Development of a Full-Body CAD Dataset for a Computational Model of the 5th Percentile Female By Matthew Logan Davis 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 2014 Winston-Salem, North Carolina Approved by: F. Scott Gayzik, PhD, Advisor, Chair Examining Committee: Kerry A. Danelson, PhD Joel D. Stitzel, PhD ACKNOWLEDGEMENTS I would first like to thank my advisor, Dr. Scott Gayzik, for all of his guidance, encouragement, and for consistently challenging me to improve and grow as a researcher. I am very thankful to be working with you and I am looking forward to the next phase of our project. I would also like to thank the other members of my committee, Dr. Joel Stitzel and Dr. Kerry Danelson, for all of your advice and aid with this project. I would like to extend a special thanks to the members of the GHBMC team: Nick, Doron, and Jeremy. You all provide a great support system and are always willing to help in any way possible. I appreciate all the time you have spent helping me complete this work. For their aid in helping me get on my feet when I first arrived, I would like to thank Dan Moreno and Ashley Hayes. You were both very instrumental in bringing me up to speed and taught me a great deal about human body modeling. I would also like to thank all of the members of the Center for Injury Biomechanics. At some point, you have each aided the efforts of this project and shared your expertise with me. It truly is an honor working with such talented researchers. Most importantly, I would like to thank my family for all of their support. To my parents, I am thankful for all of the love and commitment you have given to me help me achieve my goals. To Logan, I am very grateful for all of the sacrifices you have made and for your daily encouragement. ii TABLE OF CONTENTS Acknowledgements ........................................................................................................... ii Table of Contents ............................................................................................................. iii List of Tables ..................................................................................................................... v List of Figures ................................................................................................................... vi Abstract ........................................................................................................................... viii CHAPTER I: INTRODUCTION AND BACKGROUND ............................................ 1 Road Traffic Injuries and Human Body Modeling ....................................................... 1 Motivation for Small Female Models .............................................................................. 2 Existing Small Female Finite Element Models ............................................................... 3 Global Human Body Models Consortium (GHBMC) ................................................... 6 Chapter Summaries .......................................................................................................... 8 Chapter II: A Multi-Modality Image Set for the Development of a 5th Percentile Female Finite Element Model ......................................................................................... 8 Chapter III: Characterization of Abdominal Organ Volume and Large Vasculature Diameter for the 5th Percentile Female ........................................................................... 8 Chapter IV: A Technique for Developing CAD Geometry of Long Bones Using Clinical CT Data ............................................................................................................. 8 References .......................................................................................................................... 9 CHAPTER II: A MULTI-MODALITY IMAGE SET FOR THE DEVELOPMENT OF A 5TH PERCENTILE FEMALE FINITE ELEMENT MODEL ......................... 12 Abstract ............................................................................................................................ 13 2.1 Introduction .............................................................................................................. 14 2.2 Methods ...................................................................................................................... 17 2.2.1 Medical Imaging Protocol.................................................................................... 17 2.2.2 External Anthropometry ...................................................................................... 20 2.2.3 Segmentation........................................................................................................ 22 2.2.4 Assembly.............................................................................................................. 24 2.3 Results ........................................................................................................................ 26 2.4 Discussion................................................................................................................... 29 iii 2.5 Conclusions ................................................................................................................ 34 2.6 Acknowledgement ..................................................................................................... 34 2.7 Appendix .................................................................................................................... 35 2.8 References .................................................................................................................. 41 CHAPTER III: CHARACTERIZATION OF THORACOABDOMINAL ORGAN VOLUMES FOR THE 5TH PERCENTILE FEMALE ............................................... 45 Abstract ............................................................................................................................ 46 3.1 Introduction ............................................................................................................... 48 3.2 Methods ...................................................................................................................... 51 3.3 Results ........................................................................................................................ 56 3.4 Discussion................................................................................................................... 58 3.5 Conclusion ................................................................................................................. 62 3.6 References .................................................................................................................. 63 3.7 Appendix .................................................................................................................... 68 CHAPTER IV: A TECHNIQUE FOR DEVELOPING CAD GEOMETRY OF LONG BONES USING CLINICAL CT DATA ........................................................... 69 Abstract ............................................................................................................................ 70 4.1 Introduction ............................................................................................................... 72 4.2 Methods ...................................................................................................................... 76 4.3 Results ........................................................................................................................ 86 4.4 Discussion................................................................................................................... 90 4.5 Conclusion ................................................................................................................. 94 4.6 References .................................................................................................................. 96 CHAPTER V: CONCLUSION .................................................................................... 103 CHAPTER VI: SCHOLASTIC VITA ........................................................................ 105 iv LIST OF TABLES Table 1. Parameter values for medical image acquisition ................................................ 19 Table 2. Anthropometric measurements of the 5th percentile female volunteer ............. 27 Table 3. Volumetric organ data of the F05 from segmentation ....................................... 29 Table 4. Bone cortical thicknesses comparison to literature ............................................ 33 Table 5- 3D Rendering of Organs Used for Volume Calculations ................................... 55 Table 6- Height, Weight, and Slice Thickness for Each Subject ...................................... 57 Table 7- Volume and Diameter Data for Each Subject ................................................... 68 Table 8- Table of cortical bone thickness values from the literature. ............................... 82 Table 9- Bone Length and Maximum Cortical Thickness ................................................ 89 v LIST OF FIGURES Figure 1. CAD Representation of the GHBMC 50th Percentile Male Occupant ............... 7 Figure 2. GHBMC 50th Percentile Male Finite Element Model ........................................ 7 Figure 3. Examples of the Multi-Modality Image Set, a) Axial slice of supine CT, b) Axial slice of supine MRI, c) Axial slice of seated uMRI ................................................ 20 Figure 4. Schematic of model assembly process, a) Cervical spine masks from supine CT segmentation, b) 3D polygon surface data from segmentation placed in the quasi-seated CT image space, c) Cervical spine is transformed as a group into the model space and aligned with the skull and thoracic spine using bony landmarks (blue). .......................... 25 Figure 5. Skeletal structures segmented from supine MRI data (outlined in red) placed in the uMRI image space. ..................................................................................................... 25 Figure 6. Assembled 3D F05 Occupant. Images show the full skeleton and all organs that will be explicitly represented. Bony landmark data is shown as blue points (Table A3). ................................................................................................................................... 29 Figure 7. Comparison of F05 model assembly to F05 Hybrid III finite element model . 31 Figure 8. Masks of Liver, Spleen, Left Kidney, Right Kidney, and Heart. Images are ordered from superior to inferior. ..................................................................................... 54 Figure 9- Sagittal View of Aorta Measurements. From left to right, the images depict the superior aorta, inferior portion of the aortic arch, and the aorta at the superior level of L3. ........................................................................................................................................... 56 vi Figure 10- (Left) Average organ volumes, (Right) Average diameters for the vena cava and aorta ............................................................................................................................ 58 Figure 11- Comparison of F05 organ volumes to Geraghty et al. Organ volumes were normalized to the published target 5th percentile organ volume. ...................................... 60 Figure 12- Schematic of cortical thickness reconstruction process of the femur. (A) Initial segmentation (B) CAD rendering of the femur based on initial segmentations ............... 77 Figure 13- Process for determining characteristic cortical thickness values. (Exterior bone surface is transparent) (A) Mid-shaft with cut plane perpendicular to Z axis (B) Isometric view of interior and exterior contours of the isolated cross section of bone C) Top view of isolated cross section of bone showing exemplar outer and inner radius ......................... 79 Figure 14- Evaluation of bone transition regions about the Z axis. Z Projection, Z Transition, and Z Scan are representative of the regions of bone captured by the processes used in this study. The inner and outer radius at a representative location is shown. Radial locations used to evaluate the transition distance are also shown. ........................ 84 Figure 15- Characteristic cortical thickness values of long bones in regions above the scanner cut-off of 2.75 mm. Data for the male and female cortical thickness ranges were taken from Virtama, et al. ................................................................................................. 87 vii ABSTRACT Motor vehicle injuries and fatalities remain a leading public health concern worldwide. In 2013, the World Health Organization reported more than 1.2 million deaths as a result of motor vehicle crashes (MVCs). MVCs are also estimated to have a yearly economic cost of nearly $145 billion. To mitigate the societal impact of these crashes, researchers are using a variety of tools, including finite element (FE) models, to develop and evaluate vehicle safety devices. Such models are often developed to represent an average (50th percentile by height and weight) male occupant. However, a 2011 study from the University of Michigan Transportation Research Institute marked the first year that there were more licensed female drivers than male drivers in the United States. In order to predict the response of a greater portion of the vehicle occupants beyond the 50th percentile male, there is interest in developing such models for other cohorts among the driving public. This work focuses on the female driver in the 5th percentile of height and weight. In the past, 5th percentile female models have primarily been scaled from existing average male models, since scan data for such a specific target anthropometry is limited. However, as part of the Global Human Body Models Consortium (GHBMC) project, comprehensive image and anthropometrical data of a 5th percentile female was acquired. The long range objective of this project is to develop a detailed 5th percentile female full body finite element model in two postures: standing (pedestrian) and seated (vehicle occupant). This thesis describes the first phase of this effort, in which this multi-modality dataset is used to develop a whole-body set of CAD data for the 5th percentile female. viii Chapter I: Introduction and Background ROAD TRAFFIC INJURIES AND HUMAN BODY MODELING According to the 2013 World Health Organization’s (WHO) Report on Road Safety, road traffic injuries are the 8th leading cause of death globally [1]. This translates to more than 1.2 million deaths worldwide each year. With the increasing number of global motorists, the WHO also predicts road traffic injuries to become the 5th leading cause of worldwide death by 2030. Motor vehicle crashes (MVCs) also have significant economic costs, with a reported $518 billion being spent each year due to injuries sustained during traffic accidents. In the United States alone, MVCs account for over 30 thousand deaths, 1.5 million injuries, and $230 billion in economic costs each year [2]. To mitigate the toll of these injuries, a variety of techniques have been used to both develop and evaluate safety features for motor vehicles. One of the emerging tools for this purpose is the finite element method. Using this method, full body computational models can be used to conduct motor vehicle crash simulations with the purpose of observing how the model interacts with the vehicle and how it performs with specific safety devices, such as seatbelts and airbags. By developing high biofidelity models, researchers gain the ability to investigate localized tissue responses to impacts, thereby improving the understanding of how the body behaves during dynamic impact events and potentially providing new insights into injury mechanisms. Traditionally, human body models are developed to represent an average male (50th percentile in terms of height and weight). While these models can provide a valuable assessment of the mid-sized adult male, real world motor vehicle crashes 1 involve occupants of various size, age, and gender. This provides motivation to expand the suite of high biofidelity human body models available to researchers. MOTIVATION FOR SMALL FEMALE MODELS Historically, the National Highway Traffic Safety Administration’s (NHTSA) Federal Motor Vehicle Safety Standard (FMVSS) for the protection of occupants in the United States required the used of 50th percentile male dummies for its frontal crash tests. However, reports of air bag fatalities and other serious injuries in otherwise low severity impacts raised concerns about the effectiveness of safety devices for protecting occupants of other sizes and seat positions. One of the common factors in each of these air bag related deaths was proximity to the steering wheel. Also, stature, age, and gender seemed to play a role. NHTSA has cited that 78% of drivers who sustained fatal injuries due to air bag deployment in minor to moderate vehicle crashes were female. Of the females who died in these crashes, 82% were less than 163 cm tall [3]. As a result, the use of a small female ATD is currently required for various regulatory tests such as a frontal impact, side-pole impact, and a side-barrier impact. According to a recent study observing gender trends in motorists, there are now more licensed female drivers than male drivers in the United States [4]. Thus, it is important to represent the changing demographic. Small females are considered the most vulnerable driving population; therefore, models of these occupants can be used to provide a means of conservatively estimating the performance of safety devices. Previous research has shown that females are considered to be at a larger risk of sustaining injury during automotive accidents when compared to males [5,6]. Based on a comparison of drivers in the United States and the United Kingdom using the National 2