<%BANNER%>

Computational Assessment of Effective Dose and Patient Specific Doses for Kilovoltage Stereotactic Radiosurgery of Wet A...

Permanent Link: http://ufdc.ufl.edu/UFE0041964/00001

Material Information

Title: Computational Assessment of Effective Dose and Patient Specific Doses for Kilovoltage Stereotactic Radiosurgery of Wet Age-related Macular Degeneration
Physical Description: 1 online resource (121 p.)
Language: english
Creator: Hanlon, Justin
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: amd, anthropometric, dvh, fovea, macula, mcnpx, phantom, radiosurgery, stereotactic, voxel
Nuclear and Radiological Engineering -- Dissertations, Academic -- UF
Genre: Nuclear Engineering Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Age-related macular degeneration (AMD) is a leading cause of vision loss for people over the age of 50 in industrialized nations. Interest continues for stereotactic radiosurgery, a non-invasive treatment option for the wet form of AMD, through the development of the IRayTM (Oraya Therapeutics, Inc., Newark, CA). The goal of this modality is to destroy choroidal neovascularization beneath the pigment epithelium via three 100 kVp photon beams entering through the sclera and overlapping on the macula delivering up to 24 Gy of therapeutic dose over a span of approximately 5 minutes. A series of head phantoms was derived from CT data and employed in conjunction with the MCNPX 2.5.0 radiation transport code to simulate treatment and evaluate absorbed doses to potential tissues-at-risk. The results indicate that doses to non-targeted tissues were below thresholds for serious complications; specifically the development of cataracts and radiation-induced optic neuropathy.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Justin Hanlon.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Bolch, Wesley E.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041964:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041964/00001

Material Information

Title: Computational Assessment of Effective Dose and Patient Specific Doses for Kilovoltage Stereotactic Radiosurgery of Wet Age-related Macular Degeneration
Physical Description: 1 online resource (121 p.)
Language: english
Creator: Hanlon, Justin
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: amd, anthropometric, dvh, fovea, macula, mcnpx, phantom, radiosurgery, stereotactic, voxel
Nuclear and Radiological Engineering -- Dissertations, Academic -- UF
Genre: Nuclear Engineering Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Age-related macular degeneration (AMD) is a leading cause of vision loss for people over the age of 50 in industrialized nations. Interest continues for stereotactic radiosurgery, a non-invasive treatment option for the wet form of AMD, through the development of the IRayTM (Oraya Therapeutics, Inc., Newark, CA). The goal of this modality is to destroy choroidal neovascularization beneath the pigment epithelium via three 100 kVp photon beams entering through the sclera and overlapping on the macula delivering up to 24 Gy of therapeutic dose over a span of approximately 5 minutes. A series of head phantoms was derived from CT data and employed in conjunction with the MCNPX 2.5.0 radiation transport code to simulate treatment and evaluate absorbed doses to potential tissues-at-risk. The results indicate that doses to non-targeted tissues were below thresholds for serious complications; specifically the development of cataracts and radiation-induced optic neuropathy.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Justin Hanlon.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Bolch, Wesley E.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041964:00001


This item has the following downloads:


Full Text





COMPUTATIONAL ASSESSMENT OF EFFECTIVE DOSE AND PATIENT SPECIFIC
DOSES FOR KILOVOLTAGE STEREOTACTIC RADIOSURGERY OF WET
AGE-RELATED MACULAR DEGENERATION


















By

JUSTIN MITCHELL HANLON


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

2010

































2010 Justin Mitchell Hanlon
































To my Mom and Dad,
for all of the support they have given me over the years









ACKNOWLEDGMENTS

I would like to express my sincere gratitude to my advisor, Dr. Wesley Bolch, for

the opportunity he has afforded me and his guidance towards the completion of my

degree. I would like to extend additional thanks to the remainder of my committee: Drs.

David Hintenlang, Glenn Sjoden, Wesley "Clay" Smith, Choonsik Lee, and Erik Chell.

Erik, not only a member of my committee but an employee of Oraya, has always been a

pleasure to work with on this project. Special thanks are extended to Choonsik, who

shared his immense wealth of knowledge, and educated me to my present level of

expertise. I must also express my appreciation to the rest of the staff at Oraya

Therapeutics for both their financial and valuable research support. Michael Gertner

and Steven Hansen have had a vision for Oraya, and it has been a wonderful

opportunity to watch it grow and to have been a small part of the process. Working with

Michael "Mario" Firpo has been a pleasant experience, as he has always provided

invaluable advice and suggestions, specifically with the revision of several of my

manuscripts. Your keen eye has been an integral part of this project.

I would like to acknowledge NRE staff members Diana Dampier, Ruth Brumbaugh,

Terri Sparks, and Donna Seifert, all of whom have been helpful during my time in the

department. I would like to recognize Dr. David Gilland, a professor for a number of my

graduate courses, for his quality of instruction and engaging classes.

I would like to reflect on my past experiences and thank several others who have

assisted me in reaching my goals and dreams. I can still remember the first person that

inspired my interest in science, and for that I must acknowledge Mrs. Little, my high

school chemistry teacher. I wish to thank her for the invaluable experience of preparing

me for college level research at a very young age. I would like to extend my gratitude to









Dr. George Xu, my advisor at Rensselaer Polytechnic Institute, for the opportunity to

perform several years of undergraduate research, which undoubtedly prepared me for

graduate level work. I would also like to thank Dr. Bryan Bednarz, at the time a

graduate student of Dr. Xu's, who gave me some valuable advice and served as a role

model during my time at RPI as an undergraduate student.









TABLE OF CONTENTS

page

A C KNOW LEDG M ENTS .............................. ........................................... ............... 4

L IS T O F T A B LE S ........................ ................. ........... ..... .............................. 8

LIST O F FIG U R ES .................................................................................. 9

A BSTRACT ........................ ............................................. 11

CHAPTER

1 INTRODUCTION .............................. ............. .................. 13

1.1 AM D Disease ........................ ............. ...... ........... ........... 13
1.2 Current Treatm ents.............................. ............... 16
1.3 IRayTM ...... .. .................................................... ...............18
1.3.1 Description of Kilovoltage Stereotactic Radiosurgery for AMD.............. 18
1.3.2 The Macula and Fovea Offset .................................... .................... 19
1.4 Objectives of This Research .................. ................ ..... 20

2 DIMENSIONAL DATA FOR OCULAR ANATOMY AND THE OPTIC NERVE
PATHWAY VIA 1-mm COMPUTED TOMOGRAPHY IMAGE SETS .................. 27

2 .1 P u rp o s e ............................................................................................................ 2 7
2.2 Data Collection and Measurement Parameters ........................................... 28
2.3 Analysis of CT Data ...........................................32
2.4 Alternative Approach for Future Studies .............................. ................. 35
2.5 Conclusions Pertaining to Development of Models for AMD Treatment
Simulation ................ ......... ....................... 36

3 ANTHROPOMETRIC PHANTOMS EMPLOYED.................................................. 48

3.1 History of Computational Phantoms.................................... 48
3.2 UF NURBS Hybrid Reference Models ........... .... .................................. ..... 49
3.2.1 Head M odel Detail..................... ...................... .................... 50
3.2.2 Ocular Model Detail........................... .................... 51
3.2.3 Optic Nerve Model Detail .............................. ....... 52
3.3 Patient Specific Phantoms .................................. .. ....... 53
3.3.1 Selection and Development ........................ ........... 53
3.3.2 Expanded (3D) Angular Measurements .......................... ...... .......... 54
3.3.3 Patient Specific Treatment Planning ........ ..... ............... ............... 56
3.4 Voxelization ........................... .............. .......... ....................... 58









4 COMPUTATIONAL METHODS .............................. ..................... 70

4.1 The Monte Carlo Radiation Transport Code MCNPX ............ ............... 70
4.2 Monte Carlo Techniques Used for Treatment Modeling............................... 71
4.2.1 Cell and Surface Cards ................................. ... ............... 71
4.2.2 Source Definition .............. ....... ......... ................... 72
4.2.3 Tally Specification ..................................... 72
4.2.4 Material, Mode, and NPS Cards................................... ............... 74
4.3 Post Processing .......................... ...... ........... 75
4.3.1 Calculation of Effective Dose ............... ................................... ......... 75
4 .3.2 U utilizing M esh T ally O utput........................................... ... .. ............... 77

5 TREATMENT OUTCOME EVALUATION AND ANALYSIS................................ 79

5.1 Radiation Dose Thresholds for Complications.............. .... ................ 79
5.2 Reference Model Dose Assessment......... ................................ ....... 80
5.2.1 Tissue-specific Mean Absorbed Doses .............. .... ............... 80
5.2.2 DVH Analysis .......... ......... ......... ............... ........ ....... 81
5.2.3 Effective Dose ......... ......... ........................ 83
5.3 Patient Specific Dose Assessment ...... ............................. 84
5.4 Photon Fluence Evaluation .................. .................... ............... 86

6 C O NC LUS IO N S ............................................ ............. ................ 103

6.1 Limitations of This W ork. ....................................... ................ 103
6.2 G general C conclusions .............................................. ............. 105
6.3 Future W ork ................ ......................... .......... ............ 108

APPENDIX

A EXAM PLE OF MCNPX INPUT CODE........... .............................. .................. 109

B SAMPLES OF MATLAB CODES.............. ........... ........ ....................... 111

LIST OF REFERENCES ............. ........... ..... ....................... ..... .......... 116

B IO G R A P H IC A L S K ET C H ............. ................. ................. .................. ............... 12 1









LIST OF TABLES


Table page

2-1 Statistical summary of ocular length measurement parameters...................... 38

2-2 Statistical summary of optic nerve length measurement parameters ............... 38

2-3 Regression coefficients, number of samples, and R2 value for Gaussian
probability density function for gender-independent measurement parameters. 39

2-4 Regression coefficients, number of samples, and R2 value for Gaussian
probability density function for gender-dependent measurement parameters.... 39

3-1 Comparison of tissue masses in the UF hybrid NURBS and voxel male head
phantoms with those given in ICRP Publication 89 for the reference adult
male.............................. .... ....... ................ 60

3-2 Comparison of tissue masses in the UF hybrid NURBS and voxel female
head phantoms with those given in ICRP Publication 89 for the reference
a d u lt fe m a le .......... .................................... ....................... ..... ............. 6 1

4-1 Tissue weighting factors for the calculation of effective dose as given by
IC R P 103 ............... .... ....... .............. ............................ 78

5-1 Mean absorbed dose (mGy) to various tissues in the reference head models
for a 3 x 8 Gy Oraya Treatment to the right eye ....... ...... ............................... 87

5-2 Mean whole-body absorbed doses DT (mGy) for the estimate of the effective
dose.............................. .... ....... ................ 87

5-3 The gaze angles, voxelized optic nerve volume, and percentage of that
volume receiving more than the absorbed dose listed, representing the dose
distribution, for a cumulative 24 Gy treatment dose to the macula ................ 88

5-4 The voxelized lens volume and percentage of that volume receiving more
than the absorbed doses listed, representing the dose distribution, for a
cumulative 24 Gy treatment dose to the macula................ .... .......... 89

5-5 The voxelized macula volume and percentage of that volume receiving more
than the absorbed doses listed, representing the dose distribution, for a
cumulative 24 Gy treatment dose to the macula................ .... .......... 90

5-6 The highest tissue-averaged doses received from the set of 32 eyes
undergoing treatment simulation and the associated eye model..................... 91









LIST OF FIGURES


Figure page

1-1 Schematic of an axial cross section of ocular anatomy ............... ................... 24

1-2 IRayTM (Oraya Therapeutics, Inc., Newark, CA).............................................. 25

1-3 Schematic of the treatment geometry for a right eye............... .................... 25

1-4 I-GuideTM (Oraya Therapeutics, Inc., Newark, CA)........................................... 26

1-5 Typical retinal geography highlighting the optic disc, fovea, posterior pole,
and O raya Shift.................................................................. ......... 26

2-1 Schematic of an axial cross-section of a pair of eyes (not to scale), indicating
those anatomic parameters obtained via measurement within a single CT
im a g e s lic e ................................................................................ 4 0

2-2 Schematic of a sagittal cross-section of an eye (not to scale), indicating those
anatomic parameters obtained via measurement within multiple CT images..... 41

2-3 CT images of the right orbit of male subject A......................... ............ ......... 42

2-4 Schematic of the Frankfurt plane and normalization parameter ......................... 43

2-5 Dimensions and locations of tissue structures within the human eye as
provided in NCRP Report No. 130................ .............................. ............... 44

2-6 Correlation scatter plot for parameter M1 versus parameter Mi .......................... 44

2-7 Histograms for gender-dependent ocular measurements............................... 45

2-8 Histograms for gender-dependent optic nerve measurements........................... 46

2-9 Histograms for gender-independent measurements................. ........... 47

3-1 The whole body male (left) and female (right) reference phantoms developed
within the Advanced Laboratory for Radiological Dosimetry Studies................. 62

3-2 University of Florida NURBS male (left) and female (right) head models
based on organ masses listed in ICRP Publication 89........................... .... 63

3-3 Dimensions of the tissues structures in the NURBS eye model ......................... 63

3-4 Engineering drawings of the eye detail embedded within the reference
NURBS head m odel ................. .......... ..... .............................. 64









3-5 Male NURBS eye models with five optic nerve variations (red), the macula
targets (green), and the lenses (blue)........................................... ... .................. 65

3-6 Segmentation of the lens (blue), globe (orange), optic nerve (red), brain
(purple), and skull (teal) from a 1 mm axial CT image of the orbital region ........ 65

3-7 Patient specific models in object file format generated from three-dimensional
reconstruction of 1 mm CT data (shown without skin) .................. ............. .. 66

3-8 Scatter plots of optic nerve tilt as a function of gaze angle.............................. 67

3-9 Coronal (left) and sagittal (right) views of the head model voxelized to 1 mm3
resolution ........ ... ...... ............................ ........ ....... ......... 68

3-10 Cropped eye section voxelized to 0.5 mm3 resolution .................. ............. .. 68

3-11 Axial cross sectional view of a patient specific model voxelized to 0.5 mm3
resolution ........ ... ...... ............................ ........ ....... ......... 69

4-1 MCNPX plot of the male eye section model voxelized to 0.5 mm3 resolution..... 78

5-1 DVHs for the 'mean' optic nerves ..... ............... ................... .. ............... 91

5-2 DVHs for the extremes of male optic nerve tilt ....... ... ............................... 92

5-3 DVHs for the extremes of female optic nerve tilt........................ ..... ........ .. 93

5-4 Spatial contour map of the dose distribution within the reference eye model
of the adult m ale ............ .. ... ..... .. .. ......... ............. ... ............. 94

5-5 Dose contour maps for patient model mer............................... ..... .......... 95

5-6 Dose contour maps for patient model fkl ........ ................................... ... 96

5-7 Dose contour maps for patient model fjl ....... ..... ............ ....................... 97

5-8 Correlation scatter plots of mean absorbed dose to the optic nerve as a
function of gaze angle .......... .......... ......... .......................... ........... 98

5-9 Correlation scatter plot and linear regression of optic nerve hotspot dose as a
function of optic nerve thickness......................................... 99

5-10 Phase space diagram of energy and angular dependence of photon fluence
50 cm from the m acula target ...... ...................................................... ............... 99

5-11 Photon fluence distribution plots 50 cm from macula target ........................... 100

5-12 Photon fluence contour maps at the edge of the lattice structure................. 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

COMPUTATIONAL ASSESSMENT OF EFFECTIVE DOSE AND PATIENT SPECIFIC
DOSES FOR KILOVOLTAGE STEREOTACTIC RADIOSURGERY OF WET
AGE-RELATED MACULAR DEGENERATION

By

Justin Mitchell Hanlon

August 2010

Chair: Wesley Bolch
Major: Nuclear Engineering Sciences

Age-related macular degeneration (AMD) is a leading cause of vision loss and a

major health problem for people over the age of 50 in industrialized nations. The

current standard of care, ranibizumab, is used to help slow and in some cases stabilize

the process of AMD, but requires frequent invasive injections into the eye. Interest

continues for stereotactic radiosurgery (SRS), an option that provides a non-invasive

treatment for the wet form of AMD, through the development of the IRayTM (Oraya

Therapeutics, Inc., Newark, CA). The goal of this modality is to destroy choroidal

neovascularization beneath the pigment epithelium via delivery of three 100 kVp photon

beams entering through the sclera and overlapping on the macula delivering up to 24

Gy of therapeutic dose over a span of approximately 5 minutes. The divergent x-ray

beams targeting the fovea are robotically positioned and the eye is gently immobilized

by a suction-enabled contact lens. Device development requires assessment of patient

effective dose, reference patient mean absorbed doses to radiosensitive tissues, and

patient specific doses to the lens and optic nerve.









A series of head phantoms, including both reference and patient specific, was

derived from CT data and employed in conjunction with the MCNPX 2.5.0 radiation

transport code to simulate treatment and evaluate absorbed doses to potential tissues-

at-risk. The reference phantoms were used to evaluate effective dose and mean

absorbed doses to several radiosensitive tissues. The optic nerve was modeled with

changeable positions based on individual patient variability seen in a review of head CT

scans gathered. Patient specific phantoms were used to determine the effect of varying

anatomy and gaze.

The results showed that absorbed doses to the non-targeted tissues were below

the threshold levels for serious complications; specifically the development of radiogenic

cataracts and radiation induced optic neuropathy (RON). The effective dose

determined (0.29 mSv) is comparable to diagnostic procedures involving the head, such

as an x-ray or CT scan. Thus, the computational assessment performed indicates that

a previously established therapeutic dose can be delivered effectively to the macula

with IRayTM without the potential for secondary complications.









CHAPTER 1
INTRODUCTION

1.1 AMD Disease

Age-related macular degeneration (AMD) is a leading cause of vision loss for

people over the age of 50 in the United States and a major health problem worldwide.1

Advanced AMD has dry and wet forms which lead to blurring or blackening of the

central vision while peripheral vision is retained.2 There is no clear cut definition for

AMD; some reserve the diagnosis only for patients who experience vision loss, while

others include patients who have any change (drusen or geographic atrophy) to their

retinal pigment epithelium (RPE).3 Drusen are small yellowish-white deposits found

near the fovea. Utilizing the latter form of the definition, the majority of patients in the

early stages do not experience vision impairment until progression into the advanced

stages of the disease. In fact, drusen are present in over half of the population over 70

years of age.1 Considering the entire population in this age group, 6-8% have the

advanced form of the disease resulting in severe visual loss.4 In the latter stages,

central vision loss is associated with either the general geographic atrophy of the RPE

(dry form) or development of serious and hemorrhagic detachment of the retina and

RPE (wet form).3 The dry form accounts for about 85% of all AMD cases4; however, the

wet form accounts for 80-90% of the cases in the advanced stage resulting in severe

visual loss.5

AMD can have a profound impact on the quality of life of an individual, and

unfortunately, despite its importance and severity, there are limited treatment options.

This is because the pathophysiology of AMD is largely unknown. Research has shown

risk associated with the complement factor H gene. Factor H is a 155 kDa sialic acid









containing glycoprotein that helps regulate complement-mediated immune system

response.6 One study suggests that a single-nucleotide polymorphism in the promoter

region of HTRA1 is a major genetic risk factor for AMD.7 Another more recent study

found that the polymorphism Y402H in the complement Factor H is related to the

development of AMD, and that the pathophysiology of AMD may be related to several

other disorders, suggesting that the formation of drusen may be a systemic and

localized immune system reaction that is observed in several tissues including the eye,

kidney, and similar plaques in the brain.6 However, AMD most likely results from the

behavior of multiple genes, age, and hereditary traits.4 Other risks include

environmental factors, cigarette smoking, diet, fat intake and obesity, high cholesterol

levels, and heavy sunlight exposure.4

A brief description of ocular anatomy is essential to understanding the progression

of the disease (Figure 1-1). The shell of the eye in the posterior region largely consists

of three layers: the retina, choroid, and sclera. The sclera is a fibrous tissue that

protects and encases the eye. The choroid is a dense pigment layer that is rich with

vasculature which supplies the innermost layers of the eye with nutrients and other

essentials. The retina is a thin (~0.5 mm) sensory tissue layer that converts light signal

to nerve signal with photoreceptor cells. The macula is a region of the retinal tissue

surrounding the fovea, which is responsible for central vision. Bruch's membrane is the

inner most layer of the choroid, itself consisting of five layers: the basement membrane

of the RPE, the inner collagenous zone, a central band of elastic fibers, the outer

collagenous zone, and the basement membrane of the choriocapillaries. The RPE, a

part of the retina, provides metabolic support to the photoreceptor cells and transports









metabolic waste from photoreceptor cells through Bruch's membrane to the vasculature

in the choroid (choriocapillaries).8

Drusen are the earliest clinically detectable sign of the disease, from which vision

is retained but some patients report having trouble reading in dim light. There are many

types of drusen hard, soft, semisolid, basal laminar, and calcified all of which typically

form around the fovea.3 In general, drusen are small yellowish deposits that form

between the basement membrane of the RPE and the rest of Bruch's membrane.3

Their manifestation has been linked to free radical formation from visible light which

damages photoreceptor molecules. The RPE is unable to digest the damaged cells

correctly and metabolism is altered, resulting in an abnormal secretion of material from

the basal cell layer.9 Laboratory investigations have found partially digested RPE and

retinal cell organelles within Bruch's membrane and beneath drusen.3 However, the

formation of drusen remains unclear because these findings may be the result, and not

the cause, of drusen.10

In the advanced stage of the disease, the dry form of AMD occurs from atrophy of

the RPE layer below the retina, causing a loss of rods and cones in the central portion

of the eye. The wet (neovascular) form begins with the formation of fibrovascular tissue

from the choroids that breaks through Bruch's membrane.3 This choroidal

neovascularization (CNV) grows beneath the RPE or into the sensory retina. As a

consequence, there is often leakage and bleeding from the vessels that will lead to

increased tension at the macular lesion, resulting in serious or hemorrhagic detachment

of the RPE, fibrovascular disciform scarring, or vitreous hemorrhage.3 These events

lead to severe and rapid vision loss, ultimately causing blindness if left untreated.









1.2 Current Treatments

There are no treatment options for dry AMD; however, vitamin supplements and

antioxidants have been shown to slow the progression of the disease up 25% over a 5

year period.4 Physicians will monitor the dry form closely until the disease progresses

into the more debilitating wet form, for which there are a number of treatments used to

help slow the process of the disease. Current treatment modalities include laser

therapy,11 photodynamic therapy (PDT) using verteporfin (Visudyne, Novartis, Basel,

Switzerland),11-12 intraocular drug therapy with ranibizumab (Lucentis, Genentech, San

Francisco, CA),13 intraocular therapy using pegaptanib sodium (Macugen, OSI-

Eyetech, New York, NY),14 and brachytherapy using beta emitting radiation (Epi-

Rad90TM Ophthalmic System, NeoVista, Fremont, CA).15-16 Many consider the vascular

endothelium growth factor inhibitor (VEGF-inhibitor), ranibizumab, to be the current

standard of care. VEGF inhibitors are commonly associated with cancer treatments

since the antibodies competitively bind with VEGF to prevent the growth of proliferating

blood vessels. AMD is not cancerous but this treatment is applicable because of the

development of choroidal neovascularization that leads to vision loss. The drug is

effective but requires frequent, invasive injections into the eye for an indefinite period, a

burden on the patient and the healthcare system. Research for improved treatment

modalities and combination therapies are ongoing.

Previous to the development of ranibizumab, radiation-based modalities such as

external beam photon therapy,17 external beam proton therapy,18 and Gamma Knife

radiosurgery (GKS)19-21 were explored as potential non-invasive treatment options for

the wet form of the disease. Pilot studies involving the use of both photon and proton

external beam radiotherapy typically employed a treatment scheme where 10-20 Gy is









delivered to the macula in 2-3 Gy fractions.16' 22 Some have produced results of

reduction in vision loss, whereas others have failed to show any benefit, and in some

cases have shown deleterious effects, such as cataract formation and xerophthalmia.22

Nevertheless, there have been sufficient pilot clinical trials to suggest that photon

radiotherapy may be a viable option to treat AMD if higher fractions can be applied to

the macula target, simultaneously limiting non-target tissue toxicity.22 The most

promising study, Bergink et al,23 utilized a treatment scheme delivering a total of 24 Gy

in 4 fractions. The results of that study compare the treatment group with an

observation group and found significant difference (P<0.08) in terms of visual acuity

after 12 months follow-up with no side effects from radiation.

Fractionation schemes in radiation therapy are vital to the efficacy of treatment

and can vary widely depending on the location and type of target. The biological

effective dose (BED) to the target tissue is not only a function of total dose, but the

number of fractions and dose per fraction. Single fraction application is unusual for

cancer tumor treatment, but may have significant impact on the ability to successfully

treat the choroidal neovascularization (CNV) underlying the macula tissue. Char et al,24

reported borderline positive results by delivering a single fraction of 7.5 Gy, and the Epi-

Rad90TM (NeoVista, Fremont, CA) system has been evaluated clinically using a 24 Gy

single fraction treatment scheme.15-16

The described radiotherapy studies have not demonstrated long-term control of

the disease. Considering that brachytherapy requires surgical intervention, and trials

have shown some positive results with delivery of ~24 Gy to the macular lesion, a novel

non-invasive device for radiation treatment of AMD is being developed by Oraya









Therapeutics, Inc. that delivers a higher therapeutic dose (16 24 Gy) in a single

fraction. Preliminary experimental data recently obtained in a mini-pig animal model

suggest that single fraction kilovoltage stereotactic radiosurgery can be accomplished

without adverse effects.25

1.3 IRayTM

1.3.1 Description of Kilovoltage Stereotactic Radiosurgery for AMD

Based on the studies suggesting that radiosurgery may be a viable option for AMD

treatment, the IRayTM (Figure 1-2) has been developed by Oraya Therapeutics, Inc.

(Newark, CA) that addresses many of the inherent limitations of other systems used in

the past.26 This one-time, non-invasive treatment option will provide benefit to patients

and the medical community in terms of cost, pain, and hospital time. The goal of this

modality is to destroy the CNV beneath the RPE via delivery of three 100 kVp photon

beams entering through the pars plana to overlap on the predicted foveal center

delivering up to 24 Gy of therapeutic dose over a span of approximately 5 minutes. The

anode, with 1 mm2 focal spot, is 15 cm from the macula target with 0.75 mm Al and 0.8

mm Be filtration. The x-ray beams targeting the fovea are robotically positioned and the

eye is gently immobilized by a suction-enabled contact lens. The divergence of the

photon beam is characterized by a profile with a diameter of approximately 3.5 mm

upon scleral entry and 4 mm at the retina. The three beams, each intersecting and

delivering 8 Gy at the target, were chosen to disperse the scleral entry dose, the dose to

the edge of the lens closest to each beam, and the dose to the orbital bone and brain

tissue.

The beam geometries can be described using a spherical 3D polar coordinate

system with the z-axis aligned with the geometric axis of the eye, but transposed by an









offset described subsequently. The geometric axis of the eye is defined as the line that

intersects the point on the distal portion of the cornea and is perpendicular to the

corneal curvature at this point (Figure 1-1). The posterior pole is the point where the

geometric axis intersects the retina. For all 3 beam angles, the nominal polar angle is

30 degrees; however, if the scleral entry point of the beam is less than 4 mm from the

limbus for a given patient treatment, the system will readjust that polar angle until the 4

mm criterion is met. The azimuthal angles are described in the coronal plane of the

patient such that 0 degrees is superior to the patient and an angle of 90 degrees would

be towards the nose for a treatment of the right eye. The beam azimuthal entry angles

chosen for prototype therapies are 150, 180, and 210 degrees and are commonly

referred to as the 5, 6, and 7 o'clock beams. Schematics of the beam geometry are

presented in Figure 1-3. The treated eye of each patient is gently immobilized by a

suction-enabled contact lens with a central post and a control-yoke housing three

optically sensed fiducial markers (I-GuideTM) (Figure 1-4). Motion of the patient's eye is

substantially reduced by the I-GuideTM and the residual motion is tracked in real time

using a two-camera imaging system. Eye motion that would result in substantial dose

outside the target area triggers an interruption of the x-ray beam. Targeting error due to

patient motion is monitored and maintained below 400 pm on the retina.

1.3.2 The Macula and Fovea Offset

In ocular anatomy, the fovea is known to be offset from the posterior pole of the

eye (Figure 1-1). An evaluation of this separation was performed in house at Oraya

Therapeutics using a customized Canon CR-45 UAF non-mydriatic funds camera with

a low-power laser beacon and an anterior imaging system with added collinear at the









imaging axis. When a subject's eye is oriented such that the reflection of the laser

beacon and the center of the limbus boundary coincide with the system axis as seen in

the anterior eye view, the laser beacon position in the funds image occurs at the

intersection of the geometric axis with the retina (the posterior pole).

Analysis of these funds images from seven healthy volunteers showed that the

nominal fovea is located 1.25 mm laterally and 0.50 mm inferiorly from the posterior

pole. This fovea offset, referred to internally to the research team as the Oraya Shift, is

important because the current treatment involves aligning the system to the geometric

axis of the eye, followed by a translation of the device by the offset in order to target the

nominal fovea. A retinal geography of the fovea offset is shown in a representative

funds image in Figure 1-5. This offset is important not only for more accurate targeting

of subfoveal disease, but also because it moves the treatment beams away from the

optic disc, substantially reducing optic nerve dose.

Further studies have confirmed the validity of the offset. A total of 48 additional

eyes, both healthy and low vision, were analyzed for the position of the fovea relative to

the posterior pole.27 Similar results were obtained. As part of a larger targeting study,

thirteen cadaver eyes were dissected and the position of the fovea was compared to the

positions of needles placed through the retina at the point of treatment. The average

position of the fovea was found to be within 100 Vm in the lateral-medial direction and

100 Vm in the superior-inferior direction of the nominal fovea as defined by the offset.27

1.4 Objectives of This Research

Dosimetry characterization of this treatment is an integral part of device

development, risk assessment, and the approval process for the Federal Drug









Administration (FDA). Previous Monte Carlo radiation transport simulations were used

to provide insight into beam characterizations for optimal therapy applications including

focal spot size, maximum tube potential, and azimuthal angles of beam entry.28 In the

present work, the dosimetry characterization of the kilovoltage stereotactic AMD

radiosurgery has been expanded and enhanced in a number of areas.

Chapter 2 describes the collection and analysis of 40 head CT scans. Evaluation

of the ocular anatomy and optic nerve pathway is presented through the statistical

analysis of several measurement parameters. The results of this study provide a better

quantitative understanding of the optic nerve pathway and indicate the necessity to use

separate models for the detailed ocular anatomy of the male and female during

treatment simulations. The range of optic nerve exit tilt angles observed was utilized to

evaluate a worst-case-scenario risk assessment.

Chapter 3 highlights the creation of a reference NURBS-based model structure

and the fabrication of a 16 patient, 32 eye set of patient specific voxel models. In the

process, both a reference adult male and reference adult female head model were

constructed consistent with the anatomical data of the International Commission on

Radiological Protection's (ICRP) Publication 89.29 The detailed models include the

entire head and neck of the patient including several radiosensitive tissues at potential

risk, namely the lens, optic nerve, brain, cranial bone marrow, cranial endosteum,

thyroid, and salivary glands. These tissues allow for the assessment of effective dose

from this treatment so that comparisons of relative stochastic risk may be made against

other medical imaging and therapy procedures. The detailed ocular anatomy combines

data from Chapter 5 of Report No. 130 by the National Council on Radiation Protection









and Measurement (NCRP)30 and ICRP 89. Most importantly, the optic nerve pathways,

for both male and female and including both mean and extremes of optic nerve exit tilt,

were modeled from data in Chapter 2 of this work. Patient specific models were

designed to evaluate dose as a function of varying ocular anatomy and gaze angle.

Chapter 4 details the Monte Carlo methods used to simulate treatment and

Chapter 5 presents dosimetry calculations in the form of tissue-specific mean absorbed

dose tables, dose volume histograms (DVHs), color coded dose contour maps, and

absorbed dose distribution tables. These latter tables are an alternate data format

comparable to DVH plots, but allow for condensed presentation and listing of specific

quantitative values. Effective dose is calculated using the reference adult male and

female head phantoms for a 24 Gy treatment to the macula region. Contributions from

both the primary tube output and an estimation of leakage are included in the effective

dose calculation. For the patient specific phantoms, trends in dimensional anatomy as

a function of absorbed dose are presented and analyzed. Treatment evaluation and

analysis includes a comparison between absorbed dose to non-target tissues observed

in this study and the generally accepted thresholds for complication and debilitation,

specifically the development of cataracts and radiation induced optic neuropathy (RON).

Lastly, the energy and angular distribution of photon fluence at a radius of 50 cm from

the macula target is presented, along with photon fluence contour maps that provide a

visual representation of photon fluence at the edge of the lattice structure used during

treatment simulation. The conversion of fluence to dose rate will provide parameters

necessary for shielding calculations.









Ultimately, the work of this research will provide a better understanding of the

treatment physics and risk of non-invasive kilovoltage stereotactic radiosurgery for

AMD.

























Medial rectU


Optic dis
DI ad rm al -
Suaracnola Vt'pC-
OptX nerve Wurrouwned
wiir pia maier


Latera rectus
tendon


Figure 1-1. Schematic of an axial cross section of ocular anatomy (adapted from Snell8)



























Figure 1-2. IRayTM (Oraya Therapeutics, Inc., Newark, CA)


A B














C D


Figure 1-3. Schematic of the treatment geometry for a right eye (A) sagittal view (6
o'clock beam only) (B) sagittal view zoomed in (C) front view (D) perspective
view


b






0" ac bein



























Figure 1-4. I-GuideTM (Oraya Therapeutics, Inc., Newark, CA)


Figure 1-5. Typical retinal geography highlighting the optic disc, fovea, posterior pole,
and Oraya Shift


ri









CHAPTER 2
DIMENSIONAL DATA FOR OCULAR ANATOMY AND THE OPTIC NERVE PATHWAY
VIA 1-MM COMPUTED TOMOGRAPHY IMAGE SETS

2.1 Purpose

Dimensional data on the eye and optic nerve are critical parameters in the

development of radiation treatments for eye disease. Device development and safety

assessments require not only the central estimates, but also gender-dependent

variability in a potential patient population. Such data, however, are limited in the open

literature. Some reference data are given for ocular size measurements in Chapter 5 of

Report No. 130 by the National Council on Radiation Protection and Measurement

(NCRP),30 and in Chapter 11 of Publication 89 by the International Commission on

Radiological Protection (ICRP).29 The NCRP 130 eye model is in part based on

measurements reported by Charles and Brown,31 yet other information in the model is

unpublished.32 Furthermore, ICRP 89 reference values are limited to values of only

total eye and lens mass. In both sources, no mention is made of gender-dependent

variations in ocular structure size or optic nerve pathways.

The anatomical location and function of the optic nerve is well known in a

qualitative manner, yet limited data exist to quantify the position of the optic nerve and

its pathway within the tissue structures of the human head. The optic nerve can be

distinguished in both CT and MR images, but is difficult to image in any one slice

because of its shape and size.33 There are other imaging modalities, such as Optical

Coherence Tomography (OCT), that are specialized in imaging the optic nerve for

glaucoma-related studies, but these tend to focus on the optic disc rather than the optic

pathway.34 While there is some quantitative information on the movements of the ocular









muscles,35 limited knowledge exists regarding the optic nerve's movements as a

function of eye gaze angle.

In the past, such quantitative knowledge was not necessary for invasive surgical

medical procedures involving the eye and its orbit since these structures can easily be

localized by the ocular surgeon. Recently, the development of non-invasive stereotactic

radiosurgery has become increasingly popular in medical therapies. Non-invasive

surgery is a benefit to the patient when considering pain, medical costs, and hospital

time. Research for a new type of ocular radiosurgery is underway for the treatment of

age-related macular degeneration (AMD) which has shown to have numerous benefits

over other types of treatments for this disease.28 36 Due to attenuation of the x-ray

beams passing through the eye, the proximity of the macula to the optic nerve and disc,

and the possibility of damage to these structures from radiation during treatment, the

ability to quantify ocular and optic nerve tissue structures is now needed for medical

device development and treatment risk assessment.

To address these needs, a retrospective study of 1-mm slice resolution computed

tomography images has been undertaken for a 40-patient population of equal numbers

of males and females. The study explored differences between male and female eye

and optic nerve sizes, which may affect the attenuation of radiation beams passing

through the eye during stereotactic radiosurgery. The study also examined optic nerve

pathways in both genders to ascertain correlations with eye position, as the location of

the optic nerve is of utmost importance during radiosurgery treatment planning.

2.2 Data Collection and Measurement Parameters

With Institutional Review Board approval (IRB #481-2007 University of Florida), 40

CT scans were obtained from Shands Hospital at the University of Florida for









retrospective analysis. The gender distribution was 20 male and 20 female. The

requirements for eligible CT sets included (1) maxillofacial axial scans, (2) 1-mm slice

resolution, (3) soft tissue contrast settings, and (4) patient age over 18 years. The CT

scans were analyzed and measurements made using the image processing code 3D

DoctorTM (Able Software Corp., Lexington, MA).

Measurements pertaining to the optic nerve included (1) vertical tilt angle of the

optic nerve as it leaves the posterior region of the eye, (2) optic nerve thickness at the

posterior region of eye sclerall optic nerve thickness), (3) optic nerve thickness as it

passes through the orbit (orbital optic nerve thickness), and (4) optic nerve length from

posterior of eye to the posterior region of the orbit. These 4 optic nerve measurement

parameters will be referred to as M1 to M4, respectively. M1 is positive in the superior

direction and negative in the inferior direction.

Ocular measurements included (a) apex of cornea to lens distance (corneal

depth), (b) lens depth, (c) lens width, (d) eye depth, (e) eye width, (f) combined sclera,

choroid, and retinal thickness in the posterior hemisphere of the eye (tri-layer

thickness), (g) eye separation from apex of the right cornea to apex of the left cornea,

(h) angle between the Frankfurt Plane (defined below) and the axial plane of the CT

image (head tilt angle), and (i) vertical gaze angle. These 9 ocular measurement

parameters will be referred to as Ma to Mi, respectively. The parameters measured on

axial images are shown graphically in Figure 2-1.

M1 was calculated using the trigonometric relation tan'(X/Y) where X is obtained

by counting the number of CT slices between the axial image that showed the inferior

portion of the optic nerve exiting the posterior region of the eye and the axial image that









showed inferior portion of the optic nerve as it exits the orbit, and Y is the compressed,

2D length of the optic nerve. Measurement M4 was calculated using Pythagorean's

Theorem with X and Y given for M1. A representation of this method is shown in Figure

2-2A. In many cases, the optic nerve appeared to significantly change direction at

some point in its pathway (i.e., the optic nerve displayed some degree of slack in its

pathway at more central gaze angles). In such a case, the angle in M1 was made to the

inflection point, rather than to the posterior of the orbit to accurately depict an exit angle.

A representation is shown in Figure 2-2B. The inferior side of the optic nerve was

chosen for measurements because a clearer distinction could be made at the optic

nerve-sclera junction as compared to its superior side.

Figure 2-3 demonstrates the method for these measurements and why it is

sometimes important to measure to an alternate point in order to truly represent an

appropriate exit angle. Viewing the CT images in an inferior-to-superior order, the optic

nerve can be first seen clearly in slice 135 of Figure 2-3, but it does not connect to the

eye until slice 137, yielding X = -2 and Y represented by the black line in slice 135. If

the angle had been measured to the inferior portion of the optic nerve as it exits the

orbit, as highlighted by the black circle in slice 140, X would have been +3, yielding a

superiorly tilted angle when in fact the exit angle is inferiorly tilted.

Many of these measured structures can be seen on multiple images with 1-mm

slice CT resolution. To keep the ocular measurements (Ma through Mg) consistent from

patient to patient, a reference slice was chosen that could be easily identified within

every CT image set. Typically, the lens could be identified on ~9 CT images, and so the

median slice of this subset was selected as the reference plane. Due to right-left head









tilt, this slice may not be the same for the left and right eyes and, in such a case,

measurements to the left and right eye were made on separate reference images, and

Mg was made on either the right or left eye reference slice, whichever had the best view

of the opposing cornea. M2 and M3 were made on whichever image showed the largest

thickness.

The reference plane described is the optimal plane for measurement because it

would contain the geometric axis assuming the patient scanned had no head tilt or gaze

angle, and it is assumed that the eye is approximately rotationally symmetric around its

geometric axis. Under this assumption, the lens and eye width measured correspond to

a lens and eye diameter. However, most patients display some form of head tilt and

gaze angle within their CT images, and so these parameters were calculated as well.

To correct for patient head tilt, M1 and Mi were normalized to the Frankfurt plane.

Figure 2-4A shows the Frankfurt plane, defined as an axial plane intersecting both the

inferior point of the boney orbit and the superior point of the ear canal. The Frankfurt

plane is described as being most nearly parallel to the Earth's surface for a person in an

upright reference position. A correction factor was measured by performing 3D

reconstruction of the patient's skull using an interactive segmentation tool in 3D

DoctorTM. The resulting polygon mesh file was exported to Rhinoceros 4.0TM (McNeel

North America, Seattle, WA), as shown in Figure 2-4B, to make an angular

measurement between the Frankfurt plane and scanning plane.

Parameter Mi (vertical gaze angle) was calculated similarly to M1, by using the

trigonometric relation tan- (A/B) where A is determined from counting the number of CT

slices between the median slice of the lens (reference slice described above) and the









medial slice of the optic nerve in the posterior region of the eye (slice contains a good

approximation for the posterior pole even though it cannot be visualized in CT images),

and B is the compressed, two dimensional length between the two structures (distance

measurement made on one of the images between the lens and optic nerve).

Anatomically, A is in the superior-inferior direction (z-direction), B is in the axial plane (x-

y plane), and Mi is the angle between the two. A representation of this measurement is

shown in Figure 2-2C.

2.3 Analysis of CT Data

Parameters Mi and Mh were merely evaluations of patient head and eye

positioning during the CT imaging with little correlation between subjects and thus are

not relevant to gender-dependent discussions. The standard deviation for vertical tilt

(Mi) was quite high being 7.8 degrees for men and 8.4 degrees for women. Slice

resolution and contrast settings were potential sources of error for parameter M1,

although this parameter could also be considered arbitrary since gaze angle was not

fixed. Consequently, it is useful to give the range for this measurement, which was from

-17.6 (inferior tilt) to 15.5 degrees (superior tilt) for the male subjects and from -24.4

(inferior tilt) to 9.7 degrees (superior tilt) for the female subjects. A medial-lateral tilt

angle was estimated and ranged from +18.9 to +28.5 degrees, but was not included in

statistical analysis because of the absence of consistent anatomical landmarks within

the 2D image set to normalize amongst subjects. These angular measurements did not

indicate any patterns that would suggest a difference between right and left optic

nerves, nor between males and females.

Excluding M1, Mh, and Mi, the means, standard deviations, uncertainty, and ranges

for ocular and optic nerve parameters are given in Tables 2-1 and 2-2, respectively.









The tables also give the t-values and p-values obtained from performing an unpaired

Student's t test between the male and female sample populations. Choosing a

statistical significance threshold of 0.05 for the p-value (95% confidence) to reject the

null hypothesis, it was determined that there is a gender-dependent difference in means

for the following parameters: scleral optic nerve thickness (M2), orbital optic nerve

thickness (M3), optic nerve length (M4), eye depth (Md), eye width (Me), and eye

separation (Mg). For these parameters, gender-dependent means, standard deviations,

uncertainty propagations, and ranges are also given in Tables 2-1 and 2-2. The

measurements did not indicate patterns that would suggest a difference between right

and left eye structures. Uncertainty for each measurement was determined from the

diagonal length through a pixel (0.5 mm) and slice resolution (1 mm) of the image set,

and propagated to determine the total uncertainty in the mean.

The NCRP 130 eye dimensions are shown in Figure 2-5. Comparing current

measurements of Ma through Mf and M2 to corresponding NCRP 130 measurements,

parameters Ma, Mb, Md, and Mf are in agreement, while parameters Mc, Me, and M2

indicate that current determination of the diameters of the lens, eye, and optic nerve are

smaller in comparison to those given in the NCRP report.

To establish a relationship between vertical optic nerve tilt and vertical gaze

angle, a scatter plot and linear regression equation are shown in Figure 2-6. The

relationship established was not strong enough to determine a definitive quantitative

relationship (R2=0.3377), but there is some correlation between the two parameters.

Thus, the vertical optic nerve exit angle is related to the vertical gaze angle, which can

be visualized from the three diagrams drawn in Figure 2-2. When looking straight









ahead in the primary gaze position, there appears to be "slack" in the optic nerve. This

slack allows the optic disc and nerve to move with the eye when shifting gaze. For an

upward gaze, the optic nerve would have a superiorly tilted exit angle, and for a

downward gaze it would have an inferiorly tilted exit angle. Furthermore, the length

measurements on the anterior portion of the optic nerve give some idea of where a dip

normally occurs in the optic nerve slack (14.3 mm for men and 9.6 mm for women).

Figures 2-7 and 2-8 give histograms for gender-dependent ocular and optic nerve

parameters, respectively. Figure 2-9 gives histograms for the remaining gender-

independent parameters. The bin width in each plot was determined by the following

expression:

W = 3.49 oN-1/3 (2-1)

where W is the bin width, o is the standard deviation, and N is the number of samples.

The expression gives the optimal bin width for the most efficient unbiased estimation of

the probability density function.37 A three-parameter Gaussian probability density

function was derived for each histogram using the following model:


a exp -0.5 X 2
pdf = (2-2)
N

where a, b, Xo are fitting coefficients and N is the number of samples. These values,

along with the R2 values of the curve, are listed in Tables 2-3 and 2-4 for the ocular and

optic nerve parameters, respectively. The original Gaussian fit line was normalized by

the number of samples to give a probability density function ranging from 0 to 1. The

regression lines are statistically strong for most parameters, with R2 values above 0.95









for each parameter with the exception of M2 and M4 (for males only), which were 0.88

and 0.91, respectively.

2.4 Alternative Approach for Future Studies

While this study provides improved data for these applications, additional

improvements are warranted in future studies. CT image contrast is dependent on both

the x-ray tube potential (kVp) and tube current (mA) used during patient imaging. As

this was a retrospective study, these values were not always consistent from patient to

patient. As image contrast varies, it became difficult to differentiate structures with

similar Hounsfield numbers, such as optic nerve fibers and myelin sheath, leading to

potential variability during the measurement process. Another factor to consider in this

analysis was the 1-mm resolution of the CT images, a value chosen because it is one of

the highest resolutions readily available for head CT studies. While higher resolution

image sets are occasional taken, they are uncommonly administered in order to reduce

radiation dose to the patient. Consequently, the 1-mm resolution offered the highest

resolution available for which a relatively large patient population could be sampled.

Clearly higher resolution CT image sets would vastly improve some of the measured

parameters, namely the vertical optic nerve exit angle and the vertical gaze angle. It

should be noted, however, that 3D reconstructions of 2D images, will not provide

improved measurement accuracy, as the 3D images involve some data smoothing.

Improved accuracy will only be accomplished via higher image resolution, at the

expense of patient radiation dose.

An alternative to the retrospective study of computed tomography images

presented here would be a prospective study of patients undergoing magnetic

resonance (MR) head imaging. MR images would give improved soft-tissue contrast,









and allow for higher resolution slices to be taken without having to account for volunteer

radiation dose. Designing such a study would offer a number of other benefits as well,

such as limiting head tilt, and scanning patients with forced primary gaze angles as

would be the case during stereotactic radiosurgery for AMD. Data from such a study

would give a more direct correlation between optic nerve exit angle and gaze angle, and

may support the optic nerve movement theory presented here. Relatively long scan

times, however, would be required to achieve images approaching the 1-mm slice

resolution used in the present CT-based study.

2.5 Conclusions Pertaining to Development of Models for AMD Treatment
Simulation

There is currently a need in the medical community for gender-dependent

dimensional data for the eye and substructures, as well as for the optic nerve anatomic

position and pathway. Such data are important for the development of non-invasive

medical procedures such as stereotactic radiosurgery for treatment of wet AMD. The

present study examined forty 1-mm head CT image sets, from which gender-dependent

means, standard deviations, uncertainty, and ranges were obtained for several

anatomic parameters and a relationship between vertical optic nerve exit angle and

vertical gaze angle was established.

From the results of the sample population, it could be determined with 95%

confidence that there is a difference in gender for the total population for parameters

M2, M3, M4, Md, Me, and Mg, while there is no difference between male and female

parameters M1, Ma, Mb, Mc, and Mf. Therefore, the average male and female have the

same distribution of optic nerve exit angles in the vertical (Mi) direction, the same

position (Ma) and size (Mb and Mc) of the lens within the eye, and the same combined









sclera, choroid, and retinal thickness (Mf). As for the parameters with a notable

difference, an average male optic nerve is longer (M4) and thicker (M2 and M3) than an

average female optic nerve, making the male optic nerve slightly more difficult to avoid

during ocular stereotactic radiosurgery. An average male eye is larger in depth (Md)

and width (Me) than that in the average female, which will more heavily attenuate a

radiation beam traversing the eye. Parameter Mg was also larger for males by a more

significant margin indicating a larger skull, which corroborates information on gender-

dependent reference values given in ICRP Publication 89. As a result, a reference

computational model for pre-clinical dosimetry evaluations for AMD kilovoltage

radiosurgery should employ separate models, particularly for the optic nerve detail, for

male and female patients.











Table 2-1. Statistical summary of ocular length measurement parameters
Ma Mb Mo Md Me Mf Mg
mm mm mm mm mm mm mm
3.18 3.82 8.28 24.27 25.11 1.22 65.47
Total mean
0.06 0.06 0.06 0.06 0.06 0.06 0.08
s 0.49 0.56 0.79 0.97 1.22 0.23 4.53
min 2.0 0.5 2.7 0.5 6.6 0.5 21.8 0.5 22.0 0.5 0.8 0.5 56.6 0.5
max 4.6 0.5 5.5 0.5 10.7 0.5 26.9 0.5 28.2 0.5 2.1 0.5 77.2 0.5
t value -0.613 -0.052 1.055 2.635 2.584 1.589 3.122
p value 0.541 0.959 0.295 0.010 0.012 0.116 0.003
24.54 25.45 67.49 +
Men mean
0.08 0.08 0.11
s 0.98 1.09 4.85
min 22.4 0.5 23.8 0.5 60.0 0.5
max 26.9 0.5 28.2 0.5 77.2 0.5
23.99 24.77 63.45 +
Women mean
0.08 0.08 0.11
s 0.88 1.26 3.16
min 21.8 0.5 22.0 0.5 56.6 0.5
max 25.7 0.5 27.6 0.5 68.9 0.5


Table 2-2. Statistical summary of optic nerve length measurement parameters
M2 M3 M4
mm mm mm
Total mean 4.97 0.06 3.61 0.06 28.67 0.07
s 0.89 0.44 3.03
min 3.2 0.5 2.5 0.5 22.7 0.7
max 7.0 0.5 4.9 0.5 37.8 0.7
t value 3.760 2.034 5.412
p value <0.001 0.045 <0.001
Men mean 5.32 0.08 3.71 0.08 30.25 0.10
s 1.01 0.45 3.07
min 3.2 0.5 2.6 0.5 24.3 0.8
max 7.0 0.5 4.9 0.5 37.8 0.7
Women mean 4.63 0.08 3.51 0.08 27.10 0.09
s 0.58 0.42 2.03
min 3.7 0.5 2.5 0.5 22.7 0.7
max 6.0 0.5 4.7 0.5 30.7 0.5










Table 2-3. Regression coefficients, number of samples, and R2 value for Gaussian
probability density function for gender-independent measurement parameters
Ma Mb M, Mf M1
a 25.25 24.06 26.84 28.15 23.94
b 0.51 0.61 0.75 0.21 8.89
Xo 3.19 3.82 8.26 1.18 -1.93
N 80 80 80 80 80
R2 0.98 0.99 0.98 0.97 0.96



Table 2-4. Regression coefficients, number of samples, and R2 value for Gaussian
probability density function for gender-dependent measurement parameters
Md Me Mg M2 M3 M4

Men a 20.46 16.13 10.41 15.98 16.32 15.95
b 0.71 1.11 4.61 1.03 0.44 3.04
Xo 24.46 25.35 67.17 5.45 3.66 30.05
N40 40 20 40 40 40
R2
0.97 0.97 0.97 0.91 0.99 0.99

Women a 15.19 15.97 9.61 15.97 18.73 15.27
b 0.94 1.26 3.38 0.58 0.34 2.26
xo 23.95 24.70 62.63 4.54 3.43 27.04
N40 40 20 40 40 40
R2
0.99 0.96 0.95 0.99 0.98 0.96





























Axial View


Figure 2-1. Schematic of an axial cross-section of a pair of eyes (not to scale),
indicating those anatomic parameters obtained via measurement within a
single CT image slice















i A -A











B





_I










Sagittal View


Figure 2-2. Schematic of a sagittal cross-section of an eye (not to scale), indicating
those anatomic parameters obtained via measurement within multiple CT
images (A) superior gaze (B) primary gaze (C) inferior gaze

















I/


Figure 2-3. CT images of the right orbit of male subject A; slice progression from 134 to
142 is inferior to superior



























-'- Orbitomeatal Plane



































Figure 2-4. Schematic of the Frankfurt plane and normalization parameter (A) common
reference planes in the head (adapted from Anthoney ) (B) 3D reconstructed
polygon mesh model of the skull demonstrating measurement of head tilt with
respect to the scanning plane
.... .................. .... .................... ................... ...............................

........................


















polygoiiiiiiiiiiiiiiiiiiiiin m s o e ftes uld m n tatn e s rm n fh a itwt























05510 1
o. 1460
S 2.80.8 4.2+1 0 40

5 0+0.5
5.904 150.5 ,*"



0.5
10






Figure 2-5. Dimensions and locations of tissue structures within the human eye as
provided in NCRP Report No. 130; the geometric axis is indicated by the red
dashed line; all dimensions are in mm


Vertical Optic Nerve Tilt versus Vertical Gaze Angle


.o *




.. .
e o


0*
a,
c15


. *


*
*


M1 = 0.4521 M, + 2.4578
R2= 0.3377


-30 -20 -10 0 10 20 30


Mi (degrees)


Figure 2-6. Correlation scatter plot for parameter M1 versus parameter Mi


*^
















Eve Dfaplh fMOep


/









21 22 23 24 25 2 27
Parumnter Value (rmm)


IL ,


25




j 15









14

18



4





2

0


21 22 23 22 3 26 2 27 28 29
Prlmetur Value (mm)


5 Pamr 60 65 70mm
Parrmwtor Vlpka (mm)


E


75 80


Eye Dpth (PcenaMJ













21 22 23 24 25 28 27
Pasmrwrr VIlue (mm)


Pnl Pwnmr.w M*
Eye WMfl (FPmafd)












21 22 23 24 25 26 27 28 28
Pimranb r Value (mm)


Panramter MF
Ey* S*FiWraton (Fewmet


S5 60 6 7 75I 80
Parameter Value (rm)


Figure 2-7. Histograms for gender-dependent ocular measurements: (A, B) eye depth,
(C, D) eye width, and (E, F) eye separation for males and females,
respectively


A
0"






















A

A


3.0 3.S 4.0 4.5 5.0 56 6.0 6. 7.0
Pararnater Vdkru (mmi


2.5 30 3.5 40 4.5 5.0
Parwmeter Velu* (mrn)


20 22 24 26 28 30 32 34 32 38 40
Paramter Value (mm)


15


to





C
C


SCltanf (pfle I


4I



14


2)


15





0


0


Pormat pri N
Orwhit. Op*









D


/


I.s 3.0 .S 4.0 4.S 0m)
Parameer Value (mm)


1B
15
14
12
0I"
a


A

E 2
0


18
10
14
12


L *
a


2
0



16
'B
4
12




IL

2


Figure 2-8. Histograms for gender-dependent optic nerve measurements: (A,B) optic
nerve thickness at sclera (C,D) optic nerve thickness at orbit, and (E,F) optic
nerve length, for males and females, respectively


3A0 35 4,0 45 5&0 55 0 ,B 7,'
Paruaeter Value (mm)


Pa ameeia Ma
Optre Nerve Length











2 22 Pa4 20 rlm 32a 3 3 4mm)
Parameter Value iMm)


V






























2.5 3.0


2.0 2.5 3.0 3.5 4.0 4.5 5.0
Parmet.rt Value (mm)


Fe


3. 4 4. 4.S 5.0 S.S 6.0
Parametar Value [mm]


6 7 8 9 10 11 0, 1t0 12 14 15 18 20 22
Parwmetr Value (rmm Paramotr Value (rm)

Paramelr"M





15


10






Pararmtr Value [degrm



Figure 2-9. Histograms for gender-independent measurements: (A) corneal depth, (B)
lens depth, (C) lens width, (D) combined scleral, choroidal, and retinal layer
thickness, and (E) vertical optic nerve tilt angle


Parame&er Mb
L mos Depoi









CHAPTER 3
ANTHROPOMETRIC PHANTOMS EMPLOYED

3.1 History of Computational Phantoms

While the human body's response to ionizing radiation is well documented in

literature, direct measurement of absorbed dose in living tissue is not possible. One of

the most powerful techniques used to estimate organ doses is through the use of Monte

Carlo radiation transport codes with computation anthropometric phantoms, which

generally consist of three types: (1) stylized phantoms described by 3D geometric

surface equations, (2) voxel phantoms defined by a set of voxels segmented from

medical images, and (3) hybrid phantoms constructed from NURBS surfaces.39

The Medical Internal Radiation Dose (MIRD) phantom was the first stylized

phantom introduced in 1969, but only included three tissue regions (bone, lung, and soft

tissue).40 The evolution of stylized phantoms advanced to include several age groups

and match data from International Commission on Radiation Units (ICRU) Publication

2341 with the construction of the Oak Ridge National Laboratory (ORNL) phantoms in

1980.42 Stylized phantoms have several advantages such as smooth organ surfaces

and flexibility, but are limited to basic geometric shapes such as quadrilaterals,

ellipsoids, cylinders, and spheres. Thus, the phantom world evolved to better model the

complex and intricate nature of human anatomy through the development of voxel

model phantoms.

Voxel phantoms are generated through the contouring of tissues in medical

images. The data from these two-dimensional images are stacked to create a three-

dimensional matrix of tissue types. This technique is commonly known as segmenting

and allows for much more detailed modeling of human anatomy than stylized phantoms









can offer. Several independent groups have generated voxel models; however they are

not useful for universal distribution because they are not easily deformable and

therefore cannot be directly matched to reference values without more complex model

manipulation. As such, voxel models are mostly limited to patient specific anatomy.

Additionally, surfaces described by voxel geometry are not smooth because voxel

models are created from the compilation of many small rigid shapes (usually cubes),

and a "stair" artifact is observed near the edge of each tissue. Therefore, the quality of

a voxel phantom relies heavily on its voxel resolution, but high resolutions lead to long

computer run times during radiation transport.

The third type of computational phantoms devised, hybrid phantoms, attempt to

combine the best features of each of the previous two types. The use of non-uniform

rational B-spline surfaces (NURBS) allows for a more complex mathematic model of

organ surfaces than stylized phantoms can offer.39 NURBS-based surfaces are defined

from a series of control points that can be altered independently for non-uniform scaling

and the construction of more complex geometries. A major disadvantage of hybrid

phantoms is that they cannot be used directly with radiation transport codes and require

a process known as voxelization. Thus, for dosimetry purposes, the end result of the

hybrid phantom is a voxel model, but with the advantage of having scaled tissue

volumes to a desired reference percentile and a user defined voxel resolution. The

complete process of voxelization is described subsequently in section 3.4.

3.2 UF NURBS Hybrid Reference Models

Formulation of a male and female whole-body reference phantom was completed

within the Advanced Laboratory for Radiation Dosimetry Studies (ALRADS) research

group at the University of Florida previously (Figure 3-1). The phantoms were









constructed from (1) segmentation of patient CT images using 3D DoctorTM which allows

the regions of interest in axial CT slices to be highlighted, and (2) modification of the

resulting model to match ICRP Publication 89 50th percentile dimensions. Following

these general procedures, the skeleton, body contour, major organs and tissues were

segmented from the patient CT data and imported into Rhinoceros 4.0TM for conversion

to NURBS surfaces. The detailed methodology of the entire process has been given

previously by the ALRADS research group.39, 43

3.2.1 Head Model Detail

Two head phantoms, male and female, were extracted from the full body models to

be used for kilovoltage stereotactic radiosurgery simulation and are shown in Figure 3-

2. As noted above, two versions of hybrid phantoms exist: (1) the NURBS phantom

constructed from CT image segmentation and volumetrically adjusted to match

individual reference tissue volumes and (2) its voxelized counterpart. Table 3-1 gives

the final tissue masses for both the hybrid-NURBS and hybrid-voxel adult male head

phantoms voxelized at 1 mm x 1 mm x 1 mm resolution. Percent differences from ICRP

89 reference values are indicated along with the ICRP 89 targeted tissue mass and

reference densities as given in ICRU Report 46.44 Corresponding comparisons for the

UF hybrid female head phantom are shown in Table 3-2. Several tissues are not of

dosimetric concern in this study, but are shown for completeness. Masses for the

esophagus and spinal cord are indicated as partial masses (fractional volume contained

within the head phantoms). In general, targeted masses are achieved within 1-2% of

reference values for both genders.









3.2.2 Ocular Model Detail

The eyes have a combined reference mass of 15 g and a combined volume of

14.56 cm3, and the lenses have a combined reference mass of 0.5 g and a combined

reference volume of 0.42 cm3.29 ICRP Publication 89 only gives reference mass and

density for the lens and total eye structure, so additional ocular detail was added based

on dimensions given in NCRP Report No. 130. However, the dimensions used in

NCRP Report No. 130 are not fully compatible with ICRP Publication 89 reference

values, and so the dimensions used to design this model were modified slightly to

ensure consistency with ICRP data. The NURBS eye model with dimensions is shown

in Figures 3-3.

Anatomically, the macula is a region of the retina surrounding the fovea. Its small

size does not permit its direct segmentation in CT images, and thus it must be

computationally modeled as a separate structure for dosimetry purposes. In this study,

the macula was modeled as a cylinder 4 mm in diameter and 0.5 mm in thickness

(retinal thickness from NCRP 130 as given in Figure 2-5). The macula is placed within

the posterior region of the eye and translated according to the fovea offset described in

section 1.3.2. The cylinder is rotated to align with the curvature of the eye. Coronal and

axial views of the fovea shift are depicted in Figure 3-4A and 3-4B, respectively. In the

coronal view, the red crosshairs depict where the geometric axis would intersect the

posterior pole of the eye. As shown in this figure, the macula has been moved 1.25 mm

laterally and 0.5 mm inferiorly so that the center of the macula aligns with the fovea.

The center of the optic disc is located 3.3 mm from the geometric axis and 4.6 mm from

the center of the fovea. The axial view highlights the necessity to rotate the macula with

respect to the curvature of the eye. The optic disc is modeled as a cylinder protruding









into the vitreous humor from the end of the optic nerve, consistent with typical ocular

anatomy. The ocular models so described were applied to both the reference male and

female head phantoms.

3.2.3 Optic Nerve Model Detail

Due to the absence of available data, the optic nerve measurement parameters

described in Chapter 2 were used to design the reference optic nerve models. The

mean of the length measurements (parameters M2, M3, and M4) and the ranges of the

optic nerve tilt measurements (Mi and estimation of horizontal tilt) were used to provide

representative optic nerve models. Based on analysis presented in Chapter 2, the

length measurements of the optic nerve were gender-dependent and as such separate

models were created for both the male and female. With the IRayTM system, targeting

error due to patient motion is monitored and maintained below 400 im on the retina.

Therefore, the gaze angle of the patient was not taken into account while building these

models, which is the primary reason the standard deviations of the optic nerve tilt

parameters are larger than would be seen in a real patient population undergoing AMD

radiotherapy (where the gaze angle is fixed with the I-GuideTM). As a result, no

deviation from primary gaze was assumed during treatment simulation of the reference

geometry. However, the use of several optic nerve models, using the full range of optic

nerve tilts observed in Chapter 2, will inherently include variation of gaze angle.

The primary optic nerve was fashioned from the means of the measurements,

while the other four were formulated from the combination of extreme exit angles in the

inferior-superior direction (value of M1) and the medial-lateral direction (estimated from

2D images). For both male and female, the resulting values used in construction of the









optic nerves were -24.4 to +15.5 in the inferior-superior direction and were +18.9 to

+28.5 in the medial-lateral direction. The five optic nerve models are labeled as 'mean',

'sup-med', 'sup-lat', 'inf-med', and 'inf-lat' referring to the combination of exit angles

used for each. The male set of optic nerve models is displayed in Figure 3-5. While

creating these extreme cases for optic nerve exit angle, two control points were

maintained within head models: (1) the junction of the eye and optic nerve, and (2) the

point at the posterior region of the orbit where the optic nerve enters the cranium. With

these two points fixed, another point was needed to measure the exit angle, and so a

plane was placed behind the eye and perpendicular to the geometric axis. The plane

was placed at a distance equal to the mean of Y in Figure 2-2B, which was 14.3 mm for

males and 9.6 mm for females.

3.3 Patient Specific Phantoms

3.3.1 Selection and Development

Utilizing the CT data obtained as described in section 2.2, 16 image sets were

selected for three-dimensional reconstruction. The selection criterion was based on the

initial estimates of vertical gaze angle from measurements taken on the axial CT images

(parameter Mi). Ten patients were found to have a vertical gaze angle within 50 of

being parallel to the Frankfurt plane and additional patients were selected in increments

of 50 when available, resulting in six additional patients.

Complete three-dimensional reconstruction of the 16 patients was accomplished

similarly to the method used to construct the full body reference phantoms. The

following anatomical structures of interest were highlighted within the CT head images:

the lens, globe of the eye, optic nerve (from the posterior of the globe to the optic

foramen), brain, orbital bone, and skin (Figure 3-6). The top and back of the head are









often left out of the field-of-view (FOV) in 1 mm head CT scans, and so only the anterior

portions of the brain and skull near the orbit were segmented. This resulted in a model

consisting of a small band of tissue surround the ocular anatomy. The resulting polygon

mesh files, an example of which is shown in Figure 3-7A, were exported to Rhinoceros

4.0TM to prepare each eye for Monte-Carlo based treatment simulation. The patient

specific phantoms constructed were not converted to NURBS surfaces, and as such

cannot be accurately described as hybrid phantoms. Rather, these phantoms are

classified as voxel phantoms.

3.3.2 Expanded (3D) Angular Measurements

An alternate model with an expanded band of segmentation, an example of which

is shown in Figure 3-7B, was also exported to Rhinoceros 4.0TM to re-evaluate vertical

gaze angle measurements in 3D and measure optic nerve exit tilt in 3D. Analysis in

Chapter 2 provided initial data on vertical optic nerve tilt normalized to the Frankfurt

plane, however estimates of horizontal gaze could not be normalized and right-left head

tilt was not factored into analysis. Thus, a three-dimensional evaluation of optic nerve

tilt and gaze is warranted. The 3D angular positioning measurement parameters are

denoted as: (0) tilt angle of the optic nerve as it leaves the posterior region of the eye

and (0) gaze angle. Gaze direction was defined as the line that intersects the volume

centroids of the eye and lens. A sagittal reference plane was defined by: (1) a point on

the septum in the anterior lobe of the brain, (2) the midpoint of the ear canals, and (3)

being perpendicular to the Frankfurt plane, which was again used for the axial reference

plane. Both parameters were measured relative to the reference planes and partitioned

into vertical and horizontal components. The parameters will be referred to as Ov and

0v for the vertical component and Oh and Oh for the horizontal component.









Scatter plots and linear regression equations are shown for the measured vertical

and horizontal components of the angular measurement parameters E and 0 in Figure

3-8 along with total (angle measured in 3D environment) angular measurements. The

parameters from a linear regression indicate that the reference position of the optic

nerve is tilted 0.80 superiorly and 22.40 medially.

As in Chapter 2, the results suggest a loose correlation between the gaze angle

and optic nerve position. As gaze angle shifts, the optic nerve reacts accordingly by

being "stretched" in the opposite direction. Thus, for an upward vertical gaze, the optic

nerve would have a superiorly tilted exit angle, and for a downward vertical gaze it

would have an inferiorly tilted exit angle. For an inward (medial) horizontal gaze, the

optic nerve would increase the degree of tilt in the medial direction from its reference

position, which is already tilted approximately 22.40 in the medial direction. For an

outward (lateral) horizontal gaze, the optic nerve will reposition itself with a smaller tilt

angle with respect to its reference position in the medial direction. The R2 values are

again well below what would be characterized as a statistically significant correlation,

indicating that not all optic nerves have the same reference position and may react

differently to changes in gaze angle. Additionally, for the optic nerve to be able to react

to changing gaze, there must be some "slack" in the optic nerve in the reference or

primary gaze position. The position in which this slack comes to rest may depend on

the last direction that person gazed, or the person's head orientation with respect to

gravity if sufficient time is allowed for the optic nerve to readjust itself within the orbital

fat. Furthermore, it may seem counterintuitive that the distribution of horizontal gaze

angles is centered over a negative value, favoring a lateral gaze. However, in this









study, gaze angle was defined using the volume centroids of the lens and globe

(geometric axis) which is not coincident with the fovea (visual axis). Since the fovea is

located lateral to the posterior pole, which intersects the geometric axis, defining true

gaze angles using the visual axis would shift the distribution medially. In this scenario,

the true gaze would be dependent on the distance to the object of focus, which has no

relevance in the clinic for treatment planning.

3.3.3 Patient Specific Treatment Planning

After angular measurements were evaluated, each of the 32 eyes (left and right

done separately) was prepared for computational treatment simulation using the

Rhinoceros 4.0TM software. This was accomplished by: (1) locating the center of the

optic disc (approximated from the three-dimensionally reconstructed optic nerve), (2)

determining the position of the posterior pole (3.3 mm lateral to optic disc center), (3)

determining the position of the fovea (1.25 mm lateral and 0.5 mm inferior to the

posterior pole), (4) locating the apex of the cornea from the three-dimensionally

reconstructed globe, (5) aligning the treatment axis to intersect the fovea and to be

parallel with the geometric axis, which is defined as the intersection of the posterior pole

and the apex of cornea, (6) insertion of cylinder with 4 mm diameter and 0.5 mm

thickness coincident with the fovea representing the macula tissue, and (7) tagging

each structure with a tissue name for voxelization.

An important aspect about this method for treatment simulation should be noted

here. The position of each eye model was left as segmented to preserve the anatomy

observed directly from the CT data and was not rotated into a gaze position clinically

realistic of the stereotactic radiosurgery treatment. A range of clinically realistic vertical

gaze angles was determined from the analysis of 25 healthy volunteers whose vertical









gaze was measured with their heads situated in the IRayTM head support device

(unpublished data). The gender distribution was 15 male and 10 female. The vertical

gaze of the individual in the IRayTM system is dependent primarily on anatomical factors,

though there is one mechanical degree of freedom that contributes to the vertical gaze

angle: the chin rest can be moved anterior-posterior by 25 mm. To control for this

flexibility, the vertical gaze angles were determined at the two extremes of chin rest

position. Vertical gaze angle was measured from the Frankfurt plane to be consistent

with the computational measurements. During treatment, the patient's eye is

constrained to gaze 7 below the horizon, and the posts that hold the head rest of the

IRayTM stand normal to patient gaze, at 7 from vertical. Image analysis was used to

determine how a subject's Frankfurt plane sits in the head rest with respect to the post.

With the chin rest set all the way forward, a subject's head was set in the head rest, and

a high resolution image was taken normal to the subject's profile capturing the auditory

canal, the head rest post, and the eye and cheek. A similar image was taken with the

chin rest set all the way back. For the purpose of this measurement, the eye is allowed

to roam freely as the direction of the gaze during treatment is constrained to be

perpendicular to the post. A line representing the Frankfurt plane was drawn from the

top of the auditory canal to the infraorbital rim, based on the folds of the skin, using the

ImageJ software. A second line was drawn intersecting the first line, parallel to the

head rest post. The angle between the Frankfurt line and the post line was measured,

from which the vertical gaze angle could be determined.

Clinically, the horizontal gaze angles are small as the patient head is placed in the

system looking forward, and the eye is held forward as well. If the head were seated at









an angle in the head restraint, the clinician would re-seat the patient for placement of

the eye restraint.

The clinically relevant vertical gaze angles determined from 25 healthy volunteers

ranged from 1.70 inferior to 17.30 superior with respect to the Frankfurt plane. A

reasonable estimate for the range of clinically realistic horizontal gaze angles is within

50 of the primary gaze position, which is defined as a straight-ahead gaze.

3.4 Voxelization

As mentioned in the description of hybrid phantoms, voxelization is required for

use in conjunction with radiation transport. To accomplish this, each organ in the

NURBS models were tagged in Rhinoceros 4.0TM and then exported in raw format.

Using an in-house MATLAB code, Voxelizer 6.0, the models were voxelized to a desired

resolution in a binary file. Using another in-house MATLAB code, the binary files were

converted to lattice file format, which is readable by MCNPX.

When selecting voxel resolution, there is a tradeoff between accurately modeling

small structures (higher voxel resolution) and efficient computer run times during

radiation transport simulation (lower voxel resolution). To account for this tradeoff, three

versions of the UF NURBS reference phantom were selected and voxelized at different

resolutions. The torso of the reference phantom, to be used in the calculation of

effective dose from leakage radiation, was voxelized to 2 mm x 2 mm x 2 mm

resolution. The extracted head model was voxelized to a resolution of 1 mm x 1 mm x 1

mm; however, the limiting size of the macula (0.5 mm in thickness) determines the

resolution necessary for accurate ocular anatomy. Consequently, a finer resolution

ocular model was begot from voxelizing the full NURBS head model to 0.5 mm x 0.5

mm x 0.5 mm resolution. The finer resolution head model was carefully cropped to









include the entire optic nerve and anterior portions of the brain for the treated right eye

using the ImageJ software (NIH, Bethesda, MD). Similarly, all of the patient specific

models were cropped and voxelized to a 0.5 mm x 0.5 mm x 0.5 mm resolution. The

voxelized versions of the reference head and eye models are shown in Figure 3-9 and

3-10, respectively, and a two-dimensional cross section of a patient specific voxel model

is shown in Figure 3-11.










Table 3-1. Comparison of tissue masses in the UF hybrid NURBS and voxel male head
phantoms with those given in ICRP Publication 89 for the reference adult
male
Organ System Density ICRP 89 UFH NURBS UFH Voxel
(g/cm ) mass (g) mass (g) % Diff mass (g) % Diff


Eye Structures
Eyes (2)
Lens (2)
Macula (2)
Optic discs (2)
Optic Nerves (2)
Respiratory System
ET1 (anterior nasal layer)
ET2 (posterior nasal layer)
ET2 (oral cavity layer)
ET2 (larynx)
ET2 (pharynx)
Alimentary System
Tongue
Salivary glands
Parotid
Submaxillary
Sublingual
Tonsils
Esophagus (partial)
Skeletal System
Cranium
Mandible
Vertebrae-C
Intervertebral Discs
Additional Tissues
Brain
Ears
External nose
Pituitary Gland
Spinal Cord (partial)
Thyroid


1.03
1.07
1.03
1.04
1.04

1.03
1.03
1.03
1.07
1.03

1.05
1.03
1.03
1.03
1.03
1.03
1.03

1.38
1.38
1.38
1.10

1.04
1.10
1.05
1.03
1.04
1.05


15
0.45







28


73
85
50
25
10
3


1450.00


0.6

20


14.946
0.450
0.013
0.002
1.055

2.202
13.717
1.370
28.125
3.944

73.158
85.077
50.047
25.028
10.002
3.016
8.814

919.433
81.439
148.638
5.563

1449.641
14.879
12.985
0.602
42.333
19.956


-0.4%
0.1%







0.4%


0.2%
0.1%
0.1%
0.1%
0.0%
0.5%


14.843
0.447
0.015
0.004
1.058

2.062
8.582
1.477
27.917
3.394

72.119
84.636
49.672
24.975
9.989
2.980
8.792

905.192
81.479
147.657
3.478


0.0% 1434.722
14.874
12.892
0.3% 0.600
38.057
-0.2% 19.906


-1.0%
0.7%







-0.3%


-1.2%
-0.4%
-0.7%
-0.1%
-0.1%
-0.7%


-1.1%


0.1%

-0.5%










Table 3-2. Comparison of tissue masses in the UF hybrid NURBS and voxel female
head phantoms with those given in ICRP Publication 89 for the reference
adult female
Organ System Density ICRP 89 UFH NURBS UFH Voxel
(g/cm ) mass (g) mass (g) % Diff mass (g) % Diff


Eye Structures
Eyes (2)
Lens (2)
Macula (2)
Optic discs (2)
Optic Nerves (2)
Respiratory System
ET1 (anterior nasal layer)
ET2 (posterior nasal layer)
ET2 (oral cavity layer)
ET2 (larynx)
ET2 (pharynx)
Alimentary System
Tongue
Salivary glands
Parotid
Submaxillary
Sublingual
Tonsils
Esophagus (partial)
Skeletal System
Cranium
Mandible
Vertebrae-C
Intervertebral Discs
Additional Tissues
Brain
Ears
External nose
Pituitary Gland
Spinal Cord (partial)
Thyroid


1.02
1.07
1.03
1.04
1.04

1.07
1.02
1.02
1.07
1.02

1.05
1.02
1.02
1.02
1.02
1.02
1.03

1.38
1.38
1.38
1.10

1.04
1.10
1.05
1.02
1.04
1.05


15
0.45







19


60
70
41
21
8
3


1300


0.6

17


14.801
0.450
0.013
0.002
0.784

0.642
9.776
1.145
18.999
1.549

59.997
70.061
41.085
20.970
8.006
2.998
8.244

799.001
62.452
110.662
4.675

1302.777
9.200
16.063
0.601
11.454
16.998


-1.3%
0.1%







0.0%


0.0%
0.1%
0.2%
-0.1%
0.1%
-0.1%


14.718
0.442
0.022
0.001
0.770

0.639
9.073
3.486
18.728
1.321

59.765
69.839
40.985
20.867
7.988
2.963
8.261

790.616
62.515
110.561
3.737


0.2% 1287.817
9.178
16.638
0.1% 0.606
11.972
0.0% 16.981


-1.9%
-1.8%







-1.4%


-0.4%
-0.2%
0.0%
-0.6%
-0.2%
-1.2%


-0.9%


1.0%

-0.1%


















































Figure 3-1. The whole body male (left) and female (right) reference phantoms
developed within the Advanced Laboratory for Radiological Dosimetry Studies







62



































Figure 3-2. University of Florida NURBS male (left) and female (right) head models
based on organ masses listed in ICRP Publication 89


_-C
/---~


/

/


Ki
iirr;


NZ


\


Geometric axis j



"


Figure 3-3. Dimensions of the tissues structures in the NURBS eye model; all
dimensions are in mm










Optic Disc


3.30


4.6


Figure 3-4. Engineering drawings of the eye detail embedded within the reference
NURBS head model; (A) coronal view of the posterior region of the right eye
demonstrating the fovea offset with the posterior pole located at the
intersection of the red lines; all dimensional are in mm (B) axial view of the
poster region of the right eye demonstrating the position and rotation of the
macula with the geometric axis defined by the red line extending towards the
anterior portion of the eye


j Opre Di6C I




























Figure 3-5. Male NURBS eye models with five optic nerve variations (red), the macula
targets (green), and the lenses (blue)


Figure 3-6. Segmentation of the lens (blue), globe (orange), optic nerve (red), brain
(purple), and skull (teal) from a 1 mm axial CT image of the orbital region










































Figure 3-7. Patient specific models in object file format generated from three-
dimensional reconstruction of 1 mm CT data (shown without skin); (A) typical
size of model to be submitted for 0.5 mm3 voxelization, and (B) expanded
model that includes both the bottom of the orbit and the ear canals so that the
Frankfurt plane can be defined for evaluation of 3D measurement parameters












VerticCa Optc NIerve Ti verss Vedicair raze 4Ariy


20




0






R2 = .554
-10
W 9 OZ5


2D


10
E 10


-40 -2 0 20 4t -40
(D (degrees)


Horionlal O r: Nerve Tif versus Honzontal Gaze Angle


" *"





* = 0.249*h + 22.4
R = 0.259
-30 -10 0 10 e

4 (degrees)


Toal Opte Nave Til vwrsus Total Gaze Angle
20


25 -
20 -

a



*

S9,, =D.347* +3,502
RI = 0.41t


0,, (degrees)


Figure 3-8. Scatter plots of optic nerve tilt as a function of gaze angle; linear
regressions determine the reference positions of the optic nerve (y-axis
intercept), and the resulting parameters are shown within the plots; (A)
vertical components: values positive in the superior direction and negative in
the inferior direction (B) horizontal components: values positive in the medial
direction and negative in the lateral direction (C) total: all values positive as
they were obtained in 3D























67


S.t_







I


Figure 3-9. Coronal (left)
mm3 resolution

'p


N


and sagittal (right) views of the head model voxelized to 1


I


Figure 3-10. Cropped eye section voxelized to 0.5 mm3 resolution







































Figure 3-11. Axial cross sectional view of a patient spec
resolution


c model voxelized to 0.5 mm3









CHAPTER 4
COMPUTATIONAL METHODS

4.1 The Monte Carlo Radiation Transport Code MCNPX

Monte Carlo methods use random number generation and probability statistics to

solve a variety of physics based mathematical problems. By running a large number of

histories (samples of particle tracks), the stochastic (random) behavior of nuclear

particles is averaged and macroscopic trends can be observed. The physics model

underlying Monte Carlo techniques break down over very small distances (e.g. sub

millimeter for electrons), but provides an excellent method to simulate radiation

transport on a larger scale. This is an invaluable tool, essentially allowing radiation-

based experiments to take place within the safety of computer algorithms. Each particle

is tracked to the end of its life (or until it reaches the problem boundary) and each

individual physical event (scattering, absorption, etc) is determined by a probability

distribution function defined within the nuclear cross section libraries that are stored

within the radiation transport code.

MCNPX (Monte Carlo N-Particle eXtended) is a general purpose Monte Carlo

radiation transport code written in Fortran 90 that tracks a variety of radiation particles

over broad energy ranges.45 MCNPX began as an extension of MCNP4B in 1994 and

was developed by Los Alamos National Laboratory (LANL, Los Alamos, NM). The

extended version provides an improvement on physics simulation models, extension of

neutron, proton, and photon libraries to higher energies, addition of new particle types,

and the formation of new tally techniques. The developers of the code are so confident

in its ability to model radiation physics that they promise a $2 bill for any error found









within the code. Therefore, all simulations of ocular radiotherapy for AMD presented

within this study were performed using the MCNPX version 2.5.0.

The structure of the MCNPX code is divided into three sections and a title card.

The first two sections, the cell and surface cards, describe the problem geometry. Cells

are defined by surfaces using Boolean operators and contain information about the

material density. The last section, the data cards, contains information that defines the

material composition explicitly, the source definition, and the tally specification. A few

other cards are often given in the data card section; some are mandatory (mode and

nps cards) but many are optional and allow the user to modify the default settings, such

as energy cutoff and energy grouping methods.45

4.2 Monte Carlo Techniques Used for Treatment Modeling

An example input deck is given in Appendix A featuring the 'mean' optic nerve

model within the 0.5 mm x 0.5 mm x 0.5 mm resolution male eye section of the

reference phantom.

4.2.1 Cell and Surface Cards

Voxel model geometry was used for all input files in this work, which requires the

assignment of universe numbers to each cell within a lattice based on tissue type. Two

surface cards are necessary for such a geometry definition (besides the outer boundary

of the problem): (1) an rpp box defining the overall dimension of the lattice structure,

and (2) an rpp box defining the dimensions of each voxel. The lattice structure is filled

using data from an imported lattice (.lat) file, and the result is a lattice box filled with

several universes, each with the dimensions of a single voxel. The cell cards define the

properties of each universe: material type, density, volume, and importance. This









method is one of the most powerful ways to import complex geometries into an MCNP

input deck. A visual of voxel geometry plotted by MCNPX is given in Figure 4-1.

4.2.2 Source Definition

The relevant x-ray emission spectrum was generated for a tungsten anode tube

operated at 100 kVp with anode angle of 12 degrees and total filtration of 0.75 mm Al

and 0.8 mm Be using the computer software described in Report No. 78 of the Institute

of Physics and Engineering in Medicine.46 Using the simulated x-ray energy spectrum,

a divergent x-ray beam was modeled as a 1 mm x 1 mm area source representing a 1

mm2 focal spot. The divergence was modeled to simulate a beam collimated by an

explicit tungsten aperture with a beam diameter of 4 mm at the macula target over a

source-to-target distance of 150 mm. The nominal polar angle of 30 from the treatment

axis was accomplished using a transformation card, as were the three different

treatment azimuthal angles: 1500, 1800, and 2100.

Leakage calculation requires a different source definition. The leakage source

was defined at a point coincident with the location of the anode. The energy spectrum

of leakage radiation is generally characterized as hard, that is, higher energy photons

have a greater chance to transverse the tube housing unattenuated and the remaining

photons are harder to attenuate. The clinical energy spectrum of the leakage radiation

is unknown for this medical device; therefore an approximation was made using an 80

and 100 keV mono-energetic source of photons. The dosimetry results from both

energies were similar and so the 100 keV source was used for final reporting.

4.2.3 Tally Specification

In all, there are eight different tally specifications and four mesh tally options in

MCNPX. Tally cards can be modified by a number of keywords and other cards,









providing a versatile method for extracting useful data from the problem geometry.

Three ways to tabulate dose were utilized for this project: (1) F4 tally for cell fluence

modified by a dose response function with DE/DF cards, (2) F6 tally for energy

deposition within a cell, and (3) type 1 mesh tallies using the keyword 'pedep'.

The *F6 tally was used to calculate all tissue-specific mean absorbed doses in this

study. The *F6 tally in MCNPX reports jerks per gram per photon history. A jerk is an

MCNP unit of energy equivalent to 109 Joules, thus the reported value for each organ

tally was multiplied by a coefficient of 1012 to convert to Gy per photon history using the

FM card. Absorbed doses to the brain, thyroid, salivary glands, bone marrow, and bone

surfaces were calculated using the 1 x 1 x 1 mm3 voxel head model, while doses to the

macula, lens, optic disc, and optic nerve were calculated with the 0.5 x 0.5 x 0.5 mm3

voxel eye model. The 2 mm x 2 mm x 2 mm voxel torso model was used to calculate

the leakage contribution for several other radiosensitive organs.

The F4 tally was set up for calculation of dose to bone marrow and bone surfaces

for the cranium, mandible, and cervical vertebrae. The tally was modified by dose

response functions described in ORNL/TM-8381/V1 Table D-5.42 The MCNPX output

for this tally type is in units of photons/cm2 and the dose response functions convert

photon/m2 to Gy. Therefore, each bone tally was multiplied by a coefficient of 104 to

convert to Gy per photon history.

The type 1 mesh tally with the keyword 'pedep' was coded to exactly overlap the

lattice structure of the voxel geometry. MCNPX creates a binary output file in the form

of a 3D matrix containing the dose to each voxel. Using the built in GRIDCONV









function in MCNPX, this output file was converted to an ASCII file. This output is

necessary to beget dose volume histograms and dose contour maps.

Two other tally types were used in this study to tally photon fluence: (1) an F1 tally

modified by an eO and cO card to evaluate energy and angular dependence of fluence,

and (2) a type 1 mesh tally without any keyword (default is for fluence). An evaluation

of the distribution of photons exiting the head during treatment was necessary in

determining safety parameters and shielding design for clinical staff present during

treatment. An F1 tally sphere, with its origin at the macula target and normal vector

aligned with the treatment axis, was used to evaluate the energy and angular

dependence of photon fluence at a radius of 0.5 meters. Two-dimensional matrices

were implemented using type 1 mesh tallies, flush with the edge of the lattice structure,

which were used to characterize the spatial distribution of photons emanating from each

side of the model.

4.2.4 Material, Mode, and NPS Cards

All of the material cards defined in this project were derived from ICRU 46 material

composition and density data.

The mode 'p' was used for all input decks for this study. This mode approximates

that all secondary particles from interactions (electrons) deposit their energy locally at

the site in which it was born (KERMA approximation). This increases computing

efficiency by not creating and tracking the secondary electrons. It is a good

approximation considering the short track length of secondary electrons in the

kilovoltage energy group and was validated by Lee et al.28

A total of 107 x-ray photon histories were completed for each simulation, and the

resulting statistical errors for tissue-averaged dose tallies were found to be less than 2%









for the optic disc, less than 1% for all other tissues, and ranged from 0.6% to 2% for

each macula voxel.

4.3 Post Processing

4.3.1 Calculation of Effective Dose

Contribution from primary tube output. Effective dose was determined for a 3 x

8 Gy treatment using steps described in International Commission on Radiological

Protection (ICRP) Publication 10347 and the following expression:


E= =WTwRDTR =w, HT +HT] (4-1)
T R T 2

Tissue mean absorbed doses (DT,R) are included for the brain, thyroid, salivary glands,

active bone marrow, and bone surfaces by scaling the output from section 4.2.3 by the

number of histories necessary to deliver an absorbed dose of 8 Gy to the macula for

each treatment angle. A DT,R of 0 was assumed for all radiosensitive organs below the

neck. Radiation weighting factors (wR) can be found in Table B.4 of ICRP Publication

103 and are 1.0 for photons. The equivalent dose for tissues in the reference adult

male (HM) and female (H ) for each beam angle were summed to give a cumulative

equivalent dose for a 3 x 8 Gy treatment and sex averaged. Tissue weighting factors

(wT) are found in Table B.2. of ICRP Publication 103 and are given in Table 4-1. The

notation in this section is consistent with that in ICRP Publication 103.

Several other weighting factors were implemented for the calculation of effective

dose. The parotid, submaxillary, and sublingual portions of the salivary glands were

volume weighted to yield a single tissue dose to the target. The bone surfaces

(endosteum) were weighted by fraction of bone surface across the entire skeleton. This

tissue weighting was as follows: cranium 15.3%, mandible 0.4%, and cervical vertebrae









2.1% for male and cranium 15.8%, mandible 0.4%, and cervical vertebrae 2.2% for

female.48 The bone marrow dose was weighted by values given in Table 9.4 in ICRP

Publication 89 which reports the percentage of active marrow in each bone relative to

total body active marrow as a function of age. The maximum age of 40 year was used

here since the majority of patients undergoing AMD treatment are over the age of 50.

Contribution from leakage. The determination of effective dose from the

leakage contribution is a more complex process requiring the energy spectrum of

leakage radiation and the dose rate at some distance from the anode. The clinical

leakage energy spectrum for the IRayTM is unknown; therefore an approximation was

made using a mono-energetic source 100 keV photons. A dose rate reading in air near

the heart was taken in the clinic, found to be 12 mR/hr, and used as a conversion factor

for MCNPX output. To apply the conversion factor, modified input files were generated

filling each universe with air, thus modeling and irradiating an air phantom. The output

from the dose tally for the cells tagged as heart was converted from jerk per gram per

history to rad per history by applying a coefficient of 1014. The leakage rate was

multiplied by the treatment time per beam (~2 minutes), and converted from mR to rad

giving the total rad per beam in the clinic. The rad per beam was divided by the rad per

history from the air phantoms to give a conversion factor based on the total number of

source particles necessary to give a dose rate of 12 mR/hr near the heart. The MCNPX

inputs were re-run with a tissue filled phantom and the conversion factor was applied to

the output for dose of each tissue. The remainder of the effective dose calculation was

performed in the same manner as for the primary tube output, but with all organs in the

torso considered, and again consistent with ICRP 103.









4.3.2 Utilizing Mesh Tally Output

As mentioned, a mesh tally was coded exactly overlapping the voxel geometry

allowing for a unique manipulation of mesh tally output. An in-house MATLAB code

was written to link the dose to each voxel from the mesh tally to the original tissue types

in the voxel model. This allows the presentation of data in several useful formats

including: DVH plots, dose distribution tables, and dose contour maps. The code used

is given in Appendix B. DVH plots are created within the MATLAB program and the

dose distribution tables are exported to Excel. Further manipulation is necessary to

actualize dose contour maps using Microsoft Excel, Sigmaplot 10.0TM, and a photo

editing software.












Table 4-1. Tissue weighting factors for the calculation of effective dose as given by
ICRP 103
Tissue Number of WT Total
tissues Contribution
Lung, stomach, colon, bone marrow, breast, 6 0.12 0.72
remainder


Gonads 1 0.08 0.08


Thyroid, oesophagus, bladder, liver 4 0.04 0.16


Bone surfaces, skin, brain, salivary glands 4 0.01 0.04





J,,PO 'WINDOWi -IEI


07/05/10 14:45:48
c Reference Adult male eye
voxel model

probid = 07/05/10 14:45:12
basis: YZ
( 0.000000, 1.000000, 0.000000)
( 0.000000, 0.000000, 1.000000)
origin:
( 0.00, 0.27, -0.17)
extent = ( 4.00, 4.00)


UP RT DN LF


Edit cel 1
Cell 1
xyz = 0.00, 0.27, -0.17
CURSOR SCALES 0 CellLine
PostScript ROTATE
COLOR mat LEVEL
XY YZ ZX
LABEL sur off
MBODY on


Origin .1 .2 Zoom 5. 10

cel
imp
rho
den
vol
fcl
mas
put
mat
tmp


ext
pd
dxc
u
lat
fill
ijk
nont
pac
tal


Click here or picture or menu
Redrau Plot> End


Figure 4-1. MCNPX plot of the male eye section model voxelized to 0.5 mm3 resolution









CHAPTER 5
TREATMENT OUTCOME EVALUATION AND ANALYSIS

5.1 Radiation Dose Thresholds for Complications

The biological response of the human eye to ionizing radiation is well documented.

A review of the major findings may be found in Section 5 of Report No. 130 of the

National Council on Radiation Protection and Measurement (NCRP). Much of the

experience on the radiation response of the eye derives from studies with fractionated

and chronic regimens of low-LET radiation. The radiation absorbed doses that produce

minimally detectable changes or functional disabilities are 6, 5, 30, 15, 16, and 25 Gy,

for the lid, conjunctiva, cornea, sclera, iris, and retina, respectively. The corresponding

visually debilitating absorbed doses to these same ocular structures are 40, 35, 30, 200,

16, and 25 Gy, respectively.30

Until recently, the threshold for minimally detected changes or functional disability

to the lens was reported to be 2000 mGy and the threshold for visual debilitation was

reported as 5500 mGy.30 However, more recent studies suggest that the lens is much

more sensitive to ionizing radiation than previously thought. A recent report suggests

that the tissue-averaged dose threshold for radiation cataractogenesis could be as low

as 700 mGy.49 This limit will be used in this study in the analysis of treatment outcomes

to err on the side of greater radiological protection.

The radiological sensitivity of the optic nerve has been studied in patients whose

optic nerve was unavoidably or unintentionally irradiated as a consequence of brain or

head tumor radiotherapy showing that doses 8 Gy or higher might have some adverse

consequence.50 Another study found that doses less than 12 Gy to a short segment of

the anterior optic apparatus during stereotactic radiosurgery resulted in a low risk









(~1.1%) for radiation-induced optic neuropathy (RON); however, 3 of 4 patients in this

study that developed RON had previously received external beam radiation therapy

(EBRT) and the other had undergone two previous radiosurgery procedures.51 It is also

unclear what percentage of volume characterized the short segment of the anterior optic

apparatus. Ultimately, the authors conclude that point doses up to 12 Gy are well

tolerated by patients whose optic nerve has not been previously irradiated.

Furthermore, a recent study suggest that the optic apparatus may be more tolerant to

radiation than previously thought, able to receive up to 14 Gy without risk of developing

RON (again under the assumption that the patient has not previously undergone

radiation therapy).52

Other tissues of interest include the brain and orbital bone. The development of

necrosis in brain tissue due to radiological toxicity is summarized by Lawrence et al, and

it has been determined that the threshold for neurological toxicity is 12 Gy for a volume

of at least 5 10 cm3.53 Bone, and in this case orbital bone, contains elements with

higher atomic numbers that have a higher cross section (probability) for photon

interaction, namely the photoelectric effect, than soft tissue and fat. The resulting

secondary particles (electrons) are likely to deposit their energy locally and as such the

bone absorbs more dose than surrounding tissues. However, the skull is fairly radio-

resistant to adverse consequences and the orbital bone contains a negligible

percentage of the total active marrow in the cranium.29

5.2 Reference Model Dose Assessment

5.2.1 Tissue-specific Mean Absorbed Doses

Mean absorbed doses to several non-targeted tissues are shown in Table 5-1 for

a 3 x 8 Gy treatment to the right eye. The MCNPX output has been normalized by the









number of photons necessary to deliver a dose of 8 Gy to the macula target for each

beam. The lens received a 3-beam integral mean dose of 124 mGy and 127 mGy in the

reference male and female patient, respectively, well below the threshold for

cataractogenesis. The optic disc receives a mean dose ranging from 200 to 239 mGy,

a factor 33 to 40 times less than that for the macula target, despite its proximity to the

macula target (4.6 mm). The average absorbed dose to the optic nerve, a critical part of

this project, was found to increase with increasing beam azimuthal angle for the treated

right eye. The optic nerve used for calculation in this section was that in the 'mean'

position. The highest optic nerve mean dose recorded was 112 mGy in the reference

female patient at a 2100 beam entry angle. Other non-targeted tissue doses were also

found to be insignificant. The highest dose to the optic nerve opposite treatment was

2.4 mGy for the female at a 2100 beam entry. The mean absorbed doses to the brain

ranged from 3.0 to 4.9 mGy per treatment beam. The mean absorbed doses to the

thyroid and salivary glands were 4 to 5 orders of magnitude less, respectively, than that

to the target for all beam angles and both sexes. Considering the mean absorbed

doses to the bone structures in the head, the cranium received the highest dose to both

the active marrow and endosteal tissues, but these doses were on the order of 9 mGy

or less.

5.2.2 DVH Analysis

Dose volume histograms for the male and female mean optic nerve model are

shown in Figures 5-1 for a cumulative 3 x 8 Gy treatment. These histograms display

how the absorbed dose is distributed throughout the structure of the macula, lens, brain,

and optic nerve. Ideally the macula histogram would be a step function, but the

absorbed dose the macula is not quite evenly distributed, and the plots are thus









designed to show a treatment with a maximum dose of 24 Gy to the target. For both

sexes, a steep drop is observed for the lens, indicating that there are no hotspots. The

brain DVH also drops off steeply, but small hotspots in the anterior portions of the brain

(<10%) contribute doses ranging up to about 2.5 Gy. Nevertheless, for females only

about 5% of the brain volume receives a dose exceeding 2.5 Gy and 95% of the male

brain receives a dose less than 2 Gy.

The DVH of the 'mean' male optic nerve shows that 2-3% of that structure

receives a dose exceeding 2 Gy, and that approximately 1% receives a dose exceeding

5 Gy. Less than 1% of the optic nerve volume receives a dose over 10 Gy with a

maximum of about 12 Gy. The DVH of the 'mean' female optic nerve shows that 2-3%

of the tissue volume receives more than 1 Gy with less than 1% receiving a maximum

dose of 8 Gy. The difference in DVH's for the male and female optic nerve indicate that

the maximum dose is a function of optic nerve diameter at the posterior of the eye

(value of M2) since the mean of this measurement is 0.7 mm less for females. This

correlation will be explored more subsequently.

Due to the lack of literature on quantifying the anatomical location of the optic

nerve, and the somewhat large standard deviation of parameter M1 in this study, DVHs

were created for the combination optic nerve exit angle in the superior-inferior and

medial-lateral directions and are shown in Figure 5-2 for males and in Figure 5-3 for

females. These DVH plots are similar to the case for the 'mean' optic nerve DVH,

except for the superior-lateral combination (upper right graphs in these figures). An

optic nerve oriented in such a position is likely to come close to the beams exiting the

eye, and patients in this case may receive localized optic nerve doses as high as 17.5









Gy for males and 15 Gy for females to less than 1% of its volume. Nevertheless, only

2-3% of the male and ~1% of the female 'sup-lat' optic nerve model receives a dose

exceeding 8 Gy. This extreme situation for the reference male patient is shown in the

form of a dose contour map in Figure 5-4. As can be seen, only the outermost

periphery of the optic nerve is irradiated in this worst-case optic nerve position.

Anatomically, this region of the optic nerve would constitute the insulating myelin sheath

and would most likely not be of radiobiological importance to nerve function. Thus, the

risk of radiation optic neuropathy (RON) is hypothesized to be exceedingly small.

5.2.3 Effective Dose

The calculation of effective dose combined both contributions from primary tube

output and leakage estimation. For the primary output, the relevant absorbed tissue

doses necessary to calculate effective dose are listed in Table 5-2 for the reference

male and female head models. The calculation, based on ICRP Publication 103

recommendations, yields an effective dose of 0.281 mSv for a 3 x 8 Gy treatment to the

macula per eye. For leakage contribution, all radiosensitive organs in the head and

torso were considered, and the result of the ICRP 103 based calculation was 0.009

mSv. Thus, the total effective dose for a 3 x 8 Gy treatment is 0.29 mSv.

While the stereotactic AMD radiosurgery is therapeutic in nature, this value of

effective dose compares very favorably with radiographic imaging doses including skull

radiographs (0.1 mSv) and cervical spine radiographs (0.2 mSv), yet are much lower

than seen in CT scans of the head (2 mSv) or neck (3 mSv).54 In contrast, estimates of

effective dose in megavoltage radiotherapy are significantly higher as in treatments of

head and neck tumors (1870 mSv), brain metastases (270 mSv), and brain primary

tumors (580 mSv).55 Clinical trials are still underway and a smaller therapeutic dose









may eventually be prescribed. For example, if a 16 Gy total dose to the macula is

shown to yield good clinical outcome, the estimated effective dose would be

proportionally lower (e.g., 0.19 mSv).

5.3 Patient Specific Dose Assessment

Dose volume histograms are the most common form of data used to evaluate non-

target complication probabilities (NTCP). An alternative to DVH output was contrived

for reporting doses to non-targeted tissues in the patient specific population. The

alternative approach allows for a more condensed data format with specific quantitative

values listed in the form of dose distribution tables. The column headings in these

tables are equivalent to the x-axis of a DVH and the values in the tables correspond to

the shape of a DVH. The total voxelized volume of the optic nerve, lens, and macula

and the percent volume of that tissue over several dose regions are shown in Tables 5-

3, 5-4, and 5-5, respectively. The dose regions in the table were chosen in an effort to

most clearly depict the distribution of dose within each tissue volume. A conservative

approach was taken in reporting volumes to the optic nerve and lens; the reported

volumes include all voxels wherein the mean dose plus the computational uncertainty

surpasses the given table heading. The reverse is true for the macula table, that is, the

mean dose minus the computational uncertainty surpasses that given table heading.

The details of the MATLAB code used for this calculation are shown in Appendix B.

The abbreviation for each eye model is as follows: the first letter denotes gender (m/f),

the second letter denotes subject (A,B,C etc), and the third letter denotes left or right

eye (l/r). For example, female subject D's left eye will be denoted as patient model fdl.

Table 5-3 is organized in three parts according to clinical realism. A patient model

was selected from each of the groups for the formulation of dose contour maps which









are shown in Figures 5-5, 5-6, and 5-7. Table 3 presents the highest (of the set of 32

eye models) tissue-averaged dose to the lens and optic nerve, along with the

associated eye model. For all models, no brain voxel received a dose over 12 Gy, the

threshold for necrosis. Patient model fdl received the highest orbital bone dose with 5

voxels (1.125 mm3) receiving between 45 and 50 Gy.

Despite the variability of the location of the optic nerve observed in this study, the

highest cumulative tissue-averaged dose received was 1.3 Gy by model fkl (Table 5-6).

Dose contour maps were fabricated for this patient (Figure 5-6) and it can be seen that

the overlapping beams avoid the optic nerve. This patient demonstrated a lateral

horizontal gaze of roughly 160, outside the range of clinically relevant horizontal gaze

angles. It is somewhat unclear from the literature what maximum point dose is tolerated

by the optic nerve; nevertheless, it is reasonable to assume that the risk of developing

RON is negligible for all simulated patient models in this study given the dose volume

data presented in Table 5-3.

The highest tissue-averaged dose observed in the present study was 176 mGy by

eye model fol (Table 5-6). Eye model fjl had the highest percentage of volume receiving

doses over 300 mGy and no lens volume received a dose exceeding 400 mGy (Table 5-

4). The sagittal dose contour map and the most inferior voxelized slice that contains the

lens of patient fjl are shown in Figure 5-7 and both clearly depict that the converging

beams do not directly intersect the lens.

An attempt was made to find correlations between absorbed dose to non-targeted

tissues and the measurement parameters mentioned before. For most tissues, there

were no statistically significant relationships. Analysis of optic nerve data provided two









significant correlations. These are presented in Figures 5-8 and 5-9. The latter

relationship is intuitive; the optic nerve dose will escalate with increasing optic nerve

thickness. The former may not be intuitive at first, but becomes clearer with a better

understanding of optic nerve tilt as a function of gaze angles. The logarithmic

regressions suggest that optic nerve dose increases as vertical and lateral gazes

increase. As described in Chapters 2 and 3, an increasing vertical gaze will stretch the

optic nerve into a more vertical exit tilt, which would position the optic nerve closer to

the beams exiting the eye. As lateral gaze increases, the optic nerve tilt decreases

(approaching a limit of being positioned in parallel with the sagittal reference plane

described in section 3.3.2) and positions itself in closer proximity to the beam entering

from the lateral side (exiting medially from the eye ball). This would be the 5 o'clock

beam for the left eye and the 7 o'clock beam for the right eye.

5.4 Photon Fluence Evaluation

Figures 5-10, 5-11, and 5-12 show that photon fluence was mostly forward

directed with respect to the patient's gaze backscatteredd in reference to beam entry)

with maxima in the 50-600 and 44.4-50 keV bins. A portion of the beam traverses

through the head unattenuated, resulting in a smaller peak around 150. Color coded

contour fluence maps are presented in Figure 5-13 to supply a visual representation of

fluence. The results of this study provide a better understanding of the radiation physics

involved during treatment administration and contributed towards the design of shielding

for the treatment device.









Table 5-1. Mean absorbed dose (mGy) to various tissues in the reference head models
for a 3 x 8 Gy Oraya Treatment to the right eye; computational error was less
than 2% for the optic disc and less than 1% for all other tissues
Male Female

Beam Azimuthal Angle: 1500 1800 2100 1500 1800 2100

Right Macula 8000 8000 8000 8000 8000 8000
Right Lens 46.7 40.6 36.5 47.3 42.7 37.0
Right Optic Disc 211 200 201 239 210 212
Right Optic Nerve 48.2 56.8 93.8 56.4 68.5 112.0
Left Optic Nerve 0.92 1.11 1.40 1.84 2.00 2.44
Brain 2.96 3.72 4.25 3.82 4.86 4.94
Thyroid 0.05 0.05 0.06 0.09 0.09 0.09
Parotid 0.31 0.28 0.30 0.53 0.51 0.52
Submaxillary 0.39 0.38 0.42 0.34 0.36 0.38
Sublingual 0.23 0.24 0.26 0.59 0.58 0.57
Cranium (AM) 3.64 3.25 2.86 3.95 3.54 3.28
Mandible (AM) 0.29 0.29 0.30 0.48 0.47 0.47
Vertebrae (AM) 0.14 0.15 0.19 0.27 0.31 0.37
Cranium (BS) 8.86 7.94 7.04 9.67 8.71 8.10
Mandible (BS) 0.94 0.92 0.97 1.56 1.53 1.51
Vertebrae (BS) 0.46 0.49 0.60 0.87 0.99 1.18

Table 5-2. Mean whole-body absorbed doses DT (mGy) for the estimate of the effective
dose
Male Female
Beam Azimuthal Angle: 1500 1800 2100 1500 1800 2100
Brain 2.96 3.72 4.25 3.82 4.86 4.94
Thyroid 0.05 0.05 0.06 0.09 0.09 0.09
Salivary Glands 0.32 0.31 0.33 0.48 0.47 0.48
Bone Marrow 0.29 0.26 0.23 0.31 0.28 0.27
Bone Surfaces 1.37 1.23 1.09 1.55 1.40 1.31











Table 5-3. The gaze angles, voxelized optic nerve volume, and percentage of that
volume receiving more than the absorbed dose listed, representing the dose
distribution, for a cumulative 24 Gy treatment dose to the macula
Model Vertical Gaze Angle Horizontal Gaze Angle Optic Nerve Voxel Volume 0.5 Gy 1 Gy 5 Gy 12 Gy
degrees degrees mm3 % % % %
Both vertical and horizontal gaze are clinically realistic
fkr 9.0 -3.2 556.8 8.7 3.2 0.9 0
fnl 9.5 2.2 664 7.4 2.1 0.6 0.2
mel -0.8 -1.0 632.5 9.2 3.5 1.3 0.2
mer -1.0 -0.3 621.9 6.8 2.4 0.7 0
mkl 5.2 -2.9 599.4 10.7 3.7 1.5 0.4
mkr 8.7 1.2 702.8 9.9 4.1 1.9 0.6
mil 7.2 -4.4 523.8 11.3 4.0 0.8 0
Either vertical or horizontal gaze is clinically realistic
mlr 4.5 -6.0 640.3 16.4 8.8 4.1 0.9
fkl 8.7 -16.4 521.6 26.9 15.8 6.6 0.3
mml -1.6 -16.7 590.6 18.1 9.0 3.5 0.1
fnr 12.6 -18.1 704.9 36.7 27.7 4.8 0.3
fsr -3.6 3.3 551.3 10.6 4.0 1.6 0.4
mar -6.0 -1.5 566.8 9.0 2.8 1.1 0.2
ffl -10.9 -3.3 287.8 8.2 0.9 0 0
for -17.7 -0.5 460.5 6.1 1.6 0.4 0
fol -20.2 -2.6 547.1 3.4 0.7 0.1 0
Neither vertical or horizontal gaze is clinically realistic
ftr -2.4 -6.0 410.3 11.8 4.7 1.3 0.1
mgr -5.3 -19.0 909.1 11.8 7.1 3.8 0.9
mal -5.4 -5.9 431 11.4 4.3 1.8 0.3
mgl -6.4 8.6 686.1 8.6 3.6 1.5 0.4
mor 22.1 -14.5 445 14.3 7.4 2.0 0.3
fsl -6.8 -6.1 482.4 12.8 4.8 1.8 0.3
mfr -7.4 -18.9 621.9 11.4 5.6 2.9 0.8
ffr -7.5 -8.5 356.5 9.2 2.3 0.6 0.1
mmr -7.9 -17.5 640.1 17.8 11.4 4.9 0.2
ftl -8.5 -8.1 588.8 10.2 3.4 0.7 0.1
mfl -8.6 -5.4 776.4 10.0 4.5 2.1 0.6
mol 30.1 -29.8 480 31.4 17.5 6.0 0.2
fdr -34.4 -8.7 207.9 7.9 1.4 0.2 0
fdl -34.6 7.2 259.1 4.5 0.2 0 0
fjr -41.4 19.8 395.0 5.1 0.9 0.3 0.1
fjl -46.5 -28.0 728.6 6.7 2.2 1.1 0.4










Table 5-4. The voxelized lens volume and percentage of that volume receiving more
than the absorbed doses listed, representing the dose distribution, for a
cumulative 24 Gy treatment dose to the macula

Model Lens Voxel Volume 100 mGy 200 mGy 300 mGy 400 mGy

mm % % % %


fdl
fdr
ffl
ffr
fj
fjr
fkl
fkr
fnl
fnr
fol
for
fsl
fsr
ftl
ftr
mal
mar
mel
mer
mfl
mfr
mgl
mgr
mkl
mkr
mll
mlr
mml
mmr
mol
mor


133.6
128.3
100.4
109
135.9
115.8
74.6
94.1
123.6
114
101
94.8
142.8
146
104.5
122.9
138.9
144.3
97
97.9
118
92.9
159.4
148.1
111
102.4
139
133.9
134.4
131.1
119.8
113.4


86.0
90.7
87.5
87.6
90.2
80.6
96.1
99.3
82.5
91.6
99.9
98.5
94.0
95.3
84.4
81.3
78.2
97.6
97.0
99.5
81.9
90.8
83.5
84.5
99.0
91.0
89.0
92.9
93.1
83.9
89.0
96.3


5.7
8.2
7.0
7.5
26.1
11.9
8.2
18.9
2.0
4.9
30.3
18.2
7.2
6.8
2.9
4.4
4.8
27.5
12.1
15.8
0
0
7.8
6.8
12.0
1.3
8.4
7.5
8.1
4.9
10.3
13.3










Table 5-5. The voxelized macula volume and percentage of that volume receiving more
than the absorbed doses listed, representing the dose distribution, for a
cumulative 24 Gy treatment dose to the macula

Model Macula Voxel Volume 5 Gy 10 Gy 15 Gy 20 Gy 25 Gy

mm3%%%%
mm % % % % %


fdl
fdr
ffl
ffr
fj
fjr
fkl
fkr
fnl
fnr
fol
for
fsl
fsr
ftl
ftr
mal
mar
mel
mer
mfl
mfr
mgl
mgr
mkl
mkr
mll
mlr
mml
mmr
mol
mor


100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100


100
95.9
100
100
96.1
100
100
95.9
96.1
96.0
95.8
100
100
96.1
100
95.9
100
100
100
96.1
100
98.0
96.1
96.1
95.9
96.1
96.1
100
95.8
96.0
100
98.0











Table 5-6. The highest tissue-averaged doses received from the set of 32 eyes
undergoing treatment simulation and the associated eye model;
computational error was less than 1% for all values
Tissue Beam I Beam II Beam III Total

mGy mGy mGy mGy

macula 8000 8000 8000 24000

lens 66 (mar) 57 (fol) 69 (fol) 176 (fol)

optic nerve 747 (fkl) 276 (mol) 1100 (fnr) 1291 (fkl)


Opic nerve ilt: mean
f ,. .


j |I




ir II


Absorbed dose (Gy)


Optic rnr lilt mean
100.





I"m


.I.


II

ii- i


?1 5


Absorbed dose (Gy)


I .Ii


Figure 5-1. DVHs for the 'mean' optic nerves; (A) male (B) female













Optic nerve tit: sup-med


Absorbed dose (Gy)


60
CESO




30


Sto
















40
Tl1


Aburbd dose -Gy)
AtobsKd dose -Gy


sord dose (
Absorbed dose (Gy)


Optic nerve ilt: irnflat


R- 1 Iplic nemra
Prl mrculI
70









ID



Absorbed dose i:GyP


Figure 5-2. DVHs for the extremes of male optic nerve tilt; (A) sup-med (B) sup-lat (C)
inf-med (D) inf-lat


Optic neive bilt hifmed~



C"-- ;i l


Optic nere tilt: sup-lat


-r-------------



































Absorbed dose (Gy)


OpUc nerve ti: hf-med


-I I I,
- 1 r ,,


I. ,.
-,' .- -
Absorbed dose (Gy)


90






Esa


Optc nrrve rt sup-at

[l-*h4 rrh









I1r rIl









i

I .'
i


Absorbed dose (Giy


Optic nerve bft inf-lat



---Righ lptr: iw

'0


50



0 1
I-I


20 25


0 S 10 15
Absorbed dose (Gy)


Figure 5-3. DVHs for the extremes of female optic nerve tilt; (A) sup-med (B)
inf-med (D) inf-lat


B


sup-lat (C)






















0
-o
-5
10
S15
20
525


Figure 5-4. Spatial contour map of the dose distribution within the reference eye model
of the adult male; legend units are Gy where 0 refers to some value
approaching the limit of 0














































Figure 5-5. Dose contour maps for patient model mer; image progression is from
inferior (top left) to superior (bottom right) in 1 mm intervals, with the (center)
median slice intersecting the middle of the macula target; legend units are Gy
where 0 refers to some value approaching the limit of 0









95



























Patent: fd


Ea

40





Absorbed dose (Gy)




10


S25 C


Figure 5-6. Dose contour maps for patient model fkl; (A) sagittal image slice
intersecting the middle of the macula target, (B) DVH, and (C) axial image
slice intersecting the middle of the macula target; legend units are Gy where 0
refers to some value approaching the limit of 0
































?






Absoahbed dose (OG) B
20
25
S30 C


Figure 5-7. Dose contour maps for patient model fjl; (A) sagittal image slice intersecting
the middle of the macula target, (B) DVH, and (C) axial image slice
intersecting the middle of the macula target; legend units are Gy where 0
refers to some value approaching the limit of 0


Patient: 1I










Horizontal Gaze vs ON mean absorbed dose


0 200 400 600 800 1000 1200
Mean absorbed ON dose (mGy)


Vertical Gaze vs ON mean absorbed dose


0 200 400 600


800 1000 1200


Mean absorbed ON dose (mGy)

Figure 5-8. Correlation scatter plots of mean absorbed dose to the optic nerve as a
function of gaze angle; (A) horizontal gaze (B) vertical gaze


y = 115.3 18.96 In(x+208.4)
R2 = 0.4171


* *


A
1400


* **


y = -74.0 + 12.63 In(x-126.6)
R2 = 0.4844


R


1400















*

-6 -




O





y = 1.524 x + 4.461
R2 = 0.2342
0 -





0.0 02 0.4 06 08 1 0
% volume of ON absorbing more than 12 Gy


Figure 5-9. Correlation scatter plot and linear regression of optic nerve hotspot dose as
a function of optic nerve thickness


Angular and energy distribution of fluence for photons


100 0
i00
5 Oe-7
10 e-6
1 5e-6
2.0e-6
I 25e-6






50 0













00
-1.0 0.0 1.0
cosine


Figure 5-10. Phase space diagram of energy and angular dependence of photon
fluence 50 cm from the macula target; cosine equal to 1 is synonymous with
primary gaze; all units are #/cm2/history












Angular distribution of photon fluence


1 6e-5
**
1.4e-5 -

1.2e-5 -

U1 e-5 -

( 8 Oe-6 -
C)

o 6.0e-6

E
S4 Oe-6

2 0e-6 -

0.0
A

-1.0 0.0 1.0

cosine



Energy distribution of photon fluence


3e-5


3e-5 -
.

2e-5 -







0*



o *

B

0.0 50.0 1000

Energy (keV)



Figure 5-11. Photon fluence distribution plots 50 cm from macula target; (A) angular
dependence and (B) energy dependence




100




































w-
:....



i:_ ..


1 e-9
l1 e-8
-1 e-7
1 e-6
1 e-3


Figure 5-12. Photon fluence contour maps at the edge of the lattice structure; (A) top
(B) bottom (C) perspective (D) right side (E) front (F) back and (G) left side;
all units are #/cm2/history


101


1A




















e1e


El


Figure 5-12. Continued


102









CHAPTER 6
CONCLUSIONS

6.1 Limitations of This Work

The retrospective collection of CT data presented several challenges for this

project. Retrospective collection was warranted to limit dose to potential volunteers.

MR imaging could have been used but it would have been costly to setup a proactive

study. Because the data are collected retrospectively, and not specifically for the

purpose of this project, patients display a wide variety of gaze angles and head tilt that

would not typically be seen during stereotactic radiosurgery for wet AMD. Measures

were taken to account for this, but ideally patient specific model fabrication and

treatment planning would have been performed on a patient population potentially

undergoing SRS.

It is difficult to obtain a large number of head CT images with fine slice resolution

and facial structures intact. High resolution slices are desired for this project to

minimize the uncertainty in the measurements taken as described in Chapter 2, but

slice resolution is limited during administration because of dose considerations to the

patient. When higher slice resolution is needed for the diagnostic procedure it is

typically because there is some head trauma to the patient. To account for this tradeoff,

a 1 mm slice resolution was chosen for data collection, and, partially due to the size of

Shands Hospital at the University of Florida, enough image sets were eventually found

at this resolution without significant trauma to the facial structures including the entire

orbital region.

While the selection of this slice resolution allows for improved measurement

accuracy for orbital structures, it presents a shortcoming in data available for the entire


103









cranium and brain. At 1 mm slice resolution, the top and back of the head are often left

out of the image, again due to dose considerations to the patient. The lack of this data

prevents the calculation of mean absorbed dose and formulation of dose volume

histograms to these structures since their true total volume is unknown. Fortunately, the

area of interest where dose hotspots may form are included in the image sets and

valuable information can still be tabulated concerning absorbed dose to localized

portions of the anatomy.

Furthermore, the sample size is still too small to be highly confident that the

patient specific variations in anatomy observed are an accurate representation of the

total population. The image sets were stripped of all personal health information in

accordance with our IRB protocol to protect the privacy of subjects whose images were

collected. The sample size is further compromised considering all the images were

gathered from one location (Gainesville, FL). While Shands Hospital at the University of

Florida potentially attracts patients from throughout the state, it still limits the source of

data to one geographical location.

Voxel model geometry presents some disadvantages depending on the voxel

resolution selected for Monte Carlo simulation. Higher voxel resolutions offer superior

geometry detail but slow computer runtime and increased error in mesh tallies. The file

size of high resolution voxel models present difficulties in the following scenarios: (1) in-

house MATLAB codes crash from memory limitation when attempting to load binary

voxel model files, (2) MCNPX has trouble compiling when attempting to load lattice files,

and (3) the MCNPX built in program GRIDCONV fails when attempting to convert mdat

files from mesh tallies to ASCII files. To avoid these problems, appropriate voxel


104









resolutions were selected for the eye and head models: 0.5 mm3 and 1 mm3 voxel

resolution, respectively. The selection of these resolutions allowed for successful

completion of all MCNPX and MATLAB operations, but these resolutions restrict the

dimensions that can be accurately modeled. Specifically, the fovea offset described in

Chapter 1 cannot be modeled with the same level of accuracy as the IRayTM targeting

system. The 1.25 mm offset in the lateral direction will be approximated to 1 mm or 1.5

mm by the Voxelizer code. Despite this limitation, voxel geometry still offers numerous

benefits over stylized geometry including the ability to model tissues with complex

shape and preservation of patient specific anatomy.

6.2 General Conclusions

A new treatment for wet AMD involving kilovoltage stereotactic radiosurgery is

proposed by Oraya Therapeutics, Inc. with the establishment of the IRayTM. The

benefits of this treatment modality include non-invasive application, treatment time and

frequency, and potentially efficacy (pending results of preliminary clinical trials). A

major advancement in quantifying the position of the macula with the Oraya Shift was

incorporated into the targeting system, and subsequently the modeling involved with this

project.

The scope of this research involved the dosimetry characterization of the

treatment scheme using a variety of anthropometric models, both reference and patient

specific. Reference whole body male and female phantoms were designed by

members of the Advanced Laboratory for Radiation Dosimetry Studies (ALRADS) and

were used for this work with the improvement and addition of detail to the ocular

anatomy. Details were derived from data in ICRP 89, NCRP 130, and measurements

taken on 40 head CT scans of equal gender distribution. The statistical analysis of the


105









measurement parameters provided insight into which anatomical structure's sizes are

gender dependent and quantified the optic nerve pathways. Five reference optic nerve

models were evaluated in each head phantom based on the range of optic nerve exit

tilts observed and to indirectly account for non-primary gaze angles. To fully account for

varying gaze and anatomy, several patient specific voxel phantoms were derived from

segmentation of the CT data collected. In all, 16 patient and 32 eye models were

evaluated.

The reference models were voxelized to 0.5 mm3, 1 mm3, and 2 mm3 resolution for

the eye section, head and neck region, and torso, respectively. Each of the patient

specific eye models were voxelized to 0.5 mm3 resolution. The voxelized versions of

the phantoms were imported into MCNPX 2.5.0 Monte Carlo radiation transport code for

simulation of ocular radiotherapy. The results provided insight into the mean absorbed

dose received for several radiosensitive tissues at potential risk for a three beam

treatment cumulating in 24 Gy delivery to the macula.

Mean absorbed doses, dose volume histograms, and effective dose were

evaluated for the reference phantom. DVHs were assessed for macula target and three

non-targeted tissues: lens, optic nerve, and brain. Cumulative mean absorbed doses

to the lens were found to be 124 mGy in the reference male and 127 mGy in the

reference female for the 3-beam treatment. Integral mean absorbed doses to the optic

nerve were 200 mGy and 237 mGy in the reference male and female, respectively. The

lens and optic nerve were of utmost importance and interest, and the absorbed doses

received were below the generally accepted thresholds for cataracts and radiation

induced optic neuropathy (RON).


106









The doses to the remainder of the tissues in the reference phantom were used to

estimate an effective dose as per ICRP Publication 103 schema. The effective dose for

the proposed stereotactic AMD radiotherapy, including contribution from both the

primary tube output and leakage, is estimated to be 0.29 mSv which is a factor of ~102

to 103 lower than seen in external beam radiotherapy, a factor of ~10 lower than seen in

CT imaging, and is comparable to that seen in radiographic imaging of the head and

neck.

Considering the patient specific phantom series (n=32), the dosimetry performed

for kilovoltage stereotactic radiosurgery treatment simulation show that tissues at risk do

not receive tissue-averaged doses over the generally accepted thresholds for

complications, specifically the formation of cataracts and brain necrosis. Likewise, point

doses delivered to the optic nerve were not significant in terms of the risk associated

with developing RON. This study provided a worst-case-scenario risk assessment by

including a range of clinically unrealistic gaze angles, and correspondingly a diverse

range of optic nerve positions. The eye models receiving the highest average or point

doses were further analyzed using dose contour maps. Trends were observed for dose

as a function of gaze angle in the horizontal and vertical directions, and dose escalation

corresponded to increasing optic nerve thickness.

Ultimately, considering the results of this work, the treatment scheme employed by

the IRayTM device has the potential to deliver a therapeutic dose to the macula with

minimal irradiation of non-target tissues within a set limit of clinically realistic gaze

angles. Furthermore, the doses reported in this study could be scaled proportionally for


107









a cumulative therapeutic dose of 16 Gy to the macula tissue, the treatment scheme

currently planned for US clinical trials.

6.3 Future Work

As with any work of complex nature, there always room for improvement and

expanded investigations. The work of this research provided valuable data for the initial

phases of the project, but device development is ongoing and Phase II clinical trials are

currently being set up for the IRayTM

To date, the computational evaluation of the device has explored only variations in

anatomy and gaze, and has not taken into account the uncertainty in the targeting

system. A recent publication by Gertner et a126 claims that the uncertainly and precision

of the machine are 600 and 400 microns, respectively. Therefore, a computational

sensitivity study could be designed to answer the question, "what if we are a little bit

off?" Using the reference phantoms, the treatment axis could be transposed by 600

microns in each of the four major directions; superior, inferior, lateral, and medial to the

center of the macula target. Then, fixing the target point at the center of the macula, the

treatment axis could be rotated in each of the four major directions in increments of one

degree up to five degrees. An evaluation of non-targeted dose could be explored much

like in this present work, but also an exploration of the dose distribution to the macula.

Furthermore, experimental verification in the clinic of the computations performed

could be undertaken. A real time dosimetry system coupled with physical

anthropometric phantoms could not only validate the computational work, but also

describe how dose is deposited as a function of treatment time.


108










APPENDIX A
EXAMPLE OF MCNPX INPUT CODE

c Reference Adult male eye voxel model
c Matrix size [90,119,93]
c Voxel resolution=0.05*0.05*0.05 cm^3
c 10-22-08
c Justin Hanlon
c The University of Florida
read file=mmeanlat noecho
1001 0 -100 fill=999 imp:p=1 $ surrounding box
c ----------
c Body composition and density
c ----------
1 1 -1.03 -70 u=1 imp:p=1 vol=40.269625 $residual soft tissue
4 7 -1.04 -70 u=4 imp:p=1 vol=17.91375 $Brain
11 3 -1.1 -70 u=11 imp:p=1 vol=0.525376 $external nose
12 1 -1.03 -70 u=12 imp:p=1 vol=1.468625 $right Eye (soft tissue)
29 1 -1.03 -70 u=29 imp:p=1 vol=0.952125 $nasal layer (posterior)
57 5 -0.001205-70 u=57 imp:p=1 vol=4.273875 $Air
62 6 -1 -70 u=62 imp:p=1 vol=5.732375 $right vitreous humor (water)
64 8 -1.07 -70 u=64 imp:p=1 vol=0.21125 $right lens
66 1 -1.03 -70 u=66 imp:p=1 vol=0.005875 $right macula
68 7 -1.04 -70 u=68 imp:p=1 vol=0.001125 $right optic disc
71 101-1.525 -70 u=71 imp:p=1 vol=28.22775 $cranium
74 7 -1.04 -70 u=74 imp:p=1 vol=0.5045 $right optic nerve
c ----------
c window and outside of the window
c ----------
1002 5 -0.001205 100 -1000 #2000 imp:p=1 $Out of Voxel inside medium
1003 0 1000 imp:p=0 $OutofROI
2000 9 -19.3 -2000 2010 2030 -2020 imp:p=1 $aperture

c ----------
c surface cards
c ----------
c Matrix size [90,119,93]
c Voxel resolution=0.05*0.05*0.05
100 rpp -2.1 2.4 -2.85 3.1 -2.25 2.4 $origin at center of macula
200 rpp 0 0.05 0 0.05 0 0.05 $0.05 for voxel size
1000 so 200
70 so 200
2000 11 cy 5
2010 11 cy0.1175
2020 11 py-7.45
2030 11 py-7.7

mode p
C Material Cards
c Defined using ICRP 46 tissue compositions
C Soft tissue (male) (rho=1.03)
ml 1000 -0.105
6000 -0.256
7000 -0.027
8000 -0.602


109










11000 -0.001
15000 -0.002
16000 -0.003
17000 -0.002
19000 -0.002
c rest of the material cards omitted for space
c --------------------------------------------------------------------
c tally
c --------------------------------------------------------------------
fc6 right lens
*f6:p 64 $jerks/g
fc16 right macula
*f16:p 66
fc26 right optic disc
*f26:p 68
fc36 right optic nerve
*f36:p 74
c DVH mesh tally for right lens, macula, optic nerve, brain
tmesh
c Matrix size [90,119,93]
c rpp -2.1 2.4 -2.85 3.1 -2.25 2.4
rmeshl:p pedep
coral -2.1 89i 2.4
corbl -2.85 118i 3.1
cord -2.25 92i 2.4
endmd
c --------------------------------------------------------------------
c beam description (180 degree)
c --------------------------------------------------------------------
sdef par=2 x=d4 y=-15 z=d5 erg=d3 dir=dl vec=0 1 0 tr=l 1
# si3 sp3
0.0005 0.00000E+00
c full energy spectra omitted for space
0.1 3.04454E+01
si4 -0.05 0.05
sp4 0 1
si5 -0.05 0.05
sp5 0 1
sil h -1 0.999644633962 1
spl d 0 0.999822316981 0.000177683019
sbl d 0 010
c *tr11 0.08 0.05 0.08 30 90 60 104.4775 30 64.3411 115.6589 120 41.4096 $ beam 5 o'clock
*tr11 0.08 0.05 0.08 0 90 90 90 3060 90 12030 $ beam 6 o'clock
c *tr11 0.08 0.05 0.08 30 90 120 75.5225 30 64.3411 64.3411 120 41.4096 $ beam 7 o'clock
nps 1e7


110











APPENDIX B
SAMPLES OF MATLAB CODES

% import eye phantom binary file
% mesh files (result of mdat files after GRIDCONV) must be in same directory
% and must be named meshl, mesh2, and mesh3
clear all;
name=input('Cropped model name?','s');

%mesh file adjustment
beam=5:7;
for j=1:3;
filename=['mesh',num2str(j)];
fid=fopen(filename);

c=textscan(fid,'%s','delimiter','\n');
fclose(fid);
d=length(c{l});
e=(length(c{l})-10)./2;

%mesh w/o error generation
f=e+10;
for i=ll:f
g{l}{i-10}=c{l}{i};
end
filename=['mesh',num2str(beam(j))];
fid=fopen(filename,'w');
for i=l:e
fprintf(fid,'%s\n',g{1}{i});
end
fclose(fid);
clear f g;

%error mesh generation
f=e+ll;
for i=f:d;
g{l}{i-f+l}=c{l}{i};
end
filename=['error',num2str(beam(j))];
fid=fopen(filename,'w');
for i=l:e
fprintf(fid,'%s\n',g{1}{i});
end
fclose(fid);
clear c d e f g i j filename;
end

%phantom matrix generation
nameidl=str2num(char(name(5:6)));
nameid2=str2num(char(name(8:10)));
nameid3=str2num(char(name(12:13)));
fid = fopen(name);
phantom=reshape(fread(fid,'ubit8'),nameidl,nameid2,nameid3);
fclose(fid);
clear nameidl nameid2 nameid3;


111













phantom(find(phantom==0))=57;
s=size(phantom);


% assign organ tag and density
organ density=[
1 1.03 % residual soft tissue
2 1.04 % Brain


1.525 %
1.04 %
1 %
1.07 %
1.03 %
0.001205];


skull
optic nerve
vitreous humor
lens
macula
% air


% assign organ name
organ name=cellstr(char(...
'Residual soft tissue
'brain
'skull
'optic nerve
'vitreous humor
'lens
'macula
'air


% compose density matrix to be multiplied to flux matrix from mesh tally
density matrix=zeros(s(l),s(2),s(3));
for x=l:s(1);
for y=l:s(2);
for z=l:s(3);
density matrix(x,y,z)=organ density(find(organ density(:,1)==phantom(x,y,z)),
2);
end


end


% read mesh tally for 5, 6, and 7 o'clock beams and sum them up
total meshtally=zeros(s(1),s(2),s(3));
dose werror=zeros(s(l),s(2),s(3));
target dose=0;
for beam=5:7; % 5,6, and 7 o'clock beam direction
target dose=target dose+8;
filename=['mesh',num2str(beam)];
temp=reshape(load(filename),s(2),s(3),s(1));
meshtally=zeros(s(1),s(2),s(3));
filename2=['error',num2str(beam)];
temp2=reshape(load(filename2),s(2),s(3),s(1));
errormatrix=zeros(s(l),s(2),s(3));
for x=l:s(1);
for y=l:s(2);
for z=l:s(3);
meshtally(x,y,z)=temp(y,z,x);
errormatrix(x,y,z)=temp2(y,z,x);
% incorporate computational error
if phantom(x,y,z)==7;


112


I
1
I
1
I
1
I
1
I
1
I
1
I
1











dose werror(x,y,z)=meshtally(x,y,z)-
meshtally(x,y,z).*errormatrix(x,y,z);
else

dose werror(x,y,z)=meshtally(x,y,z)+meshtally(x,y,z).*errormatrix(x,y,z);
end
end
end
end

% unit conversion
% MeV/cm3 / density(g/cm3) 1.6e-13 le3 = Gy/particle
dose werror=dose werror./density matrix.*1.6e-13.*le3;
dose werror(find(phantom==57))=0;
total meshtally=total meshtally+dose werror;
end

% DVH plotting --------------------------------------------------------
% modify dvh organ array for organ tags you're interested in
% currently the number of organs for DVH is limited to 5 which is enough
graph color=cellstr(char('-k','-.b','--r',':k','--b'));
dvh organ=[7 6 4 2];
maximum macula dose=max(total meshtally(find(phantom==7)));
hold off
for i=l:size(dvh organ,2)
legend title(i)=organ name(find(organ density(:,1)==dvh organ(i)));
maximum dose=max(total meshtally(find(phantom==dvh organ(i))));
dose=[0:maximum dose/50:maximum dose];
dvh temp=hist(total meshtally(find(phantom==dvh organ(i))),dose);
dvh temp(51)=0;
for j=size(dose,2):-1:2;
dvh temp(j-l)=dvh temp(j-1)+dvh temp(j);
end
plot(dose/maximum macula dose*target dose,dvh temp./dvh temp(1)*100,char(grap
h color(i)),'LineWidth',2.5);
hold on
end
titlename=['Patient: ',char(name(1:3))];
title(titlename,'fontsize',14);
xlabel('Absorbed dose (Gy) ','fontsize',14);
ylabel('Volume (%)','fontsize',14);
legend(char(legend title),2);

%dose distribution code
%scales doses to give max macula dose 24Gy
maximum macula dose=max(total meshtally(find(phantom==7)));
sourceparticlesneeded=24/maximum macula dose;
dosematrix=total meshtally.*sourceparticlesneeded;

braindose=zeros(1,5);
skulldose=zeros(1,5);
ONdose=zeros(1,5);
lensdose=zeros(1,5);
macdose=zeros(1,5);
headingl=[0.5,1,2,5,10,1000];
heading2=[0.1,0.2,0.3,0.4,0.5,1000];


113












heading3=[5,10,15,20,25,1000];
heading4=[0.5,1,5,12,15,1000];
heading5=[25,40,45,47.5,50,1000];


for i=1:5
brain=0;
skull=0;
ON=0;
lens=0;
mac=0;
for x=l:s(1);
for y=l:s(2);
for z=l:s(3);
if phantom(x,y,z)==4;
ON=ON+1;
if dosematrix(x,y,z)>=heading4(i);
ONdose(l,i)=ONdose(l,i)+l;
end
end
if phantom(x,y,z)==2;
brain=brain+l;
if dosematrix(x,y,z)>=headingl(i);
braindose(l,i)=braindose(l,i)+l;
end
end
if phantom(x,y,z)==6;
lens=lens+l;
if dosematrix(x,y,z)>=heading2(i);
lensdose(l,i)=lensdose(l,i)+l;
end
end
if phantom(x,y,z)==7;
mac=mac+l;
if dosematrix(x,y,z)>=heading3(i);
macdose(l,i)=macdose(l,i)+l;
end
end
if phantom(x,y,z)==3;
skull=skull+l;
if dosematrix(x,y,z)>=heading5(i);
skulldose(l,i)=skulldose(l,i)+l;
end
end
end
end
end
end


%converts #voxels to mm^3
brain=brain.*0.125;
skull=skull.*0.125;
ON=ON.*0.125;
lens=lens.*0.125;
mac=mac.*0.125;
braindose=braindose.*0.125;
skulldose=skulldose.*0.125;


114











ONdose=ONdose.*0.125;
lensdose=lensdose.*0.125;
macdose=macdose.*0.125;

%writes to excel
menu={'OAR','Total Vol','Threshold Vol'};


xlswrite('DVH
xlswrite('DVH
xlswrite('DVH
xlswrite('DVH

xlswrite('DVH
xlswrite('DVH
xlswrite('DVH

xlswrite('DVH
xlswrite('DVH
xlswrite('DVH

xlswrite('DVH
xlswrite('DVH
xlswrite('DVH-

xlswrite('DVH
xlswrite('DVH
xlswrite('DVH


Table.xls',menu(l,:),l,'al:cl');
Table.xls','O',l,'a7');
Table.xls',ON,1,'b7');
Table.xls',ONdose(1,:)1,l 'c7:g7');


Table.xls','B',l,'a5');
Table.xls',brain,l,'b5'
Table.xls',braindose(1,


),1,'c5:g5' ) ;


Table.xls','L', 'a3');
Table.xls',lens,l,'b3');
Table.xls',lensdose(1,:),, 'c3:g3');

Table.xls','M',l,'all');
Table.xls',mac,l,'bll');
Table.xls',macdose(l, :) 'cll:gll');


Table.xls','S',l,'a9');
Table.xls',skull,l,'b9'
Table.xls',skulldose(1,


);
:), i,'c9:g9');


Dose map code
Must find the slice desired and input into dosematrix(x,:,:) part
Must enter the other 2 dimensions of your matrix in the reshape part
Copy .xls output to Sigmaplot and create a 'many Z' contour map

% -------------------------------------------------------
% Axial
% -------------------------------------------------------
two d matrix=dosematrix(:,:,23);
two d matrix=fliplr(two d matrix);



Sagittal


two d matrix=dosematrix(26,:,:);
two d matrix=reshape(two d matrix,114,61);

xlswrite('Dosemap.xls',two d matrix);


115









LIST OF REFERENCES


1. H. Leibowitz, D. E. Krueger and L. R. Maunder, "The Framingham Eye Study
Monograph: an ophthalmological and epidemiological study of cataract,
glaucoma, diabetic retinopathy, macular degeneration, and visual acuity in a
general population of 2631 adults," Surv Ophthalmol 24, 335-610 (1980).

2. R. P. Murphy, "Age-related macular degeneration," Ophthalmology 93, 969-971
(1986).

3. N. M. Bressler, S. B. Bressler and S. L. Fine, "Age-related macular
degeneration," Surv Ophthalmol 32, 375-413 (1988).

4. S. Haddad, C. A. Chen, S. L. Santangelo and J. M. Seddon, "The genetics of
age-related macular degeneration: a review of progress to date," Surv
Ophthalmol 51, 316-363 (2006).

5. J. Gass, Stereoscopic Atlas of Macular Disease and Treatment. (CV Mosby co,
St. Louis, 1985).

6. L. A. Donoso, T. Vrabec and H. Kuivaniemi, "The role of complement Factor H in
age-related macular degeneration: a review," Surv Ophthalmol 55, 227-246
(2010).

7. A. DeWan, L. Mugen, S. Hartman, S. S. Zhang, D. Liu, C. Zhao, P. Tam, W. M.
Chan, D. Lam, M. Snyder, C. Barnstable, C. P. Pang and J. Hoh, "HTRA1
promoter polymorphism in wet age-related macular degeneration," Science 314,
989-992 (2006).

8. R. S. Snell and M. A. Lemp, Clinical anatomy of the eye. (Blackwell Science,
Maiden, MA, 1998).

9. R. W. Young, "Pathophysiology of age-related macular degeneration," Surv
Ophthalmol 31 (1987).

10. J. Gass, "Drusen and disciform macular detachment and degeneration," Arch
Ophthalmol 90, 206-217 (1973).

11. Macular Photocoagulation Study Group, "Visual outcome after laser
photocoagulation for subfoveal choroidal neovascularization secondary to age-
related macular degeneration. The influence of initial lesion size and initial visual
acuity.," Arch Ophthalmol 112, 480-488 (1994).

12. Verteporfin in Photodynamic Therapy Study Group, "Photodynamic therapy of
subfoveal choroidal neovascularization in pathologic myopia with verteporfin. 1-
year results of a randomized clinical trial--VIP report no. 1," Ophthalmology 108,
841-852 (2001).


116









13. P. J. Rosenfeld, D. M. Brown, J. S. Heier, D. S. Boyer, P. K. Kaiser, C. Y. Chung
and R. Y. Kim, "Ranibizumab for neovascular age-related macular degeneration,"
N Engl J Med 355, 1419-1431 (2006).

14. P. A. Quiram, K. A. Drenser, M. M. Lai, A. Capone and M. T. Trese, "Treatment
of vascularly active familial exudative vitreoretinopathy with pegaptanib sodium
(Macugen)," Retina 28, S8-S12 (2008).

15. M. P. Avila, M. E. Farah, A. Santos, J. P. Duprat, B. W. Woodward and J. Nau,
"Twelve-month short-term safety and visual-acuity results from a multicentre
prospective study of epiretinal strontium-90 brachytherapy with bevacizumab for
the treatment of subfoveal choroidal neovascularisation secondary to age-related
macular degeneration," Br J Ophthalmol 93, 305-309 (2009).

16. M. P. Avila, M. E. Farah, A. Santos, Z. Kapran, J. P. Duprat, B. W. Woodward
and J. Nau, "Twelve-month safety and visual acuity results from a feasibility
study of intraocular, epiretinal radiation therapy for the treatment of subfoveal
CNV secondary to AMD," Retina 29, 157-169 (2009).

17. H. Churei, K. Ohkubo, M. Nakajo, H. Hokotate, Y. Baba, J. Ideue, K. Miyagawa,
H. Nakayama, Y. Hiraki, T. Kitasato and N. Yabe, "External-beam radiation
therapy for age-related macular degeneration: two years' follow-up results at a
total dose of 20 Gy in 10 fractions," Radiat Med 22, 398-404 (2004).

18. H. J. Zambarakji, A. M. Lana, E. Ezra, D. Gauthier, M. Goitein, J. A. Adams, J. E.
Munzenrider, J. W. Miller and E. S. Gragoudas, "Proton beam irradiation for
neovascular age-related macular degneration," Ophthalmology 113, 2012-2019
(2006).

19. A. Haas, G. Papaefthymiou, G. Langmann, O. Schrottner, B. Feigl, K. A. Leber,
R. Hanselmayer and G. Pendl, "Gamma knife treatment of subfoveal, classic
neovascularization in age-related macular degeneration: a pilot study," J
Neurosurg 93, 172-176 (2000).

20. M. Hayashi, M. Chernov, M. Usukura, K. Abe, Y. Ono, M. Izawa, S. Hori, T. Hori
and K. Takakura, "Gamma knife surgery for choroidal neovascularization in age-
related macular degeneration. Technical note.," J Neurosurg 102, 200-203
(2005).

21. M. A. Henderson, S. Valluri, S. S. Lo, T. C. Witt, R. M. Worth, R. P. Danis and R.
D. Timmerman, "Gamma knife radiosurgery in the treatment of choroidal
neovascularization (wet-type macular degeneration)," Stereotact Funct
Neurosurg 85, 11-17 (2007).

22. S. V. Goverdhan, F. A. Gibbs and A. J. Lotery, "Radiotherapy for age-related
macular degeneration: no more pilot studies please.," Eye 19, 1137-1141 (2005).


117









23. G. J. Bergink, C. B. Hoyng, R W van der Maazen, J. R. Vingerling, WA van Daal
and A. F. Deutman, "A randomized controlled clinical trial on the efficacy of
radiation therapy in the control of subfoveal choroidal neovascularization in age-
related macular degeneration: radiation versus observation," Graefes Arch Clin
Exp Ophthalmol 236, 321-325 (1998).

24. D. H. Char, A. I. Irvine, M. D. Posner, J. Quivey, T. L. Phillips and S. Kroll,
"Randomized trial of radiation for age-related macular degeneration," Am J
Ophthalmol 127, 574-578 (1999).

25. R. P. Singh, D. Moshfeghi, E. M. Shusterman, S. A. McDormick and M. Gertner,
"Evaluation of transconjunctival collimated external beam radiation for age-
related macular degeneration (ARMD)," AAO Abstract 08-PP-30018701-AAO
(2008).

26. M. Gertner, E. Chell, K. H. Pan, S. Hansen, P. K. Kaiser and D. M. Moshfeghi,
"Stereotactic targeting and dose verification for age-related macular
degeneration," Med Phys 37, 600-606 (2010).

27. M. E. Arnoldussen, M. Shusterman, D. Fletcher, L. Renninger, L. Dang, I.
Koruga, M. Firpo, J. Liang and M. Gertner, "Quantitative Measurements of
Retinal Structures Relative to the Geometric Axis of the Eye," ARVO E-Abstract
3789 50 (2009).

28. C. Lee, E. Chell, M. Gertner, S. Hansen, R. W. Howell, J. Hanlon and W. E.
Bolch, "Dosimetry characterization of a multibeam radiotherapy treatment for
age-related macular degeneration," Med Phys 35, 5151-5160 (2008).

29. ICRP, "ICRP Publication 89: Basic anatomical and physiological data for use in
radiological protection reference values," Ann ICRP 32, 1-277 (2002).

30. NCRP, "Biological effects and exposure limits for hot particles," National Council
on Radiation Protection and Measurements Report No. 130 (1999).

31. M. W. Charles and N. Brown, "Dimensions of the human eye relevant to radiation
protection," Phys Med Biol 20, 202-218 (1975).

32. B. V. Worgul, The Edward S. Harkness Eye Institute Resident's Basic Science
Study Guide. (Columbia University, New York, NY, 1991).

33. R. Unsold, J. DeGroot and T. H. Newton, "Images of the optic nerve: anatomic-
CT correlation," AJR Am J Roentgenol 135, 767-773 (1980).

34. J. A. Rogers, A. G. Podoleanu, G. M. Dobre, D. A. Jackson and F. W. Fitzke,
"Topography and volume measurements of the optic nerve using en-face optical
coherence tomography," Optics Express 9, 533-545 (2001).


118









35. T. Krzizok and B. Schroeder, "Quantification of recti eye muscle paths in high
myopia," Strabismus 11, 213-220 (2003).

36. J. Hanlon, C. Lee, E. Chell, M. Gertner, S. Hansen, R. W. Howell and W. E.
Bolch, "Kilovoltage stereotactic radiosurgery for age-related macular
degeneration: assessment of optic nerve dose and patient effective dose," Med
Phys 36, 3671-3681 (2009).

37. D. Scott, "On optimal and data-based histograms," Biometrika 66, 605-610
(1979).

38. T. Anthoney, Neuroanatomy and the Neurologic Exam. (CRC Press, New York,
NY, 1994).

39. C. Lee, C. Lee, D. Lodwick and W. E. Bolch, "NURBS-based 3-D
anthropomorphic computational phantoms for radiation dosimetry applications,"
Radiat Prot Dosimetry 127, 227-232 (2007).

40. W. S. Snyder, "Estimates of absorbed fractions for monoenergetic photon
sources uniformly distributed in various organs of a heterogeneous phantom,"
Society of Nuclear Medicine MIRD Pamphlet No. 5 (1969).

41. ICRP, "ICRP Publication 23: Report on the Task Group on Reference Man," Ann
ICRP (1975).

42. M. Cristy and K. F. Eckerman, "Specific absorbed fractions of energy at various
ages from internal photon sources," Oak Ridge National Laboratory Report No.
ORNL/TM-8381/Volumes I-VII (1987).

43. C. Lee, D. Lodwick, D. Hasenauer, J. L. Williams, C. Lee and W. E. Bolch,
"Hybrid computational phantoms of the male and female newborn patient:
NURBS-based whole-body models," Phys Med Biol 52, 3309-3333 (2007).

44. ICRU, "Photon, electron, proton and neutron interaction data for body tissues,"
International Commission on Radiation Units and Measurements Report No. 46
(1992).

45. D. B. Pelowitz, "MCNPX User's Manual Version 2.5.0," (Los Alamos National
Laboratory, Los Alamos, NM, 2005).

46. K. Cranley, B. J. Gilmore, G. W. A. Fogarty and L. Desponds, "Catalogue of
diagnostic x-ray spectra and other data," The Institute of Physics Report No. 78
(1997).

47. ICRP, "ICRP Publication 103: Recommendations of the International
Commission on Radiological Protection," Ann ICRP 37, 1-332 (2007).


119









48. M. Zankl, K. F. Eckerman and W. E. Bolch, "Voxel-based models representing
the male and female ICRP reference adult--the skeleton," Radiat Prot Dosimetry
127, 174-186 (2007).

49. B. V. Worgul, Y. I. Kundiyev, N. M. Sergiyenko, V. V. Chumak, P. M. Vitte, C.
Medvedovsky, E. V. Bakhanova, A. K. Junk, O. Y. Kyrychenko, N. V.
Musijachenko, S. A. Shylo, O. P. Vitte, S. Xu, X. Xue and R. E. Shore, "Cataracts
among Chernobyl clean-up workers: implications regarding permissible eye
exporsures," Radiat Res 167, 233-243 (2007).

50. C. A. Girkin, C. H. Comey, L. D. Lunsford, M. L. Goodman and L. B. Kline,
"Radiation optic neuropathy after stereotactic radiosurgery," Ophthalmology 104,
1634-1643 (1997).

51. S. L. Stafford, B. E. Pollock, J. A. Leavitt, R. L. Foote, P. D. Brown, M. J. Link, D.
A. Gorman and P. J. Schomberg, "A study on the radiation tolerance of the optic
nerves and chiasm after stereotactic radiosurgery," Int J Radiat Oncol Biol Phys
55, 1177-1181 (2003).

52. T. Hasegawa, T. Kobayashi and Y. Kida, "Tolerance of the optic apparatus in
single-fraction irradiation using stereotactic radiosurgery: evaluation in 100
patients with craniopharyngioma," Neurosurgery 66, 688-694 (2010).

53. Y. R. Lawrence, X. A. Li, I. el Naga, C. A. Hahn, L. B. Marks, T. E. Merchant and
A. P. Dicker, "Radiation dose-volume effects in the brain," Int J Radiat Oncol Biol
Phys 76, S20-S27 (2010).

54. F. A. Mettler, W. Huda, T. T. Yoshizumi and M. Mahesh, "Effective doses in
radiology and diagnostic nuclear medicine: a catalog," Radiology 248, 254-263
(2008).

55. NCRP, "Ionizing Radiation Exposure of the Population of the United States,"
National Council on Radiation Protection and Measurements Report No. 160
(2009).


120









BIOGRAPHICAL SKETCH

Justin Hanlon was born in 1985 in Nashua, New Hampshire. The older of two

children, he grew up in Auburn, New Hampshire, and graduated from Pinkerton

Academy in 2003. He earned his B.S., a dual degree in nuclear engineering and

engineering physics, at Rensselaer Polytechnic Institute, in 2007.

In 2007, he was granted admission to the Ph.D. program within the Nuclear and

Radiological Engineering Department at the University of Florida to perform research as

a graduate assistant in computational medical physics. Since September 2007, he has

worked closely with Oraya Therapeutics Incorporated, a medical device development

company based in Newark, California, which funded his research.

He completed his Ph.D. degree at the University of Florida in August 2010.


121





PAGE 1

1 COMPUTATIONAL ASSESSMENT OF EFFECTIVE DOSE AND PATIENT SPECIFIC DOSES FOR KILOVOLTAGE STEREOTACTIC RADIOSURGERY OF WET AGE RELATED MACULAR DEGENERATION By JUSTIN MITCHELL HANLON A DISSERTATION PRESENTED TO THE GRADUATE SC HOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Justin Mitchell Hanlon

PAGE 3

3 To my Mom and Dad, for all of the support they have given me ove r the years

PAGE 4

4 ACKNOWLEDGMENTS I would like to express my sincere gratitude to my advisor, Dr. Wesley Bolch, for the opportunity he has afforded me and his guidance toward s the completion of my degree. I would like to extend additional thanks to the rem ainder of my committee: Drs. Erik, not only a member of my c ommittee but an employee of Oraya has always been a pleasure to work with on this project Special thanks are extended to Choonsik, who shared his immense wealth of knowledge and educated me to my present level of expertise. I must also express my appreciation to the rest of the staff at Oraya Therapeutics for both their financial and valuable research support. Michael Gertner and Steven Hansen have had a vision for Oraya and it has been a wonderful opportunity to watch it grow and to have been a small part of the process. W orking with has been a pleasant experience, as he has al ways prov ided invaluable advice and suggestions specifically with the revision of several of my manuscrip ts. Your keen eye has been an integral part of this project I would like to acknowledge NRE staff members Dian a Dampier, Ruth Brumba u gh, Terri Sparks, and Do nna Seifert all of whom have been helpful during my time in the department I would like to recognize Dr. David Gilland, a professor for a number of my graduate courses, for his quality of instruction and engaging classes. I would like to reflect on my past experiences and thank several others who have assisted me in reach ing my goals and dreams I can still remember the first person that inspired my interest in science, and for that I must acknowledge Mrs. Little, my high sch ool chemistry teacher. I w ish to thank her for the invaluable experience of preparing me for college level research at a very young age. I would like to extend my gratitude to

PAGE 5

5 Dr. George Xu, my ad visor at Rensselaer Polytechnic Institute for the opportunity to perform several yea rs of undergraduate research which undoubtedly prepared me for graduate level work. I would also like to thank Dr. Bryan Bednarz, at the time a valuable advice and s erved as a role model during my time at RP I as an undergrad uate student.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIG URES ................................ ................................ ................................ .......... 9 ABSTRACT ................................ ................................ ................................ ................... 11 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 13 1.1 AMD Disease ................................ ................................ ................................ .... 13 1.2 Current Treatments ................................ ................................ ........................... 16 1.3 IRay TM ................................ ................................ ................................ ............... 18 1.3.1 Description of Kilovoltage Stereotac tic Radiosurgery for AMD .............. 18 1.3.2 The Macula and Fovea Offset ................................ ............................... 19 1.4 Objectives of This Research ................................ ................................ ............. 20 2 DIMENSIONAL DATA FOR OCULAR ANATOMY AND THE OPTIC NERVE PATHWAY VIA 1 mm COMPUTED TOMOGRAPHY IMAGE SETS ...................... 27 2.1 Purpose ................................ ................................ ................................ ............ 27 2.2 Data Collection and Measurement Parameters ................................ ................ 28 2.3 Analysis of CT Data ................................ ................................ .......................... 32 2.4 Alternative Approach for Future Stud ies ................................ ........................... 35 2.5 Conclusions Pertaining to Development of Models for AMD Treatment Simulation ................................ ................................ ................................ ...... 36 3 ANTHROPOMETRIC PHANTOMS EMPLOYED ................................ .................... 48 3.1 History of Computational Phantoms ................................ ................................ .. 48 3.2 UF NURBS Hybrid Reference Models ................................ .............................. 49 3.2.1 Head Model Detail ................................ ................................ ................. 50 3.2.2 Ocular Model Detail ................................ ................................ ............... 51 3.2.3 Optic Nerve Model Detail ................................ ................................ ...... 52 3.3 Patient Specific Phantoms ................................ ................................ ................ 53 3.3.1 Selection and Development ................................ ................................ .. 53 3.3.2 Expanded (3D) Angular Measurements ................................ ................ 54 3.3.3 Patient Specific Treatment Planning ................................ ..................... 56 3.4 Voxelization ................................ ................................ ................................ ...... 58

PAGE 7

7 4 COMPUTA TIONAL METHODS ................................ ................................ .............. 70 4.1 The Monte Carlo Radiation Transport Code MCNPX ................................ ....... 70 4.2 Monte Carlo Techniques Used for Treatment Modeling ................................ .... 71 4.2.1 Cell and Surface Cards ................................ ................................ ......... 71 4.2.2 Source Definition ................................ ................................ ................... 72 4.2.3 Tally Specif ication ................................ ................................ ................. 72 4.2.4 Material, Mode, and NPS Cards ................................ ............................ 74 4.3 Post Processing ................................ ................................ ................................ 75 4.3.1 Calculation of Effective Dose ................................ ................................ 75 4.3.2 Utilizing Mesh Tally Output ................................ ................................ .... 77 5 TREATMENT OUTCOME EVALUATION AND ANALYSIS ................................ .... 79 5.1 Radiation Dose Thresholds for Complications ................................ .................. 79 5.2 Reference Model Dose Assessment ................................ ................................ 80 5.2.1 Tissue specific Mean Absorbed Doses ................................ ................. 80 5.2.2 DVH Analysis ................................ ................................ ........................ 81 5.2.3 Effective Dose ................................ ................................ ....................... 83 5.3 Patient Specific Dose Assessment ................................ ................................ ... 84 5.4 Photon Fluence Evaluation ................................ ................................ ............... 86 6 CONCLUSIONS ................................ ................................ ................................ ... 103 6.1 Limitations of This Work ................................ ................................ .................. 103 6.2 General Conclusions ................................ ................................ ....................... 105 6.3 Future Work ................................ ................................ ................................ .... 108 APPENDIX A EXAMPLE OF MCNPX INPUT CODE ................................ ................................ .. 109 B SAMPLES OF MATLAB CODES ................................ ................................ .......... 1 11 LIST OF RE FERENCES ................................ ................................ ............................. 116 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 121

PAGE 8

8 LIST OF TABLES Table page 2 1 Statistical summary of ocular length measurement parameters ......................... 38 2 2 Statistical summary of optic nerve length measurement parameters ................. 38 2 3 Regression coeffic ients, number of samples, and R 2 value for Gaussian probability density function for gender independent measurement parameters 39 2 4 Regression coefficients, number of samples, and R 2 value for Gaussian probability density function for gender dependent measurement parameters .... 39 3 1 Comparison of tissue masses in the UF hybrid NURBS and voxel male head phantoms with those given in ICRP P ublication 89 for the reference adult male ................................ ................................ ................................ .................... 60 3 2 Comparison of tissue masses in the UF hybrid NURBS and voxel female head phantoms with those given in ICRP Publication 89 for the reference adult female ................................ ................................ ................................ ........ 61 4 1 Tissue weighting factors for the calculation of effective dose as given by ICRP 103 ................................ ................................ ................................ ............ 78 5 1 Mean absorbed dose (mGy) to various tissues in the reference head models for a 3 x 8 Gy Oraya Treatment to the right eye ................................ ................. 87 5 2 Mean whole body absorbed doses D T (mGy) for the estimate of the effective dose ................................ ................................ ................................ .................... 87 5 3 The gaze angles, voxelized optic nerve volume, and percentage of that volume receiving more than the absorbed dose listed, representing the dose distribution, for a cumulative 24 Gy treatment dose to the macula ..................... 88 5 4 The voxelized lens volume and percentage of that volume receiving more than the absorbed doses listed, representing the dose distribution, for a cumulative 24 Gy treatment dose to t he macula ................................ ................. 89 5 5 The voxelized macula volume and percentage of that volume receiving more than the absorbed doses listed, representing the dose distribution, for a cumulative 24 Gy treatment dose to the macula ................................ ................. 90 5 6 The highest tissue averaged doses received from the set of 32 eyes undergoing treatment simulation and the associated eye model ........................ 91

PAGE 9

9 LIST OF FIGURES Figure page 1 1 Schematic of an axial cross section of ocular anatomy ................................ ...... 24 1 2 IRay TM (Oraya Therapeutics, Inc., Newa rk, CA) ................................ ................. 25 1 3 Schematic of the treatment geometry for a right eye ................................ .......... 25 1 4 I Guide TM (Oraya Therapeutics, Inc., Newark, CA) ................................ ............. 26 1 5 Typical retinal geography highlighting the optic disc, fovea, posterior pole, and Oraya Shift ................................ ................................ ................................ ... 26 2 1 Schematic of an axial cross section o f a pair of eyes (not to scale), indicating those anatomic parameters obtained via measurement within a single CT image slice ................................ ................................ ................................ .......... 40 2 2 Schematic of a sagittal cross section of an eye (not to scale) indicating those anatomic parameters obtained via measurement within multiple CT images ..... 41 2 3 CT images of the right orbit of male subject A ................................ .................... 42 2 4 Schematic of the Frankfurt plane and normalization parameter ......................... 43 2 5 Dimensions and locations of tissue structures within the human eye as provided in NCRP Report No. 130 ................................ ................................ ...... 44 2 6 Correlation scatter plot for parameter M 1 versus parameter M i .......................... 44 2 7 Histograms for gender dependent ocular measurements ................................ ... 45 2 8 Histograms for gender dependent optic nerve measurements ........................... 46 2 9 Histograms for gender independent measurements ................................ ........... 47 3 1 The whole body male (left) and female (right) reference phantoms developed within the Advanced Laboratory for Radiological Dosimetry Studies .................. 62 3 2 Universit y of Florida NURBS male (left) and female (right) head models based on organ masses listed in ICRP Publication 89 ................................ ....... 63 3 3 Dimensions of the tissues structures in the NURBS eye model ......................... 63 3 4 Engineering drawings of the eye detail embedded within the reference NURBS head model ................................ ................................ ........................... 64

PAGE 10

10 3 5 Male NURBS eye models with five optic nerve vari ations (red), the macula targets (green), and the lenses (blue) ................................ ................................ 65 3 6 Segmentation of the lens (blue), globe (orange), optic nerve (red), brain (purple), and skull (teal) from a 1 mm axial CT im age of the orbital region ........ 65 3 7 Patient specific models in object file format generated from three dimensional reconstruction of 1 mm CT data (shown without skin) ................................ ........ 66 3 8 Scatter plots of optic nerve tilt as a function of gaze angle ................................ 67 3 9 Coronal (left) and sagittal (right) views of the head model voxelized to 1 mm 3 resolution ................................ ................................ ................................ ............ 68 3 10 Cropped eye section voxelized to 0.5 mm 3 resolution ................................ ........ 68 3 11 Axial cross sectional view of a patient specific model voxelized to 0.5 mm 3 resolution ................................ ................................ ................................ ............ 69 4 1 MCNPX plot of the male eye section model voxelized to 0.5 mm 3 resolution ..... 78 5 1 nerves ................................ ................................ ....... 91 5 2 DVHs for the extremes of male optic nerve tilt ................................ .................. 92 5 3 DVHs for the extremes of female optic nerve tilt ................................ ................. 93 5 4 Spatial contour map of the dose distribution within the reference eye model of the adult male ................................ ................................ ................................ 94 5 5 Dose contour maps for patient model mer ................................ .......................... 95 5 6 Dose contour maps for patient model fkl ................................ ............................ 96 5 7 Dose contour maps for patient model fjl ................................ ............................. 97 5 8 Correlation scatter plots of mean absorbed dose to the optic nerve as a function of gaze angle ................................ ................................ ........................ 98 5 9 Correlation scatter plot and linear regression of optic nerve hots pot dose as a function of optic nerve thickness ................................ ................................ ......... 99 5 10 Phase space diagram of energy and angular dependence of photon fluence 50 cm from the macula target ................................ ................................ ............. 99 5 11 Photon fluence distribution plots 50 cm from macula target ............................. 100 5 12 Photon fluence contour maps at the edge of the lattice structure ..................... 101

PAGE 11

11 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 COMPUTATIONAL ASSESSMENT OF EFFECTIVE DOSE AND PATIENT SPECIFIC DO SES FOR KILOVOLTAGE STEREOTACTIC RADIOSURGERY OF WET AGE RELATED MACULAR DEGENERATION By Justin Mitchell Hanlon August 2010 Chair: Wesley Bolch Major: Nuclear Engineering Sciences Age related macular degenerat ion (AMD) is a leading cause of vision loss and a major health problem for people over the age of 50 in industrialized nations The current standard of care, ranibizumab, is used to help slow and in some cases stabili ze the process of AMD, but require s frequent invasive injections into the eye. I nterest continues for stereotactic radiosurgery (SRS), an option that provi des a non invasive treatment for the wet form of AMD through the development of the IRay TM (Oraya Therapeutics, Inc., Newark, CA) The goal of this modality is to destroy choroida l neovascularization beneath the pigment epithelium via delivery of three 100 kVp photon beams entering through the sclera and overlapping on the macula delivering up to 24 Gy of therapeutic dose over a span of approximately 5 minutes. The divergent x ray beams targeting the fovea are robotically positioned and the eye is gently immobilized by a suction enabled contact lens. Device development requires assessment of patient effective dose, reference patient mean absorbed doses to radiosensitive tissues a nd patient specific doses to the lens and optic nerve.

PAGE 12

12 A series of head phantoms, including both reference and patient specific, w as derived from CT data and employed in conjunction with the MCNPX 2.5.0 radiation transport code to simulate treatment and ev aluate absorbed doses to potential tissues at risk. The r eference phantoms were used to evaluate effective dose and mean absorbed doses to several radiosensitive tissues The optic nerve was modeled with changeable positions based on individual patient v ariability seen in a review of head CT scans gathered. Patient specific phantoms were used to determine the effect of varying anatomy and gaze. The results showed that absorbed doses to the non targeted t issues were below the threshold levels for serious complications; specifical ly the development of radiogenic cataracts and radiation induced optic neuropathy (RON) The e ffective dose determined (0.2 9 mSv) is comparable to diagnostic procedures involving the head, such as an x ray or CT scan. Thus, t he c omputational assessment performed indicates that a previously established therapeutic dose can be delivered effectively to the macula with IRay TM without the potential for secondary complications

PAGE 13

13 CHAPTER 1 INTRODUCTION 1.1 AMD Disease Age related macula r degeneration (AMD) is a leading cause of vision loss for people over the age of 50 in the United States and a major health problem worldwide. 1 Advanced AMD has dry and wet forms which lead to blurring or blackening of the central vision while peripheral vision is retained. 2 There is no clear cut definitio n for AMD; some reserve the diagnosis only for patients who experience vision loss, whi le others include patients who have any change ( drusen or geographic atrophy) to their retinal pigment epithelium (RPE) 3 Drusen are small yellowish white deposits found ne ar the fovea. Utilizing the latter form of the definition, the majority of patient s in the early stages do not experience vision impairment until progression into the advanced stages of the disease. In fact, d rusen are present in over half of the populat ion over 70 years of age 1 Considering the entire population in this age group 6 8% have the advanced form of the disease resultin g in severe visual loss. 4 In th e latter stages, central vision loss is associated with either the general geographic atrophy of the RPE (dry form) or development of serous and hemorrhagic detachment of the retina and RPE (wet form). 3 The dry form accounts for about 85% of all AMD cases 4 ; h owever, the wet form accounts for 80 90% of the cases in the advanced stage resulting in severe visual loss 5 AMD can have a profound impact on the quali ty of life of an individual, and unfortunately, despite its importance and severity there are limited treatment options. This is because the pathophysiology of AMD is largely unknown. Research has shown risk associated with the complement factor H gene. Factor H is a 155 kDa sialic acid

PAGE 14

14 containing glycoprotein that helps regulate complement mediated immune system response. 6 One study suggests that a single nucleotide polymorphism in the promoter region of HTRA1 is a major genetic risk factor for AMD 7 Another more recent study found that the polymorphism Y402H in the complement Factor H is related to the development of AMD, and that the pathophysiology of AMD may be related to several other disorders suggesting that the formation of drusen may be a systemic and localized immune system reaction that is observed in several tissues including the eye, kidney, and similar plaques in the brain. 6 However, AMD most likely results from the behavior of multiple genes, age, and hereditary traits. 4 Other risk s include environmental factors, cigarette smoking, diet, fat intake and obesity, high cholesterol levels and heavy sunlight exposure. 4 A brief description of ocular anatomy is essential to understanding the progression of the disease (Figure 1 1) The shell of the eye in the posterior region largely consists of three layers: the retina, choroid and sclera. The sclera is a fibrous tissue that protects and encases the eye The choroid is a dense pigment layer that is rich with vasculature which supplies the inner most layers of the eye with nutrients and other essentials. The retina is a thin ( ~0.5 mm) sensory tissue layer that converts light signal to nerve signal with photoreceptor cells. The macula is a region of the retinal tissue surrounding the fovea which is responsible for central vision inner most layer of th e choroid, itself consisting of five layers: the basement membrane of the RPE, the inner collagenous zone, a central band of elastic fibers, the outer collagenous zone, and the basement membrane of the choriocapillaries. The RPE a part of the retina, pro vides metabolic support to the photoreceptor cells and transports

PAGE 15

15 metabolic waste from photoreceptor cells in th e choroid (choriocapillaries). 8 Drusen are the e arliest clinically detectable sign of the disease, fr om which vision is retained but some patients report having trouble reading in dim light. There are many types of drusen hard, soft, semisoli d, basal laminar, and calcified all of which typically fo rm around the fovea 3 In general, d rusen are small yellowis h deposits that form between the basement membrane of the RPE and the rest of 3 Their manifestation has been linked to free radical formation from visible light which damages photoreceptor molecules. The RPE is unable to digest the damaged cells correctly and metabolism is altered, resulting in a n abnormal secretion of material from the basal cell layer 9 Laboratory investigations have found partially diges ted RPE and retinal cell organelles within 3 However, the form ation of drusen remains unclear because these findings may be the result, and not the cause, of drusen. 10 In the advanced stage of the disease, the dry form of AMD occurs from atrophy of the RPE layer below the retina, causing a loss of rods and cones in the central portion of the eye. The wet (neovascular) form begins with the formation of fibrovascular tissue from the choroids that break s through 3 This choroidal neovascularization (CNV) grows beneath the RPE or into the sensory retina. As a consequence, there is often leakage and bleeding from the vessels that will lead to increase d tension at the macula r lesion, re sulting in serous or hemorrhagic detachment of the RPE, fibro vascular disciform scarring, or vitreous hemorrhage. 3 These events lead to severe and rapid vision loss, ultimately causing blindness if left untreated.

PAGE 16

16 1.2 Current Treatments T here are no treatment option s for dry AMD ; howeve r, vitamin supplements and a ntioxidants have been shown to slow the progression of the disease up 25% over a 5 year period 4 Physicians will monitor the dry form closely until the disease progresses into the more debilitating wet form, for which t here are a number of tre atments used to help slow the process of the disease Current treatment modalities include laser therapy, 11 photodynamic therapy (PDT) using verte porfin ( Visudyne Novartis, Bas e l, Switzerland) 11 12 intraocular drug therapy with ranibiz umab ( Lucentis Genentech, San Francisco, CA) 13 intraocular therapy using pegaptanib sodium ( Macugen OSI Eyetech, New York, NY) 14 and brachytherapy using beta emitting radiation (Epi NeoVista, Fremont, CA) 15 16 Many consider the vascular endothelium growth factor inhibitor (VEGF inhibitor), ranibizumab, to be the curr ent standard of care. VEGF inhibitors are commonly associated with cancer treatments since the antibodies competitively bind with VEGF to prevent the growth of proliferating blood vess e ls. AMD is not cancerous but this treatment is applicable because of the development of choroidal neovascularization that leads to vision loss. The drug is effective but requires frequent, invasive injections into the eye for an indefinite period a burden on the patient and the healthcare system. Research for improved t reatment modalities and combination therapies are ongoing. Previous to the development of ranibizumab radiation based modalities such as external beam photon therapy 17 external beam proton therapy, 18 and Gamma Knife radiosurgery (GKS) 19 21 were explored as potential non invasive treatment options for the wet form of the disease P ilot studies involving the use of both photon and proton external bea m radiotherapy typically employed a treatment scheme where 10 20 Gy is

PAGE 17

17 delivered to the macula in 2 3 Gy fractions. 16, 22 Some have produced results of reduction in vision loss, whereas others have failed to show any benefit, and in some cases have shown deleterious effects, such as cataract formation and xerophthalmia. 22 Nevertheless, there have been sufficient pilot clinical trials to suggest that photon radiotherapy may be a viable option to treat A MD if higher fractions can be applied to the macula ta rget, simultaneously limiting non target tissue toxicity. 22 The most promising study, Bergink et al 23 utilized a treatment scheme delivering a total of 24 Gy in 4 fractions. The results of that study compare the treatment group with an observation group and found significant difference (P<0.08) in terms of visual acuity after 12 months follow up with no side effects from radiation. Fractionation schemes in radiation therapy are vital to the efficacy of treatment and can vary widely depending on the location and type of target. The biological effe ctive dose (BED) to the target tissue is not only a function of total dose, but the number of fractions and dose per fraction. Single fraction application is unusual for cancer tumor treatment, but may have significant impact on the ability to successfull y treat the choroidal neovascularization ( CNV) underlying the macula tissue. Char et al 24 reported borderline positive results by delivering a single fraction of 7.5 Gy, and the Epi ( NeoVista, Fremon t, CA) system has been evaluated clinically using a 24 Gy single fraction treatment scheme. 15 16 The described radiotherapy studies have not demonstrated long term control of the disease. Considering that brachythera py requires surgical intervention, and trials have shown some positive results with delivery of ~24 Gy to the macular lesion a novel non invasive device for radiation treatment of AMD is being developed by Oraya

PAGE 18

18 Therapeutics, Inc. that delivers a higher t herapeutic dose ( 16 24 Gy) in a single fraction Preliminary experimental data recently obtained in a mini pig animal model suggest that single fraction kilovoltage stereotactic radiosurgery can be accomplished without adverse effects. 25 1.3 IRay TM 1.3.1 Description of Kilovoltage Stereotactic Radiosurgery for AMD Based on the studies suggesting that radiosurgery may be a viable option for AMD treatment, the I Ray TM (Figure 1 2 ) has been developed by Oraya Therapeutics, Inc. (Newark, CA) that addresses many of the inherent limitations of other systems used in the past. 26 This one time, non invasive treatment option will provide benefit to patients and the me dical community in terms of cost, pain, and hospital time. The goal of this modality is to destroy the CNV beneath the RPE via delivery of three 100 kVp photon beams entering through the pars plana to overlap on the predicted foveal center delivering up t o 24 Gy of therapeutic dose over a span of approximately 5 minutes. The anode, wit h 1 mm 2 focal spot is 15 cm from the macula target with 0.75 mm Al and 0.8 mm Be filtration. The x ray beams targeting the fovea are robotically positioned and the eye is gently immobilized by a suction enabled contact lens. The divergence of the photon beam is characterized by a profile with a diameter of approximately 3.5 mm upon scleral entry and 4 mm at the retina. The three beams, each intersecting and delivering 8 Gy at the target, were chosen to disperse the scleral entry dose, the dose to the edge of the lens closest to each beam, and the dose to the orbital bone and brain tissue. The beam geometries can be described using a spherical 3D polar coordinate system wi th the z axis aligned with the geometric axis of the eye but transposed by an

PAGE 19

19 offset described subsequently. The geometric axis of the eye is defined as the line that intersects the point on the distal portion of the cornea and is perpendicular to the co rneal curvature at this point (Figure 1 1). The posterior pole is the point where the geometric axis intersects the retina. For all 3 beam angles, the nominal polar angle is 30 degrees; however, if the scleral entry point of the beam is less than 4 mm fr om the limbus for a given patient treatment, the system will readjust that polar angle until the 4 mm criterion is met. The azimuthal angles are described in the coronal plane of the patient such that 0 degrees is superior to the patient and an angle of 9 0 degrees would be towards the nose for a treatment of the right eye. The beam azimuthal entry angles chosen for prototype therapies are 150, 180, and 210 degrees and are commonly referred to as the Schematics of the beam geom e try are presented in Figure 1 3 The treated eye of each patient is gently immobilized by a suction enabled contact lens with a central post and a control yoke housing three optically sensed fiducial markers (I Guide TM ) (Figure 1 4 ) Motion of the patien substantially reduced by the I Guide TM and the residual motion is tracked in real time using a two camera imaging system. Eye motion that would result in substantial dose outside the target area triggers an interruption of the x ray beam. Targ eting error due to patient motion is monitored and maintained below 400 m on the retina. 1.3.2 The Macula and Fovea Offset In ocular anatomy, the fovea is known to be offset from the posterior pole of the eye (Figure 1 1) An evaluation of this separat ion was performed in house at Oraya Therapeutics using a customized Canon CR 45 UAF non mydriatic fundus camera with a low power laser beacon and an anterior imaging system with added collinear at the

PAGE 20

20 at the reflection of the laser beacon and the center of the limbus boundary coincide with the system axis as seen in the anterior eye view, the laser beacon position in the fundus image occurs at the intersection of the geom etric axis with the retina ( the posterior pole). Analysis of these fundus images from seven healthy volunteers showed that the nominal fovea is located 1.25 mm laterally and 0.50 mm inferiorly from the posterior pole. This fovea offset, referred to internally to the research team as t he Oraya Shift is important because the current treatment involves aligning the system to the geometric axis of the eye, followed by a translation of the device by the offset in order to target the nominal fovea. A retinal geography of the fovea offset i s shown in a representative fundus image in Figure 1 5 This offset is important not only for more accurate targeting of subfoveal disease, but also because it moves the treatment beams away from the optic disc, substantially reducing optic nerve dose. Further studies have confirmed the validity of the offset. A total of 48 additional eyes, both healthy and low vision, were analyzed for the position of the fovea relative to the posterior pole. 27 Similar results were obtained. A s part of a larger targeting study, thirteen cadaver eyes were dissected and the position of the fovea was compared to the positions of needles placed t hrough the retina at the point of treatment. The average position of the fovea was found to be within 100 m in the lateral medial direction and 100 m in the superior inferior direction of the nominal fovea as defined by the offset. 27 1.4 Objectives of T his Research Dosimetry characterization of this treatment is an integral part of device development risk assessment, and the ap proval process for the Federal Drug

PAGE 21

21 Administration (FDA). Previous Monte Carlo radiation transport simulations were used to provide insight into beam characterizations for optimal therapy applications including focal spot size, maximum tube potential, and azimuthal angles of beam entry. 28 In the present work the dosimetry characterization of the kilovoltage stereotactic AMD radiosurgery has been expanded and enhanced in a number of areas. Chapter 2 describes the collection and analysis of 40 head CT scans. Evaluation of the ocular anatomy and optic nerve pathway is presented through the statistical analysis of several measurement parameters. The results of this stu dy provide a better quantitative understanding of the optic nerve pathway and indicate the necessity to use separate models for the detailed ocular anatomy of the male and female during treatment simulations. The range of optic nerve exit tilt angles obse rved was utilized to evaluate a worst case scenario risk assessment. Chapter 3 highlights the creation of a reference NURBS based model structure and the fabrication of a 16 patient 32 eye set of pat ient specific voxel models In the process, both a refe rence adult male and reference adult female head model were constructed consistent with the anatomical data of the International Commission on 29 The detailed models include the entire head and neck of the patient including several radiosensitive tissues at potential risk, namely the lens, optic nerve, brain, cranial bone marrow, cranial endosteum, thyroid, and salivary glands These tissues allow for the assessment of effective dose from this treat ment so that comparisons of relative stochastic risk may be made against other medical imaging and therapy procedures. The detailed ocular anatomy combines data from Chapter 5 of Report No. 130 by the National Council on Radiation Protection

PAGE 22

22 and Measureme nt (NCRP) 30 and ICRP 89. Most importantly, the optic nerve pathways, for both male and female and including both mean and extreme s of optic nerve exit tilt, were modeled from data in Chapter 2 of this work. Patient specific models were designed to evaluate dose as a function of varying ocular anat omy and gaze angle. Chapter 4 de tails the Monte Carlo methods used t o simulate treatment and Chapter 5 presents dosimetry calculations in the form of tissue specific mean absorbed dose tables, dose volume histograms (DVHs) color coded dose contour maps, and absorbed dose distribution tables. These latter tables are an alternate data format comparable to DVH plots, but allow for condensed presentation and listing of specific quantitative values. Effective dose is calculated using the reference adult mal e and female head phantoms for a 24 Gy treatment to the macula region. Contributions from both the primary tube output and an estimation of leakage are included in the effective dose calculation. For the patient specific phantoms, t rends in dimensional a natomy as a function of absorbed dose are presented and analyzed. Treatment evaluation and analysis includes a comparison between absorbed dose to non target tissues observed in this study and the generally accepted thresholds for complication and debilita tion, specifically the development of cataracts and radiation induced optic neuropathy (RON). Lastly, the energy and angular distribution of photon fluence at a radius of 50 cm from the macula target is presented, along with photon fluence contour maps th at provide a visual representation of photon fluence at the edge of the lattice structure used during treatment simulation. The conversion of fluence to dose rate will provide parameters necessary for shielding calculations.

PAGE 23

23 Ultimately, the work of this r esearch will provide a better understand ing of the treatment physics and risk of non invasive kilovoltage stereotactic radiosurgery for AMD.

PAGE 24

24 Figure 1 1. Schematic of an axial cross section of ocular anatomy (adapted from Snell 8 )

PAGE 25

25 Figure 1 2 IRay TM (Oraya Therapeutics, Inc., Newark, CA) Figure 1 3 Schematic of the treatment geometry for a right eye (A) sagittal view (6 (B) sagittal view zoomed in (C) front view (D) perspective view C D A B

PAGE 26

26 Figure 1 4 I Guide TM (Oraya Therapeutics, Inc., Newark, CA) Figure 1 5 Typical retinal geography highlighting the optic disc, fovea, posterior pole and Oraya Shift

PAGE 27

27 CHAPTER 2 DIMENSIONAL DATA FOR OCULAR ANATOMY AND T HE OPTIC NER VE PATHWAY VIA 1 MM COMPUTED TOMOGRAP HY IMAGE SETS 2 .1 Purpose Dimensional data on the eye and optic nerve are critical parameters in the development of radiation treatments for eye disease. Device development and safety assessment s require not only the c entral estimates, but also gender dependent variability in a potential patient population. Such data, however, are limited in the open literature. Some reference data are given for ocular size measurements in Chapter 5 of Report No. 130 by the National C ouncil on Radiation Protection and Measurement (NCRP), 30 and in Chapter 11 of Publication 89 by the International Commission on Radiological Protection (ICRP). 29 The NCRP 130 eye model is in part based on measurements reported by Charles and Brown, 31 yet other information in the model is unpublished. 32 Furthermore, ICRP 89 reference values are limited to values of only total eye and lens mass. In both sources, no mention is made of gender dependent variations in ocular structure size or optic nerve pathways. The anatomi cal location and function of the optic nerve is well known in a qualitative manner, yet limited data exist to quantify the position of the optic nerve and its pathway within the tissue structures of the human head. The optic nerve can be distinguished in both CT and MR images, but is difficult to image in any one slice because of its shape and size. 33 There are other imaging modalities such as Optical Coherence Tomography (OCT), that are specialized in imaging the optic nerve for glaucoma related studies, but these tend to focus on the opti c disc rather than the optic pathway. 34 While there is some quantitative information on the movements of the ocular

PAGE 28

28 muscles 35 function of eye gaze angle. In the past, such quantitative knowledge was not necessary for invasive surgical medical procedures invo lving the eye and its orbit since these structures can easily be localized by the ocular surgeon. Recently, the development of non invasive stereotactic radiosurgery has become increasingly popular in medical therapies. Non invasive surgery is a benefit to the patient when considering pain, medical costs, and hospital time. Research for a new type of ocular radiosurgery is underway for the treatment of age related macular degeneration (AMD) which has shown to have numerous benefits over other types of tr eatments for this disease. 28, 36 Due to attenuation of the x ray beams passing thr ough the eye, the proximity of the macula to the optic nerve and disc, and the possibility of damage to these structures from radiation during treatment, the ability to quantify ocular and optic nerve tissue structures is now needed for medical device deve lopment and treatment risk assessment. T o address these needs, a retrospective study of 1 mm slice resolution computed tomography images has been undertaken for a 40 patient population of equal numbers of males and females. The study explored differences between male and female eye and optic nerve sizes, which may affect the attenuation of radiation beams passing through the eye during stereotactic radiosurgery. The study also examined optic nerve pathways in both genders to ascertain correlations with ey e position, as the location of the optic nerve is of utmost importance during radiosurgery treatment planning. 2 .2 Data Collection and Measurement Parameters With Institutional Review Board approval (IRB #481 2007 University of Florida), 40 CT scans were o btained from Shands Hospital at the University of Florida for

PAGE 29

29 retrospective analysis. The gender distribution was 20 male and 20 female. The requirements for eligible CT sets included (1) maxillofacial axial scans, (2) 1 mm slice resolution, (3) soft tiss ue contrast settings, and (4) patient age over 18 years. The CT scans were analyzed and measurements made using the image processing code 3D D octor TM (Able Software Corp., Lexington, MA) Measurements pertaining to the optic nerve included ( 1 ) vertical tilt angle of the optic nerve as it leaves the posterior region of the eye, (2) optic nerve thickness at the posterior region of eye (sc leral optic nerve thickness), (3) optic nerve thickness as it passes through the orbit (orbital optic nerve thickness), and ( 4 ) optic nerve length from posterior of eye to the posterior region of the orbit These 4 optic nerve measurement parameters will be referred to as M 1 to M 4 respectively. M 1 is positive in the superior direction and negative in the inferior directi on Ocular measurements included ( a ) apex of cornea to lens distance (corneal depth), ( b ) lens depth, ( c ) lens width, ( d ) eye depth, ( e ) eye width, ( f ) combined sclera, choroid, and retinal thickness in the posterior hemisphere of the eye (tri layer thickn ess), ( g ) eye separation from apex of the right cornea to apex of the left cornea, ( h ) angle between the Frankfurt Plane (defined below) and the axial plane of the CT image (head tilt angle), and ( i ) vertical gaze angle. These 9 ocular measurement parame ters will be referred to as M a to M i respectively. The parameters measured on axial images a re shown graphically in Figure 2 1 M 1 was calculated using the trigonometric relation tan 1 (X/Y) where X is obtained by counting the number of CT slices between the axial image that showed the inferior portion of the optic nerve exiting the posterior region of the eye and the axial image that

PAGE 30

30 showed inferior portion of the optic nerve as it exits the orbit, and Y is the compressed, 2D length of the optic nerve. M easurement M 4 Theorem with X and Y given for M 1 A representation of this method is shown in Figure 2 2A. In many cases, the optic nerve appeared to significantly change direction at some point in its pathway (i.e., the optic nerve displayed some degree of slack in its pathway at more central gaze angles). In such a case, the angle in M 1 was made to the inflection point, rather than to the posterior of the orbit to accurately depict an exit angle. A repr esentation is s hown in Figure 2 2B The inferior side of the optic nerve was chosen for measurements because a clearer distinction could be made at the optic nerve sclera junction as compared to its superior side. Figure 2 3 demonstrates the method for these measurement s and why it is sometimes important to measure to an alternate point in order to truly represent an appropriate exit angle. Viewing the CT images in an inferior to superior order, the optic nerve can be first seen clearly in slice 135 of Figure 2 3 but i t does not connect to the eye until slice 137, yielding X = 2 and Y represented by the black line in slice 135. If the angle had been measured to the inferior portion of the optic nerve as it exits the orbit, as highlighted by the black circle in slice 1 40, X would have been +3, yielding a superiorly tilted angle when in fact the exit angle is inferiorly tilted. Many of these measured structures can be seen on multiple images with 1 mm slice CT resolution. To keep the ocular measurements (M a through M g ) consistent from patient to patient, a reference slice was chosen that could be easily identified within every CT image set. Typically, the lens could be identified on ~9 CT images and so the median slice of this subset was selected as the reference plane Due to right left head

PAGE 31

31 tilt, this slice may not be the same for the left and right eyes and, in such a case, measurements to the left and right eye were made on separate reference images, and M g was made on either the right or left eye reference slice, whichever had the best view of the opposing cornea. M 2 and M 3 were made on whichever image showed the largest thickness. The reference plane described is the optimal plane for measurement because it would contain the geometric axis assuming the patient sc anned had no head tilt or gaze angle, and it is assumed that the eye is approximately rotationally symmetric around its geometric axis. Under this assumption, the lens and eye width measured correspond to a lens and eye diameter. However, most patients d isplay some form of head tilt and gaze angle within their CT images, and so these parameters were calculated as well. To correct for patient head tilt, M 1 and M i were normalized to the Frankfurt plane. Figure 2 4A shows the Frankfurt plane, defined as an axial plane intersecting both the inferior point of the boney orbit and the superior point of the ear canal. The Frankfurt plane upright reference position. A correctio n factor was measured by performing 3D 3D Doctor TM The resulting polygon mesh file was exported to Rhinoceros 4.0 TM (McNeel North America, Seattle, WA) as shown in Figure 2 4 B, to make an angular measurement between the Frankfurt plane and scanning plane. Parameter M i (vertical gaze angle) was calculated similarly to M 1 by using the trigonometric relation tan 1 (A/B) where A is determined from counting the number of CT slices between the median slice of the lens (reference slice described above) and the

PAGE 32

32 medial slice of the optic nerve in the posterior region of the eye (slice contains a good approximation for the posterior pole even though it cannot be visualized in CT images) and B is the compressed, two dimensional length between the two structures (distance measurement made on one of the images between the lens and optic nerve). Anatomically, A is in the superior inferior direction (z direction), B is in the axial plane (x y plane), and M i is the angle between the two. A representation of this measurement is shown in Figure 2 2 C. 2 .3 Analysis of CT D ata Parameters M i and M h were merely evaluations of patient head and eye positioning during the CT imaging with little corre lation between subjects and thus are not relevant to gender dependent discussions. The standard deviation for vertical tilt (M 1 ) was quite high being 7.8 degrees for men and 8.4 degrees for women. Slice resolution and contrast settings were potential sou rces of error for parameter M 1 although this parameter could also be considered arbitrary since gaze angle was not fixed. C onsequently, it is useful to give the range for this measurement, which was from 17.6 (inferior tilt) to 15.5 degrees (superior ti lt) for the male subjects and from 24.4 (inferior tilt) to 9.7 degrees (superior tilt) for the female subjects. A medial lateral tilt angle was estimated and ranged from +18.9 to +28.5 degrees but was not included in statistical analysis because of the absence of consistent anatomical landmarks within the 2D image set to normalize amongst subjects The se angular measurements did not indicate any patterns that would suggest a difference between right and left optic nerves, nor between males and females. Excluding M 1 M h and M i t he means, standard deviations, uncertainty, and ranges for ocular and optic nerve parameters are given in Tables 2 1 and 2 2 respectively.

PAGE 33

33 The tables also give the t values and p values obtained from performing an unpaired St statistical significance threshold of 0.05 for the p value (95% confidence) to reject the null hypothesis, it was determined that there is a gender dependent difference in means for the following parameters : scleral optic nerve thickness (M 2 ), orbital optic nerve thickness (M 3 ), optic nerve length (M 4 ), eye depth (M d ), eye width (M e ), and eye separation (M g ). For these parameters, gender dependent means, standard deviations, uncerta inty propagations, and ranges are also given in Tables 2 1 and 2 2. The measurements did not indicate patterns that would suggest a difference between right and left eye structures. Uncertainty for each measurement was determined from the diagonal length through a pixel (0.5 mm) and slice resolution (1 mm) of the image set, and propagated to determine the total uncertainty in the mean. The NCRP 130 eye dimensions are shown in Figure 2 5. Comparing current measurements of M a through M f and M 2 to corresp onding NCRP 130 measurements, parameters M a M b M d and M f are in agreement, while parameters M c M e and M 2 indicate that current determination of the diameters of the lens, eye, and optic nerve are smaller in comparison to those given in the NCRP report To establish a relationship between vertical optic nerve tilt and vertical gaze angle, a scatter plot and linear regression equation are shown in Figure 2 6. The relationship established was not strong enough to determine a definitive quantitative relat ionship ( R 2 =0.3377) but there is some correlation between the two parameters. Thus, the vertical optic nerve exit angle is related to the vertical gaze angle, which can be visualized from the three diagrams drawn in Figure 2 2 When looking straight

PAGE 34

34 ahe slack allows the optic disc and nerve to move with the eye when shifting gaze. For an upward gaze, the optic nerve would have a superiorly tilted exit angle, and for a downward gaze it would have an inferiorly tilted exit angle. Furthermore, the length measurements on the anterior portion of the optic nerve give some idea of where a dip normally occurs in the optic nerve slack (14.3 mm for men and 9.6 mm for women). F igures 2 7 and 2 8 give histograms for gender dependent ocular and optic nerve parameters, respectively. Figure 2 9 gives histograms for the remaining gender independent parameters. The bin width in each plot was determined by the following expression: ( 2 1) where W is the bin width, is the standard deviation, and N is the number of samples. The expression gives the optimal bin width for the most efficient unbiased estimation of the probability density function. 37 A three parameter Gaussian probability density function was derived for each histogram using the following model: (2 2 ) where a, b, x o are fitting coefficients and N is the number of samples. These values, along with the R 2 values of the curve, are listed in Tables 2 3 and 2 4 for the ocular and optic nerve parameters, respectively. The original Gaussian fit line was normalized by the number of samples to give a probability density function ranging from 0 to 1. The regression lines are statistically strong for most parameters, with R 2 values above 0.95

PAGE 35

35 for each parameter with the exception of M 2 and M 4 (for males only), which were 0.88 and 0.91, respectively. 2 .4 Alternative Approach for Future Studies While this study provides improved data for these applications, additional improvements are warranted in future studies. CT image contrast is dependent on both the x ray tube potential (kVp) and tube current (mA) used during patient imaging. As this was a retrospective study, these values were not always consistent from patient to patient. As image contrast varies, it became difficult to differentiate structures w ith similar Hounsfield numbers, such as optic nerve fibers and myelin sheath, leading to potential variability during the measurement process. Another factor to consider in this analysis was the 1 mm resolution of the CT images, a value chosen because it is one of the highest resolutions readily available for head CT studies. While higher resolution image sets are occasional taken, they are uncommonly administered in order to reduce radiation dose to the patient. Consequently, the 1 mm resolution offered the highest resolution available for which a relatively large patient population could be sampled. Clearly higher resolution CT image sets would vastly improve some of the measured parameters, namely the vertical optic nerve exit angle and the vertical g aze angle. It should be noted, however, that 3D reconstructions of 2D images, will not provide improved measurement accuracy, as the 3D images involve some data smoothing. Improved accuracy will only be accomplished via higher image resolution, at the ex pense of patient radiation dose. An alternative to the retrospective study of computed tomography images presented here would be a prospective study of patients undergoing magnetic resonance (MR) head imaging. MR images would give improved soft tissue con trast,

PAGE 36

36 and allow for higher resolution slices to be taken without having to account for volunteer radiation dose. Designing such a study would offer a number of other benefits as well, such as limiting head tilt, and scanning patients with forced primary gaze angles as would be the case during stereotactic radiosurgery for AMD. Data from such a study would give a more direct correlation between optic nerve exit angle and gaze angle, and may support the optic nerve movement theory presented here. Relative ly long scan times, however, would be required to achieve images approaching the 1 mm slice resolution used in the present CT based study. 2 .5 Conclusion s Pertaining to Development of Models for AMD Treatment Simulation There is currently a need in the med ical community for gender dependent dimensional data for the eye and substructures, as well as for the optic nerve anatomic position and pathway. Such data are important for the development of non invasive medical procedures such as stereotactic radiosurg ery for treatment of wet AMD. The present study examined forty 1 mm head CT image sets, from which gender dependent means, standard deviations, uncertainty, and ranges were obtained for several anatomic parameters and a relationship between vertical optic nerve exit angle and vertical gaze angle was established. From the results of the sample population, it could be determined with 95% confidence that there is a difference in gender for the total population for parameters M 2 M 3 M 4 M d M e and M g while there is no difference between male and female parameters M 1 M a M b M c and M f Therefore, the average male and female have the same distribution of optic nerve exit angles in the vertical (M 1 ) direction, the same position (M a ) and size (M b and M c ) of t he lens within the eye, and the same combined

PAGE 37

37 sclera, choroid, and retinal thickness (M f ). As for the parameters with a notable difference, an average male optic nerve is longer (M 4 ) and thicker (M 2 and M 3 ) than an average female optic nerve, making the m ale optic nerve slightly more difficult to avoid during ocular stereotactic radiosurgery. An average male eye is larger in depth (M d ) and width (M e ) than that in the average female, wh ich will more heavily attenuate a radiation beam traversing the eye. P arameter M g was also larger for males by a more significant margin indicating a larger skull, which corroborates information on gender dependent reference values given in ICRP Publication 89. As a result, a reference computational model for pre clinical d osimetry evaluations for AMD kilovoltage radiosurgery should employ separate models particularly for the optic nerve detail, for male and female patients.

PAGE 38

38 Table 2 1. Statistical summary of ocular length measurement parameters M a M b M c M d M e M f M g mm mm mm mm mm mm mm Total mean 3.18 0.06 3.82 0.06 8.28 0.06 24.27 0.06 25.11 0.06 1.22 0.06 65.47 0.08 s 0.49 0.56 0.79 0.97 1.22 0.23 4.53 min 2.0 0.5 2.7 0.5 6.6 0.5 21.8 0.5 22.0 0.5 0.8 0.5 56.6 0.5 max 4.6 0.5 5 .5 0.5 10.7 0.5 26.9 0.5 28.2 0.5 2.1 0.5 77.2 0.5 t value 0.613 0.052 1.055 2.635 2.584 1.589 3.122 p value 0.541 0.959 0.295 0.010 0.012 0.116 0.003 Men mean 24.54 0.08 25.45 0.08 67.49 0.11 s 0.98 1.09 4.85 min 22 .4 0.5 23.8 0.5 60.0 0.5 max 26.9 0.5 28.2 0.5 77.2 0.5 Women mean 23.99 0.08 24.77 0.08 63.45 0.11 s 0.88 1.26 3.16 min 21.8 0.5 22.0 0.5 56.6 0.5 max 25.7 0.5 27.6 0.5 68.9 0.5 Table 2 2. St atistical summ ary of optic nerve length measurement parameters M 2 M 3 M 4 mm mm mm Total mean 4.97 0.06 3.61 0.06 28.67 0.07 s 0.89 0.44 3.03 min 3.2 0.5 2.5 0.5 22.7 0.7 max 7.0 0.5 4.9 0.5 37.8 0.7 t value 3.760 2.034 5.412 p value < 0.00 1 0.045 < 0.00 1 Men mean 5.32 0.08 3.71 0.08 30.25 0.10 s 1.01 0.45 3.07 min 3.2 0.5 2.6 0.5 24.3 0.8 max 7.0 0.5 4.9 0.5 37.8 0.7 Women mean 4.63 0.08 3.51 0.08 27.10 0.09 s 0.58 0.42 2.03 min 3.7 0.5 2. 5 0.5 22.7 0.7 max 6.0 0.5 4.7 0.5 30.7 0.5

PAGE 39

39 Table 2 3. Regression coefficients, number of samples, and R 2 value for Gaussian probability density function for gender independent measurement parameters M a M b M c M f M 1 a 25.25 24.06 26.84 2 8.15 23.94 b 0.51 0.61 0.75 0.21 8.89 x 0 3.19 3.82 8.26 1.18 1.93 N 80 80 80 80 80 R 2 0.98 0.99 0.98 0.97 0.96 Table 2 4 Regression coefficients, number of samples, and R 2 value for Gaussian probability density function for gender dependent measu rement parameters M d M e M g M 2 M 3 M 4 Men a 20.46 16.13 10.41 15.98 16.32 15.95 b 0.71 1.11 4.61 1.03 0.44 3.04 x 0 24.46 25.35 67.17 5.45 3.66 30.05 N 40 40 20 40 40 40 R 2 0.97 0.97 0.97 0.91 0.99 0.99 Women a 15.19 15.97 9.61 15.97 18.73 15.27 b 0.94 1.26 3.38 0.58 0.34 2.26 x 0 23.95 24.70 62.63 4.54 3.43 27.04 N 40 40 20 40 40 40 R 2 0.99 0.96 0.95 0.99 0.98 0.96

PAGE 40

40 Figure 2 1. Schematic of an axial cross section of a pair of eyes (not to scale), indicating those anatomic parameters o btained via measurement within a single CT image slice

PAGE 41

41 Figure 2 2. Schematic of a sagittal cross section of an eye (not to scale), indicating those anatomic parameters obtained via measurement within multiple CT image s (A) superior gaze (B) primary g aze (C) in ferior gaze A B C

PAGE 42

42 Figure 2 3 CT images of the r ight orbit of male subject A; s lice progression from 134 to 142 is inferior to superior

PAGE 43

43 Figure 2 4 Schematic of the Frankfurt plane and normalization parameter (A) c ommon ref erence planes in th e head ( adapted from Anthoney 38 ) (B) 3D reconstructed polygon mesh model of the skull demonstrating measurement of head tilt with respect to the scanning plane A B

PAGE 44

44 Figure 2 5 Dimensions and locations of tissue structures within the human eye as provided in NCRP Report No. 130; t he geometric axis is indicated by the red dashed line ; all dimensions are in mm Figure 2 6 Correlation scatter plot for parameter M 1 versus parameter M i

PAGE 45

45 Figure 2 7 Histograms for gende r dependent ocular measurements: (A, B) eye depth, (C, D) eye width, and (E, F) eye separation for males and females, respectively A B C D E F

PAGE 46

46 Figure 2 8 Histograms for gender dependent optic nerve measurements: ( A,B ) optic nerve thickness at sclera ( C,D ) o ptic nerve thickness at orbit and ( E,F ) optic nerve length, for males and females, respectively B F E A C D

PAGE 47

47 Figure 2 9 Histograms for gender independent measurements: (A) corneal depth, (B) lens depth, (C) lens width, (D) combined scleral, choroidal, and re tinal layer thickness, and (E) vertical optic nerve tilt angle E D C A B

PAGE 48

48 CHAPTER 3 ANTHROPOMETRIC PHANTOMS EMPLOYED 3.1 History of Computational Phantoms literature, direct measurement of absorbed dose in living tissue is not possible. One of the most powerful techniques used to estimate organ doses is through the use of Monte Carlo radiation transport codes with computation anthropometric phantoms, which generally consist of three types: (1) stylized phantoms described by 3D geometric surface equations (2) voxel phantoms defined by a set of voxels segmented from medical images, and (3) hybrid phantoms constructed from NURBS surfaces. 39 The Medical Internal Rad iation Dose (MIRD) phantom was the first stylized phantom introduced in 1969, but only included three tissue regions (bone, lung, and soft tissue). 40 The evolution of stylized phantoms advanced to include several age groups and match data from I nternational C ommissi on on Radiation Units (ICRU) Publication 23 41 with the construction of the Oak Ridge National Laboratory (ORNL) phantoms in 1980. 42 Stylized phantoms have sev eral advantages such as smooth organ surfaces and flexibility, but are limited to basic geometric shapes such as quadrilaterals, ellipsoids, cylinders, and spheres. Thus, the phantom world evolved to better model the complex and intricate nature of human anatomy through the development of voxel model phantoms. Voxel phant oms are generated through the contouring of tissues in medical images The data from these two dimensional images are stacked to create a three dimensional matrix of tissue types. This technique is commonly known as segmenting and allows for much more detailed modeling of human anatomy than stylized phantoms

PAGE 49

49 can offer Several i ndependent groups have generated voxel models; h owever they are not useful for universal distribution because t hey are not easily deformable and therefore cannot be directly matched to reference values without more complex model manipulation. As such, voxel models are mostly limited to patient specific anatomy. Additionally, surfaces described by voxel geometry are not smooth because voxel models are created from the compilation of many small rigid shapes (usually cubes) and r the edge of each tissue. Therefore the quality of a voxel phantom relies heavily on its voxel resolu tion, but high resolutions lead to long computer run times during radiation transport. The third type of computational phantoms devised hybrid phantoms, attempt to combine the best features of each of the previous two types. The use of non uniform rati onal B spline surfaces (NURBS) allows for a more complex mathematic model of organ surfaces than stylized phantoms can offer. 39 NURBS based surfaces are defined from a series of control points that can be altered independently for non uniform scaling and the construction of more complex geometries. A major disadvantage of hybrid phantoms is that they cannot be used directly with radiation transport codes and require a process known as voxelization Thus, for dosimetry purposes, t he end result of the hybrid phantom is a voxel model, but with the advantage of having scaled tissue volumes to a desired reference percentile and a user defined voxel resolution. The complete process of voxelization is described subsequently in section 3 .4. 3 2 UF NURBS Hybrid Reference Model s Formulation of a male and female whole body reference phantom was completed within the Advanced Laboratory for Radiation Dosimetry Studies (ALRADS) research group at the University of Florida previously (Figure 3 1) The phantom s were

PAGE 50

50 constructed from (1) s egmentation of patient CT images using 3D D octor TM which allows the regions of interest in ax ial CT slices to be highlighted, and ( 2) m odification of the resulting model to match ICRP Publication 89 50 th percen tile dimensions Following these general procedures, t h e skeleton, body contour, major organs and tissues were segmented from the patient CT data and imported into Rhinoceros 4.0 TM for conversion to NURBS surfaces The d etailed methodology of the entire process has been given previously by the ALRADS research group 39, 43 3 2 .1 Head M odel D etail Two head phantoms, male and female, were extracted from the full body models to be used for kilovol tage stereotactic radiosurgery simulation and are shown in Figure 3 2 As noted above, two versions of hybrid phantoms exist: (1) the NURBS phantom constructed from CT image segmentation and volumetrically adjusted to match indi vidual reference tissue vol umes and (2) its voxelized counterpart. Table 3 1 gives the final tissue masses for both the hybrid NURBS and hybrid voxel adult male head phantoms voxelized at 1 mm x 1 mm x 1 mm resolution Percent differences from ICRP 89 reference values are indicate d along with the ICRP 89 targeted tissue mass and reference densities as given in ICRU Report 46. 44 Corresponding comparisons for the UF hybrid female head phantom are shown in Table 3 2. Several tissues are not of dosimetric concern in this study, but are shown for completeness. Masses for the esophagus and spinal cord are ind icated as partial masses ( fractional volume contained within the head phantoms). In general, targeted masses are achieved within 1 2% of reference values for both genders

PAGE 51

51 3 2 .2 O cular M odel D etail The ey es have a combined reference mass of 15 g and a combined volume of 14.56 cm 3 and the lenses have a combined reference mass of 0.5 g and a combined reference volume of 0.42 cm 3 29 ICRP Publication 89 only gives reference mass and density for the lens and total eye structure, so additional ocular detail was added based on dimensions given in NCRP Report No. 130. However, the dimensions used in NCRP Report No. 130 are not fully compatible with ICRP Publication 89 reference values, and so the dimens ions used to design this model were modified slightly to ensure consistency with ICRP data. The NUR BS eye model with dimensions is shown in Figures 3 3. Anatomically, the macula is a region of the retina surrounding the fovea. Its small size does not p ermit its direct segmentation in CT images, and thus it must be computationally modeled as a separate structure for dosimetry purposes. In this study, the macula was modeled as a cylinder 4 mm in diameter and 0.5 mm in thickness (retinal thickness from NC RP 130 as given in Figure 2 5) The macula is placed within the posterior region of the eye and tr anslated according to the fovea offset described in section 1.3.2 The cylinder is rotated to align with the curvature of the eye. Coronal and axial views of the fovea shift are depicted in Figure 3 4A and 3 4B, respectively. In the coronal view, the red crosshairs depict where the geometric axis would intersect the posterior pole of the eye. As shown in this figure, the macula has been moved 1.25 mm later ally and 0.5 mm inferiorly so that the center of the macula aligns with the fovea. The center of the optic disc is located 3.3 mm from the geometric axis and 4.6 mm from the center of the fovea. The axial view highlights the n ecessity to rotate the macul a with respect to the curvature of the eye. The optic disc is modeled as a cylinder protruding

PAGE 52

52 into the vitreous humor from the end of the optic nerve, consistent with typical ocular anatomy. The ocular models so described were applied to both the refere nce male and female head phantoms. 3 2 .3 O ptic N erve M odel D etail Due to the absence of available data, t he optic nerve measurement parameters described in Chapter 2 were used to design t he reference optic nerve models. The mean of the length measurements (parameters M 2 M 3 and M 4 ) and the ranges of the optic nerve tilt measurements (M 1 and estimation of horizontal tilt) were used to provide representative optic nerve models. Based on analysis presented in Chapter 2, the length measurements of the optic n erve were gender dependent and as such separate models were created for both the male and female. With the IRay TM system, targeting error due to patient motion is monitored and maintained below 400 m on the retina. Therefore, the gaze angle of the patie nt was not taken into account while building these models which is the primary reason the standard deviations of the optic nerve tilt parameters are larger than would be seen in a real patient population undergoing AMD radiotherapy (where the gaze angle i s fixed with the I Guide TM ). As a result, no deviation from primary gaze was assumed during treatment simulation of the reference geometry. However, the use of several optic nerve models, using the full range of optic nerve tilts observed in Chapter 2, wi ll inherently include variation of gaze angle. The primary optic nerve was fashioned from the means of the measurements, while the other four were formulated from the combination of extreme exit angles in the inferior superior direction (value of M 1 ) and t he medial lateral direction ( estimated from 2D images ). For both male and female, t he resulting values used in construction of the

PAGE 53

53 optic nerves were 24.4 to +15.5 in the inferior superior direction and were +18.9 to +28.5 in the medial lateral direction. used for each The male set of optic nerve models is displayed in Figure 3 5. While creating these extreme cas es for optic nerve exit angle, two control points were maintained within head models: (1) the junction of the eye and optic nerve, and (2) the point at the posterior region of the orbit where the optic nerve enters the cranium. With these two points fixed another point was needed to measure the exit angle, and so a plane was placed behind the eye and perpendicu lar to the geometric axis. The plane was placed at a distance equ al to the mean of Y in Figure 2 2B which was 14.3 mm for males and 9.6 mm for fe males. 3 3 Patient Specific Phantom s 3. 3 .1 Selection and D evelopment Utilizing the CT data obtain ed as described in section 2.2 16 image sets were selected for three dimensional reconstruction The selection criterion was based on the initial estimates of vertical gaze angle from measurements taken on the axial CT images ( parameter M 1 ) Ten patients were found to have a vertical gaze angle within 5 o of being parallel to the Frankfurt plane and additional patients were selected in increments of 5 o when available, resulting in six additional patients. Complete three dimensional reconstruction of the 16 patients was accomplished similarly to the method used to construct the full body reference phantoms. The following anatomical structures of interest we re highlighted within the CT head images: t he lens, globe of the eye, optic nerve (from the posterior of the globe to the optic foramen), brain, orbital bone, and skin (Figure 3 6) The top and back of the head are

PAGE 54

54 often left out of the field of view (FOV ) in 1 mm head CT scans, and so only the anterior portions of the brain and skull near the orbit were segmented. This resulted in a model consisting of a small band of tissue surround the ocular anatomy. The resulting polygon mesh files, an example of wh ich is shown in Figure 3 7A were exported to Rhinoceros 4.0 TM to prepare each eye for Monte Carlo based treatment simulation. The patient specific phantoms constructed were not converted to NURBS surfaces, and as such cannot be accurately described as hy brid phantoms. Rather, thes e phantoms are classified as voxel phantoms. 3 3 .2 Expanded ( 3D ) A ngular M easurements An alternate model with an expanded band of segmentation, an example of which is shown in Figure 3 7B, was also exported to Rhinoceros 4.0 TM to re evaluate vertica l gaze angle measurements in 3D and measure optic nerve exi t tilt in 3D. Analysis in Chapter 2 provided initial data on vertical optic nerve tilt normalized to the Frankfurt plane, however estimates of horizontal gaze could not be n ormalized and right left head tilt was not factored into analysis. Thus, a three dimensional evaluation of optic nerve tilt and gaze is warranted. The 3D angular positioning measurement parameters are denoted as: ( ) tilt angle of the optic nerve as it leaves the posterior region of the eye and ( ) gaze angle. Gaze direction was defined as the line that intersects the volume centroids of the eye and lens. A sagittal reference plane was defined by: (1) a point on t he septum in the anterior lobe of the brain, (2) the midpoint of the ear canals, and (3) being perpendicular to the Frankfurt plane, which was again used for the axial reference plane. Both parameters were measured relative to the reference planes and par titioned into vertical and horizontal components. The parameters will be referred to as v and v for the vertical component and h h for the horizontal component.

PAGE 55

55 Scatter plots and linear regression equations are shown for the measured vertical and horizontal components of the angular measurement parameters and in Figure 3 8 along w ith total ( angle measured in 3D environment ) angular measurements The parameters from a linear regression indicate that the reference position of the optic nerve is tilted 0.8 o superiorly and 22.4 o medially. As in Chapter 2, t he results suggest a loose correlation between the gaze angle and optic nerve position. As gaze angle shifts, the optic nerve reacts accordingly by nerve would have a superiorly tilted exit angle, and for a downward vertical gaze it would have an inferiorly tilted exit angle. For an inward (medial) horizontal gaze, the optic nerve would increase the degree of tilt in the medial direction from its reference position, which is already tilted a pproximately 22.4 o in the medial direction. For an outward (lateral) horizontal gaze, the optic nerve will reposition itself with a smaller tilt angle with respect to its reference position in the medial direction. The R 2 values are again well below what would be characterized as a statistically significant correlation, indicating that not all optic nerves have the same reference position and may react differently to changes in gaze angle. Additionally, for the optic nerve to be able to react to changing primary gaze position. The position in which this slack comes to rest may depend on gravity if sufficient time is allowed for the optic nerve to readjust itself within the orbital fat. Furthermore, it may seem counterintuitive that the distribution of horizontal gaze angles is centered over a negative value, favoring a lateral gaze. However, in th is

PAGE 56

56 study gaze angle was defined using the volume centroids of the lens and globe (geometric axis) which is not coincident with the fovea (visual axis). Since the fovea is located lateral to the posterior pole, which intersect s the geometric axis, definin g true gaze angles using the visual axis would shift the distribution medially. In this scenario, the true gaze would be dependent on the distance to the object of focus, which has no relevance in the cl inic for treatment planning. 3.3.3 Patien t S pecific T reatment P lanning After angular measurements were evaluated, each of the 32 eyes (left and right done separately) was prepared for computational treatment simulation using the Rhinoceros 4.0 TM software This was accomplished by: (1) locating the center o f the optic disc (approximated from the three dimensionally reconstructed optic nerve), (2) determining the position of the posterior pole (3.3 mm lateral to optic disc center), (3) determining the position of the fovea (1.25 mm lateral and 0.5 mm inferior to the posterior pole), (4) locating the apex of the cornea from the three dimensionally reconstructed globe, (5) aligning the treatment axis to intersect the fovea and to be parallel with the geometric axis, which is defined as the intersection of the po sterior pole and the apex of cornea, (6) insertion of cylinder with 4 mm diameter and 0.5 mm thickness coincident with the fovea representing the macula tissue, and (7) tagging each structure with a tissue name for voxelization. An important aspect about t his method for treatment simulation should be noted here. The position of each eye model was left as segmented to preserve the anatomy observed directly from the CT data and was not rotated into a gaze position clinically realistic of the stereotactic rad iosurgery treatment. A range of clinically realistic vertical gaze angles was determined from the analysis of 25 healthy volunteers whose vertical

PAGE 57

57 gaze was measured with their heads situated in the IRay TM head support device (unpublished data) The gende r distribution was 15 male and 10 female. The vertical gaze of the individual in the IRay TM system is dependent primarily on anatomical factors, though there is one mechanical degree of freedom that contributes to the vertical gaze angle: the chin rest ca n be moved anterior posterior by 25 mm. To control for this flexibility, the vertical gaze angles were determined at the two extremes of chin rest position. Vertical gaze angle was measured from the Frankfurt plane to be consistent with the computational constrained to gaze 7 below the horizon, and the posts that hold the head rest of the IRay TM stand normal to patient gaze, at 7 from vertical. Image analysis was used to ankfurt plane sits in the head rest with respect to the post. canal, the he ad rest post, and the eye and cheek. A similar image was taken with the chin rest set all the way back. For the purpose of this measurement, the eye is allowed to roam freely as the direction of the gaze during treatment is constrained to be perpendicula r to the post. A line representing the Frankfurt plane was drawn from the top of the auditory canal to the infraorbital rim, based on the folds of the skin, using the ImageJ software. A second line was drawn intersecting the first line, parallel to the h ead rest post. The angle between the Frankfurt line and the post line was measured, from which the vertical gaze angle could be determined. Clinically, the horizontal gaze angles are small as the patient head is placed in the system looking forward, and t he eye is held forward as well. If the head were seated at

PAGE 58

58 an angle in the head restraint, the clinician would re seat the patient for placement of the eye restraint. The clinically relevant vertical gaze angles determined from 25 healthy volunteers range d from 1.7 o inferior to 17.3 o superior with respect to the Frankfurt plane. A reasonable estimate for the range of clinically realistic horizontal gaze angles is within 5 o of the primary gaze position, which is defined as a straight ahead gaze. 3. 4 Voxe lization As mentioned in the description of hybrid phantoms, voxelization is required for use in conjunction with radiation transport. To accomplish this, each organ in the NURBS model s were tagged in Rhinoceros 4.0 TM and then exported in raw format. Usi ng an in house MATLAB code, Voxelizer 6 .0 the models were voxelized to a desired resolution in a binary file. Using another in house MATLAB code, the binary file s were conver ted to lattice file format which is readable by MCNPX. When selecting voxel r esolution, there is a tradeoff between accurately modeling small structures (higher voxel resolution) and efficient computer run times during radiation transport simulation (lower voxel resolution). To account for this tradeoff, three versions of the UF N URBS reference phantom were selected and voxelized at different resolutions. The torso of the reference phantom, to be used in the calculation of effective dose from leakage radiation, was voxelized to 2 mm x 2 mm x 2 mm resolution. The extracted head mo del was voxelized to a resolution of 1 mm x 1 mm x 1 mm ; however, the limiting size of the macula (0.5 mm in thickness) determines the resolution necessary for accurate ocular anatomy Consequently, a finer resol ution ocular model was begot from voxelizin g the full NURBS head model to 0.5 mm x 0.5 mm x 0.5 mm resolution. The finer resolution head model was carefully cropped to

PAGE 59

59 include the entire optic nerve and anterior portions of the brain for the treated right eye using the ImageJ software (NIH, Bethesd a, MD) Similarly, all of the patient specific models were cropped and voxelized to a 0.5 mm x 0.5 mm x 0.5 mm resolution The voxelized versions of the reference head and eye models are shown in Figure 3 9 and 3 10 respectively, and a two dimensional c ross section of a patient specific voxel model is shown in Figure 3 11

PAGE 60

60 Table 3 1. Comparison of tissue masses in the UF hybrid NURBS and voxel male head phantoms with those given in ICRP Publication 89 for the reference adult male Organ System Densit y ICRP 89 UFH NURBS UFH Voxel (g/cm 3 ) mass (g) mass (g) % Diff mass (g) % Diff Eye Structures Eyes (2) 1.03 15 14.946 0.4% 14.843 1.0% Lens (2) 1.07 0.45 0.450 0.1% 0.447 0.7% Macula (2) 1.03 0.013 0.015 Optic discs (2) 1.04 0.002 0.004 Optic Nerves (2) 1.04 1.055 1.058 Respiratory System ET1 (anterior nasal layer) 1.03 2.202 2.062 ET2 (posterior nasal layer) 1.03 13.717 8.582 ET2 (oral cavity layer) 1.03 1.370 1.477 ET2 (larynx) 1.07 28 28.125 0.4% 27.917 0.3% ET2 (pharynx) 1.03 3.944 3.394 Alimentary System Tongue 1.05 73 73.158 0.2% 72.119 1.2% Salivary glands 1.03 85 85.077 0.1% 84.636 0.4% Parotid 1.03 50 50.047 0.1% 49.672 0.7% Submaxillary 1.03 25 25.028 0.1% 24.975 0.1% Sublingua l 1.03 10 10.002 0.0% 9.989 0.1% Tonsils 1.03 3 3.016 0.5% 2.980 0.7% Esophagus (partial) 1.03 8.814 8.792 Skeletal System Cranium 1.38 919.433 905.192 Mandible 1.38 81.439 81.479 Vertebrae C 1.38 148.638 147.657 Intervertebral D iscs 1.10 5.563 3.478 Additional Tissues Brain 1.04 1450.00 1449.641 0.0% 1434.722 1.1% Ears 1.10 14.879 14.874 External nose 1.05 12.985 12.892 Pituitary Gland 1.03 0.6 0.602 0.3% 0.600 0.1% Spinal Cord (partial) 1.04 42.333 38.057 Thyroid 1.05 20 19.956 0.2% 19.906 0.5%

PAGE 61

61 Table 3 2. Comparison of tissue masses in the UF hybrid NURBS and voxel female head phantoms with those given in ICRP Publication 89 for the reference adult female Organ System Density ICRP 89 UFH NURBS UFH Voxel (g/cm 3 ) mass (g) mass (g) % Diff mass (g) % Diff Eye Structures Eyes (2) 1.02 15 14.801 1.3% 14.718 1.9% Lens (2) 1.07 0.45 0.450 0.1% 0.442 1.8% Macula (2) 1.03 0.013 0.022 Optic discs (2) 1.04 0.002 0.001 Optic Nerves (2) 1.04 0.784 0.770 Respiratory System ET1 (anterior nasal layer) 1.07 0.642 0.639 ET2 (posterior nasal layer) 1.02 9.776 9.073 ET2 (oral cavity layer) 1.02 1.145 3.486 ET2 (larynx) 1.07 19 18.999 0.0% 18.728 1.4% ET2 (pharynx) 1. 02 1.549 1.321 Alimentary System Tongue 1.05 60 59.997 0.0% 59.765 0.4% Salivary glands 1.02 70 70.061 0.1% 69.839 0.2% Parotid 1.02 41 41.085 0.2% 40.985 0.0% Submaxillary 1.02 21 20.970 0.1% 20.867 0.6% Sublingual 1.02 8 8.006 0.1% 7.9 88 0.2% Tonsils 1.02 3 2.998 0.1% 2.963 1.2% Esophagus (partial) 1.03 8.244 8.261 Skeletal System Cranium 1.38 799.001 790.616 Mandible 1.38 62.452 62.515 Vertebrae C 1.38 110.662 110.561 Intervertebral Discs 1.10 4.675 3.737 Additional Tissues Brain 1.04 1300 1302.777 0.2% 1287.817 0.9% Ears 1.10 9.200 9.178 External nose 1.05 16.063 16.638 Pituitary Gland 1.02 0.6 0.601 0.1% 0.606 1.0% Spinal Cord (partial) 1.04 11.454 11.972 Thyroid 1.05 17 16.998 0.0 % 16.981 0.1%

PAGE 62

62 Figure 3 1. The whole body male (left) and female (right) reference phantoms developed within the Advanced Laboratory for Radiological Dosimetry Studies

PAGE 63

63 Figure 3 2 University of Florida NURBS male (left) and female (right) head models based on organ masses listed in ICRP Publication 89 Figure 3 3 Dimensions of the tissues structures in the NURBS eye model ; all dimensions are in mm

PAGE 64

64 Figure 3 4 Engineering drawings of the eye detail embedded within the reference NURBS head model; (A) c oronal view of the posterior region of the ri ght eye demonstrating the fovea offset with t he posterior pole located at the intersection of the red lines; all dimensional are in mm (B) a xial view of the poster region of the right eye demon strating the position and rotation of the macula with the geometric axis defined by t he red line extending towards the anterior portion of the eye A B

PAGE 65

65 Figure 3 5 Male NURBS eye models w ith five optic nerve variations (red), t he macula target s (gree n ) and the lenses (blue) Figure 3 6 Segmentation of the lens (blue), globe (orange), optic nerve (red), brain (purple), and skull (teal) from a 1 mm axial CT image of the orbital region

PAGE 66

66 Figure 3 7 P atient specific model s in object file format generat ed from three dimensional reconstruction of 1 mm CT data ( shown without skin ) ; (A) t ypical size of model to be submitted for 0.5 mm 3 voxelization, and (B) e xpanded model that include s both the bottom of the orbit and the ear canals so that the Frankfurt pl ane can be defined for evaluation of 3D measurement parameters A B

PAGE 67

67 Figure 3 8. Scatter plots of optic nerve t ilt as a function of gaze angle; l inear regression s determine the reference position s of the optic nerve (y axis intercept), and the resulting parameters are shown within the plot s ; (A) v ertical components: values positive in the superior direction and negative in the inferior direction (B) h orizontal components: values positive in the medial direction and negativ e in the lateral direction (C) t o tal: all values positive as they were obtained in 3D C B A

PAGE 68

68 Figure 3 9. Coronal (left) and sagittal (right) views of the head model voxelized to 1 mm 3 resolution Figure 3 10. Cropped eye section voxelized to 0.5 mm 3 resolution

PAGE 69

69 Figure 3 11. Axial cross se ctional view of a patient specific model voxelized to 0.5 mm 3 resolution

PAGE 70

70 CHAPTER 4 COMPUTATIONAL METHODS 4 .1 The Monte Carlo R adiation T ransport C ode MCNPX Monte Ca rlo methods use random number generation and probability statistics to solve a variety of physics based mathematical problems. By running a large number of histories (samples of particle tracks ), the stochastic (random) behavior of nuclear particles is averaged and macroscopic trends can be observed. T he physics model underlying Monte Carlo t echniques break down over very small distances ( e.g. sub millimeter for electrons ), but provides an excellent method to simulate radiation transport on a larger scale. This is an invaluable tool essentially allowing radiation based experiments to take pl ace within the safety of computer algorithms. Each particle is tracked to the end of its life (or until it reaches the problem boundary) and each individual physical event (scattering, absorption, etc) is determined by a probability distribution function defined within the nuclear cross section libraries that are stored within the radiation transport code. MCNPX (Monte Carlo N Particle eXtended) is a general purpose Monte Carlo radiation transport code written in Fortran 90 that tracks a variety of radia tion particles over broad energy ranges. 45 MCNPX began as an extension of MCNP4B in 1994 and was developed by Los Alamos National Laboratory (LANL, Los Alamos, NM). The extended version provides an improvement on physics simulation models, extension of neutron, proton, and photon libraries to higher energies, addition of new particle types, and the formation of new tally techniques The developers of the code are so confident in its ability to model radiation physics that they promise a $2 bill for any error found

PAGE 71

71 within the code. Therefore, all s imulations of ocular radiotherapy for AMD presented within this study were performed using the MCNPX version 2.5.0. The structure of the MCNPX code is divided into three sections and a title card. The first two sections, the cell and surface cards, describe the problem geometry. Cells are defined by surfaces using Boolean operators and c ontain information about the material density. The last section, the data cards, contains information that defines the material composition explicitly, the source definitio n, and the tally specification. A few other cards are often given in the data card section ; some are mandatory (mode and nps cards ) but many are optional and allow the user to modify the default settings such as energy cutoff and energy grouping methods. 45 4.2 Monte Carlo T echniques U sed for T reatment M odeling An example input deck is given in Appendix A featuring the mean optic nerve model within the 0.5 mm x 0.5 mm x 0.5 mm resolution male eye section of the reference phantom 4.2.1 Cell and S urface C ards Voxel model geometry was used for all input files in this work, which requires the assignment of universe numbers to each cell within a lattice based on tissue type Two surface cards are necessary for suc h a geometry definition (besides the outer boundary of the problem): (1) an rpp box defining the overall dimension of the lattice structure, and (2) an rpp box defining the dimensions of each voxel. The lattice structure is filled using data from an impor ted lattice (.lat) file, and the result is a lattice box filled with several universes, each with the dimensions of a single voxel. The cell cards define the properties of each universe: material type, density, volume, and importance. This

PAGE 72

72 method is one of the most powerful ways to import complex geometries into an MCNP input deck. A visual of voxel geometry plotted by MCNPX is given in Figure 4 1. 4.2.2 Source D efinition The relevant x ray emission spectrum was generated for a tungsten anode tube operat ed at 100 kVp with anode angle of 12 degree s and total filtration of 0.75 mm Al and 0.8 mm Be using the computer software described in Report No. 78 of the Institute of Physics and Engineering in Medicine. 46 Using the simulated x ray energy spectrum, a divergent x ray beam was modeled as a 1 mm x 1 mm area source represent ing a 1 mm 2 focal spot The divergen ce was modeled to simulate a beam collimated by an explicit tungsten aperture with a beam diameter of 4 mm at the macula target over a source to target distance of 150 mm. The nominal polar angle of 30 from the treatment axis was accomplished using a transformation card, as were the three different treatment azimuth al angle s : 150 180 and 210 L eakage calculation require s a different source definition. The leakage sour ce was defined at a point coincident with the location of the anode. The energy spectrum of leakage radiation is generally characterized as hard, that is, higher energy photons have a greater chance to transverse the tube housing unattenuate d and the rema ining photons are harder to attenuate. The clinical energy spectrum of the leakage radiation is unknown for this medical device; therefore an approximation was made using an 80 and 100 keV mono energetic source of photons. The dosimetry results from both energies were similar and so the 100 keV source was used for final reporting. 4.2.3 Tally S pecification In all, there are eight different tally specification s and four mesh tally options in MCNPX. Tally cards can be modified by a number of keywords and o ther cards,

PAGE 73

73 providing a versatile method for extracting useful data from the problem geometry. Three ways to tabulate dose were utilized for this project: (1) F4 tally for cell fluence modified by a dose response function with DE/DF cards, (2) F6 tally fo r energy The F6 tally was used to calculate all tissue specific mean absorbed doses in this study The *F6 tally in MCNPX reports jerks per gram per photon history. A j er k is an MCNP unit of energy equivalent to 10 9 Joules, thus the reported value for each organ tally was multiplied by a coefficient of 10 12 to co nvert to Gy per photon history using the FM card. Absorbed doses to the brain, thyroid, salivary glands, bone m arrow, and bone surfaces were calculated using the 1 x 1 x 1 mm 3 voxel head model, while doses to the macula, lens, optic disc, and optic nerve were calculated with the 0.5 x 0.5 x 0.5 mm 3 voxel eye model. The 2 mm x 2 mm x 2 mm voxel torso model was used to calculate the leakage contribution for several other radiosensitive organs The F4 tally was set up for calculation of dose to bone marrow and bone surfaces for the cranium, mandible, and cervical vertebrae. The tally was modified by dose response fun ctions described in ORNL/TM 8381/V1 Table D 5 42 The MCNPX output for this tally type is in units of photons/cm 2 and the dose response functions convert photon/m 2 to Gy. Therefore, each bone tally was multiplied by a coefficient of 10 4 to convert to Gy per photon history. lattice structure of the voxel geometry. MCNPX creates a binary output file in the form of a 3D matrix containing the dose to each voxel. Using th e built in GRIDCONV

PAGE 74

74 function in MCNPX, this output file was converted to an ASCII file. Th is output is necessary to beget dose volume histograms and dose contour maps. Two other tally types were used in this study to tally photon fluence: (1) an F1 tally modified by an e0 and c0 card to evaluate energy and angular dependence of fluence, and (2) a type 1 mesh tally without any keyword (default is for fluence). An evaluation of the distribution of photons exiting the head during treatment was necessary in d etermining safety parameters and shielding design for clinical staff present during treatment. An F1 tally sphere, with its origin at the macula target and normal v ector aligned with the treatment axis, was used to evaluate the energy and angular dependen ce of photon fluence at a radius of 0.5 meters. Two dimensional matrices were implemented using type 1 mesh tallies, flush with the edge of the lattice structure which were used to characterize the spatial distribution of photons emanating from each sid e of the model. 4.2.4 Material, Mode, and NPS C ards All of the material cards defined i n this project were derived from ICRU 46 material composition and density data study This mode approximates tha t all secondary particles from interactions (electrons) deposit their energy locally at the site in which it was born (KERMA approximation) This increases computing efficiency by not creating and tracking the secondary electrons. It is a good approximat ion considering the short track length of secondary electrons in the kilovoltage energy group and was validated by Lee et al 28 A total of 10 7 x ray photon histories we re completed for each simulation, and the resulting statistical errors for tissue averaged dose tallies were found to be less than 2%

PAGE 75

75 for the optic disc, less than 1% for all other tissues and ranged from 0.6% to 2% for each macula voxel. 4.3 Post Proce ssing 4.3.1 Calculation of Effective Dose Contribution from primary tube output. Effective dose was determined for a 3 x 8 Gy treatment using steps described in International Commission on Radiological Protection ( ICRP ) Publication 103 47 and the following expression: (4 1) Tissue mean absorbed doses (D T ,R ) are inc luded for the brain, thyroid, salivary glands, active bone marrow, and bone surfaces by scaling the output from section 4.2.3 by the number of histories necessary to deliver an absorbed dose of 8 Gy to the macula for each treatment angle A D T ,R of 0 was assumed for all radiosensitive organs below the neck. Radiation weighting factors ( w R ) can be found in Table B.4 of ICRP Publication 103 and are 1.0 for photons. The equivalent dose for tissues in the referenc e adult male ( ) and female ( ) for each beam angle were summed to give a cumulative equivalent dose for a 3 x 8 Gy treatment and sex averaged Tissue weighting factors (w T ) are found in Table B.2. of ICRP Publication 103 and are given in Table 4 1 The notation in this section is consistent with that in ICRP Publication 103. Several other weighting factors were implemented for the calculation of effective dose. The parotid, submaxillary, and sublingual portions of the salivary glands were volume we ighted to yield a single tissue dose to the target. The bone surfaces (endosteum) were weighted by fraction of bone surface across the entire skeleton. This tissue weighting was as follows: cranium 15.3%, mandible 0.4%, and cervical vertebrae

PAGE 76

7 6 2.1% for ma le and cranium 15.8%, mandible 0.4%, and cervical vertebrae 2.2% for female. 48 The bone marrow dose was weighted by values given in Table 9.4 in ICRP Publication 89 which reports the percentage of active marrow in each bone relative to total body active marrow as a func tion of age. The maximum age of 40 year was used here since the majority of patients undergoing AMD treatment are over the age of 50. Contribution from leakage. The determination of effective dose from the leakage contribution is a more complex process requiring the energy spectrum of leakage radiation and the dose rate at some distance from the anode The clinical leakage energy spectrum for the IRay TM is unknown; therefore an appr oximation was made using a mono energetic source 100 keV photons. A do se rate reading in air near the h eart was taken in the clinic, found to be 12 mR/hr, and used as a conversion factor for MCNPX output. To apply the conversion factor, modified input files were generated filling each universe with air, thus modeling and ir radiating an air phantom. The output from the dose tally for the cells tagged as heart was converted from jerk per gram per history to rad per history by applying a coefficient of 10 14 The leakage rate was multiplied by the treatment time per beam (~2 m inutes), and converted from mR to rad giving the total rad per beam in the clinic The rad per beam was divided by the rad per history from the air phantoms to give a conversion factor based on the total number of source particles necessary to give a dose rate of 12 mR/hr near the heart. The MCNPX inputs were re run with a tissue filled phantom and the conversion factor was applied to the output for dose of each tissue. The remainder of the effective dose calculation was performed in the same manner as f or the primary tube output, but with all organs in the torso considered, and again consistent with ICRP 103.

PAGE 77

77 4.3.2 Utilizing Mesh Tally Output As mentioned a mesh tally was coded exactly overlapping the voxel geometry allowing for a unique manipulation of mesh tally output. An in house MATLAB code was written to link the dose to each voxel from the mesh tally to the original tissue types in the voxel model. This allows the presentation of data in several useful formats including: DVH plots, dose distribu tion tables, and dose contour maps. The code used is given in Appendix B. DVH plots are created within the MATLAB program and the dose distribution tables are exported to Excel. Further man ipulation is necessary to actualize dose contour maps using Micr osoft Excel, Sigmaplot 10.0 TM and a photo editing software.

PAGE 78

78 Table 4 1. Tissue weighting factors for the calculation of effective dose as given by ICRP 103 Tissue Number of tissues w T Total Contribution Lung, stomach, colon, bone marrow, breast, remai nder 6 0.12 0.72 Gonads 1 0.08 0.08 Thyroid, oesophagus, bladder, liver 4 0.04 0.16 Bone surfaces, skin, brain, salivary glands 4 0.01 0.04 Figure 4 1. MCNPX plot of the male eye section model voxelized to 0.5 mm 3 resolution

PAGE 79

79 CHAPTER 5 TREATMENT OUTCOME EVALUATION AND ANALY SIS 5 .1 Radiation Dose Threshold s for Complications The biological response of the human eye to ionizing radiation is well documented. A review of the major findings may be found in Section 5 of Report No. 130 of the National Council on Radiation Protection and Measurement (NCRP) Much of the experience on the radiation response of the eye derives from studies with fractionated and chronic regimens of low LET radiation. The radiation absorbed doses that produce minimally dete ctable changes or functional disabilities are 6, 5 30, 15, 16 and 25 Gy, for the lid, conjunc tiva, cornea, sclera, iris and retina, respectively The corresponding visually debilitating absorbed doses to these same ocular structur es are 40, 35, 30, 200 16, and 25 Gy, respectively 30 Until rece ntly the threshold for minimally detected changes or functional disability to the lens was reported to be 2000 mGy and the threshold for visual debilitation was reported as 5500 mGy. 30 However, more recent studies suggest that the lens is much more sensitive to ionizing radiation than previously t hought. A recent report suggests that t he tissue averaged dose threshold f or radiation cataractogenesis could be as low as 700 mGy. 49 This limit will be used in this study in the analysis of treatment outcomes to err on the side of greater radiological protection. The radiological sensitivity of the optic nerve has been studied in patients whose optic nerve was unavoidably or unintentionally irradiated as a consequence of brain or head tumo r radiotherapy showing that doses 8 Gy or higher might have some adverse consequence. 50 Another study found that doses less than 12 Gy to a short segment of the anterior optic apparatus during stereotactic radiosurgery resulted in a lo w risk

PAGE 80

80 (~1.1%) for radiation induced optic neuropathy (RON); however, 3 of 4 patients in this study that developed RON had previously received external beam radiation therapy (EBRT) and the other had undergone two previous radiosurgery procedures. 51 It is also unclear what percentage of volume characterized the short segment of the anterior optic apparatus. Ultimately, the authors conclude that point doses up to 12 Gy are well tolerated by patients whose optic nerve has not been previously irradiated. Furthermore, a recent study suggest that the optic apparatus may be more tolerant to radiation than previously thought, able to receive up to 14 Gy without risk of developing RON (again under the assumption that the patient has not previously undergone radiation therapy). 52 Other tissues of interest include the brain and orbital bone. The development of necrosis in brain tissue due to radiological toxicity is summarized by Lawrence et al and it has been determined that the threshold for neurological toxicity is 12 Gy for a volume of at least 5 10 cm 3 53 Bone, and in this case orbital bone, contains elements with higher atomic numbers that have a higher cross section (p robability) for photon interaction, namely the photoelectric effect, than soft tissue and fat. The resulting secondary particles (electrons) are likely to deposit their energy locally and as such the bone absorbs more dose than surrounding tissues. Howev er, the skull is fairly radio resistant to adverse consequences and the orbital bone contains a negligible percentage of the total active marrow in the cranium. 29 5.2 Reference Model Dose Assessment 5.2. 1 Tissue specific Mean A bsorbed D oses Mean a bsorbed doses to several non targe ted tissues are shown in Table 5 1 for a 3 x 8 Gy treatment to the right eye. The MCNPX output has been normalized by the

PAGE 81

81 number of photons necessary to deliver a dose of 8 Gy to the macula target for each beam The lens received a 3 beam integral mean dose of 124 mGy and 127 mGy in the reference male and female patient, respectively well below the threshold for cataractogenesis. The optic disc receives a mean dos e ranging from 200 to 239 mGy, a factor 33 to 40 times le ss than that for the macula target despite its proximity to the macula target (4.6 mm) The average absorbed dose to the optic nerve, a critical part of this project, was found to increase with increasing beam azimuthal angle for the treated right eye position. The highest optic nerve mean dose recorded was 112 mGy in the reference female patient at a 210 0 beam entry angle. Other non targeted tissue doses were also found to be insignificant. The highest dose to the optic nerve opposite treatment was 2.4 mGy for the female at a 210 0 beam entry. The mean absorbed doses to the brain ranged from 3.0 to 4.9 mGy per treatment beam. The mean absorbed doses to the thyroid and saliva ry glands were 4 to 5 orders of magnitude less, respectively, than that to the target for all beam angles and both sexes. Considering the mean absorbed doses to the bone structures in the head, the cranium received the highest dose to both the active marr ow and endosteal tissues, but these doses were on the order of 9 mGy or less. 5.2. 2 DVH A nalysis Dose volume histograms for the male and female mean optic nerve model are shown in Figures 5 1 for a cumulative 3 x 8 Gy treatment. These histograms display how the absorbed dose is distributed throughout the structure of the macula, lens, brain, and optic nerve. Ideally the macula histogram would be a step function but the absorbed dose the macula is not quite evenly distributed, and the plots are thus

PAGE 82

82 desi gned to show a treatment with a maximum dose of 24 Gy to the target. For both sexes, a steep drop is observed for the lens, indicating that there are no hotspots. The brain DVH also drops off steeply, but small hotspots in the anterior portions of the br ain (<10%) contribute doses ranging up to about 2.5 Gy. Nevertheless, for females only about 5% of the brain volume receives a dose exceeding 2.5 Gy and 95% of the male brain receives a dose less than 2 Gy. hat 2 3% of that structure receives a dose exceeding 2 Gy, and that approximately 1% receives a dose exceeding 5 Gy. Less than 1% of the optic nerve volume receives a dose over 10 Gy with a shows that 2 3% of the tissue volume receives more than 1 Gy with less than 1% receiving a maximum the maximum dose is a function of optic nerve diameter at the poste rior of the eye (value of M 2 ) since the mean of this measurement is 0.7 mm less for females. This correlation will be explored more subsequently. Due to the lack of literature on quantifying the anatomical location of the optic nerve, and the somewhat lar ge standard deviation of parameter M 1 in this study, DVHs were created for the combination optic nerve exit angle in the superior inferior and medial lateral directions and are shown in Figure 5 2 for males and in Figure 5 3 for females. The se DVH plots a DVH except for the superior lateral combination (upper right graphs in these figures). An optic nerve oriented in such a position is likely to come close to the beams exiting the eye, and patients in this case may receive localized optic nerve doses as high as 17.5

PAGE 83

83 Gy for males and 15 Gy for females to less than 1% of its volume. Nevertheless, only 2 exceeding 8 Gy. This e xtreme situation for the reference male patient is shown in the form of a dose contour map in Figure 5 4 As can be seen, only the outermost periphery of the optic nerve is irradiated in this worst case optic nerve position. Anatomically, this region of the optic nerve would constitute the insulating myelin sheath and would most likely not be of radiobiological importance to nerve function. Thus the risk of radiation optic neuropathy (RON) is hypothesized to be exceedingly small 5.2. 3 Effective D ose Th e calculation of effective dose combined both contributions from primary tube output and leakage estimation. For the primary output, t he relevant absorbed tissue doses necessary to calculate effective dose are listed in Table 5 2 for the reference male an d female head models. The calculation, based on ICRP Publica tion 103 recommendations, yield s an effective dose of 0.28 1 mSv for a 3 x 8 Gy treatment to the macula per eye For leakage contribution, all radiosensitive organs in the head and torso were con sidered, and the result of the ICRP 103 based calculation was 0.009 mSv. Thus, the total effective dose for a 3 x 8 Gy treatment is 0.29 mSv. While the stereotactic AMD radiosurgery is therapeutic in nature, this value of effective dose compares very favo rably with radiographic imaging doses including skull radiographs (0.1 mSv) and cervical spine radiographs (0.2 mSv), yet are much lower than seen in CT scans of the head (2 mSv) or neck (3 mSv). 54 In contrast, estimates of effective dose in megavoltage radiothe rapy are significantly higher as in treatments of head and neck tumors (1870 mSv), brain metastases (270 mSv), and brain primary tumors (580 mSv). 55 Clinical trials are still underway and a smaller therapeutic dose

PAGE 84

84 may eventually be prescribed. For example, if a 16 Gy total dose to the m acula is shown to yield good clinical outcome, the estimated effective dose would be proportionally lower (e.g., 0.19 mSv). 5. 3 Patient Specific Dose Assessment Dose volume histograms are the most common form of data used to evaluate non target complicatio n probabilities (NTCP). An alternative to DVH output was contrived for reporting doses to non targeted tissues in the patient specific population. The alternative approach allows for a more condensed data format with specific quantitative valu es listed i n the form of dose distribution tables. The column headings in these tables are equivalent to the x axis of a DVH and the values in the tables correspond to the shape of a DVH. T he total voxelized volume of the optic nerve lens, and macula and the perce nt volume of that tissue over several do se regions are shown in Tables 5 3, 5 4, and 5 5 respectively. The dose regions in the table were chosen in an effort to most clearly depict the distribution of dose within each tissue volume. A conservative appro ach was taken in reporting volumes to the optic nerve and lens ; the reported volumes include all voxels wherein the mean dose plus the computational uncertain ty surpasses the given table heading The reverse is true for the macula table, that is the mean dose minus the computational uncertainty surpasses that given table heading. The details of the MATLAB code used for this calculation are shown in Appendix B. The abbreviation for each eye model is as follows: the first letter denotes gender (m/f), the second letter denotes subject (A,B,C etc), and the third letter denotes left or right fdl Table 5 3 is organized in three parts according to clinical realism. A patie nt model was selected from each of the groups for the formulation of dose contour maps which

PAGE 85

85 are shown in Figures 5 5, 5 6, and 5 7 Table 3 presents the highest (of the set of 32 eye models) tissue averaged dose to the lens and optic nerve, along with th e associated eye model. For all models, no brain voxel received a dose over 12 Gy the threshold for necrosis Patient model fdl received the highest orbital bone dose with 5 voxels (1.125 mm 3 ) receiving between 45 and 50 Gy. Despite the variability of t he location of the optic nerve observed in this study, the highest cumulative tissue averaged dose received was 1.3 Gy by model fkl (Table 5 6). Dose contour maps were fabricated for this patient (Figure 5 6 ) and it can be seen that the overlapping beams avoid the optic nerve. This patient demonstrated a lateral horizontal gaze of roughly 16 o outside the range of clinically relevant horizontal gaze angles. It is somewhat unclear from the literature what maximum point dose is tolerated by the optic nerve ; nevertheless, it is reasonable to assume that the risk of developing RON is negligible for all simulated patient models in this study given the dose volume data presented in Table 5 3 The highest tissue averaged dose observed in the present study was 17 6 mGy by eye model fol (Table 5 6 ). Eye model fjl had the highest percentage of volume receiving doses over 300 mGy and no lens volume received a dose exceeding 400 mGy (Table 5 4 ). The sagittal dose contour map and the most inferior voxelized slice that contains the lens of patient fjl are shown in Figure 5 7 and both clearly depict that the converging beams do not directly intersect the lens. An attempt was made to find correlations between absorbed dose to non targeted tissues and the measurement param eters mentioned before. For most tissues, there were no statistically significant relationships. Analysis of optic nerve data provided two

PAGE 86

86 significant correlations. These are presented in Figures 5 8 and 5 9. The latter relationship is intuitive; the o ptic nerve dose will escalate with increasing optic nerve thickness. The former may not be intuitive at first, but becomes clearer with a better understanding of optic nerve tilt as a function of gaze angles. The logarithmic regressions suggest that opti c nerve dose increases as vertical and lateral gazes increase. As describe d in Chapters 2 and 3, an increasing vertical gaze will stretch the optic nerve into a more vertical exit tilt, which would position the optic nerve closer to the beams exiting the eye. As lateral gaze increases, the optic nerve tilt decreases (approaching a limit of being positioned in parallel with the sag ittal reference plane described in section 3.3.2) and positions itself in closer proximity to the beam entering from the latera 5. 4 Photon Fluence Evaluation Figures 5 10, 5 11, and 5 12 show that p hoton fluence was mostly forward directed with r with maxima in the 50 60 o and 44.4 50 keV bins. A portion of the beam traverses through the head unattenuated, resulting in a smaller peak around 150 o Color coded contour fluence ma ps are presented in Figure 5 13 to supply a visual representation of fluence. The results of this study provide a better understanding of the radiation physics involved during treatment administration and contribute d towards the design of shielding for th e treatment device.

PAGE 87

87 Table 5 1 Mean absorbed dose (mGy) to various tissues in the reference head models for a 3 x 8 Gy Oraya Treatment to the right eye ; computational error was less than 2% for the optic disc and less than 1% for all other tissues Ma le Female Beam Azimuthal Angle : 150 o 180 o 210 o 150 o 180 o 210 o Right Macula 8000 8000 8000 8000 8000 8000 Right Lens 46.7 40.6 36.5 47.3 42.7 37.0 Right Optic Disc 211 200 201 239 210 212 Right Optic Nerve 48.2 56.8 93.8 56.4 68.5 112.0 Left Optic Nerve 0.92 1.11 1.40 1.84 2.00 2.44 Brain 2.96 3.72 4.25 3.82 4.86 4.94 Thyroid 0.05 0.05 0.06 0.09 0.09 0.09 Parotid 0.31 0.28 0.30 0.53 0.51 0.52 Submaxillary 0.39 0.38 0.42 0.34 0.36 0.38 Sublingual 0.23 0.24 0.26 0.59 0.58 0.57 Cranium (AM) 3.64 3.25 2.86 3.95 3.54 3.28 Mandible (AM) 0.29 0.29 0.30 0.48 0.47 0.47 Vertebrae (AM) 0.14 0.15 0.19 0.27 0.31 0.37 Cranium (BS) 8.86 7.94 7.04 9.67 8.71 8.10 Mandible (BS) 0.94 0.92 0.97 1.56 1.53 1.51 Vertebrae (BS) 0.46 0.49 0.60 0.87 0.99 1.18 Tab le 5 2. Mean whole body absorbed doses D T (mGy) for the estimate of the effective dose Male Female Beam Azimuthal Angle : 150 0 180 0 210 0 150 0 180 0 210 0 Brain 2.96 3.72 4.25 3.82 4.86 4.94 Thyroid 0.05 0.05 0.06 0.09 0.09 0.09 Salivary Glands 0.32 0.31 0.33 0.48 0.47 0.48 Bone Marrow 0.29 0.26 0.23 0.31 0.28 0.27 Bone Surfaces 1.37 1.23 1.09 1.55 1.40 1.31

PAGE 88

88 Table 5 3. The gaze angles, voxelized optic nerve volume, and percentage of that volume receiving more than the absorbed dose listed, repres enting the dose distribution, for a cumulative 24 Gy treatment dose to the macula Model Vertical Gaze Angle Horizontal Gaze Angle Optic Nerve Voxel Volume 0.5 Gy 1 Gy 5 Gy 12 Gy degrees degrees mm 3 % % % % Both vertical and horizontal gaze are clinicall y realistic fkr 9.0 3.2 556.8 8.7 3.2 0.9 0 fnl 9.5 2.2 664 7.4 2.1 0.6 0.2 mel 0.8 1.0 632.5 9.2 3.5 1.3 0.2 mer 1.0 0.3 621.9 6.8 2.4 0.7 0 mkl 5.2 2.9 599.4 10.7 3.7 1.5 0.4 mkr 8.7 1.2 702.8 9.9 4.1 1.9 0.6 mll 7.2 4.4 523.8 11.3 4.0 0.8 0 Either vertical or horizontal gaze is clinically realistic mlr 4.5 6.0 640.3 16.4 8.8 4.1 0.9 fkl 8.7 16.4 521.6 26.9 15.8 6.6 0.3 mml 1.6 16.7 590.6 18.1 9.0 3.5 0.1 fnr 12.6 18.1 704.9 36.7 27.7 4.8 0.3 fsr 3.6 3.3 551.3 10.6 4.0 1.6 0.4 m ar 6.0 1.5 566.8 9.0 2.8 1.1 0.2 ffl 10.9 3.3 287.8 8.2 0.9 0 0 for 17.7 0.5 460.5 6.1 1.6 0.4 0 fol 20.2 2.6 547.1 3.4 0.7 0.1 0 Neither vertical or horizontal gaze is clinically realistic ftr 2.4 6.0 410.3 11.8 4.7 1.3 0.1 mgr 5.3 19.0 909.1 11.8 7.1 3.8 0.9 mal 5.4 5.9 431 11.4 4.3 1.8 0.3 mgl 6.4 8.6 686.1 8.6 3.6 1.5 0.4 mor 22.1 14.5 445 14.3 7.4 2.0 0.3 fsl 6.8 6.1 482.4 12.8 4.8 1.8 0.3 mfr 7.4 18.9 621.9 11.4 5.6 2.9 0.8 ffr 7.5 8.5 356.5 9.2 2.3 0.6 0.1 mmr 7.9 17.5 640.1 17.8 11.4 4.9 0.2 ftl 8.5 8.1 588.8 10.2 3.4 0.7 0.1 mfl 8.6 5.4 776.4 10.0 4.5 2.1 0.6 mol 30.1 29.8 480 31.4 17.5 6.0 0.2 fdr 34.4 8.7 207.9 7.9 1.4 0.2 0 fdl 34.6 7.2 259.1 4.5 0.2 0 0 fjr 41.4 19.8 395.0 5.1 0.9 0.3 0.1 fjl 46.5 28.0 728.6 6.7 2.2 1.1 0.4

PAGE 89

89 Table 5 4 The voxelized lens volume and percentage of that volume receiving more than the absorbed doses listed, representing the dose distribution, for a cumulative 24 Gy treatment dose to the macula Model Lens Voxel Volume 100 mGy 200 mGy 300 mGy 400 mGy mm 3 % % % % fdl 133.6 86.0 5.7 0 0 fdr 128.3 90.7 8.2 0 0 ffl 100.4 87.5 7.0 0 0 ffr 109 87.6 7.5 0 0 fjl 135.9 90.2 26.1 3.1 0 fjr 115.8 80.6 11.9 0.3 0 fkl 74.6 96.1 8.2 0 0 fkr 94.1 99.3 18.9 0 0 fnl 123 .6 82.5 2.0 0 0 fnr 114 91.6 4.9 0 0 fol 101 99.9 30.3 2.4 0 for 94.8 98.5 18.2 0 0 fsl 142.8 94.0 7.2 0 0 fsr 146 95.3 6.8 0 0 ftl 104.5 84.4 2.9 0 0 ftr 122.9 81.3 4.4 0 0 mal 138.9 78.2 4.8 0 0 mar 144.3 97.6 27.5 2.5 0 mel 97 97.0 12.1 0 0 m er 97.9 99.5 15.8 0 0 mfl 118 81.9 0 0 0 mfr 92.9 90.8 0 0 0 mgl 159.4 83.5 7.8 0 0 mgr 148.1 84.5 6.8 0 0 mkl 111 99.0 12.0 0 0 mkr 102.4 91.0 1.3 0 0 mll 139 89.0 8.4 0 0 mlr 133.9 92.9 7.5 0 0 mml 134.4 93.1 8.1 0 0 mmr 131.1 83.9 4.9 0 0 mol 119.8 89.0 10.3 0 0 mor 113.4 96.3 13.3 0 0

PAGE 90

90 Table 5 5 The voxelized macula volume and percentage of that volume receiving more than the absorbed doses listed, representing the dose distribution, for a cumulative 24 Gy treatment dose to the macula Model Macula Voxel Volume 5 Gy 10 Gy 15 Gy 20 Gy 25 Gy mm 3 % % % % % fdl 6.3 100 100 100 100 0 fdr 6.1 100 100 100 95.9 0 ffl 6.3 100 100 100 100 0 ffr 6.3 100 100 100 100 0 fjl 6.4 100 100 100 96.1 0 fjr 6.1 100 100 10 0 100 0 fkl 6.5 100 100 100 100 0 fkr 6.1 100 100 100 95.9 0 fnl 6.4 100 100 100 96.1 0 fnr 6.3 100 100 100 96.0 0 fol 6.0 100 100 100 95.8 0 for 6.3 100 100 100 100 0 fsl 6.3 100 100 100 100 0 fsr 6.4 100 100 100 96.1 0 ftl 6.3 100 100 100 100 0 ftr 6.1 100 100 100 95.9 0 mal 6.3 100 100 100 100 0 mar 6.3 100 100 100 100 0 mel 6.1 100 100 100 100 0 mer 6.4 100 100 100 96.1 0 mfl 6.1 100 100 100 100 0 mfr 6.3 100 100 100 98.0 0 mgl 6.4 100 100 100 96.1 0 mgr 6.4 100 100 100 96.1 0 mkl 6.1 100 100 100 95.9 0 mkr 6.4 100 100 100 96.1 0 mll 6.4 100 100 100 96.1 0 mlr 6.4 100 100 100 100 0 mml 6.0 100 100 100 95.8 0 mmr 6.3 10 0 100 100 96.0 0 mol 6.5 100 100 100 100 0 mor 6.3 100 100 100 98.0 0

PAGE 91

91 Table 5 6 The highest tissue averaged doses received from the set of 32 eyes undergoing treatment simulation and the associated eye model ; computational error was les s than 1% for all values Tissue Beam I Beam II Beam III Total mGy mGy mGy mGy macula 8000 8000 8000 24000 lens 66 (mar) 57 (fol) 69 (fol) 176 (fol) optic nerve 747 (fkl) 276 (mol) 1100 (fnr) 1291 (fkl) Figure 5 1 ; (A) male (B) female B A

PAGE 92

92 Figure 5 2. DVH s for the extremes of male optic nerve tilt ; (A) sup med (B) sup lat (C) inf med (D) inf lat B D C A

PAGE 93

93 Figure 5 3. DVH s for the extremes of female optic nerve tilt ; (A) sup med (B) sup lat (C) inf m ed (D) inf lat B D C A

PAGE 94

94 Figure 5 4. Spatial contour map of the dose distribution within the reference eye model of the adult male ; legend units are Gy where 0 refers to some v alue approaching the limit of 0

PAGE 95

95 Figure 5 5. Dose con tour maps for patient model mer; i mage progression is from inferior (top left) to superior (b ottom right ) in 1 mm intervals, with the (c enter) median slice intersecting the midd le of the macula target; l egend units are Gy where 0 refers to some v alue approaching the limit of 0

PAGE 96

96 Figure 5 6. Dose con tour maps for pa tient model fkl; (A) sagittal image slice intersecting the middle of the macula target, (B) DVH, and (C) axial image slice intersect ing the middle of the macula target; legend units are Gy where 0 refers to some v alue approaching the limit of 0 C B A

PAGE 97

97 Figure 5 7. Dose con tour maps for patient model fjl; (A) sa gittal image slice intersecting the middle of the macula target, (B) DVH, and (C) axial image slice intersect ing the middle of the macula target; legend units are Gy where 0 refers to some v alue approaching the limit of 0 C B A

PAGE 98

98 Figure 5 8 Correlation scatter plots of mean absorbed dose to the optic nerve as a function of gaze angle ; (A) horizontal g aze (B) v ertical g aze A B

PAGE 99

99 Figure 5 9. Correlation scatter plot and linear regression of optic nerve hotspot dose as a function of optic nerve thickness Figure 5 10. Phase space diagram of energy and angular depe ndence of photon fluen ce 50 cm from the macula target; c osine equal to 1 is synonymous with primary gaze; a ll units are #/cm 2 /history

PAGE 100

100 Figure 5 11. Photon fluence distribution plots 50 cm from macula target ; (A) angular dependence and (B) energy depe ndence B A

PAGE 101

101 Figure 5 12. Photon fluence contour maps at the edge of the lattice structure ; (A) top (B) bottom (C) perspective (D) right side (E) front (F) back and (G) left side; a ll units are #/cm 2 /history C B A

PAGE 102

102 Figure 5 12. Continued G F E D

PAGE 103

103 CHAPTER 6 CONCLUSIONS 6.1 Limitations of T his Work The retrospectiv e collection of CT data presented several challenges for this project. Retrospective collectio n was warranted to limit dose to potential volunteers. MR imaging could have be en used but it would hav e been costly to setup a proactive study. Because the data are collected retrospectively, and not specifically for the purpose of this project, patients display a wide variety of gaze angles and head tilt that would not typically be seen during stereotact ic radiosurgery for wet AMD Measures were taken to account for this, but ideally patient specific model fabrication and treatment planning would have been performed on a patient population potentially undergoing SRS. It is difficult to obtain a large num ber of head CT images with fine slice resolution and facial structures intact. High resolution slices are desired for this project to minimize the uncertainty in the measurements taken as described in Chapter 2 but slice resolution is limited during admi nistration because of dose considerations to the patient. When higher slice resolution is needed for the diagnostic procedure it is typically because there is some head trauma to the patient To account for this tradeoff, a 1 mm slice resolution was chos en for data collection and partially due to the size of Shands Hospital at the University of Florida, enough image sets were eventually found at this resolution without significant trauma to the facial structures including the entire orbital region. W hile the selection of this slice resolution allows for improved measurement accuracy for orbital structures, it presents a shortcoming in data available for the entire

PAGE 104

104 cranium and brain. At 1 mm slice resolution, the top and back of the head are often lef t out of the image, again due to dose considerations to the patient. The lack of this data prevents the calculation of mean absorbed dose and formulation of dose volume histograms to these structures since their true total volume is unknown. Fortunately the area of interest where dose hotspots may form are included in the image sets and valuable information can still be tabulated concerning absorbed dose to localized portions of the anatomy Furthermore, the sample size is still too small to be highly co nfident that the patient specific variations in anatomy observed are an accurate representation of the total population. The image sets were stripped of all personal health information in accordance with our IRB protocol to protect the privacy of subjects whose images were collected. The sample size is further compromised considering all the images were gathered from one location (Gainesville, FL). While Shands Hospital at the University of Florida potentially attracts patients from throughout the state, it still limits the source of data to one geographical location. Voxel model geometry presents some disadvantages depending on the voxel resolution selected for Monte Carlo simulation. Higher voxel resolution s offer super ior geometry detail but slow comp uter runtime and increase d error in mesh tallies. The file size of high resolution voxel models present difficulties in the following scenarios : ( 1) i n house MATLAB codes crash from memory limitation when attempting to load binary voxel model files, ( 2) MCNPX h as trouble compiling when attempting to load lattice file s and ( 3) the MCNPX built in program GRIDCONV fails when attempting to convert mdat files from mesh tallies to ASCII files. To avoid these problems, appropriate voxel

PAGE 105

105 resolutions were select ed for the eye and head models: 0.5 mm 3 and 1 mm 3 voxel resolution, respectively. The selection of these resolutions allowed for successful completion of all MCNPX and MATLAB operation s but t hese resolutions restrict the dimensions that can be accurately modeled. Specifically, the fovea offset described in Chapter 1 cannot be modeled with the same level of accuracy as the I Ray TM targeting system. The 1.25 mm offset in the lateral direction will be approxi mated to 1 mm or 1.5 mm by the Voxelizer code. D espite this limitation, voxel geometry still offers numerous benefits over stylized geometry including the ability to model tissues with complex shape and preservatio n of patient specific anatomy. 6.2 General Conclusions A new treatment for wet AMD involvi ng kilovoltage stereotactic radiosurgery is proposed by Oraya Therapeutics, Inc. with the establish ment of the IRay TM The benefits of this treatment modality include non invasive application treatment time and frequency, and potentially efficacy (pendin g results of preliminary clinical trials). A major advancement in quantifying the position of the macula with the Oraya Shift was incorporated into the targeting system, and subsequently the modeling involved with this project. The scope of this researc h involved the dosimetry characterization of the treatment scheme using a variety of anthropometric models both reference and patient specific Reference whole body male and female phantoms were designed by members of the Advanced Laboratory for Radiatio n Dosimetry Studies (ALRADS) and were used for this work with th e improvement and addition of detail to the ocular anatomy. Details were derived from data in ICR P 89, NCRP 130, and measurements taken on 40 head CT scan s of equal gender distribution. The statistical analysis of the

PAGE 106

106 gender dependent and quanti fied the optic nerve pathways. Five reference optic nerve models were evaluated in each head phantom based on the ra nge of optic nerve exit tilts observed and to indirectly account for non primary gaze angles. To fully account for varying gaze and anatomy, several patient specific voxel phantoms were derived from segmentation of the CT data collected. In all, 16 patie nt and 32 eye models were evaluated. The reference models were voxelized to 0.5 mm 3 1 mm 3 and 2 mm 3 resolution f or the eye section, head and neck region and torso, respectively. Each of the patient specific eye models were voxelized to 0.5 mm 3 resoluti on. The voxelized versions of the phantoms were imported into MCNPX 2.5.0 Monte Carlo radiation transport code for simulation of ocular radiotherapy. The results provided insight into the mean absorbed dose received for several radiosensitive tissues at potential risk for a three beam treatment cumulating in 24 Gy delivery to the macula. Mean absorbed doses, dose volume histograms, and effective dose were evaluated for the reference phantom. DVHs were assessed for macula target and three non targeted t issues: lens, optic nerve, and brain. Cumulative m ean absorbed doses to the lens were found to be 124 mGy in the reference male and 127 mGy in the reference female for the 3 beam treatment. Integral mean absorbed doses to the optic nerve were 200 mGy an d 237 mGy in the reference male and female, respectively. The lens and optic nerve were of utmost importance and interest, and the absorbed doses received were below the generally accepted thresholds for cataracts and radiation induced optic neuropathy (R ON).

PAGE 107

107 The doses to the remainder of the tissues in the reference phantom were used to estimate an effective dose as per ICRP Publication 103 schema. The effective dose for the proposed stereotactic AMD radiotherapy, including contribution from both the p rimary tube output and leakage, is estimated to be 0.29 mSv which is a factor of ~10 2 to 10 3 lower than seen in external beam radiotherapy, a factor of ~10 lower than seen in CT imaging, and is comparable to that seen in radiographic imaging of the head an d neck. Considering the patient specific phantom series (n=32) t he dosimetry performed for kilovoltage stereotactic radiosurg ery treatment simulation show that tissues at risk do not receive tissue averaged doses over the generally accepted thresholds for complications, specifically the formation of cataracts and brain necrosis. Likewise, point doses delivered to the optic nerve were not significant in terms of the risk associated with developing RON. This study provided a worst case scenario risk assess ment by including a range of clinically unrealistic gaze angles, and correspondingly a diverse range of optic nerve positions. The eye models receiving the highest average or point doses were further analyzed using dose contour maps. Trends were observed for dose as a function of gaze angle in the horizontal and vertical directions, and dose escalation corresponded to increasing optic nerve thickness. Ultimately, considering the results of this work, the treatment scheme employed by the IRay TM device ha s the potential to deliver a therapeutic dose to the macula with minimal irradiation of non target tissues within a set limit of clinically realistic gaze angles. Furthermore, the doses reported in this study could be scaled proportionally for

PAGE 108

108 a cumulativ e therapeutic dose of 16 Gy to the macula tissue, the treatment scheme currently planned for US clinical trials. 6. 3 Future Work As with any work of complex nature, there always room for improvement and e xpanded investigations. The work of this research p rovided valuable data for the initial phases of the project, but device development is ongoing and Phase II clinical trials are currently being set up for the IRay TM To date, the computation al evaluation of the device has explored only variations in ana tomy and gaze and has not taken into account the uncertainty in the targeting system. A recent publication by Ger t ner et a l 26 claims that the uncertainly and precision of the machine are 6 00 and 4 00 micro ns, respectively. Therefore, a computation al sensitivity study could be designed could be transposed by 6 00 microns in each of the four major dire ctions; superior, inferior, lateral, and medial to the center of the macula target. Then, fixing the target point at the center of the macula, the treatment axis could be rotated in each of the four major directions in increments of one degree up to five degrees. An evaluation of non targeted dose could be explored much like in this present work, but also an exploration of the dose distribution to the macula. Furthermore, experimental verification in the clinic of the computations performed could be und ertaken. A real time dosimetry system coupled with physical anthropometric phantoms could not only validate the computational work, but also describe how dose is deposited as a function of treatment time.

PAGE 109

109 APPENDIX A EXAMPLE OF MCNPX INPUT CODE c R eference Adult male eye voxel model c Matrix size [90,119,93] c Voxel resolution=0.05*0.05*0.05 cm^3 c 10 22 08 c Justin Hanlon c The University of Florida read file=mmeanlat noecho 1001 0 100 fill=999 imp:p=1 $ surrounding box c ---------------------------c Body composition and density c ---------------------------1 1 1.03 70 u=1 imp:p=1 vol=40.269625 $residual soft tissue 4 7 1.04 70 u=4 imp:p=1 vol=17.91375 $Bra in 11 3 1.1 70 u=11 imp:p=1 vol=0.525376 $external nose 12 1 1.03 70 u=12 imp:p=1 vol=1.468625 $right Eye (soft tissue) 29 1 1.03 70 u=29 imp:p=1 vol=0.952125 $nasal layer (posterior) 57 5 0.001205 70 u=57 imp:p=1 vol=4.273875 $Air 62 6 1 70 u=62 imp:p=1 vol=5.732375 $right vitreous humor (water) 64 8 1.07 70 u=64 imp:p=1 vol=0.21125 $right lens 66 1 1.03 70 u=66 im p:p=1 vol=0.005875 $right macula 68 7 1.04 70 u=68 imp:p=1 vol=0.001125 $right optic disc 71 101 1.525 70 u=71 imp:p=1 vol=28.22775 $cranium 74 7 1.04 70 u=74 imp:p=1 vol=0.5045 $right optic nerve c ---------------------------c window and outside of the window c ---------------------------1002 5 0.001205 100 1000 #2000 imp:p=1 $Out of Voxel inside medium 1003 0 1000 imp: p=0 $Out of ROI 2000 9 19.3 2000 2010 2030 2020 imp:p=1 $aperture c ---------------------------c surface card s c ---------------------------c Matrix size [90,119,93] c Voxel resolution=0.05*0.05*0.05 100 rpp 2.1 2.4 2.85 3.1 2.25 2.4 $origin at center of macula 200 rpp 0 0.05 0 0.05 0 0.05 $0.05 for voxel size 1000 so 200 70 so 200 2000 11 cy 5 2010 11 cy 0.1175 2020 11 py 7.45 2030 11 py 7.7 mode p C Material Cards c Defined using ICRP 4 6 tissue compositions C Soft tissue (male) (rho=1.03) m1 1000 0.105 6000 0.256 7000 0.027 8000 0.602

PAGE 110

110 11000 0.001 15000 0.002 16000 0.003 1700 0 0.002 19000 0.002 c rest of the material cards omitted for space c -------------------------------------------------------------------c tally c -------------------------------------------------------------------fc6 right lens *f6:p 64 $jerks/g fc16 right macula *f16:p 66 fc26 right optic disc *f26:p 68 fc36 right optic nerve *f36:p 74 c DVH mesh tally for right lens, macula, optic nerve, brain tmesh c Matrix size [90,119,93] c rpp 2.1 2.4 2.85 3.1 2.25 2.4 rmesh1:p pedep cora1 2.1 89i 2.4 corb1 2.85 118i 3.1 corc1 2.25 92i 2.4 endmd c -------------------------------------------------------------------c beam description (180 degree) c -------------------------------------------------------------------sdef par=2 x=d4 y= 15 z=d5 erg=d3 dir=d1 vec=0 1 0 tr=11 # si3 sp3 0.0005 0.00000E+00 c full energy spectra omitted for space 0.1 3.04454E+01 si4 0.05 0.05 sp4 0 1 si5 0.05 0.05 sp5 0 1 si1 h 1 0.999644633962 1 sp1 d 0 0.999822316981 0.000177683019 sb1 d 0 0 10 c *tr11 0.08 0.05 0.08 30 90 60 104.4775 30 64.3411 115.6589 120 41.4096 $ beam 5 o'clock *tr 11 0.08 0.05 0.08 0 90 90 90 30 60 90 120 30 $ beam 6 o'clock c *tr11 0.08 0.05 0.08 30 90 120 75.5225 30 64.3411 64.3411 120 41.4096 $ beam 7 o'clock nps 1e7

PAGE 111

111 APPENDIX B SAMPLES OF MATLAB CO DES % import eye phantom binary file % mesh files (result of mdat files after GRIDCONV) must be in same directory % and must be named mesh1, mesh2, and mesh3 clear all ; name=input( 'Cropped model name?' 's' ); %mesh file adjustment beam=5:7; for j=1:3; filename=[ 'mesh' ,num2str(j)]; fid=fopen(filename); c=textscan(fid, '%s' 'delimiter' \ n' ); fclose(fid); d=length(c{1}); e=(length(c{1}) 10)./2; %mesh w/o error generation f=e+10; for i=11:f g{1}{i 10}=c{1}{i}; end filename=[ 'mesh' ,nu m2str(beam(j))]; fid=fopen(filename, 'w' ); for i=1:e fprintf(fid, '%s \ n' ,g{1}{i}); end fclose(fid); clear f g ; %error mesh generation f=e+11; for i=f:d; g{1}{i f+1}=c{1}{i}; end filename=[ 'error' ,n um2str(beam(j))]; fid=fopen(filename, 'w' ); for i=1:e fprintf(fid, '%s \ n' ,g{1}{i}); end fclose(fid); clear c d e f g i j filename ; end %phantom matrix generation nameid1=str2num(char(name(5:6))); nameid2=str2num(char(name(8:10)) ); nameid3=str2num(char(name(12:13))); fid = fopen(name); phantom=reshape(fread(fid, 'ubit8' ),nameid1,nameid2,nameid3); fclose(fid); clear nameid1 nameid2 nameid3 ;

PAGE 112

112 phantom(find(phantom==0))=57; s=size(phantom); % assign organ tag and density organ_density =[ 1 1.03 % residual soft tissue 2 1.04 % Brain 3 1.525 % skull 4 1.04 % optic nerve 5 1 % vitreous humor 6 1.07 % lens 7 1.03 % macula 57 0.001205]; % air % assign organ name organ_name=cellstr(char( ... 'Residual soft tissue ... 'brain ... 'skull ... 'optic nerve ... 'vitreous humor ... 'lens ... 'macula ... 'air )); % compose density matrix to be multiplied to flux matrix from mesh tally density_matrix=zeros(s(1),s(2),s(3)); for x=1:s(1); for y= 1:s(2); for z=1:s(3); density_matrix(x,y,z)=organ_density(find(organ_density(:,1)==phantom(x,y,z)), 2); end end end % read mesh tally for 5, 6, and 7 o'clock beams and sum them up total_meshtally=zeros(s(1),s(2),s(3)); dose_werror=zero s(s(1),s(2),s(3)); target_dose=0; for beam=5:7; % 5,6, and 7 o'clock beam direction target_dose=target_dose+8; filename=[ 'mesh' ,num2str(beam)]; temp=reshape(load(filename),s(2),s(3),s(1)); meshtally=zeros(s(1),s(2),s(3)); filename2=[ 'error' ,num2str(b eam)]; temp2=reshape(load(filename2),s(2),s(3),s(1)); errormatrix=zeros(s(1),s(2),s(3)); for x=1:s(1); for y=1:s(2); for z=1:s(3); meshtally(x,y,z)=temp(y,z,x); errormatrix(x,y,z)=temp2(y,z,x); % incorporate comp utational error if phantom(x,y,z)==7;

PAGE 113

113 dose_werror(x,y,z)=meshtally(x,y,z) meshtally(x,y,z).*errormatrix(x,y,z); else dose_werror(x,y,z)=meshtally(x,y,z)+meshtally(x,y,z).*errormatrix(x,y,z); end end end end % unit conversion % MeV/cm3 / density(g/cm3) 1.6e 13 1e3 = Gy/particle dose_werror=dose_werror./density_matrix.*1.6e 13.*1e3; dose_werror(find(phantom==57))=0; total_meshtally=total_meshtally+dose_werror; end % DVH pl otting -------------------------------------------------------% modify dvh_organ array for organ tags you're interested in % currently the number of organs for DVH is limited to 5 which is enough graph_color=cellstr(char( k' .b' -r' ':k' -b' )); dv h_organ=[7 6 4 2]; maximum_macula_dose=max(total_meshtally(find(phantom==7))); hold off for i=1:size(dvh_organ,2) legend_title(i)=organ_name(find(organ_density(:,1)==dvh_organ(i))); maximum_dose=max(total_meshtally(find(phantom==dvh_organ(i)))); dose=[0:maximum_dose/50:maximum_dose]; dvh_temp=hist(total_meshtally(find(phantom==dvh_organ(i))),dose); dvh_temp(51)=0; for j=size(dose,2): 1:2; dvh_temp(j 1)=dvh_temp(j 1)+dvh_temp(j); end plot(dose/maximum_macula_dose*target_do se,dvh_temp./dvh_temp(1)*100,char(grap h_color(i)), 'LineWidth' ,2.5); hold on end titlename=[ 'Patient: ,char(name(1:3))]; title(titlename, 'fontsize' ,14); xlabel( 'Absorbed dose (Gy)' 'fontsize' ,14); ylabel( 'Volume (%)' 'fontsize' ,14); legend(char(legend_titl e),2); %dose distribution code %scales doses to give max macula dose 24Gy maximum_macula_dose=max(total_meshtally(find(phantom==7))); sourceparticlesneeded=24/maximum_macula_dose; dosematrix=total_meshtally.*sourceparticlesneeded; braindose=zeros(1,5); skulldose=zeros(1,5); ONdose=zeros(1,5); lensdose=zeros(1,5); macdose=zeros(1,5); heading1=[0.5,1,2,5,10,1000]; heading2=[0.1,0.2,0.3,0.4,0.5,1000];

PAGE 114

114 heading3=[5,10,15,20,25,1000]; heading4=[0.5,1,5,12,15,1000]; heading5=[25,40,45,47.5,50,1000]; for i=1: 5 brain=0; skull=0; ON=0; lens=0; mac=0; for x=1:s(1); for y=1:s(2); for z=1:s(3); if phantom(x,y,z)==4; ON=ON+1; if dosematrix(x,y,z)>=heading4(i); ONdose(1,i)=ONdose(1,i)+1; end end if phantom(x,y,z)==2; brain=brain+1; if dosematrix(x,y,z)>=heading1(i); braindose(1,i)=b raindose(1,i)+1; end end if phantom(x,y,z)==6; lens=lens+1; if dosematrix(x,y,z)>=heading2(i); lensdose(1,i)=lensdose(1,i)+1; end end if phantom(x,y,z)==7; mac=mac+1; if dosematrix(x,y,z)>=heading3(i); macdose(1,i)=macdose(1,i)+1; end end if phantom(x,y,z)==3; skull=skull+1; if dosematrix(x,y,z)>=heading5(i); skulldose(1,i)=skulldose(1,i)+1; end end end end end end %converts #voxels to mm^3 brain=brain.*0.125; skull=skull.*0.125; ON=ON.*0.125; lens=lens.*0.125; mac=mac.*0.125; braindose=braindose.*0.125; skulldose=skulldose.*0.125;

PAGE 115

115 ONdose=ONdose.*0.125; lensdose=lensdose.*0.125; macdose=macdose.*0.125; %w rites to excel menu={ 'OAR' 'Total Vol' 'Threshold Vol' }; xlswrite( 'DVH_Table.xls' ,menu(1,:),1, 'a1:c1' ); xlswrite( 'DVH_Table.xls' 'O' ,1, 'a7' ); xlswrite( 'DVH_Table.xls' ,ON,1, 'b7' ); xlswrite( 'DVH_Table.xls' ,ONdose(1,:),1, 'c7:g7' ); xlswrite( 'DVH_Table.xls' B' ,1, 'a5' ); xlswrite( 'DVH_Table.xls' ,brain,1, 'b5' ); xlswrite( 'DVH_Table.xls' ,braindose(1,:),1, 'c5:g5' ); xlswrite( 'DVH_Table.xls' 'L' ,1, 'a3' ); xlswrite( 'DVH_Table.xls' ,lens,1, 'b3' ); xlswrite( 'DVH_Table.xls' ,lensdose(1,:),1, 'c3:g3' ); xlswrite( 'DVH_Table. xls' 'M' ,1, 'a11' ); xlswrite( 'DVH_Table.xls' ,mac,1, 'b11' ); xlswrite( 'DVH_Table.xls' ,macdose(1,:),1, 'c11:g11' ); xlswrite( 'DVH_Table.xls' 'S' ,1, 'a9' ); xlswrite( 'DVH_Table.xls' ,skull,1, 'b9' ); xlswrite( 'DVH_Table.xls' ,skulldose(1,:),1, 'c9:g9' ); % Dose map c ode % Must find the slice desired and input into dosematrix(x,:,:) part % Must enter the other 2 dimensions of your matrix in the reshape part % % -----------------------------------------------------% % Axial % % ------------------------------------------------------% two_d_matrix=dosematrix(:,:,23); % two_d_matrix=fliplr(two_d_matrix); % ------------------------------------------------------% Sagittal % -----------------------------------------------------% two_d_matrix=dosematrix(26,:,:); % two_d_matrix=reshape(two_d_matrix,114,61);

PAGE 116

116 LIST OF REFERENCES 1. H. Leibowitz, D. E. Krueger and L. R. Maunder, "The Framing ham Eye Study Monograph: an ophthalmological and epidemiological study of cataract, glaucoma, diabetic retinopathy, macular degeneration, and visual acuity in a general population of 2631 adults," Surv Ophthalmol 24 335 610 (1980). 2. R. P. Murphy, "Age related macular degeneration," Ophthalmology 93 969 971 (1986). 3. N. M. Bressler, S. B. Bressler and S. L. Fine, "Age related macular degeneration," Surv Ophthalmol 32 375 413 (1988). 4. S. Haddad, C. A. Chen, S. L. Santangelo and J. M. Seddon, "The g enetics of age related macular degeneration: a review of progress to date," Surv Ophthalmol 51 316 363 (2006). 5. J. Gass, Stereoscopic Atlas of Macular Disease and Treatment (CV Mosby co, St. Louis, 1985). 6. L. A. Donoso, T. Vrabec and H. Kuivaniemi, "The role of complement Factor H in age related macular degeneration: a review," Surv Ophthalmol 55 227 246 (2010). 7. A. DeWan, L. Mugen, S. Hartman, S. S. Zhang, D. Liu, C. Zhao, P. Tam, W. M. Chan, D. Lam, M. Snyder, C. Barnstable, C. P. Pang and J. Hoh, "HTRA1 promoter polymorphism in wet age related macular degeneration," Science 314 989 992 (2006). 8. R. S. Snell and M. A. Lemp, Clinical anatomy of the eye (Blackwell Science, Malden, MA, 1998). 9. R. W. Young, "Pathophysiology of age related ma cular degeneration," Surv Ophthalmol 31 (1987). 10. J. Gass, "Drusen and disciform macular detachment and degeneration," Arch Ophthalmol 90 206 217 (1973). 11. Macular Photocoagulation Study Group, "Visual outcome after laser photocoagulation for subfov eal choroidal neovascularization secondary to age related macular degeneration. The influence of initial lesion size and initial visual acuity.," Arch Ophthalmol 112 480 488 (1994). 12. Verteporfin in Photodynamic Therapy Study Group, "Photodynamic ther apy of subfoveal choroidal neovascularization in pathologic myopia with verteporfin. 1 year results of a randomized clinical trial -VIP report no. 1," Ophthalmology 108 841 852 (2001).

PAGE 117

117 13. P. J. Rosenfeld, D. M. Brown, J. S. Heier, D. S. Boyer, P. K. Kai ser, C. Y. Chung and R. Y. Kim, "Ranibizumab for neovascular age related macular degeneration," N Engl J Med 355 1419 1431 (2006). 14. P. A. Quiram, K. A. Drenser, M. M. Lai, A. Capone and M. T. Trese, "Treatment of vascularly active familial exudative v itreoretinopathy with pegaptanib sodium (Macugen)," Retina 28 S8 S12 (2008). 15. M. P. Avila, M. E. Farah, A. Santos, J. P. Duprat, B. W. Woodward and J. Nau, "Twelve month short term safety and visual acuity results from a multicentre prospective study of epiretinal strontium 90 brachytherapy with bevacizumab for the treatment of subfoveal choroidal neovascularisation secondary to age related macular degeneration," Br J Ophthalmol 93 305 309 (2009). 16. M. P. Avila, M. E. Farah, A. Santos, Z. Kapran, J P. Duprat, B. W. Woodward and J. Nau, "Twelve month safety and visual acuity results from a feasibility study of intraocular, epiretinal radiation therapy for the treatment of subfoveal CNV secondary to AMD," Retina 29 157 169 (2009). 17. H. Churei, K. Ohkubo, M. Nakajo, H. Hokotate, Y. Baba, J. Ideue, K. Miyagawa, H. Nakayama, Y. Hiraki, T. Kitasato and N. Yabe, "External beam radiation therapy for age related macular degeneration: two years' follow up results at a total dose of 20 Gy in 10 fractions," Radiat Med 22 398 404 (2004). 18. H. J. Zambarakji, A. M. Lana, E. Ezra, D. Gauthier, M. Goitein, J. A. Adams, J. E. Munzenrider, J. W. Miller and E. S. Gragoudas, "Proton beam irradiation for neovascular age related macular degneration," Ophthalmology 113 2012 2019 (2006). 19. A. Haas, G. Papaefthymiou, G. Langmann, O. Schrottner, B. Feigl, K. A. Leber, R. Hanselmayer and G. Pendl, "Gamma knife treatment of subfoveal, classic neovascularization in age related macular degeneration: a pilot study," J Ne urosurg 93 172 176 (2000). 20. M. Hayashi, M. Chernov, M. Usukura, K. Abe, Y. Ono, M. Izawa, S. Hori, T. Hori and K. Takakura, "Gamma knife surgery for choroidal neovascularization in age related macular degeneration. Technical note.," J Neurosurg 102 2 00 203 (2005). 21. M. A. Henderson, S. Valluri, S. S. Lo, T. C. Witt, R. M. Worth, R. P. Danis and R. D. Timmerman, "Gamma knife radiosurgery in the treatment of choroidal neovascularization (wet type macular degeneration)," Stereotact Funct Neurosurg 85 11 17 (2007). 22. S. V. Goverdhan, F. A. Gibbs and A. J. Lotery, "Radiotherapy for age related macular degeneration: no more pilot studies please.," Eye 19 1137 1141 (2005).

PAGE 118

118 23. G. J. Bergink, C. B. Hoyng, R W van der Maazen, J. R. Vingerling, W A van D aal and A. F. Deutman, "A randomized controlled clinical trial on the efficacy of radiation therapy in the control of subfoveal choroidal neovascularization in age related macular degeneration: radiation versus observation," Graefes Arch Clin Exp Ophthalmo l 236 321 325 (1998). 24. D. H. Char, A. I. Irvine, M. D. Posner, J. Quivey, T. L. Phillips and S. Kroll, "Randomized trial of radiation for age related macular degeneration," Am J Ophthalmol 127 574 578 (1999). 25. R. P. Singh, D. Moshfeghi, E. M. Shu sterman, S. A. McDormick and M. Gertner, "Evaluation of transconjunctival collimated external beam radiation for age related macular degeneration (ARMD)," AAO Abstract 08 PP 30018701 AAO (2008). 26. M. Gertner, E. Chell, K. H. Pan, S. Hansen, P. K. Kaiser and D. M. Moshfeghi, "Stereotactic targeting and dose verification for age related macular degeneration," Med Phys 37 600 606 (2010). 27. M. E. Arnoldussen, M. Shusterman, D. Fletcher, L. Renninger, L. Dang, I. Koruga, M. Firpo, J. Liang and M. Gertner, "Quantitative Measurements of Retinal Structures Relative to the Geometric Axis of the Eye," ARVO E Abstract 3789 50 (2009). 28. C. Lee, E. Chell, M. Gertner, S. Hansen, R. W. Howell, J. Hanlon and W. E. Bolch, "Dosimetry characterization of a multibeam radiotherapy treatment for age related macular degeneration," Med Phys 35 5151 5160 (2008). 29. ICRP, "ICRP Publication 89: Basic anatomical and physiological data for use in radiological protection reference values," Ann ICRP 32 1 277 (2002). 30. NC RP, "Biological effects and exposure limits for hot particles," National Council on Radiation Protection and Measurements Report No. 130 (1999). 31. M. W. Charles and N. Brown, "Dimensions of the human eye relevant to radiation protection," Phys Med Biol 20 202 218 (1975). 32. B. V. Worgul, The Edward S. Harkness Eye Institute Resident's Basic Science Study Guide (Columbia University, New York, NY, 1991). 33. R. Unsold, J. DeGroot and T. H. Newton, "Images of the optic nerve: anatomic CT correlation," AJR Am J Roentgenol 135 767 773 (1980). 34. J. A. Rogers, A. G. Podoleanu, G. M. Dobre, D. A. Jackson and F. W. Fitzke, "Topography and volume measurements of the optic nerve using en face optical coherence tomography," Optics Express 9 533 545 (2001).

PAGE 119

119 35. T. Krzizok and B. Schroeder, "Quantification of recti eye muscle paths in high myopia," Strabismus 11 213 220 (2003). 36. J. Hanlon, C. Lee, E. Chell, M. Gertner, S. Hansen, R. W. Howell and W. E. Bolch, "Kilovoltage stereotactic radiosurgery for age related macular degeneration: assessment of optic nerve dose and patient effective dose," Med Phys 36 3671 3681 (2009). 37. D. Scott, "On optimal and data based histograms," Biometrika 66 605 610 (1979). 38. T. Anthoney, Neuroanatomy and the Neurologi c Exam (CRC Press, New York, NY, 1994). 39. C. Lee, C. Lee, D. Lodwick and W. E. Bolch, "NURBS based 3 D anthropomorphic computational phantoms for radiation dosimetry applications," Radiat Prot Dosimetry 127 227 232 (2007). 40. W. S. Snyder, "Estimate s of absorbed fractions for monoenergetic photon sources uniformly distributed in various organs of a heterogeneous phantom," Society of Nuclear Medicine MIRD Pamphlet No. 5 (1969). 41. ICRP, "ICRP Publication 23: Report on the Task Group on Reference Man ," Ann ICRP (1975). 42. M. Cristy and K. F. Eckerman, "Specific absorbed fractions of energy at various ages from internal photon sources," Oak Ridge National Laboratory Report No. ORNL/TM 8381/Volumes I VII (1987). 43. C. Lee, D. Lodwick, D. Hasenauer, J. L. Williams, C. Lee and W. E. Bolch, "Hybrid computational phantoms of the male and female newborn patient: NURBS based whole body models," Phys Med Biol 52 3309 3333 (2007). 44. ICRU, "Photon, electron, proton and neutron interaction data for body ti ssues," International Commission on Radiation Units and Measurements Report No. 46 (1992). 45. D. B. Pelowitz, "MCNPX User's Manual Version 2.5.0," (Los Alamos National Laboratory, Los Alamos, NM, 2005). 46. K. Cranley, B. J. Gilmore, G. W. A. Fogarty a nd L. Desponds, "Catalogue of diagnostic x ray spectra and other data," The Institute of Physics Report No. 78 (1997). 47. ICRP, "ICRP Publication 103: Recommendations of the International Commission on Radiological Protection," Ann ICRP 37 1 332 (2007).

PAGE 120

120 48. M. Zankl, K. F. Eckerman and W. E. Bolch, "Voxel based models representing the male and female ICRP reference adult -the skeleton," Radiat Prot Dosimetry 127 174 186 (2007). 49. B. V. Worgul, Y. I. Kundiyev, N. M. Sergiyenko, V. V. Chumak, P. M. Vit te, C. Medvedovsky, E. V. Bakhanova, A. K. Junk, O. Y. Kyrychenko, N. V. Musijachenko, S. A. Shylo, O. P. Vitte, S. Xu, X. Xue and R. E. Shore, "Cataracts among Chernobyl clean up workers: implications regarding permissible eye exporsures," Radiat Res 167 233 243 (2007). 50. C. A. Girkin, C. H. Comey, L. D. Lunsford, M. L. Goodman and L. B. Kline, "Radiation optic neuropathy after stereotactic radiosurgery," Ophthalmology 104 1634 1643 (1997). 51. S. L. Stafford, B. E. Pollock, J. A. Leavitt, R. L. Foo te, P. D. Brown, M. J. Link, D. A. Gorman and P. J. Schomberg, "A study on the radiation tolerance of the optic nerves and chiasm after stereotactic radiosurgery," Int J Radiat Oncol Biol Phys 55 1177 1181 (2003). 52. T. Hasegawa, T. Kobayashi and Y. Kid a, "Tolerance of the optic apparatus in single fraction irradiation using stereotactic radiosurgery: evaluation in 100 patients with craniopharyngioma," Neurosurgery 66 688 694 (2010). 53. Y. R. Lawrence, X. A. Li, I. el Naga, C. A. Hahn, L. B. Marks, T. E. Merchant and A. P. Dicker, "Radiation dose volume effects in the brain," Int J Radiat Oncol Biol Phys 76 S20 S27 (2010). 54. F. A. Mettler, W. Huda, T. T. Yoshizumi and M. Mahesh, "Effective doses in radiology and diagnostic nuclear medicine: a catal og," Radiology 248 254 263 (2008). 55. NCRP, "Ionizing Radiation Exposure of the Population of the United States," National Council on Radiation Protection and Measurements Report No. 160 (2009).

PAGE 121

121 BIOGRAPHICAL SKETCH Justin Hanlon was born in 1985 in Nashua, New Hampshire. The older of two children, he grew up in Auburn, New Hampshire, and graduated from Pinkerton Academy in 2003. He earned his B.S., a dual degree in n uclear e ngineering and e ngineering p hysics at Rensselaer Polytechnic Institute, in 2007. In 2007, he was granted admission to the Ph.D. program within the Nuclear and Radiological Engineering Department at the University of Florida to perform research as a graduate assistant in computational medical physics. Since September 2007 he has worked closely with Oraya Therapeutics Incorporated, a medical device development company based in Newark, California, which funded his research. He completed his Ph.D. degree at the University of Florida in August 2010.