Computational Modeling of Eye Trauma for Different Orbit Anthropometries and Different Projectile

Impacts

By

Ashley Anne Weaver

A Thesis Submitted to the Graduate Faculty of

VIRGINIA TECH – WAKE FOREST UNIVERSITY

SCHOOL OF BIOMEDICAL ENGINEERING & SCIENCES

In Partial Fulfillment of the Requirements

for the Degree of

MASTER OF SCIENCE

Biomedical Engineering

May 2010

Winston-Salem, North Carolina

Approved by: Joel D. Stitzel, PhD, Advisor, Chair _____________________________ Examining Committee: Stefan M. Duma, PhD _____________________________ Warren N. Hardy, PhD _____________________________

ii

Acknowledgements First and foremost, I would like to thank my advisor, Dr. Joel Stitzel, whose

creativity, vision, leadership, and dedication are an inspiration to me. Thank you for your

guidance and encouragement on this thesis and other projects I’ve worked on. I truly

appreciate the numerous opportunities you have provided me with in graduate school thus

far, and all your time and effort that is largely responsible for any success I have

achieved.

Thank you to the remainder of my committee and to the Center for Injury

Biomechanics lab. Special thanks to Kathryn Loftis who I collaborated with on much of

the work outlined in this thesis. Thanks also to Josh Tan for help with TeraRecon and

Mao Yu for programming assistance.

To my husband, Neel, whose southern charm is a breath of fresh air when I come

home from the “big city.” Thank you for keeping me balanced and for making me smile

and laugh on a daily basis. Your love and support are truly a blessing. Thanks for all

your help around the house and for taking care of the “kids” (Kit-Kat and Woody).

Finally, I need to thank my family for their unconditional love and support

throughout my life. I could not have asked for better parents and without them I would

not be the person I am today. To my dad, whose strong work ethic and selflessness I

admire, thank you for instilling in me the value of education and for always letting me

know you are proud of my accomplishments. Thanks to my mom for all the phone calls

and “notes from home” that always remind me how much you care. Thanks also to

Lindsay, Pat, Steve, Brooke, and all other family and friends that have supported me over

the years.

iii

Table of Contents Acknowledgements............................................................................................................. ii

Table of Contents............................................................................................................... iii

List of Tables ...................................................................................................................... v

List of Figures .................................................................................................................... vi

Abstract ............................................................................................................................ viii

Chapter 1: Introduction ....................................................................................................... 1

Eye Anatomy .................................................................................................................. 1

Eye Injury ....................................................................................................................... 3

Research Goals ............................................................................................................... 5

References....................................................................................................................... 7

Chapter 2: CT Based Three-Dimensional Measurement of Orbit and Eye Anthropometry

........................................................................................................................................... 10

Abstract......................................................................................................................... 11

Introduction................................................................................................................... 12

Methods ........................................................................................................................ 14

Results ........................................................................................................................... 19

Discussion..................................................................................................................... 23

References..................................................................................................................... 26

Chapter 3: Biomechanical Modeling of Eye Trauma for Different Orbit Anthropometries

........................................................................................................................................... 28

Abstract......................................................................................................................... 29

Introduction................................................................................................................... 30

Methods ........................................................................................................................ 32

iv

Test Matrix Development ......................................................................................... 32

Finite Element Models .............................................................................................. 35

Simulations ............................................................................................................... 40

Statistical Analysis.................................................................................................... 41

Results ........................................................................................................................... 42

Discussion..................................................................................................................... 47

Acknowledgements....................................................................................................... 52

References..................................................................................................................... 53

Chapter 4: Evaluation of Different Projectiles in Matched Experimental Eye Impact

Simulations ....................................................................................................................... 57

Abstract......................................................................................................................... 58

Introduction................................................................................................................... 59

Methods ........................................................................................................................ 61

Results ........................................................................................................................... 65

Discussion..................................................................................................................... 74

Acknowledgements....................................................................................................... 77

References..................................................................................................................... 78

Appendix....................................................................................................................... 81

Summary of Research ....................................................................................................... 83

Scholastic Vita .................................................................................................................. 84

v

List of Tables Table 1. Demographics and ocular and orbital measurements. ........................................ 20

Table 2. Pearson product-moment correlation coefficients and p-values......................... 22

Table 3. Pattern of variation in ocular and orbital measurements between genders......... 23

Table 4. Test matrix defining eye and orbit parameters. .................................................. 34

Table 5. Magnitude and timing of stress, strains, contact forces, and pressure. Initial component contacted by ball: orbital brow (B) or cornea (C). Bolded stress and pressure values exceeded the globe rupture criteria........................................................................ 43

Table 6. Stress and pressure averages for anthropometric extremes. ............................... 46

Table 7. Multiple regression results: t ratio and p-value of each factor’s contribution to the regression model describing the response, and adjusted R-squared values describing response variation accounted for by the regression model. .............................................. 47

Table 8. Summary of simulations test matrix. .................................................................. 62

Table 9. Results of Student’s t-test comparing computational simulations grouped by experimental globe rupture. .............................................................................................. 66

Table 10. Results of Student’s t-test comparing projectile shape..................................... 66

Table 11. Pearson product-moment correlation coefficients and p-values relating stress and pressure to projectile variables................................................................................... 72

Table 12. Parameter estimates for predictor variables in the multiple regression model. 73

Table 13. Simulation results ............................................................................................. 81

Table 14. Publication plan for research outlined in this thesis. ........................................ 83

vi

List of Figures Figure 1. Anatomical structures of the eye. ........................................................................ 2

Figure 2. Ocular injuries as a percentage of all combat injuries......................................... 4

Figure 3. Alignment with the crista galli (a), falx cerebri (b), and nasion-sella turcica plane (c). Aligned 3D image used to measure orbital aperture (d). ................................. 16

Figure 4. a) Lateral eye protrusion (LP) and lateral distance (LD). b) Superior eye protrusion (SP) and superior distance (SD). ..................................................................... 18

Figure 5. Brow protrusion angle and medial (M), lateral (L), superior (S), and inferior (I) orbital rim coordinate points............................................................................................. 19

Figure 6. a) Orbital width (OW), orbital height (OH), superior distance (SD), and lateral distance (LD). b) Brow protrusion angle (BP) and eye protrusion (EP). ........................ 33

Figure 7. Variation in measured and modeled anthropometries. ...................................... 35

Figure 8. Lagrangian mesh depicting corneoscleral shell, lens, zonules, and ciliary body and Eulerian mesh depicting initially filled volume (dark gray) and initially unfilled volume (light gray). .......................................................................................................... 36

Figure 9. Small and large orbital apertures with incorporated eye model. Left photo: Model 13 (Orbital width and height: 32.48, 26.82 mm); Right photo: Model 15 (Orbital width and height: 40.04, 36.40 mm). ................................................................................ 39

Figure 10. Orbits with less and more brow protrusion with incorporated eye model. Left photo: Model 11 (Brow protrusion: 19.75 deg); Right photo: Model 17 (Brow protrusion: 34.35 deg). ........................................................................................................................ 39

Figure 11. Models with less and more eye protrusion. Left photo: Model 5 (Eye protrusion: 5.89 mm); Right photo: Model 14 (Eye protrusion: 12.06 mm). ................... 39

Figure 12. Distribution of peak corneoscleral stresses in 27 simulations......................... 43

Figure 13. Distribution of peak pressures observed in 27 simulations ............................. 43

Figure 14. Select simulations depicting effect of anthropometric variation on eye response............................................................................................................................. 44

Figure 15. Peak corneoscleral stress and strain element locations. .................................. 45

Figure 16. Geometry for sphere and cylinder projectiles. ................................................ 63

Figure 17. Projectile mass and diameter comparison. ...................................................... 63

vii

Figure 18. Gelatin and orbit surrounding lagrangian-eulerian meshes of the eye model. Lagrangian eye mesh (on left) depicting corneoscleral shell, lens, zonules, and ciliary body and Eulerian eye mesh (on right) depicting initially filled volume (dark gray) and initially unfilled volume (light gray)................................................................................. 64

Figure 19. Location of peak stresses in corneoscleral shell for different projectiles........ 67

Figure 20. Principal stress distribution, MPa, at time of peak stress for eight different projectiles. Note the different fringe levels for upper four projectiles versus the lower four projectiles.......................................................................................................................... 68

Figure 21. Maximum principal stress versus normalized velocity for different projectiles............................................................................................................................................ 69

Figure 22. Maximum principal stress versus kinetic energy (logarithmic scale) for different projectiles. .......................................................................................................... 70

Figure 23. Maximum principal stress versus maximum normalized kinetic energy. ....... 70

Figure 24. Maximum principal stress versus area-normalized kinetic energy (logarithmic scale) for different projectiles. .......................................................................................... 71

Figure 25. Maximum principal stress actual versus predicted plot with 95% confidence curves for multiple regression model with area-normalized energy and relative size predictor variables............................................................................................................. 73

viii

Abstract Over 1.9 million eye injuries occur each year in the United States, resulting in

30,000 cases of blindness annually. Common causes of eye trauma include motor vehicle

crashes, military operations, and ocular impacts with sporting equipment. Research

assessing eye injury risk with human variation and for diverse impact scenarios is

important to the advancement of eye injury prevention and mitigation. For this thesis, a

computational eye model was used to determine the biomechanical response of the eye

for various impacts.

Orbit anthropometric variation within the normal population was measured using

CT images, and age and gender effects were evaluated. Measurements provided three-

dimensional information on orbit geometry and were used to develop orbit models of

varying anthropometries. Baseball impact simulations were conducted to assess the

effect of orbit anthropometry on eye injury metrics.

Simulations with eight different projectiles (airsoft pellet, baseball, BB, blunt

impactor, paintball, aluminum, foam, and plastic rods) were run to characterize effects of

the projectile size, mass, geometry, material properties, and velocity on eye model

response. This study presents a matched comparison of experimental test results and

computational model outputs including stress, energy, and pressure that can be used to

evaluate risk of eye injury.

1

Chapter 1: Introduction This chapter provides an overview of eye anatomy, eye injury, and research goals

of this thesis. Subsequent chapters of this thesis are planned or submitted manuscripts

detailing separate eye research efforts.

EYE ANATOMY

The eye is the organ responsible for the sense of sight. The human eye has an

approximately spherical shape with a diameter of 25 mm and is composed of various

external and internal tissues (Figure 1). The fibrous exterior of the eye, termed the

corneoscleral shell, is comprised of two structures: the cornea and the sclera. The cornea

is a transparent structure on the anterior surface of the eye that is the primary refractory

medium of the eye. A thin transition region, termed the limbus, joins the cornea to the

sclera. The sclera is a tough, opaque, white structure that encompasses the remainder of

the globe surface1.

Internal structures of the eye can be divided into two segments: the anterior and

posterior segments. The anterior segment consists of the anterior and posterior chambers

of the eye which contain the aqueous humor, iris, lens, zonules, and ciliary body. The

anterior chamber is the space between the cornea and iris, while the posterior chamber is

defined anteriorly by the iris and posteriorly by the lens and ciliary body. Aqueous

humor is the volume of fluid present in both the anterior and posterior chambers. The

ciliary body functions to support the lens through the zonules and provide the mechanism

for focusing the lens. The iris, located just anterior to the lens, contracts and relaxes to

control the amount of light transmitted through the pupil1.

2

The posterior segment of the eye contains the vitreous body, choroid, retina, and

optic nerve. The vitreous body contains a watery fluid termed the vitreous humor that

transmits light and supports the retina and lens. The choroid is the layer between the

sclera and retina that serves as the vascular supply to the retina. The retina is the neural

layer of the eye that is sensitive to light. Visible light is refracted while passing through

the cornea, aqueous humor, lens, and vitreous humor. Light reaching the retina is

converted to electrical signals that are transmitted to the brain via the optic nerve

enabling visual perception1.

Vitreousbody

Cornea

Sclera

AnteriorChamber

Iris

Zonules

Lens

CiliaryBody

Limbus

Retina

Optic nerve

PosteriorChamber Vitreous

body

Cornea

Sclera

AnteriorChamber

Iris

Zonules

Lens

CiliaryBody

Limbus

Retina

Optic nerve

PosteriorChamber

Cornea

Sclera

AnteriorChamber

Iris

Zonules

Lens

CiliaryBody

Limbus

Retina

Optic nerve

PosteriorChamber

Figure 1. Anatomical structures of the eye.

The eyeball is surrounded by a bony orbit and occupies most of the anterior

portion of this cavity. A total of six extraocular muscles suspend the eye within the orbit

and are responsible for controlling eye movement. The four rectus muscles originate at

the posterior wall of the orbit, insert anterior to the eye equator, and are named for the

region of the eye in which they insert (superior, inferior, medial, lateral). The rectus

muscles elevate, depress, abduct, or adduct the eye. Two additional muscles are termed

the superior and inferior oblique muscles. The superior oblique muscle originates at the

posterior orbit, passes around a bony structure called the trochlea, and inserts in a medial-

3

lateral direction across the superior eye. The inferior oblique originates at the medial

orbit wall and inserts in a medial-lateral direction across the inferior eye. The superior

oblique muscle elevates the back of the eye and abducts the eye, while the inferior

oblique muscle depresses the back of the eye and adducts the eye1.

EYE INJURY

Nearly 2 million eye injuries occur each year in the United States2. Eye injuries

are classified on the Abbreviated Injury Scale (AIS) as mild (AIS 1) or moderate (AIS 2).

Corneal abrasions and hyphema are AIS 1 eye injuries, while AIS 2 eye injuries include

retinal detachment, corneal/scleral laceration, globe rupture, and eye enucleation. These

injuries can greatly affect a patient’s ability to perform everyday tasks easily, with 30,000

patients becoming blind in one eye per year in the United States as a result of trauma3. A

high frequency of visual loss is associated with globe ruptures, which account for over

9,000 eye injuries in the United States each year4. There is a multitude of causes for eye

trauma, but common causes include motor vehicle crashes5-12, military operations13-18,

and ocular impacts with sporting equipment and consumer products19-25.

In 2007, approximately 66,000 people sustained vehicle-related eye injuries in the

United States2. As many as 15% to 25% of vehicle-related eye injuries are severe eye

injuries that require surgery or result in blindness or long-term visual impairment7, 11.

Globe ruptures account for up to 45% of these severe eye injuries6. The implementation

of the airbag in motor vehicles has reduced severe and fatal injuries, but the risk of less

severe injuries such as eye injuries has increased. In a study of the NASS-CDS database

from 1993-2000, the incidence of eye injury was found to be 3.1% for occupants exposed

to an airbag deployment versus 2.0% for an occupant not exposed to an airbag

4

deployment7. In addition to airbags, other vehicle components such as flying glass or

foam particles from the vehicle’s dashboard can impact the eye during a crash and result

in eye injury.

In military operations, projectile or goggle loading can result in severe eye

injuries. There has been a significant increase this century in the percentage of eye

injuries sustained by soldiers in war relative to the total number of combat injuries13, 14, 26,

27. The percentage of eye injuries has dramatically increased from 2% in World War I

and World War II to over 13% in Operation Desert Storm (Figure 2). Penetrating eye

injuries caused by shrapnel and other debris are common in combat, but nearly 25% of

eye injuries result from blunt trauma sustained in motor vehicle and helicopter crashes,

falls, and impacts with blunt objects14, 15.

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

Amer

ican

Civ

il W

ar(1

861-

1865

)

Wor

ld W

ar I

(191

4-19

18)

Wor

ld W

ar II

(Am

eric

an)

(194

1-19

45)

Wor

ld W

ar II

(Sov

iet)

(194

1-19

45)

Kore

an W

ar(1

950-

1953

)

6-D

ay W

ar(1

967)

Oct

ober

197

3 W

ar(1

973)

Leba

non

War

(198

2)

Viet

nam

War

(196

2-19

72)

Ope

ratio

n D

eser

tSt

orm

(199

0-19

91)

Ocu

lar I

njur

ies

as a

Pe

rcen

tage

of A

ll In

jurie

s

Figure 2. Ocular injuries as a percentage of all combat injuries.

Eye trauma due to sporting equipment and consumer products are also common.

Each year in the United States, more than 600,000 sports-related eye injuries occur, with

40,000 of these requiring emergency care19. Frequent agents of blunt eye trauma in

5

sports include baseballs, golf balls, squash balls, tennis balls, paintballs, and hockey

sticks20-24. Projectiles from fireworks, firearms such as BB and pellet guns, and other

consumer products can result in severe eye injuries such as globe ruptures22, 28, 29.

RESEARCH GOALS

Previous studies have investigated ocular trauma using experimental tests and

computational modeling11, 30-40 31, 41-48. This research has yielded information on the

biomechanics of eye trauma and eye injury risk for different types of impacts. Ocular

trauma continues to be a pertinent research topic for the purposes of preventing and

mitigating eye injuries. Research efforts are warranted to improve eye protection across

populations and for a variety of impact scenarios.

Anthropometric variation in the population may be an important factor to consider

when investigating trauma and predicting injury risk for individuals. Particular

anthropometries could make an individual more prone to injury. Orbit shape and size, as

well as eye placement, varies significantly between persons of different gender, age, and

ethnicity49-57. These differences in orbit and eye anthropometry are suspected to affect

the biomechanical response of the eye when subjected to a traumatic impact. To the

authors’ knowledge, the effect of orbit anthropometry on risk of eye injury has not been

previously investigated. Goals of the current research are to measure orbit and eye

anthropometry variation in the population and determine the effect of this variation on the

response of the eye when subjected to impact.

Sources of eye injury vary in the population, with the common sources being

motor vehicle crashes, sports-related impacts, and military operations2, 13, 19. For each of

these three sources of trauma, there are a variety of different projectile impacts and

6

loading scenarios that actually cause the eye injury. In motor vehicle crashes, eye injury

can result from impact with an airbag, flying glass, or foam particles from the vehicle’s

dashboard5-12, 49. Sporting equipment such as balls or hockey sticks that impact the eye

can result in severe eye injury20-24. Eye injuries in the military can be caused by shrapnel,

debris, motor vehicle and helicopter crashes, falls, and impacts with blunt objects14, 15.

The injury response of the eye subjected to these different types of impacts likely varies.

Determining the biomechanical response of the eye to a variety of impact scenarios has

implications in the arenas of automotive safety design, military goggle design, and design

and regulation of sporting eye protective equipment. The goal of the current research is

to computationally model a variety of projectile experimental impact tests to determine

the relationship between projectile parameters and the biomechanical response of the eye.

7

REFERENCES

1. Moore KL, Dalley, A.F., Agur, A.M.R. Clinically Oriented Anatomy. 6th edition ed. Baltimore: Lippincott Williams & Wilkins; 2010. 2. McGwin G, Jr., Xie A, Owsley C. Rate of eye injury in the United States. Arch Ophthalmol 2005;123:970-976. 3. Parver LM. Eye trauma. The neglected disorder. Arch Ophthalmol 1986;104:1452-1453. 4. Smith D, Wrenn K, Stack LB. The epidemiology and diagnosis of penetrating eye injuries. Acad Emerg Med 2002;9:209-213. 5. Lueder GT. Air bag-associated ocular trauma in children. Ophthalmology 2000;107:1472-1475. 6. Kuhn F, Collins P, Morris R, Witherspoon CD. Epidemiology of motor vehicle crash-related serious eye injuries. Accid Anal Prev 1994;26:385-390. 7. Duma SM, Jernigan MV, Stitzel JD, et al. The effect of frontal air bags on eye injury patterns in automobile crashes. Arch Ophthalmol 2002;120:1517-1522. 8. Duma SM, Kress TA, Porta DJ, et al. Airbag-induced eye injuries: a report of 25 cases. J Trauma 1996;41:114-119. 9. Duma SM, Rath AL, Jernigan MV, Stitzel JD, Herring IP. The effects of depowered airbags on eye injuries in frontal automobile crashes. Am J Emerg Med 2005;23:13-19. 10. Muller-Jensen K, Allmaras W. [Eye injuries by safety glass (windshield)]. Hefte Unfallheilkd 1969;99:259-263. 11. Duma SM, Crandall JR. Eye injuries from airbags with seamless module covers. J Trauma 2000;48:786-789. 12. Fukagawa K, Tsubota K, Kimura C, et al. Corneal endothelial cell loss induced by air bags. Ophthalmology 1993;100:1819-1823. 13. Heier JS, Enzenauer RW, Wintermeyer SF, Delaney M, LaPiana FP. Ocular injuries and diseases at a combat support hospital in support of Operations Desert Shield and Desert Storm. Arch Ophthalmol 1993;111:795-798. 14. Biehl JW, Valdez J, Hemady RK, Steidl SM, Bourke DL. Penetrating eye injury in war. Mil Med 1999;164:780-784. 15. Mader TH, Aragones JV, Chandler AC, et al. Ocular and ocular adnexal injuries treated by United States military ophthalmologists during Operations Desert Shield and Desert Storm. Ophthalmology 1993;100:1462-1467. 16. Mader TH, Carroll RD, Slade CS, George RK, Ritchey JP, Neville SP. Ocular war injuries of the Iraqi insurgency, January-September 2004. Ophthalmology 2006;113:97-104. 17. Thach AB, Johnson AJ, Carroll RB, et al. Severe eye injuries in the war in Iraq, 2003-2005. Ophthalmology 2008;115:377-382. 18. Weichel ED, Colyer MH, Ludlow SE, Bower KS, Eiseman AS. Combat Ocular Trauma Visual Outcomes during Operations Iraqi and Enduring Freedom. Ophthalmology 2008;115:2235-2245. 19. Hecker S. More than half a million Americans suffer eye injuries from sports-related accidents: Lack of proper eye protection can lead to painful injuries, vision loss and even blindness. In: America PB (ed), Vision News. Chicago; 2007:1-2.

8

20. Cassen JH. Ocular trauma. Hawaii Med J 1997;56:292-294. 21. Rodriguez JO, Lavina AM, Agarwal A. Prevention and treatment of common eye injuries in sports. Am Fam Physician 2003;67:1481-1488. 22. Chisholm L. Ocular injury due to blunt trauma. Appl Ther 1969;11:597-598. 23. Pardhan S, Shacklock P, Weatherill J. Sport-related eye trauma: a survey of the presentation of eye injuries to a casualty clinic and the use of protective eye-wear. Eye (Lond) 1995;9 ( Pt 6 Su):50-53. 24. Vinger PF, Sparks JJ, Mussack KR, Dondero J, Jeffers JB. A program to prevent eye injuries in paintball. Sports Vision 1997;33:33-40. 25. Thach AB, Ward TP, Hollifield RD, et al. Ocular injuries from paintball pellets. Ophthalmology 1999;106:533-537. 26. Stone W. Ocular injuries in the Armed Forces. Journal of the American Medical Association 1950;142:151. 27. Wong TY, Smith GS, Lincoln AE, Tielsch JM. Ocular trauma in the United States Army: hospitalization records from 1985 through 1994. American journal of ophthalmology 2000;129:645-650. 28. Berger RE. A model for evaluating the ocular contusion injury potential of propelled objects. J Bioeng 1978;2:345-358. 29. Kuhn FC, Morris RC, Witherspoon DC, et al. Serious fireworks-related eye injuries. Ophthalmic epidemiology 2000;7:139-148. 30. Delori F, Pomerantzeff O, Cox MS. Deformation of the globe under high-speed impact: its relation to contusion injuries. Invest Ophthalmol 1969;8:290-301. 31. Stitzel JD, Duma SM, Cormier JM, Herring IP. A nonlinear finite element model of the eye with experimental validation for the prediction of globe rupture. Stapp Car Crash J 2002;46:81-102. 32. Kennedy EA, Ng, TP, Duma, SM. Evaluating eye injury risk of Airsoft pellet guns by parametric risk functions. Biomed Sci Instrum 2006;42:7-12. 33. Kennedy EA, Ng TP, McNally C, Stitzel JD, Duma SM. Risk functions for human and porcine eye rupture based on projectile characteristics of blunt objects. Stapp Car Crash J 2006;50:651-671. 34. Scott WR, Lloyd WC, Benedict JV, Meredith R. Ocular injuries due to projectile impacts. Annu Proc Assoc Adv Automot Med 2000;44:205-217. 35. Green RP, Jr., Peters DR, Shore JW, Fanton JW, Davis H. Force necessary to fracture the orbital floor. Ophthal Plast Reconstr Surg 1990;6:211-217. 36. Weidenthal DT, Schepens CL. Peripheral fundus changes associated with ocular contusion. American journal of ophthalmology 1966;62:465-477. 37. Weidenthal DT. Experimental Ocular Contusion. Arch Ophthalmol 1964;71:77-81. 38. McKnight SJ, Fitz J, Giangiacomo J. Corneal rupture following radial keratotomy in cats subjected to BB gun injury. Ophthalmic Surg 1988;19:165-167. 39. Duma SM, Ng TP, Kennedy EA, Stitzel JD, Herring IP, Kuhn F. Determination of significant parameters for eye injury risk from projectiles. J Trauma 2005;59:960-964. 40. Vinger PF, Duma SM, Crandall J. Baseball hardness as a risk factor for eye injuries. Arch Ophthalmol 1999;117:354-358.

9

41. Power ED, Duma SM, Stitzel JD, et al. Computer modeling of airbag-induced ocular injury in pilots wearing night vision goggles. Aviat Space Environ Med 2002;73:1000-1006. 42. Stitzel JD, Hansen GA, Herring IP, Duma SM. Blunt trauma of the aging eye: injury mechanisms and increasing lens stiffness. Arch Ophthalmol 2005;123:789-794. 43. Uchio E, Kadonosono K, Matsuoka Y, Goto S. Simulation of air-bag impact on an eye with transsclerally fixated posterior chamber intraocular lens using finite element analysis. J Cataract Refract Surg 2004;30:483-490. 44. Uchio E, Ohno S, Kudoh J, Aoki K, Kisielewicz LT. Simulation model of an eyeball based on finite element analysis on a supercomputer. Br J Ophthalmol 1999;83:1106-1111. 45. Uchio E, Ohno S, Kudoh K, Kadonosono K, Andoh K, Kisielewicz LT. Simulation of air-bag impact on post-radial keratotomy eye using finite element analysis. J Cataract Refract Surg 2001;27:1847-1853. 46. Kisielewicz LT, Kodama N, Ohno S, Uchio E. Numerical Prediction of Airbag Caused Injuries on Eyeballs After Radial Keratotomy. Society of Automotive Engineers 1998. 47. Kennedy EA, Stitzel, JD, Duma, SM. Matched experimental and computational simulations of paintball eye impacts. Biomed Sci Instrum 2008;44:243-248. 48. Weaver A, Loftis K, Duma S, Stitzel J. Biomechanical Modeling of Eye Trauma for Different Orbit Anthropometries. Ophthalmology (submitted for publication) 2010. 49. Rao SK, Greenberg PB, Filippopoulos T, Scott IU, Katsoulakis NP, Enzer YR. Potential impact of seatbelt use on the spectrum of ocular injuries and visual acuity outcomes after motor vehicle accidents with airbag deployment. Ophthalmology 2008;115:573-576.

10

Chapter 2: CT Based Three-Dimensional Measurement of Orbit and Eye Anthropometry

Accepted for publication at Investigative Ophthalmology and Visual Science, April 2010

Ashley A. Weaver1,2, Kathryn L. Loftis1,2, Josh C. Tan1, Stefan M. Duma2,3, Joel D. Stitzel1,2

1Wake Forest University School of Medicine, Winston-Salem, NC 2Virginia Tech – Wake Forest University Center for Injury Biomechanics,

Winston-Salem, NC 3Virginia Polytechnic Institute and State University, Blacksburg, VA

This chapter is formatted according to the requirements of the journal to which the paper

was submitted. Ashley Weaver assisted with data collection, analyzed the data, and

prepared the manuscript. Kathryn Loftis was responsible for the majority of the data

collection and assisted with the manuscript preparation. Josh Tan assisted with the study

design. Dr. Stefan Duma and Dr. Joel Stitzel acted in an advisory and editorial capacity.

11

ABSTRACT

Purpose: To measure eye and orbit anthropometric variation within the normal

population using CT images, and investigate the effects of age and gender on eye and

orbit anthropometry. Quantification of eye and orbit anthropometric variation within the

normal population and between persons of different age and gender is important for the

prediction and prevention of eye injury.

Methods: A systematic method was developed to three-dimensionally align head CT

images and measure ocular and orbital parameters for 39 subjects. Twenty-four

measurements along the orbital rim were collected to quantify orbital aperture.

Protrusion of the brow and the eye were measured, along with relative distance

measurements to describe location of the eye within the orbit.

Results: The orbit widened with age and significant relationships were identified relating

subject height to orbital aperture and eye location measurements. Orbital aperture and

eye location measurements varied significantly between genders.

Conclusions: The comprehensive set of measurements collected in this study provides

three-dimensional information on orbit geometry, as well as placement of the eye within

the orbit. These measurements and the methodology employed in this study will

contribute to the development of finite element models of the orbit and eye for

computational modeling purposes and may be useful for the design of eye protection

equipment.

12

INTRODUCTION

Over 1.9 million eye injuries occur each year in the United States1. Common

causes of eye trauma include motor vehicle crashes2-9, military operations10-15, and ocular

impacts with sporting equipment16-22. In military, automotive, and sporting safety, there

is concern over eye protection for different individuals. Literature shows there are

significant differences in ocular and orbital geometry among individuals of different ages,

genders, and ethnicities23-32. Differences in eye and orbit anthropometry are suspected to

affect the response of the eye when subjected to a traumatic impact.

Age and gender related ocular and orbital changes have been previously studied.

Ocular protrusion decreases with age in both males and females23. This study found the

mean protrusion for 21-30 year olds to be 20.2 mm and the mean protrusion for 71-80

year olds to be 16.9 mm, with an average reduction of 0.06 mm/yr. Ocular protrusion

and interpupillary distance have been shown to vary significantly between genders with

males having greater measurements than females28. Eyelid fissure has been shown to

lengthen by more than 10% between the ages of 12 and 25 years and shorten by a similar

amount after 45 years of age26. Aging also causes a downward shift of the lower eyelid,

which is more prominent in males than females. Another study found the orbital rim and

cheek mass move posteriorly with respect to the anterior cornea with age31. Pessa and

Chen measured orbital aperture on human skulls by measuring distances to the superior

and inferior orbital rims from a horizontal centerline through the orbit27. These superior

and inferior orbital rim distances were found to increase with age, especially at the

medial superior orbital rim and lateral inferior orbital rim. It was hypothesized that this

change in orbital aperture with age could also affect eye measurements.

13

Ethnic differences in ocular and facial anthropometry have also been noted in

previous literature. The interpupillary distance, outer canthal distance, and inner canthal

distance for African-American males and females from birth to 24 years old were found

to be statistically significant when compared to Caucasian population measurements25. In

another study, the interpupillary distance, palpebral fissure width, and eye protrusion

have been shown to vary significantly between African-American and Caucasian

populations28. Inner canthal and interpupillary distance were found to be greater in

American-born Japanese compared to Caucasians of a similar age30. Hispanics were also

shown to have a wider interpupillary distance compared with Caucasians. Studies have

shown eye protrusion is greater for African-Americans than Caucasians, and eye

protrusion is smaller for Hispanics compared with Caucasian and African-Americans24, 28,

29.

The majority of previous studies have used external measurements or photographs

to measure surface anatomy of the eye and surrounding facial regions23-26, 28-30. However,

these methods are not capable of describing the underlying bony and soft tissue anatomy

that may be important in the protection of the eye during traumatic impact. Human skulls

have been used to measure orbital aperture, but this method does not allow for

measurement of the location or protrusion of the eye within the orbit27. With CT images,

accurate measurements can be collected for soft tissue structures such as the eye, as well

as the underlying bony structures that surround and protect the eye. Select distance

measurements describing the eye’s relation to surrounding bone and soft tissue, orbital

volume, and orbital wall morphology have been collected using CT images, but further

measurements are necessary to fully describe eye and orbit anatomy31, 33-36. Accurate

14

measurement of eye and orbit anthropometry is valuable in the design of eye protective

equipment and modeling of facial impacts for injury prediction purposes.

METHODS

For this study, a systematic method was developed to collect several ocular and

orbital measurements from CT images. Twenty-four measurements along the orbital rim

were collected to quantify the orbital aperture. The protrusion of the brow and the eye

were measured, along with distance measurements to describe the location of the eye

within the orbit.

Head CT scans of 39 Caucasian subjects from Wake Forest University Baptist

Medical Center were obtained. Research adhered to the tenets of the Declaration of

Helsinki. Collected scans were ruled to be grossly normal in head/skull anatomy and of

adequate imaging quality to obtain geometrical measurements. Measurements of the left

orbit and eye for each subject were collected using multiple software programs. Images

were aligned within TeraRecon AquariusNET Server version 1.8.1.6 (TeraRecon, Inc.,

San Mateo, CA) medical imaging software and measurements were collected from

screenshots of the aligned images using ImageJ (National Institutes of Health, Bethesda,

MD) software. Coordinates along the orbital rim were collected using a separate

software program, Amira (Visage Imaging, San Diego, CA).

Within TeraRecon, a slab view at maximal thickness was used to align the CT

correctly for each subject and a Hounsfield Unit (HU) based bone window was initially

selected for viewing the image (Window Width (WW) 2200, Window Level (WL) 200).

The anterior-posterior axis was aligned with the crista galli in the axial view (Figure 3a).

The coronal view (window levels: WW 350, WL 75) was used to align the superior-

15

inferior axis along the falx cerebri (Figure 3b). Due to former alignment in the other

viewing windows, the nasion and sella turcica were visible in the sagittal view (window

levels: WW 2200, WL 200). The axial plane was realigned with the nasion-sella plane

(Figure 3c), an anatomical plane used in a former study of the orbit27.

The aligned CT images were rendered in three-dimensions (3D) and were

optimized with the following parameters: Mask 1 (WW 500, WL 400, opacity 100%,

right_up linear volume-rendering curve shape), Mask 2 (WW 200, WL 200, opacity 20%,

right_recline triangle volume-rendering curve shape). A screenshot of the aligned 3D

bone reconstruction was captured and a grid template was overlaid to collect orbital

aperture measurements (Figure 3d). The grid was rotated so the horizontal axis was

parallel with a line connecting the superior rims of both orbits. The grid was positioned

with the left horizontal endpoint on the nasion and the right horizontal endpoint on the

lateral edge of the orbit. The eleven vertical lines of the grid were constructed to be

equally spaced along the horizontal axis. Twenty-four orbital rim measurements were

collected from each of the horizontal and vertical gridlines. The superior orbital rim

perimeter was calculated by summing the distances between each of the twelve superior

orbital rim measurements. The inferior orbital rim perimeter was calculated using the

same method and was summed with the superior orbital rim perimeter to compute the

total orbital rim perimeter. The two segments corresponding to orbital width (OW) and

orbital height (OH) are depicted in Figure 3d. For the statistical analysis, these

measurements were also normalized to adjust for size. To normalize, each measurement

was divided by the subject’s height in millimeters and multiplied by 100. This allowed

16

orbital width and height measurements that are relative to the subject height to be

investigated across different genders and ages.

Crista galliCrista galli

a)

Falx cerebriFalx cerebri

b)

Nasion

Sellaturcica

Nasion

Sellaturcica

c)

Orbital width

Orbital height

1 2 3 4 5 6 7 8 9 10 11

Orbital width

Orbital height

1 2 3 4 5 6 7 8 9 10 11 d)

Figure 3. Alignment with the crista galli (a), falx cerebri (b), and nasion-sella turcica plane (c). Aligned 3D image used to measure orbital aperture (d).

17

Further alignment of the images allowed the protrusion and location of the eye to

be measured. An abdomen window was selected for viewing the eyeball (window levels:

WW 350, WL 75). In the axial view, the anterior-posterior axis was rotated until it was

aligned with the cornea and the center of the optic canal. The axial slice with the most

eye protrusion was selected and the perpendicular (medial-lateral) axis was placed on the

edge of the lateral orbital rim, as shown in Figure 4a. Lateral eye protrusion (LP) was

measured from the cornea to the axes intersection. This protrusion measurement

corresponded to the protrusion of the eye relative to the lateral orbital rim. A lateral

distance (LD) measurement describing the eye’s location relative to the lateral orbital rim

was collected by measuring from the lateral edge of the orbit to the axes intersection.

The sagittal view of the previous alignment was used to collect superior eye protrusion

(SP) and superior distance (SD) measurements (Figure 4b). A line was drawn from the

cornea back through the center of the eye and another line was drawn connecting the

superior and inferior orbital rims (white lines in Figure 4b). Superior eye protrusion was

measured from the cornea to intersection of these two lines to describe protrusion of the

eye relative to the superior orbital rim. A superior distance measurement describing the

eye’s location relative to the superior orbital rim was collected by measuring from the

superior edge of the orbit to the intersection of the two white lines.

The CT scans were imported into Amira software to record point coordinate

values on the orbital rim. The bony anatomy was rendered in 3D and aligned with the

coordinate system of the 3D reconstruction in Figure 3d. As described earlier, the crista

galli, falx cerebri, and nasion-sella plane were used to establish a landmark-defined

coordinate system for each subject. Four coordinate points were recorded: the medial

18

orbital rim point at the nasion, the lateral orbital rim point horizontally across from the

nasion, and the center orbital rim point located superiorly and inferiorly (Figure 5). The

angle of brow protrusion (BP) was calculated using Equation 1, where OH is the orbital

height and D is the distance between the superior and inferior coordinates depicted in

Figure 5. The formula is chosen so superior orbital rim points anterior to the inferior

orbital rim point result in positive angles.

⎟⎠⎞

⎜⎝⎛= −

DOHBP 1cos Equation 1

LD

LP

LD

LP

a)

SD

SP

SD

SP

b)

Figure 4. a) Lateral eye protrusion (LP) and lateral distance (LD). b) Superior eye protrusion (SP) and superior distance (SD).

19

M

S

I

Lat. node

LM

S

I

Lat. node

L

Figure 5. Brow protrusion angle and medial (M), lateral (L), superior (S), and inferior (I) orbital rim coordinate points.

RESULTS

Select subject demographics and ocular and orbital measurements are reported in

Table 1. For some subjects, CT scan slices did not extend to the inferior orbit and it was

not possible to collect all the measurements. Missing data was accounted for in the

statistical analysis.

20

Table 1. Demographics and ocular and orbital measurements.

Subject Sex Age Height, cm

OW, mm

OH, mm

LP, mm

LD, mm

SP, mm

SD, mm

BP, deg

1 M 17 180 36.6 - 7.7 18.6 - - - 2 M 20 172 34.0 29.6 10.2 19.2 4.4 10.5 29.33 M 21 175 35.3 33.2 12.0 19.0 4.9 16.3 29.34 M 21 185 32.1 31.0 6.1 16.1 1.9 15.0 32.65 M 23 177 36.4 32.1 12.6 19.4 3.6 13.5 25.66 F 24 162 36.4 35.1 13.1 16.2 3.9 16.7 22.87 F 24 152 38.1 28.8 9.0 16.9 1.4 14.7 31.98 M 25 177 36.1 32.7 11.4 21.2 2.4 18.3 - 9 F 29 173 35.1 34.1 9.1 18.5 5.0 16.8 20.1

10 M 31 185 36.1 - 16.0 18.7 - - - 11 F 32 167 40.3 27.5 5.5 22.2 3.0 10.7 27.312 M 33 180 37.2 33.3 16.7 17.9 8.0 18.2 26.213 F 34 162 36.6 27.7 13.4 17.1 3.5 14.4 28.614 M 37 185 37.9 - 17.0 17.0 - - - 15 F 38 157 39.4 33.7 14.0 15.5 6.9 14.7 21.316 F 40 172 35.6 35.6 13.7 15.2 4.4 24.2 27.317 F 42 157 34.8 30.4 9.4 18.3 6.5 14.9 22.918 F 42 160 36.6 - 12.6 19.0 - - - 19 F 42 165 35.1 31.9 12.1 17.4 8.3 14.6 28.720 F 43 167 34.4 31.1 11.4 17.3 5.9 16.3 - 21 F 47 165 36.5 32.2 13.0 17.2 4.2 14.8 20.622 M 49 183 38.5 34.5 10.8 20.4 5.0 20.6 37.823 F 51 165 38.9 30.7 14.9 16.9 2.2 15.5 34.424 F 51 165 36.2 31.4 11.6 18.0 3.9 17.2 22.025 F 54 163 36.6 - 6.7 19.2 - - - 26 F 57 172 39.3 34.8 12.2 16.1 5.0 17.4 20.227 M 59 185 39.6 36.2 15.5 17.6 6.3 20.6 26.828 F 60 149 34.9 33.0 18.7 18.3 13.4 15.4 32.729 M 60 190 41.8 34.0 19.7 18.8 11.4 17.9 - 30 M 60 180 38.5 32.0 12.1 22.0 4.1 16.0 27.631 F 60 160 36.6 - 13.9 16.7 5.3 17.3 - 32 F 60 157 36.6 34.0 11.5 17.1 5.1 14.3 28.433 M 62 166 39.6 29.2 14.4 20.7 2.8 16.9 - 34 F 63 155 33.4 28.0 11.2 16.8 2.9 15.1 26.535 M 65 177 39.4 32.2 14.9 18.1 6.1 16.5 26.936 M 66 190 39.5 31.9 10.6 19.4 2.3 18.2 25.937 F 66 167 37.6 - 13.4 17.4 - - - 38 F 74 170 36.5 31.5 6.8 18.3 1.1 15.7 30.239 F 76 167 36.4 31.9 11.4 19.9 2.1 18.1 20.9

Avg. 45.1 170. 2 36.9 32.0 12.2 18.2 4.8 16.3 27.0S.D. 16.8 10.7 2.1 2.3 3.2 1.7 2.7 2.6 4.5

21

A multivariate correlation analysis was used to examine relationships between

subject demographics (age and subject height) and the ocular and orbital measurements

collected (OW, OH, LP, SP, LD, SD, orbital rim perimeters, and height-normalized OW

and OH). Pearson product-moment correlation coefficients and p-values were computed

for all combinations, comparing demographics to measurements, as well as

measurements to each other. Results are presented in Table 2 for correlations that were

significant (p<0.05) or mildly significant (p<0.10). Noteworthy findings included

significant positive correlations with age for the raw and height-normalized OW

measurements, as well as a mildly significant relationship between age and the SD

measurement. Several measurements (OH, height-normalized OW and OH, SD, and the

inferior and total rim perimeters) were significantly correlated with subject height,

suggesting that subject height is a key factor in explaining variation in orbit

anthropometry across individuals. In addition, mild correlations with subject height were

identified for the OW and LD measurements. Many of the ocular and orbital

measurements were significantly correlated with each other. Measurements

characterizing orbital aperture (OW, OH, and orbital rim perimeters) were found to be

significantly correlated with eye protrusion (LP and SP measurements) and location of

the eye within the orbit (LD and SD measurements). Eye protrusion measurements (LP,

SP) were not only significantly correlated with orbital aperture measurements, but also

with the SD measurement describing eye location within the orbit. No correlations with

p-values less than 0.10 were identified for the brow protrusion angle.

22

Table 2. Pearson product-moment correlation coefficients and p-values.

Variable by Variable Correlation p-value OW Age 0.34 0.03 *OW LP 0.33 0.04 *OW Subject Height 0.27 0.09

Norm. OW Subject Height -0.67 < 0.001 *Norm. OW OW 0.53 < 0.001 *Norm. OW Age 0.34 0.03 *Norm. OW Norm. OH 0.34 0.05 *

OH SD 0.65 < 0.001 *OH LP 0.41 0.02 *OH SP 0.40 0.02 *OH Subject Height 0.36 0.04 *

Norm. OH OH 0.63 < 0.001 *Norm. OH LD -0.50 0.004 *Norm. OH Subject Height -0.49 0.005 *Norm. OH SP 0.43 0.01 *Norm. OH LP 0.35 0.05 *

LD Subject Height 0.28 0.08 SD Subject Height 0.39 0.02 *SD LP 0.36 0.04 *SD Age 0.30 0.09 SP LP 0.68 < 0.001 *

Inferior rim perimeter Subject Height 0.55 0.001 *Inferior rim perimeter OW 0.51 0.003 *Inferior rim perimeter LD 0.42 0.02 *Inferior rim perimeter OH 0.31 0.08 Superior rim perimeter OH 0.51 0.003 *Superior rim perimeter Norm. OH 0.47 0.007 *Superior rim perimeter SD 0.46 0.007 *Superior rim perimeter Norm. OW 0.37 0.02 *Superior rim perimeter OW 0.35 0.03 *Superior rim perimeter LP 0.29 0.07 Superior rim perimeter LD -0.29 0.08

Total rim perimeter OW 0.75 < 0.001 *Total rim perimeter Inferior rim perimeter 0.67 < 0.001 *Total rim perimeter OH 0.66 < 0.001 *Total rim perimeter Superior rim perimeter 0.55 0.001 *Total rim perimeter SD 0.50 0.004 *Total rim perimeter Subject Height 0.47 0.007 *Total rim perimeter LP 0.35 0.05 *

*indicates statistical significance (p<0.05)

23

One-way Analysis of Variance (ANOVA) was used to assess the effect of gender

on each of the ocular and orbital measurements. ANOVA F test statistics and p-values

are reported in Table 3. P-values less than 0.05 indicate a statistically significant

difference in sample means between genders. As Table 3 shows, the lateral distance

measurement and inferior orbital rim perimeter vary significantly between males and

females. Orbital width and height measurements normalized by subject height vary

significantly between genders. When these measurements (OW, OH) are not normalized,

no significant differences are detected between genders, suggesting that normalizing by

height unmasks the gender effect on orbital anthropometry.

Table 3. Pattern of variation in ocular and orbital measurements between genders.

Male (n=16) Female (n=16)

Measurement Mean SD Mean SD

F Ratio p-value

OW (mm) 37.42 2.44 36.60 1.71 1.55 0.22 OH (mm) 32.44 1.89 31.75 2.51 0.71 0.41 Normalized OW 2.08 0.14 2.25 0.13 16.03 <0.001 * Normalized OH 1.80 0.08 1.95 0.15 9.29 0.005 * LP (mm) 12.98 3.60 11.69 2.94 1.51 0.23 LD (mm) 19.01 1.55 17.63 1.53 7.68 0.009 * SP (mm) 4.85 2.64 4.70 2.76 0.02 0.88 SD (mm) 16.81 2.75 15.94 2.53 0.87 0.36 BP (deg) 28.80 3.81 25.94 4.69 2.71 0.11 Inferior rim perimeter (mm) 64.97 3.83 61.37 4.89 4.95 0.03 * Superior rim perimeter (mm) 49.30 4.39 50.94 3.79 1.54 0.22 Total rim perimeter (mm) 114.74 5.49 112.15 5.44 1.74 0.20

* indicates statistical significance (p<0.05)

DISCUSSION

The current study presents a systematic method to align head CT images and

collect orbital aperture, eye protrusion, eye location, and brow protrusion measurements

that can be compared across individuals. This study documented variation in normal eye

24

and orbit anatomy using 39 subjects. Contrary to previous studies that examined

variation in surface anatomy using external measurements and photographs23-26, 28-30, this

study investigates variation in bony and soft tissue anatomy of the eye and orbit. Only a

few studies have collected measurements to quantify orbital geometry and the location of

the eye within the orbit27, 31. The collection of additional measurements was warranted to

fully characterize orbit and eye anatomy and quantify variation across individuals.

Statistical results of this study suggest the orbit widens with age. Results also

showed significant relationships exist between subject height and the orbital aperture and

eye location measurements. These findings suggest variation in orbital anthropometry

can partially be attributed to differences in subject height and normalizing by subject

height may reveal other effects on orbit anthropometry such as age, gender, and ethnicity.

Normalizing orbital width and height by subject height, the statistical analysis showed

significant differences between genders exist for these measurements. Normalized

orbital width and height was greater in females compared to males, suggesting that

relative to subject height, the female’s orbital aperture is proportionately larger than the

male’s orbital aperture. However, without normalization, the means of all the ocular and

orbital measurements were greater in males than females. Although statistical

significance was not reached when comparing each measurement across genders, the

lateral distance describing eye location and the inferior orbital rim perimeter were both

significantly greater in males than females. Future studies with a larger sample size may

find significant differences between genders for other parameters such as eye and brow

protrusion.

25

The smaller sample size used in this study may have affected the ability to obtain

statistical significance when examining age and gender effects. Contrary to the literature,

no significant decrease in eye protrusion was observed with age, but sample size was

substantially smaller than the number of subjects in previous work and may not have

been sufficient to obtain statistical significance23. Ethnic variation was not accounted for

in this study, thus the effect of ethnicity on the ocular and orbital measurements could not

be assessed at this time. The method developed could be used in future studies to collect

measurements for a larger number of subjects and quantify ocular and orbital

anthropometry across individuals of varying ages, genders, and ethnicities.

The comprehensive set of measurements collected in this study provides detailed

information on orbital geometry, as well as placement of the eye within the orbit. This

set of measurements will be used for the development of finite element models of the

orbit for computational modeling purposes and may be useful for the design of eye

protective equipment. Although sample size limited the ability to obtain statistically

significant relationships regarding orbital anthropometry, the significance of the effect of

orbital variation on eye injury has yet to be assessed. Incorporating variation in orbit

anthropometry and information about eye location into a finite element model of the

human eye37, computational simulations will be used to study interaction between the eye

and orbit during traumatic impact to evaluate risk for eye injury and orbital fractures.

When normal variation in orbital anthropometry is implemented into a finite element

model of the eye, statistically significant injury risk variation may result. It is well-

known that anthropometric variation exists in the population and determining these injury

risk functions will be valuable in mitigating eye injury across individuals.

26

REFERENCES

1. McGwin G, Jr., Xie A, Owsley C. Rate of eye injury in the United States. Arch Ophthalmol 2005;123:970-976. 2. Lueder GT. Air bag-associated ocular trauma in children. Ophthalmology 2000;107:1472-1475. 3. Kuhn F, Collins P, Morris R, Witherspoon CD. Epidemiology of motor vehicle crash-related serious eye injuries. Accid Anal Prev 1994;26:385-390. 4. Duma SM, Jernigan MV, Stitzel JD, et al. The effect of frontal air bags on eye injury patterns in automobile crashes. Arch Ophthalmol 2002;120:1517-1522. 5. Duma SM, Kress TA, Porta DJ, et al. Airbag-induced eye injuries: a report of 25 cases. J Trauma 1996;41:114-119. 6. Duma SM, Rath AL, Jernigan MV, Stitzel JD, Herring IP. The effects of depowered airbags on eye injuries in frontal automobile crashes. Am J Emerg Med 2005;23:13-19. 7. Muller-Jensen K, Allmaras W. [Eye injuries by safety glass (windshield)]. Hefte Unfallheilkd 1969;99:259-263. 8. Duma SM, Crandall JR. Eye injuries from airbags with seamless module covers. J Trauma 2000;48:786-789. 9. Fukagawa K, Tsubota K, Kimura C, et al. Corneal endothelial cell loss induced by air bags. Ophthalmology 1993;100:1819-1823. 10. Heier JS, Enzenauer RW, Wintermeyer SF, Delaney M, LaPiana FP. Ocular injuries and diseases at a combat support hospital in support of Operations Desert Shield and Desert Storm. Arch Ophthalmol 1993;111:795-798. 11. Biehl JW, Valdez J, Hemady RK, Steidl SM, Bourke DL. Penetrating eye injury in war. Mil Med 1999;164:780-784. 12. Mader TH, Aragones JV, Chandler AC, et al. Ocular and ocular adnexal injuries treated by United States military ophthalmologists during Operations Desert Shield and Desert Storm. Ophthalmology 1993;100:1462-1467. 13. Mader TH, Carroll RD, Slade CS, George RK, Ritchey JP, Neville SP. Ocular war injuries of the Iraqi insurgency, January-September 2004. Ophthalmology 2006;113:97-104. 14. Thach AB, Johnson AJ, Carroll RB, et al. Severe eye injuries in the war in Iraq, 2003-2005. Ophthalmology 2008;115:377-382. 15. Weichel ED, Colyer MH, Ludlow SE, Bower KS, Eiseman AS. Combat Ocular Trauma Visual Outcomes during Operations Iraqi and Enduring Freedom. Ophthalmology 2008;115:2235-2245. 16. Hecker S. More than half a million Americans suffer eye injuries from sports-related accidents: Lack of proper eye protection can lead to painful injuries, vision loss and even blindness. In: America PB (ed), Vision News. Chicago; 2007:1-2. 17. Cassen JH. Ocular trauma. Hawaii Med J 1997;56:292-294. 18. Rodriguez JO, Lavina AM, Agarwal A. Prevention and treatment of common eye injuries in sports. Am Fam Physician 2003;67:1481-1488. 19. Chisholm L. Ocular injury due to blunt trauma. Appl Ther 1969;11:597-598.

27

20. Pardhan S, Shacklock P, Weatherill J. Sport-related eye trauma: a survey of the presentation of eye injuries to a casualty clinic and the use of protective eye-wear. Eye (Lond) 1995;9 ( Pt 6 Su):50-53. 21. Vinger PF, Sparks JJ, Mussack KR, Dondero J, Jeffers JB. A program to prevent eye injuries in paintball. Sports Vision 1997;33:33-40. 22. Thach AB, Ward TP, Hollifield RD, et al. Ocular injuries from paintball pellets. Ophthalmology 1999;106:533-537. 23. Ahmadi H, Shams PN, Davies NP, Joshi N, Kelly MH. Age-related changes in the normal sagittal relationship between globe and orbit. J Plast Reconstr Aesthet Surg 2007;60:246-250. 24. Bolanos Gil de Montes F, Perez Resinas FM, Rodriguez Garcia M, Gonzalez Ortiz M. Exophthalmometry in Mexican adults. Rev Invest Clin 1999;51:341-343. 25. Pivnick EK, Rivas ML, Tolley EA, Smith SD, Presbury GJ. Interpupillary distance in a normal black population. Clin Genet 1999;55:182-191. 26. van den Bosch WA, Leenders I, Mulder P. Topographic anatomy of the eyelids, and the effects of sex and age. Br J Ophthalmol 1999;83:347-352. 27. Pessa JE, Chen Y. Curve analysis of the aging orbital aperture. Plast Reconstr Surg 2002;109:751-755; discussion 756-760. 28. Barretto RL, Mathog RH. Orbital measurement in black and white populations. Laryngoscope 1999;109:1051-1054. 29. Dunsky IL. Normative data for hertel exophthalmometry in a normal adult black population. Optom Vis Sci 1992;69:562-564. 30. Pryor HB. Objective measurement of interpupillary distance. Pediatrics 1969;44:973-977. 31. Pessa JE, Desvigne LD, Lambros VS, Nimerick J, Sugunan B, Zadoo VP. Changes in ocular globe-to-orbital rim position with age: implications for aesthetic blepharoplasty of the lower eyelids. Aesthetic Plast Surg 1999;23:337-342. 32. Shufelt C, Fraser-Bell S, Ying-Lai M, Torres M, Varma R. Refractive error, ocular biometry, and lens opalescence in an adult population: the Los Angeles Latino Eye Study. Investigative ophthalmology & visual science 2005;46:4450-4460. 33. Song WK, Lew H, Yoon JS, Oh MJ, Lee SY. Role of medial orbital wall morphologic properties in orbital blow-out fractures. Investigative ophthalmology & visual science 2009;50:495-499. 34. Ye J, Kook KH, Lee SY. Evaluation of computer-based volume measurement and porous polyethylene channel implants in reconstruction of large orbital wall fractures. Investigative ophthalmology & visual science 2006;47:509-513. 35. Regensburg NI, Kok PH, Zonneveld FW, et al. A new and validated CT-based method for the calculation of orbital soft tissue volumes. Investigative ophthalmology & visual science 2008;49:1758-1762. 36. Forbes G, Gehring DG, Gorman CA, Brennan MD, Jackson IT. Volume measurements of normal orbital structures by computed tomographic analysis. Ajr 1985;145:149-154. 37. Stitzel JD, Duma SM, Cormier JM, Herring IP. A nonlinear finite element model of the eye with experimental validation for the prediction of globe rupture. Stapp Car Crash J 2002;46:81-102.

28

Chapter 3: Biomechanical Modeling of Eye Trauma for Different Orbit Anthropometries

Submitted to Ophthalmology, March 2010

Ashley A. Weaver1,2, Kathryn L. Loftis1,2, Stefan M. Duma2,3, Joel D. Stitzel1,2

1Wake Forest University School of Medicine, Winston-Salem, NC 2Virginia Tech – Wake Forest University Center for Injury Biomechanics,

Winston-Salem, NC 3Virginia Polytechnic Institute and State University, Blacksburg, VA

This chapter is formatted according to the requirements of the journal to which the paper

was submitted. Ashley Weaver assisted with the study design, set up the computational

simulations, analyzed the simulation results, and prepared the manuscript. Kathryn Loftis

collect anthropometric measurements and was involved with the study design. Dr. Joel

Stitzel provided assistance with the computational modeling efforts. Dr. Joel Stitzel and

Dr. Stefan Duma also acted in an advisory and editorial capacity on this project.

29

ABSTRACT

Objectives: In military, automotive, and sporting safety, there is concern over eye

protection and the effects of facial anthropometry differences on risk of eye injury. The

objective of this study is to investigate differences in orbital geometry and analyze their

effect on eye impact injury.

Design: Experimental study.

Participants: 27 different orbital anthropometries were modeled with an incorporated

eye model.

Methods: Clinical measurements of the orbital aperture, brow protrusion angle, eye

protrusion, and the eye location within the orbit were used to develop a matrix of

simulations. A finite element model of the orbit was developed from a CT scan of an

average male and transformed to model 27 different anthropometries. Impacts were

modeled using an eye model incorporating lagrangian-eulerian fluid flow for the eye,

representing a full eye for evaluation of omnidirectional impact and interaction with the

orbit. Computational simulations of a Little League (CD25) baseball impact at 30.1 m/s

were conducted to assess the effect of orbit anthropometry on eye injury metrics.

Main Outcome Measures: Parameters measured include stress and strain in the

corneoscleral shell, internal dynamic eye pressure, and contact forces between the orbit,

eye, and baseball. The location of peak stresses and strains was also assessed.

Results: Two simulations exceeded the globe rupture threshold for stress, while seven

simulations exceeded the pressure threshold. On average, higher stresses and pressures

were observed in simulations with more eye protrusion, less brow protrusion, and larger

orbital aperture. Main effects and interaction effects identified in the statistical analysis

30

illustrate the complex relationship between the anthropometric variation and eye

response.

Conclusions: The results of the study showed that the eye is more protected from impact

with smaller orbital apertures, more brow protrusion, and less eye protrusion, provided

that the orbital aperture is large enough to deter contact of the eye with the orbit.

INTRODUCTION

There are over 1.9 million eye injuries each year in the United States1. Mild eye

injuries (AIS 1) include corneal abrasions and hyphema, while moderate eye injuries

(AIS 2) include retinal detachment, corneal/scleral laceration, globe rupture, and eye

enucleation. Over 9,000 globe ruptures and 30,000 cases of blindness occur each year in

the United States as the result of trauma2, 3.

Motor vehicle crashes4-14, military operations15-21, and ocular impacts with

sporting equipment and consumer products22-28 are common causes of eye injuries. In

motor vehicle crashes, severe eye injury can result when the eye is impacted by an airbag,

flying glass, or foam particles from the vehicle’s dashboard. Approximately 66,000

people sustained vehicle-related eye injuries in the United States in 2007 and as many as

15% to 25% of vehicle-related eye injuries require surgery or result in long-term visual

impairment1, 7, 11. In military operations, projectile or goggle loading can result in severe

eye injuries. There has been a significant increase this century in the percentage of eye

injuries sustained by soldiers in war relative to the total number of combat injuries, with

over 13% of all combat injuries in Operation Desert Storm being ocular injuries15, 16.

More than 600,000 sports-related eye injuries occur each year in the United States, with

31

40,000 of these requiring emergency care22. Frequent agents of blunt trauma in sports

include baseballs, golf balls, squash balls, tennis balls, paintballs, and hockey sticks23-27.

Experimental eye impact tests with a variety of blunt objects have previously been

conducted and used to determine injury tolerance of the eye11, 29-39. Projectiles tested in

these studies included foam particles, BBs, aluminum rods, plastic rods, blunt impactors,

baseballs, and paintballs. Experimental impact tests with BBs and projectile shooting

toys have also been conducted with the Facial and Ocular Countermeasure Safety

(FOCUS) headform which contains a simulated orbit, synthetic eye, and an eye load cell

that records forces for injury prediction purposes40, 41. Computational models have been

used to model a variety of eye impacts and analyze injury potential30, 42-48. The Virginia

Tech – Wake Forest University (VT-WFU) Eye Model is a finite element model of the

eye validated to predict globe rupture for dynamic blunt eye impacts30. Experimental eye

impact tests with foam projectiles, BBs, and baseballs were performed to validate the

model and determine stress and pressure thresholds for globe rupture.

In military, automotive, and sporting safety, there is concern over eye protection

for different individuals. Literature shows that there are significant differences in orbit

shape and size, as well as eye placement, between persons of different gender, age, and

ethnicity49-57. These anthropometric differences are suspected to affect the response of

the eye when subjected to a traumatic impact. To the authors’ knowledge, previous

experimental eye impact tests and computational simulations have not investigated the

effect of orbit anthropometry on the risk of eye injury. The objective of the current study

is to model differing orbital anthropometries and eye placement to study the

biomechanical response of the eye when subjected to a blunt impact. Results of these

32

simulations will explain eye injury risk variation across individuals and are valuable to

the design of eye protection equipment.

METHODS

Test Matrix Development

Head computed tomography (CT) scans from Wake Forest University Baptist

Medical Center were obtained for 19 Caucasian subjects ages 18-50. The Institutional

Review Board (IRB) of Wake Forest University School of Medicine approved use of

deidentified scans for research (IRB#: BG05-483). These scans were ruled to be grossly

normal in head anatomy and of adequate imaging quality to obtain geometrical

measurements. A systematic method, reported separately, was used to align the CT

images to an anatomically-defined coordinate system and measurements of the left orbit

and eye were collected (Weaver AA et al., “CT Based Three-Dimensional Measurement

of Orbit and Eye Anthropometry, submitted to Invest Ophthalmol Vis Sci). The orbital

width (OW) and height (OH) was measured to define orbital aperture (Figure 6a).

Lateral distance (LD) and superior distance (SD) measurements were collected to

describe the eye’s location relative to the lateral and superior orbital rims. Protrusion of

the eye (EP) relative to the lateral orbital rim was measured, along with the angle of brow

protrusion (BP) relative to the inferior orbital rim (Figure 6b).

33

LD

SDOW

OH

EP

BP

a) b)

LD

SDOW

OH

EP

BP

a) b)

Figure 6. a) Orbital width (OW), orbital height (OH), superior distance (SD), and lateral distance (LD). b) Brow protrusion angle (BP) and eye protrusion (EP).

The mean and standard deviation of each measurement was used to develop a test

matrix for transforming a finite element model of the orbit. For each parameter, the mean

and the value 2 standard deviations above and below the mean was used within the

matrix to accommodate for different orbital widths, orbital heights, brow protrusion

angles, and eye protrusion measurements. Location of the eye within the orbit was also

defined in the test matrix. Orbital width and height were combined into a single

parameter termed orbital aperture consisting of two coordinates. Orbital aperture, eye

protrusion, and brow protrusion were varied with each other to create 27 models. The

parameters defining each of the 27 models are listed in the test matrix (Table 4).

34

Table 4. Test matrix defining eye and orbit parameters.

Model Orbital Width (mm)

Orbital Height (mm)

Brow Protrusion

(deg)

Lateral Distance

(mm)

Superior Distance

(mm)

Eye Protrusion

(mm) 1 32.48 26.82 19.75 16.06 13.14 5.89 2 36.26 31.61 19.75 17.93 15.49 5.89 3 40.04 36.40 19.75 19.80 17.84 5.89 4 32.48 26.82 27.05 16.06 13.14 5.89 5 36.26 31.61 27.05 17.93 15.49 5.89 6 40.04 36.40 27.05 19.80 17.84 5.89 7 32.48 26.82 34.35 16.06 13.14 5.89 8 36.26 31.61 34.35 17.93 15.49 5.89 9 40.04 36.40 34.35 19.80 17.84 5.89

10 32.48 26.82 19.75 16.06 13.14 12.06 11 36.26 31.61 19.75 17.93 15.49 12.06 12 40.04 36.40 19.75 19.80 17.84 12.06 13 32.48 26.82 27.05 16.06 13.14 12.06 14 36.26 31.61 27.05 17.93 15.49 12.06 15 40.04 36.40 27.05 19.80 17.84 12.06 16 32.48 26.82 34.35 16.06 13.14 12.06 17 36.26 31.61 34.35 17.93 15.49 12.06 18 40.04 36.40 34.35 19.80 17.84 12.06 19 32.48 26.82 19.75 16.06 13.14 18.23 20 36.26 31.61 19.75 17.93 15.49 18.23 21 40.04 36.40 19.75 19.80 17.84 18.23 22 32.48 26.82 27.05 16.06 13.14 18.23 23 36.26 31.61 27.05 17.93 15.49 18.23 24 40.04 36.40 27.05 19.80 17.84 18.23 25 32.48 26.82 34.35 16.06 13.14 18.23 26 36.26 31.61 34.35 17.93 15.49 18.23 27 40.04 36.40 34.35 19.80 17.84 18.23

Orbital width, brow protrusion, and eye protrusion are plotted in Figure 7 to

depict variation in the anthropometries measured and the 27 anthropometries modeled.

The green box indicates the range of anthropometries modeled and the majority of the

measured anthropometries fall within these ranges. Similar results were produced when

plotting orbital height, brow protrusion, and eye protrusion, with the majority of the

measured anthropometric variation falling within the modeled anthropometry ranges.

35

3234

3638

4042

15

20

25

30

355

10

15

20

Orbital Width (mm)Brow Protrusion (deg)

Eye

Prot

rusi

on (m

m)

Measured AnthropometryModeled Anthropometry

Figure 7. Variation in measured and modeled anthropometries.

Finite Element Models

The previous quarter symmetry VT-WFU eye model30 was updated to a full eye

model for incorporation into computational simulations to study the eye response during

blunt impact. The full eye model was developed with Hypermesh (Hyperworks v 9.0,

Altair, Troy, MI) in conjunction with LS-Prepost (Livermore Software Technology

Corporation, Livermore, CA). The geometry utilized cross sectional dimensions identical

to those in the quarter symmetry eye model. The model is based on two spherical

geometries forming the anterior and posterior chambers of the eye and includes the

corneoscleral shell (consisting of the cornea, sclera, and limbus), as well as internal eye

structures including the ciliary body, zonules, and lens. Thicknesses are defined at the

corneal apex, equator, limbus and posterior pole and distributed evenly around the

circumference of the eye. The model has a total of 29,660 solid and shell elements and

36

includes Lagrangian and Eulerian meshes to accurately represent the mechanics of both

solid and fluid interactions. Material properties and definitions are maintained from the

previous eye model. There are two significant differences in the new model. One is that

it represents the entire eye and therefore does not require symmetry planes and will

handle impact directions other than centered anterior-posterior impacts with radially

symmetric objects. The other is a coarsening of the mesh to maintain reasonable run-

time and time step since the full eye is four times the volume of the original eye model.

The mesh of the model is depicted in Figure 8. This eye model has been previously

validated with experimental data to predict globe rupture when peak stresses in the

corneoscleral shell exceed 23 MPa and local dynamic pressures exceed 2.1 MPa.

Lens

Zonules

Corneoscleral Shell

Anterior Chamber

Ciliary Body

Vitreous

Anterior Chamber

Vitreous

Lens

Zonules

Corneoscleral Shell

Anterior Chamber

Ciliary Body

Vitreous

Anterior Chamber

Vitreous

Figure 8. Lagrangian mesh depicting corneoscleral shell, lens, zonules, and ciliary body and Eulerian mesh depicting initially filled volume (dark gray) and initially

unfilled volume (light gray).

A finite element model of the orbit was developed from a head CT of a 21 year

old male subject with height and weight measurements similar to the 50th percentile male.

37

This subject had a complete CT scan of the left orbit and eye and no reported pathology

to the orbit or surrounding regions. Mimics software v 12.01 (Materialise, Ann Arbor,

MI) was used to segment the CT scan. Thresholding and region growing operations were

used to select the bony structure of the left orbit and surrounding facial regions. The

orbit was enclosed by filling in the optic canal, and the superior and inferior orbital

fissures. A three-dimensional rendering of the orbit was smoothed to reduce artificial

texture from pixellation. The model was remeshed to improve the quality of the meshing

and further simplified to isolate only the orbit and surrounding frontal surface regions.

The test matrix was used to transform the orbit model in LS-Prepost to generate

27 different models of varying anthropometries. Each of these models represents a

unique combination of the following parameters: orbital width and height, brow

protrusion angle, lateral and superior distance measurements, and eye protrusion. A

script was written that used a series of scaling, rotation, and translation operations to

transform the orbit model to the correct position in relation to the full eye model. The

orbit model was first aligned to match the orientation of the subject’s three-dimensional

CT reconstruction used to collect the orbital aperture measurements. The detailed

method of aligning the CT images and collecting measurements is reported separately,

but involved aligning the images with the crista galli, falx cerebri, and nasion-sella plane

to establish a landmark-defined coordinate system for each subject (Weaver AA et al.,

“CT Based Three-Dimensional Measurement of Orbit and Eye Anthropometry, submitted

to Invest Ophthalmol Vis Sci). Four coordinate points were selected on the orbit model

to be used during the transformation: the medial orbital rim node at the nasion, the lateral

38

orbital rim node horizontally across from the nasion, and the center orbital rim nodes

located superiorly and inferiorly.

First, the orbit was translated so that the point midway between the medial and

lateral nodes and midway between the superior and inferior nodes was located at the

origin. The orbit was globally scaled along the superior-inferior and medial-lateral axes

to model the orbital height and width. Figure 9 contrasts two orbits with equivalent brow

and eye protrusion that were scaled to model different size orbital apertures. Following

scaling, the orbit was then rotated about the medial-lateral axis at the inferior node to

model the appropriate brow protrusion angle. Figure 10 depicts orbits with differing

degrees of brow protrusion, but equal orbital apertures and eye protrusion. Next, the

orbit was translated in the medial-lateral and superior-inferior directions to locate the

orbit in the correct position relative to the eye model using the lateral and superior

distance measurements. The orbit was then translated in the superior-inferior, medial-

lateral, and anterior-posterior directions to locate the orbit in the correct position relative

to the eye model using the eye protrusion measurements. Models with differing degrees

of eye protrusion and equal orbital apertures and brow protrusion are contrasted in Figure

11. Translation of the orbit to model eye protrusion was the final step necessary in

creating the resulting orbit model to use in the simulation.

39

Figure 9. Small and large orbital apertures with incorporated eye model. Left

photo: Model 13 (Orbital width and height: 32.48, 26.82 mm); Right photo: Model 15 (Orbital width and height: 40.04, 36.40 mm).

Figure 10. Orbits with less and more brow protrusion with incorporated eye model. Left photo: Model 11 (Brow protrusion: 19.75 deg); Right photo: Model 17 (Brow

protrusion: 34.35 deg).

Figure 11. Models with less and more eye protrusion. Left photo: Model 5 (Eye

protrusion: 5.89 mm); Right photo: Model 14 (Eye protrusion: 12.06 mm).

40

After creating the 27 orbit models using the method described above, some

overlap between the eye and the inferior orbital rim was seen in two of the models.

These models had the smallest orbital aperture, least eye protrusion, and larger brow

protrusion angles. To adjust for the overlap in these two models, the orbits of Models 4

and 7 were manually translated 1 mm and 2.5 mm, respectively, in the inferior direction.

Simulations

After transforming all the orbit models according to the parameters in the test

matrix, a high speed blunt impact simulation was run in LS-Dyna (Livermore Software

Technology Corporation, Livermore, CA) for each orbit model with the incorporated eye

model. In each simulation, a Little League (CD25) baseball traveling at 30.1 m/s was

used to simulate a blunt impact to the eye and orbit. The ball was placed at a set distance

from the cornea of the eye so that it was not interacting with the eye or orbit prior to the

simulation. The same distance was used for all simulations, allowing for comparisons

regarding the timing of the peak stresses, strains, contact forces, and pressures. Contact

automatic single surface was utilized to form contacts between the eye and ball, the eye

and orbit, and the ball and orbit. Three separate contacts were established in order to

track the contact forces on these, as they were expected to vary with each simulation.

Soft=2 was utilized as it aided stability of the contacts. All contact parameters were the

same as in Stitzel et al, 200230. The orbit was modeled as a rigid body to provide contact

surfaces for the ball and eye. Most simulations terminated due to large deformation of

the ciliary body, but after the metric of interest, peak stresses, were reached in the

corneoscleral shell. Simulations were run to 1 ms to accommodate different timing of

41

interaction with the orbit and eye, a time duration well in excess of that needed to model

the entire event.

Stress and force data was filtered with SAE 3000 Hz, while strain and pressure

data was filtered with SAE 10000 Hz58. Within the corneoscleral shell, the maximum

principal stress and mid-surface principal strains in the meridional and circumferential

directions were obtained from the filtered data for each model. Peak contact forces

between the ball, eye, and orbit were recorded, and peak pressure was obtained by

averaging the peak pressure of four elements located at the center of the Eulerian portion

of the eye model. The timing of the peak stresses, strains, and contact forces was also

obtained. The first component (orbit, brow or cornea) contacted by the ball was recorded

for each simulation to characterize differences in interaction order between the ball, eye,

and orbit. Location of the peak stresses and strains in the corneoscleral shell was also

recorded.

Statistical Analysis

Multiple regression analyses were performed by analyzing the data on the basis of

independent variables orbital width (OW), brow protrusion (BP), and eye protrusion (EP)

and interaction variables OW*BP, BP*EP, OW*EP, OW*BP*EP. Orbital height was

paired with orbital width in each orbital model and therefore was not a separate

independent variable in the analysis. Thus, the OW factor can be considered to represent

the orbital aperture when analyzing results. Dependent variables were peak stress,

strains, contact forces, pressure, and timing of these peak values. The overall regression

model significance was assessed using the p-value obtained from the Analysis of

Variance (ANOVA) F-test. The adjusted R-squared value was computed to approximate

42

the variability in the response (dependent variable) that is accounted for by the factors

(independent and interaction variables) in the model. For example, an adjusted R-

squared value of 0.54 indicates the factors in the model account for 54% of the variation

in response. Parameter estimates were computed to predict the increase in response for a

unit increase in the factor. The t ratios were computed from the parameter estimates and

standard errors. A two-tailed t-test with an alpha of 0.05 was used to test the null

hypothesis that a particular parameter estimate was equal to zero. The null hypothesis is

rejected for p-values less than 0.05 and the associated factor was said to contribute

significantly to the model.

RESULTS

Results of the simulations are summarized in Table 5, Figure 12, and Figure 13.

Stress and pressure values exceeding the previously established globe rupture criteria

(stress threshold: 23 MPa, pressure threshold: 2.1 MPa) are bolded in Table 5.

Three simulations are depicted in Figure 14 to illustrate the effect of varying

anthropometry on the eye response. In Figure 14a, increased brow protrusion and

decreased eye protrusion resulted in a peak corneoscleral stress well below the globe

rupture criterion. Average orbital aperture, brow protrusion, and eye protrusion

parameters are modeled in Figure 14b, resulting in a peak stress of 21.9 MPa. The

simulation depicted in Figure 14c illustrates a model with a small orbital aperture, less

brow protrusion, and more eye protrusion that experienced a peak stress meeting the

globe rupture criterion. Compression of the eye against the orbit during the simulation is

also visible in Figure 14c.

43

Table 5. Magnitude and timing of stress, strains, contact forces, and pressure. Initial component contacted by ball: orbital brow (B) or cornea (C). Bolded stress

and pressure values exceeded the globe rupture criteria.

Stress Meridional Strain

Circum. Strain

Ball-Eye Force

Ball-Orbit Force

Orbit-Eye Force

Model Initial Ball

Contact Mag. (MPa)

Time (ms) Mag. Time

(ms) Mag. Time (ms)

Mag. (N)

Time (ms)

Mag. (N)

Time (ms)

Mag. (N)

Time (ms)

Pressure (MPa)

1 B 19.6 0.83 0.11 0.85 0.25 0.80 504 0.78 5009 0.84 169 0.75 1.4 2 C 21.8 0.83 0.13 0.81 0.30 0.81 494 0.75 4072 0.86 82 0.78 2.4 3 C 22.7 0.86 0.13 0.82 0.32 0.86 616 0.73 2756 0.86 0 0.86 2.2 4 B 17.8 0.90 0.04 0.85 0.20 0.85 385 0.90 6311 0.90 113 0.81 1.3 5 B 19.0 0.87 0.08 0.85 0.24 0.87 465 0.86 5195 0.86 108 0.86 1.2 6 B 20.5 0.86 0.11 0.82 0.28 0.82 454 0.73 4210 0.86 15 0.80 1.5 7 B 11.4 1.00 0.01 1.00 0.09 0.98 175 0.87 6384 0.93 17 1.00 0.6 8 B 18.0 0.84 0.03 0.79 0.23 0.89 402 0.92 5601 0.85 217 0.92 1.0 9 B 18.4 0.88 0.05 0.84 0.24 0.88 397 0.88 5112 0.88 78 0.82 1.4

10 C 21.8 0.83 0.11 0.80 0.33 0.83 616 0.73 2300 0.82 85 0.71 1.9 11 C 22.6 0.86 0.16 0.82 0.31 0.82 680 0.74 1633 0.86 16 0.86 2.2 12 C 22.7 0.86 0.14 0.82 0.32 0.84 678 0.74 851 0.85 0 0.48 1.4 13 B 23.8 0.87 0.16 0.82 0.38 0.87 427 0.74 4052 0.86 214 0.86 1.7 14 C 21.9 0.86 0.13 0.83 0.33 0.86 531 0.73 3153 0.86 84 0.86 1.9 15 C 22.7 0.86 0.14 0.83 0.32 0.86 626 0.72 2343 0.85 0 0.48 1.9 16 B 20.2 0.89 0.11 0.88 0.27 0.89 390 0.89 5255 0.89 179 0.89 1.3 17 B 20.8 0.87 0.13 0.86 0.31 0.87 418 0.77 4529 0.86 103 0.86 1.2 18 B 20.5 0.86 0.09 0.81 0.28 0.82 441 0.74 3861 0.85 2 0.85 1.7 19 C 23.0 0.84 0.17 0.84 0.33 0.84 748 0.77 548 0.84 232 0.82 2.4 20 C 21.0 0.83 0.15 0.83 0.30 0.83 672 0.75 18 0.85 77 0.85 2.1 21 C 21.1 0.82 0.12 0.82 0.30 0.80 668 0.74 0 0.56 0 0.80 1.7 22 C 18.0 0.76 0.16 0.76 0.26 0.76 697 0.74 1592 0.82 168 0.75 2.0 23 C 22.3 0.86 0.16 0.83 0.32 0.83 723 0.75 1338 0.86 64 0.75 2.0 24 C 22.2 0.86 0.15 0.82 0.31 0.83 665 0.74 739 0.85 0 0.85 1.9 25 B/C 21.6 0.85 0.12 0.82 0.30 0.83 482 0.73 3185 0.85 79 0.72 1.9 26 C 22.9 0.86 0.13 0.82 0.33 0.83 554 0.73 2795 0.86 11 0.70 2.2 27 C 22.8 0.86 0.14 0.82 0.33 0.83 646 0.73 2324 0.86 0 0.77 2.1

0

3

6

9

12

15

11-13 13-15 15-17 17-19 19-21 21-23 23-25Peak Corneoscleral Stress (MPa)

Num

ber

of S

imul

atio

ns

0

3

6

9

12

15

0-0.7 0.7-1.4 1.4-2.1 2.1+Peak Internal Eye Pressure (MPa)

Num

ber

of S

imul

atio

ns

Figure 12. Distribution of peak corneoscleral stresses in 27 simulations

Figure 13. Distribution of peak pressures observed in 27 simulations

44

0 3 6 9 12 15 18 21 24 27 30

Fringe Levels (MPa)

0 3 6 9 12 15 18 21 24 27 30

Fringe Levels (MPa) Figure 14. Select simulations depicting effect of anthropometric variation on eye response.

a) 11.4 MPa peak stress experienced by Model 7 (Orbital width and height: 32.48, 26.82 mm; Brow protrusion: 34.35 deg; Eye protrusion: 5.89 mm).

b) 21.9 MPa peak stress experienced by Model 14 (Orbital width and height: 36.26, 31.61 mm; Brow protrusion: 27.05 deg; Eye protrusion: 12.06 mm).

c) 23.0 MPa peak stress experienced by Model 19 (Orbital width and height: 32.48, 26.82 mm; Brow protrusion: 19.75 deg; Eye protrusion: 18.23 mm)

Figure 15 displays the element locations within the corneoscleral shell where the

peak stresses and strains occurred for the 27 simulations. High stress-strain elements

were typically located along the equator of the eye where equatorial expansion occurs.

Orbit model numbers are provided in the figure for each element location and can be

cross-referenced with Table 4 to investigate anthropometric trends. Peak stress and strain

45

locations appear to be grouped by eye protrusion. Of the nine orbit models with the most

eye protrusion, peak values were located in the superior, lateral quadrant of the eye in six

of these models (Models 19-24). These models also had least or average brow protrusion

values. High stresses and strains were located in the inferior, lateral quadrant of the eye

in seven of the nine orbit models with the least eye protrusion (Models 1-6, 9). All but

one of these models had least or average brow protrusion values. Peak values for all orbit

models with average eye protrusion were identified in the inferior half of the eye (Models

10-18).

Anterior

Superior

Inferior

Posterior

Medial

Lateral

19

2122

23 24

204

51

92

7

183,625

8

1513,16

1117

2614,27

1012

Anterior

Superior

Inferior

Posterior

Medial

Lateral

19

2122

23 24

204

51

92

7

183,625

8

1513,16

1117

2614,27

1012

Figure 15. Peak corneoscleral stress and strain element locations.

To investigate anthropometric effects, the average stress and pressure was

calculated for each anthropometric extreme (Table 6). The largest differences were seen

in the nine simulations with more eye protrusion (18.23 mm) which had stress and

46

pressure averages that were 2.9 MPa and 0.6 MPa higher than the nine simulations with

less eye protrusion (5.89 mm). Higher average stresses and pressures were also observed

in models with less brow protrusion and larger orbital apertures.

Table 6. Stress and pressure averages for anthropometric extremes.

Anthropometric Parameter Stress (MPa)

Pressure (MPa)

More eye protrusion (18.23 mm) 21.7 2.0 Less eye protrusion (5.89 mm) 18.8 1.4 Less brow protrusion (19.75 deg) 21.8 2.0 More brow protrusion (34.35 deg) 19.6 1.5 Larger orbital aperture (40.04, 36.40 mm) 21.5 1.8 Smaller orbital aperture (32.48, 26.82 mm) 19.7 1.6 Results of the multiple regression analysis are reported in Table 7. P-values

obtained from the ANOVA F-test indicate all regression models are statistically

significant except for the regression model for the orbit-eye force timing. Adjusted R-

squared values are reported in Table 7 indicating the percent variation in response that is

accounted for by factors in the regression model. The t ratios and p-values indicating the

statistical significance of the parameter estimates are reported. Factors with p-values less

than 0.05 were found to be significantly different from zero and contribute significantly

to the overall regression model.

47

Table 7. Multiple regression results: t ratio and p-value of each factor’s contribution to the regression model describing the response, and adjusted R-squared values

describing response variation accounted for by the regression model.

Factors Response Statistic OW BP EP OW*BP BP*EP OW*EP OW*BP*EP

Adj. R2

t ratio 2.23 -2.71 3.54 1.08 3.10 -1.55 -0.18 Stress p-value 0.038* 0.014* 0.002* 0.294 0.006* 0.137 0.861

0.54

t ratio 1.04 -4.36 6.55 0.66 2.95 -2.33 1.10 Merid. Strain p-value 0.310 <0.001* <0.001* 0.519 0.008* 0.031* 0.283

0.73

t ratio 1.71 -2.24 3.73 1.14 2.52 -1.85 -0.20 Circum. Strain p-value 0.104 0.037* 0.001* 0.268 0.021* 0.080 0.845

0.50

t ratio 3.57 -8.26 9.16 1.96 1.34 -2.00 0.93 Ball-Eye Force p-value 0.002* <0.001* <0.001* 0.065 0.197 0.060 0.362

0.87

t ratio -10.85 19.06 -28.00 0.77 2.65 3.59 -1.69 Ball-Orbit Force p-value <0.001* <0.001* <0.001* 0.450 0.016* 0.002* 0.107

0.98

t ratio -5.21 0.11 -0.75 1.60 -1.54 -1.51 -0.52 Orbit-Eye Force p-value <0.001* 0.910 0.462 0.126 0.140 0.147 0.608

0.52

t ratio 1.56 -4.27 5.33 2.31 3.62 -3.06 1.46 Pressure p-value 0.134 <0.001* <0.001* 0.032* 0.002* 0.006* 0.160

0.73

t ratio -0.51 2.89 -2.82 -1.91 -1.22 2.35 2.28 Stress Time p-value 0.614 0.009* 0.011* 0.072 0.239 0.030* 0.034*

0.49

t ratio -1.48 1.65 -1.97 -1.93 -1.60 2.32 1.84 Merid. Strain Time p-value 0.156 0.116 0.064 0.069 0.127 0.032* 0.081

0.39

t ratio -1.09 3.99 -4.04 -2.60 -3.50 1.30 3.12 Circum. Strain Time p-value 0.289 0.001* 0.001* 0.018* 0.002* 0.209 0.006*

0.69

t ratio -2.48 3.28 -4.58 -0.53 -3.64 1.36 -0.28 Ball-Eye Force Time p-value 0.023* 0.004* <0.001* 0.603 0.002* 0.189 0.782

0.64

t ratio -1.67 2.48 -2.48 0.93 1.36 -1.05 2.73 Ball-Orbit Force Time p-value 0.111 0.023* 0.023* 0.364 0.188 0.305 0.013*

0.43

t ratio -1.33 1.37 -1.30 -0.08 -1.70 0.57 1.19 Orbit-Eye Force Time p-value 0.201 0.187 0.208 0.936 0.105 0.577 0.248

0.10

* indicates statistical significance (p<0.05); Abbreviations: OW (Orbital Width), BP (Brow Protrusion), EP (Eye Protrusion) DISCUSSION

A wide range of peak stresses, strains, contact forces, and pressures were seen

among the 27 simulations, demonstrating the differences in eye response that exist when

varying orbital aperture, brow protrusion, and eye protrusion. Peak stresses in the

corneoscleral shell ranged from 11 to 24 MPa, with 2 of the 27 simulations exceeding the

23 MPa globe rupture prediction established for the model30. The peak internal eye

48

pressure ranged from 0.6 to 2.4 MPa, with 7 of the 27 simulations exceeding the 2.1 MPa

globe rupture prediction criterion. Only a few models exceeded the globe rupture

thresholds, but additional models experienced stresses and pressures near the thresholds

suggesting that the eye was in danger of rupturing in these simulations. Slightly more

than 50% of the simulations fell within the 21-23 MPa stress range and seven additional

models experienced pressures within 0.2 MPa of the pressure threshold.

Peak stresses and strains in the simulations were primarily located along the

equator of the eye where equatorial expansion occurs. Experimental studies have shown

the equator to be the principal location for globe rupture following internal pressurization

of the eye or impacts with blunt objects such as baseballs30, 59-61. Results of this study

and previous work suggest the weakest portion of the eye is the equator. Analysis of

anthropometric trends suggest that peak stresses and strains occur in the superior, lateral

quadrant of the eye for less protective orbits (those with greater eye protrusion and less

brow protrusion). Peak stresses and strains shift to the inferior half of the eye when there

is less eye protrusion.

The first component contacted by the ball during the simulation varied among

different anthropometries. The ball contacted the cornea first in 11 simulations and

contacted the brow first in 15 simulations. In one simulation, the ball contacted the

cornea and brow at approximately the same time. In simulations where the ball contacted

the cornea first, the peak stress and pressure averages (22 MPa and 2.0 MPa,

respectively) were near the globe rupture thresholds. Lower average peak stress and

pressure values (19 MPa and 1.3 MPa, respectively) were observed in simulations where

the ball contacted the brow first.

49

Contact between the eye and the orbit during the simulation often resulted in high

stresses and pressures as the deformable eye was compressed against the rigid orbit

during impact. Defining contact between the orbit and eye as an orbit-eye contact force

greater than 4 N, contact was identified in 20 of the 27 simulations (74%). Statistically,

OW was found to be the sole factor contributing to this orbit-eye contact. Since OW and

OH were paired, we conclude that the orbital aperture contributes to the contact seen

between the eye and the orbit during impact.

The multiple regression analysis showed the BP and EP factors each have a

significant effect on the stress, meridional and circumferential strains, ball-eye force,

ball-orbit force, and pressure. The t ratio signs suggest that a decrease in BP and an

increase in EP results in a decrease in ball-orbit force and an increase in stress, strain,

ball-eye force, and pressure. The OW factor was found to significantly affect stress and

the three contact forces. The t ratios suggest that an increase in orbital aperture results in

an increase in stress and ball-eye force and a decrease in ball-orbit and orbit-eye forces.

Collectively, these results suggest eye protection improves with increases in BP and

decreases in EP and OW. However, a decrease in OW may also result in increased orbit

and eye contact as the eye is compressed against the orbit during impact.

No interaction effects were detected for the ball-eye and orbit-eye contact forces,

suggesting that the above discussion of the main effects may adequately describe these

contact forces. However, interaction effects were found to be statistically significant for

several dependent variables, suggesting that particular combinations of the independent

variables are responsible for some of the response variability. BP*EP was the sole

interaction effect contributing to the stress and circumferential strain, suggesting

50

particular combinations, such as more brow protrusion and less eye protrusion, contribute

to the overall stress and circumferential strain in the eye. BP*EP and OW*EP interaction

effects contributed significantly to the meridional strain, ball-orbit force, and pressure,

and an additional interaction effect, OW*BP, contributed to the pressure response.

Timing of the peak responses was found to be affected by a variety of factors. BP

and EP factors were found to have a significant effect on the timing of the peak stresses,

circumferential strains, and ball-eye and ball-orbit contact forces. The OW factor was

found to have an additional effect on the timing of the ball-eye force. Interaction effects

were identified for every timing response except for the orbit-eye force timing,

suggesting that particular combinations of orbital aperture, brow protrusion, and eye

protrusion parameters contribute to the overall timing of peak stresses, strains, and

contact forces.

Failure is not incorporated into the eye model, thus the model was not capable of

actually rupturing. Also, the orbit model was a fixed rigid object that was not capable of

deforming. This presents a limitation because injuries such as orbital fractures are not

represented in the simulations. Although the eye model has been experimentally

validated to predict globe rupture, experimental eye impact tests with orbits of different

anthropometries could be conducted for validation purposes.

A baseball impact was simulated for this study, but future studies could

investigate the effect of orbital anthropometry on eye injury caused by different projectile

types. Baseball impacts are clinically relevant, as baseball is the leading cause of sports-

related eye injuries in young persons (ages 5-14)39. For the current study, a larger

projectile (Little League baseball) that interacted with the eye as well as the orbit and

51

surrounding facial regions was chosen to evaluate anthropometric effects on eye

response. Modeling of other projectile eye impacts could yield useful information

regarding orbital anthropometry, even if the projectile is small and interacts solely with

the eye. For instance, a small projectile such as a BB may not directly impact the orbit,

but contact between the eye and orbit during the simulation could radically differ among

orbits of different sizes and shapes. The eye might be more likely to be compressed into

a smaller orbit, producing high corneoscleral stresses.

Average anthropometric measurements and values two standard deviations from

the mean were used to establish parameters for the test matrix in this study and account

for orbital variation in the population. This ensured a wide range of values were

represented for each parameter investigated: orbital aperture, brow protrusion, and eye

protrusion. When plotted, nearly all the measured anthropometries did fall within the

range of anthropometries modeled in this study (Figure 7). While all anthropometries

modeled may not be representative of particular individuals in the population, they

allowed the averages and extremes of each parameter to be studied to investigate the

effect on eye injury. Orbital anthropometries well outside the norm do exist in the

population. Graves’ disease or an orbital tumor can cause exophthalmos, an excessive

protrusion of the eyeball as a result of an increase in orbital contents within a normal

orbit. Other conditions such as Crouzon syndrome can result in excessive protrusion of

the eye that is accompanied by decreased orbital capacity. Recession of the eyeball

within the orbit, endophthalmos, may result from a congenital anomaly, Horner’s

syndrome, or trauma. Additionally, congenital anomalies can produce other forms of

abnormal orbital anthropometry, such as a more or less prominent brow.

52

In summary, 27 different orbital anthropometries were modeled and eye impact

simulations were conducted to study the effect of orbital anthropometry on eye injury.

Results suggest the eye is more protected from impact with greater brow protrusion, less

eye protrusion, and smaller orbital apertures, provided the aperture is large enough to

deter contact between the orbit and the eye. The relationship between the anthropometric

variation and eye response is complex, as the statistical analysis reported significant

interaction effects that suggest some anthropometric parameters may have a synergistic

effect on eye response. Results of this study aid in explaining eye injury risk variation

across individuals and are valuable to the design of eye protection equipment and the

mitigation of eye injuries.

ACKNOWLEDGEMENTS

The authors wish to thank Mao Yu for programming and post-processing

assistance and Josh Tan of Center for Biomolecular Imaging for assistance with

TeraRecon software. The United States Army Aeromedical Research Laboratory funded

this effort.

53

REFERENCES

1. McGwin G, Jr., Xie A, Owsley C. Rate of eye injury in the United States. Arch Ophthalmol 2005;123(7):970-6. 2. Parver LM. Eye trauma. The neglected disorder. Arch Ophthalmol 1986;104(10):1452-3. 3. Smith D, Wrenn K, Stack LB. The epidemiology and diagnosis of penetrating eye injuries. Acad Emerg Med 2002;9(3):209-13. 4. Rao SK, Greenberg PB, Filippopoulos T, et al. Potential impact of seatbelt use on the spectrum of ocular injuries and visual acuity outcomes after motor vehicle accidents with airbag deployment. Ophthalmology 2008;115(3):573-6. 5. Lueder GT. Air bag-associated ocular trauma in children. Ophthalmology 2000;107(8):1472-5. 6. Kuhn F, Collins P, Morris R, Witherspoon CD. Epidemiology of motor vehicle crash-related serious eye injuries. Accid Anal Prev 1994;26(3):385-90. 7. Duma SM, Jernigan MV, Stitzel JD, et al. The effect of frontal air bags on eye injury patterns in automobile crashes. Arch Ophthalmol 2002;120(11):1517-22. 8. Duma SM, Kress TA, Porta DJ, et al. Airbag-induced eye injuries: a report of 25 cases. J Trauma 1996;41(1):114-9. 9. Duma SM, Rath AL, Jernigan MV, et al. The effects of depowered airbags on eye injuries in frontal automobile crashes. Am J Emerg Med 2005;23(1):13-9. 10. Muller-Jensen K, Allmaras W. [Eye injuries by safety glass (windshield)]. Hefte Unfallheilkd 1969;99:259-63. 11. Duma SM, Crandall JR. Eye injuries from airbags with seamless module covers. J Trauma 2000;48(4):786-9. 12. Fukagawa K, Tsubota K, Kimura C, et al. Corneal endothelial cell loss induced by air bags. Ophthalmology 1993;100(12):1819-23. 13. Anderson SK, Desai UR, Raman SV. Incidence of ocular injuries in motor vehicle crash victims with concomitant air bag deployment. Ophthalmology 2002;109(12):2356-8. 14. Lehto KS, Sulander PO, Tervo TMT. Do motor vehicle airbags increase risk of ocular injuries in adults? Ophthalmology 2003;110(6):1082-8. 15. Heier JS, Enzenauer RW, Wintermeyer SF, et al. Ocular injuries and diseases at a combat support hospital in support of Operations Desert Shield and Desert Storm. Arch Ophthalmol 1993;111(6):795-8. 16. Biehl JW, Valdez J, Hemady RK, et al. Penetrating eye injury in war. Mil Med 1999;164(11):780-4. 17. Mader TH, Aragones JV, Chandler AC, et al. Ocular and ocular adnexal injuries treated by United States military ophthalmologists during Operations Desert Shield and Desert Storm. Ophthalmology 1993;100(10):1462-7. 18. Mader TH, Carroll RD, Slade CS, et al. Ocular war injuries of the Iraqi insurgency, January-September 2004. Ophthalmology 2006;113(1):97-104. 19. Weichel ED, Colyer MH, Ludlow SE, et al. Combat Ocular Trauma Visual Outcomes during Operations Iraqi and Enduring Freedom. Ophthalmology 2008;115(12):2235-45.

54

20. Thach AB, Johnson AJ, Carroll RB, et al. Severe eye injuries in the war in Iraq, 2003-2005. Ophthalmology 2008;115(2):377-82. 21. Colyer MH, Chun DW, Bower KS, et al. Perforating Globe Injuries during Operation Iraqi Freedom. Ophthalmology 2008;115(11):2087-93. 22. Hecker S. More than half a million Americans suffer eye injuries from sports-related accidents: Lack of proper eye protection can lead to painful injuries, vision loss and even blindness. In: America PB, ed. Vision News. Chicago, 2007. 23. Cassen JH. Ocular trauma. Hawaii Med J 1997;56(10):292-4. 24. Rodriguez JO, Lavina AM, Agarwal A. Prevention and treatment of common eye injuries in sports. Am Fam Physician 2003;67(7):1481-8. 25. Chisholm L. Ocular injury due to blunt trauma. Appl Ther 1969;11(11):597-8. 26. Pardhan S, Shacklock P, Weatherill J. Sport-related eye trauma: a survey of the presentation of eye injuries to a casualty clinic and the use of protective eye-wear. Eye (Lond) 1995;9 ( Pt 6 Su):50-3. 27. Vinger PF, Sparks JJ, Mussack KR, et al. A program to prevent eye injuries in paintball. Sports Vision 1997;33:33-40. 28. Thach AB, Ward TP, Hollifield RD, et al. Ocular injuries from paintball pellets. Ophthalmology 1999;106(3):533-7. 29. Delori F, Pomerantzeff O, Cox MS. Deformation of the globe under high-speed impact: its relation to contusion injuries. Invest Ophthalmol 1969;8(3):290-301. 30. Stitzel JD, Duma SM, Cormier JM, Herring IP. A nonlinear finite element model of the eye with experimental validation for the prediction of globe rupture. Stapp Car Crash J 2002;46:81-102. 31. Kennedy EA, Ng, TP, Duma, SM. Evaluating eye injury risk of Airsoft pellet guns by parametric risk functions. Biomed Sci Instrum 2006;42:7-12. 32. Kennedy EA, Ng TP, McNally C, et al. Risk functions for human and porcine eye rupture based on projectile characteristics of blunt objects. Stapp Car Crash J 2006;50:651-71. 33. Scott WR, Lloyd WC, Benedict JV, Meredith R. Ocular injuries due to projectile impacts. Annu Proc Assoc Adv Automot Med 2000;44:205-17. 34. Green RP, Jr., Peters DR, Shore JW, et al. Force necessary to fracture the orbital floor. Ophthal Plast Reconstr Surg 1990;6(3):211-7. 35. Weidenthal DT, Schepens CL. Peripheral fundus changes associated with ocular contusion. Am J Ophthalmol 1966;62(3):465-77. 36. Weidenthal DT. Experimental Ocular Contusion. Arch Ophthalmol 1964;71:77-81. 37. McKnight SJ, Fitz J, Giangiacomo J. Corneal rupture following radial keratotomy in cats subjected to BB gun injury. Ophthalmic Surg 1988;19(3):165-7. 38. Duma SM, Ng TP, Kennedy EA, et al. Determination of significant parameters for eye injury risk from projectiles. J Trauma 2005;59(4):960-4. 39. Vinger PF, Duma SM, Crandall J. Baseball hardness as a risk factor for eye injuries. Arch Ophthalmol 1999;117(3):354-8. 40. Bisplinghoff JA, Duma SM. Evaluation of eye injury risk from projectile shooting toys using the focus headform. Biomed Sci Instrum 2009;45:107-12.

55

41. Kennedy EA, Inzana JA, McNally C, et al. Development and validation of a synthetic eye and orbit for estimating the potential for globe rupture due to specific impact conditions. Stapp Car Crash J 2007;51:381-400. 42. Power ED, Duma SM, Stitzel JD, et al. Computer modeling of airbag-induced ocular injury in pilots wearing night vision goggles. Aviat Space Environ Med 2002;73(10):1000-6. 43. Stitzel JD, Hansen GA, Herring IP, Duma SM. Blunt trauma of the aging eye: injury mechanisms and increasing lens stiffness. Arch Ophthalmol 2005;123(6):789-94. 44. Uchio E, Kadonosono K, Matsuoka Y, Goto S. Simulation of air-bag impact on an eye with transsclerally fixated posterior chamber intraocular lens using finite element analysis. J Cataract Refract Surg 2004;30(2):483-90. 45. Uchio E, Ohno S, Kudoh J, et al. Simulation model of an eyeball based on finite element analysis on a supercomputer. Br J Ophthalmol 1999;83(10):1106-11. 46. Uchio E, Ohno S, Kudoh K, et al. Simulation of air-bag impact on post-radial keratotomy eye using finite element analysis. J Cataract Refract Surg 2001;27(11):1847-53. 47. Kisielewicz LT, Kodama N, Ohno S, Uchio E. Numerical Prediction of Airbag Caused Injuries on Eyeballs After Radial Keratotomy. Society of Automotive Engineers 1998. 48. Kennedy EA, Stitzel, JD, Duma, SM. Matched experimental and computational simulations of paintball eye impacts. Biomed Sci Instrum 2008;44:243-8. 49. Ahmadi H, Shams PN, Davies NP, et al. Age-related changes in the normal sagittal relationship between globe and orbit. J Plast Reconstr Aesthet Surg 2007;60(3):246-50. 50. Barretto RL, Mathog RH. Orbital measurement in black and white populations. Laryngoscope 1999;109(7 Pt 1):1051-4. 51. Bolanos Gil de Montes F, Perez Resinas FM, Rodriguez Garcia M, Gonzalez Ortiz M. Exophthalmometry in Mexican adults. Rev Invest Clin 1999;51(6):341-3. 52. Dunsky IL. Normative data for hertel exophthalmometry in a normal adult black population. Optom Vis Sci 1992;69(7):562-4. 53. Pessa JE, Chen Y. Curve analysis of the aging orbital aperture. Plast Reconstr Surg 2002;109(2):751-5; discussion 6-60. 54. Pessa JE, Desvigne LD, Lambros VS, et al. Changes in ocular globe-to-orbital rim position with age: implications for aesthetic blepharoplasty of the lower eyelids. Aesthetic Plast Surg 1999;23(5):337-42. 55. Pivnick EK, Rivas ML, Tolley EA, et al. Interpupillary distance in a normal black population. Clin Genet 1999;55(3):182-91. 56. Pryor HB. Objective measurement of interpupillary distance. Pediatrics 1969;44(6):973-7. 57. van den Bosch WA, Leenders I, Mulder P. Topographic anatomy of the eyelids, and the effects of sex and age. Br J Ophthalmol 1999;83(3):347-52. 58. SAE. SAE J211-1: Instrumentation for Impact Test-Part 1-Electronic Instrumentation. SAE International, 2007. 59. Bisplinghoff JA, McNally C, Duma SM. High-rate internal pressurization of human eyes to predict globe rupture. Arch Ophthalmol 2009;127(4):520-3.

56

60. Kennedy EA, Voorhies KD, Herring IP, et al. Prediction of severe eye injuries in automobile accidents: static and dynamic rupture pressure of the eye. Annu Proc Assoc Adv Automot Med 2004;48:165-79. 61. Voorhies KD. Static and dynamic stress/strain properties for human and porcine eyes. Mechanical Engineering. Blacksburg: Virginia Tech, 2003; v. Master of Science.

57

Chapter 4: Evaluation of Different Projectiles in Matched Experimental Eye Impact Simulations

Submission planned for Journal of Biomechanics

Ashley A. Weaver1,2, Eric A. Kennedy4, Stefan M. Duma2,3, Joel D. Stitzel1,2

1Wake Forest University School of Medicine, Winston-Salem, NC 2Virginia Tech – Wake Forest University Center for Injury Biomechanics,

Winston-Salem, NC 3Virginia Polytechnic Institute and State University, Blacksburg, VA

4Bucknell University, Lewisburg, PA

This chapter is formatted according to the requirements of the journal to which the paper

is to be submitted. Ashley Weaver assisted with post-processing of the computational

simulations, performed the statistical analysis, and prepared the manuscript. Eric

Kennedy performed the experimental tests that were computationally modeled.

Computational modeling work was performed by Dr. Joel Stitzel. Dr. Joel Stitzel and Dr.

Stefan Duma also acted in an advisory and editorial capacity on this project.

58

ABSTRACT

Eye trauma results in 30,000 cases of blindness each year in the United States and

is the second leading cause of monocular visual impairment. Eye injury is caused by a

wide variety of projectile impacts and loading scenarios with common sources of trauma

being motor vehicle crashes, military operations, and sporting impacts. For the current

study, experimental eye impact tests in the literature were computationally modeled to

analyze global and localized responses of the eye to a variety of blunt projectile impacts.

Simulations were run with eight different projectiles (airsoft pellet, baseball, BB, blunt

impactor, paintball, aluminum, foam, and plastic rods) to characterize effects of the

projectile size, mass, geometry, material properties, and velocity on eye response. This

study presents a matched comparison of experimental test results and computational

model outputs including stress, energy, and pressure used to evaluate risk of eye injury.

In general, the computational results agreed with the experimental results. Locations of

peak stresses were located at the center of the cornea, the limbus, or the equator

depending on the type of projectile impacting the eye. Area-normalized kinetic energy

was a good predictor of peak stress and pressure. Additional incorporation of a relative

size parameter that relates the projectile area to the area of the eye reduced stress

response variability and may be of importance in eye injury prediction. Results of this

study are relevant to the design and regulation of safety systems and equipment to protect

against eye injury.

59

INTRODUCTION

Eye trauma results in over 9,000 globe ruptures, 30,000 cases of blindness, and

1.9 million eye injuries each year in the United States (McGwin et al., 2005; Parver,

1986; Smith et al., 2002). In motor vehicle crashes, severe eye injury can result when the

eye is impacted by an airbag, flying glass, or foam particles from the vehicle’s dashboard

(Duma and Crandall, 2000; Duma et al., 2002; Duma et al., 1996; Duma et al., 2005b;

Fukagawa et al., 1993; Kuhn et al., 1994; Lueder, 2000; Muller-Jensen and Allmaras,

1969; Rao et al., 2008). Approximately 66,000 people sustained vehicle-related eye

injuries in the United States in 2007 (McGwin et al., 2005). As many as 15% to 25% of

vehicle-related eye injuries are severe eye injuries (Duma and Crandall, 2000; Duma et

al., 2002), with up to 45% of these severe eye injuries being globe ruptures (Kuhn et al.,

1994). The rate of eye injuries in the military has dramatically increased this century,

with over 13% of all combat injuries in Operation Desert Storm being ocular injuries

(Heier et al., 1993). While penetrating eye injuries caused by shrapnel and other debris

are common in combat, nearly 25% of eye injuries result from blunt trauma sustained in

motor vehicle and helicopter crashes, falls, and impacts with blunt objects (Biehl et al.,

1999; Mader et al., 1993). More than 600,000 sports-related eye injuries occur each year

in the United States, with 40,000 of these requiring emergency care (Hecker, 2007).

Frequent agents of blunt trauma in sports include baseballs, golf balls, squash balls,

tennis balls, paintballs, and hockey sticks (Cassen, 1997; Chisholm, 1969; Pardhan et al.,

1995; Rodriguez et al., 2003; Vinger et al., 1997).

Experimental eye impact tests with a variety of blunt objects have previously been

conducted and used to determine injury tolerance of the eye (Delori et al., 1969; Duma

60

and Crandall, 2000; Duma et al., 2005a; Green et al., 1990; Kennedy et al., 2006a;

Kennedy et al., 2006b; McKnight et al., 1988; Scott et al., 2000; Stitzel et al., 2002;

Vinger et al., 1999; Weidenthal, 1964; Weidenthal and Schepens, 1966). Impacts were

performed on human and animal (porcine, monkey, and cat) eyes in these studies and

projectiles tested included foam particles, BBs, aluminum rods, plastic rods, blunt

impactors, baseballs, and paintballs. Experimental impact tests with the Facial and

Ocular Countermeasure Safety (FOCUS) headform have been used to predict eye injury

for BBs and projectile shooting toys (Bisplinghoff and Duma, 2009; Kennedy et al.,

2007). In these studies, a projectile was fired at the FOCUS simulated orbit and synthetic

eye and peak force was recorded by the FOCUS eye load cell for injury prediction

purposes.

Computational models of the eye have been used to simulate a variety of impacts

and analyze injury potential (Kennedy et al., 2008; Kisielewicz et al., 1998; Power et al.,

2002; Stitzel et al., 2002; Stitzel et al., 2005; Uchio et al., 2004; Uchio et al., 1999; Uchio

et al., 2001; Weaver et al., 2010). The Virginia Tech – Wake Forest University (VT-

WFU) Eye Model is a finite element model of the eye validated to predict globe rupture

for dynamic blunt eye impacts (Stitzel et al., 2002). This model was validated with 22

matched experimental tests and computational simulations of eye impacts with foam

projectiles, BBs, and baseballs. Computational simulations with the VT-WFU Eye

Model have been matched with experimental paintball tests to assess eye injury risk

(Kennedy et al., 2008). The eye model has been used in simulations to model airbag,

steering wheel, and foam particle impacts in motor vehicle crashes to assess the risk of

eye injury with aging and increasing lens stiffness (Stitzel et al., 2005). Baseball impact

61

simulations have also been conducted with the eye model to investigate effects of orbital

anthropometry on eye injury (Weaver et al., 2010).

Experimental studies have used kinetic energy, velocity, mass, and normalized

energy of the projectile to predict risk of eye injury (Berger, 1978; Duma and Crandall,

2000; Duma et al., 2005a; Kennedy et al., 2006b; Scott et al., 2000). In computational

models, injury can be assessed from localized stresses and strains the eye sustains during

an impact simulation (Kisielewicz et al., 1998; Power et al., 2002; Stitzel et al., 2002;

Stitzel et al., 2005; Uchio et al., 2004; Uchio et al., 1999; Uchio et al., 2001). For the

current study, experimental eye impact tests in the literature were computationally

modeled using the VT-WFU Eye Model to analyze global and localized responses of the

eye to a variety of blunt projectile impacts. Simulations were run with eight different

projectiles to characterize effects of the projectile size, mass, geometry, material

properties, and velocity on eye model response. This study presents a matched

comparison of experimental test results and computational model outputs including

stress, energy, and pressure that can be used to evaluate risk of eye injury.

METHODS

Existing data from the literature was compiled from blunt ocular impact studies

with human eyes (Delori et al., 1969; Kennedy et al., 2006b; Stitzel et al., 2002). A total

of eight projectile geometries were simulated to recreate the experiments, including

varying sizes of several geometries, for a total of 79 cases. The geometries and material

properties of projectiles modeled are summarized in Table 8. As depicted in Figure 16,

projectile shape was either spherical (baseball, paintball, airsoft pellet, and BB) or

cylindrical (blunt impactor, aluminum, plastic, and foam rods). A bar chart showing

62

diameter and mass of each projectile is shown in Figure 17 (note the logarithmic scale) to

demonstrate the relative size and weight of the various projectiles. The items varied in

diameter, mass and density as well as elastic modulus depending on the particular

material. For each projectile, a range of impact velocities is considered. A separate finite

element model was created for each projectile geometry using LS-Dyna software

(Livermore Software Technology Corporation, Livermore, CA) with the appropriate

material properties and running parameters.

Table 8. Summary of simulations test matrix.

Object # of

Cases Run

Shape Diameter (mm)

Mass (g)

Velocity Lower Limit (m/s)

Velocity Upper Limit (m/s)

Material Modulus (N/mm2)

Aluminum Rod (large) 4 Cylinder 11.16 5.19 43.71 53.21 Aluminum 70000

Aluminum Rod (small) 4 Cylinder 9.25 3.57 42.13 59.20 Aluminum 70000

Baseball 5 Sphere 76.10 146.50 30.10 42.80 Baseball 12

BB (large) 9 Sphere 4.50 0.38 53.00 122.40 Steel with copper 200000

BB (small) 5 Sphere 4.37 0.34 10.18 11.47 Steel with copper 200000

Blunt Impactor 3 Cylinder 19.90 112.55 8.53 8.72 Delrin 3000

Foam (large) 12 Cylinder 6.35 0.08 4.30 31.00 Foam 2.208

Foam (small) 6 Cylinder 4.50 0.04 38.06 47.23 Foam 2.208

Paintball 12 Sphere 17.30 3.13 65.50 112.50 Paintball 12

Plastic Rod (large) 4 Cylinder 9.75 0.69 17.83 20.16 HDPE 1000

Plastic Rod (small) 4 Cylinder 7.62 0.35 23.50 24.46 HDPE 1000

Airsoft Pellet (heavy) 7 Sphere 6.00 0.21 73.03 87.35 PVDF 1800

Airsoft Pellet (light) 4 Sphere 6.00 0.12 88.32 117.80 HDPE 1000

Total Runs 79

63

Figure 16. Geometry for sphere and cylinder projectiles.

0.01

0.10

1.00

10.00

100.00

1000.00

Alum. ro

d (lg)

Alum. ro

d (sm

)

Baseba

ll

BB (lg)

BB (sm)

Blunt Im

pacto

r

Foam (lg

)

Foam (s

m)

Paintba

ll

Plastic

Rod (lg

)

Plastic

Rod (s

m)

Airsoft

(heav

y)

Airsoft

(ligh

t)

Mass (g)Diameter (mm)

Figure 17. Projectile mass and diameter comparison.

Projectile impacts were simulated using an existing quarter-cylinder finite

element model of the eye validated to predict globe rupture with corneoscleral shell

stresses exceeding 23 MPa and local dynamic pressures exceeding 2.1 MPa (Stitzel et al.,

2002). The detailed eye model includes the corneoscleral shell (consisting of the cornea,

sclera, and limbus), as well as internal eye structures including the ciliary body, zonules,

and lens (Figure 18). The model has a total of 10,020 solid and shell elements and

64

includes Lagrangian and Eulerian meshes to accurately represent the mechanics of both

solid and fluid interactions. A box structure filled with a gelatin material was modeled to

represent the orbit and soft tissue surrounding the eye.

Gelatin

Anterior Chamber

Corneoscleral Shell

Lens

Zonules

Vitreous

Ciliary Body

Vitreous

Anterior Chamber

Orbit

Gelatin

Anterior Chamber

Corneoscleral Shell

Lens

Zonules

Vitreous

Ciliary Body

Vitreous

Anterior Chamber

Orbit

Gelatin

Anterior Chamber

Corneoscleral Shell

Lens

Zonules

Vitreous

Ciliary Body

Vitreous

Anterior Chamber

Orbit

Figure 18. Gelatin and orbit surrounding lagrangian-eulerian meshes of the eye

model. Lagrangian eye mesh (on left) depicting corneoscleral shell, lens, zonules, and ciliary body and Eulerian eye mesh (on right) depicting initially filled volume

(dark gray) and initially unfilled volume (light gray).

Dynamic modeling was performed using LS-Dyna software. In order to

efficiently analyze the large number of cases required and summarize the results more

easily, a simulation was set up utilizing Isight optimization software v. 9.0 (Engineous

Software, Cary, NC) in conjunction with LS-Dyna, Microsoft Excel, and Matlab (The

Mathworks, Inc, Natick, MA). Principal stress and strain data was collected at an interval

of 0.01 ms (a typical run is 0.3-0.6 ms). The maximum principal stress in the

corneoscleral shell was calculated for each case and the associated element location was

recorded. Pressure data was collected from four elements at the center of the Eulerian

portion of the eye model at an interval of 0.005 ms and filtered with SAE 10000 Hz

65

(SAE, 2007). An average of the maximum pressure in each of the four elements was

computed to represent the maximum pressure in each case.

Statistical analysis was performed using JMP software (SAS Institute Inc., Cary,

NC). A Student’s t-test was used to compare the mean stress and pressure in

computational simulations grouped by experimental globe rupture and projectile shape.

Oneway Analysis of Variance (ANOVA) was also used to compare mean stresses across

simulations grouped by the peak stress location. A multivariate correlation analysis was

used to identify projectile parameters that were significantly correlated with peak stress

and pressure. Finally, a multiple regression analysis was used to develop statistically

significant regression models and identify significant predictors for stress and pressure.

RESULTS

Results from all computational simulations are summarized in Table 13 in the

Appendix. Based on previously published globe rupture levels, the computational results

agreed well with the experimental results. A Student’s t-test revealed mean stresses and

pressures were significantly higher and exceeded the rupture thresholds in simulations

matched to experiments where globe rupture occurred (Table 9). Based on the 23 MPa

stress threshold for globe rupture, the computational results matched the experimental

results with a specificity and sensitivity of 1 and 0.46, respectively. For the 2.1 MPa

pressure threshold, computational and experimental results matched with a specificity of

0.94 and sensitivity of 0.54.

66

Table 9. Results of Student’s t-test comparing computational simulations grouped by experimental globe rupture.

Rupture No RuptureMetric, MPa Avg SD Avg SD t Ratio p-value

Stress 24.2 6.1 10.0 5.1 10.99 <0.0001 Pressure 2.2 0.8 0.5 0.5 11.05 <0.0001

A Student’s t-test was used to compare the mean stresses and pressures for

projectiles of different shapes. Taking only the projectile shape into account, the cases in

this study with spherical projectiles had significantly higher stresses and pressures

compared to the cylindrical projectile cases (Table 10).

Table 10. Results of Student’s t-test comparing projectile shape.

Cylinder Sphere Metric, MPa Avg SD Avg SD t Ratio p-value

Stress 9.8 5.3 18.9 8.8 5.48 <0.0001 Pressure 0.6 0.6 1.5 1.1 4.56 <0.0001

Peak stresses were located in different regions of the eye for different projectiles

(Figure 19). Peak stresses occurred in the center of the cornea for the BB and in the

limbus for the plastic, foam, and some large aluminum rods. Other projectiles (airsoft

pellet, paintball, blunt impactor, baseball, and most aluminum rods) resulted in peak

stresses located in the sclera near the equator due to equatorial expansion of the eye.

Oneway ANOVA identified a significantly lower average peak stress in the limbus (8.0

MPa) versus the cornea and the sclera (19.0 and 18.1 MPa, respectively). Principal stress

distribution in the corneosclearal shell at the time of peak stress is depicted in Figure 20

for each of the eight different projectiles.

67

Peak Stress Location

Alum. rod (lg)Alum. rod (sm)

BaseballBB (lg)

BB (sm)Blunt impactor

Foam (lg)Foam (sm)

PaintballPlastic rod (lg)

Plastic rod (sm)Softair (heavy)

Softair (light)

Cornea

Sclera

Airsoft (light)Airsoft (heavy)Plastic rod (sm)Plastic rod (lg)

PaintballFoam (sm)Foam (lg)

Blunt impactorBB (sm)BB (lg)

BaseballAlum. rod (sm)Alum. rod (lg)

Proj

ectil

e

Peak Stress Location

Alum. rod (lg)Alum. rod (sm)

BaseballBB (lg)

BB (sm)Blunt impactor

Foam (lg)Foam (sm)

PaintballPlastic rod (lg)

Plastic rod (sm)Softair (heavy)

Softair (light)

Cornea

Sclera

Airsoft (light)Airsoft (heavy)Plastic rod (sm)Plastic rod (lg)

PaintballFoam (sm)Foam (lg)

Blunt impactorBB (sm)BB (lg)

BaseballAlum. rod (sm)Alum. rod (lg)

Proj

ectil

e

Figure 19. Location of peak stresses in corneoscleral shell for different projectiles.

68

Foam rod, small diameter vel = 47.23 m/s, t = 0.16 ms

Blunt impactor vel = 8.62 m/s, t = 0.59 ms

Plastic rod, small diameter vel = 23.50 m/s, t = 0.12 ms

Airsoft pellet, heavy vel = 75.00 m/s, t = 0.32 ms

Baseball

vel = 30.10 m/s, t = 0.49 ms Aluminum rod, large diameter

vel = 53.21 m/s, t = 0.39 ms

Paintball

vel = 97.88 m/s, t = 0.29 ms BB, large diameter

vel = 91.70 m/s, t = 0.1 ms

Figure 20. Principal stress distribution, MPa, at time of peak stress for eight different projectiles. Note the different fringe levels for upper four projectiles versus

the lower four projectiles.

69

There are a number of comparisons to be made among the projectiles. Some of

the parameters to compare include the effect of size (mass and area), modulus, and

velocity. The peak principal stress for all the projectiles over their full velocity range is

shown in Figure 21 versus a normalized velocity scale (normalized by maximum velocity

for each projectile, V/Vmax). Linear relationships between peak stress and velocity are

clearly shown here, with the slope of the linear relationship varying depending on the

projectile.

Kinetic energy (KE) is a useful parameter to compare across cases because it

accounts for the mass and velocity of the projectile. In Figure 22, kinetic energy is

plotted on a logarithmic scale due to the large variation between the cases. Normalizing

the kinetic energy can be done a number of ways. It can be normalized by the maximum

kinetic energy for each case, KE/KEmax (Figure 23), or by the frontal area of the

projectile (Figure 24).

5

10

15

20

25

30

35

Max

Prin

cipa

l Stre

ss (M

Pa)

0 0.2 0.4 0.6 0.8 1 1.2Normalized Velocity (V/Vmax)

Pr

( g)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Airsoft (heavy)Airsoft (light)

5

10

15

20

25

30

35

Max

Prin

cipa

l Stre

ss (M

Pa)

0 0.2 0.4 0.6 0.8 1 1.2Normalized Velocity (V/Vmax)

Pr

( g)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Airsoft (heavy)Airsoft (light)

( g)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Airsoft (heavy)Airsoft (light)

Figure 21. Maximum principal stress versus normalized velocity for different

projectiles.

70

5

10

15

20

25

30

35

Max

Prin

cipa

l Stre

ss (M

Pa)

0.01 0.1 1 10 100 1000Kinetic Energy (J)

Pr

( g)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Airsoft (heavy)Airsoft (light)

5

10

15

20

25

30

35

Max

Prin

cipa

l Stre

ss (M

Pa)

0.01 0.1 1 10 100 1000Kinetic Energy (J)

Pr

( g)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Airsoft (heavy)Airsoft (light)

( g)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Airsoft (heavy)Airsoft (light)

Figure 22. Maximum principal stress versus kinetic energy (logarithmic scale) for

different projectiles.

5

10

15

20

25

30

35

Max

Prin

cipa

l Stre

ss (M

Pa)

0 0.2 0.4 0.6 0.8 1 1.2Normalized Kinetic Energy (KE/KEmax)

P

( g)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Airsoft (heavy)Airsoft (light)5

10

15

20

25

30

35

Max

Prin

cipa

l Stre

ss (M

Pa)

0 0.2 0.4 0.6 0.8 1 1.2Normalized Kinetic Energy (KE/KEmax)

P

( g)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Airsoft (heavy)Airsoft (light)

( g)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Airsoft (heavy)Airsoft (light)

Figure 23. Maximum principal stress versus maximum normalized kinetic energy.

71

5

10

15

20

25

30

35

Max

Prin

cipa

l Stre

ss (M

Pa)

10 100 1000 10000 100000Normalized Kinetic Energy (J/m2)

P

( g)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Airsoft (heavy)Airsoft (light)

5

10

15

20

25

30

35

Max

Prin

cipa

l Stre

ss (M

Pa)

10 100 1000 10000 100000Normalized Kinetic Energy (J/m2)

P

( g)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Airsoft (heavy)Airsoft (light)

( g)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Airsoft (heavy)Airsoft (light)

Figure 24. Maximum principal stress versus area-normalized kinetic energy

(logarithmic scale) for different projectiles.

Peak stress and pressure correlations with projectile parameters such as the

geometry, velocity, kinetic energy, and normalized kinetic energy were investigated and

statistical significance was assessed (Table 11). Area-normalized kinetic energy had the

highest Pearson correlation coefficients and was found to be a much better predictor of

peak stress and pressure than kinetic energy alone. This supports previous findings that

normalized kinetic energy is a better predictor of globe rupture (Duma et al., 2005a;

Kennedy et al., 2006b).

72

Table 11. Pearson product-moment correlation coefficients and p-values relating stress and pressure to projectile variables.

Stress (MPa) Pressure (MPa) Projectile Variable

Correlation P-value Correlation P-value Area (m2) 0.28 0.0110* 0.35 0.0017* Density (g/mm3) 0.32 0.0041* -0.03 0.8014 Depth (mm) 0.02 0.9294 0.04 0.8079 Diameter (m) 0.35 0.0018* 0.47 <0.0001* Kinetic Energy (J) 0.36 0.0012* 0.45 <0.0001* KE/KEmax -0.05 0.6685 -0.02 0.8820 KE/Relative size (J) 0.69 <0.0001* 0.62 <0.0001* Mass (g) 0.23 0.0421* 0.34 0.0023* Modulus (N/mm2) 0.28 0.0138* -0.08 0.4964 Norm. KE (J/m2) 0.89 <0.0001* 0.66 <0.0001* Relative size 0.43 0.0001* 0.64 <0.0001* Velocity (m/s) 0.69 <0.0001* 0.60 <0.0001* V/Vmax 0.03 0.7609 0.05 0.6765 Volume (mm3) 0.23 0.0444* 0.28 0.0129*

Significant correlations with stress and pressure included the projectile area,

diameter, mass, velocity, and kinetic energy, all of which are essentially accounted for in

the normalized kinetic energy parameter which was found to be the best single predictor

of peak stresses and pressure. Volume and the relative size (area of the projectile relative

to the area of the eye) were significantly correlated with stress and pressure and density

and modulus were significantly correlated with stress. A multiple regression analysis

was performed to regress each of the dependent variables (peak stress and pressure) on

the predictor variables (area-normalized kinetic energy, relative size, volume, density,

and modulus). The results of the ANOVA F-test revealed the overall regression model

was statistically significant (p-value < 0.0001). Area-normalized energy and relative size

were statistically significant predictor variables for stress and pressure, while volume was

also a significant predictor for pressure. Regression models were then created for stress

73

and pressure incorporating just the significant predictor variables. For stress, the

regression model with normalized energy and relative size variables had an R-square of

0.91, an improvement over the 0.79 R-square for normalized energy alone (Figure 25).

In the case of pressure, the R-square of the regression model with normalized energy,

relative size, and volume predictor variables was 0.85 versus 0.44 for the regression

model with normalized energy as a single predictor. This suggests relative size of the

projectile and volume (in the case of pressure) accounts for some of the variability in the

predicted and actual stress and pressure responses.

Table 12. Parameter estimates for predictor variables in the multiple regression model.

Stress (MPa) Pressure (MPa) Term Estimate Std Error t Ratio p-value| Estimate Std Error t Ratio p-value Intercept 4.94E+00 6.45E-01 7.66 <0.0001* 3.71E-02 9.69E-02 0.38 0.7034 Norm. KE (J/m2) 2.26E-04 1.18E-05 19.10 <0.0001* 1.87E-05 1.77E-06 10.52 <0.0001*Relative Size 1.23E+01 2.22E+00 5.53 <0.0001* 3.20E+00 3.34E-01 9.58 <0.0001*Vol (mm3) -5.58E-06 9.77E-06 -0.57 0.5694 -7.44E-06 1.47E-06 -5.07 <0.0001*Modulus (N/mm2) -2.49E-05 2.38E-05 -1.05 0.2989 -3.34E-07 3.58E-06 -0.09 0.9259 Density (g/mm3) 9.46E+02 6.58E+02 1.44 0.1552 -2.31E+01 9.89E+01 -0.23 0.8159

0

10

20

30

Max

imum

Prin

cipa

l Stre

ss (M

Pa) A

ctua

l

0 10 20 30Max Principal Stress (MPa) Predicted

P<.0001 RSq=0.91

( g)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

0

10

20

30

Max

imum

Prin

cipa

l Stre

ss (M

Pa) A

ctua

l

0 10 20 30Max Principal Stress (MPa) Predicted

P<.0001 RSq=0.91

( g)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

( g)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)

Alum. rod (lg)Alum. rod (sm)BaseballBB (lg)BB (sm)Blunt impactorFoam (lg)Foam (sm)PaintballPlastic rod (lg)Plastic rod (sm)Softair (heavy)Softair (light)

Figure 25. Maximum principal stress actual versus predicted plot with 95%

confidence curves for multiple regression model with area-normalized energy and relative size predictor variables.

74

DISCUSSION

This study determined the stress and pressure response of the eye through

computational modeling of a variety of projectiles and loading conditions, providing a

wealth of new computational technology and injury information for the eye model. In

general, the computational results agreed strongly with the matched experimental results,

further validating the ability of the eye model to predict globe rupture in diverse loading

conditions. Computational simulations matched to experiments where globe rupture

occurred experienced significantly higher stresses and pressures with a mean response

that exceeded the previously published stress and pressure globe rupture levels (Stitzel et

al., 2002). In comparison, mean stress and pressure fell below the globe rupture levels in

simulations matched to experiments where no globe rupture occurred.

Comparing the experimental and computational results for each case, specificity

was high and virtually no simulations falsely predicted a globe rupture, indicating a low

type I error rate. There is room for improvement in the sensitivity of the stress and

pressure thresholds to predict globe rupture. Further inspection revealed seven of the

simulations that falsely predicted no globe rupture would occur were those with an

aluminum rod projectile. In four simulations with a large diameter BB projectile, the

peak stress correctly predicted a globe rupture would occur, but peak pressures fell below

the 2.1 MPa threshold. Finally, in two baseball simulations and three paintball

simulations that experienced globe rupture, the peak stress was less than 1 MPa below the

23 MPa threshold. Perhaps computational results nearing the rupture thresholds should

be more closely inspected and a percent chance of injury may be a more appropriate

predictor of globe rupture. Overall, improvements could be instituted to improve globe

75

rupture prediction in the aluminum rod simulations, but stress appears to sufficiently

predict globe rupture in the remaining simulations.

The statistical analysis revealed spherical projectiles (baseball, BB, paintball,

airsoft pellet) resulted in higher stresses and pressures in the eye compared to cylindrical

projectiles (blunt impactor, aluminum, foam, and plastic rods). However, there are likely

other factors besides the projectile shape affecting the stresses and pressure observed.

Further investigation controlling for the projectile material, mass, and velocity is

warranted to fully characterize the effect of projectile shape on eye response.

Principal stress distribution in the eye varied greatly depending on the type of

projectile impacting the eye. High stresses near or exceeding the globe rupture thresholds

occurred either in the center of the cornea or the equator of the sclera. For a small, hard

object with a high velocity, such as the BB, the center of the cornea experiences the

highest stresses and is the most likely region of the eye to be injured in this type of

impact scenario. In general, for larger, high velocity projectiles with a substantial mass

(baseball, paintball, airsoft pellet, blunt impactor, and aluminum rods), injury is more

likely to occur at the equator of the sclera due to equatorial expansion. Projectiles with

less mass and lower velocity (foam, plastic, and some aluminum rods) resulted in peak

stresses that were well below the globe rupture thresholds and occurred near the limbus

of the eye.

In general, the computational results support well the notion that area-normalized

kinetic energy of the projectile is a much better predictor of peak stress in the

corneoscleral shell than kinetic energy alone. It is from a computational standpoint a

much better predictor of peak stress and of globe rupture. This is evidenced by the

76

tighter grouping of stress data when plotted with area-normalized energy versus kinetic

energy. Maximum principal stresses versus area-normalized kinetic energy slopes are

more similar between different projectiles than they are when stress is plotted against

normalized velocity or peak-normalized kinetic energy. Normalized energy accounts for

the mass, velocity, and frontal area of the projectile and essentially reflects the amount of

energy presented to a given area of the eye. However, the amount of normalized energy

presented to the eye is underpredicted for the baseball simulations since the baseball is

larger in diameter than the eye. Thus, normalized energy-based estimation of injury risk

is low (1-18%) for the baseball impacts that resulted in globe rupture experimentally and

experienced peak stresses near or exceeding 23 MPa. In addition to normalized energy,

incorporating a relative size parameter that relates the projectile area to the area of the

eye reduced variability in the stress response and may be of importance in eye injury

prediction. Addition of this parameter may be particularly important for large projectile

impacts such as the baseball.

The eye model or the projectiles were not capable of failing in the simulations,

presenting a limitation to the study. Also, simplified models of the orbit and surrounding

soft tissue were used. Orbital anthropometry was not modeled at this time, but has been

investigated for the baseball projectile in a previous study (Weaver et al., 2010). All of

the stress data presented in this report is unfiltered and conclusions were made based on

assumptions about where in the trace to take the data. High frequency activity was not

observed in the majority of the simulations. An exception was the airsoft pellets, which

created artificially high stresses early on in the event. These high frequency stresses are

likely non-physical in nature, resulting in artificially high levels of stress in the model

77

over very short time durations. In the case of the airsoft pellets, the peak stresses

followed a more recognizable pattern (the lower frequency activity) later in the event, and

the peak stresses achieved toward the end of the event were used for comparison between

different test severities.

Results presented in this study have important implications in the prevention and

mitigation of ocular trauma in the civilian and military sectors. The variety of projectiles

and loading conditions modeled provide information on the corneoscleral stress

distribution and vitreous pressure during blunt eye impacts commonly seen in motor

vehicle crashes, military operations, and sports-related impacts. Compared to

experimental tests, the computational simulations quantify in more detail the local and

global responses of eye during impact. Stress and pressure data obtained from the

simulations could be useful in the design and validation of anthropometric test devices

(ATDs) that predict eye injury such as the FOCUS headform with an eye load cell

(Kennedy et al., 2007). Vitreous pressure data could be useful in designing a fluid-filled

eye with pressure sensor, while corneoscleral stress data could be used to design an ATD

that measures load at multiple regions across the frontal surface of the eye. In

conclusion, agreement of the computational and experimental results in the current study

validates the ability of the eye model to predict injury under a variety of loading

scenarios and illustrates the value of using the eye model in future studies to model other

impact scenarios.

ACKNOWLEDGEMENTS

The United States Army Aeromedical Research Laboratory provided funding for

this work.

78

REFERENCES

Berger, R. E., 1978. A model for evaluating the ocular contusion injury potential of propelled objects. J Bioeng 2, 345-58.

Biehl, J. W., Valdez, J., Hemady, R. K., Steidl, S. M., and Bourke, D. L., 1999. Penetrating eye injury in war. Mil Med 164, 780-4.

Bisplinghoff, J. A., and Duma, S. M., 2009. Evaluation of eye injury risk from projectile shooting toys using the focus headform. Biomed Sci Instrum 45, 107-12.

Cassen, J. H., 1997. Ocular trauma. Hawaii Med J 56, 292-4. Chisholm, L., 1969. Ocular injury due to blunt trauma. Appl Ther 11, 597-8. Delori, F., Pomerantzeff, O., and Cox, M. S., 1969. Deformation of the globe under high-

speed impact: its relation to contusion injuries. Invest Ophthalmol 8, 290-301. Duma, S. M., and Crandall, J. R., 2000. Eye injuries from airbags with seamless module

covers. J Trauma 48, 786-9. Duma, S. M., Jernigan, M. V., Stitzel, J. D., Herring, I. P., Crowley, J. S., Brozoski, F. T.,

and Bass, C. R., 2002. The effect of frontal air bags on eye injury patterns in automobile crashes. Arch Ophthalmol 120, 1517-22.

Duma, S. M., Kress, T. A., Porta, D. J., Woods, C. D., Snider, J. N., Fuller, P. M., and Simmons, R. J., 1996. Airbag-induced eye injuries: a report of 25 cases. J Trauma 41, 114-9.

Duma, S. M., Ng, T. P., Kennedy, E. A., Stitzel, J. D., Herring, I. P., and Kuhn, F., 2005a. Determination of significant parameters for eye injury risk from projectiles. J Trauma 59, 960-4.

Duma, S. M., Rath, A. L., Jernigan, M. V., Stitzel, J. D., and Herring, I. P., 2005b. The effects of depowered airbags on eye injuries in frontal automobile crashes. Am J Emerg Med 23, 13-9.

Fukagawa, K., Tsubota, K., Kimura, C., Hata, S., Mashita, T., Sugimoto, T., and Oguchi, Y., 1993. Corneal endothelial cell loss induced by air bags. Ophthalmology 100, 1819-23.

Green, R. P., Jr., Peters, D. R., Shore, J. W., Fanton, J. W., and Davis, H., 1990. Force necessary to fracture the orbital floor. Ophthal Plast Reconstr Surg 6, 211-7.

Hecker, S., 2007. More than half a million Americans suffer eye injuries from sports-related accidents: Lack of proper eye protection can lead to painful injuries, vision loss and even blindness. In: Vision News. (P. B. America, Ed.), pp. 1-2, Chicago.

Heier, J. S., Enzenauer, R. W., Wintermeyer, S. F., Delaney, M., and LaPiana, F. P., 1993. Ocular injuries and diseases at a combat support hospital in support of Operations Desert Shield and Desert Storm. Arch Ophthalmol 111, 795-8.

Kennedy, E. A., Inzana, J. A., McNally, C., Duma, S. M., Depinet, P. J., Sullenberger, K. H., Morgan, C. R., and Brozoski, F. T., 2007. Development and validation of a synthetic eye and orbit for estimating the potential for globe rupture due to specific impact conditions. Stapp Car Crash J 51, 381-400.

Kennedy, E. A., Ng, T. P., and Duma, S. M., 2006a. Evaluating eye injury risk of Airsoft pellet guns by parametric risk functions. Biomed Sci Instrum 42, 7-12.

Kennedy, E. A., Ng, T. P., McNally, C., Stitzel, J. D., and Duma, S. M., 2006b. Risk functions for human and porcine eye rupture based on projectile characteristics of blunt objects. Stapp Car Crash J 50, 651-71.

79

Kennedy, E. A., Stitzel, J. D., and Duma, S. M., 2008. Matched experimental and computational simulations of paintball eye impacts. Biomed Sci Instrum 44, 243-248.

Kisielewicz, L. T., Kodama, N., Ohno, S., and Uchio, E., 1998. Numerical Prediction of Airbag Caused Injuries on Eyeballs After Radial Keratotomy. Society of Automotive Engineers.

Kuhn, F., Collins, P., Morris, R., and Witherspoon, C. D., 1994. Epidemiology of motor vehicle crash-related serious eye injuries. Accid Anal Prev 26, 385-90.

Lueder, G. T., 2000. Air bag-associated ocular trauma in children. Ophthalmology 107, 1472-5.

Mader, T. H., Aragones, J. V., Chandler, A. C., Hazlehurst, J. A., Heier, J., Kingham, J. D., and Stein, E., 1993. Ocular and ocular adnexal injuries treated by United States military ophthalmologists during Operations Desert Shield and Desert Storm. Ophthalmology 100, 1462-7.

McGwin, G., Jr., Xie, A., and Owsley, C., 2005. Rate of eye injury in the United States. Arch Ophthalmol 123, 970-6.

McKnight, S. J., Fitz, J., and Giangiacomo, J., 1988. Corneal rupture following radial keratotomy in cats subjected to BB gun injury. Ophthalmic Surg 19, 165-7.

Muller-Jensen, K., and Allmaras, W., 1969. [Eye injuries by safety glass (windshield)]. Hefte Unfallheilkd 99, 259-63.

Pardhan, S., Shacklock, P., and Weatherill, J., 1995. Sport-related eye trauma: a survey of the presentation of eye injuries to a casualty clinic and the use of protective eye-wear. Eye (Lond) 9 ( Pt 6 Su), 50-3.

Parver, L. M., 1986. Eye trauma. The neglected disorder. Arch Ophthalmol 104, 1452-3. Power, E. D., Duma, S. M., Stitzel, J. D., Herring, I. P., West, R. L., Bass, C. R.,

Crowley, J. S., and Brozoski, F. T., 2002. Computer modeling of airbag-induced ocular injury in pilots wearing night vision goggles. Aviat Space Environ Med 73, 1000-6.

Rao, S. K., Greenberg, P. B., Filippopoulos, T., Scott, I. U., Katsoulakis, N. P., and Enzer, Y. R., 2008. Potential impact of seatbelt use on the spectrum of ocular injuries and visual acuity outcomes after motor vehicle accidents with airbag deployment. Ophthalmology 115, 573-576.

Rodriguez, J. O., Lavina, A. M., and Agarwal, A., 2003. Prevention and treatment of common eye injuries in sports. Am Fam Physician 67, 1481-8.

SAE, 2007. SAE J211-1: Instrumentation for Impact Test-Part 1-Electronic Instrumentation. SAE International.

Scott, W. R., Lloyd, W. C., Benedict, J. V., and Meredith, R., 2000. Ocular injuries due to projectile impacts. Annu Proc Assoc Adv Automot Med 44, 205-17.

Smith, D., Wrenn, K., and Stack, L. B., 2002. The epidemiology and diagnosis of penetrating eye injuries. Acad Emerg Med 9, 209-13.

Stitzel, J. D., Duma, S. M., Cormier, J. M., and Herring, I. P., 2002. A nonlinear finite element model of the eye with experimental validation for the prediction of globe rupture. Stapp Car Crash J 46, 81-102.

Stitzel, J. D., Hansen, G. A., Herring, I. P., and Duma, S. M., 2005. Blunt trauma of the aging eye: injury mechanisms and increasing lens stiffness. Arch Ophthalmol 123, 789-94.

80

Uchio, E., Kadonosono, K., Matsuoka, Y., and Goto, S., 2004. Simulation of air-bag impact on an eye with transsclerally fixated posterior chamber intraocular lens using finite element analysis. J Cataract Refract Surg 30, 483-90.

Uchio, E., Ohno, S., Kudoh, J., Aoki, K., and Kisielewicz, L. T., 1999. Simulation model of an eyeball based on finite element analysis on a supercomputer. Br J Ophthalmol 83, 1106-11.

Uchio, E., Ohno, S., Kudoh, K., Kadonosono, K., Andoh, K., and Kisielewicz, L. T., 2001. Simulation of air-bag impact on post-radial keratotomy eye using finite element analysis. J Cataract Refract Surg 27, 1847-53.

Vinger, P. F., Duma, S. M., and Crandall, J., 1999. Baseball hardness as a risk factor for eye injuries. Arch Ophthalmol 117, 354-8.

Vinger, P. F., Sparks, J. J., Mussack, K. R., Dondero, J., and Jeffers, J. B., 1997. A program to prevent eye injuries in paintball. Sports Vision 33, 33-40.

Weaver, A. A., Loftis, K. L., Duma, S. M., and Stitzel, J. D., 2010. Biomechanical Modeling of Eye Trauma for Different Orbit Anthropometries. Ophthalmology (submitted for publication).

Weidenthal, D. T., 1964. Experimental Ocular Contusion. Arch Ophthalmol 71, 77-81. Weidenthal, D. T., and Schepens, C. L., 1966. Peripheral fundus changes associated with

ocular contusion. Am J Ophthalmol 62, 465-77.

81

APPENDIX

Table 13. Simulation results

Experimental Test Computational Simulation

Projectile Velocity (m/s)

Norm. KE (J/m2)

Risk of Globe Rupture Globe rupture?

Max Principal

Stress (MPa)

Max Stress Time (ms)

Pressure (MPa)

Alum. rod (lg) 43.71 50646 98% yes 18.77 0.39 1.64 Alum. rod (lg) 44.10 51557 98% yes 19.00 0.35 1.59 Alum. rod (lg) 44.69 52958 99% yes 19.17 0.34 1.76 Alum. rod (lg) 53.21 75075 100% yes 21.02 0.39 2.28 Alum. rod (sm) 42.13 47227 95% yes 17.21 0.41 1.57 Alum. rod (sm) 43.58 50531 98% yes 17.68 0.40 1.57 Alum. rod (sm) 44.10 51752 98% yes 17.80 0.40 1.67 Alum. rod (sm) 50.40 67585 100% yes 19.02 0.39 1.62 Baseball 30.10 14591 0% no 20.50 0.49 1.34 Baseball 34.40 19057 1% yes 22.14 0.45 2.12 Baseball 35.50 20296 2% yes 22.65 0.47 2.42 Baseball 41.20 27336 11% yes 24.96 0.42 2.57 Baseball 42.80 29501 18% yes 25.42 0.42 2.45 BB (lg) 53.00 33116 35% no 17.07 0.12 0.68 BB (lg) 53.80 34123 41% no 17.44 0.11 0.76 BB (lg) 55.80 36707 58% no 17.85 0.14 0.76 BB (lg) 59.70 42018 84% no 19.42 0.12 1.00 BB (lg) 71.93 56930 100% yes 24.49 0.11 1.01 BB (lg) 85.20 85579 100% yes 30.06 0.10 1.18 BB (lg) 90.40 96344 100% yes 31.42 0.10 1.25 BB (lg) 91.70 99135 100% yes 31.56 0.10 1.35 BB (lg) 122.40 176624 100% yes 46.11 0.08 2.13 BB (sm) 10.18 1179 0% no 5.49 0.30 0.07 BB (sm) 10.98 1370 0% no 5.98 0.28 0.10 BB (sm) 11.19 1425 0% no 6.12 0.28 0.12 BB (sm) 11.37 1469 0% no 6.19 0.28 0.14 BB (sm) 11.47 1496 0% no 6.26 0.28 0.13 Blunt impactor 8.53 13167 0% no 9.32 0.60 0.61 Blunt impactor 8.62 13435 0% no 9.35 0.59 0.56 Blunt impactor 8.72 13761 0% no 9.45 0.58 0.51 Foam (lg) 4.30 22 0% no 1.68 0.34 0.03 Foam (lg) 5.40 35 0% no 2.06 0.30 0.04 Foam (lg) 6.00 44 0% no 2.27 0.28 0.04 Foam (lg) 10.60 137 0% no 3.65 0.23 0.07 Foam (lg) 14.20 245 0% no 4.46 0.21 0.12 Foam (lg) 14.30 249 0% no 4.51 0.21 0.12 Foam (lg) 18.90 434 0% no 5.71 0.19 0.13 Foam (lg) 23.00 643 0% no 6.48 0.18 0.15

82

Continued Experimental

Test Computational Simulation

Projectile Velocity (m/s)

Norm. KE (J/m2)

Risk of Globe Rupture Globe rupture?

Max Principal

Stress (MPa)

Max Stress Time (ms)

Pressure (MPa)

Foam (lg) 26.70 867 0% no 7.25 0.18 0.19 Foam (lg) 26.80 873 0% no 7.23 0.18 0.18 Foam (lg) 28.60 994 0% no 7.49 0.17 0.20 Foam (lg) 31.00 1168 0% no 7.70 0.16 0.24 Foam (sm) 38.06 1780 0% no 7.42 0.17 0.18 Foam (sm) 41.16 2081 0% no 7.60 0.17 0.17 Foam (sm) 43.65 2341 0% no 7.79 0.16 0.20 Foam (sm) 44.66 2450 0% no 7.87 0.16 0.20 Foam (sm) 45.73 2569 0% no 7.94 0.16 0.17 Foam (sm) 47.23 2740 0% no 8.05 0.16 0.19 Paintball 65.50 28604 14% yes 22.11 0.38 1.97 Paintball 68.02 31370 26% no 22.35 0.38 2.18 Paintball 70.77 34032 41% no 22.85 0.37 2.10 Paintball 71.10 34461 43% no 22.83 0.37 2.39 Paintball 71.28 34613 44% yes 22.88 0.38 2.51 Paintball 71.73 35070 47% yes 22.88 0.37 2.48 Paintball 72.00 35478 50% yes 22.94 0.36 2.40 Paintball 73.28 36645 57% yes 23.20 0.37 2.22 Paintball 75.29 38684 69% yes 23.47 0.36 2.64 Paintball 97.88 65030 100% yes 26.74 0.29 3.84 Paintball 108.00 79493 100% yes 28.34 0.27 3.37 Paintball 112.50 86503 100% yes 29.04 0.26 4.29 Plastic rod (lg) 17.83 1460 0% no 9.60 0.12 0.39 Plastic rod (lg) 19.23 1699 0% no 9.79 0.11 0.42 Plastic rod (lg) 19.26 1704 0% no 9.72 0.11 0.40 Plastic rod (lg) 20.16 1867 0% no 9.86 0.11 0.43 Plastic rod (sm) 23.50 2089 0% no 9.66 0.12 0.35 Plastic rod (sm) 23.75 2134 0% no 9.69 0.11 0.39 Plastic rod (sm) 23.96 2171 0% no 9.76 0.11 0.37 Plastic rod (sm) 24.46 2264 0% no 9.79 0.11 0.34 Airsoft (heavy) 73.03 19694 2% no 11.18 0.32 0.78 Airsoft (heavy) 75.00 20591 2% no 11.30 0.32 0.74 Airsoft (heavy) 76.29 20330 2% no 11.35 0.31 0.65 Airsoft (heavy) 78.30 22032 3% no 11.52 0.32 0.69 Airsoft (heavy) 81.19 23981 5% no 11.78 0.32 0.74 Airsoft (heavy) 85.33 26292 9% no 12.05 0.32 0.79 Airsoft (heavy) 87.35 27849 12% no 12.16 0.32 0.89 Airsoft (light) 88.32 15987 1% no 10.52 0.30 0.60 Airsoft (light) 91.62 17531 1% no 10.67 0.30 0.85 Airsoft (light) 92.12 17364 1% no 10.67 0.30 0.71 Airsoft (light) 93.63 17798 1% no 10.69 0.30 0.65

83

Summary of Research The research presented in this thesis has yielded significant contributions to the

field of injury biomechanics, particularly to the field of ocular trauma. The research

outlined in this thesis has fulfilled the following objectives:

1. A systematic method utilizing anatomical landmarks has been developed to

align head CT images and collect orbit and eye anthropometric measurements

to quantify variation in the normal population.

2. Twenty-seven models of varying orbital anthropometries have been developed

and used in computational simulations to demonstrate the relationship

between orbital anthropometry and the biomechanical response of the eye

upon impact.

3. A variety of projectile experimental impact tests have been computationally

modeled and the relationship between projectile parameters and the

biomechanical response of the eye has been demonstrated.

Research outlined in chapters two through four is expected to be published in

scientific journals and presented at relevant scientific conferences as shown (Table 14).

Table 14. Publication plan for research outlined in this thesis.

Chapter Topic Journal / (Conference) 2 CT Based Three-Dimensional

Measurement of Orbit and Eye Anthropometry

Investigative Ophthalmology and Visual Science†

3 Biomechanical Modeling of Eye Trauma for Different Orbit Anthropometries

Ophthalmology*

4 Evaluation of Different Projectiles in Matched Experimental Eye Impact Simulations

Journal of Biomechanics (American Society of Biomechanics)

†Accepted *Submitted

84

Scholastic Vita Ashley Anne Weaver Graduate Student & Research Engineer Virginia Tech-Wake Forest University School of Biomedical Engineering and Sciences Office: Virginia Tech-Wake Forest University Center for Injury Biomechanics Medical Center Boulevard, MRI 2 Winston-Salem, North Carolina 27157 Work: (336) 716-0944; Cell: (828) 446-3998

Email: asweaver@wfubmc.edu

EDUCATION 2008-Present

Wake Forest University, M.S. and Ph.D. Candidate, Biomedical Engineering Virginia Tech-Wake Forest University Center for Injury Biomechanics Advisor: Dr. Joel Stitzel

2004-2008 North Carolina State University, B.S., Biomedical Engineering (Minors: Genetics, Biological Sciences) Summa Cum Laude, Valedictorian, GPA 4.0 Senior Project: Pressure-shifting device for geriatric chair for the prevention of pressure sores.

PROFESSIONAL EMPLOYMENT 2008-present Research Engineer Virginia Tech - Wake Forest Center for Injury Biomechanics Wake Forest University School of Medicine Winston-Salem, North Carolina 2005-2008 Draftsman Ruggles Engineering, PC Catawba, North Carolina Summer 2007 Research Intern Virginia Tech - Wake Forest Center for Injury Biomechanics Wake Forest University School of Medicine Winston-Salem, North Carolina 2007 Engineering Intern Gilero, LLC Raleigh, North Carolina 2005-2006 Teaching Assistant North Carolina State University, Physics Department Raleigh, North Carolina

85

PROFESSIONAL CERTIFICATION AND WORKSHOPS 1. Biomedical Engineering Society Student Leadership Workshop, Pittsburgh, PA,

October 2009. 2. Biomedical Engineering Society Student Chapter Development Workshop,

Pittsburgh, PA, October 2009. 3. Materialise Innovation Course, Ann Arbor, MI, February 2009. 4. Fundamentals of Engineering, Certification received, 2008.

COMPUTING SKILLS Amira, Carecast, ImageJ, JMP, LS-Dyna, Matlab, Microsoft Office, Mimics, TeraRecon

LEADERSHIP POSITIONS BMES National Student Affairs Committee 2010-present VT-WFU BMES Chapter Co-President 2009-present NCSU BMES Chapter President 2007-2008 NCSU Engineering World Health Chapter President 2007-2008 NCSU BMES Chapter Community Service Chair 2006-2007 NCSU Engineering World Health Chapter Treasurer 2006-2007

STUDENT ORGANIZATION MEMBERSHIPS Biomedical Engineering Society, VT-WFU chapter 2009-present IEEE Engineering in Medicine and Biology Society, WFU chapter 2008-present University Scholars Program, NCSU 2004-2008 Biomedical Engineering Society, NCSU chapter 2004-2008 Genetics Club, NCSU 2006-2008 Engineering World Health, NCSU chapter 2006-2008

PROFESSIONAL MEMBERSHIPS AND SERVICE Biomedical Engineering Society 2006-present Society of Automotive Engineers 2009-present IEEE 2009-present Triad Biotech Alliance 2009-present Honor Society of Phi Kappa Phi 2006-2008 Engineering World Health 2006-2008

AWARDS AND HONORS 1. National Science Foundation Graduate Research Fellowship, April 2010.

2. Student Travel Award, Ohio State University Injury Biomechanics Symposium, May 2009.

3. Student Paper Award, 3rd place - Weaver AA, Gilmartin TD, Anz AW, Stubbs AJ, Stitzel JD. “A Method to Measure Acetabular Metrics from Three Dimensional Computed Tomography Pelvis Reconstructions.” 46th Rocky Mountain Bioengineering Symposium, Milwaukee, WI, April 2009.

86

4. Best Clinical Research Award - Anz AW, Frino J, Lang JE, Weaver AA, Stitzel JD, Stubbs AJ. “Do the Classical Measurements for Acetabular Dysplasia Delineate Volume: A Correlation with CT Reconstruction?” Wake Forest University Department of Orthopaedic Surgery Resident Research Day, Winston-Salem, NC, April 2009.

5. Virginia Tech-Wake Forest University School for Biomedical Engineering and Sciences Graduate Fellowship, 2008-present.

6. Valedictorian and Summa Cum Laude, North Carolina State University, 2008.

7. Service and Citizenship Award, Biomedical Engineering Department, North Carolina State University, 2008.

8. Semester Dean’s List 8 of 8 semesters, North Carolina State University, 2004-2008.

9. Health Advocates Alliance Scholarship Recipient, private scholarship based on academic merit and character, 2008.

10. Dean’s Merit Scholarship, merit-based College of Engineering scholarship, North Carolina State University, 2004-2008.

11. Summer Research Opportunities Program Fellowship, undergraduate research fellowship, Wake Forest University, 2007.

12. Thomas Jackson Martin Scholarship, merit-based College of Engineering scholarship, North Carolina State University, 2004-2005.

13. Salutatorian, Bandys High School, 2004.

BIBLIOGRAPHY

Papers in Refereed Journals 1. Weaver AA, Loftis KL, Tan JC, Duma SM, Stitzel JD. “CT Based Three-

Dimensional Measurement of Orbit and Eye Anthropometry.” Paper accepted for publication by Investigative Ophthalmology and Visual Science, April 2010.

2. Weaver AA, Loftis KL, Duma SM, Stitzel JD. “Biomechanical Modeling of Eye Trauma for Different Orbit Anthropometries.” Paper submitted to Ophthalmology, March 2010.

3. Anz AW, Frino J, Lang JE, Weaver AA, Stitzel JD, Stubbs AJ. “Traditional Radiographic Measurements of Acetabular Hip Dysplasia Do Not Correlate with Three-Dimensional Computed Tomographic Reconstructions of Volume and Surface Areas.” Paper submitted to CORR, February 2010.

Papers in Refereed Conference Publications 1. Stitzel JD, Kilgo PD, Weaver AA, Loftis KL, Martin RS, Meredith JW. “Age Thresholds for

Increased Mortality of Predominant Crash Induced Thoracic Injuries.” Paper submitted to Annu Proc Assoc Adv Automot Med., April 2010.

87

2. Weaver AA, Gilmartin TD, Anz AW, Stubbs AJ, Stitzel JD. “A Method to Measure Acetabular Metrics from Three Dimensional Computed Tomography Pelvis Reconstructions.” Biomed Sci Instrum. 2009;45: 155-60.

3. Weaver AA, Gayzik FS, Stitzel JD. “Biomechanical Analysis of Pulmonary Contusion in Motor Vehicle Crash Victims: A Crash Injury Research and Engineering Network (CIREN) Study.” Biomed Sci Instrum. 2009;45: 364-9.

Other Conference Papers (abstract style or non-refereed)/Scientific Exhibits 1. Weaver AA, Gilmartin TD, Anz AW, Stubbs AJ, Stitzel JD. “A Method to Measure

Acetabular Metrics from 3D Computed Tomography Pelvis Reconstructions.” Biomedical Engineering Society Annual Fall Scientific Meeting, Pittsburgh, PA, October 2009.

2. Kim H, Weaver AA, Stitzel JD. “Approach For Measuring Cortical Thickness of the Ribs Across Populations Using Computed Tomography.” Biomedical Engineering Society Annual Fall Scientific Meeting Poster Session, Pittsburgh, PA, October 2009.

3. Colonna AL, Ennis TM, Martin RS, Weaver AA, Crane D, Mowery NT, Stitzel JD, Hoth JJ. “Percent of Pulmonary Contusion Predicts Development of ARDS.” The American Association for the Surgery of Trauma 68th Annual Meeting Poster Session, Pittsburgh, PA, October 2009.

4. Anz AW, Frino J, Lang JE, Weaver AA, Stitzel JD, Stubbs AJ. “Do the Classical Measurements for Acetabular Dysplasia Delineate Volume: A Correlation with CT Reconstruction?” North Carolina Orthopaedic Association 2009 Annual Meeting, Pinehurst, NC, September 2009.

5. Kim H, Weaver AA, Stitzel JD. “Approach For Measuring Cortical Thickness of the Ribs Across Populations Using Computed Tomography.” Summer Research Opportunities Program Poster Session, Wake Forest University Baptist Medical Center, Winston-Salem, NC, July 2009.

6. Anz AW, Frino J, Lang JE, Weaver AA, Stitzel JD, Stubbs AJ. “Do the Classical Measurements for Acetabular Dysplasia Delineate Volume: A Correlation with CT Reconstruction?” Eastern Orthopaedic Association 40th Annual Meeting, Paradise Island, Bahamas, June 2009.

7. Weaver AA, Gayzik FS, Stitzel JD. “Biomechanical Analysis of Pulmonary Contusion in Motor Vehicle Crash Victims: A Crash Injury Research and Engineering Network (CIREN) Study.” Virginia Tech-Wake Forest University School of Biomedical Engineering and Sciences Symposium Poster Session, Virginia Tech, Blacksburg, VA, May 2009.

8. Weaver AA, Gayzik FS, Stitzel JD. “Biomechanical Analysis of Pulmonary Contusion in Motor Vehicle Crash Victims: A Crash Injury Research and Engineering Network (CIREN) Study.” Ohio State University Injury Biomechanics Symposium Poster Session, Ohio State University, Columbus, OH, May 2009.

88

9. Anz AW, Frino J, Lang JE, Weaver AA, Stitzel JD, Stubbs AJ. “Do the Classical Measurements for Acetabular Dysplasia Delineate Volume: A Correlation with CT Reconstruction?” Wake Forest University Department of Orthopaedic Surgery Resident Research Day, Winston-Salem, NC, April 2009.

10. Ruggles AA, Gayzik FS, Stitzel JD. “Biomechanical Analysis of Pulmonary Contusion in Motor Vehicle Crash Victims: A Crash Injury Research and Engineering Network (CIREN) Study.” Summer Research Opportunities Program Poster Session, Wake Forest University Baptist Medical Center, Winston-Salem, NC, July 2007.

Additional Presentations at Professional Meetings/Conferences 1. Weaver AA, Loftis KL, Duma SM, Stitzel JD. “Modeling Human Variation: Orbit

Anthropometry and Effect on Eye Injury Metrics.” Advanced Technologies and New Frontiers in Injury Biomechanics with Military and Aerospace Applications, Washington DC, August 2009.

2. Ruggles AA, Stuenkel E, West A, Perkins L, Srinivas R. “Pressure-shifting Prototype for a Geriatric Chair to Aid in the Prevention of Pressure Sores.” North Carolina State University/University of Chapel Hill Biomedical Research Design Symposium, Research Triangle Park, NC, April 2008.

Technical Reports 1. Stitzel, JD, Weaver AA, Loftis KL, Yu, MM. “Development of a Non-symmetric

Finite Element Model of the Eye and Development of an Anthropometrically Accurate Model of the Orbit for Countermeasure Effectiveness Evaluation.” Report to the United States Army Aeromedical Research Laboratory, Ft. Rucker, Alabama, June 2009.