good cooperative relationship with the child by initiating a conversation with the child

about things that interested him or her. Chatting often enables the child to relax and









increases their willingness to help with the study. The technologists always explained the

procedure fully to the patient no matter how young the child was, in words that could be

understood. The technologists felt this knowledge not only helped to lower the patient's

anxiety, but also increased their credibility. Children like to be talked to and feel reassured

and comforted by participating in an interesting conversation. This technique also worked

with infants, who seemed to feel more secure when they heard a friendly voice.

A major aid utilized by all the surveyed facilities to decrease parent and patient

anxiety was to furnish them with a simple, printed explanation of the procedure when the

appointment is initially made. The technologists felt this proved to be a great help in

getting the patient to relax and be cooperative. Scheduling the appointment to coincide

with the child's routine was also important, although obviously not always practical. All

technologists expressed the preference to not x-ray children under the age of five between

noon and 4:00 P.M., as that is when most children are eating lunch and taking naps. The

technologists preferred to perform the procedures in the morning when the children were

rested, refreshed, not a lot of doctors had poked at them, and nothing really overwhelming

had happened to them yet. Children cannot handle the procedure as well later in the

afternoon because they are tired, stressed out and they have reached their limit for the day.
In obtaining the child's cooperation, it is often helpful to give him/her a small

degree of control over the circumstances. For example, a teddy bear or a security blanket

may be placed as the child wishes, or the choice of a particular Band-Aid or color of

hospital gown may give the child the feeling that he/she can still influence this new

environment. Another method of helping children feel in control of the situation is to tell

them what they're going to do during the exam. The most successful technologists

observed would say "You're going to stand up here and do this, then you're going to turn

around and do that, I'm going to do this, and then you're finished!" rather than "Come on
in here and let's take some pictures." Another example often recalled by the technologists

included emphasizing what's positive, e.g. asking a child to show them how long they










could hold still rather than telling the child "Don't move." It is worthwhile for the child to

understand that he/she is an active participant in this important procedure that is being

performed on his/her behalf. Children often want to help adults, and the sick child is no

exception. The phrase "Now I need you to help me" was heard frequently during the site

surveys.

Parent involvement in the procedure varied from site to site and appeared to be a

controversial issue. At most of the facilities, parents were invited to be present in the

radiography suite for all plain film examinations, as most young children are more

cooperative during radiographic procedures when a parent is present. Separation from

parents can be very stressful for young children and the presence of the parent may make

the child more cooperative by diminishing the fear of being deserted in a strange place.

During the explanation of the exam, the technologist can determine the child's response to

the new environment and decide which parent will make the most effective assistant. It

was common to invite only one parent to accompany the patient into the radiographic

suite. A careful explanation of the procedure enhances the ability of the parent to assist

the technologist and provides reassurance to the child. Parents sometimes feel guilty that

their child has suffered an accident or illness and allowing them to assist may help to ease

their anxiety as well. The technologists also made it clear that if the parent became upset

with the procedure or were disruptive to the achievement of good rapport with the child,

they would not hesitate to ask them to leave the room. The technologists indicated that

the possibility that a parent may be asked to leave is often sufficient encouragement for the

child to cooperate in the study, as is the thought of having to repeat the procedure. The

technologist also made it clear to the child that the examination he/she is undergoing is

necessary and inevitable and that he/she has no choice in the matter. The idea that the

technologist is "the boss" was also made clear to the child. The technologists routinely

stressed the importance of immobilization to the parent. If immobilization devices are

used, the parent can assist in placing the child in the device, although most technologists









seemed to prefer to do this themselves so that the parent just provided moral support and

could "come to the rescue" when the procedure was over. Parents can also demonstrate

to the child what will take place during the exam. The technologists indicated that many

children balk at instructions from an instinct of self-preservation, as children will not turn

their backs to complete strangers in complete medical settings, especially if they have had

rectal exams, lumbar punctures or shots Getting the child into the appropriate position

was often accomplished with the use of stickers on the wall at the point where children of

differing heights would be looking. The children were then told to look at and count the

stickers appropriate to their height. Often, animal stickers would be scattered about the

mural previously discussed. The children are then instructed to try and spot as many

stickers as possible and choose a favorite, they are then told to hold very still because if

they move, the animal sticker will run away. Bargaining with lollipops and stickers is a

common technique (however, many children possess uncanny negotiating skills). Many

such games were observed. Of additional benefit to the technologist, the parent can serve

as an extra set of hands and eyes to ensure the child's safety while on a radiography table

or in a restraining device, since toddlers are very active.

Finally, the child needs to be rewarded for good behavior and for contributing to

the success of the exam. The technologists felt this was important for symbolic value and

hopefully left the child with a relatively good feeling about a potentially negative

experience. Verbal praise was observed to be administered copiously throughout the

procedure and subsequently in the parent's presence at the completion of the exam. To

quote Mary Francis Sheppard, Supervisor of Pediatric Radiology at Shands at UF and lead

pediatric radiologist, "It's of great benefit and costs nothing." All facilities also distributed

a variety of stickers, lollipops, coupons, or small toys.

Younger children are sometimes inconsolable and will cry or struggle even under

optimal circumstances. In these situations, it was observed that performing the exam

quickly and efficiently is the best thing that can be done. Pacifiers appeared to be helpful,










as was a soothing continuously talking voice. It was observed that the technologist would

always talk to the child, even if he/she was screaming, crying, looking away, or pretending

to talk to someone else. Frequently, the children who seemed to be ignoring the

technologist were actually paying close attention. A common ploy used by the

technologist would be to keep asking the child questions during all the screaming and slide

in the question if they wanted to go home and if they answered, then the technologist

knew they were listening. Although most parents would tell their child to not cry, the

technologist would tell them that crying is preferable to biting, kicking or spitting as a

method to relieve their stress.

To consistently obtain high-quality radiographs in agitated children is a technically

demanding task, but the overall observation from the site surveys was that success was

possible with thought, care, time and patience. The following poem by pediatric

radiologist Leonard E. Swischuk, M.D. (1997), was posted in a majority of the surveyed

facilities and accurately depicts the approach a facility should follow with pediatric

patients.
A bundle of movement so awkward to hold,
Crying and crying for minutes untold.
Perpetual motion, so hard to contain,
Your charisma and patience, it truly can strain

Your very first thought may prompt a retreat,
You'll want to give up and concede to defeat.
But before you pack up and go screaming away,
Just think for a moment, this babe needs you today.

Hold him awhile, help him adjust,
Restrain his momentum, it's almost a must.
You'll not really hurt him, if his arms you must bind,
Just stay in command, be gentle and kind.

There is nothing so fine as to take a babe home,
And nothing so hard as to leave him alone.
So help little Jack, and somebody's Jill,
Their mothers and fathers sure hope that you will.









Patient and Examination Trends


Although the workload distribution of examinations were significantly different

between hospitals, the same examinations tended to predominate in each hospital. As

shown by the survey data in Appendix C, the numerical workload was dominated by the

data for the hospitals with the most beds and any statistical comparison among the

hospitals would be of little value. Looking only at the examinations identified from the

initial survey at Shands at UF, it was seen that these examinations listed in Table 2-4 were

represented in each hospital sample. The purpose of this survey was only to characterize

the examinations that were most commonly employed and not to establish their precise

frequencies.

In adult examinations, the assumption that the field size equals the film size is

usually made for organ dose calculations. This assumption is not representative of

pediatric radiography, where an unexposed margin was seen on the majority of films in the

various hospitals surveyed. The field sizes used on the phantom in this survey represent

the fields used in actual clinical practice

This section illustrates the exams included in the site survey in each of the figures.

The left hand illustration is an actual patient x-ray obtained from Dr. Jonathan Williams'

teaching and case file and the right hand illustration is the comparable view of the one-

year-old anthropomorphic phantom. Note that the dosimeters, leads and drill holes are

frequently visible in the phantom images.

The anatomical features shown in the projection in Figure 5-1 demonstrate the

petrous ridges that are projected through the lower third of the orbits. The organs of

hearing and balance are housed in the petrous pyramids. The petrous ridges are the upper

border of these pyramids.




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ESNMQ45XC_VR6CE0 INGEST_TIME 2013-02-14T14:17:29Z PACKAGE AA00012983_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES