You are on page 1of 222

COMPUTATIONAL MODELING OF HIP JOINT MECHANICS

by Andrew Edward Anderson

A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

Department of Bioengineering The University of Utah April 2007

Copyright © Andrew Edward Anderson 2007 All Rights Reserved

subjectspecific FE models of the pelvis and entire hip joint were developed. Unfortunately. Quantifying the relevant mechanical parameters in the joint (i. It can be credibly argued that prior computational models of the hip joint do not have the ability to predict cartilage and bone mechanics with sufficient accuracy for clinical application. The etiology of hip OA is unknown however. was assessed using phantom . abnormal. hip finite element models to date have used simplified geometry and have not been validated. such as the finite element method. cortical bone and cartilage thickness. The accuracy of model geometry.e. due to the high loads this joint supports. While its efficient structure may lend to a lifetime of mobility. open the possibility of predicting hip joint biomechanics noninvasively and could revolutionize the way pathological hips are diagnosed and treated. Toward this objective. repetitive loading of the hip is thought to result in osteoarthritis (OA). Experimental studies have provided valuable insight into baseline hip joint biomechanics but they require a protocol that is inherently invasive. The aim of this dissertation is to develop and validate methods that will facilitate patient-specific modeling of hip joint biomechanics. Numerical modeling techniques. mechanics have been implicated as the primary factor.ABSTRACT The hip joint is one of the largest weight bearing structures in the human body.e. i. cartilage and bone stresses) appears to be central to an enhanced understanding of this disease.

including improved diagnosis of pathology. FE predictions were compared directly with experimental data for purposes of validation. patientspecific approaches to treatment.and post-operative mechanics. The developed modeling methodologies have a number of potential longer-term uses and benefits.based imaging studies. the validated modeling protocols were extended to analyze patient-specific models to demonstrate the general feasibility of the approach and to quantify differences in hip joint biomechanics between a normal and dysplastic hip joint. Finally. and prediction of the success rate of corrective surgeries based on pre. The sensitivity of the models to errors in assumed and measured model inputs was quantified. recognizing that acetabular dysplasia may be the single most important contributor of hip joint OA. v .

Anderson “It matters not how long we live but how” Philip James Bailey .To my father: Richard E.

.....................................................25 Bone Material Properties .................................... INTRODUCTION ............5 References......................................................................................................................................................................................................................................................................................... Muscles........................................................... iv LIST OF TABLES......................................................TABLE OF CONTENTS ABSTRACT...........................................................................35 Computational Modeling of the Hip Joint ..... ix LIST OF FIGURES ............................................................................................1 Research Goals.....................................20 Osteoarthritis....................29 In-Vivo Studies of Hip Joints .........................................................................32 Numerical Modeling of Hip Joint Biomechanics ...................................................................................... Ligaments..... BACKGROUND ..................10 Forward ...........22 Experimental Hip Joint Biomechanics.................................................20 Hip Dysplasia..............................................................................................................................................................................................................................43 ...........4 Summary of Chapters .....................................................................................11 Cartilage.....................10 Hip Joint Structure and Function ...................................................................................8 2........................................................................................................................................................................................................................... xiii CHAPTER 1......................................................................................13 Labrum...25 Cartilage Material Properties ..x ACKNOWLEDGEMENTS...............................................................................................1 Motivation........................................................................................................36 References...............................................................11 Pelvic and Femoral Bone ...................................................................................................................................................................................................................15 Hip Joint Pathology........................................................35 Analytical Modeling of the Hip Joint ..................................................................................................27 In-Vitro Studies of Hip Joints .............. Capsule................................

..........110 Contrast Enhanced Scans................................102 Phantom Description............................76 Cortical Bone Thickness .......................... FACTORS INFLUENCING CARTILAGE THICKNESS MEASUREMENTS WITH MULTI-DETECTOR CT A PHANTOM STUDY..................................................................................................................................................77 Trabecular Bone Elastic Modulus.........122 ......................................................................................................................................106 Error Analysis ................................3.................................................76 Reconstructions of Pelvic Geometry ....................................................................................................................................................................................................................73 Data Analysis ....................85 References..102 CT Imaging Protocol..............................................................................117 Study Limitations.....................119 Practical Applications .................................................75 Results ...........97 Abstract ......................................... VALIDATION AND SENSITIVITY STUDIES........................60 Materials and Methods...................................................................................................................................................................69 Assessment of Cortical Bone Thickness.........105 Image Segmentation.......................................72 Sensitivity Studies........92 4..............................................................................99 Materials and Methods......................63 Geometry Extraction and Mesh Generation .....................................................................................................................................................................................................................................................................58 Abstract ....................66 Position-Dependent Cortical Thickness.................................97 Introduction..........................................................................................................................80 FE Model Predictions .................109 Results ......................................................63 Experimental Study...................................................................................120 References............................................................................................................................................................................................71 Material Properties and Boundary Conditions...............................................................................................................115 Discussion ...................................................................................................................................................................................................................................................................................................................................................... A SUBJECT-SPECIFIC FINITE ELEMENT MODEL OF THE PELVIS: DEVELOPMENT............................................................................................................................................. and Measurement of Thickness ................................................. Surface Reconstruction....................80 Discussion ...............................................................58 Introduction.......................................110 Non-Enhanced Scans .....................................................

...189 7................. VALIDATION OF FINITE ELEMENT PREDICTIONS OF CARTILAGE CONTACT PRESSURE IN THE HUMAN HIP JOINT.......................165 Introduction........................................152 References............................................................................................137 Results ...141 Contact Patterns ........................................167 Methods.....................................................170 CT Arthrography............................................................................Cartilage Material Properties and Thickness.....................................................................................................................172 FE Boundary and Loading Conditions ......................... PATIENT-SPECIFIC FINITE ELEMENT MODELING PROOF OF CONCEPT ......................................................................................................................................................................................................................142 Misalignment and Magnitude Errors ..................................126 Introduction...............................................171 FE Mesh Generation................................................................................149 Discussion ........147 Sensitivity Studies............................126 Abstract .................................................................................................................................................200 References................................130 Experimental Protocol .............................................................. Contact Area...........170 Subject-Selection ........FE Boundary Conditions ...................................................175 Data Analysis .....................................................................................................................................................................................................5.................................................................................................................................................140 Peak Pressure......................................................................................... Average Pressure.....................................................................................................................................133 Sensitivity Studies....................140 FE Mesh Characteristics ........................... Cartilage Thickness.........................................................................130 Computational Protocol ......................................174 Sensitivity Studies...............................................................................................................................................205 ...................170 Image Data Analysis and Segmentation ........................... Material Properties.......................159 6..............165 Abstract ....................................182 References.......................................... DISCUSSION .................128 Methods..............................................................145 Sensitivity Studies..................................................................................................................................................................................................................................................................................195 Summary ..........................................................................................................................................................................................................................................195 Limitations and Future Work......177 Discussion ..............................................175 Results ............................................................................................................................136 Data Analysis ....

.....3.........................146 Differences in FE predicted mechanics between a normal and dysplastic hip joint ....................................................84 Misalignment and magnitude errors of FE predicted cartilage contact pressures ............... viii .LIST OF TABLES Table 3.....................................................................1......2............179 5................................................... 5......75 Reconstruction errors for simulated cortical bone ..2......................................... 6....................1...................... 3.................... Page Models studied for sensitivity analysis .................1....146 Differences of center of pressures between FE and experimental results ...............78 Results for all sensitivity models ................... 3............................

..................... Page Photograph of plastic hip joint.70 Cortical bone thickness phantom ....... 3....79 Distribution of pelvic Von-Mises stress ............ experimentally measured strains for subject-specific and sensitivity models .....................65 3D reconstruction of pelvis from CT image data...4.....17 Illustration of iliofemoral ligament................6.. 2.................... 3....6........................................................................................................16 Illustration of hip joint capsule ...........8.71 Schematic showing length measurements obtained from cadaveric pelvis and accuracy of geometry reconstruction..........3................................. 3..................................67 Finite element mesh of pelvis with close-up of acetabulum................12 Histological cross-section of cartilage ....................................................14 Illustration of hip joint with labrum.................................................. 3...........2..............................24 Schematic of pelvis loading fixture ......................1..................... 2.......5..................76 Contours of position dependent cortical bone thickness of the pelvis............LIST OF FIGURES Figure 2...................... 3.....5......... 2..........7.......................................................... ...........18 Volumetric CT scan of patient with acetabular retroversion ......... 3.................1....83 3.............. 3.......................4.3..........2.....................82 Finite element predicted vs...........68 Schematic illustrating the special cases considered in in determination of cortical thickness ...............................................................9..... 3............................................................. 2............. 2.......................................................

................................................... 5...6......... Schematic of phantom used to assess accuracy of cartilage thickness reconstructions ...........................3......... 5........... 4...................1..................111 Simulated cartilage RMS and mean residual reconstruction errors for the transverse contrast enhanced CT scans as a function of imaging direction and resolution ........................................................................2..4....104 Simulated cartilage RMS and mean residual reconstruction errors for the transverse contrast enhanced CT scans as a function of agent concentration................ 4..........1..........140 Finite element vs........... agent concentration.........................................................................................150 4.4.........8......5................. imaging direction and resolution ...............2.....3....................................... 5......................... average pressure and and contact area due to changes in cartilage material properties and geometry.......132 Finite element mesh of the entire hip joint in the walking position with a close-up of the acetabulum..........................112 Simulated cartilage RMS and mean residual reconstruction errors for the transverse contrast enhanced CT scans as a function of joint spacing.......... average pressure and and contact area due to changes in model boundary conditions........... 5.............114 Simulated cartilage RMS and mean residual reconstruction errors for the transverse non-contrast enhanced CT scans as a function of imaging direction and resolution ......................... xi .........141 Qualitative comparison between finite element and experimentally measured contact pressure .......143 Finite element predicted pressures relative to the femur and acetabulum ............................... 5................................. experimentally measured average pressure and contact area ..................................7...................135 Contours of cartilage thickness....4...................5... 5.....148 Percent changes in peak pressure......................................................................................................................116 Experimental setup for loading of hip joint . 5....... 5............ 4.................................144 Percent changes in peak pressure....

.................................9........2....................5.................................... Contours of cartilage pressure predicted by the baseline and rigid bone finite element models........180 6............ xii ............ 6....... 6............................177 Finite element predicted acetabular cartilage pressures as a function of anatomical and joint reaction force orientation for the normal and dysplastic hip joint ..................173 Comparison of finite element predicted acetabular cartilage pressures between a normal and dysplastic hip joint .151 Fringe plots of acetabular and femoral cartilage thickness for the normal and dysplastic hip joints........................1..3.......

University of Utah Funding Incentive Seed Grant. . the Orthopaedic Research and Education Foundation. and from NIH Grant #F31-EB00555 is gratefully acknowledged.ACKNOWLEDGEMENTS Financial support from the University of Utah Department of Orthopaedics. Scientific Computing Imaging Institute (SCI) and Brad Maker of Lawrence Livermore National Lab are also given acknowledgement for their contributions. The University of Utah Department of Radiology (CAMT).

resulting in osteoarthritis (OA) [4-7]. While supporting the majority of the human body (~ 2/3 of total bodyweight) the joint must simultaneously facilitate smooth articulation of the lower limbs to enable bi-pedal gait. While THA has enjoyed a high rate of success in elderly patients (less than 10% require revision THA [10]).CHAPTER 1 INTRODUCTION MOTIVATION The hip joint serves a very important biomechanical function. abnormal.9]. social and psychological effects.5 times bodyweight are transferred between the femur and pelvis [1-3]. the surgery is generally avoided in the younger population due to the limited lifespan of implants and . During routine daily activities. While its efficient structure may lend to a lifetime of mobility. Hip joint OA represents a significant burden to society via financial. nearly 193. forces on the order of 5. repetitive loading of the hip is thought to result in the breakdown of articular cartilage. It is estimated that nearly 40 million Americans currently have joint osteoarthritis (~18% of the population) of which nearly 3% of the cases originate at the hip joint [8.000 osteoarthritic hips are replaced annually in the United States by way of total hip arthroplasty (THA) [10]. To alleviate pain and return the hip to at least the most basic functioning state.

body weight. In particular. in part because it takes so long to develop. quantifying the relevant mechanical parameters in the joint (i. Research on the biomechanics of the hip is not new to orthopedic medicine. substantial voids remain. factors such as abnormal joint geometry (i. simplifying model assumptions have often resulted in model . the advent. along with a tremendous evolution of computing power. they require a protocol that is inherently invasive. have lead to substantial growth in the field of computational biomechanics.2 unfavorable results of revision THA. Unlike experimental investigations. In particular.e. computational modeling opens the possibility of predicting hip joint biomechanics noninvasively. the etiology of hip OA is unknown. Nevertheless. Although a common misconception by the general public. Thus.a decade or more likely passes before cartilage has fissured to the point where bone contact initiates pain [14]. However. and prior injury appear to play major roles independently of age [11-13]. While previous studies have provided valuable insight into baseline hip joint biomechanics. due to the high loads this joint supports. These attractive points. mechanical factors have been implicated as a primary causes [4-7]. increased availability and resolution medical imaging modealities provide a means to develop detailed computational models that are based on individual patient geometry. cartilage and bone stresses) appears to be central to an enhanced understanding of this disease [14]. hip OA is not confined to elderly patients. hip dysplasia). occupation.e. As detailed later. Although computational models have provided substantial additional insight to hip biomechanics above and beyond that obtained via experimental studies.

3 predictions that are inconsistent with experimental measurements. and 2) an assessment of model error and uncertainty is accounted for in an effort to gauge how inaccuracies could be propagated due to erroneous model inputs and assumptions. without having to validate each model independently. . For an analyst to develop patient-specific models. Most models have not been validated by direct comparison with experimental data. it is necessary to demonstrate that: 1) the computational protocol yields results that predict known/measured quantities with sufficient accuracy.

4 RESEARCH GOALS

The overall aim of this dissertation is to develop and validate methods that will facilitate patient-specific modeling of hip joint biomechanics. It is clear that patientspecific computational models have the potential to revolutionize the way that disorders of the hip joint are diagnosed and treated. However, first rigorous and validated

protocols that incorporate both computational and experimental techniques must be established. In this dissertation research, model predictions are compared directly with experimental data and errors in assumed and measured model inputs, i.e. material properties, constitutive behavior, geometry, and boundary conditions, are discussed as they pertain to errors in the developed model and in the context of future patient modeling efforts. Second, recognizing that acetabular dysplasia may be the single most important contributor of hip joint OA [17-20], the validated subject-specific finite element modeling protocol is extended to analyze patient-specific models to demonstrate the general feasibility of the approach and to quantify differences in hip joint biomechanics between a normal subject and a patient with acetabular dysplasia. Finally, limitations of the modeling protocol as well as unforeseen challenges in modeling individual patients non-invasively are presented in the context of general hip joint modeling applications but with special emphasis to the study of acetabular dysplasia.

5 SUMMARY OF CHAPTERS

The structure of this dissertation has been organized to answer the essential questions that arise when developing a protocol to generate patient-specific models of the hip joint. The primary objective of Chapter 2 is to provide a working knowledge base that will serve as reference to the topics covered in the remaining Chapters. Chapter 2 discusses hip joint structure and function and provides further motivation for developing computational models, in particular, by application of the finite element method. Topics such as hip osteoarthritis, approaches to quantify hip biomechanics (experimental and computational), and the importance of computational model verification, validation, and sensitivity studies are presented. In the context of hip joint OA, the most fundamental mechanics of interest are bone and cartilage stresses and strains. In this dissertation computational models are developed and validated to study the mechanics of bone and cartilage separately in an effort to maintain simplicity and to control confounding factors (Chapter 3 and Chapter 5, respectively). Chapter 3 discusses the development and validation of a subject-specific finite element model of the pelvis. Computational predictions of bone strains were validated by direct comparison to experimentally measured strains using tri-axial strain gauges attached to the cortical surface of the pelvis. The influence of measured and assumed model inputs was assessed using sensitivity studies and the geometrical accuracy of the model, in particular, cortical bone thickness, was quantified. Since computational models often use medical image data as the basis for creating model geometry, it is crucial to demonstrate that the reconstructed geometry is an

6 accurate representation of the true continuum. Although this is important for all

computational studies, it becomes absolutely essential when modeling joint contact mechanics between layers of articulating cartilage. Although volumetric computed

tomography (CT) data are often used as the basis for constructing joint models, the lower bounds for detecting articular cartilage thickness and the influence of imaging parameters on the ability to image cartilage have yet to be reported. Thus, it was necessary to quantify the accuracy and detection limits of cartilage geometry with CT. To this end, a phantom based imaging study was conducted and is presented in Chapter 4. While the results of this study have application to subject-specific models using cadaveric joints (Chapter 5), the primary focus of this work was centered around quantifying the accuracy of CT arthrography since this procedure is required to visualize cartilage in live patients, which has direct relevance to patient-specific modeling (Chapter 6). The results of this study are discussed in terms of general clinical applications of CT for imaging articular cartilage in diarthrodial joints. Chapter 5 presents results for a subject-specific finite element model of the mechanics of cartilage in the hip joint. Finite element predictions of cartilage contact pressures were validated by comparing computational predictions to experimentally measured contact pressures using pressure sensitive film. A physiological experimental and computational loading protocol was employed in an effort to lay the groundwork necessary for creating realistic patient-specific models. Sensitivity studies were included in the analysis to determine the influence of measured and assumed model inputs on the

7 ability to predict cartilage contact pressures. Chapter 5 concludes with a discussion of the model’s limitations. As discussed in Chapter 2, acetabular dysplasia may be the primary etiology of hip joint OA. Therefore, it was of interest to demonstrate the feasibility of the modeling protocol developed in Chapters 3-5 for application to individual patients with dysplasia. Chapter 6 presents two patient-specific finite element models: one for a normal volunteer and one for a patient with acetabular dysplasia. Differences in computationally predicted bone and cartilage mechanics are reported and clinical implications of these data are discussed. Chapter 6 concludes with a discussion of the challenges involved with patientspecific modeling and makes recommendations for circumventing these issues in the context of future modeling efforts. Finally, Chapter 7 summarizes the entire dissertation by highlighting the important contributions that were made to the field of hip computational biomechanics along with a discussion of limitations and future research directions.

." N Engl J Med.. Setton. A. 2001.. T. "The Reaction of Articular Cartilage to Injury and Osteoarthritis (Second of Two Parts). J. 2007." Ann Internal Med.. pp. "Factors Associated with Osteoarthritis of the Knee in the First National Health and Nutrition Examination Survey (Hanes I). M. T. S. Lawrence. G. American Academy of Orthopaedic Surgeons. pp. pp. R. F. W.. and Dieppe. Race. 34.. 2000.. D. Heller.. Harris.. pp. A. R.. [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] . 726-737. L. 635646. G. T. Part 2.. 291. 1335-40. Carlson." N Engl J Med. 179-89. and Dieppe. 2879-83. J. A. L. Treatment Approaches."Questions and Answers about Hip Replacement." Am J Epidemiol. Fijan. "Mechanical Factors in Articular Cartilage and Their Role in Osteoarthritis. C. P. "Hip Contact Forces and Gait Patterns from Routine Activities. 2000. Deuretzbacher.. J. W. R. 291. R. Felson. 1995. C. Guilak." Proc Natl Acad Sci U S A. Bergmann... "Osteoarthritis: New Insights. "Contact Pressures in the Human Hip Joint Measured in Vivo. Mow. A. R. N. Evidence for an Association with Overweight. Bergmann. D. 1998. Poole. Mankin..... F. 1285-92. and Mann. G.. Part I the Disease and Its Risk Factors.. 1986. "Imbalances of Anabolism and Catabolism of Cartilage Matrix Components in Osteoarthritis. A. H.. Berlin. K. Mankin. 1974. G. H.. 133. American Academy of Orthopaedic Surgeons. 133..8 REFERENCES [1] Hodge. and Felson. 83. H. W.. and Duda. Rohlmann." Ann Internal Med. "The Reaction of Articular Cartilage to Injury and Osteoarthritis (First of Two Parts). J.. Burgess. G.. D. Graichen." in Osteoarthritic Disorders. 1988. Felson." in Osteoarthritic Disorders. Strauss.. "Osteoarthritis: New Insights. pp. pp. Anderson. P.. Lawrence. pp. A. R. AAOS. 1974. 1995. 128. and Ratcliffe. V. A. C. and Physical Demands of Work. J. Free University and Humboldt University." National Institute of Arthritis and Musculoskeletal and Skin Disorders. Hip98: Data Collection of Hip Joint Loading on CD-Rom." J Biomech. 859-71.

"Verification.217 Persons.. P. O. 116. 810-24.. 1993. Validation and Sensitivity Studies in Computational Biomechanics. pp. Solomon.. C... 64. "Patterns of Osteoarthritis of the Hip. Heliovaara. A. 38.. 8. 1974. pp." J Biomech Eng. "Do Occupation-Related Physical Factors Contribute to Arthritis?.. 1986. "Etiology of Osteoarthritis of the Hip.9 [12] [13] Felson. K. J." Br J Radiol. and Hirsch. Makela. Maryland... H. W. Oberkampf. MO. T." Clin Orthop. Laurel. Aromaa.. T. W.. "Verification. 63-77. L.. Johns Hopkins University." J Bone Joint Surg Br. [14] [15] [16] [17] [18] [19] [20] . D." presented at Proceedings of the second open scientific meeting of the Hip Society. St Louis. Stulberg. A Health Survey of 7.. 1976. A. "Cartilage Stresses in the Human Hip Joint. Knekt. Harris. D. T. 2033. 176-83. M. "Acetabular Dysplasia and Development of Osteoarthritis of the Hip. R.. 1994.. 2006. Murray. Macirowski. H. Trauma and Workload with Coxarthrosis. R. 10-8. G. and Sievers. Impivaara. Validation.. Anderson. 513-8." Computer Methods in Biomechanics and Biomedical Engineering.. Trucano. 2002. S. W. pp. and Predictive Capability in Computational Engineering and Physics." Acta Orthop Scand.. J.. 1965. and Mann. and Weiss. S. M. pp. 58. "The Aetiology of Primary Osteoarthritis of the Hip. and Harris. pp. Tepic. B." presented at Foundations for Verification and Validation in the 21st Century Workshop. L. A. O. pp. 1994. Ellis.. "Association of Overweight. W. In Press Dec 2006." Baillieres Clin Rheumatol.

Furthermore. This chapter presents the most relevant background information.CHAPTER 2 BACKGROUND FORWARD As discussed below. However. and numerical modeling of hip joint biomechanics. . it can be credibly argued that most previous hip joint computational models do not have the ability to predict cartilage and bone mechanics with sufficient accuracy for clinical application. The purpose of this dissertation is to present novel methods to develop and validate computational models of the hip joint that result in models that indeed have direct clinical applicability to study diseases such as osteoarthritis (OA) and hip dysplasia. hip pathologies. models that have included more realistic geometries have not been validated by direct comparison to experimental data. including: hip joint structure and function. experimental hip joint biomechanics. based on the available in-vivo and in-vitro experimental data. computational models have the potential to non-invasively estimate hip mechanics for living subjects.

11 HIP JOINT STRUCTURE AND FUNCTION The hip is a ball and socket joint formed by the articulation of the spherical head of the femur and the concave acetabulum of the pelvis. Pelvic and Femoral Bone The pelvis forms a girdle which protects the digestive and female reproductive organs. which fuse together to form the ox coxae. It is formed from three bones: the ilium. cylindrical structure. It forms the primary connection between the lower limbs and skeleton of the upper body. The entire hip joint is surrounded by a fibrous. neck and lateral bony . Both the femur and acetabulum are covered with a layer of cartilage to provide smooth articulation and to absorb load. The diaphysis of femur is a simplistic. Many large nerves and blood vessels pass through the pelvis to the lower limbs. Muscles and tendons provide actuation forces for extension/flexion. while the proximal femur is irregular in shape. ischium and pubis.1). At the point of fusing they form the acetabulum (Figure 2. consisting of a spherical head. Several ligaments connect the pelvis to femur to further stabilize the joint and capsule. and two condyles (medial and lateral) distally.1). or innominate bone (Figure 2. It consists of a head and a neck proximally. a diaphysis (or shaft). adduction/abduction and internal/external rotation. The femur is the longest and strongest bone in the human body.1). Joints adjacent to the pelvis include the sacroiliac (SI) and pubis joint (Figure 2. flexible capsule to permit large ranges of motion but to prohibit the proximal femur from dislocation.

The trochanters serve as the site of major muscle attachment.2 mm thick.12 Sacro-Iliac Joint Ilium Acetabulum Pubis Pubis Joint Femur Ischium Figure 2. consisting of a dense.7 to 3. The lateral location of these structures offers a mechanical advantage to assist with abducting the hip [2]. Cortical bone along the diaphysis of femur is much thicker (up to ~ 7 mm [4]) and supports the large tensile and compressive loads that develop as a result of hip loading. protrusions termed the greater and lesser trochanters. Trabecular . The bony structure of the pelvis is similar to a sandwich composite material. Photograph of a plastic hip showing the individual bones and joints.1. thin shell of cortical bone (0. The spherical head of the femur has a thin layer of subchondral bone where cartilage attaches. stiff. which is less stiff than cortical bone. [3]) filled with much less dense trabecular bone.

Cartilage is avascular and anueral [6]. The most common GAGs are keratin sulfate and chondroitin sulfate [7]. The thickness of femoral cartilage ranges from 1. Nutrients diffuse through the matrix by way of interstitial fluid.2).3 mm thick in normal adults [9]. In addition. Cartilage has remarkable mechanical properties in that it is strong but flexible and has extremely low coefficients of friction [5]. Cartilage Cartilage is composed of collagenous fibers and chondrocytes embedded in a firm gel (Figure 2. The solid matrix represents nearly 60% of the dry weight and is composed of proteoglycans. which are large proteins with a protein backbone and glycosaminoglycan (GAG) side chain [7]. Matrix fibers make up the remaining dry weight and are composed of collagen type II [7].5 in the normal adult [9]. which makes up nearly 60-80% of the total weight of cartilage [7].2) [5]. The entire head of the femur is covered with a smooth layer of hyaline cartilage of varying thickness.0 to 2. except for a small depression called the fovea capitis femoris that gives attachment to the ligamentum teres. ionic charges in the fluid are thought to facilitate nutrient flow [8]. Chondrocytes and their precursor chondroblasts are the only cells in cartilage (Figure 2.2 to 2. . The acetabulum is covered with a horseshoe shaped layer of articulating hyaline cartilage ranging from 1.13 bone is found throughout the pelvis and proximal femur but is not as prevalent along the diaphysis of the femur as this area primarily contains marrow.

Copyright Lutz Slomianka 1998-2006.14 Calcified Cartilage Cartilage Matrix Figure 2. Chondrocytes appear as small dots. Subchondral Bone . Through the thickness histological photograph of bovine articular cartilage (stained with Alician Blue).2.

It is the strongest ligament in the human body and allows one to maintain posture for extended periods without extensive muscular fatigue. Ligaments. The pubofemoral ligament is the most anterior and inferior portion of the fibrous capsule. The iliofemoral ligament attaches from the pelvis to the femur and resists excessive extension (Figure 2. The capsule consists of two sets of fibers: longitudinal and circular fibers. It attaches to the pubic bone and passes inferolaterally to merge with the iliofemoral portion of the fibrous capsule (attaching to intertrochanteric line).5).4). Capsule. Muscles The hip joint labrum is a ring of fibrocartilagenous tissue that surrounds the rim of the acetabulum (Figure 2. to prevent dislocation between the femoral head and acetabulum and is thought to act as a seal to prevent loss of interstitial fluid. Overabduction of the hip joint is prevented by this ligament. The entire hip joint is surrounded by the hip capsule (Figure 2.3). The ischiofemoral ligament attaches from the ischial part of the acetabulum to . The capsule is further defined by ligaments. It helps to guide normal motion. The capsule attaches proximally to entire periphery of the acetabulum. It also covers the femoral head and neck like a sleeve. An in vitro study where the labrum was removed have demonstrated that cartilage is strained to a greater degree than joints with intact labrums for a given load [10]. suggesting that greater fluid flow in the latter scenario causes increased cartilage matrix deformation.15 Labrum. Circular fibers form a collar around the femoral neck whereas longitudinal retinacular fibers travel along the neck and carry blood vessels. beyond the acetabular labrum. eventually attaching to the base of the femoral neck.

16 the femur and supports the posterior aspect of hip capsule. Adopted from [1]. . Finally. Illustration of hip joint with labrum. Although it supports little or no loads it may serve as an important conduit to supply blood to the head of the femur in some individuals. the ligament teres connects the head of the femur to the acetabulum (Figure 2.3.3). Labrum Figure 2.

Adopted from [1].4.17 Figure 2. . Illustration of hip joint capsule.

Adopted from [1].5.18 Figure 2. Illustration of iliofemoral ligament. .

gemelli superior and inferior. the adductores longus. The rotators that cause inward rotation are the glutaeus minimus. The adductors include the adductores magnus. anterior fibers of the glutaeus medius. and brevis. Finally. and rotate using several muscles attached between the pelvis and femur. brevis. Hip joint extensors include the glutaeus maximus. adductores longus and brevis. longus. the piriformis. and magnus. the pectineus. and the upper part of the glutaeus maximus. Many of the hip muscles are responsible for more than one type of movement in the hip as different areas of the muscle act on tendons in different ways. the gracilis. iliacus. lower part of the glutaeus maximus. and the Sartorius.19 The hip can flex. abduct. pectineus. tensor fasciae latae and the iliacus and psoas major. the glutaei medius and minimus. glutaeus maximus. assisted by the hamstring muscles and the ischial head of the adductor magnus. extend. sartorius. and the anterior fibers of the glutaei medius and minimus. inward rotators include the posterior fibers of the glutaeus medius. adduct. obturatores externus and internus. the pectineus. . Hip joint flexors include the psoas major. quadratus femoris. rectus femoris.

OA is classically associated with more advanced age but is being seen and treated more frequently in younger patients [12]. Gradual loss of the matrix components is thought to be caused by a loss of proteoglycans. OA is further characterized by an increase in vascularity of . traumatic The most common disorders include arthritis arthritis. Perthes disease). with eburnation of the underlying subchondral bone and a proliferative response characterized by osteophytosis [13. synovium. although some changes in the integrity of collagen network may be necessary to initiate the disease [15]. Osteoarthritis and hip dysplasia are the primary thrust of this dissertation. It is also the most common cause of musculoskeletal pain in the United States [11]. and capsule [12].20 HIP JOINT PATHOLOGY The hip is subject to disease due to improper development. impingement and dysplasia. osteoarthritis). acute trauma and prolonged mechanical wear and tear. slipped epiphysis.14]. avascular necrosis (osetochondritis dissecans. bone. Hip OA is a disorder of the entire joint. It is also thought that OA disrupts the mechanism for fluid support. bursitis. Osteoarthritis Osteoarthritis is the most common type of arthritis in the hip and is intimately associated with other disorders such as avascular necrosis. (rheumatoid arthritis. developmental dysplasia of the hip and femoro-acetabular impingement. which may exacerbate solid matrix degeneration [12]. involving cartilage. and thus will be the focus of further discussion. labrum. OA is characterized by a loss of articular cartilage in the predominately load bearing areas of the joint. slipped epiphysis.

it is unknown whether degenerative . which may lead to further degeneration. considerable clinical. epidemiological. No experimental studies to date have allowed one to directly separate the contributions of hydrostatic and deviatoric stresses generated by joint loading. Therefore. Therefore. Excessive mechanical stress can directly damage articular cartilage and subchondral bone and can adversely alter chondrocyte function including the balance between synthetic and degradative activity [19-25]. and pneumatic drill operators. metal workers. However.21 surrounding bone. OA was once thought to be primary or idiopathic in nature. They may also cause outgrowths of cartilage as well as osteophyte formation. Opinions differ concerning the relative contributions of direct mechanical trauma to articular cartilage versus elevation of articular stresses secondary to stiffening of subchondral bone. miners. the mechanics of bone in areas adjacent to cartilage may be important to developing a comprehensive understanding of hip OA. suggest that repetitive intense loading is associated with early onset of joint degeneration [16-18]. Surveys of individuals with physically demanding occupations. including farmers. construction workers. Cysts often form within the surrounding bone and are accompanied by changes to the joint margin. and experimental evidence supports the concept that mechanical demand greater than some critical level has a major role in the development and progression of joint degeneration in all forms of OA. the precise mechanism remains controversial. Although joint loading or overloading can lead to cartilage degeneration.

Subluxation caused by dysplasia of the hip joint is a primary cause of degenerative joint disease and clinical disability [30]. In the broadest sense. including replacement of the prosthesis. Nearly 193. it is thought that the altered biomechanics cause cartilage and bone of the hip to break down prematurely.000 total hip arthroplasties (THA) are performed each year in the United States [26]. THA is highly successful in relieving pain and restoring movement. However.22 changes are primarily due to excessive loads applied during normal activities or shearing of cartilage during abnormal motions involving local instability. or completely dislocated [28]. malformed. In the elderly patient population. Surgical correction of the anatomic abnormalities associated with hip dysplasia is performed on younger patients via pelvic . ongoing problems with wear and particulate debris may eventually necessitate further surgery. Hip Dysplasia Developmental dysplasia of the hip (DDH) describes a broad spectrum of problems. Consequently. including hips that are unstable. hip OA is treated by prosthetic replacement of the hip joint. The chance of a hip replacement lasting 20 years is about 80% [27]. DDH is a developmental deformity characterized by malorientation and a reduction of contact area between the femur and acetabulum [29]. subluxated (incomplete dislocation). leading to early hip osteoarthritis. OA due to hip dysplasia is commonly treated by THA. Subluxation leads to increased stresses across the hip joint each time the hip is loaded during gait [31]. Men and patients who weight more than 165 pounds have higher rates of failure [27].

Retroversion is difficult to diagnose by the untrained clinician since a standard hip radiograph shows that the hip is normal. In contrast to these reports. in patients that went unrecognized before. It is likely that the obscurity of the retroverted acetabulum has often persuaded the clinician into prescribing passive treatments for a potentially aggressive pathology. These discrepancies highlight the need for an improved understanding of hip dysplasia. Several studies have shown that mild developmental dysplasia. . can be diagnosed by evaluation of a 2-D planar radiograph of the hip. may indeed be the leading cause of osteoarthritis in the hip [39-43]. referred to herein as traditional dysplasia. In the retroverted acetabulum. Redirectional acetabular osteotomy (periacetabular osteotomy) involves cutting the socket free from the pelvis and rotating it to a new orientation [32-37]. [29] estimated that 76% of patients with OA of the hip have some type of untreated acetabular dysplasia. has been identified [38. Recently.6). The most common type of dysplasia. the acetabular opening and its proximal roof lie at an angle of retroversion with respect to the sagittal plane. the anatomy of the retroverted acetabulum is still pathologic. other studies have failed to find a statistically significant relationship between acetabular dysplasia and the risk of hip OA [45-49]. posterior coverage of the femoral head is lost (Figure 2. Wilson and Poss [44] reported that deformity of the acetabulum is found in 2535% of adult cases of OA of the hip. a specific variant of dysplasia. referred to herein as retroversion of the acetabulum.39]. which preserves the hip joint and associated articular cartilage and delays the need for prosthetic replacement of the hip.23 osteotomy. nevertheless. Michaeli et al. Thus.

Peters. The lines indicate the edge of the acetabulum. The patient’s right hip was considered normal. . Courtesy of Christopher L. Note excessive forward progression of anterior edge of left acetabulum in image A) and lack of posterior wall coverage of left femur in image B). Anterior (A) and Posterior (B) view of a volumetric CT scan from a patient with acetabular retroversion of the left hip.24 A) Anterior Right Left B) Left Posterior Right Figure 2.6.

the tissues of primary interest in the context of studying OA and hip dysplasia are bone and cartilage. ligaments. However. pressure sensitive film. In-vitro experimental studies are conducted using whole cadaveric hip joint bones or individual tissue samples that are instrumented with sensing devices such as strain gauges. The converse was also stated to hold true. In the mid 20th century Dempster and Liddicoat [52] demonstrated that cortical bone exhibits different moduli of elasticity when loaded in different directions. several studies have employed instrumented hip prostheses implanted at the time of THA in patients with OA. to study the effects of interventions and diseases. Bone Material Properties Studies of the material properties of bone date back to the mid 1800s when Wertheim measured the strength and elasticity of human cortical bone specimens [50]. In-vivo studies are difficult to conduct as access to the hip joint is extremely intrusive. and to quantify normal cartilage and bone mechanics. the bone will remodel itself over time to become stronger to resist that mechanism of loading. However. The dependence of the elastic properties on the basic lamellar unit of cortical bone was . and hip capsule serve as vital components when studying general hip joint biomechanics. In the latter half of the 19th century Julius Wolff published pioneering work regarding bone remodeling (termed Wolff’s Law. [51]) by stating that if loading on a particular bone increases.25 EXPERIMENTAL HIP JOINT BIOMECHANICS The contribution of muscles. tendons. The objectives of these tests are to ascertain material properties. and force transducers.

Lang et al. Building on this work. The combined data made it possible to establish empirical relations between apparent density. It has also been argued that trabecular alignment corresponds to principal stress directions [51. The reported Young’s moduli for cortical bone have been shown to be about 20 – 22 GPa along the axis of long bone and about 12 – 14 GPa transverse to it [61. Early work by Chalmers and Weaver [68] and Galante et al. cortical bone possesses orthotropic elastic properties. Therefore.67]. Dalstra et al. Direct mechanical tests further confirmed that cortical bone can be reliably considered as a transversely isotropic material [59. and nondestructive mechanical testing was used to obtain Young's moduli and Poisson's ratios in three orthogonal directions for cubic specimens of pelvic trabecular bone [75]. To investigate the transverse isotropy assumption further. [55. [58] measured the anisotropic moduli using ultrasound.60]. They showed that.54]. used dual-energy quantitative computer tomography (DEQCT) to investigate the distribution of bone densities in pelvic bone.69-74]. but stiffness in various directions normal to the Haversian canals did not deviate more than 10%. [68] showed that porosity was far more important than mineral content in determining material properties.26 recognized early by Evans et al. apparent rather than calcium-equivalent densities are most often used to identify the remodeling state of bone [65. [53.66. in general.56] measured cortical bone moduli by assuming transverse isotropy (the plane normal to the Haversian canals being the plane of isotropy).62]. calcium . van Buskirk and Ashman [57] and Katz et al. It is widely accepted that trabecular bone exhibits orthotropic material behavior (three preferred material directions) [63-65].

He described the ability of cartilage to deform under pressure and to regain its original shape when the pressure was removed. suggesting that it a viscoelastic material [76-79]. Later work confirmed that collagen fibers are oriented nearly perpendicular to the calcified interface and change orientation gradually until they are nearly parallel to the articulating surface [84.79] demonstrated that the stiffness of trabecular bone specimens increased significantly as the loading rate was increased incrementally.e. [81] demonstrated the viscoelastic behavior of bone through creep and relaxation tests.85]. Schoenfeld et al. Linde and Hvid [78. found that pelvic trabecular bone stiffness ranged from 100 – 250 MPa when using density as a primary predictor [75].27 equivalent density and elastic modulus for pelvic trabecular bone. He further described how the collagen fibers anchored in the underlying bone ran vertically through the cartilage as: “a mass of short and nearly parallel fibers rising from the bone. Cartilage Material Properties Cartilage structure and function was described as early as 1743 when William Hunter presented the paper “Of the Structure and Disease of Articulating Cartilage” [82]. Bone exhibits rate-dependent material behavior. . Collagen fiber orientation was investigated further in the late 1800’s when India ink studies demonstrated that cartilage split lines (i. and terminating at the external surface of the cartilage”. Dalstra et al. path of collagen fiber alignment) had a tendency to lay parallel to the articulating surface and to extend radially [83]. [80] determined the relaxation of trabecular bone and Zilch et al.

107.87] and through the depth [88-92]. which is essentially incompressible.108]. Cartilage is viscoelastic due to its high water content and relative mobility of the fluid phase relative to the solid phase [104-106]. The equilibrium modulus of cartilage is very low. McCutchen proposed a self pressurizing. some have argued that fluid could flow into the cartilage during loading causing a “boosting” effect [114]. In addition. Cartilage exhibits a tension-compression nonlinearity wherein cartilage has higher stiffness values in tension than in compressive [88. on the order of 0. Nevertheless. yet contact stresses measured invivo routinely exceed 2. . because cartilage in the normal joint is ~70% liquid. cartilage is stiffer when loaded along the split line direction compared to perpendicular to this direction [93-97]. corroborating the original “weeping” mechanism [113].0 MPa [109-111].99]. In contrast. Several theories have been proposed to explain how cartilage can routinely support loads higher than what the solid matrix can withstand [112].95. and because the cartilage layers indeed consolidate under load ~10% [115] it is more likely that fluid flows out of cartilage layers. It is thought that the higher stiffness in tension versus compression allows cartilage to resist radial expansion under axial compressive loading and results in increased fluid pressurization and dynamic stiffness [93.89.100-103]. “weeping” mechanism whereby synovial joints are supported mainly by the hydrostatic fluid pressure [113].28 Experimental studies demonstrate that cartilage is an inhomogenous tissue.98. Cartilage modulus varies extensively depending on the location on the articulating surface [86.3 – 1.5 MPa [104.

g. during automobile side impacts or femur fracture in the elderly).7 MPa.119-121].1 – 40 Hz. Minimum and maximum moduli ranged from 14.29 Under cyclic loading at physiological frequencies.g. Studies with particular relevance to this dissertation are related to the measurement of bone and cartilage stresses and strains in intact hips or hips with simulated dysplasia as they provide baseline experimental data and detail proven methodologies that can be useful when developing and validating computational models of the hip joint. but still nonlinear.. Stress strain curves became markedly steeper. These data demonstrate the ability of cartilage to routinely maintain physiological levels of contact stress [105]. In-Vitro Studies of Hip Joints In vitro experimental studies have served to elucidate hip biomechanics on the macro scale. respectively. These studies have helped to elucidate plausible modes of failure (e. Hysteresis (energy dissipation) was reported as zero at 40 Hz and was not substantial at rates higher than 1 Hz. Park et al. at higher loading rates. peri-prosthetic bone shielding).. . implications of prosthetic replacement (e. It was suggested that the nonlinear stress response of cartilage under loading was in part due to the tension-compression nonlinearity. [105] demonstrated the rate response of cartilage tissue samples under load controlled unconfined compression using frequencies ranging form 0.6 – 65. interstitial fluid pressures remain elevated [116-118]. The dynamic cartilage modulus under these conditions is orders of magnitude greater than the equilibrium modulus [117. and magnitudes of normal bone and cartilage stresses and strains.

Cortical bone strains ranged from 1700 – 2300 µE.134-136]. Finlay et al. [128] measured the distribution of strain in the proximal femur under conditions of simulated single-leg stance using strain gauges applied to the cortex.133. Kim et al. [127] dissected four cadaveric pelvi free of soft tissue and instrumented each hemipelvis with ten rosette strain gauges to measure normal pelvic strains in-vitro.2 MPa and Poisson’s ratio of 0. they found that strain increased from proximal to distal in the intact femora under load. spherical femoral heads).e.130-138]. Strains decreased from proximal to distal in the intact femora under load.138].3 for cortical bone. piezoelectric sensors [131. In contrast to Oh et al.122127]. More recently.132]. Pressure sensitive film is often the measurement technique of choice as it is inexpensive. or instrumented prostheses [115. Static loading was applied through the intact hip joint to simulate single leg stance.30 Several studies have used strain gauges to quantify strains in the pelvis [3. and has . [129] affixed strain gauges in the proximal femur and subjected it to loads of 900 N. These studies used pressure sensitive film [130.5kN of force directed from the femur into the joint.. [123] subjected pelvi with 2. accommodates various geometries when cut (i. and the highest values were in the calcar area.137. This strain pattern was consistent with bending applied to the ilium from the action of the abductor and joint reaction forces. Ries et al. Oh et al. The medial portion of the pelvis was under tension directed vertically and the lateral ilium was in compression. Numerous in vitro studies have investigated cartilage hip joint stresses in the intact hip joint [115. Normal pelvic maximum principal stresses reached ~12 MPa assuming an elastic modulus of 6.

Rushfeldt et al. [139]). Local stresses were sensed by arrays of 24 compliant miniature transducers inset superficially in the femoral head cartilage.92 MPa and 8.80 MPa. Adams et al. Contact stress magnitude was usually found to rise nearly linearly with applied joint loads in excess of about 1000 N. The full contact stress patterns were irregular and complex. Using pressure sensitive film. respectively. suggesting that high cartilage stresses are likely to occur throughout the thickness of hyaline cartilage in the hip joint.57 MPa at the interface between acetabular cartilage and subchondral bone. The sites of maximum local stress were found to underlie the general region of the acetabular dome. [138] measured the in vitro distribution of pressure on the cartilage surface of the human acetabulum using a modified endoprosthesis with fourteen integral pressure transducers. Peak pressures at 2250 N of load were ~11 MPa and decreased with time while the contact area increased. oriented in an approximately anterior-to-posterior direction. Brown et al.135].31 proven to be reasonably accurate (± 10% error. It was concluded that the highly irregular . [133] measured the time variant distributions of intra-articular contact stress from direct measurement of seventeen grossly normal fresh cadaveric hips. For a resultant joint load of 2700 N. von Eisenhart-Rothe measured peak contact stresses of ~9 MPa in intact femoral heads when loaded directly into the acetabulum at 300% bodyweight [134. The pressure distribution was neither uniform nor axisymmetric about the load vector. Maximum pressure ranged from 4. but most commonly the general feature was a central band or “ridge” of pressure elevation.93 to 9. implanted eleven piezoelectric pressure transducers through the bone of the acetabulum [132]. the spatial mean contact stress and peak local contact stress averaged 2.

for studying specific Another limitation is that to analyze individual disorders such as hip OA and dysplasia one would need to obtain cadaveric specimens that exhibited the pathology of interest. Using a combined experimental and computational protocol they found that. in vitro experimentation may not be the most appropriate technique for the study of specific hip pathologies. However. if not impossible endeavor. Given the challenges of obtaining donor tissue this task would be a very difficult. In-Vivo Studies of Hip Joints No known methods exist to measure pelvic and femoral bone strains in-vivo. [140] was one of the first to describe the development of a radio telemetrized femoral . Therefore. [115] used a similar prosthesis to measure the total surface on acetabular cartilage when step-loaded by an instrumented hemiprosthesis. Macirowski et al. One limitation of experimental studies is that measurements of stress and strain are only obtained at the location of the sensing device. Based on these limitations. Their results provided further support for the “weeping” mechanism proposed by McCuthen [113]. a priori knowledge is required to place sensors in meaningful positions. Carlson et al. a considerable amount of work has been devoted to measuring hip joint contact pressures and joint reaction forces using instrumented prostheses.32 pressure profiles observed are due primarily to cartilage thickness distribution and irregularities at the calcified cartilage interface. fluid pressure supported nearly 90% of the load within the cartilage network stresses. even after long-duration application of physiological force. However. disorders this information may be unknown.

Pressure magnitudes were synchronized with body-segment kinematic data and foot-floor force measurements to locate transduced pressure areas on the natural acetabulum and to correlate movement kinematics and dynamics with local cartilage pressures. stair climbing. More recently. the procedure is invasive and limited to a small sample of patients who have already undergone treatment to correct . with abrupt spatial and temporal gradients. Joint reaction forces were as high as 5. and rising from a chair.5 times bodyweight) during walking. Bergmann et al. rehabilitation. [111] implanted a prosthesis into a patient and measured contact stress at 10 discrete locations. Data from instrumented prostheses have yielded contact pressure data that are consistent with experimental studies. However. recovery. joint reaction forces and joint kinematics. but were generally lower (2.5 times bodyweight when the subjects rose from a chair. and descending stairs.110]. suggesting that this technique has the ability to accurately measure hip joint mechanics in vivo. Data were acquired during surgery. Gait analysis was performed on each patient and contact pressure data and equivalent joint reaction force were evaluated in parallel with joint kinematics during a variety of daily activities such as walking. The data revealed very high local (up to 18 MPa) and non-uniform pressures. descending stairs. In 1985 Hodge et al. stair climbing. and normal activity. implanted similar prostheses in 5 patients who underwent THA surgery for treatment of hip OA [109. Their data also suggested that contact stresses and joint reaction forces correlated well with foot-floor force measurements and demonstrated large inter-subject variation in contact stresses. for longer than 1 year after surgery.33 prosthesis in 1974.

. as with in vitro studies. Finally. In addition. the technique is limited to the measurement of joint reaction forces associated with implant contact mechanics rather than cartilage stresses [141]. which may play a vital role in the development and progression of diseases such as OA and hip dysplasia.34 OA. data from instrumented prostheses only yield mechanical estimates at the contact site rather than an estimate of the mechanics throughout the joint.

35 NUMERICAL MODELING OF HIP JOINT BIOMECHANICS

Analytical Modeling of the Hip Joint Analytical approaches to predicting hip joint biomechanics generally use the equations of statics to solve for resultant joint reaction forces. For clarity, “analytical” studies will be described herein as those that can be solved using simplified mathematical equations that do not require discretization of geometry or stipulation of material properties. Several analytical models have been developed to estimate hip joint

mechanics [29,142-145]. In previous efforts, the geometry of the acetabulum and femur was assumed spherical by calculating the average radius of the femur and acetabulum by measuring 2-D radiographs [29,142-145]. The equivalent joint reaction force was

estimated by summing zero a vertical body force (generally 5/6 bodyweight) and nonvertical abductor muscle force. Contact pressures were found by distributing the force over the estimated area. Two analytical models have been developed to compare mechanics between normal joints and those affected by acetabular dysplasia [29,142]. Mavcic et al. used a mathematical model of static, single-leg stance based on AP radiographs of normal and dysplastic subjects [142]. Dysplastic hips had significantly larger peak contact stresses than healthy hips (7.1 kPa/N and 3.5 kPa/N, respectively). Michaeli and co-workers also demonstrated notable differences in the location and magnitude of contact stresses between normal cadaveric pelvi and plastic pelvi with simulated dysplasia [29]. Although these studies further support the notion of pathological biomechanics and in particular increased contact stresses in the dysplastic hip, they neglected several

36 important aspects of the biomechanics. For example, idealized geometry was used to represent all or part of the hip articulation, neglecting the issues of regional and patientspecific congruency between the femoral and acetabular cartilage layers. Ignoring

cartilage geometry and assuming joints to be concentric likely lead to erroneous estimates (underestimation) of contact pressure.

Computational Modeling of the Hip Joint For clarity, “computational” models will be described as a subclass of numerical models, which requires the geometry of interest to be discretized into smaller mathematical problems. Constitutive equations and boundary conditions also must be stipulated during the development of a computational model. The use of computational modeling is an attractive method for studying hip joint biomechanics. Computer models have the ability to predict bone and cartilage stresses and strains throughout the continuum of interest rather than at select measurement locations. With the advent and availability of medical imaging techniques, individual patient models can be developed by segmenting image data, which contains detailed geometry and estimates of mechanical properties. Therefore, gross simplifying assumptions regarding hip joint geometry (i.e. spherical geometry, concentric articulation of cartilage) are not necessary. Finally, with the exception of radiation exposure during CT, computational models can be developed noninvasively using living subjects, allowing the analysis of individual patients.

37 Constitutive Models for Bone and Cartilage Besides providing insight into the contribution of different tissue components to overall tissue material behavior, constitutive models are necessary to represent the properties of tissue in computational models. Constitutive equations with particular relevance to this dissertation will be discussed further. Although viscoelastic, bone can be considered as an elastic material for many applications [146]. Therefore, under the assumptions associated with linearized

elasticity, the material behavior is characterized by a fourth-order elasticity tensor C in the generalized Hooke’s law that relates the Cauchy stress T to the infinitesimal strain tensor ε :

T = C : ε ⇔ Tij = Cijkl ε kl .

(1.1)

In its most general form, the elasticity tensor involves 21 independent elastic coefficients that must be determined experimentally. In the case of orthotropy, three orthogonal planes of symmetry exist, leaving nine independent coefficients. The number of

independent elastic coefficients is reduced further to five for the case of transverse isotropy and to two for the case of isotropy. Cartilage has been modeled as isotropic-elastic [147], isotropic biphasic [99,148], transversely isotropic biphasic [149], poroviscoelastic [150], and as a fibril reinforced poroelastic material [103]. When cartilage is loaded instantaneously, the response is equivalent to that of an incompressible material [99,102,148,151-154]. In such instances the use of a linearized elastic constitutive model may be appropriate. However, the accuracy of model predictions will degrade as strains increase since the assumption of

38 infinitesimal strain results in spurious strains when the continuum undergoes rigid rotations. Hyperelasticity is based on the existence of a strain energy potential, which generally results in a nonlinear relationship between stress and strain. Poroelastic

constitutive models describe the relative contributions of solid and fluid phases to the overall material behavior of cartilage. They were originally developed to describe the mechanics of soils [155,156] and were extended to cartilage using the biphasic theory developed by Mow et al. [99].

Finite Element Modeling of Hip Joint Biomechanics The finite element (FE) method is a proven technique that has been used extensively to evaluate biomechanical systems. The FE method allows an analyst to obtain a solution for the stress and strain distribution throughout a continuum when the applied loads, boundary conditions and material properties are known. FE model

construction can be divided into three distinct steps: 1) pre-processing, 2) stipulation of boundary and loading conditions, and 3) analysis and post-processing. Pre-processing involves discretization of the geometry of interest into small finite elements and has historically been the most challenging aspect when constructing FE models of human joints. However, with the development of commercial segmentation programs this task has become much less daunting as the process is generally automated [157,158]. Next, loading and boundary conditions are applied to the model to govern displacements of individual elements. Constitutive equations and associated material coefficients are

specified data. The discretized equations of motion, based on minimization of potential energy, are solved to obtain the displacement field. Strains and then stresses are

However. In this study predicted pressures were irregularly distributed over the surface of the femoral head with values of peak pressure on the order of 4 MPa. The earliest FE contact model was reported by Brown and DiGioia [162]. Peak cortical bone von-Mises stresses were well below 1 MPa. used measurements from a 3D coordinate measuring machine to generate contours of the horizontal sections of the iliac bone. they apparently made a mistake in calculating the magnitude of load from kgf to N whereby instead of multiply by 9.161-163] with either rigid [115. model predictions are post-processed to facilitate visualization and data analysis. Dalstra et al. Finally. A few 3D FE models have been developed to predict bone strains in intact hips [3. . making their results almost a factor of 100 too small [3].8. they divided by 9. [3] used CT image data to develop a realistic 3D model of the human pelvis. Cortical bone was assigned a spatially varying thickness. Simulated muscle forces and bodyweight were applied. FE predictions of von-Mises stress were on the order of ±4 MPa and were in fair agreement with experimental data although no statistical tests were conducted to quantify model accuracy.163]. Nearly all FE hip joint modeling studies that have analyzed cartilage contact have used two-dimensional. An FE mesh of the pelvis was constructed using these contours.8.161] or deformable bones [162.160].39 computed from the displacement field. FE predictions of cortical bone stress were compared to those measured on the cadaveric pelvis for purposes of validating the model. based on measurements from CT image data.159. Oonishi et al. plane strain models [115. Strain gauges were attached to a pelvis to experimentally measure cortical bone strains.

it is evident that the cartilage contact FE models developed thus far have a tendency to predict lower contact pressures when compared to in-vitro studies that have loaded cadaveric joints with similar forces. steep pressure gradients were measured. pressures were on the order of 5 MPa and a rather uniform contact distribution was observed. assuming femoral acetabular surfaces to be spherical and congruent. Macirowski et al. This is the only FE study to date to explicitly model the acetabular cartilage thickness.2 mm) likely influenced the cartilage sealing process since during early step loading high local maxima and irregular. [163] developed a similar model based on geometry from a radiograph.40 Rapperport et al. At 1000 N of applied load peak. [115] utilized a combined experimental/analytical approach to model fluid flow and matrix stresses in a biphasic contact model of a cadaveric acetabulum. At the instant load was applied peak contact pressures measured by the prosthesis were on the order of 5 MPa. With the exception of Macirowski’s study. who concluded that the highly irregular pressure profiles observed during an instrumented prosthesis in-vitro study were likely due to the cartilage thickness distribution and irregularities at the calcified cartilage interface [138].75 MPa. The . This argument parallels that made by Rushfeldt et al. The acetabulum was step loaded to 900 N using an instrumented femoral prosthesis. An important conclusion made in this study was that even small variations in sphericity (up to 0. When the experimentally measured total surface stress was applied to the FE model average predicted pressures (solid stress + fluid pressures) were approximately 1. Rigid bone models yielded predictions only slightly different than the deformable bone model.

88 MPa.135. Discrete Element Analysis Discrete element analysis (DEA) (a variant of the FE method) has been used to analyze hip joint contact mechanics [164. Peak contact pressures for dysplastic and asymptomatic hips (contra-lateral joints) ranged from 3. . [168] constructed patientspecific. They concluded that bone irregularities caused localized pressure elevations and that asymptomatic hips had pathological mechanics.56 – 9. Although the computational modeling literature suggests that normal joints may be modeled as spherical structures with concentric articulation [164. There were significant differences between the normal control and the asymptomatic hips and a trend towards significance between asymptomatic and symptomatic joints.165]. While these models represent the current state of the art for modeling patient-specific hip joint contact mechanics it remains unknown if the predictions were accurate as the computational modeling protocol was not validated.166. contact FE models using CT arthrography image data. Very recently patient-specific FE models have been developed to elucidate the biomechanics of dysplastic hip joints [168]. Russell et al.167]. non-linear.41 discrepancy between predictions is most likely attributed to the fact that prior computational studies assumed spherical geometry and concentric articulation. it has been well documented that the hip joint is neither spherical nor has cartilage with uniform thickness [9. Cartilage layers were represented using a series of discrete compressive spring elements and tangential shear elements.171].170. A normal model was also generated which was based on geometry from the Visible Human Project [169].115.

and stair climbing were relatively low in comparison to previous experimental measurements. DEA may be effective [172]. . They determined that the joint contact area and normalized peak contact pressure were significantly different between men and women. gross simplifying assumptions are made using this technique (e. which provides an explanation of why DEA estimates of peak pressure substantially underestimate peak pressure when compared to experimental data. which again was substantially lower than experimental data form the literature under similar loading conditions.5 MPa.42 Yoshida et al. [171] developed a dynamic DEA model to investigate the distribution of hip joint contact pressures using in-vivo data from the literature. The normalized peak contact pressure was around 1. generation of 3D models from 2D radiographs). descending stairs. However. because direct experimental validation has not been performed for this technique. [170] generated 3D contact hip DEA models using 2-D radiographs by assuming that the femoral head and the acetabular surface were spherical in shape.g. In a similar study Genda et al. For certain limited instances. Furthermore. The model assumed spherical geometry for the femur and acetabulum and concentric articulation. Peak pressures during simulated walking. it is unclear whether or not DEA has the ability to accurately predict hip joint contact mechanics.

1999. and van Erning. M. "Structure and Function of Articular Cartilage and Meniscus. D. Hayes." Ann Internal Med. "Aging Changes in Thickness of the Proximal Femoral Cortex. A. and Seedhom. Gore. 2004. "Physicochemical Properties of Articular Cartilage. C.. XIII. Ganz. B. 272-8.. pp. Lawrence.. 1995. 635646. L. London. C. 58. C. J. K. pp. 215290.. Amsterdam.. 1997. R. M. 948-53. V... Correlations among Roentgenographic Measurements and Hip Strength and Mobility. Springer New York.. IOS. Murray. 1979. Part I the Diease and Its Risk Factors. Gray's Anatomy of the Human Body. pp. E. pp. pp. vol. T. pp. Maroudas. R. A. and Dieppe.a Practical Guide to Assessment and Treatment. Mow and W. Felson. M. "Thickness of Human Articular Cartilage in Joints of the Lower Limb. 113-177. Pitman Medical. Shepherd. pp. and Sepic. S. 1918. [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] ." Am J Phys Anthropol. 59. P. Huiskes. 2002. Gardner." J Bone Joint Surg Am.43 REFERENCES [1] Gray. 2000. 27-34. "Osteoarthritis: New Insights. A. "Roentgenographic Measurements after Muller Total Hip Replacement.. P. Philadelphia. Bryant. V. Dalstra. B. R.. 2. The Netherlands. Lippincott-Raven.." J Biomech Eng. Mow. NY. 121-30. 20 ed. Ed. Ferguson... Freeman. pp. 133.. Articular Cartilage Lesions.. F. Philadelphia." in Adult Articular Cartilage. 1977.. D. S. G. "Development and Validation of a Three-Dimensional Finite Element Model of the Pelvic Bone." Ann Rheum Dis. H. B. M. A. 117. 2003.." J Biomech. Lea and Febiger. R. 59... 1982. and Ratcliffe." in Basic Orthopaedic Biomechanics. D. C. J. M. Eds. Ericksen. 171-8. Mechanobiology: Cartilage and Chondrocyte vol. and Ito. "An in Vitro Investigation of the Acetabular Labral Seal in Hip Joint Mechanics. 36. T.

. and Howell. S. and Bayliss.." Semin Arthritis Rheum. 322... P. E.. 536-7." Semin Arthritis Rheum. Haut. E. A. G. 221-34. D. Rosenberg. 21.. "Mechanical Properties of Canine Articular Cartilage Are Significantly Altered Following Transection of the Anterior Cruciate Ligament.." J Orthop Res." Acta Orthop Scand Suppl. PA. Trauma and Workload with Coxarthrosis. 2007. F.. Mow. 1994. Knekt. D. M. C. 63-77. O. 1994. 8. D. A. D. 64. "Factors Associated with Osteoarthritis of the Knee in the First National Health and Nutrition Examination Survey (Hanes I). Radin. V. T." Arch Biochem Biophys. 120. 179-89. and Boyd. 13-5. C. "An Investigation of Biphasic Failure Criteria for Impact-Induced Fissuring of Articular Cartilage. 1995. B. Anderson. Brown. 451-63. Disuse. L. Gregory.44 [12] Callaghan. Jeffrey. T.. B. J.. D. J. 1984. Felson. Caterson. 6-9. pp. B. Makela.. C. L.. L. A. E. Setton. J. R. Aromaa.217 Persons. Burr. E." Am J Epidemiol. and Altiero. Pita. I.. and Rubash. Burr. Impivaara.." J Rheumatol Suppl. S. and Aspden.. and Felson.. A Health Survey of 7. Boyd... 1991. [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] . 20... 87-96. Evidence for an Association with Overweight. 2nd ed. "Mechanical Determinants of Osteoarthrosis. "Osteoarthritis and Articular Cartilage Use.. 513-8. T. E. R. R. L. J. Radin. pp. 128.. Caterson. D.. pp. B. M. "Association of Overweight. T. J. C.. Fyhrie. pp. Buckwalter. pp... Race. and Physical Demands of Work. J. The Adult Hip. Muller. 1990. D. and Goodwin. D. "Do Occupation-Related Physical Factors Contribute to Arthritis?. 43. pp. J. J." J Biomech Eng. "Proteoglycans of Articular Cartilage: Changes in Aging and in Joint Disease. 12-33.. Martin." Acta Orthop Scand. M. D. D." Baillieres Clin Rheumatol." J Orthop Res. M.. Heliovaara. 1998. 1995. 1988. R. pp. and Sievers. W. 12.. Philadelphia. B. Lippincott Williams and Wilkins. and Abuse: Experimental Studies. "Osteoarthrosis--the Orthopedic Surgeon's Perspective.. A. 1995.. K.. "Effects of Mechanical Loading on the Tissues of the Rabbit Knee.. 2. pp. 266. Atkinson. 1993. pp.. "Matrix Damage and Chondrocyte Viability Following a Single Impact Load on Articular Cartilage. T. Hardingham. vol. R. 12-21. Radin. H.. N. J. pp.. pp..

Trousdale. M. S. O.. K." Med Eng Phys. A. "Acetabular Dysplasia in the Adult. B. o. W."Hip Implants. Philadelphia. T.. PA. T. and Lark. AAOS. M... H. Park.. A. T. W.. "Surgical Treatment of Developmental Dysplasia of the Hip in Adults: I. pp. 181-7. 205-12. Peters. S. Y. D. A. L. Berry. pp." in The Hip and Its Disorders.. Cooperman. 24. Nonarthroplasty Options. and Stulberg. Morrissy.. E. pp. S. 2002. Matsuda. "Rotational Acetabular Osteotomy for Acetabular Dysplasia.. and Kurosawa. "Long-Term Results of Rotational Acetabular Osteotomy in Young Patients with Advanced Osteoarthrosis of the Hip. T. Springer-Verlag. M. James. pp.. A. 2001." J Biomech Eng." J Bone Joint Surg Br. H. M. K. Ed. [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] . 1997. 2007. Maezawa.. A... "Bernese Periacetabular Osteotomy. E. S. M. Nuttall. WB Saunders. o. S. Badger. 1983. N. 321-33. 245-6. R. Kurz." Clin Orthop. Moro." J Am Acad Orthop Surg.. J. S.. Michaeli. M. A. Steinberg. S. Frank. 1998.. 19. Takatori. C... Murphy.. "Injurious Mechanical Compression of Bovine Articular Cartilage Induces Chondrocyte Apoptosis. R. M. 180-6..." Clin Orthop.. D. 336-41. Ninomiya. I. Bombelli. 5. S." J Orthop Sci. M. 1991.. K.." American Academy of Orthopaedic Surgeons AAOS. and Altiero. 1983. J. A Follow-up for More Than Ten Years. A. pp. A. and Dunn.. Coleman. Loening... B. J. K. T. R. R. I. pp. H... E. 2007. Nozawa... 2000. 120. "Congenital Dislocation of the Hip. pp. Grodzinsky.. "Comparison of Predicted and Measured Contact Pressures in Normal and Dysplastic Hips.. Germany. E. Blake. A. E. and Nagai." National Institute of Arthritis and Musculoskeletal and Skin Disorders.. M.. Haut... D. Osteoarthritis of the Hip. Sanchez-Sotelo. R." Arch Biochem Biophys. Berlin.. 2000. K. 10. B. 565-9.... "Triple Innominate Osteotomy in Young Adults for the Treatment of Acetabular Dysplasia: A 9-Year Follow-up Study. 59-65. Takatori."Questions and Answers About Hip Replacement. "Impact-Induced Fissuring of Articular Cartilage: An Investigation of Failure Criteria. and Hipp. D. 381... Hung. and Cabanela. E. 79-85. AAOS. J. J. Shitoto. Fukushima. Levenston. pp. 2001. pp. R. Nakamura. 2002. S. S.45 [24] Atkinson. C. 84. Y. Morimoto." Orthopedics.. O.. H. S. H. Wallensten.

2001.. F. "The Aetiology of Primary Osteoarthritis of the Hip. A. and Courpied.. 88. O.. and Coggon. "Prevalence of Hip Osteoarthritis and Acetabular Dysplasia in French and Japanese Adults. 239-45. 810-24.. pp. 621-50. Kim. G. D. J. 2000. Wickham. Moskowitz. 1999. Lam. K. J. Ali-Gombe. C. M.. J. D... pp. 1986. P. T. Philadelphia. pp. A.. 39. J. MO. 1995. and Coggon. "Patterns of Osteoarthritis of the Hip. Harris. S. and Croft. S." J Bone Joint Surg Br.. Ushiyama... Inoue.. C. C.. 58.. P. W. D. P. Hukuda. pp. "Osteoarthritis of the Hip and Acetabular Dysplasia. 1976. W.46 [37] Siebenrock. W. 2033. [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] . 745-8. H. L. and H. "Acetabular Dysplasia and Development of Osteoarthritis of the Hip. H. "Periacetabular Osteotomy: The Bernese Experience. Croft. Eds. and Ganz.." presented at Proceedings of the second open scientific meeting of the Hip Society. 81. 38. R. and Harris. "Is the Hip Involved in Generalized Osteoarthritis?. Leunig. J.. 308-10. E.. R. D." R. 325-8. "Retroversion of the Acetabulum. G. D. 50. pp. W. A.. "Osteoarthritis: Diagnosis and Medical/Surgical Treatment. 1992.. "Osteoarthritis of the Hip and Acetabular Dysplasia in Nigerian Men. M.. 1965. 512-5. Bircher.. D. M. pp. and Allen.. Huang.. 1991. pp. Howell. St Louis." Clin Orthop. 1974. Croft. 727-9.. Cooper. Stulberg. Cooper." Br J Rheumatol. A. Goldberg. K. and Silman. 1992. Wicart." J Bone Joint Surg Br. 31. "The Relationship between Acetabular Retroversion and Osteoarthritis of the Hip.... pp. P. Y.. pp. Croft. 743-4. Wilson. C. 176-83. Solomon. Silman." Ann Rheum Dis. "Hip Osteoarthritis and Dysplasia in Chinese Men.." Ann Rheum Dis..." Instr Course Lect. pp.. Lau." J Bone Joint Surg Br. Andrew. pp. pp." Rheumatology (Oxford)." J Rheumatol.. P. 23. V.. D. M. R. WB Saunders. "Etiology of Osteoarthritis of the Hip. Kawasaki.. 50. P. 2006. 965-9. C. Mankin. R." Br J Radiol. Hutchinson. Lin. and Poss. 54. T. 1996. E.. S. Wickham. Murray. M.

AMD 36. NY.. pp. J. S. W.. Ed. G.. S." Am J Anat. S.. L." J Biomech. and Bang. pp. 17. vol. C. S. "Ultrasonic Wave Propagation in Human Cortical Bone--Ii. 383-96. pp. New York.. T. P. pp. Yoon. Van Buskirk. G. pp. 1969.. de Chim et de Phys. B. P." IEEE Trans Biomed Eng. and Vincentelli. S. H. "Compact Bone as a Non-Isotropic Material. Springer. 1847.47 [50] Wertheim. Zioupos.. pp. 21. Wolff. Hirschwald. 1970. "Experimentally Determined Microcracking around a Circular Hole in a Flat Plate of Bone: Comparison with Predicted Stresses. A. Currey.. S. pp. S. C. D. M. Maharidge. Measurements of Elastic Properties and Microhardness." J Biomech. 8.. Meunier. Das Gesetz Der Transformation Der Knochen. Berlin. R. Evans." Am. 331-62. "Relation of Collagen Fiber Orientation to Some Mechanical Properties of Cortical Bone. T. pp.. H.. Reilly. Lang. and Burstein. G. Dempster... H. pp. Evans.. T. and Christel. R.. J. "Elastic Coefficients of Animal Bone. "Physical and Histological Differences between Human Fibular and Femoral Compact Bone. pp. 63-71. C. 9. D. W. 1995. 459-64.." in Studies of the Anatomy and Function of Bone and Joints." J Biomech. 1984. "Ultrasonic Method for Measuring Elastic Coefficients of Bone and Results on Fresh and Dried Bovine Bones. pp. 1976. and Barton.." Philos Trans R Soc Lond B Biol Sci. Cowin.. "Memoire Sur Pelasticite Et Al Cohesion Des Principaux Tissus Du Corps Humain. S31-6. L. R. and Ashman. J. "The Elastic Moduli of Bone. 91. "The Effects of Remodeling on the Elastic Properties of Bone. 36 Suppl 1. F. Berlin. 1981. 165." Calcif Tissue Int." Science.. Katz." in Mechanical Properties of Bone. 2. Ed. 1892. 393-405. B. J. ASME. Lang. 1975. R. 1952.. A. B. 385-414. and Liddicoat. G. Lipson. [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] . Yoon. S. "The Elastic and Ultimate Properties of Compact Bone Tissue. 1966. D. Evans. 2878. F. 142-155. 347. 101-5. 1969. and Katz. F. 131-143. Mirza.

" J Orthop Res.. 1966. "A Continuous Wave Technique for the Measurement of the Elastic Properties of Cortical Bone. "An Approach for TimeDependent Bone Modeling and Remodeling-Application: A Preliminary Remodeling Simulation. T. Biomechanics of the Locomotor Apparatus. D. C. 20. H. 1989. and Weaver. 33.. C. 1325-30. An Evaluation of Some Methods for Measuring Osteoporosis..48 [62] Ashman. 1981. Fudala." J Biomech. 231-244.. Beaupre... and Snyder.. J. Hayes. Odgaard.. B.. F.. G. 17. 662-70. pp. C."Numerical Simulation of Bone Adaptation to Mechanical Loading. J." J Elast. R. 651-61.... C. 20. M. P. 2000. D. Ed. T. "Toward a Quantitative Formulation for Wolff's Law in Trabecular Bone. E. 1990. 1994. H.. Schiffers.. and Rice. 1135-50. New York. G. pp. Cowin. NY. W.. Springer. 299-308. T. 349-61. 1998. R. Orr. C." J Biomech.." J Biomech. D. Wirtz. Grootenboer. pp. R.. Chalmers.. pp. 43-68. Carter. and Cowin. Beaupre. and Carter." J Orthop Res. J. pp. D." J Biomech.. N. "The Anisotropic Hooke's Law for Cancellous Bone and Wood. Weinans.. Jacobs. ASME. K. 53.. and Carter. J." Stanford University. 1984. Yang.. "Critical Evaluation of Known Bone Material Properties to Realize Anisotropic Fe-Simulation of the Proximal Femur.. Fyhrie. Orr. R. "Cancellous Bone: Its Strength and Changes with Aging and an Evaluation of Some Methods for Measuring Its Mineral Content. 48. R. 1987. "Relationship between Loading History and Femoral Cancellous Bone Architecture. pp. and Fyhrie. van Rietbergen.. 8." in Mechanical Properties of Bone. pp. E." J Biomech. D. Carter. S. Kabel. Huiskes. S. B. R.. vol.. D. and Forst. AMD Vol 45. J. D. R.... Berlin. E. Weichert. 125-46. pp. C.. 785-94. E. Dalstra. R. 22. Pauwels.. T. S. Huiskes. R. B. G. 8. T. D. and Whalen.. "An Approach for TimeDependent Bone Modeling and Remodeling--Theoretical Development. J." J Bone Joint Surg Am. C. S. P. pp. "Adaptive Bone-Remodeling Theory Applied to Prosthetic-Design Analysis. D. S.. [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] . "Trabecular Bone Density and Loading History: Regulation of Connective Tissue Biology by Mechanical Energy. Pandorf. 1990. 1980. Van Buskirk.. Radermacher.. pp. K. 1987. Cowin. T. and Slooff.Ii. A. W. B. R... Orr.

313-7. 1974. L. D." Med Biol Eng.. E. 97. Hunter... M." Philosophical Transactions. de Meester. 24. 248-256. "The Matrix. "Uber Die Spaltrichtungen Der Gelenkknorpel. H. and Mulier. and van Erning.. P.. Dalstra.. 523-35. 1898. 42.. 1991. Huiskes. J.. Meachim. Denham. Odgaard. pp. pp. 12. R. Huiskes. P. "Material Properties of Femoral Cancellous Bone in Axial Loading. 1993.." J Bone Joint Surg Am. Jr. London. G. 12. K." Dan Med Bull. F. 257-62. 41. pp. Linde. 1994.. pp. L.. Dependency on Strain Rate.. J. "Collagen Alignments and Artificial Splits at the Surfaces of Human Articular Cartilage. Schoenfeld. A. K.. R. pp. P.. C. "Mechanical Properties of Trabecular Bone. Pitman. F. Part Ii: Time Dependent Properties. 119-38. and Soballe." J Biomech. 514-521. E.." J Biomech.. 1994.. 12. A. H. 1992. 803-9. 26. and Meyer." in Adult Articular Cartilage. "Comparative Study of the Intrinsic Mechanical Properties of the Human Acetabular and Femoral Head Cartilage. Ed. 101-118." J Biomech. A.. Bergmann. Lautenschlager. pp. A. P. R. Hultkranz. Zilch. pp. pp. "Mechanical and Textural Properties of Pelvic Trabecular Bone.. 1743. "The Compressive Behavior of Bone as a Two-Phase Porous Structure." J Orthop Res. C.. R. H.. "Of the Structure and Diseases of Articulating Cartilages. G. Meachim. "The Mechanical Behaviour of Intracondylar Cancellous Bone of the Femur at Different Loading Rates. and Emery.... Heymans. pp. W.. A. I. H. R. R." Arch Orthop Trauma Surg.. Odgaard.. 118. 1977. C. 59. Aernoudt.. 1425-41.. "Mechanical Properties of Human Cancellous Bone in the Femoral Head. 1974. G..49 [74] Weinans. Linde. M.. 25." Verh Anat Ges. D. 954-62. J. [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] . 1977. and Stockwell. P. 747-62.. R. and Kolbel.. W. 10. A. and Dzida. A. Carter. Agarwal. pp. pp... "Elastic and Viscoelastic Properties of Trabecular Bone by a Compression Testing Approach. 1979. and Grootenboer. F. pp. pp. Norgaard. 340-9. J. and Hayes. I. M. 1-67. Martens.. 2nd ed. Hvid. Freeman. Ducheyne." J Biomech. "The Behavior of Adaptive Bone-Remodeling Simulation Models. M. Athanasiou. 1980." J Anat.. W. Rohlmann..

W... C. C. 437-46. 1102-17. "Anisotropy. S. pp. Huang. L. pp. M. 557-67. and Seedhom.. C. E. M.." J Biomech. Influence of Ionic Conditions. 12. S. 9. C. Schinagl. W. pp. 499-506. V.. T. 456-72. Muller. C. 1127-8. H. 339-53. "The Tensile Properties of the Cartilage of Human Femoral Condyles Related to the Content of Collagen and Glycosaminoglycans. N.. Pita. 38. 1997.. T.. F." J Biomech.. B. pp. A. 785-91. and Manicourt. "Measurements of Nonhomogeneous. D." Biochim Biophys Acta. pp.. D.. H. pp. D. V... 124. "The 'Instantaneous' Compressive Modulus of Human Articular Cartilage in Joints of the Lower Limb. and Mow. 1968. C. G. and Tuke. Weight Bearing. Jemmott.. 1976. Wang. Kuei. S.. 2002. 4. "An Automated Approach for Direct Measurement of Two-Dimensional Strain Distributions within Articular Cartilage under Unconfined Compression. O. C.. A. Gomez. 799-809. 38. J. pp. C.. F." J Bone Joint Surg Am.. Chen. C. Gurskis. 1979. 2003. pp. "Large Deformation Nonhomogeneous and Directional Properties of Articular Cartilage in Uniaxial Tension.. 1980. R." J Orthop Res. Deng... "The Intrinsic Tensile Behavior of the Matrix of Bovine Articular Cartilage and Its Variation with Age.. Kempson. A. Woo. S. and Tension-Compression Nonlinearity of Human Glenohumeral Cartilage in Finite Deformation. "Tensile Properties of Articular Cartilage.. L.. and Ateshian. 1973. D. C. G. S. Lubock. Stankiewicz. 1986. Freeman." J Biomech Eng." Rheumatology (Oxford). J. pp. "Tensile Properties of Human Knee Joint Cartilage: I. C. and Sah. Chahine. pp. C. H. G. L. Ateshian. and Mow. 36. E.50 [87] Shepherd. G. A. R. S. F." J Orthop Res. Mow. A. V. "DepthDependent Confined Compression Modulus of Full-Thickness Bovine Articular Cartilage. B. 379-92. Roth. V. and Akeson. G. 297. and Jemmott. M.. Woo.. 124-32. P... Hung...." J Biomech.. C. and Swanson. Akizuki. 15. M. Inhomogeneity. A. Ateshian. 2005. Directional Mechanical Properties of Articular Cartilage in Tension. and Fibrillation on the Tensile Modulus. E. "Optical Determination of Anisotropic Material Properties of Bovine Articular Cartilage in Compression. Wang. [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] . A. pp. Howell." J Biomech. Kempson.. Muir.. C. G. Akeson. and Hung.. Pollard." Nature... 62. 220.... A. Y. H. 1999. M. G.

M. 1984. W. [102] Soltz.. J. B. J. P. "A Conewise Linear Elasticity Mixture Model for the Analysis of Tension-Compression Nonlinearity in Articular Cartilage.. C. 16. [101] Fortin. M... S. "Biomechanical and Structural Characteristics of Canine Femoral and Tibial Cartilage.. 88-94. C. [105] Park. 99-107. pp. C. and Helminen." Eng Med. C. 65-73. [107] Arokoski.. M. and Jurvelin. [108] Mow. 1999. pp. [99] [100] Cohen. J. M. M. H. 102. "Volumetric Changes of Articular Cartilage During Stress Relaxation in Unconfined Compression. B. Kuei. Tammi.." Osteoarthritis Cartilage. G. 1980. "Fluid Transport and Mechanical Properties of Articular Cartilage: A Review.. Kiviranta. Mow. "Biphasic Creep and Stress Relaxation of Articular Cartilage in Compression: Theory and Experiments. V. Holmes. "Variations in the Intrinsic Mechanical Properties of Human Articular Cartilage with Age. [106] Wong.. M." J Biomech. Hung. D. J.." J Biomech Eng. and Armstrong. J. 120. and Mow.. Hyttinen. A.. 576-86. A.. Soulhat. D. J. S. G. and Ateshian." J Biomed Mater Res." J Bone Joint Surg Am." J Biomech Eng. A.. 1987. M.. "Mechanical Response of Bovine Articular Cartilage under Dynamic Unconfined Compression Loading at Physiological Stress Levels.. V. V. 2004.. Arokoski. M. 491-6. M. 122. 2000. pp. 17.." J Biomech Eng. 121. pp. V. 1049-54. Kovanen. J. 15-22. 2000. and Buschmann. pp. 37794.. A. Buschmann." J Biomech Eng.. G.. .. pp. S. W.. C. 1999. 189-95.. Lai.. 33. "A Transversely Isotropic Biphasic Model for Unconfined Compression of Growth Plate and Chondroepiphysis. and Jurvelin. and Lai. H. G. Ponticiello. pp. 73-84. pp. Hunziker. pp. 1998. V. H.. 48. M.. and Shirazi-Adl. pp. M.. I.. and Water Content.. M. C. W. J.. Degeneration. "Indentation Study of the Biochemical Properties of Articular Cartilage in the Canine Knee. 12. C. M.51 [98] Armstrong." J Biomech.. Lai." J Biomech Eng. 64. and Mow.. A. S. J. [103] Soulhat. pp.. E.. T. 340-7. 122. and Ateshian. Shirazi-Adl.. 2000. "Unconfined Compression of Articular Cartilage: Nonlinear Behavior and Comparison with a Fibril-Reinforced Biphasic Model. [104] Jurvelin. C. 1982. "A Fibril-NetworkReinforced Biphasic Model of Cartilage in Unconfined Compression. Helminen.

[114] Walker. 1998. [113] McCutchen. R. D. and Ateshian. Rohlmann." J Biomech Eng. 1968. Kim. Ed. [116] Huang. 512-520. V. D. D. W. "The Role of Cartilage Streaming Potential.. A.." J Biomech Eng.. Graichen. 1995.. Fijan. "Contact Pressures in the Human Hip Joint Measured in Vivo. E. 2001. G. Grodzinsky. Wong.. S. Y. W. 10-8. [115] Macirowski. and Grodzinsky. G. Longfield.. 1-7. and Ateshian. Berlin. Y. H. 1055-66. 123." in The Biomedical Engineering Handbook. Hunziker.... Deuretzbacher. L.. K. pp. and Wright.. pp. pp. E. S.. A. A. "The Role of FlowIndependent Viscoelasticity in the Biphasic Tensile and Compressive Responses of Articular Cartilage. 83. J. R. J. J. G." J Biomech. 2879-83. "Joint Lubrication. Tepic.." J Biomech. G. M.. W. T. 1981. [112] Furey. R. Frank.. pp. and Duda. V. pp. 28. J. "Stimulation of Aggrecan Synthesis in Cartilage Explants by Cyclic Loading Is Localized to Regions of High Interstitial Fluid Flow. Heller. Dowson. 2001. M. 2000. M. B. J." J Biomech Eng. A." Nature. 34. J. J. Y. 366.52 [109] Bergmann.. [120] Kim.. [117] Lee. and Grodzinsky." Proc Natl Acad Sci U S A. C. Carlson.. 27. A. R. Bonassar. D. C. L. and Mann. R. pp. 28. C." Ann Biomed Eng. 333-351. D. A. "Interstitial Fluid Pressurization During Confined Compression Cyclical Loading of Articular Cartilage. and Mann. [118] Soltz. 1999. A.. 1994.. 150-9. pp. 280-92. J. 1995. W. 103..." Ann Rheum Dis. Fluid Flow and Pressure in the Stimulation of Chondrocyte Biosynthesis During Dynamic Compression. [110] Bergmann. Free University and Humboldt University." Arch Biochem Biophys. "Hip Contact Forces and Gait Patterns from Routine Activities.. 116. C. 1284-1285. R. 410-7. "Sponge-Hydrostatic and Weeping Bearings.. F. A. L.. Frank. pp. J. S.. FL. "Cartilage Stresses in the Human Hip Joint. M. and Roylance. Strauss. M.. 184. . [119] Buschmann. pp. "Oscillatory Compressional Behavior of Articular Cartilage and Its Associated Electromechanical Properties. H. 1959. Hip98: Data Collection of Hip Joint Loading on Cd-Rom.. pp.. N... 1986. [111] Hodge.. Bronzino. G. M.... G. "'Boosted Lubrication' in Synovial Joints by Fluid Entrapment and Enrichment... Harris. Mow. Boca Raton. pp.. E. CDC Press. G. 859-71... Burgess. W. K.

. "Strain Distribution in the Proximal Human Femur. Malsizing of Endoprostheses. [123] Finlay. S. D. Huggler. 75-85. pp. 536-41. Huggler. [129] Kim. Miller. 1975. C.. Bourne.. 1989. P. D. Bhatia. 1986. and Andreae. L. "The Effects of Prosthetic Acetabular Replacement on Strains in the Pelvis. Y. Au." presented at Proceedings of the 30th Annual Orthopaedic Research Society. [124] Jacob. A.. G. P. Doong. W. Kim.. GA. S. Schreiber. [130] Afoke.. Chakrabarty. and Jacob.. H." Bull Prosthet Res. P. "Contact Pressures in the Human Hip Joint. [122] Dietschi. "Proximal Strain Distribution in the Loaded Femur.." Acta Orthop Belg. R. [128] Oh. G. Y.. Kerin.." J Biomech." J Orthop Res. "Experimental Investigation of Deformation of the Weight-Bearing Acetabulum.. pp. pp. H. Landsberg. H. 1999. D. Kim. A. J. R. 295301. Walker. 153-7. Grodzinsky. S... 619-36. Plaas. A. "Pelvic Stresses in Vitro--I. A.. and Sandy. [127] Ries. "Mechanical Function of Subchondral Bone as Experimentally Determined on the Acetabulum of the Human Pelvis. [125] Lionberger.. P. A. Dietschi.. B. [131] Adams. An in Vitro Comparison of the Distributions in the Intact Femur and after Insertion of Different Hip-Replacement Femoral Components. and Cho...." J Biomech. H. 9. 703-14.. pp.. 80-9. M." J Biomed Eng.... and Harris. 1989. G. M. C. 1984. 11.. J. J.. H.. Gurtowski. 19. pp. [126] Petty. L. 398-402. and Schreiber. and Dee. Atlanta. J.... 1980. pp.. J. Y. A.. 7. H. and Dolan. 14. J. R.. C. A. pp." J Bone Joint Surg Br. 1976.. A.. 69. and Hutton. 41 Suppl 1. "Experimental Determination of Stress Distributions in Articular Cartilage before and after Sustained Loading." Clin Biomech (Bristol... J. J. R. Y. A.. H.. S." J Bone Joint Surg Br. J. 83. W. C. and Granholm. "In Vitro Evaluation of the Effect of Acetabular Prosthesis Implantation on Human Cadaver Pelves. 625-7. 88-96. J. I. W.53 [121] Sah. P. N. pp. .. J. H. Avon). 60. Byers. An in Vitro Comparison in the Intact Femur and after Insertion of Reference and Experimental Femoral Stems. "Normal Pelvic Strain Pattern in Vitro. pp. 1987. "Biosynthetic Response of Cartilage Explants to Dynamic Compression. J. and Piotrowski. pp. 2001." J Bone Joint Surg Am.. Pugh. 10-33. B. 1978.

" J Biomech. 44. R.. [141] Mann. 164-167. 1983. R. "Contact Stress Gradient Detection Limits of Pressensor Film. Rock. J. 2005. pp. "Improved Techniques for Measuring in Vitro the Geometry and Pressure Distribution in the Human Acetabulum. Mann. 1979. and Brown.. and Shaw. 1981. R. W. Steinlechner M. 1999.." Anat Embryol (Berl)." Science. 204. T. C. "Contact Pressures in the Flexed Hip Joint During Lateral Trochanteric Loading. A. H. "Direct Measurement of Local Pressures in the Cadaveric Human Hip Joint During Simulated Level Walking. . H. D. P. D.. 14. pp. 195. "Comment On "Quantitative Determination of Joint Incongruity and Pressure Distribution During Simulated Gait and Cartilage Thickness in the Human Hip Joint". [137] Rushfeldt. 658-66. D. "Direct Comparison of Contact Areas. E. Alonso. 413-5. S. Mann. T. W. and Eberhardt. pp." Journal of Orthopaedic Research. D.. 279-88. Mann. E. [140] Carlson.. D. 532-539. 16. 315-23. M. "Influence of Cartilage Geometry on the Pressure Distribution in the Human Hip Joint. [136] Sparks. A. W. pp. J. 359-66." Journal of Orthopaedic Research.." J Biomech. W. Muller-Gerbl.... and Harris. C. J. Landgraf. B. 1997. "Quantitative Determination of Joint Incongruity and Pressure Distribution During Simulated Gait and Cartilage Thickness in the Human Hip Joint. pp.. R. Contact Stress and Subchondral Mineralization in Human Hip Joint Specimens. [134] von Eisenhart-Rothe. and Harris. Etheridge. [133] Brown.. R. W.. R.. [135] von Eisenhart-Rothe R. pp. 21. A. R... and Putz. 257-64. D. 373-84..." J Biomech Eng. Muller-Gerbl M and Eckstein F. pp. and Swanson. 1992. pp. 114. W. P. [138] Rushfeldt. 352-7. D. P.. pp... F.54 [132] Adams. Ii Instrumented Endoprosthesis Measurement of Articular Surface Pressure Distribution. 1974. D. T." J Orthop Res. "In Vitro Contact Stress Distributions in the Natural Human Hip. Beason. and Harris. Eckstein. 23. 18... pp. 1985. [139] Hale. E. 7. W..." Ann Rheum Dis. H. C. S. "A Radio Telemetry Device for Monitoring Cartilage Surface Pressures in the Human Hip." IEEE Trans Biomed Eng. W. 2002.

V.. R. 120. C..." J Biomech.. [150] Suh.. C. 2000. L. 1972.. [152] Ateshian.. "A Mathematical Analysis for Indentation Tests of Articular Cartilage. 16573. V. P. 2000. Herrmann.. 195201. "Finite Element Formulation of Biphasic Poroviscoelastic Model for Articular Cartilage. Pedersen. pp. M.. "An Analysis of the Unconfined Compression of Articular Cartilage." J Biomech. 440.. and Pedersen..... . C. Iglic. Keer. 1653-7. W. Odgaard. R. V. V.. G. M. pp.. V. 32. L. pp... R177-8. F. D. V. Brand.. "Stress Distribution on the Hip Joint Articular Surface During Gait. Theoretical Analysis. and Mow... B. and Mow. M. R. 1998.. F. Daniel. S.. V. Zhu. A. Lai. [151] Armstrong. 1025-1030. Iglic." J Biomech Eng. "Direct Mechanics Assessment of Elastic Symmetries and Properties of Trabecular Bone Architecture. pp. and Mow. M. V. B.. 5. 27.." J Biomech. M.. and Srakar.. 1984. G.55 [142] Mavcic. pp. K. A. A. 2002. V. [145] Ipavec. B. C. W. "Mathematical Modelling of Stress in the Hip During Gait. and Huiskes.. Kabel. Antolic.. Iglic. Kralj-Iglic. Good.. and Kralj-Iglic. pp." Journal of Orthopaedic Research. "A New Method to Determine the Tensile Properties of Articular Cartilage Using the Indentation Test. V. 1229-35.. W. 20. B. pp." presented at Transactions of the 47th Annual Meeting of the Orthopaedic Research Society. 29. 1347-60. Kralj-Iglic.. R275-6. Mavcic. 106.. J. Antolic. A. 431. 1987. 703-714. Lai." Pflugers Arch.. A. B.. 20. [148] Mak. [143] Mavcic. "Biphasic Indentation of Articular Cartilage." J Biomech." Pflugers Arch. 1996. 1996. R. F. A. C." J Biomech. D. [144] Ipavec. pp. Iglic. 1999. 541-51.. [147] Hayes. Brand. [146] Van Rietbergen. C. B. "Peak Contact Stress in Human Hip During Gait. pp. and Bai. G... A." J Biomech Eng. T. A. and Gardner. Pompe. "Mathematical Estimation of Stress Distribution in Normal and Dysplastic Hips. and Iglic.. K. and Mockros. "An Asymptotic Solution for the Contact of Two Biphasic Cartilage Layers.I. 1994. [149] Mow.. pp. W... M. J. R.

... 114. 39-44. M." Comput Biol Med. B. Konishi. Springer-Verlag." Laryngoscope. 953-60. 3. NY. 33.. 31.. [160] Oonishi.. Zeng. C. R. [164] Genda. J. 8.. "A Computer Simulation Study of Normal and Abnormal Hip Joint Contact Pressure.. V. S. [156] Terzaghi.. Ganz. 435-46. and Liang. 711-6. "Mechanical Analysis of the Human Pelvis and Its Application to the Artificial Hip Joint--by Means of the Three Dimensional Finite Element Method.. C.. and Hasegawa. 2006. [159] Goel. D. L. K. Chen. F.. "The Influence of the Acetabular Labrum on Hip Joint Cartilage Consolidation: A Poroelastic Finite Element Model. N. Theoretical Soil Mechanics. D... A. [158] Wang. T. pp. pp. N. pp. New York." J Biomech.. Li.. and Ito. T. H." J Biomech." Arch Orthop Trauma Surg. H. Isha. Valliappan. Mow.56 [153] Bachrach. pp. Bryant. John Wiley and Sons." in Handbuck Der Physik. Bai... W. Y. S. Xia. and Toupin." J Orthop Res. R. E. "A Contact-Coupled Finite Element Analysis of the Natural Adult Hip. 1995. 91-104. 19. pp. pp. 427-44. N. D." J Biomech. "ThreeDimensional Reconstruction and Modeling of Middle Ear Biomechanics by HighResolution Computed Tomography and Finite Element Analysis. Carter. J. "Stresses in the Normal Pelvis. R. C. "Experimental Determination of the Linear Biphasic Constitutive Coefficients of Human Fetal Proximal Femoral Chondroepiphysis. 1986. Z.. 1998. D.. M. K. Chen. T. Y. 1943. 1985. 2000. R. Berlin. 1983. D. J. "Contact Finite Element Stress Analysis of the Hip Joint. Z. H. K. and Guilak. [155] Truesdell. and Liu. and Miura.. W. 1960. [157] Lee. .. J. pp.... and Schurman. 445-51.. R. J. Ye. J.. V.. and DiGioia.. [154] Brown... pp... C. T. P. 202-6. 1984. Lee. 116. pp. and Svensson. "Incompressibility of the Solid Matrix of Articular Cartilage under High Hydrostatic Pressures. J. "ThreeDimensional Reconstruction on Pc-Windows Platform for Evaluation of Living Donor Nephrectomy. 17. 2007. [163] Rapperport. F.. and Singerman. 86. pp." J Biomech. [161] Ferguson. 437-48. 226-793. [162] Brown. 3rd. H. T.. 1978. T." Comput Methods Programs Biomed.. pp. "The Classical Field Theories. Hasegawa. 597-605. F. 16." J Biomech.

. pp. "Cartilage Contact Pressure Elevations in Dysplastic Hips: A Chronic Overload Model.. J. M. L.. 1318-33. 1. J. Y.. A. [166] Rushfeldt.. N. K. 39.. A. J. N... pp. pp. pp.. M. J. M. D.... M. 1195-1198.. S." Nature. Kitagawa. E. 2006. E.. [170] Genda. V. D. B. E. and Chao. 217. G. N.." J Bone Joint Surg Am.. T. P. D. A. W. R. N. and Pedersen." J Biomech." Journal of Orthopaedic Surgery and Research. Grosland. A. Spitzer. [169] Ackerman. G. and Harris. [168] Russell.. Goodfellow. and Grosland. H. Fetto." J Biomech...57 [165] Konishi. 2004. "Incongruent Surfaces in the Human Hip Joint. 895905. 34. [167] Bullough. Barrance. 1993. Y. P." Medinfo. Mann. A New Computerized Technique. Sphericity and Cartilage Thickness. Greenwald. "New Methods for Assessing Cartilage Contact Stress after Articular Fracture. Scherzinger. D. M. "Improved Techniques for Measuring in Vitro the Geometry and Pressure Distribution in the Human Acetabulum--I." J Biomech. 14. and Mieno. 1996-2004. J.. 1981... 2006. M. H. R. T. J. . Shivanna. "Three-Dimensional Dynamic Hip Contact Area and Pressure Distribution During Activities of Daily Living. W.. Wilckens. and Chao.. E. 1968. Iwasaki. Rudert. 2001.. MacWilliams. [171] Yoshida. pp. and Whitlock... 75. 253-60. "Determination of Acetabular Coverage of the Femoral Head with Use of a Single Anteroposterior Radiograph. M. H. P. 8. pp.. 1290. 1995. pp. Li. [172] Brown. Ultrasonic Measurement of Acetabular Surfaces. and O'Connor... J. 52-8.. Faust. "Normal Hip Joint Contact Pressure Distribution in Single-Leg Standing--Effect of Gender and Anatomic Parameters. "The Visible Human Data Set: An Image Resource for Anatomical Visualization." Clin Orthop Relat Res.

and Sensitivity Studies”. 2005. Validation.L. at the patient-specific level. Although finite element (FE) models of the pelvis have been developed. and to assess the sensitivity of FE strain predictions to assumptions regarding cortical bone thickness as well as bone and cartilage material properties. The objectives of this study were to develop and validate a realistic FE model of the pelvis using subjectspecific estimates of bone geometry. B. location-dependent cortical thickness and trabecular bone elastic modulus. Anderson.. and previous models used simplifying assumptions regarding geometry and material properties. Vol. may lead to improved treatment modalities. 127. C. Tuttle. Peters.. with kind permission from the ASME. validation by direct comparison with subject-specific strains has not been performed. pp: 364-373. Cortical bone strains Reprinted from Journal of Biomechanical Engineering. 3.A.E. 1 . Weiss. “A Subject-Specific Finite Element Model of the Pelvis: Development.CHAPTER 3 A SUBJECT-SPECIFIC FINITE ELEMENT MODEL OF THE PELIVS: DEVELOPMENT. A FE model of a cadaveric pelvis was created using subject-specific CT image data. VALIDATION AND SENSITIVITY STUDIES1 ABSTRACT A better understanding of the three-dimensional mechanics of the pelvis..D. J.. No. A. Acetabular loading was applied to the same pelvis using a prosthetic femoral stem in a fashion that could be easily duplicated in the computational model.

824). Changes to cortical bone thickness and elastic modulus had the largest effect on cortical bone strains. The FE model was less sensitive to changes in all other parameters. . baseline FE predictions were strongly correlated with experimental results (r2 = 0. with a best-fit line that was not statistically different than the line y=x (Experimental strains=FE predicted strains).59 were monitored with rosette strain gauges in ten locations on the left hemi-pelvis. FE strain predictions were compared directly with experimental results for validation. The methods developed and validated in this study will be useful for creating and analyzing patient-specific FE models to better understand the biomechanics of the pelvis. Overall.

and force telemetry data have been used to estimate joint contact forces at the acetabulum [1-3. are notorious for generating pelvic fractures. reported that nearly 76% of THA recipients exhibited signs of a dysplastic joint .a condition that went unrecognized prior to surgery [3].1421]. it has been hypothesized that subtle alterations in pelvic geometry (i.e. such as undiagnosed pelvic dysplasia. with the thin layers of cortical bone carrying most of the load. In addition to pelvic fractures. the relationship between pelvic dysplasia and osteoarthritis remains controversial since there is no direct quantitative evidence linking the two together. Simplified mathematical models. appear to be more prevalent among candidates for total hip arthroplasty (THA) than primary arthritis [10-13].5].37% [4. It would be wise to develop methods capable of quantifying the mechanics beyond the acetabular contact .. secondary causes of osteoarthritis. Michaeli et al. Despite its efficient structure. In fact. The fracture itself often causes multiple internal trauma leading to a mortality rate on the order of 12 . the pelvis can become damaged due to altered loading. Nevertheless. pelvic dysplasia) lead to osteoarthritis [6-11]. such as those generated in car accidents. Forces as high as 5.60 INTRODUCTION The acetabulum and adjoining pelvic bones are one of the most important weight bearing structures in the human body. These studies provide valuable information concerning overall joint mechanics but do not yield estimates of the surrounding bone stresses and strains. experimental contact analyses.5 times body weight are transferred from the femur to the acetabulum during activities such as running and stair climbing [1-3]. Side impact forces. The structure of the pelvis is a sandwich material.

The potential benefit of patient-specific FE analysis becomes clear when one considers how difficult (if not impossible) it would be to assemble a population of donor tissue that exhibits a specific pathology such as pelvic dysplasia. location-dependent cortical thickness and trabecular bone elastic modulus. and may present the framework necessary for preoperative surgical planning. cortical bone thickness. 2) A subject-specific FE model . which can accommodate large inter-subject variations in bone geometry and material properties. an analysis of the stress distribution in and around the pelvic joint may clarify the mechanical relationship between pelvic geometry and predisposition to osteoarthritis. or via experiments with cadaveric tissue. An alternative approach to analyze pelvic mechanics is the finite element (FE) method. diagnosis.61 interface since there is evidence to suggest that the surrounding bone plays a pivotal role in the progression of diseases such as osteoarthritis [22-24]. and to assess the sensitivity of FE cortical strain predictions to cortical bone thickness and bone and cartilage material properties. The following hypotheses were tested: 1) A FE model of the pelvis that incorporated subjectspecific geometry. and position dependent trabecular bone elastic modulus would accurately predict cortical bone strains. surgical approaches. Specifically. It is difficult to assess the stress and strain distribution throughout the entire pelvis using simplified mathematical models. A better understanding of the mechanics for the entire pelvis could lead to improved implant designs. The objectives of this study were to develop and validate a FE model of the pelvis using subject-specific estimates of bone geometry. implanted prostheses.

62 of the pelvis would be more accurate than models that assume average cortical bone thickness and trabecular elastic modulus. .

63 MATERIALS AND METHODS A combined experimental and computational protocol was used to develop and validate a subject-specific three-dimensional model of a 68 y/o female cadaveric pelvis (International Bioresearch Solutions. 20 mg/cm3 increments) was also scanned with the same field of view and energy settings. The block allowed for spatial registration of experimental and FE coordinate systems. Tucson. were removed. with the exception of articular cartilage. Kyoto Kagaku Co. [26].44x0. CT data from the BMD phantom were averaged over each block to obtain a relationship between CT scanner pixel intensity and calcium equivalent bone density. screened for large-scale osteoarthritis prior to the study. in-plane The pelvic joint was visually resolution=0.. A CT scan (512x512 acquisition matrix. The mounting and loading of the pelvis followed a protocol similar to that described by Dalstra et al. 354 slices) was obtained in a superior to inferior fashion using a Marconi-MX8000 scanner (Philips Medical Systems. Japan). Kyoto. AZ). slice thickness=0.6 mm. The iliac crests were submerged in a mounting pan of . A registration block and wires were attached to the iliac crest.400 mg/cm3. FOV=225 mm. while the wires served as a guide to reproduce the boundary conditions used in the experimental model [25].44 mm. consisting of 21 rectangular blocks of urethane with varying concentrations of hydroxyapaptite (0 . Bothell. Experimental Study The sacroiliac joint and all soft tissues. A bone mineral density (BMD) phantom (BMD-UHA. WA).

and 1.1). ischium. GA) to the depth defined by the iliac guide wires. The average time to reach quasi-static equilibrium for each loading scenario was 6 minutes.1) 3D coordinates of the strain gauges and registration block were determined in a laboratory reference frame using an electromagnetic digitizer (Model BE-3DX. Mar-Hyde.25. Immersion Corp.50. (3. ε2.0 X body weight (559 N) were applied to the acetabulum by displacing a femoral prosthesis. pubis. were attached to the left hemi-pelvis at locations around the acetabulum. The femoral implant was displaced continuously until the appropriate load was reached at which time the displacement was held constant. representing 30 channels of data.75. Atlanta. and ilium. attached to a linear actuator (Figure 3.25 N/sec for at least 60 seconds. allowing stress relaxation. CA). Vertically oriented loads of 0. NC). . 0. Ten three-element rectangular rosette strain gauges (WA-060WR-120. Geometric features of the pelvis were digitized to determine the accuracy of the geometry reconstruction. ε3) for the last 10 seconds of the equilibrium period was obtained and then converted to in-plane principal strains (εP. Raleigh. San Jose. until the load relaxed to a value greater than 95% of the original with a load-time slope less than 0. εQ) using the relationship [27]: ε P. Vishay Measurements Group.Q = ε1 + ε 3 2 ± 1 (ε1 − ε 2 )2 + (ε 2 − ε 3 )2 2 . 0. An average of the rosette gauge readings (ε1..64 catalyzed polymer resin (Bondo.

(F) mounting pan for embedding pelvis. Schematic of fixture for loading the pelvis via a femoral implant component. (B) load cell.65 LOAD A B C D E F G Figure 3.1. . (D) femoral component. (E) pelvis. and (G) lockable X-Y translation table. (C) ball joint. (A) actuator.

A 4-node. registration block. which was then decimated [29] and smoothed [30] to form the final surface using VTK (Kitware Inc. This element has three translational and rotational degrees of freedom at each node. NM). Albuquerque. Cortical bone was represented with quadratic 3-node shell elements [33]. Clifton Park. The elements were based on the Hughes-Liu shell [34. Mesh refinement tests were performed with this element using a model of a cantilever beam under a tip load with a thickness that was 10% of the beam length. 24 degree of freedom tetrahedral element was used to represent trabecular bone [32]. with selective-reduced integration to suppress zero-energy modes [36]. This approach resulted in an overlap of one cortical bone thickness between the tetrahedral solid .2). and guide wires were extracted from the CT data via manual segmentation (Figure 3. on the outer surface of the pelvis. The shell reference surface and shell element normal were defined so that the cortical thickness pointed inward toward the interface between cortical and trabecular bone. NY) [31] (Figure 3.2). The geometry of the shells was based on the nodes of the outside faces of the tetrahedral elements. which has three translational and rotational degrees of freedom per node..66 Geometry Extraction and Mesh Generation Contours for the outer cortex and the boundary of the cortical and trabecular bone. A volumetric tetrahedral mesh was created from the final surface to represent the outer cortex (CUBIT. FE-predicted tip deflections reached an asymptote of 4% error with respect to an analytical solution when at least 3 tetrahedral elements were used through the thickness of the beam.35]. Points comprising the contours were triangulated [28] to form a polygonal surface. Sandia National Laboratories.

anterior.polygonal surface after decimation to reduce the number of polygons and smoothing to reduce highfrequency digitizing artifact.the original polygonal surface representing the cortical bone was reconstructed by Delaunay triangulation of the points composing the segmented contours. A) P B) C) M L A Figure 3. A) . A . S .67 element and thin shell element. L . I inferior.CT image slice at the level of the ilium.lateral.superior.medial. B) . C) .posterior. P .2. showing the registration block (arrow) and the distinct boundary between cortical and trabecular bone. M . The elastic modulus for all tetrahedral element nodes in this region of overlap was set to 0 MPa. . Mesh refinement tests showed that the 3-node shell was nearly as accurate as using three tetrahedral elements through the thickness of the beam (< 5% error with respect to analytical solution).

B) close-up view of the mesh at the acetabulum. determined by averaging the distance between the implant and acetabulum in the neutral kinematic position. The final FE model consisted of 190. composed of 190. Acetabular cartilage was represented with 350 triangular shell elements with a constant thickness of 2 mm.000 shell elements.000 shell elements for cortical bone (Figure 3. A) FE mesh of the pelvis.000 tetrahedral elements and 31. A) B) Figure 3. .3).000 tetrahedral elements for trabecular bone and 31. The final surface mesh density was 0.5 shell elements/mm2 with a volumetric density of 2.5 tetrahedral elements/mm3.68 The density of the FE mesh was adjusted until it was at or above the beam mesh density required to achieve an error of 4%.3.

The algorithm was tested using polygonal surfaces representing parallel planes. concentric spheres.5 times the smallest distance between the nodes. The minimum value of nodal thickness was assumed to be 0. Vectors were constructed between each node on the cortical surface and the 100 nearest nodes on the surface defining the corticaltrabecular boundary.4). Cortical thickness was determined by minimizing both the distance between the nodes of the surfaces and the angle of the dot product between the surface normal of the cortical surface with that of each corresponding trabecular vector. When the above-described algorithm reported a thickness value that exceeded 1. the smallest distance between nodes on the two surfaces was used. . special consideration of thickness was necessary (Figure 3. and layered boxes with varying mesh densities. In areas of high curvature (such as the acetabular rim).44 mm/pixel).69 Position-Dependent Cortical Thickness An algorithm was developed to determine the thickness of the cortex based on the distances between the polygonal surfaces representing the outer cortex and the boundary between the cortical and trabecular bone. FOV/512 = 0.44 mm or the width of one pixel (FOV= 225.

70 A) B) C) n n n Figure 3. Case A) the smallest angle of the dot product between the cortical node and nearest trabecular node neighbor yields the desired thickness measurement. Case B) the smallest distance between nodes provides the desired thickness measurement. Case C) the normal vector (n) from the cortical node does not intersect the trabecular surface. Nodes on the cortical surface are represented as open circles. For cases B and C. a weighting scheme was applied such that the smallest distance between the nodes was taken as the cortical thickness when the originally reported thickness value exceeded 1.4. Schematics illustrating the special cases considered in determination of cortical thickness. . while nodes on the trabecular surface are shown as filled circles.5 X the smallest distance between nodes on the two surfaces. Both the distance between the surfaces and the angle of the dot product between the normal vector (n) with that of the vector created by subtracting the trabecular and cortical node coordinates were considered.

A) B) Figure 3. Lucite disc. Ten aluminum tubes (wall thickness 0. B) crosssectional CT image of the cortical bone phantom. The inner and outer circumferences of the tubes were segmented from the CT image data using the same technique to extract the pelvic geometry.127– 2. The centers of the aluminum tubes were filled with Lucite rods so that both the inner and outer surfaces of the tubes were surrounded by a soft tissue equivalent material [38. The phantom was scanned with a CT scanner and manually segmented to determine the accuracy of cortical bone reconstruction. The phantom was scanned with the same CT scanner field of view and energy settings used for the cadaveric pelvis and bone mineral density phantom.921 mm) were fit into a 70 mm dia. Aluminum has x-ray attenuation coefficient that is similar to cortical bone [37].39]. . A) Tissue equivalent phantom containing 10 aluminum tubes used to simulate cortical bone with varying thickness.5) [37]. The surfaces were meshed and the thickness algorithm was used to determine wall thickness.71 Assessment of Cortical Bone Thickness A custom-built phantom was used to assess the accuracy of cortical thickness measurements (Figure 3. Changes in thickness can be seen for the thicker tubes but become less apparent as the tube wall thickness decreases.5. The z-axis of the scanner was aligned flush with the top edge of the tissue phantom to prevent volume averaging between successive slices.

4095). a relationship was used to convert calcium equivalent density ( ρ ca ) to apparent bone density ( ρ app ) [40]: ρ app = ρca 0.29 [26]. Nodal moduli were averaged to assign an element modulus.1 MPa and 0.626 .0008i INT − 0.2) Here ρ ca is the calcium equivalent density of trabecular bone (g/cm3) and INT is the CT scanner intensity value (0 . (3. . A linear relationship was established between CT scanner pixel intensity and calcium equivalent density using the CT image data from the BMD solid phantom: ρca = 0. (3.41 MPa.4) where E is the elastic modulus (MPa) and ρ app is the apparent density of the trabecular bone (g/cm3). Baseline material properties for cortical bone were E = 17 GPa and Poisson’s ratio (ν) = 0.4 [42].3 ( ρ app ) 2. Coefficients C1 and C2 were selected as 4. Next.3) Finally.72 Material Properties and Boundary Conditions The femoral implant was represented as rigid while cortical and trabecular bone were represented as isotropic hypoelastic.46 . respectively with Poisson’s ratio=0.8037 (r 2 = 0. Acetabular articular cartilage was represented as a hyperelastic Mooney-Rivlin material [41].9938) (3. an empirical relationship was used to convert apparent density of pelvic trabecular bone to elastic modulus for each node [40]: E = 2017.

1). a transformation was applied to the FE model to align it with the experimental coordinate system. Livermore. The trabecular elastic modulus and cortical thickness were varied to reflect the median and inter-quartile range estimated computationally. Contact was enforced between the femoral implant and cartilage while load was applied to the implant using the same magnitude and direction measured experimentally. trabecular bone Poisson’s ratio. Each model required Sensitivity Studies Sensitivity studies were performed to assess the effects of variations in assumed and estimated material properties and cortical thickness on predicted cortical surface strains.1 GB of memory. Variations in assumed parameters were based on standard deviations from the literature (Table 3. The assumed parameters were cortical bone Poisson’s ratio. The estimated parameters were trabecular elastic modulus and cortical bone thickness. The FE models included . approximately 3 hours of wall clock time and 1. CA) on a Compaq Alphaserver DS20E (2 667 MHz processors). To establish the neutral kinematic position. A corresponding coordinate system was established for the experimental measurements using the digitized coordinates of the registration block [25]. and cortical bone elastic modulus. cartilage elastic modulus. Analyses were performed with the implicit time integration capabilities of LS-DYNA (Livermore Software Technology Corporation.73 A FE coordinate system was created from the polygonal surface of the reconstructed registration block. Nodes superior to the iliac guide wires and nodes along the pubis synthesis joint were constrained to simulate the experiment.

The sensitivity of each model. constant trabecular elastic modulus (CTEM). with alterations in cortical bone Poisson’s ratio (SSCV).5) The numerator in (5) is the percent change in slope of the best-fit lines between the sensitivity model and baseline subject-specific model. the change in constant model input parameters was used in the denominator.74 constant cortical shell thickness (CST). trabecular bone Poisson’s ratio (SSTV). cortical elastic modulus (SSCM). A sensitivity model (OVERLAP) was analyzed to determine the cortical surface strain effects due to overlap between the cortical shell and tetrahedral elements. For the overlap model the tetrahedral surface nodes were assigned the maximum elastic modulus estimated from the cadaveric CT image data (3829 MPa). was defined as: S= % change in slope . The denominator is the percent change in the model input parameter between the sensitivity model and the baseline subject-specific model. constant shell thickness and elastic modulus (CST/CTEM) and subject-specific models (position dependent trabecular elastic modulus and cortical thickness). . S. For those sensitivity models that investigated constant inputs such as cortical thickness and trabecular bone elastic modulus. articular cartilage thickness (ACT) and articular cartilage elastic modulus (ACEM). % change in input parameter (3. The surface nodes were averaged to estimate the elastic modulus for the each tetrahedral element as was done in the subject-specific model.

456 MPa (Quartiles) Thickness = 1.1.2.0. Statistical tests (α=0. representing each strain gauge.5.29 E = ± 1 SD (1. A rectangular perimeter. and r2 values for each sensitivity model with the baseline subjectspecific model [43]. 1 SD (0.0 mm (Min/Max) E = 1. Nodes = Max Trabecular Modulus Reference EXP EXP EXP [57] [55] [58] EXP [59] NA Data Analysis FE predictions of cortical principal strains were averaged over elements that were located beneath each strain gauge. 0.62 GPa) Thickness = 0. Type CST CTEM CST/CTEM SSCV SSTV SSCM ACT ACEM OVERLAP Models Analyzed Thickness = ± 0.49 mm) E = 45. was created on the surface of the FE mesh using digitized points from the experiment.39 ν=0. 164. Models studied for FE sensitivity analysis. Deviations in material properties and cortical thickness were taken from experimentally measured/estimated values (EXP) as well as data reported in the literature. ν=0. Statistical tests were used to test differences between the slope of the best-fit line. 7. 4.41 mm.36. FE predicted strains were plotted against experimental strains.79 MPa (Min/Max) Surface Tet. Strains for a shell were included in the average if at least 50% of its area was within the perimeter. E = 164 MPa ν=0. .75 Table 3.05) were performed to compare the slope and y-intercept of the subject-specific best-fit line with the line y=x (Experimental Strains=FE Strains) to test the null hypothesis: there was no significant difference between FE predicted strains and experimental strains [43]. Best-fit lines and r2 values were reported for each model at all loads.

. B) excellent agreement was observed between experimental measurements and the FE mesh dimensions. Experimental Measurements (mm) 180 160 140 120 100 80 60 40 20 0 0 20 40 60 80 100 120 140 160 180 y = 0.998 FE Measurements (mm) Figure 3. Measurements were based on identifiable anatomical features of the iliac wing. yielding a total error of less than 3%.6). A) schematic showing the length measurements that were obtained from the cadaveric pelvis with an electromagnetic digitizer. ischium.16.97x + 2. pubis. and acetabulum. obturator foramen. Correlation between measurements on the cadaveric pelvis with the corresponding FE mesh was strong (r2=0.76 RESULTS Reconstruction of Pelvic Geometry The geometry reconstruction techniques yielded a faithful reproduction of the measured geometric features of the pelvis (Figure 3.6.998). There was no statistical difference between the slope and y-intercept of the regression line and the line y = x. r2= 0.

Cortical bone thickness ranged from 0.2). Cortical thickness was highest along the iliac crest. The above errors are based on polygonal surfaces with a resolution similar to the pelvis FE mesh.00 mm (mean 1. errors were ±0.635 mm. the ascending pubis ramus.762 and 2.49 mm (SD)) (Figure 3. the ischial tuberosity.44 . The reported standard deviation in nodal thickness for these tubes was also less than 10% of the average nodal thickness (Table 3. the iliac fossa and the area surrounding the pubic tubercle.004%. For all surfaces. Cortical bone was thin at the acetabular cup. errors decreased with increasing surface resolution.7).127 and 0. errors in thickness increased progressively for tubes with wall thickness between 0. Therefore individual nodal thickness values did not deviate much from the average nodal thickness. However. For parallel planes and concentric spheres.9210 mm (Table 3.77 Cortical Bone Thickness The thickness algorithm accurately predicted thickness using simple polygonal surfaces with known distances between the surfaces.41 ± 0. at the gluteal surface and around the acetabular rim.4. the RMS error was ±2%. .2). For the layered boxes. The thickness algorithm estimated aluminum tube wall thickness accurately (less than ±10% error) for tubes with thicknesses between 0.

Measurement of aluminum tube wall thickness from CT data.709 ± 0.111 163 0.270 2.2 1.554 ± 0.815 ± 0.079 60 0.669 ± 0.317 ± 0.063 8.762 1.3 1.825 ± 0.089 67 0.381 0.071 12 0.127 0.5 2.635 0.78 Table 3.254 0.762 mm.921 Estimated Thickness (mm) (mean ± SD) Error (%) 0.7 1.2.039 ± 0.032 2.108 -4.016 1.781 ± 0.077 3.094 336 0. Errors in wall thickness were less than 10% for thicknesses greater than or equal to 0.078 -2. Errors increased progressively as the wall thickness decreased.088 2.982 ± 0.508 0.8 . True Thickness (mm) 0.638 ± 0.

75 mm 0. B) medial view. Cortical thickness beneath the surface of the strain gauges was similar to the average model thickness of 1. . A) anterior view.79 A) B) 2.44 mm Anterior Oblique Medial Figure 3. at the gluteal surface and around the acetabular rim.7. Cortical thickness was highest along the iliac crest. Areas of thin cortical bone were located at the acetabular cup.41 mm but deviated less. the iliac fossa and the area surrounding the pubic tubercle. the ischial tuberosity. the ascending pubis ramus. Contours of position dependent cortical bone thickness with rectangles indicating the locations of the 10 strain gauges used during experimental loading.

5 . Coefficients of determination and y-intercept values were not statistically different than the subject-specific model for all sensitivity models analyzed (Table 3. models representing the median (164 MPa) and lower bound (45 MPa) of . was significantly stiffer (lower strains) than the subject-specific model (Figure 3.3). representing the upper bound (456 MPa) of the inter-quartile range. pubis joint. superior acetabular rim.824) with experimental measurements (Figure 3. FE Model Predictions FE predicted von Mises stresses for the subject-specific model ranged from 0-44 MPa and were greatest near the pubis-symphasis joint.80 Trabecular Bone Elastic Modulus Trabecular elastic modulus ranged from 2. and along the ischial tuberosity and the interior of the ilium. and on the ilium just superior to the acetabulum for each load applied (Figure 3.8).3). Data were significantly skewed to the right (positively skewed) so the median and bounds of the inter-quartile range were used for sensitivity models rather than the arithmetic mean and standard deviation.9 B) (Table 3.9 A) and had a best-fit line that was not statistically different than y=x (Experimental Strains=FE Strains). Although not statistically significant. median = 164 MPa. Areas of low modulus were located near the sacroiliac joint.0 MPa (mean = 338 MPa. The sensitivity model with constant trabecular elastic modulus. (Table 3.456 MPa).3829. inter-quartile range = 45 .3). Areas of high modulus were predominately near muscle insertion sites and within the subchondral bone surrounding the acetabulum. Baseline FE predictions of principal strains showed strong correlation (r2=0.

and trabecular bone . FE predictions were significantly stiffer than the subject-specific model predictions when both average thickness and trabecular elastic modulus were used (Table 3.9 C).5. Sensitivity values for the remaining models were also comparable to those of the constant trabecular bone modulus. values of the sensitivity parameter for the cortical bone modulus models were actually greater than those for changes to cortical thickness. Using a ratio of average sensitivities. The model with average cortical thickness predicted strains that were statistically similar to subject-specific model results (Table 3.81 trabecular elastic modulus were also stiffer than the subject-specific model (Figure 3. which suggests that FE predicted strains were not very sensitive to changes in cartilage modulus. 1 SD (Table 3.9 B). However.3). On average.3). Changes to the cortical bone elastic modulus were significantly different than the subject-specific model for E=15. (Table 3. for both ± 0.62 MPa. The remaining sensitivity models had best-fit lines that were not statistically different than the subject-specific model (Table 3.3). Changes in the thickness of the cortical bone had a profound effect on cortical strains (Figure 3.38 MPa but were not for E=18.3). cartilage thickness. FE predicted strains were 15 times more sensitive to alterations to the cortical bone elastic modulus than they were to changes in the trabecular bone elastic modulus. cortical surface strains were approximately 10 times more sensitive to changes in cortical thickness than to alterations to trabecular bone elastic modulus (Table 3.3).3). This suggests that the pelvic FE model was very sensitive to changes in cortical bone modulus despite the fact that statistical significance was not obtained for both models. cortical bone Poisson’s ratio.

The best-fit line for the overlap sensitivity model was nearly identical to the subject-specific model. A) anterior view.3).8. B) medial view. and on the ilium just superior to the acetabulum. . which suggests that FE predicted surface strains were not sensitive to overlap between the cortical shell and trabecular tetrahedral element. A) B) 5 MPa 0 MPa Anterior Oblique Medial Figure 3. Areas of greatest stress were near the pubis-symphasis joint. superior acetabular rim. Distribution of Von-Mises stress at 1 X body weight.82 Poisson’s ratio (Table 3.

.83 A) B) C) Figure 3. Changes to the trabecular modulus did not have as significant of an effect on the resulting cortical bone strains as did changes to cortical bone thickness. experimental cortical bone principal strains. B) constant trabecular modulus.9. For the subjectspecific model there was strong correlation between FE predicted strains with those that were measured experimentally with a best-fit line that did not differ significantly from the line y=x (Experimental strains=FE predicted strains). A) subjectspecific. C) constant cortical thickness. FE predicted vs.

193x + 2.272x + 8.41 1.334 0.39 ν =0.754 0. Higher values of sensitivity indicate a greater sensitivity to alterations in the model input/parameter.004 NA .05.824 0.187 0.840x + 6.670** y = 1.460 y = 0.622 y = 0.094 y = 1.820* y = 1.62 0. Lines with slopes significantly different than the subject-subject model are indicated (* p < 0.024x + 5.770 0.395x – 2.823 y = 1.294 y = 1. Cortical Thick.832 Model Type Subject-Specific Const.41.0 1.01).728 0.590 y =1. Results for all FE models including best-fit lines.788 0.821 0.36 7. ** p< 0.841 0.056 0.824 0.019x + 4.824 0. 164 ν = 0. r2 0.66 1.371 y = 1.2 ν = 0.265* y = 0.767 0. (mm) Cartilage E (MPa) Overlap (MPa) Value NA 1.204x + 2.951 0.79 Esuface nodes = 3829 Best-Fit Line y = 1.015x + 4.827 0.059x + 5. Y intercepts for all lines shown were not significantly different from zero.764 0.370** y = 1. Trabecular E (MPa) Const.84 Table 3. & E (mm.732 0.743 0.059** y = 0.559* y = 0.17 1.780 0.952x + 12.507 y = 0.833 0.3.789 0.001 0.890x – 2. MPa) Cortical ν Trabecular ν Cortical E (GPa) Cartilage Thick.013x + 4.92 164 45 456 1. Thick.709 y = 1.82 0.810 0.100 0.107x + 4.720x – 2.064 NA 0.142x + 7.0 4.931 0.119 Sensitivity NA NA 0.715 y = 1.062 0. (mm) Const.956x + 9.914 0.841 0.777 0.38 18.248** y = 1.29 15.90 0.005 1.294 y = 1.898x + 3. All r2 values were not significantly different than the subject-specific model. r2 values and sensitivity parameters.072x + 5.015x + 4.711 y = 1.054x – 2. Best-fit lines were generated in reference to experimentally measured values of strain.911 NA 0.

FE predictions of cortical strains were not statistically different than predictions from the subject-specific model when an average cortical thickness was used. there was an overlap between the shell and tetrahedral solid elements. This choice was based on compatibility with the tetrahedral elements used for the trabecular bone and considerations of element accuracy. Since the sensitivity parameter showed that cortical bone strains were very sensitive to changes in cortical thickness (Table 3. When constant cortical bone thickness and trabecular bone elastic modulus were used. In theory. this overlap could .85 DISCUSSION The most accurate FE model predictions were obtained when position-dependent cortical thickness and elastic modulus were used.3). Tetrahedral elements were used for the trabecular bone because they allow automatic mesh generation based on Delaunay tesselation.27 mm (SD)). so our second hypothesis was accepted.38 ± 0. but showed less deviation (1. Cortical shell thicknesses at the locations of the strain gauges were very close to the average thickness for the pelvis. the model was significantly stiffer than the subject-specific model. Cortical bone was represented using 3-node shell elements.7). Thick shells (wedges) and pentahedral (prismatic) solid elements were considered for the cortex but were later rejected since they produced inaccurate predictions of tip deflection when modeling cantilever beam bending. this suggests that the similarity in results was most likely attributable to comparable thickness estimates (Figure 3. with a shell reference surface positioned to align with the top of the cortical surface. However. Since the geometry of the model was based on the outer cortical surface.

high stresses were confined to a very small area that represented the location of contact between the head of the prosthetic femur and acetabulum and were still well below published values for ultimate stress [44]. However. If this were the case then the sensitivity model that assigned the maximum trabecular elastic modulus to all surface tetrahedral nodes would have been stiffer than the subject-specific model. the Von Mises stresses for cortical bone in the region of contact changed by 29% and 38% when cartilage thickness was reduced to 0 mm or increased to 4 mm. the results showed that this was not the case. However. material properties and applied loading make it impossible to compare FE predictions of stresses and strains in this study with previous investigations. repetitive stresses [44]. a layer of thin shells could be placed at the interface between cortical and trabecular bone with thickness defined towards the outer cortex of the pelvis. the slopes of the regression lines for these sensitivity models . However. The peak values of Von-Mises stress in this study appear to be unrealistic since bone would degenerate under such high. FE studies that aim to investigate the mechanics at the interface between cortical bone and trabecular bone should consider modeling the cortex without overlap. It is likely that a more physiological loading condition would generate better femoral head coverage and thus reduce the peak stresses at the contact interface. However.86 produce inaccurate estimations of cortical surface strain. Differences in boundary conditions. respectively. On average. this approach would not represent the surface topology of the pelvis as accurately as meshing the outer surface with tetrahedral elements. To remove the overlap.

the average stresses for areas of strain gauge attachment. Subject-specific FE strains were compared to models that assumed constant cortical thickness and elastic modulus.45. FE model accuracy was more dependent on cortical bone thickness than trabecular elastic modulus. average material properties and/or did not validate FE predictions of stress and strain. were very similar to those reported by Dalstra et al [26]. but subject-specific experimental measurements were not performed. although statistical tests were not performed to support this conclusion. Although cortical surface strains were not sensitive to cartilage thickness. Early models of the pelvis were either simplified 2-D [46-48] or axisymmetric models [49. Different cadaveric specimens were used for FE mesh generation and experimental tests.3). Most three-dimensional FE models [26. was the first and only attempt to develop and validate a three-dimensional FE model of the pelvis using subject-specific geometry and material properties [26]. In fact. The work of Dalstra et al. away from the applied load.51-56] used simplified pelvic geometry. The FE model was validated using experimental measures of strain in the peri-acetabular region of a cadaveric pelvis.50]. Nevertheless. The value for peak Von-Mises stress was also consistent with Schuller and co-workers who conducted a FE investigation to model single-leg stance in which peak values of Von-Mises stresses were as high as 50 MPa [45]. the local stresses and strains could be highly dependent on cartilage material properties and thickness.87 were very similar to the subject-specific model (Table 3. it was reported that the acetabulum of the experimental test sample was 45 mm whereas that of the specimen used for FE geometry and material properties was 62 mm [26]. The .

it is absolutely crucial to use accurate pelvic morphology if the research objective is to study diseases in which geometry is abnormal such as pelvic dysplasia. results for position- . which was ten times the thickness used in this study [26]. For example. Previously developed FE models of the pelvis have been based on coarse geometric representations. but it is likely that FE strain predictions would change substantially if an idealized geometry was used rather than a faithful representation of the external geometry. one should refrain from concluding that a position-dependent trabecular modulus is not important since it was shown that the model that assumed average cortical thickness and trabecular modulus was not as accurate as the baseline model. the small slice thickness and robust surface reconstruction techniques yielded a very accurate representation of the original geometry (Figure 3. In addition.88 effect of using average estimates was not investigated. While it may be acceptable to model the pelvis with idealized geometry for some applications. For this study deviations to the trabecular elastic modulus only had a significant effect on cortical surface strains when the upper interquartile range of trabecular bone elastic modulus was assessed. However. The present approach allowed cortical bone thickness to be estimated without laborious hand digitization [26]. In this study. the effects of alterations in other bone and cartilage material properties were not investigated.6). Moreover. FE model predictions of cortical strain were relatively insensitive to most model inputs (except cortical thickness and modulus). The relative importance of model input parameters will depend heavily on the FE model predictions that are of interest. Dalstra et al. handdigitized 6 mm thick CT slices.

This change in model displacement did not result in significant deviations of strain for the cortex beneath the gauges but could have altered the surface strain at other locations. it is recommended that a position-dependent trabecular bone modulus be included to improve overall FE model accuracy. Although the results of the sensitivity studies suggest that changes in material properties (except for under/over-estimation of cortical bone elastic modulus) were not likely to produce significant changes in cortical bone surface strains. A limitation to this study was the fact that the contralateral hemipelvis was not incorporated in the FE model. Finally. Nodes along the pubis joint were constrained. However. For this reason. although the slopes were not significantly different over the entire interquartile range. FE predictions of overall model displacement were altered considerably when a constant trabecular modulus was used (data not shown). it is likely that strains would be more sensitive to changes in the boundary conditions and applied loading conditions. The use of a well-defined experimental loading configuration allowed accurate replication of the loading conditions in the FE model. Therefore. If this were the case. strains were found to be greatest at the . the strains near the pubis joint and along the ischium should have been much lower than other areas around the acetabular rim. it was not the intent of this proof of concept study to replicate physiological loading conditions. Future studies will investigate pelvic mechanics under physiological loading conditions using additional experimental data. but some deflection may have occurred at the pubis joint in the experiment.89 dependent thickness and constant trabecular modulus were stiffer than the subjectspecific model.

Results demonstrated that cortical thickness could be measured down to approximately 0. Deviations in other assumed and estimated input parameters had little effect on the predicted cortical strains. In conclusion. Measurement accuracy depends largely on the axial and longitudinal resolution of the acquisition matrix and CT scanner collimation. Errors increased progressively for cortices that were less than the minimum collimation [37]. it was probably minimal since deflection to this joint would act as an immediate strain relief to the pubis and ischium. Our approach has the potential for application to individual patients based on volumetric CT scans.7 mm thick with less than 10% error. In the present study.7 mm for their scanner. If compression did occur at the pubis joint. The accuracy also depends on the energy settings. which was approximately 0.90 pubis joint and ischium during the experimental study. This will provide a means to examine . which was then confirmed by the FE results. our approach for subject-specific FE modeling of the pelvis has the ability to predict cortical bone strains accurately during acetabular loading. Cortical bone strains were most sensitive to changes in cortical thickness and cortical bone elastic modulus. and reconstruction algorithm. pitch. determined that cortical bone thickness could be estimated within 10% for cortices that were equal to or greater than the minimum collimation of the CT scanner. Palpation of the pubic cartilage demonstrated that the joint appeared to be an extension of the trabecular bone. Prevrhal et al. which suggests that the joint was relatively stiff. a cortical bone phantom was used to assess the measurement limits of the CT scanner and segmentation procedure simultaneously. CT is notorious for overestimating the thickness of cortical bone.

such as in the case of pelvic dysplasia. .91 the biomechanics of the pelvis for cases when subject-specific geometry is important.

. 1978. pp. Rohlmann. D.. and Perry... F. Springer-Verlag.. "Cortical Bone-Loss with Age. Stulberg. T.. Fischer. H.. Strate.. pp. MO. and Harris. Cooper. W. pp... 1997. pp. Adams. 1993. R. A. 969-90. A. Strauss.. 34. and Champion. 810-24. D. F. A. pp. P. H. Graichen. 36.. Ochsner.. D. Berlin. "Associated Aortic RupturePelvic Fracture: An Alert for Orthopedic and General Surgeons. Deuretzbacher. and Coggon. 859-71. pp. 2001. G. Murphy. [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] . 1976. W.." Surgery. G. 84.." Br J Radiol.. R. 176-83. Cole. A. DiPasquale. J. C. G.... "The Aetiology of Primary Osteoarthritis of the Hip. "Hip Joint Loading During Walking and Running. S. Bombelli.." Lancet.92 REFERENCES [1] [2] Bergmann. A. R. Davies. 19. 1965. C.." J Biomech. "Comparison of Predicted and Measured Contact Pressures in Normal and Dysplastic Hips. S123-8. "Hip Contact Forces and Gait Patterns from Routine Activities... P. and Duda. L.. 33. 1983." Clin Orthop. Measured in Two Patients. pp. 1992. Hoffman. 2. G. St Louis. F. Parfitt. M. W.." J Biomech.. Rothenberger." J Trauma. A." presented at Proceedings of the second open scientific meeting of the Hip Society. A. Solomon. Harris." Calcif Tissue Int. R. G. F. Webster. Velasco. 26. Jr. G.. Murray.. 58. 2033. D.... S. and Sweetnam. M. "Etiology of Osteoarthritis of the Hip. Rozycki. pp. 38. 429-34. Wickham. 1971." Ann Rheum Dis.. J. P. Croft. 1201-2. G. 1991.. Heller." J Bone Joint Surg Br.. Bergmann. Germany... 180-6. B. 1984. and Rohlmann. J.. H. J. D. P. P. "Osteoarthritis of the Hip and Acetabular Dysplasia. N. M. pp. pp. "Acetabular Dysplasia and Development of Osteoarthritis of the Hip. "Patterns of Osteoarthritis of the Hip. R.." Med Eng Phys. S. Jr. "Age-Related Structural Changes in Trabecular and Cortical Bone: Cellular Mechanisms and Biomechanical Consequences.. D. 1974.. 308-10. Michaeli. and Hipp. 356-61. O. R. Jr. 1986. Osteoarthritis of the Hip. Graichen. 50. G. "The Mortality Associated with Pelvic Fractures. pp..

"Contact Pressures in the Loaded Human Cadaver Hip... A. Moore. B. pp. D. 2879-83. K. Brand. W. 331-5. Hodge. B. T. and Mann. Duryea.. and Iglic." J Biomech. T. 69. A. 1999. "Radiographic Evaluation of Osteoarthritis. 431. K.. SanchezPernaute.. B. Iglic. R... Tomaino.. F. 9. 32. "High-Resolution Mri Detects Cartilage Swelling at the Early Stages of Experimental Osteoarthritis. Iglic. "Comparative Study of Iliac Crest and Subchondral Femoral Bone in Osteoarthritic Patients. Iglic. 2004. Mavcic.. M..." J Bone Joint Surg Br. Fazzalari. Swanson." Pflugers Arch. A. 1983.." Proc Natl Acad Sci U S A. R. V. J. M.. 11-41." J Biomech. R. W.. T.. Iglic. J." Osteoarthritis Cartilage. Manthey.. H. Ipavec. Fischer.. A. 463-72. Pedersen. "Peak Contact Stress in Human Hip During Gait. J.. K. 1992. V." Journal of Orthopaedic Research. Brand. pp. H. pp. "Stress Distribution on the Hip Joint Articular Surface During Gait.. B. and Vernon-Roberts.. and Pedersen. "Contact Pressures in the Human Hip Joint." J Bone Joint Surg Br. Mavcic. W. 20. S. and Kralj-Iglic. R177-8. "Mathematical Modeling of Stress in the Hip During Gait.. E. pp. and Shaw. 302-13. R. V. Egido. D.. J. 34.. A. 1229-35. L.. D. A.. S... 2001. D. Antolic. and Hutton.. 13. A. Pompe. 2001. Daniel. B." Pflugers Arch.. Pfaeffle. O. G." J Biomech. B. Byers.. Palacios. pp. Calvo. pp 1025-30.. 440. Kralj-Iglic.. P.. 83. L.. 16.. M.. J. G. I. pp. Antolic. M.. N. Ruiz-Cabello. 373-84. R. B.. "Contact Pressures in the Human Hip Joint Measured in Vivo.. M. pp.. 57... V. and Woo. Kralj-Iglic.. W. Hernandez. 42.. Delgado. N. A. pp. V. "Mathematical Estimation of Stress Distribution in Normal and Dysplastic Hips. P. A.. 377-83. and Herrero-Beaumont. M. Carlson. J. 2000. 1975. Fijan. "A Method for Measuring Joint Kinematics Designed for Accurate Registration of Kinematic Data to Models Constructed from Ct Data.. 1987. pp. "In Vitro Contact Stress Distributions in the Natural Human Hip.. [18] [19] [20] [21] [22] [23] [24] [25] . 2002. Manson. Day. L.. 1986. Ipavec. R. 536-41. W.. R. Burgess. N. R. R275-6." Radiol Clin North Am. H. and Weissman. E.. Gupta..93 [14] [15] [16] [17] Afoke. K." Bone. Harris. v.. pp. Mavcic. S. Y. and Freeman. B. T. and Srakar. pp. 1996.. C.. V. D.. Brown.

G. G. 44. Huiskes. M. Taubin.. S.. and Kalender. Avila. Three Dimensional Shells. Prevrhal. Boissonnat. 1992. 1-29. pp... K. 593-610. 1995. "Optimal Surface Smoothing as Filter Design.. W. 2002. "Nonlinear Finite Element Analysis of Shells: Part I." Computer Graphics (Proceedings of SIGGRAPH). 419-451.. Kitware Inc. Engelke. Zarge. L. "Analysis of Thick and Thin Shell Structures. 3 ed. The Visualization Toolkit: An Object Orientated Approach to Computer Graphics.94 [26] Dalstra. W. pp..-D. Pacific Grove. pp.. 117. J. and van Erning.. "Nonlinear Finite Element Analysis of Shells: Part Ii. "Solid Elements with Rotational Degrees of Freedom: Part Ii Tetrahedron Elements. W. 1981. K. pp. A. and Liu." Stanford University IBM RC-20404. 167-181. and Lorensen. Thomson. J. Zhang. Pawlak. Hughes. J.. and Hoffman. J. R. 27. 1970. 751-64. T. T. 1981. Hughes.. "Development and Validation of a Three-Dimensional Finite Element Model of the Pelvic Bone... J." International Journal for Numerical Methods in Engineering. and Golub." Computational Methods in Applied Mechanics. pp. Hughes. pp. [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] .. and Kiusalaas." International Journal for Numerical Methods in Engineering. CA.. W. 14. E. K." Comp Vis Graph Img Proc. J.. "Generalization of Selective Integration Procedures to Anisotropic and Nonlinear Media... T. "Decimation of Triangle Meshes. 1988." Computational Methods in Applied Mechanics. J. pp. W. vol.. pp. 1999. Two Dimensional Shells. L.. 272-8. M. Schroeder.. "Accuracy Limits for the Determination of Cortical Width and Density: The Influence of Object Size and Ct Imaging Parameters." Phys Med Biol. 2. S. 9. Mechanics of Materials. T. 331-362." J Biomech Eng. Pytel. S.. 1980. W. and Yunus. 25. Ahmad. 1996." International Journal for Numerical Methods in Engineering. and Liu. 31. 2003. "Shape Reconstruction from Planar Cross-Sections. 27. Schroeder. W. T.. 44.. P. S. 1991. A. J. 15.

pp. pp.. Schuller. L.. 111-9..." J Biomech. D. 1986.. Okumura. Z. pp.. 165-70.. and Cooke. W. R. A. L." J Biomech Eng." J Orthop Res. 64. 1993. Before and after Total Joint Replacement. Huiskes. M. 1001-6. R.. D.. 305-15.." J. D." J Biomech. Crowninshield.. pp. R. "An Axisymmetric Model of Acetabular Components in Total Hip Arthroplasty. 468-74. IL. pp. G." J Biomech. 2003... R. H. 15. Cohen. S. J. Pedersen. and Marti. Phys. Brand. R. 1982.. "A Theory of Large Elastic Deformation. T." J Biomech. A Finite Element Study. Yamamuro.. E. Carter. Huiskes. Wevers. Dalstra. Odgaard. "Stress Distributions in the Acetabular Region--Ii. P. R. Suzuki. 15. Appl. K. Lawrence Erlhaum Associates. F. M. A.. 75. Dutta.. H... Vasu. pp. Publishers. W. H. 3rd ed. F. Cohen. M. 15. "Total Hip Reconstruction in Acetabular Dysplasia. and Harris. "Stress Distributions in the Acetabular Region--I. D. W... D.. L. G. "Influence of Phantom Diameter. and Schurman. R.95 [38] Nickoloff.. R. 3. R. Effects of Cement Thickness and Metal Backing of the Total Hip Acetabular Component. and Johnston.. Siu.. 30. 11. "Mechanical and Textural Properties of Pelvic Trabecular Bone. Mooney. H.. Dalstra. 155-64. "Quantitative Computed Tomography: Comparative Study Using Different Scanners with Two Calibration Phantoms. T." J Bone Joint Surg Br. and Aiken. I. Springfield. Little. R. D. 2003. and Lu. H. 1991.. 1993. J. R.. pp.. Applied Multiple Regression Analysis for the Behavioral Sciences." Br J Radiol.. D. A.. D. 26." Med Phys.. 1982. D. pp. pp. Rapperport. Evans. Carter. Vasu. K. and Harris. 395-402. B. [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] . 1985.. West.. 1982. Mahwah.. S.. 108. 1940. Kvp and Scan Mode Upon Computed Tomography Dose Index. pp... Mechanical Properties of Bone. and Yamamoto.. J. S. "A ThreeDimensional Finite Element Analysis of the Upper Tibia. R. 523-35. M. and van Erning. R. 1973. NJ. Thomas. "Contact Finite Element Stress Analysis of the Hip Joint. C. 582-92. Carter. 435-46.

58. L. V. B. R. "Load Transfer across the Pelvic Bone. I. 31. Lappi. F. 422-31.. 1995.. S. "Analysis of the Structural Behavior of the Pelvis During Lateral Impact Using the Finite Element Method. 1983.. pp.. 620-5. 109-19. and McAndrew.. "Determination of Elastic Constants for Human Femurs.. M. T.. 715-24. E. S. Hille. P.. M. Armstrong. pp." J Orthop Res. D. pp. M. 1979. Khmelniker. Seral. L. "Mechanical Analysis of the Human Pelvis and Its Application to the Artificial Hip Joint--by Means of the Three Dimensional Finite Element Method." J Biomech.96 [50] [51] [52] Huiskes. 516-22. 1987. Schneider. Dalstra. M. 2001. "The Effect of Interfacial Parameters on Cup-Bone Relative Micromotions. D. "Three-Dimensional Finite Element Analysis of Several Internal and External Pelvis Fixations. pp." J Biomech Eng. 61. "Finite Element Analysis of Acetabular Reconstruction. M. 427-44. Seral. J. I. Palanca. A. Oonishi.. Konosu. 9.." J Bone Joint Surg Am." presented at SAE International. A Finite Element Investigation.. 101. and Morlock.. 113-20. J." J Biomech Eng. Spears.. 34. H. G. A. 28. and Hasegawa. "Development of a Biofidelic Human Pelvic Fe-Model with Several Modifications onto a Commercial Use Model for Lateral Loading Conditions. M. pp. Noncemented Threaded Cups... "In Vitro Measurement of Articular Cartilage Deformations in the Intact Human Hip Joint under Load. and Gardner." J Biomech. S. 16. Isha. C. 1999. [53] [54] [55] [56] [57] [58] [59] . "Estimation of Mechanical Properties of Cortical Bone by Computed Tomography. pp. and Lemay. M.. E. and Gracia. M. H. 1991.. Doblare. and Schneider. B. Snyder. M. 744-55. 193-97." J Biomech." Acta Orthop Scand. 122. pp. pp.. Dawson. and Huiskes.. King. Garcia. V." Accid Anal Prev. G.. 2003. pp. R.. 2000. M.. Pfleiderer. 1979.. Bahrani. E.... R..

5 mm. .J.CHAPTER 4 FACTORS INFLUENCING CARTILAGE THICKNESS MEASUREMENTS WITH MULTI-DETECTOR CT: A PHANTOM STUDY1 ABSTRACT The purpose of this study was to prospectively assess in a phantom the accuracy and detection limits of cartilage thickness measurements from MDCT arthrography as a function of contrast agent concentration. Images were reconstructed at intervals of both 1. A. imaging plane.E. B. Each chamber had a nylon cylinder encased by sleeves of aluminum and polycarbonate to simulate trabecular bone. Anderson. Ellis..L. joint space and tube current. The phantom was scanned with and without contrast agent on three separate days. Simulated cartilage thickness was determined from image segmentation. C.. J. Variations in simulated cartilage thickness and joint space were assessed. Weiss. Root mean squared and mean 1 Reprint of article In Press: “Factors Influencing Cartilage Thickness Measurements with Multi-Detector CT: A Phantom Study”.0 and 0. Radiology. Peters. A phantom with nine chambers was manufactured. Accepted February 16th 2007. spatial resolution. with chamber axes both perpendicular and parallel to the scanner axis..A. cortical bone. and cartilage. using known measurements as the reference standard. Contrast agent concentration and tube current were varied.

5 mm caused slight increases in error. scanning plane.98 residual errors were used to characterize the measurements. Measurements from anisotropic image data were less accurate than those for isotropic data. This study establishes lower bounds for accuracy and repeatability of cartilage thickness measurement using MDCT arthrography. joint spacing. Decreasing the simulated joint space to 0. segmentation reproducibility were determined.0 mm when no contrast agent or a low concentration of contrast agent (25%) was used. below 0. Errors grew as concentration of contrast agent increased.5 mm errors grew exponentially. CT scanner and image Simulated cartilage was accurately reconstructed (<10% error) for thicknesses >1. . Altering tube current did not affect accuracy. and provides data pertinent to choosing contrast agent concentration. and spatial resolution to optimize accuracy.

thinner beam collimation. multi-planar scanning and higher temporal and spatial resolution. which may be useful for preoperative surgical planning (e. 23].99 INTRODUCTION Three-dimensional reconstruction of volumes and surfaces from medical image data has become increasingly common. and providing guidelines for interpreting the results of biomechanical models which aim to investigate joint contact mechanics. less attention has been given to CT arthrography [6.g. The advent and availability of multi-detector CT has yielded substantial advances over single detector CT by providing shorter data acquisition times. Of particular interest is the ability to visualize the spatially varying thickness of articular cartilage in diarthrodial joints. tracking of cartilage degeneration over time. precise quantification of cartilage loss due to osteoarthritis (epidemiological studies). cartilage transfer procedures). While a substantial body of research has examined MRI cartilage reconstruction errors (e. both for the diagnosis and treatment of patients with joint pathologies and to define patient-specific geometry for computational models and computer-assisted surgery. Nevertheless. although fat-suppressed spoiled gradient-echo in the steady state (FS-SPGR) is still considered the best imaging protocol for imaging articular cartilage [7-16]. [1722]). estimates of cartilage thickness determined via MRI image data are often validated by . With a-priori knowledge of the reconstruction error the clinician or researcher can make informed interpretations regarding cartilage thickness. 18. Recent evidence suggests that multi-detector CT (MDCT) arthrography may be more sensitive than MRI for detecting cartilaginous lesions [1-5] and quantifying cartilage thickness [6].g.

23]. to quantify cartilage thickness using CT arthrography in patients who complain of pain that may be related to OA but do not have direct evidence of radiographic thinning or localized defects. an understanding of the spatial variation in thickness of articular cartilage in joints without visible damage could prove to be a valuable clinical tool. It would be useful. To our knowledge only one study compared quantitative measurements of cartilage thickness between reconstructed MDCT arthrography images to excised tissue samples [6]. the use of harvested cartilage plugs in this study limited the range of cartilage thickness that could be analyzed [6]. . CT imaging and subsequent three-dimensional reconstruction of articular cartilage has a variety of clinical and basic science applications.100 direct comparison with CT arthrography results [18. 19]. therefore. However. For example. Studies that have compared cartilage thickness measurements estimated from CT arthrography data to those obtained from physical measurements of anatomical sections have generally been qualitative assessments [18. which may erroneously imply that CT arthrography is the “reference standard” for such estimations. From the point of view of experimental investigations of cartilage contact mechanics using cadaveric tissues. recent evidence suggests that cartilage may actually swell in the early stages of OA [24]. Although CT arthrography is most often performed in an effort to diagnose articular cartilage damage rather than to quantify cartilage thickness. quantification of differences in reconstruction accuracy between standard CT and CT arthrography would clarify whether cadaveric joints should be completely dissected and imaged with air or if the joint capsule should be left intact to obtain the highest accuracy.

101 The bounds of cartilage thickness detection and hence the ultimate reconstruction error remains unknown for MDCT arthrography. . joint space and tube current. spatial resolution. imaging plane. In addition. using known measurements as the reference standard. Thus. the purpose of our study was To prospectively assess in a phantom the accuracy and detection limits of cartilage thickness measurements from MDCT arthrography as a function of contrast agent concentration. the influence of imaging parameters on the ability to detect and reconstruct articular cartilage from MDCT arthrography image data has not been assessed.

nylon threaded caps were used to seal the fluid in the chambers.1 B). Fullerton. the cylindrical sleeve of aluminum represented cortical bone. An outer polycarbonate four-prong spacer was press-fit into each of the chambers between the outer layer of simulated cartilage and adjacent nylon phantom body (Figure 4. Figure 4.00 mm) with constant simulated cartilage thickness of 2.1). and 1. 0.. The central nylon cylinder simulated trabecular bone.00 mm (Phantom Chambers 1-6.1 A) was held constant at 2.1 A).25. and the outer cylindrical sleeve simulated cartilage (Figure 4.50.75.50. A micrometer with accuracy of ±0.. The joint space in phantom chambers 1-6 (Figure 4. CA). The phantom body was constructed using nylon (Natural Cast Nylon.00 mm was used in the remaining three compartments (Phantom Chambers 7-9. The polycarbonate cylindrical sleeves were machined to wall thickness values of 0. Ogden.1 B). and 4. 1.00. Professional Plastics Inc. All aluminum cylinders were machined to a wall thickness of 1. 0.102 MATERIALS AND METHODS Phantom Description An imaging phantom was designed and manufactured to quantify the error in reconstructing cartilage thickness (CNA Precision Machine. Nine chambers were drilled into the phantom body (Figure 4. Figure 4. 0. UT) (Figure 4.00.25. A varying joint space (0. Fullerton.00 mm to represent cortical bone with constant thickness.0 mm.1).01 mm was .1 A). Each chamber was composed of a central nylon cylinder encased by cylindrical sleeves of aluminum and polycarbonate (Standard Polycarbonate. Finally. The spacer held the central cylinders securely in place and provided a “joint space” that could be filled with contrast agent. Professional Plastics Inc. 2. CA).

polycarbonate and aluminum were chosen because their x-ray attenuation values are similar to trabecular bone. cartilage and cortical bone. respectively [25-28]. respectively. The size of the phantom (250 x 250 mm) was representative of a typical field of view (FOV) for imaging human diarthrodial joints. The range of cartilage thickness (0.103 used by the manufacturer to determine the wall thickness tolerance of the aluminum and polycarbonate cylindrical sleeves.07 mm. representing cortical bone and cartilage. The tolerance was reported to be within ± 0. Nylon. . 33].00 mm) was chosen to represent the range reported in the literature for human articular cartilage [32. This range of cylinder diameters is similar to that reported in the literature for human femoral and humeral heads [29-31].25–4. The outer diameter of each compartment (outer boundary of simulated cartilage) was kept constant at 52 mm while the diameter of the aluminum sleeve and central nylon cylinder were adjusted between 38–46 mm to accommodate differences in cartilage thickness and joint spacing.

0. 0. . 2.5. Simulated cartilage thicknesses of 4. The longitudinal (L) imaging plane is also shown. (5) polycarbonate fourpronged spacer for creating the joint space. Number call-outs correspond to the same details provided above. respectively. A constant thickness of 2. A) schematic of phantom used to assess accuracy and detection limits of MDCT in the transverse plane. C) CT scan of the phantom with contrast agent and inset showing image details of chamber #1.1.25 mm with constant joint space of 2.0. and 0.0 mm with joint spaces of 1. 0. (4) joint space. 0.0. and 0. 1. and (6) bulk of the phantom body made using nylon.75. (2) 1 mm thick aluminum sleeve to represent cortical bone. respectively.5.25 mm were used in chambers 7-9. (3) polycarbonate sleeve to represent cartilage.104 L 1 A) 4 7 L 2 B) 4 1 2 5 8 3 6 9 3 6 5 C) 2 1 6 5 3 4 Figure 4.0 mm were used in chambers 1-6.0. B) exploded view of chamber #1 detailing: (1) nylon center cylinder to represent trabecular bone.

105

CT Imaging Protocol All phantom scans were performed with a Siemens SOMATOM® Sensation 64 CT Scanner (Siemens Medical Solutions USA, Malvern, PA). This scanner makes use of a periodic motion of the focal spot in the longitudinal direction to double the number of simultaneously acquired slices with the goal of attaining improved spatial resolution and elimination of spiral artifacts regardless of spiral pitch. Constant scanning parameters for this study were: 120 kVp, 512 x 512 matrix, 300 mm FOV, and 1 mm slice thickness. A total of 6 fluid scans and 4 non-enhanced scans were performed. The imaging protocol detailed below was performed on three separate days to assess the reproducibility of the CT scanner and segmentation procedure.

Contrast Enhanced Scans Contrast agent (Omnipaque 350 mgI/ML, GE Healthcare, Princeton, NJ) was mixed with 1% lidocaine HCL (Hospira Inc., Lake Forest, IL) in separate concentrations of 25, 50, and 75%. The phantom was scanned using a tube current of 200 mAs for each of the three concentrations (n = 3 scans) in the “transverse” or frontal plane (Figure 4.1 A). The laser guide was used to align the CT slice axis perpendicular to the phantom chambers longitudinal axes, thereby minimizing volumetric averaging between slices. Additional transverse scans were conducted with tube currents of 150 and 250 mAs using the phantom filled with 50% contrast agent (n = 2 scans). A scan with tube current of 200 mAs was performed on the phantom filled with 50% contrast agent parallel to the phantom chambers “longitudinal” axes (Figure 4.1 A) to intentionally introduce volumetric averaging.

106

Non-Enhanced Scans The phantom was scanned without fluid to estimate the error in cartilage thickness reconstruction for disarticulated, dissected cadaveric joints. Non-enhanced scans were performed in the transverse plane using tube currents of 150, 200, and 250 mAs (n = 3 scans). A final non-enhanced scan was performed with a tube current of 200 mAs parallel to the phantom chambers longitudinal axes to intentionally introduce volumetric averaging between successive slices.

Image Segmentation, Surface Reconstruction, and Measurement of Thickness Phantom image data were transferred to a Linux workstation for post-processing. Image data were re-sampled post-CT using 0.5 mm slice intervals for the contrast enhanced and non-enhanced longitudinal scans to assess changes in accuracy between an anisotropic spatial resolution (0.586 x 1.0 x 0.586 mm) and near isotropic resolution (0.586 x 0.5 x 0.586 mm). Thinner post-scan reconstructions in the transverse plane would have been ambiguous since the curvature of the phantom chambers did not change as slices were taken through this direction. Separate splines for the outer surface of the aluminum cylinder, representing cortical bone, and the boundary between the polycarbonate cylinder and air (nonenhanced scan) or contrast agent (enhanced scan), representing the outer layer of simulated cartilage, were extracted from the image data. Both automatic and semiautomatic thresholding techniques were employed using commercial segmentation software (Amira 4.1, Mercury Computer Systems, Chelmsford, MA).

107

Each dataset was automatically thresholded using a masking technique available in Amira 4.1, which allows the user to highlight pixels over a range of defined intensities. For datasets with contrast agent included, the mask was adjusted incrementally until all of the pixels representing nylon (the bulk of the phantom body) were excluded. Thus, pixels with intensities greater than this value were masked as contrast agent and simulated cortical bone whereas values less were defined as simulated cartilage. The same masking procedure was used for the non-enhanced scan datasets to define the simulated cortical bone boundary; however, the boundary between simulated cartilage and air was defined by reversing the mask such that all pixels representing the nylon body of the phantom were included. As mentioned above, the masking procedure was performed for each CT dataset separately to ensure that the appropriate threshold range was chosen independently of alterations in tube current, contrast agent concentration, spatial resolution or scanner direction. Following masking of all of the datasets it was later determined that inter-scan threshold values varied by less than 5%. Due to CT volumetric averaging it was necessary to utilize a semi-automatic thresholding technique for datasets where contrast agent was included. However, this procedure was only required for phantom chambers with simulated cartilage thickness of 0.5 and 0.25 mm (chambers 5 and 6, Figure 4.1 A); simulated cartilage thicker than this was effectively segmented by the automatic method, regardless of contrast agent concentration, tube current, spatial resolution or scanner direction. For the 0.5 and 0.25 mm chambers the baseline automatic threshold value was first used to define a general segmentation spline. Next, regions where pixels blended together were separated using a

108

paintbrush tool available in Amira 4.1 such that the resulting spline followed the general boundary between simulated cartilage and contrast agent. Although volumetric

averaging was present, the intensity gradient between contrast agent and simulated cartilage was strong enough to allow for easy visual separation. To ensure uniformity, all of the semi-automatic segmentations were performed by the senior author, A.E.A. Splines were stacked upon one another and triangulated using the Marching Cubes algorithm [34] to form surfaces that represented the outer surfaces of simulated cortical bone and cartilage. To preserve the native splines of the CT image data, the resulting polygonal surfaces were not altered via decimation or smoothing. A published algorithm was used to assign thickness to each of the nodes defining the simulated cartilage surface [35]. The algorithm has been tested for accuracy using concentric cylinders with known thickness. Reported errors were less than 2% [35].

Specifically. Systat Software Inc.. Chicago. (4. SPSS Inc.e. IL). joint spacing.5 for Windows 2002. quantification of means and standard deviations that cannot ascertain statistical significance) were calculated using statistical software (SPSS 11. The overall thickness accuracy for each phantom chamber was assessed using the root mean squared (RMS) error criteria: 2 ⎡ n ⎤ RMS = ⎢ ∑ ( ti CT − t Phantom ) n ⎥ . ⎣ i =1 ⎦ 12 (4. contrast agent concentration. CA) with standard deviation error bars to indicate the inter-scan variation in reconstruction accuracy.. and imaging plane.1) where the summation is over the number of surface nodes n and tPhantom is a constant thickness that was assessed by direct manufacturer measurement of the phantom. The mean residual error was calculated to determine the directionality of the error: Mean Residual = ∑ (t n i =1 CT i − t Phantom ) n . RMS and mean residual errors were averaged for the three days that CT scans were conducted.109 Error Analysis Thickness values were analyzed to determine the accuracy and detection limit of MDCT and to investigate the influence of tube current.2) Statistical Analysis Descriptive statistics (i.0. San Jose. The resulting means were plotted (SigmaPlot 8. .

3).2 A).75 mm (Figure 4.0 mm (Figure 4. An increase in contrast agent concentration resulted in a greater tendency for simulated cartilage to be underestimated for values between 1.2). The simulated cartilage of the phantom was accurately reconstructed (less than 10% RMS and mean residual error) for thickness greater than 1. Substantial differences in average reconstruction errors were also noted when the scanner direction and spatial resolution were altered (Figure 4. . a shift in error from under to overestimation occurred as the thickness approached the spatial resolution of the image data (0.0 and 4.0 mm thick (Figure 4. The anisotropic longitudinal reconstructions at 50% concentration produced RMS and mean residual errors greater than the corresponding transverse and near-isotropic longitudinal dataset at 50% concentration for simulated cartilage thicker than 1.2).586 mm) (Figure 4. However. altering the tube current resulted in negligible differences over the range of simulated cartilage thickness analyzed (data not shown).110 RESULTS Contrast Enhanced Scans There were notable differences in the average RMS and mean residual error due to alterations in contrast agent concentration (Figure 4.3).2 B).586 x 0. Transverse scan RMS errors grew progressively as the concentration was increased from 25% – 75% for values of thickness greater than 0. Finally.0 mm when the lowest concentration of contrast agent (25%) was used and the direction of the CT scan was transverse to the phantom (Figure 4.2 B).

0 3.5 2.0 3.2.75 mm). RMS errors grew progressively as the contrast agent concentration increased (for thickness > 0. Simulated cartilage RMS (A) and mean residual (B) reconstruction errors for the transverse contrast enhanced scan datasets as a function of contrast agent concentration.0 True Thickness (mm) 120 Mean Residual Error (% of true thickness) 100 80 50 40 30 20 10 0 -10 -20 0.0 2.0 2.5 1.0 1.5 1.5 2.5 25% 50% 75% B) 4.111 Root Mean Squared Error (% of true thickness) 140 120 100 60 50 40 30 20 10 0 0. The directionality of the error was dependent on the contrast agent concentration and simulated cartilage thickness.0 0.0 True Thickness (mm) Figure 4.5 25% 50% 75% A) 4.0 0.5 3. .0 1.5 3.

0 2.5 4.5 3.3.0 3.5 1.0 0.0 50% Longitudinal 50% Longitudinal Isotropic 50% B) True Thickness (mm) Figure 4.0 1.0 0.5 Transverse 50% Longitudinal 50% Longitdinal Isotropic 50% 3.5 1.0 2. .5 4.0 True Thickness (mm) Mean Residual Error (% of true thickness) 100 90 80 70 60 50 40 30 20 10 0 -10 -20 -30 0. Errors were greatest for the anisotropic longitudinal data reconstructions.112 Root Mean Squared Error (% of true thickness) 120 110 100 90 80 70 60 50 40 30 20 10 0 0.5 2.5 2.0 3. The longitudinal isotropic reconstructions yielded errors more consistent with the transverse scan results. Simulated cartilage RMS (A) and mean residual (B) reconstruction errors for the transverse contrast enhanced scan datasets at 50% concentration as a function of imaging plane direction and spatial resolution.0 1.

5 mm thick and error bars overlapped adjacent data points.2 – 4.113 There were differences in RMS errors over the range of joint spaces studied due to changes in contrast agent concentration and scanner direction (Figure 4.4 indicated a high level of reproducibility for simulated cartilage between 0. below 0.0 – 0. Errors increased as the concentration of contrast agent was increased (Figure 4.4).4). Examination of the standard deviation error bars in Figures 4.78 – 4. The standard deviation error bars also did not overlap adjacent results within this range.25 – 0.0 mm thick.4) but not due to alterations in tube current (data not shown).5 mm errors grew exponentially (Figure 4. however. .5 mm reconstruction) over the full range of joint spaces analyzed (Figure 4.4). Mean residual error analysis indicated that simulated cartilage thickness was underestimated for all datasets and that these errors were the smallest for the transverse 25% scan (data not shown). The anisotropic longitudinal dataset (1.0 mm reconstruction) produced greater RMS errors than the corresponding transverse and near-isotropic longitudinal scan datasets (0. Standard deviations were much larger for simulated cartilage 0.5 mm. RMS errors for each individual transverse scan increased slightly when the joint space was decreased from 2.

25 0.25 1. Errors increased as contrast agent concentration increased.00 True Joint Space(mm) Figure 4.50 0. and spatial resolution. contrast agent concentration.4. Simulated cartilage RMS errors as a function of joint space thickness.75 2.114 Root Mean Squared Error (% of true thickness) 35 30 25 20 15 10 5 0 0.00 1. Reconstructions from the isotropic longitudinal dataset were more accurate than the anisotropic dataset in the same imaging plane. .00 Transverse 25% Transverse 50% Transverse 75% Longitudinal 50% Longitudinal Isotropic 50% 0.50 1.75 1. imaging plane direction.

0 mm thick (Figure 4.25 mm was due to a shift from an underestimation to overestimation of cartilage thickness. standard deviation error bars of the nonenhanced scans were negligible for thicker simulated cartilage (0. however.115 Non-Enhanced Scans Reconstructions of the non-enhanced transverse scan at 200 mAs resulted in RMS errors less than 10% for thickness values greater than 1.0 and 4. however errors grew exponentially below 2.0 mm thick simulated cartilage.5 A).5 A). the lack of increase in RMS error at 0. The leveling point in the RMS plot aligned well with corresponding points of inflection on the mean residual error plot (Figure 4.78 – 4.5 – 0. Errors for the longitudinal anisotropic scan were substantially greater than the transverse and nearisotropic longitudinal scans for simulated cartilage less than 2.78 – 4. RMS errors for the longitudinal and near-isotropic longitudinal scan datasets were similar to the transverse scan for 2.0 mm (Figure 4.0 mm thick) but increased when thickness was decreased below this range.5 A).5 A). RMS errors for the non-enhanced scans were within 2% of those reported for the contrast enhanced transverse scans at 25% contrast agent concentration for simulated cartilage 0. Therefore. the RMS error leveled out between 0.0 mm thick.0 mm (Figure 4. RMS errors grew exponentially for simulated cartilage less than 1.5 B). .25 mm (Figure 4. Altering the tube current from 150 – 250 mAs did not have an appreciable effect on the RMS or mean residual errors in the transverse plane (data not shown).78 mm. As with the contrast enhanced scans. Standard deviation error bars also did not overlap at adjacent data points within this range but did for thicknesses less than 0.0 mm thick.

5 4.0 3.0 1.0 0. .5 4.5 3.5 Transverse Longitudinal Longitudinal Isotropic B) 3.0 2.5 1.5 2.0 1.0 Transverse Longitudinal Longitudinal Isotropic A) True Thickness (mm) 140 Mean Residual Error (% of true thickness) 120 100 50 40 30 20 10 0 -10 -20 0. RMS errors for the longitudinal isotropic datasets were consistently less than the anisotropic dataset for simulated cartilage less than 2. Simulated cartilage RMS (A) and mean residual (B) reconstruction errors for the non-enhanced scan datasets at 200 mAs as a function of imaging plane direction and spatial resolution.0 2.5.0 3.5 1.0 True Thickness (mm) Figure 4.5 2.0 0.116 Root Mean Squared Error (% of true thickness) 160 140 120 60 50 40 30 20 10 0 0.0 mm thick.

scan direction. The entire protocol was repeated on separate days and only minor inter-scan variation was noted for simulated cartilage between 0. The simulated cartilage of the phantom was accurately reconstructed (less than 10% RMS and mean residual error) for thicknesses greater than 1.0 mm for both contrast enhanced and non-enhanced scans. The results of our study also demonstrated that the accuracy of CT cartilage reconstructions were dependent on the concentration of contrast agent.78 – 4.07 mm. tube voltage. any differences noted in reconstruction error within this range were due to the intervention studied rather than from confounding factors such as thresholding procedure or CT scanner variability. spatial resolution and to a lesser extent. separate scans were done whenever a new intervention (e. .0 mm when either no contrast agent or a low concentration of contrast agent (25%) was used.117 DISCUSSION To our knowledge our study is the first quantify the detection limits and accuracy of MDCT using a phantom. In addition. Alterations in the scanner tube current did not affect the accuracy of simulated cartilage thickness reconstructions in the range tested for both contrast enhanced and non-enhanced scans. Care was taken to control confounding factors in our study. imaging plane direction. thus variations in the true thickness of the phantom would not have a substantial influence on the perceived values of phantom thickness as assessed by CT. Therefore.g. joint spacing. The physical thickness of the phantom was measured to a tolerance of ± 0. and contrast agent concentration) was performed in an effort to isolate these effects.

[6] compared ankle cartilage measurements obtained from MDCT double-contrast arthrography and three-dimensional FS-SPGR MRI to physical measurements of excised plugs from cadaveric ankles (ranging from 1 – 2 mm thick) and found that CT was more accurate than MRI.586 mm). This was due to the fact that thickness could not drop much below the width of a single pixel without extensive surface decimation and smoothing. 26]. El-Khoury et al. joint geometry. the results of our study demonstrate that the direction of the error is dependent on the concentration of the joint fluid and spatial resolution of the image data when CT arthrography is used. An explanation for this finding is as follows: as the concentration was increased larger pixel intensity gradients were established at the boundary between cartilage and contrast agent. Therefore. El-Khoury’s best-fit line of physical plug measurements plotted against MDCT estimates indicated that CT underestimated cartilage thickness by approximately 5% [6]. although CT has been shown to overestimate the thickness of thin structures [25.586 x 0. Nevertheless. and arthrography technique (double contrast) between this study and our work make exact comparisons impossible. Differences in segmentation methodology. This initiated more intense volumetric averaging at this boundary and resulted in a greater tendency for cartilage thickness to be underestimated. However.118 The results of the contrast enhanced scan reconstructions showed a direct relationship between contrast agent concentration and reconstruction error. a shift in error from under to overestimation occurred as the thickness approached the spatial resolution of the image data (0. which has the same .

Reconstruction errors from these chambers could have been influenced by user technique.25 and 0. and measurement technique. It is well known that articular cartilage exhibits depth and location dependent inhomogeneities in material structure [36-38]. Study Limitations The results of our study must be interpreted in light of the inherent differences between measurements obtained from a phantom to those taken from experimental studies that use real cartilage specimens. a total of three scans were performed and descriptive statistics utilized to assess reproducibility. In addition. Finally. However. the magnitude of the standard deviation error bars for thicknesses within this range were very similar to the non-enhanced scan deviation bars.5 mm it was necessary to use a semi-automatic method to segment simulated cartilage from the contrast enhanced scan datasets. The . tissue homogeneity.119 direction and similar magnitude of error as the 25% concentration agent results of our study over the same range of thickness. For chambers with simulated thicknesses of 0. although similar [25-28]. small differences will exist between x-ray attenuation values of real tissue to that of the materials used in our phantom. our approach allowed us to eliminate potential confounding factors such as geometry. In addition. and these factors were not part of our study design. which was a statistical methodology consistent with a phantom study related to MRI slice thickness [39]. diarthrodial joints such as the shoulder and hip have spherical geometry but the phantom chambers were cylindrical. and a purely automatic segmentation technique was used for the latter datasets. Nevertheless.

Thus. which would eliminate simulated cortical bone as a confounding factor to our study. it appears that there is little difference between contrast enhanced and nonenhanced scans for the thinnest phantom chambers. It must be emphasized that the phantom reconstruction errors are likely a best case result since confounding factors were controlled. The phantom was designed to simulate the interface between cartilage and cortical bone. Failure to do so will result in increased errors when the joint space reaches a critical threshold (occurring at 0.120 standard deviations for simulated cartilage less than 0. any errors introduced would be consistent over all datasets. Practical Applications The results of our study provide minimum bounds for the errors in cartilage thickness measurement using MDCT and provide guidelines for practical use. In addition. therefore. with bars overlapping adjacent data points. A lower concentration of contrast agent is likely to reduce the amount of volumetric averaging between cartilage and contrast agent since it was shown that higher concentrations caused the simulated cartilage to appear thinner than its true thickness. however the wall thickness of aluminum was held constant for each phantom cylinder and the thresholding protocol was not biased to changes between phantom chambers or whole datasets. joint spacing should be maximized prior to scanning.78 thick were substantial for both contrast enhanced and non-enhanced scans. which can be done by completely filling the joint capsule with the diluted contrast solution and/or applying traction to the joint.5 mm in our . There was likely some image thinning of polycarbonate due to volumetric averaging between the polycarbonate and adjacent aluminum cylindrical sleeve.

MDCT (unlike its predecessors) offers the ability to do this without increased radiation dosage to the patient [40]. the accuracy of MDCT with and without arthrography is dependent on several factors including the contrast agent concentration. In conclusion. Finally.121 phantom study). given the additional technical challenges of keeping joint fluid within the capsule of a dissected joint. imaging plane. one could expect similar cartilage reconstruction accuracies when cadaveric tissues are CT scanned with or without contrast agent since reconstruction errors were similar between non-enhanced and contrast enhanced scan (25% concentration) datasets. CT image reconstructions should be chosen such that isotropic or near-isotropic spatial resolutions are achieved. it seems more appropriate to scan the specimen without contrast agent. interpretation of biomechanical models. Fortunately. However. From a basic science point of the view the following conclusion can be made: assuming that sufficient joint space is maintained and a low contrast agent concentration is used. . and spatial resolution. and design of epidemiological studies aimed to investigate changes in cartilage thickness. An improved understanding of the detection limits and accuracy of MDCT cartilage reconstructions will assist in the diagnosis of joint pathologies. joint spacing.

and Yoshikawa. Westhoff. E. E.. C. V. "Effect of gradient and section orientation on quantitative analysis of knee joint cartilage. K. Schmid.. Jamart. Schulte. B. K. "Evaluation of patellar cartilage surface lesions: comparison of CT arthrography and fat-suppressed FLASH 3D MR imaging... F.. 25(1). Kiefer.. Hodler.. 32(5). K. Daenen. "Cartilage lesions in the ankle joint: comparison of MR arthrography and CT arthrography. G. Pfirrmann. 11(4). 233(3). Nishii. Tanaka. F. Link. and Malghem." J Magn Reson Imaging. "Assessment of normal patellar cartilage volume and thickness using MRI: an analysis of currently available pulse sequences.. B. F. T.. 981-985.. B. A. 170(3). M. 2000.. pp. D.. R.. S. J.. 15(4). T. H. Lengele. M. and Saltzman. S. T. K. and Woertler. 768-773. Bruegel. pp.. Englmeier. J. M. Nakanishi. B. J. Maldague. A. T. Sittek." Skeletal Radiol. Englmeier..122 REFERENCES [1] Vande Berg.. P. M. J. 469-470.. 2002... Eckstein. Waldt. W. pp. Sittek. Maag. Alliman. Stammberger... El-Khoury. 1996. pp. H... J. Eckstein. H. Gavazzeni. Marcelis. B. R.. Lecouvet. 55-62. Ferrara. 1998. Haubner.. N.. 2005.. 430-436.. "Fat-suppressed 3D spoiled gradient-echo MRI and MDCT arthrography of articular cartilage in patients with hip dysplasia. 593-597.. H." Eur Radiol. Ganter. pp. 379-385.. K. H. J.. 259-265. P. pp." Skeletal Radiol. 185(2)... R. Poilvache. pp. J... L. M. pp... pp." AJR Am J Roentgenol. M. Brown. H. M. 2005. Rummeny. Milz." Radiology.. 1998. and Zanetti." AJR Am J Roentgenol. K. "Comparison of multislice CT arthrography and MR arthrography for the detection of articular cartilage lesions of the elbow. and Dondelinger. Vienne. and Reiser. H. Lundberg. "Cartilage thickness in cadaveric ankles: measurement with double-contrast multi-detector row CT arthrography versus MR imaging.. H. F. Eckstein. [2] [3] [4] [5] [6] [7] [8] [9] . S. "Assessment of knee cartilage in cadavers with dual-detector spiral CT arthrography and MR imaging. and Reiser... K. and Reiser. Y. 784-791.. Sugano. M.. J. 2004.. E. "In vivo reproducibility of threedimensional cartilage volume and thickness measurements with MR imaging.. P... Rudert. M.... 8(6). S... Materne. C. Kuhn. 2003." Radiology. 222(2)... Faber. C. J.. M. Priebsch. R. Miki. F." Eur Radiol.

and Peterfy... A. Gold.. Eckstein.. and Hospodar. G. Cohen. J. Shimakawa.0 T with IDEAL balanced steadystate free precession--initial experience. S. Sartoris. M. and Brittain. McCarthy. S. Legrand. 127-132. R. and contact areas from MRI: in-vitro calibration and in-vivo measurements. G. Shimakawa. Gold. Vasanawala. and Putz. Kwak. C. G. Reeder.. P. 1997. C. pp.. D. F. C. Adam. S. 198(1). pp. M. Marcelis. 2006. 238(2). Wirth. 1996. Eckstein. 187(2). McCauley. P. 7(1). D. J. 30(3).. "Detection of knee hyaline cartilage defects using fat-suppressed three-dimensional spoiled gradient-echo MR imaging: comparison with standard MR imaging and correlation with arthroscopy. H. McCauley. F. W. R. 1995.. Schulte." Osteoarthritis Cartilage. Pathria. M. "Magnetic resonance imaging (MRI) of articular cartilage in knee osteoarthritis (OA): morphological assessment. Milz. Disler.. H.123 [10] Recht." Radiology. W.. H. 712-718. D. Disler.." Radiology. E. and Beaulieu. G. 2006.. 473-478. 546-551... J... Webb. T. [11] [12] [13] [14] [15] [16] [17] [18] .... Beaulieu. Hargreaves.. "Abnormalities of articular cartilage in the knee: analysis of available MR techniques. E.. R. Recht. J.. 377-382. and Belhobek. 240(2). J. D. R. Cicuttini.. C. 167(1). N.." Radiology. and Resnick. 209-212. P. C. G... C. "Articular cartilage of the knee: rapid three-dimensional MR imaging at 3... and Ateshian. thickness... D. C. pp. Becker. M.. G. J. J.. 1996. T. Reiser. Waterton... Raynauld.. A. D." Radiology. Johnson.... E.. D. "Knee cartilage topography. A4675. Kelman. F. Kramer. Wirth. "Non-invasive determination of cartilage thickness throughout joint surfaces using magnetic resonance imaging.. F. pp. B. M. 95-109. Ciaccio. L...." AJR Am J Roentgenol.. J... Fogarasi. Kornaat. 1993.. Paletta. F. S. E. J. pp. M. S. Ratner.. A. "Accuracy of fat-suppressed three-dimensional spoiled gradient-echo FLASH MR imaging in the detection of patellofemoral articular cartilage abnormalities. M. "Articular cartilage of the knee: evaluation with fluctuating equilibrium MR imaging--initial experience in healthy volunteers. Pelc. G. Piraino. Brittain... C.. P. pp. pp. Z. D. A. P.. H.. G.." J Biomech. Yu. A. P.. R. Haghighi. H.. 165(2). 14 Suppl A. S. M.." AJR Am J Roentgenol. P. N. D. Schils. S.. J. A. J. P. F. Trudell. S. Fuchs. Sittek. B. C. "Fat-suppressed three-dimensional spoiled gradient-echo MR imaging of hyaline cartilage defects in the knee: comparison with standard MR imaging and arthroscopy. 1999.. pp." Osteoarthritis Cartilage. 2006.. D. pp. P. D. and Fuchs. 285-289.

124

[19]

Haubner, M., Eckstein, F., Schnier, M., Losch, A., Sittek, H., Becker, C., Kolem, H., Reiser, M., and Englmeier, K. H., 1997, "A non-invasive technique for 3dimensional assessment of articular cartilage thickness based on MRI. Part 2: Validation using CT arthrography," Magn Reson Imaging, 15(7), pp. 805-813. Losch, A., Eckstein, F., Haubner, M., and Englmeier, K. H., 1997, "A noninvasive technique for 3-dimensional assessment of articular cartilage thickness based on MRI. Part 1: Development of a computational method," Magn Reson Imaging, 15(7), pp. 795-804. McGibbon, C. A., Bencardino, J., Yeh, E. D., and Palmer, W. E., 2003, "Accuracy of cartilage and subchondral bone spatial thickness distribution from MRI," J Magn Reson Imaging, 17(6), pp. 703-715. Hardy, P. A., Nammalwar, P., and Kuo, S., 2001, "Measuring the thickness of articular cartilage from MR images," J Magn Reson Imaging, 13(1), pp. 120-126. Eckstein, F., Schnier, M., Haubner, M., Priebsch, J., Glaser, C., Englmeier, K. H., and Reiser, M., 1998, "Accuracy of cartilage volume and thickness measurements with magnetic resonance imaging," Clin Orthop Relat Res(352), pp. 137-148. Wu, J. Z., Herzog, W., and Epstein, M., 2000, "Joint contact mechanics in the early stages of osteoarthritis," Med Eng Phys, 22(1), pp. 1-12. Prevrhal, S., Engelke, K., and Kalender, W. A., 1999, "Accuracy limits for the determination of cortical width and density: the influence of object size and CT imaging parameters," Phys Med Biol, 44(3), pp. 751-764. Prevrhal, S., Fox, J. C., Shepherd, J. A., and Genant, H. K., 2003, "Accuracy of CT-based thickness measurement of thin structures: modeling of limited spatial resolution in all three dimensions," Med Phys, 30(1), pp. 1-8. Nickoloff, E. L., Dutta, A. K., and Lu, Z. F., 2003, "Influence of phantom diameter, kVp and scan mode upon computed tomography dose index," Med Phys, 30(3), pp. 395-402. Suzuki, S., Yamamuro, T., Okumura, H., and Yamamoto, I., 1991, "Quantitative computed tomography: comparative study using different scanners with two calibration phantoms," Br J Radiol, 64(767), pp. 1001-1006. McPherson, E. J., Friedman, R. J., An, Y. H., Chokesi, R., and Dooley, R. L., 1997, "Anthropometric study of normal glenohumeral relationships," J Shoulder Elbow Surg, 6(2), pp. 105-112.

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

125

[30]

Mavcic, B., Pompe, B., Antolic, V., Daniel, M., Iglic, A., and Kralj-Iglic, V., 2002, "Mathematical estimation of stress distribution in normal and dysplastic hips," Journal of Orthopaedic Research, 20(5), pp. 1025-1030. Takase, K., Yamamoto, K., Imakiire, A., and Burkhead, W. Z., Jr., 2004, "The radiographic study in the relationship of the glenohumeral joint," J Orthop Res, 22(2), pp. 298-305. Ateshian, G. A., Soslowsky, L. J., and Mow, V. C., 1991, "Quantitation of articular surface topography and cartilage thickness in knee joints using stereophotogrammetry," J Biomech, 24(8), pp. 761-776. Shepherd, D. E., and Seedhom, B. B., 1999, "Thickness of human articular cartilage in joints of the lower limb," Ann Rheum Dis, 58(1), pp. 27-34. Lorenson, W. E., and Cline, H. E., 1987, "Marching Cubes: a high resolution 3D surface reconstruction algorithm," Computer Graphics, 21, pp. 163-169. Anderson, A., Peters, C. L., Tuttle, B. D., and Weiss, J. A., 2005, "SubjectSpecific Finite Element Model of the Pelvis: Development, Validation and Sensitivity Studies," Journal of Biomech Eng, 127, pp. 364-373. Chen, S. S., Falcovitz, Y. H., Schneiderman, R., Maroudas, A., and Sah, R. L., 2001, "Depth-dependent compressive properties of normal aged human femoral head articular cartilage: relationship to fixed charge density," Osteoarthritis Cartilage, 9(6), pp. 561-569. Huang, C. Y., Stankiewicz, A., Ateshian, G. A., and Mow, V. C., 2005, "Anisotropy, inhomogeneity, and tension-compression nonlinearity of human glenohumeral cartilage in finite deformation," J Biomech, 38(4), pp. 799-809. Shepherd, D. E., and Seedhom, B. B., 1999, "The 'instantaneous' compressive modulus of human articular cartilage in joints of the lower limb," Rheumatology (Oxford), 38(2), pp. 124-132. Lemke, A. J., Kazi, I., de Bary, P., Bernhardt, U., Foerster, P. I., Foerster, M., and Felix, R., 2006, "Development and evaluation of phantom for verification of section thickness at thin-section MR imaging," Radiology, 240(2), pp. 552-558. Hsieh, J., 2003, "Image artifacts: appearances, causes and corrections.," Computed Tomography: principles, design, artifacts, and recent advances, SPIE Press, Bellingham, WA, pp. 167-240.

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

CHAPTER 5 VALIDATION OF FINITE ELEMENT PREDICTIONS OF CARTILAGE CONTACT PRESSURE IN THE HUMAN HIP JOINT

ABSTRACT Methods to predict contact stresses can provide an improved understanding of load distribution in the normal and pathologic hip. The objectives of this study were to develop and validate a three-dimensional finite element (FE) model for predicting cartilage contact stresses in the human hip using subject-specific geometry from computed tomography image data, and to assess the sensitivity of model predictions to boundary conditions, cartilage geometry, and cartilage material properties. Loads based on in vivo data were applied to a cadaveric hip joint to simulate walking, descending stairs, and stair-climbing. Contact pressures and areas were measured using pressure sensitive film. CT image data were segmented and discretized into FE meshes of bone and cartilage. FE boundary and loading conditions mimicked the experimental testing. Fair to good qualitative correspondence was obtained between FE predictions and experimental measurements for simulated walking and descending stairs, while excellent agreement was obtained for stair-climbing. Experimental peak pressures, average

pressures, and contact areas were 10.0 MPa (limit of film detection), 4.4-5.0 MPa and

127 321.9-425.1 mm2, respectively, while FE predicted peak pressures, average pressures and contact areas were 10.8-12.7 MPa, 5.1-6.2 MPa and 304.2-366.1 mm2, respectively. Misalignment errors, determined as the difference in root mean squared error before and after alignment of FE results, were less than 10%. Magnitude errors, determined as the residual error following alignment, were approximately 30% but decreased to 10-15% when the regions of highest pressure were compared. Alterations to the cartilage shear modulus, bulk modulus, or thickness resulted in ±25% change in peak pressures, while changes in average pressures and contact areas were minor (±10%). When the pelvis and proximal femur were represented as rigid, there were large changes, but the effect depended on the particular loading scenario. Overall, the subject-specific FE predictions compared favorably with pressure film measurements and were in good agreement with published experimental data. The validated modeling framework provides a foundation for development of patient-specific FE models to investigate the mechanics of normal and pathological hips.

causing pain. loss of mobility and often leading to the need for total hip arthroplasty.10. [6.g. and experimental evidence supports the concept that mechanical factors play a major role in the development and progression of OA [2-5].128 INTRODUCTION It is estimated that 3 percent of all adults over the age of 30 in the United States have osteoarthritis (OA) of the hip [1]. testing protocols are inherently invasive. To this end. For example. There are no experimental methods available to assess hip contact mechanics noninvasively on a patient-specific basis.13. mechanical data are limited to the measurement area. However.8-14]). Experimental studies have been based on either in vitro loading of cadaveric specimens [8. The ability to evaluate hip joint contact mechanics on a patient-specific basis could lead to improvements in the diagnosis and treatment of hip OA. both experimental and computational approaches have been developed to measure and predict hip contact mechanics (e. epidemiological.14] or in vivo loading using instrumented femoral prostheses implanted into live patients [11. and specific joint pathologies cannot be readily studied. Considerable clinical. While in vitro experimental studies have provided baseline values of hip joint contact pressure. .7].15-17]. the method is highly invasive and existing data are from older patients who have already been treated for advanced OA. The use of instrumented prostheses represents the current state of the art for experimental study of in vivo hip mechanics [11.15-17]. it has been demonstrated that a combination of duration and magnitude of contact pressures and shear stresses on the acetabular and femoral cartilage of hips with acetabular dysplasia can predict the onset of OA [6.

These models have proven useful in the context of parametric or phenomenological investigations.129 Computational modeling is an attractive alternative to experimental testing since it is currently the only method that has the potential to predict joint contact mechanics noninvasively.12. Before computational models can be applied to the study of patient-specific hip joint contact mechanics. However. it is necessary to demonstrate that the chosen modeling strategy can produce accurate predictions and to quantify the sensitivity of model predictions to variations in known and unknown model inputs [22]. Prior computational approaches have included the discrete element analysis (DEA) technique [18-20] and the finite element (FE) method [9. The objectives of this study were: 1) to develop and validate a finite element (FE) model of hip joint contact mechanics using experimental measurements of cartilage contact pressure under physiological loading. . their ability to accurately predict patient-specific contact mechanics is questionable due to over-simplification of joint geometry and an absence of model validation. and 2) to assess model sensitivity to several measured and assumed model inputs.21].

26]. abduction. The joint was screened for osteoarthritis and the cartilage was determined to be in excellent condition (Tonnis Grade 0) [23]. reported hip joint anatomical orientations (flexion. rotation) and equivalent hip joint forces (magnitude and direction) during routine daily activities .28] and bone [25] when geometry is at least 1. 320 mm field of view. A volumetric CT scan of the hip was obtained (512 x 512 acquisition matrix.16]. Experimental loading was based on published data for in vivo hip loads [15. The femur was dislocated from the acetabulum to ensure separation between the acetabular and femoral cartilage in the image data. Bothell.6 mm slice thickness) using a Marconi MX8000 CT scanner (Phillips Medical Systems.75 mm thick.16]. The aforementioned scanner settings produce thickness errors of less than 10% for simulated cartilage [27.0 and 0. The acetabular labrum was dissected free from cartilage..625 mm. Japan) was included to correlate CT voxel intensities with calcium equivalent bone density [25. Kyoto Kagaku Co. in plane resolution = 0.130 METHODS A combined experimental and computational protocol was used to develop and validate a subject-specific FE model of a 25 y/o male cadaveric hip joint (body weight = 82 kg). respectively. The blocks were used to define anatomical axes for referencing joint loading angles using Bergmann’s coordinate system definition [15. WA). Kinematic blocks were attached to the femur and pelvis for spatial registration between FE and experimental coordinate systems [24]. Experimental Protocol All soft tissue with the exception of articular cartilage was removed. 0. Kyoto.625 X 0. A solid bone mineral density phantom (BMD-UHA. Bergmann et al.

[15.131 for 4 patients with instrumented femoral prostheses [15. stair-climbing and descending stairs. Data for the average patient were used in the present study to simulate walking.. The .085 mm [29]. The iliac crests of the pelvis were cemented into a mounting pan in neutral anatomical orientation (anterior iliac spine in plane with plane of pubis symphysis. CA) with a measured positional accuracy of ±0. different film sizes were cut into a rosette pattern using a knife plotter. Three-dimensional orientation of the joint was achieved by flexing. NJ) was used to measure joint contact pressures. abducting and rotating the femur relative to the pelvis.7-10 MPa) was best suited for measuring pressures produced during applied loading. and angles between the planes were calculated. Madison.16]) and attached to a lockable rotation frame (Figure 5. Pressure sensitive film (Sensor Products Inc. The frame was flexed and abducted relative to the vertical axis of the actuator to simulate the orientation of the equivalent hip joint force vector for each loading scenario.1 A). A custom loading apparatus was developed to apply the kinematics that corresponded to these loading conditions (Figure 5. The femur was potted and attached to a lockable ball joint (Figure 5.1). Prior to dissection.1 B).16]. The digitized points were fit to planes. Equivalent joint reaction force angle and anatomical orientation were confirmed by digitizing the loading fixture surfaces (joint force) and planes of the kinematic blocks (anatomical orientation) using a Microscribe-3DX digitizer (Immersion Corp. The orientation of the pelvis and femur fixtures were adjusted until the direction of the joint reaction force vector and anatomical orientation angles were within ± 1° of those reported by Bergmann’s average subject. San Jose.. Pilot testing demonstrated that low sensitivity range film (1.

Experimental setup for loading of hip joint. A) schematic of lockable rotation frame and cement pan used to constrain and orient the pelvis relative to the actuator plane. Polyethylene sheets were used to keep the pressure film dry.1 C). on the surface of the femoral cartilage. and medial aspects of the rosette to reference the location of contact pressures relative to the hip joint. Small notches were cut in the anterior. C) pressure sensitive film. posterior.132 rosette that maximized contact area and minimized overlap was chosen (Figure 5. A) B) Load Cell C) Pelvis Mounting Pan Ball Joint Femur Pot Locks Figure 5. cut into a rosette pattern. B) femur pot attached to a lockable ball and socket joint. .1.

1. Splines representing the outer surface of cortical bone were obtained from automatically thresholded images [25]. the rate of actuator displacement was adjusted until peak loads were achieved within 0. representative of the time required by the average subject reported by Bergmann et al [15. The entire protocol was repeated 3 times for each loading scenario. cartilage and kinematic blocks in the CT image data (Amira 4.133 Peak loads for each activity were simulated by displacing the femur into the acetabulum at a constant rate. The specimen was allowed to recover between trials for over 100 times the interval that was needed to reach peak load. Three to four cycles of preconditioning were necessary to obtain the correct displacement rate. The boundary between trabecular and cortical bone was segmented both . Computational Protocol Commercial software was used to segment surfaces of the cortical bone. The films were stored in a dark location for 48 hours following testing [30] and then scanned and converted to digital grayscale images. MA).33 sec. For each activity. Planes of the kinematic blocks were digitized to establish an experimental coordinate system in neutral orientation. trabecular bone. Cartilage was segmented from air using a threshold value that was found to achieve the greatest accuracy for reconstructing simulated cartilage in a phantom based imaging study [27. An independent calibration curve was established to relate pixel intensity to pressure [31].16]. The three notches on the film were digitized. The actuator was returned to its starting position at the same displacement. The pressure film was then attached to the head of femur between sheets of polyethylene.28]. Mercury Computer Systems. The femur was then displaced into the acetabulum until the target load was achieved. Boston.

The resulting pelvis and femur cortical meshes consisted of 13. respectively (Figure 5. based on the distance between adjacent trabecular bone boundary nodes [25]. XYZ Scientific. and 4. Cartilage surfaces were decimated and smoothed slightly to remove visible triangular irregularities and segmentation artifact. Triangular surfaces were generated for each structure using the Marching Cubes algorithm [32]. Livermore.176 elements. The meshes for acetabular and femoral cartilage were considered converged if there was less than a 5% change in contact area. The final pelvis and femur trabecular bone tetrahedral meshes consisted of 227. . CA) and hexahedral element meshes were created. which were densities consistent with prior pelvic FE modeling [25]. The triangular surface mesh for cortical bone was converted to a quadratic 3-node shell element mesh [25. The interiors of the cortical shell meshes were filled with tetrahedral elements to represent trabecular bone of the femur and pelvis [25]. representative of the mesh density that has been shown to produce accurate predictions of cortical bone strains in prior pelvic FE modeling [25].2). When cortical and trabecular bone blended together in the image data they were manually separated.196 elements.3335].562. The outer cortical surface facets were decimated to a density consistent with our previous study [25]. peak pressure and average pressure between subsequent meshes. Convergence studies were performed by increasing the number of elements through the thickness incrementally while the overall aspect ratios were held constant by adjusting the in-plane mesh resolution. Acetabular and femoral cartilage surfaces were imported into FE preprocessing software (TrueGrid. respectively.108 and 82.134 automatically and semi-automatically. Position dependent shell thickness was assigned to each node.

Overlap between the shell and tetrahedral solid elements [25] was accounted for by assigning an elastic modulus of 0 MPa to all tetrahedral elements that shared nodes with shell elements. Trabecular bone was represented as isotropic hypoelastic with ν=0.8 MPa [37]. homogenous. To establish the neutral kinematic position of each loading scenario.2. An average elastic modulus was calculated for each tetrahedral element using empirical relationships from the literature [25. Cortical bone was represented as hypoelastic.20 [26]. Triangular shell elements indicate cortical bone.135 A) B) Figure 5. with three elements through the thickness. neo-Hookean hyperelastic material [36] with shear modulus G=6. Cartilage was represented with a hexahedral element mesh. Cartilage was represented as an incompressible. and isotropic with elastic modulus E=17 GPa and Poisson’s ratio ν=0. B) close-up at the acetabulum. Incompressibility was enforced using the augmented Lagrangian method [38]. A) finite element mesh of the entire hip joint in the walking kinematic position.29 [39]. the FE model was transformed from the CT coordinate system to the appropriate experimental .26] and the BoneMat software [40].

136 reference system [24]. the baseline threshold value used to segment cartilage was adjusted by ± 50%.452) were analyzed. SI and cement line were fully constrained. bulk to shear modulus ratios of 100:1 (ν=0. thickness and FE model boundary conditions affected predictions of cartilage contact mechanics. Updated cartilage FE hexahedral meshes were generated based on these surfaces. The baseline cartilage shear modulus was altered by ± 1 SD (G=10. three separate cases were analyzed: 1) bones were assumed rigid. while the nodes at the pubis. 2) the rigid constraint at the pubis joint was removed. Sensitivity Studies Sensitivity studies were performed to investigate how changes in assumed cartilage material properties.495) and 10:1 (ν=0. Contact between the femoral and acetabular cartilage was enforced using the penalty method [43].28].45 and 2. The Mortar method was used to tie acetabular and femoral cartilage to the acetabulum and femoral head. Separate models were generated for each loading activity. those residing within the sacroiliac (SI) and pubis joint. respectively [41.42]. and 3) trabecular bone was removed so that deformation of only the cortical bone was considered. To ascertain the effects of the assumption of cartilage incompressibility.68 MPa) using standard deviations for human cartilage [44]. The rigid femur nodes were constrained to move only in the direction of applied load. All analyses were performed using NIKE3D [43]. . and those inferior to the cement line of the potted femur were defined rigid according to anatomical boundaries determined experimentally. Nodes superior to the pelvis cement line. To account for differences in segmentation threshold intensity between real and simulated cartilage [27. yielding a total of 27 models. To quantify the affects of model boundary conditions.

FE pressure predictions were transformed into a synthetic film image with the same dimensions. Only those pixel intensities within a user specified range were compared. and 2) differences in the magnitude of contact. Further constraints were made in the calculation of RMS error . The program allowed for the investigation of two types of error: 1) misalignment between FE and experimental results. A root mean squared (RMS) error criterion was used to assess the degree of similarity by comparing pixel intensity values between FE and experimental images. Separate synthetic FE images were created and aligned for each experimental image since the pressure films were not placed in the exact same anatomical position between loading trials. Finally. as the pressure films. This range was taken as the full sensing range of the film (1. the program converted the grayscale images of pressure to fringed color using the calibration curve. Surface nodes of the femur cartilage FE mesh were fit to a sphere and then flattened by a spherical-to-rectilinear coordinate transformation. The synthetic image was aligned with the pressure film image using the experimentally digitized notches on the pressure film. Next.7-10 MPa) but was also determined for smaller 2 MPa bins of pressure to assess the ability of the FE models to predict pressures within specific ranges. First. a difference image between each synthetic and experimental image was created.137 Data Analysis A program was developed to compare FE predicted cartilage contact pressures with results from pressure sensitive film. The rosette cuts on the experimental film were duplicated in the synthetic FE pressure film image by moving the FE pressure results circumferentially. including rosette cuts. according to the wedge angle of the rosette.

the experimental film data were considered the “truth”. contact area. if a pixel in the synthetic FE image indicated a pressure within the specified range but the corresponding experimental film pixel did not.138 because. Pixel intensities that indicated pressures within the film range (1. then the pixel was not included in the calculation.7-10 MPa) were used to determine FE and experimental average pressures. posterior. Peak pressure. and center of pressure were calculated for each experimental film and synthetic FE image after the synthetic FE films were aligned with experimental images to minimize RMS error. However. Contact area was calculated by multiplying the number of pixels within the detectable pressure range of the film by the area of each pixel (0. Misalignment error could occur due to fitting the FE mesh to a sphere. along with the re-calculated anterior. Misalignment error was also distinguished from magnitude error.0154 mm2). Misalignment error was quantified independently by calculating the RMS error real time while the synthetic FE image was rigidly rotated about a spherical coordinate system. average pressure. or from FE model inaccuracies. FE peak pressure was determined by recording the maximum FE pressure value within the region of experimentally measured contact. The rotations required to minimize RMS error. then the pixel was included. from inaccurate digitization of the notches used to align the results. in this study. and lateral positions of the notches were recorded. The center . if an experimental pixel was within the range but its corresponding FE pixel was not. Specifically. Misalignment error was then expressed as the difference in pre and post alignment RMS error whereas magnitude error was taken as the post alignment error. Experimental peak pressures were calculated as the maximum experimental film pixel intensity.

average pressure and contact area were reported as combined results from the three loading scenarios analyzed. For the cartilage sensitivity studies.experimental film COP) was expressed as the anatomical difference in anterior/posterior and medial/lateral positions over the detectable range of the film.10 MPa) since the pressure range was no longer limited by the film. .139 of pressure (COP) was found by determining the center of the image. and contact areas were calculated per the methodology discussed above. First. but a larger pressure range was used (0. For the boundary condition sensitivity studies. percent changes in peak pressure. RMS errors were calculated per above. a baseline image was created for each baseline FE model from which subsequent sensitivity study predictions were compared.5 . Similar analyses were performed for the sensitivity studies. The difference in the centers of pressure between images (FE synthetic film COP . which was weighted according to pixel intensity. results were reported as percent changes with respect to each loading scenario. Changes in peak and average pressure.

5±0. respectively as estimated using the cortical bone thickness algorithm [25] (Figure 5.2).000 and 23.140 RESULTS FE Mesh Characteristics Cortical bone thickness in the pelvis and femur averaged 1.415 elements.3. 2.9±2.4 and 1. respectively.0 mm 0.2). The resulting trabecular bone modulus in the pelvis and femur averaged 270±188 and 295±198 MPa. Cartilage thickness in the acetabular and femoral cartilage meshes was 1. respectively (Figure 5. Three elements through the cartilage thickness were necessary to yield converged FE predictions. The final mesh for acetabular and femoral cartilage consisted of 15.8 mm Figure 5. respectively.3 mm.8±0.3). and each analysis took on the order of 2 hours of wall clock time.6±0.5 mm. respectively.8 and 2. The final FE model was composed of ~400k elements (Figure 5. . Contours of cartilage thickness. Femoral and acetabular cartilage was thickest in the anterormedial and superior regions.

61 MPa for walking. 8 7 6 500 Average FE Pressure Average Exp. There was excellent agreement between FE predictions and experimental measurements for stairclimbing.0 MPa and 321. FE predictions of peak pressure were 10. Error bars indicate standard deviation.2 MPa and 304.45. respectively. FE predicted and experimentally measured average pressure (left y-axis) and contact area (right y-axis). FE models tended to overestimate average pressure and to underestimate contact area during simulated walking and descending stairs.4.2-366. .7-10.1-6. Experimental average pressure and contact area ranged from 4.73 and 11.1 mm2.78. However. respectively. Average Pressure.0 MPa (range of film detection).4). All pressure films recorded pressures at the upper limit of film detection (10 MPa). descending stairs and stairclimbing. Contact Area 400 350 Average Contact Area (mm2) Average Pressure (MPa) 5 4 3 2 300 250 200 150 100 1 0 50 Walking Descending Stairs Stair-Climbing 0 Figure 5. Contact Area Experimental pressures ranged from 1.1 mm2.141 Peak Pressure. less than 5% of the total pixels fell into this category. Pressure 450 Average FE Contact Area Average Exp.9-425. respectively (Figure 5. 12. while FE predicted average pressure and contact area ranged from 5.

Overall.6 bottom row). Experimental pressure distributions were similar during simulated walking and descending stairs. The majority of contact occurred along the lateral aspect of the acetabulum for all three loading activities (Figure 5. The contact area moved from anterior to posterior as the resultant load vector changed from shallow extension during descending stairs to more moderate flexion angles during walking and stair-climbing (Figure 5. However. with a horseshoe shaped bicentric contact pattern directed anterorlaterally to posterormedially. As the femur was rotated internally during simulated stair-climbing the contact pattern was oriented in a lateral to medial direction (Figure 5. . Patterns of FE-predicted contact varied with the loading activity (Figure 5.5 middle row).6 bottom row).142 Contact Patterns The experimental pressure recordings revealed bi-centric patterns of contact during simulated walking and descending stairs and a more or less mono-centric pattern during simulated stair-climbing (Figure 5.6 B). experimental bi-centric contact patterns during simulated walking and descending stairs were not predicted by the FE models (Figure 5. the magnitude and location of FE predicted contact pressures corresponded well with experimental measures.5 top row).5 top row).

Bottom row .experimental film contact pressures (representative results are shown).143 Walking Experimental Descending Stairs Posterior Stair-Climbing 10 MPa Medial Difference Lateral FE 0 MPa Anterior Figure 5. Models predicted mono-centric.FE synthetic films. Note: FE synthetic films and difference images are shown prior to manual alignment with experimental results. The best qualitative correspondence was during stairclimbing. Middle row . . Top row . Bi-centric patterns of contact were observed during simulated walking and descending stairs. indicating locations where contact was not predicted by the models.5. while a mono-centric pattern was observed during stair-climbing. irregularly shaped patterns of contact.difference images.

144 Walking Descending Stairs Stair-Climbing 8 MPa Superior-Anterior View 0 MPa Oblique Lateral View Figure 5. FE predicted contact pressures on the femur (top) and acetabulum (bottom). . The highest contact pressures primarily occurred near the lateral region of the acetabulum. Acetabular cartilage contact pressures moved from anterior to posterior as the equivalent joint reaction force vector changed from shallow extension during descending stairs to deep flexion during stair-climbing.6.

This finding indicates that FE models were best suited for predicting the higher stressed regions of cartilage. The rotations and resulting translations of the experimental film fiducials required to minimize RMS errors were less than 3° and 5 mm for walking and stair-climbing but were substantially higher for the descending stairs case (22° and 9 mm) (Table 5. When RMS error was plotted in 2 MPa pressure bins. The smallest difference in the COP occurred for stair-climbing.2). Overall. . Differences in contact pressure were greatest for descending stairs and were least during stair-climbing (Figure 5.1). residual RMS errors were on the order of 30%. corroborating the good qualitative correspondence between FE synthetic and experimental films in these locations (Figure 5. As suggested by the difference images.5). misalignment errors were less than 7% (Table 5. calculated prior to manual alignment between FE synthetic and experimental films. Differences in center of pressure locations.145 Misalignment and Magnitude Errors Difference images. In general. Following manual alignment. RMS errors decreased to around 10-15% at the maximum bound of pressure analyzed (8-10 MPa). as calculated over the entire film detection range. errors were greatest for descending stairs and least for stair-climbing (Figure 5. were less than 10 mm (Table 5. COPs for the FE models were directed more lateral (-) and anterior (+) to experimentally measured COPs.1). further clarified the degree of qualitative agreement between FE synthetic and experimental films (Figure 5.5 bottom row). while the largest difference occurred during descending stairs.5 bottom row).5 bottom row).

09 (0.26 (2.90) 26.93) Walking Descending Stairs Stair-Climbing .2.86) Rot Z (°) (±SD) -0.14) Rot X (°) (±SD) 2.65 (7.49 (2.74 (0.92 (0.196) 3.75 (2.1.70) ∆ Position mm (±SD) 2.00) 1.14 (0.11) 6.31 (0.37 (1.86) 0.51) Magnitude RMS Error (%) (±SD) 32. FE misalignment and magnitude errors. Center of Pressure Difference (mm) Medial/Lateral (±SD) Anterior/Posterior (±SD) -6.22 (1.06) 2.77 (0.00) Table 5.24 (0.39) -7. Differences in centers of pressure between synthetic FE and experimental films (negative = lateral/posterior.08 (0.55 (1. Misalignment RMS Error (%) (±SD) Walking Descending Stairs Stair-Climbing 0. The rotations required to align results and associated changes in the positions of the film fiducials are shown.22) 9.13) 3.88 (1.60 (1.38) -22.59 (2.34) 7. positive = medial/anterior).34) 4. Misalignment error was calculated as the reduction in total RMS error after the synthetic films were manually rotated to minimize RMS error between FE and experimental films.146 Table 5.03 (0. Magnitude error was the residual error following alignment.26) 33.05 (6.32) 8.73 (0.58) 2.

495 did not have an appreciable effect (Figure 5. However.7 C).5%.5 to ν=0. while changes in average pressure and contact area were less than 10% (Figure 5.7 A). . Lowering the cartilage Poisson’s ratio from ν=0. indicating that the spatial distributions and magnitudes of contact pressure did not change substantially. a further decrease in the Poisson’s ratio to 0.452 resulted in a 25% decrease in peak pressures.7 B). RMS differences between baseline and all cartilage sensitivity study results were approximately 6. Altering the thickness of femoral and acetabular cartilage (~ 10% change average cartilage thickness) resulted in less than a ±10% change in FE predictions (Figure 5.7 B). while changes in average pressure and contact area were around ±10% (Figure 5.147 Sensitivity Studies .Cartilage Material Properties and Thickness Changes of ±50% to the shear modulus resulted in approximately a ±30% change in FE predictions of peak pressures.

7. A) effects of changes to the shear modulus by ±1 SD. B) effects of changes to cartilage compressibility (100:1. .148 A) 40 30 20 0 -5 B) Percent Change 10 0 Percent Change +1S -1SD -10 -15 100:1 -10 -20 -30 -40 Peak Pressure Average Pressure Contact Area -20 -25 -30 Peak Pressure Average Pressure Contact Area 10:1 Bulk:Shear Shear Modulus -35 10 8 6 C) +50 % -50% Percent Change 4 2 0 -2 -4 -6 -8 Peak Pressure Average Pressure Contact Area -10 Cartilage Thickness Figure 5. average pressure and contact area due to alterations in cartilage material properties and geometry. Error bars indicate standard deviations over the three loading activities analyzed. Percent changes in peak pressure. C) effects of altering the cartilage thickness. 10:1 bulk to shear ratios).

Average RMS differences between baseline results and the latter boundary condition sensitivity studies were only 3. FE predictions of peak pressure. FE predictions changed on the order of -15 to +5% (Figure 5.5%.e. Representing the bones as rigid structures affected both the magnitude (Figure 5. changes in FE predictions ranged from -25 to +5% (Figure 5.8 A) and spatial distribution (Figure 5.8 A).FE Boundary Conditions Rigid bone models decreased computation times from ~2 hours to <10 minutes.1%.8 C). average pressure and contact area were altered but also varied according to the loading scenario analyzed (Figure 5. When the rigid constraint on the pubis joint was removed. only the cortical shells supported the cartilage.149 Sensitivity Studies . .2±5. but the degree of effect depended on the loading activity.8 B). RMS differences in synthetic films between baseline and rigid bone models averaged 29. i. Finally. when the trabecular bone was removed.9) of cartilage contact pressure.

W.150 A) 120 100 80 Peak Pressure Average Pressure Contact Area 5 B) W 0 DS SC Percent Change 60 40 20 0 Percent Change W DS SC Rigid Bones -5 -20 -40 -60 -10 Peak Pressure Average Pressure Contact Area -15 Pubis Joint Free C) 10 5 0 W DS SC Percent Change -5 -10 -15 -20 -25 -30 Peak Pressure Average Pressure Trabecular Bone Removed Contact Area Figure 5. . DS.8. respectively. B) effects of removing the pubis joint constraint. C) effects of removing the trabecular bone from the FE analysis. descending stairs. average pressure and contact area due to alterations in boundary conditions. Percent changes in peak pressure. and stair-climbing. SC indicate walking. A) effects of a rigid bone material assumption.

.9. Contours of cartilage contact pressure predicted by the baseline (top row) and rigid bone FE models (bottom row) for the three loading activities.151 Walking Baseline FE Descending Stairs Posterior Stair-Climbing 10 MPa Lateral Medial Rigid 0 MPa Anterior Figure 5. The largest effect of the rigid bone assumption occurred for simulated walking and descending stairs.

It was necessary to move the linear actuator up by ~20 mm to access the film markers. This offset was minor during walking and stair-climbing but was greater during descending stairs. The FE model provided very reasonable predictions of both the spatial distribution and magnitude of cartilage contact pressure under the simulated loading conditions. the femoral neck would have contacted the edge of the acetabulum. The femur was in extension for this loading activity and when the translation was applied.152 DISCUSSION To our knowledge this is the first study to validate FE predictions of cartilage contact pressure with experimental measurements in the human hip joint. Excellent predictions were obtained for simulated stair-climbing. In contrast. It was assumed that this displacement resulted in a perfect vertical translation for purposes of defining the marker coordinates. but when the coordinates were plotted relative to the translated model they did not reside on the surface of the cartilage. the descending stairs case required a substantial amount of manual rotation (Table 1). It is likely that the majority of misalignment error was due to the method of digitizing the film fiducials during the experiment. the magnitude of pressure in these locations was low in comparison to the anterior region where the FE models provided more reasonable correspondence. resulting in an offset of the film marker coordinates. The posterior aspect of the bi-centric experimental contact pattern was not predicted by the FE model for walking and descending stairs. Nevertheless. Contact in this location . Small manual rotations of the pressure film were necessary to minimize RMS errors for simulated walking and stair-climbing.

This finding demonstrates that peak pressure prediction requires more accurate model inputs for cartilage geometry and material properties than for average pressure prediction. Therefore. the modeling strategies developed herein may be well suited for predicting the primary region of contact. Computational models of the hip have often represented bones as rigid structures [12. which may be sufficient for many patientspecific modeling applications. which is an attractive simplification because solution times are greatly reduced.45]. bulk modulus or thickness (±10%). However. The results of the sensitivity study that removed the pubis constraint demonstrated only minor differences in FE predicted . The effect is modulated by the specific boundary and loading conditions in the model. The finding that RMS magnitude errors decreased when the bounds of pressure were increased suggest that the models were best suited for predicting localized “hot spots”. greater changes in peak pressure were noted (up to ~25%). The present study demonstrated that the assumption of rigid bones can alter predictions of cartilage contact stresses in the hip. investigators should use caution when representing the bones as rigid for modeling cartilage contact mechanics in the human hip. FE predictions of average pressure and contact area were not overly sensitive to changes in the cartilage shear modulus. the FE models assumed that the pubis joint was rigid. Although the contralateral pelvis was left intact in the experimental study.153 would not have occurred with the hip in moderate and deep flexion during walking and stair-climbing. Because the consequence of the rigid bone assumption cannot be assessed without a direct comparison to the case of deformable bones.

13. While this simplification was warranted for the current study. Large differences in material properties.12. Bi-centric. horseshoe shaped patterns extended from the anterior to posterior aspect of the femur were noted [13]. in fair agreement with the results of the current study. [13] ranged from 7 MPa at 50% body weight to 9 MPa at 300% body weight.21. simulations of side-impact loading [46]).45] with either . for patient-specific modeling applications it may be appropriate to exclude trabecular bone assuming that similar boundary and loading conditions are assigned. Experimental studies have used pressure sensitive film to measure hip joint contact pressures under similar loading conditions [8.154 cartilage contact mechanics. but some general trends can be identified. The results of the sensitivity study suggest that it plays little mechanical role with regard to cartilage contact stresses. non-symmetric pressure distributions [8.14]. Since the reported elastic modulus of trabecular bone is orders of magnitude less than cortical bone. thereby giving credence to this model assumption. we investigated whether or not trabecular bone needed to be represented in the models. Nearly all FE hip joint modeling studies to date have used two-dimensional. Afoke et al. It is worth mentioning that all of the aforementioned studies predicted irregular. Therefore.14]. Peak pressures measured by von Eisenhart-Rothe et al.13. plane strain models [9.g. [8] measured peak pressures on the order of 10 MPa at 350% body weight and the anterorsuperior surface of the cartilage was identified as an area of high pressure [8]. geometry and boundary conditions make it impossible to directly compare the FE predictions from this study with prior modeling studies. it may not be appropriate for models where load is directed more medially (e.

To our knowledge. yielding peak contact pressures on the order of 5 MPa.47±0. which is in good agreement with Macirowski et al’s predictions. Qualitatively our predictions of . The model assumed spherical geometry and concentric articulation.29 MPa over the three loading scenarios analyzed. The acetabulum was step loaded to 900 N using an instrumented femoral prosthesis. When the experimentally measured total surface stress was applied to the FE model. [12] used a combined experimental/analytical approach to model fluid flow and matrix stresses in a biphasic contact model of a cadaveric acetabulum. scaling the applied load of our model to 900 N and assuming a lower bound of 0. [20] developed a dynamic DEA model to investigate the distribution of hip joint contact pressures using the Bergmann gait data. FE predicted pressures were irregularly distributed over the surface of the femoral head.21]. making it impossible to directly compare average results. The lower range of pressure used to determine average pressures was not specified.45] or deformable bones [9. average predicted pressures (solid stress + fluid pressures) were approximately 1. the earliest FE contact model was reported by Brown and DiGioia [9].155 rigid [12. Macirowski et al. This is the only previous FE study to explicitly model the acetabular cartilage thickness. Rapperport et al. In this study.75 MPa. Yoshida et al.) yields average pressures of 2. Values of peak pressure were on the order of 4 MPa at loads representative of those applied in the current study.3 MPa to calculate average pressure (lowest pressure isobar indicated by Macirowski et al. [21] developed a similar model that predicted peak pressures on the order of 5 MPa at 1000 N of load. Using rigid bone models resulted in predictions only slightly different than the deformable bone model. However.

. 3. The aforementioned computational models assumed a cartilage modulus ranging from 10-15 MPa [9.48. and stair-climbing were substantially less than those predicted in the current study (3. This discrepancy is most likely due to the assumptions of spherical geometry and concentric articulation in the prior computational studies. the results of our sensitivity studies demonstrate that this is not the case. the FE models developed herein predicted higher contact pressures than previous FE and DEA studies.49].77.18-20].16] the use of average data loading data likely did not . Several limitations of the current study must be mentioned. peak pressures predicted in this study were still nearly double those reported previously [9. While one might expect that a higher modulus would result in higher contact pressures. Even with a 25% reduction. With the exception of the study by Macirowski et al. experimental loads were based on average in vivo data from older patients who had already undergone treatment for advanced hip OA. the hip joint is not spherical and cartilage thickness is nonuniform [12.156 primary contact during simulated walking. respectively. Peak pressures during walking. Given the large inter-patient variation in joint kinematics observed by Bergmann et al.47]. First. respectively).8 MPa) in the current study.71 MPa. and stair-climbing are in good agreement with the results of this study. [15. as changes in the cartilage shear modulus of ±50% resulted in only a ±25% and ±10% change in peak and average contact pressures.21] yet cartilage was given a baseline modulus of ~40 MPa (G=6. Although the literature suggests that normal hips may be modeled as spherical structures with concentric articulation [19. 5. descending stairs. but the spatial distributions of contact were markedly different. descending stairs.26.

making the use of pressure film appropriate in the current study.157 accurately represent the actual kinematics for the specimen in this study. The primary focus of the present research was to quantify cartilage contact pressures in the peri-acetabular region rather than bone stresses in areas where muscles were attached. it was not feasible to do so using larger rosettes as they caused excessive overlap and crinkle artifact during pilot testing. . However. While it would be desirable to capture the entire region of contact. making its inclusion a potential confounding factor.53]. Although actions of individual muscles were not considered. The acetabular labrum was removed in this study.53]. Therefore. we could justifiably model the action of all muscles as a single equivalent force vector acting through the hip joint. film measurements have been shown to be equivalent to the contact stresses resulting from an incompressible elastic analysis [52]. Our approach is justified since the objective of the experimental protocol was to apply realistic loading and boundary conditions that could be reproduced in the FE simulations for model validation. The labrum was removed because its material properties are unknown.51].16]. the equivalent joint reaction force was based on in vivo data [15. A limitation of film pressure measurement is that the technique records a “high watermark” rather than measurements of time-loading history [50. yet it has been suggested that the labrum helps to maintain fluid pressurization in the hip [45. Results from the pressure measurements indicate that contact occurred beyond the perimeter of the film during simulated walking and descending stairs. It is possible that leaving the labrum intact would have altered the experimentally measured and computationally predicted contact pressures [45.

60].12-14. it was represented as incompressible hyperelastic in this study. The sensitivity studies established the modeling inputs and assumptions that are important for predicting contact pressures. Predictions were in good agreement with other experimental studies that used pressure film. perhaps better.59. our approach for subject-specific FE modeling of the hip joint produced very reasonable predictions of cartilage contact pressures and areas when compared directly to pressure film measurements. variation in stiffness over its surface [44. In conclusion. We recently demonstrated the equivalence between biphasic and incompressible hyperelastic FE predictions during instantaneous loading [55]. . Cartilage also exhibits depth dependent material properties [56]. In vitro studies suggest that fluid flow is minimal during fast loading [12. Future modeling efforts should investigate the importance of these effects via sensitivity studies.53]. piezoelectric sensors and instrumented prostheses [8. making our assumption of incompressibility warranted given the loading rates used in the experiments. The validated FE modeling procedures developed in this study provide the basis for the future analysis of patient-specific FE models of hip cartilage mechanics. predictions of contact stress magnitude and distribution.158 Although cartilage exhibits biphasic material behavior [54].57] and tension-compression nonlinearity [58]. Incorporating these aspects might have resulted in different.

N. R. K. J. D. Harris. "Contact Pressures in the Human Hip Joint Measured in Vivo. Y.. Byers." J Biomech. T. Burgess. "The Effect of Contact Pressure Elevations and Aseptic Necrosis on Long Term Outcome of Congenital Hip Dislocation. [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] . W. and Mann. pp. S. A. M.. 69. A." J Orthop Res... pp. 1974. F. 373-84. 1983. R. C. D. T. Jordan... R. and DiGioia. 437-48. Brown. Weinstein. 1995. A. C.. 291. H." in Osteoarthritic Disorders." J Biomech. Kington. 1995. and Weinstein. McAlindon.. A.. "The Reaction of Articular Cartilage to Injury and Osteoarthritis (First of Two Parts). R. E. 536-41. 17. D. "Imbalances of Anabolism and Catabolism of Cartilage Matrix Components in Osteoarthritis." J Bone Joint Surg Br. Brown... T. Lawrence.. Guilak.D. A.. Hadley. M. S... 133. S. T. J. J. P.. Sharma.159 REFERENCES [1] Felson. "Contact Pressures in the Human Hip Joint.. American Academy of Orthopaedic Surgeons.. A. and Shaw. 159-66. G. pp." Proc Natl Acad Sci U S A. 504-510. V... R. "Mechanical Factors in Articular Cartilage and Their Role in Osteoarthritis. A. 2000. R." Ann Intern Med. Lane. D. Sowers. L. pp. Mow. G. K.. A. W.. R. Ateshian. A. 1974. J. 1285-92. W. D. 291. H.. "Osteoarthritis: New Insights. Afoke. 1335-40. "A Contact-Coupled Finite Element Analysis of the Natural Adult Hip. and Fries. R. Hodge. G. 16.. 1995.. pp... 2879-83. D.. "The Reaction of Articular Cartilage to Injury and Osteoarthritis (Second of Two Parts). 635-46.. M. T.. 8. 83. J." N Engl J Med. Zhang. Poole... C. N. American Academy of Orthopaedic Surgeons.. pp.. S. F. Hirsch. T. S. Carlson. T. Part 1: The Disease and Its Risk Factors. L... 3rd. Y. Z. 1986. T. Yanovski. T. pp. Fijan. 1984. Brandt." in Osteoarthritic Disorders." N Engl J Med. M. "In Vitro Contact Stress Distributions in the Natural Human Hip. Buckwalter.. 1990. D. "Chronic Stress Tolerance Levels for Human Articular Cartilage: Two Nonuniform Contact Models Applied to Long-Term Follow-up of Cdh. Helmick... N. Poole. L. A.. Brown. 1987. L.L. Brown. and Ratcliffe. C. Spector. 28. C. Dieppe. Mankin. pp... Mankin.. Maxian. W. P. and Hutton. H. Nevitt. pp." J Biomech. Setton. D..

A... 34. 2001.. 1998. pp. R.. "Three-Dimensional Dynamic Hip Contact Area and Pressure Distribution During Activities of Daily Living. J. Rapperport. T.. 21.. Tepic. Tönnis...160 [12] [13] Macirowski. Wilckens. and Miura.. 2006.. 116. P. Landgraf. Validation and Sensitivity Studies in Computational Biomechanics. Eckstein. F. M. N. E.. E. Springer-Verlag. D. M. pp. J." J Biomech.. Iwasaki.. Congenital Dysplasia and Dislocation of the Hip in Children and Adults. Rohlmann. Y.. and Duda. "Quantitative Determination of Joint Incongruity and Pressure Distribution During Simulated Gait and Cartilage Thickness in the Human Hip Joint.. Heller.... 1999. 257-64. Ellis... Konishi. E. Carter.. J. and Putz... MacWilliams. Genda. "Contact Finite Element Stress Analysis of the Hip Joint. pp. Hasegawa. 2006. 279-88. "Cartilage Stresses in the Human Hip Joint. pp. Bergmann. R.. R. 1985. Li. Steinlechner M. Contact Stress and Subchondral Mineralization in Human Hip Joint Specimens. E.. Rock. C. pp. Y.. 7. J. Bergmann. 10-8. 1987. Mann. 39. W. Hip98: Data Collection of Hip Joint Loading on Cd-Rom. Barrance. 1996-2004. Genda. Anderson. pp. J. Deuretzbacher. and Chao. G.. G. 202-6. Muller-Gerbl.. B. 859-71. A. 34. 532-539. Free University and Humboldt University. N.. von Eisenhart-Rothe. "Hip Contact Forces and Gait Patterns from Routine Activities. Kitagawa.." J Biomech.." Anat Embryol (Berl).. A.. Yoshida. W. Berlin. N. 1997. 1994. pp. 435-46.. [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] . Faust. 2001. J. R... B. C. J. "A Radio Telemetry Device for Monitoring Cartilage Surface Pressures in the Human Hip. A. In Press Dec 2006. D. pp. D. Fetto. Y. Carlson. Strauss." Journal of Orthopaedic Research. M. R. Muller-Gerbl M and Eckstein F. 1995. and Mann. "Normal Hip Joint Contact Pressure Distribution in Single-Leg Standing--Effect of Gender and Anatomic Parameters. W. G. 1974. F. pp." Arch Orthop Trauma Surg.. 895905." J Biomech.. G. "Verification. 3. J. E. Graichen.. H." J Orthop Res. D. and Chao." J Biomech Eng. C.." Computer Methods in Biomechanics and Biomedical Engineering. "A Computer Simulation Study of Normal and Abnormal Hip Joint Contact Pressure. and Weiss. A.. 114. and Harris. G. H.. "Direct Comparison of Contact Areas. A.. and Schurman." IEEE Trans Biomed Eng. J. S. T. Berlin. von Eisenhart-Rothe R.. 195.

2005. 2005. "SubjectSpecific Finite Element Model of the Pelvis: Development. Lujan.. J. E. A. L... A. pp. San Diego. J.. A. T. T. J. A. B.. D. CA.. A. Chapman and Hall.. Two Dimensional Shells. J. A.. M. 27. Anderson. 127. [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] . Lake. "Factors Influencing Cartilage Thickness Measurements with Multi-Detector Ct: A Phantom Study. 359-66. and Eberhardt." J Biomech Eng." in Optical Measurement Methods in Biomechanics..161 [24] Fischer.. J. "Contact Pressures in the Flexed Hip Joint During Lateral Trochanteric Loading. L. 9. D. Ellis. P. Beason. C.. and van Erning. D. A. 23. Lorensen.. 167-181. "Marching Cubes: A High Resolution 3d Surface Construction Algorithm.. E. Liggins. pp. and Weiss. Hughes. S. T. 26. pp. 1980.. T. K.. 193-7. pp. Odgaard. A. and Weiss. 364-73.. M. J. R. Alonso. Ellis. "The Practical Application of Fuji Prescale PressureSensitive Film. Validation and Sensitivity Studies." J Orthop Res.." Interational Journal for Numerical Methods in Engineering. C. Etheridge. Pfaeffle. 2007. In Press.. S. 21. and Liu. A. "Nonliner Finite Element Analysis of Shells: Part I. B. Ellis. "Generalization of Selective Integration Procedures to Anisotopic and Nonlinear Media.. T. H. Peters. 127. 163-169. J." presented at Transactions of the 53rd Annual Meeting of the Orthopaedic Research Society.. B. 15. B.. B.. R. T. "Factors Influence Cartilage Thickness Measurements Using Multi-Detector Ct: A Phantom Study. E." Comput. J. J. W." J Biomech.. Feb 2007. Huiskes. Dalstra. pp. J. Graph. pp. J. A.. L.. M. 523-35.." J Biomech Eng. K.. 2005. S.. 2001. 2007. 1997. pp.. J. pp. L. "Mechanical and Textural Properties of Pelvic Trabecular Bone. and Weiss." Radiology. Tuttle. W. and Weiss. C. and Woo. Peters... and Cline. H. Hughes. "Simultaneous Measurement of Three-Dimensional Joint Kinematics and Ligament Strains with Optical Methods.. E.. 1987." J Biomech... A. B. Manson. J. W." Compuational Methods in Applied Mechanics. 34. London. Anderson. Peters... 377-83. L. Sparks. 1993. P. Anderson. 1981. "A Method for Measuring Joint Kinematics Designed for Accurate Registration of Kinematic Data to Models Constructed from Ct Data. Plaizier. Tomaino.

J." J Biomech Eng. 630-9.. Hestenes. P. R.O." International Journal for Numerical Methods in Engineering. 2004. J.. 65-73. 20. R. S. 2004.. Park. M. A." J Orthop Res. Mantovani.... and Prasad. pp. Ganz. W. Ramaniraka. and van Erning. and Farron." Osteoarthritis Cartilage. J. Three Dimensional Shells. 2002. J.... Maker. 193." Computer Methods in Applied Mechanics and Engineering.. 4. 33. pp. Dalstra. Athanasiou. Huiskes. Puso. Ferencz. H... M. A.. T... 953-60. 12. "Multiplier and Gradient Methods. 1995. P. R. Three-Dimensional Finite Element Code for Solid and Structural Mechanics.. 601-29.. 315-36." J Biomech. M. 2004. A. N... 59. 1981." Lawrence Livermore National Laboratory Technical Report. pp. "A Finite Element Model of the Shoulder: Application to the Comparison of Normal and Osteoarthritic Joints. pp. C. Avon). P.. 735-40. and Liu. Agarwal. 117. Bryant. and Dzida. Zannoni. G.. pp. R. pp.162 [35] Hughes." Med Eng Phys. UCRL-MA.. "Comparative Study of the Intrinsic Mechanical Properties of the Human Acetabular and Femoral Head Cartilage.. S. pp. and Laursen. 1998. Ferguson. M.. Buchler. Rouhana.. 2006. Iannotti. pp. "Nike3d: A Nonlinear. S.. "Material Properties Assignment to Finite Element Models of Bone Structures: A New Method. Ruan.. 331-362.. and Ateshian. and Viceconti. A. Barbat. K.. 303-320." Journal of Optimization Theory and Applications." Computational Methods in Applied Mechanics. L. 1969. C. A. 1990. M. and Ito. "The Influence of the Acetabular Labrum on Hip Joint Cartilage Consolidation: A Poroelastic Finite Element Model. and J. K. 1994. "Nonlinear Finite Element Analysis of Shells: Part Ii. R. N. S. L. Rakotomanana.. T. F.. "A 3d Mortar Method for Solid Mechanics. T. Hung. 27. El-Jawahri. A. pp. 2000. K. J." Clin Biomech (Bristol. J. S. R. R. 12. Puso. "Analysis and Evaluation of the Biofidelity of the Human Body Finite Element [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] . "Mechanical Response of Bovine Articular Cartilage under Dynamic Unconfined Compression Loading at Physiological Stress Levels. 272-8. "Development and Validation of a Three-Dimensional Finite Element Model of the Pelvic Bone. Implicit. "A Mortar Segment-to-Segment Contact Method for Large Deformation Solid Mechanics. T. 340-9.. W. A. pp. M... A.. B. 17.

. [47] Konishi. M. D. T. Falcovitz. 1999. 27. and Sah. E. B. S. "Determination of Acetabular Coverage of the Femoral Head with Use of a Single Anteroposterior Radiograph. G. J.. D. 58. H. 2006. "Equivalence between ShortTime Biphasic and Incompressible Elastic Material Response. B. L. Mow. pp. H. [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] . Rudert. D.. 38." J Biomech. Lai." Ann Rheum Dis. 2001. P. and Harris. and Mow. Chen." Journal of Biomechanical Engineering.. Kuei. Mann. Lai. "The 'Instantaneous' Compressive Modulus of Human Articular Cartilage in Joints of the Lower Limb. G. pp." J Biomech." Osteoarthritis Cartilage. K. 75.. A. pp. 27-34. V. Ateshian. D.. "Thickness of Human Articular Cartilage in Joints of the Lower Limb. A. and Armstrong.. and Grosland." J Biomech. 9.. pp. W.. Zhu. W." J Biomech Eng. "New Methods for Assessing Cartilage Contact Stress after Articular Fracture. J. M. M. T. pp. 253-60. R. B. E. C. 491-507. pp. 52-8. B. and Ito.. J. Schneiderman. J. 1980. 1999. M. Ganz. "An in Vitro Investigation of the Acetabular Labral Seal in Hip Joint Mechanics. Ellis. Brown... 171-8." J Biomech Eng. and Mieno. 14. Ateshian. and Brown.. In Press Nov 2006.. W. 114. Hale. "Depth-Dependent Compressive Properties of Normal Aged Human Femoral Head Articular Cartilage: Relationship to Fixed Charge Density. 50. C.. "Biphasic Creep and Stress Relaxation of Articular Cartilage in Compression: Theory and Experiments. W." Stapp Car Crash J. C. C. "Contact Stress Gradient Detection Limits of Pressensor Film. 73-84. R.... 2004.. Ferguson.163 Model in Lateral Impact Simulations According to Iso-Tr9790 Procedures. W. pp. E. Sphericity and Cartilage Thickness..." Rheumatology (Oxford). and Weiss.. 36.. B. T. "Improved Techniques for Measuring in Vitro the Geometry and Pressure Distribution in the Human Acetabulum--I. B. pp. 352-7." J Bone Joint Surg Am. J. Ultrasonic Measurement of Acetabular Surfaces. S." Clin Orthop Relat Res. Bryant. 2003. 1318-33. S. Shepherd. 1347-60. pp. and Seedhom. 1993. Shepherd. Maroudas. pp. T.. 1992. R. 1994. A.. "An Asymptotic Solution for the Contact of Two Biphasic Cartilage Layers. and Seedhom.. R. 124-32. A. S. 561-9. A New Computerized Technique. N. pp.. 102. Rushfeldt.. D. J. G.. Y. N.. 1981. V.

"Improved Techniques for Measuring in Vitro the Geometry and Pressure Distribution in the Human Acetabulum. P. C. S. Rushfeldt. and Swanson. 1981. R." Ann Rheum Dis. pp. 2004." J Biomech. "Direct Measurement of Local Pressures in the Cadaveric Human Hip Joint During Simulated Level Walking.. Adams. G.. A.. O. 658-66.164 [58] Chahine. H. [59] [60] . D.. 1251-61. Ii Instrumented Endoprosthesis Measurement of Articular Surface Pressure Distribution.. Mann. "Anisotropic Strain-Dependent Material Properties of Bovine Articular Cartilage in the Transitional Range from Tension to Compression.. Wang. and Ateshian. and Harris. pp. T. W. 1985. W. C. A. 44. pp. N. Hung.. 14. C. 315-23. D.." J Biomech. 37.

4. 6.6. and descending stairs [14.6 MPa. and 6. stair-climbing. Cartilage stresses were substantially elevated in the dysplastic joint. respectively whereas the normal model predicted values of 18.8] and cartilage [9.7. Mean cartilage peak pressures. Institutional review board approval was obtained to perform CT arthrography under fluoroscopic control.15].CHAPTER 6 PATIENT-SPECIFIC FINITE ELEMENT MODELING: PROOF OF CONCEPT ABSTRACT Acetabular dysplasia may be the leading cause of premature osteoarthritis (OA) of the hip [1-6]. average pressures. and shear stresses in the dysplastic joint were 28. Computed tomography (CT) image data were segmented and meshed using a previously validated protocol [11-13].10] associated with dysplasia and the resulting cartilage biomechanics are poorly understood. The objectives of this study were to demonstrate the feasibility of patient-specific finite element (FE) modeling by investigating differences in cartilage biomechanics between a normal and dysplastic hip joint and to assess the sensitivity of model predictions to changes in joint loading. the relationship between geometric alterations to the acetabular geometry [7. However. Model loading was based on in-vivo data from the literature and simulated conditions of walking. .

1 . providing the motivation for efforts to analyze a population of subjects. The results from this proof of concept study suggest that patients with dysplasia may have altered cartilage biomechanics.166 3.6. and 3. respectively.6 mm2).1143. . Surprisingly. contact areas in the dysplastic hip joint (976.8 MPa.3mm2) were greater than those predicted in the normal joint (512.3 579.

[2] estimated that 40% of patients with hip OA have some type of untreated hip dysplasia. Aronson [1] and Bombelli et al. Consequently. The discrepancies in the literature motivate the need for an improved understanding of the biomechanics of the dysplastic hip. [17. other studies have failed to find a statistically significant relationship between acetabular dysplasia and the risk of hip OA [19-24]. leading to early hip osteoarthritis.16]. In contrast to these reports. It is believed that subluxation leads to increased stresses across the articulating surface each time the hip is loaded during gait [17. Harris et al.16. It is believed that the biomechanics of the hip plays an important role in the development of osteoarthritis in the dysplastic joint [2. but few attempts have been made to quantify these differences. Subluxation (incomplete dislocation) caused by dysplasia of the hip joint is a primary cause of degenerative joint disease and clinical disability [25]. Mathematical and computational models are an attractive methodology for studying hip dysplasia since they may have the ability to non-invasively predict joint .167 INTRODUCTION Several studies have shown that mild developmental dysplasia may indeed be the leading cause of osteoarthritis in the hip [2-6.18] estimate that as many as 76% of patients with hip OA have some type of untreated acetabular dysplasia. Wilson and Poss [6] found deformity of the acetabulum in 25-35% of adult cases of hip OA.25-27]. it is thought that altered biomechanics causes cartilage and bone of the hip to degenerate prematurely.18]. Differences in joint biomechanics between normal and dysplastic hips may have important implications for predicting the development and progression of hip OA and for developing treatment strategies.

31]. bodyweight.33]. and contact area [27-29]. neglecting the issues of regional and patient-specific congruency between the femoral and acetabular cartilage layers. Most models have made simplifying assumptions regarding model geometry and have utilized modeling protocols that have not been validated. and estimated muscles forces to solve for equivalent joint reaction forces. However. This is a significant advantage since hip dysplasia is a three-dimensional pathology [30. they neglect several important aspects. Unlike analytical approaches. contact pressures. Rigid body spring models (RBSM) and the finite element (FE) method have been applied to illustrate mechanical differences between normal and dysplastic joints [32. computational models have the ability to predict joint mechanics using calculations that are based on three-dimensional geometry. Analytical studies use simplified mathematical statics calculations based on 2-D geometry. These studies used idealized geometry to represent all or part of the hip articulation. prior to patient-specific modeling it is necessary to demonstrate that the chosen computational protocol produces accurate models [34].168 mechanics on a patient-specific basis. Although analytical studies further support the notion of pathological biomechanics and in particular increased contact stresses in the dysplastic hip. The objectives of this study were to demonstrate the feasibility of patient-specific finite element (FE) modeling by investigating differences in cartilage biomechanics between a normal and dysplastic hip joint and to assess the sensitivity of model . A handful of analytical and computational studies have inferred cartilage contact pressures in dysplastic hips as a means to differentiate their mechanical environment [27-29].

.169 predictions to changes in joint loading. Patient-specific FE models were constructed and analyzed using a protocol that has been subjected to extensive validation [11-13].

NJ) and 1% Lidocaine (Hospira Inc. CT Arthrography CT arthrography was performed to enable detection of cartilage layers in the image data. acetabular angle of Sharp [25] of 47°.. Approximately 20 ml of solution was injected into each subject’s hip . IL) was injected into each hip joint capsule. Princeton. There was no evidence of visible joint narrowing in the pre-operative radiograph. Nycomed Amersham. An 18gauge needle was inserted into the hip joint capsule under fluoroscopic control.170 METHODS Subject Selection Approval to recruit and image subjects for CT arthrography was obtained from the University of Utah Institutional Review Board. The pathological hip was from a 35 y/o female patient (body weight = 149 lbs. A small amount of epinephrine (0. A solution consisting of equal parts of contrast agent (Omnipaque. Lake Forest. and pre-operative Harris Hip Score [2.25] of 57. A Tonnis grade of 1 was assigned to this patient (increased sclerosis of the head and acetabulum and slight lipping at the joint margins [35]). Informed consent was obtained from a patient with dysplasia and a normal control subject. The traditionally dysplastic patient had a center edge angle [25] of 0°.1 ml) was included to keep the fluid within the capsule.) who had been diagnosed with an excessively anteverted hip. The normal control was a 27 y/o male (BW = 147 lbs) with no history of dysplasia or hip pain. The pathologic joint (left hip) was injected for the subject with dysplasia whereas a random assignment was made for the normal control subject (left hip).

FOV = 300 mm. abducted. The data were re-sampled at 0. The scan started at the superior iliac crest on the pelvis and ended at the superior third of the femur.586 x 0. PA) scanner was used to acquire volumetric CT data of each subject’s pelvis and proximal femur in a supine position (120 kVp. ~400 slices).0 mm voxel resolution.586 x 1. Cartilage boundaries were delineated using a threshold value from a phantom-based imaging study of cartilage [12]. However. and boundary between cortical and trabecular bone were automatically and semi-automatically segmented using commercial software (Amira 4. 512 x 512 matrix. Mercury Computer Systems. Malvern.586 x 0.5 mm slice thicknesses to achieve a near isotropic voxel resolution (0. outer cortex. Chelmsford.) Pixel threshold values assigned to bone were based on our previous modeling efforts [34]. 1 mm slices. Immediately prior to scanning the joint was passively flexed.5 mm). In most regions bone and cartilage could be segmented automatically. Manual segmentation tools available in Amira were used to further define .171 capsule. MA. 0. and rotated to ensure that the articulating surfaces of cartilage were coated with contrast agent [36]. Image Data Analysis and Segmentation DICOM images were transferred to a Linux workstation for post-processing.585 x 0. requiring semi-automatic segmentation. 250 mAs.1. A Siemans Somatom 64 CT (Siemens Medical Solutions USA. which has been shown to produce more accurate reconstructions of cartilage thickness than anisotropic resolutions [12]. some regions demonstrated volumetric averaging. Separate splines representing the outer layer of cartilage. Following injection the subject was transferred to the CT scanner via a wheelchair to limit joint loading and associated fluid loss.

Although volumetric averaging was present. Position dependent shell thickness was assigned to each node. neo-Hookean hyperelastic material [40] with shear modulus G=6. the boundaries between tissue structures could be easily discerned by visual inspection. The outer cortical bone surface facets were decimated to a density consistent with our previous study [11. The final normal and dysplastic hip joint FE meshes (Figure 6. Cartilage was represented as an incompressible. Triangular surfaces were generated for each structure using the Marching Cubes algorithm [37].34. Cartilage thickness for the acetabular and femoral cartilage mesh geometries was also assessed using the cortical bone thickness algorithm [34].34].8 MPa [41]. . Incompressibility was enforced using the augmented Lagrangian method [42]. based on the distance between adjacent trabecular bone boundary nodes [34].38. Cartilage Thickness. XYZ Scientific.39].172 boundaries in these regions. Trabecular bone was neglected in this study as we previously demonstrated that it has little effect on the prediction of cartilage contact pressures and cortical bone strains [11. CA) and hexahedral element meshes were created [11]. Acetabular and femoral cartilage surfaces were imported into FE preprocessing software (TrueGrid. Material Properties The triangular surface mesh for cortical bone was converted to a quadratic 3-node shell element mesh [11.34].1) had densities consistent with our prior hip joint FE modeling studies [11. Cortical bone was represented as hypoelastic. Cartilage surfaces were decimated and smoothed slightly to remove visible triangular irregularities and segmentation artifact [11].34]. Livermore. FE Mesh Generation.

Figure 6.1. . Note lack of anterior femoral head coverage and incongruence in the dysplastic hip. Patient-specific FE meshes of normal (top) and dysplastic (bottom) hip joints.173 homogenous.29 [43]. and isotropic with elastic modulus E=17 GPa and Poisson’s ratio ν=0. Insets show details of FE discretization.

34]. and rotated relative to the pelvis so that 3D kinematic alignment was identical to that for Bergmann’s average subject. while the Zp axis is aligned upwards. Data at peak loading for each scenario (~2.15]. the Xp axis originates at the center of a vector passing through the center of the left and right femoral heads and is aligned along that vector. The Mortar method was used to tie acetabular and femoral cartilage to the acetabulum and femoral . while the nodes at the pubis and SI joints were fully constrained [11. walking. The femur was flexed.15]. the entire hip joint was flexed and abducted so that a vertically directed force in the Z-axis would result in the correct equivalent joint reaction force vector for Bergmann’s average subject [11. Reaction force data were scaled based on patient body weight.174 FE Boundary and Loading Conditions Stair-climbing. passing through the center of the L5-S1 vertebral body. The CT coordinate axes of each subject’s FE mesh were transformed into an embedded joint coordinate system that used the convention described by Bergmann [14.14. and descending stairs were simulated using average invivo instrumented prostheses gait data [14. abducted. Next.34]. The Yp axis is perpendicular to Xp and Zp and points in the ventral direction. Specifically.15]. These data describe both the anatomical orientation of the hip as well as the magnitude and direction of the equivalent joint reaction force vector throughout the gait cycle. The rigid nodes at the base of the proximal femur were constrained to move only in the direction of applied load. Nodes residing within the sacroiliac (SI) and pubis joint and those at the base of the proximal femur diaphysis were defined as rigid [11.5 times bodyweight) were used to drive the FE simulations.

2 to assess the effects of variations in joint reaction force (± 1 SD) and 2 to assess variations in joint kinematics (± 1 SD of flexion. Contact between the femoral and acetabular cartilage was enforced using the penalty method [46].33]. Contact area was also calculated using a method described in our prior work [11] since this measure may provide indirect insight into material stresses and studies have suggested that contact area is reduced in the dysplastic joint [28. Therefore. this implies that mechanical measures of stress are the critical factors defining the overload conditions.29. 15 cases were analyzed per subject.45]. One standard deviation in the kinematic and reaction force data represented nearly 50% of the baseline values. peak contact pressure. abduction.175 head. implicit FE code NIKE3D [46]. respectively [44. Data Analysis If damage due to habitual mechanical overload is the instigating factor in the onset of OA. For each loading scenario there were 5 models: 1 baseline. All analyses were performed using the nonlinear. rotation). average pressure and maximum shear stress for cartilage were determined from each FE model. All quantitative results were . Sensitivity Studies Sensitivity studies were performed for each model to assess the effects of altering joint kinematic angles and direction of equivalent joint reaction force using the standard deviations reported by Bergmann [14. Qualitative anatomical differences in the location of cartilage contact were also interpreted. yielding a total of 30 FE analyses.15].

176 reported as means and standard deviations based on combined data for the baseline model and all sensitivity models for each loading scenario. .

177 RESULTS Computational measurements of cartilage thickness demonstrated that cartilage in the dysplastic hip was slightly thicker than that of the normal hip.2 left column).2 right column).14±0.01±0.47 and 1. and 1.2. Fringe plots of femoral (left column) and acetabular (right column) cartilage thickness in the normal (top row) and dysplastic (bottom row) hip joints. Acetabular and femoral cartilage thickness was 1. Femur Posterior NORMAL Acetabulum Superior 1.75 mm Posterior DYSPLASI Medial 0.2 left column). Both the dysplastic and normal subject had thicker cartilage in the medial aspect of the femur (Figure 6.35 mm for the normal hip (Figure 6.2). .22±0.36 and 1. The normal subject had thicker cartilage in the superorlateral region of the aceteabulum whereas cartilage was thicker in the posterior aspect in the dysplastic acetabulum (Figure 6.0 mm Anterior Inferior Figure 6.36±0.31 mm for the dysplastic hip. A small region of thick cartilage was also noted along the lateral aspect of both femurs (Figure 6.

1). average pressure and maximum shear stress for the dysplastic patient were consistently higher than those for the normal subject during all three activities (Table 6. . contact areas were larger for the dysplastic hip joint (Table 6.178 Peak cartilage contact pressure.1). Surprisingly.

0) 512.5) 1084.0 (0.6) 4.3 (89.3) 19.3 top).5) 5. descending stairs) and 12 sensitivity models (4 for each loading scenario).3 (0.3 (62. . the distal anterior horn of the acetabular cartilage was consistently in contact during all three loading scenarios (Figure 6.9) 20.3 (2.0 (249.4) 3.1 (141.6) 3.1 (0.7 (0.0) Fringe plots of cartilage contact pressures elucidated the origins of altered loading in the dysplastic hip (Figure 6.9 (0.3 (99.0 (1. (MPa) (±SD) DYSPLASIA Walking Stair Climbing Descending Stairs NORMAL Walking Stair Climbing Descending Stairs 30.179 Table 6.5) 7.8 (0. Anterior wall cartilage pressures were nearly nonexistent in the normal model where most of the load was distributed along the superior aspect of the acetabulum (Figure 6.9) 29. (MPa) (±SD) 6.2) 16.3 baseline (walking.3 bottom).8 (0.5 (1.3) Max Shear (MPa) (±SD) 8.1 (0.9) 6.1.1 (0. for the dysplastic hip.7) Average Press. Average FE model results (± SD) for the dysplastic and normal hip.3 bottom).3) 3.3 (6.8 (1.1) 25.6 (2.1) Contact Area (mm2) (±SD) 1143.6 (86.6 (1. Press.42) 976.6) 3. Max.4) 579. A total of 15 models were analyzed for each subject. stair-climbing.9) 569.3 top).6 (1.5) 3. In contrast.8) 5. The pattern of contact demonstrated was more or less mono-centric over the three loading scenarios analyzed (Figure 6.7 (1.

Bottom row) normal hip. Dysplastic contact pressures were higher and the location of contact was altered when compared to the normal hip joint model.180 Walking DYSPLASIA Stair-Climbing Descending-Stairs 20 MPa NORMAL 0 MPa Figure 6. . Lateral inferior oblique view of FE predicted contact pressures for the three loading scenarios simulated. Note that the results shown are for the best case (lowest peak and average pressures) for each of the subjects.3. Top row) dysplastic hip.

1. Kinetics -1 SD Flexion. Avg. Lateral inferior oblique view of FE predicted contact pressures for 10 of the 30 sensitivity studies analyzed (walking loading scenario only).4). The location of contact did not change substantially following alteration of the joint angles and angles of the equivalent force vector (separated by columns). Rotation. Rotation. Kinetics Avg. Avg. but those analyses also demonstrated little variation as indicated by the small standard deviations in Table 6. Fringe plots of the additional 20 analyses are not shown due. Kinetics +1 SD Angles of Joint Reaction Force Avg. Abduction. Bottom row) dysplastic patient. . Kinematics Avg. Kinematics -1 SD Angles of Joint Reaction Force Avg.181 When compared to the baseline model.4. Abduction. Kinematics 20 MPa DYSPLASIA NORMAL 0 MPa Figure 6. the location of contact did not change substantially when the joint kinematics and kinetics were altered in all of the sensitivity models for each subject. Top row) normal subject. As shown in the fringe plots of cartilage contact pressure for the normal and dysplastic joints for 10 of the 30 cases (Figure 6. +1 SD Flexion. the general location and magnitude of contact did not change substantially intra-subject following alteration of both the joint angles and angles of the equivalent force vector.

which was contradictory to several studies reported in the literature [28.33]. Anteversion of the acetabulum creates poor anterior support of the femur. Model predictions suggested that cartilage stresses were elevated in the dysplastic hip joint. The posterior aspect of the acetabulum was also contacted resulting in a distinct bi-centric contact pattern which was likely due to incongruent mating between the femur and acetabulum (Figure 6. which causes the anterior wall to be loaded more than normal.1). This is the first patient-specific modeling study to our knowledge to utilize a computational protocol that had been previously validated by comparing subject-specific FE model predictions directly with experimental measurements [11-13]. The fringe plots of cartilage stress indicated that this was the case for the anteverted joint where a localized “hot spot” of cartilage stress was noted. resulting in a mono-centric pattern of contact. The dysplastic FE model predicted contact areas that were larger than the normal control. which may provide a mechanical explanation as to why persons with hip dysplasia frequently develop hip joint OA.29.182 DISCUSSION This study developed realistic three-dimensional FE models of the hip joint to compare joint biomechanics between a normal and dysplastic hip. . most of the cartilage contact occurred on the superior wall of the joint. The normal acetabulum was not excessively tilted anteriorly and was more spherical in shape when qualitatively compared to the dyplastic joint. Therefore. The predicted patterns of contact pressure and bone stress can be interpreted directly in terms of the type of acetabular dysplasia.

whereas the normal joint had substantial support perpendicular to the load.183 A possible explanation for this was that the location of contact in the dysplastic joint was predominately confined to areas that were nearly parallel to the principal direction of loading. the mean changes in contact area were not statistically significant. the results of this study suggest that the location of contact may be more important than the absolute value of contact area when assessing joint mechanics in the dysplastic hip.5 kPa/N. They developed pre and post-operative computational models to analyze changes in peak pressures and contact areas following peri-acetabular osteotomy and found that while 9 of 12 cases showed decreased peak pressure after surgery. Therefore. Several attempts have been made to quantify the biomechanics associated with dysplasia.1 kPa/N and 3. [47]. [27] used a mathematical model of static. Dysplastic hips had significantly larger peak contact stresses than healthy hips (7. Although limited by the small sample size. Mavcic et al. single-leg stance based on AP radiographs of normal and dysplastic subjects. Michaeli and co-workers combined experiments and computer modeling to predict the contact stress in the human hip joint and to investigate differences in location and magnitude of contact stresses between normal cadaveric pelvi and plastic pelvi with simulated dysplasia [29]. resulting in larger contact areas. At each surface patch . Their computational model was based on 3-D reconstructions from CT image data that was fit to spheres to represent femoral and acetabular geometry. respectively). greater compressive and shear strains were required to carry the applied load in the dysplastic joint. The surface of each sphere was discretized into 0. This conclusion augments recent data reported by Armand et al.5 mm2 patches.

Mean contact pressures for the dysplastic hips were not reported. This discrepancy is likely attributed to the assumption of spherical joints and uniform cartilage thickness.53] and were substantially less than those predicted in the current study. At 4.8. Geometric models were made from conventional anteroposterior radiographs with the assumption that the acetabular surface was spherical. were nearly 7 times less than those measured by the pressure film in the cadaveric and plastic pelvi. Nevertheless. Peak pressures for dysplastic hips were on the order of 7 MPa. Anthropometric studies have demonstrated that the hip joint is neither concentric nor has uniform cartilage thickness [54-56] and that even normal joints demonstrate irregular patterns of cartilage .5 times bodyweight the mean peak pressures for normal hips was around 3 MPa.48-52] and in vivo [15. [28] applied this computational model to analyze joint contact pressures for 70 dysplastic and 12 normal hips. Peak pressures reported in the aforementioned studies consistently underestimated those measured in vitro [7.184 the dot product between the applied load and a vector normal to the patch was calculated and the total load was divided among the patches. Hipp et al. Computational predictions of contact pressure. [32] used a three-dimensional rigid body spring model (RBSM) to compare hip joint mechanics between 112 normal and 66 dysplastic hip joints. Genda et al. as calculated by dividing the force magnitude by the surface area of each patch. They concluded that the technique was not proven to yield accurate predictions of contact stress in dysplastic joints [29]. Contact areas were 26% smaller and contact pressures were 23% higher in dysplastic hips. however the model with the largest peak pressure predicted a value of 16 MPa.

One of the reasons for this discrepancy could be that Russell et al. suggesting that the contralateral hip in DDH is also affected.89 MPa. Recently.185 contact with steep pressure gradients [7. Russell et al. both patient-specific FE models were loaded using identical kinetics . [33] generated realistic three-dimensional FE models of six dysplastic. Therefore. the primary objective of this paper was to demonstrate the feasibility of the FE method for modeling hip dysplasia. Several limitations of the current study must be mentioned. In addition. First. five asymptomatic and one normal subject using CT arthrogram data. Peak contact pressures for the symptomatic and asymptomatic hips ranged from 3. Future modeling efforts could use this patient-specific modeling protocol to investigate a more reasonable sample size. The peak pressures in the current study were substantially elevated in comparison to these results despite very similar boundary and loading conditions. which is roughly 4 times less than the modulus used in the current study.49-52]. only one patient-specific FE model was analyzed for each subject group. They reported significant differences between the normal control and the asymptomatic hips and trend towards significance between the asymptomatic hips and the symptomatic hips of patients afflicted with DDH. applied a substantial amount of smoothing to the articulating surface of femoral and acetabular cartilage. In addition. eliminating the ability to perform statistical tests. an accurate representation of the three-dimensional geometry of the hip joint may be a necessary precursor to predicting accurate biomechanics.88 MPa whereas normal hip peak pressures ranged from 1.56 – 9.8. However.75 – 1. they assumed a cartilage elastic modulus of 10 MPa.

10. it is possible that some intermediate kinematic position would place the joint in an optimal position that could result in reduced peak bone and cartilage stresses. Model accuracy would likely degrade if patients with thinner cartilage were studied. Nevertheless.51. Average cartilage thicknesses for the models analyzed in our study demonstrated that cartilage was thicker in the dysplastic joint and was above the critical threshold for obtaining accurate reconstructions for both models analyzed.48-52].60]. This will be investigated in future studies.59]. This could present a challenge while studying OA since the disease is predominately associated with a loss of cartilage [58.186 and kinematics based on average in-vivo data. Our prior work [12] has demonstrated that cartilage thickness reconstruction errors based on CT arthrogram image data increase exponentially when cartilage thickness reaches a critical value (around 1 mm). .33. and thus it is presently recommended that our modeling technique should not be applied to patients who have cartilage thicknesses that are predominantly less than 1 mm. studies have shown that cartilage in patients with dysplasia is actually thicker than normal joints during the early onset of symptoms [9.8. While inaccurate kinematics could be a source of modeling error. This in fact may be the reason why our patientspecific models predicted peak cartilage pressures that were higher than those predicted by previous models [27-29. our sensitivity studies demonstrate that the spatial distribution of contact pressures are clearly more dependent on joint geometry than variations in kinematics or loading vector direction.57] and measured in-vitro [7. Nevertheless.

In addition. it is well known that trabecular bone architecture and density are sensitive to localized stresses (i.e. density and mechanics may provide valuable insight into the biomechanics of hip dysplasia. yet it has been suggested that the labrum helps to maintain fluid pressurization in the hip [62. homogenous material. a better understanding of trabecular bone orientation. Wolff’s Law [61]). Future patient-specific modeling efforts should investigate the importance of these effects via sensitivity studies. Because we did not explicitly validate trabecular bone stresses and strains [13]. we are hesitant to predict trabecular bone mechanics on a patient-specific basis. The modeling protocol could be improved by . Although actions of individual muscles were not considered. the acetabular labrum was removed in this study. cartilage exhibits depth dependent material properties [64]. However.g. It is possible that leaving the labrum intact would have altered the predicted contact patterns and magnitudes. It is unlikely that mathematical or computational models that use simplified or idealized representations of the articular surface geometry (e. the equivalent joint reaction force was based on in vivo data [14. Our protocol allows for realistic three-dimensional models to be developed and analyzed non-invasively using patient-specific CT arthrogram image data..187 Trabecular bone was not modeled in this study since we demonstrated that it has little effect on the prediction of cartilage stresses in the hip [11].15]. Finally. variation in stiffness over its surface [65. which justifies our method of modeling the action of all muscles as a single equivalent force vector. [27-29]) provide accurate insight into differences between normal and dysplastic hips.66] and tension-compression nonlinearity [67] yet was modeled as an isotropic. Therefore.63].

the results from this proof of concept study suggest that patients with dysplasia may have altered joint biomechanics and motivates future modeling efforts to analyze a greater number of subjects. . more sophisticated cartilage constitutive equations and by incorporating the acetabular labrum. In conclusion.188 including patient-specific kinematics.

D. 1965. 650-7. 35. [2] [3] [4] [5] [6] [7] [8] [9] [10] . R. 1976. Contact Stress and Subchondral Mineralization in Human Hip Joint Specimens. "Etiology of Osteoarthritis of the Hip. Eckstein... 1997." Br J Radiol. W. T. 1986. H. 621-50... J. Rock. and Poss. 532-539. R.. pp." Osteoarthritis Cartilage. Miki. pp. 810-24. Landgraf.." J Bone Joint Surg Br. N. Anderson. C. vol... 1974. and Yoshikawa.. and Putz. Steinlechner M. G. J.. "Osteoarthritis: Diagnosis and Medical/Surgical Treatment. A. Moskowitz. and Babyn. pp. von Eisenhart-Rothe. "Quantitative Determination of Joint Incongruity and Pressure Distribution During Simulated Gait and Cartilage Thickness in the Human Hip Joint. "Osteoarthritis of the Young Adult Hip: Etiology and Treatment. Mankin.. MO. 2033. F. Harris. S.. Wilson. 279-88. D. Nishii. "Acetabular Dysplasia and Development of Osteoarthritis of the Hip. pp. J. Howell. 1999. Sato." AJR Am J Roentgenol. 195.. Ed. 1993. and Harris. pp. "Direct Comparison of Contact Areas. Stulberg." in Instructional Course Lectures. 38. 176-83. St Louis. H. 58. C." R. 12. Eds. "Sonographic Diagnosis of Developmental Dysplasia of the Hip: Importance of Increased Thickness of Acetabular Cartilage. M." Journal of Orthopaedic Research. S. Solomon. Y. H. and H.189 REFERENCES [1] Aronson. Soboleski. Philadelphia. Sugano. 2004. "Three-Dimensional Distribution of Acetabular Cartilage Thickness in Patients with Hip Dysplasia: A Fully Automated Computational Analysis of Mr Imaging. 7. Tanaka.. D. L. MO. R. W.. 1992. "Patterns of Osteoarthritis of the Hip. "The Aetiology of Primary Osteoarthritis of the Hip. 161. I." presented at Proceedings of the second open scientific meeting of the Hip Society. D. WB Saunders. pp. von Eisenhart-Rothe R.. pp.. Goldberg.. R. M." Anat Embryol (Berl).. ICV Mosby. O. H. Muller-Gerbl. 839-42. H. Muller-Gerbl M and Eckstein F. St. M. 1986. A. W. P.. Murray. pp. Louis. V." Clin Orthop.

"SubjectSpecific Finite Element Model of the Pelvis: Development. and Coggon. "Validation of Finite Element Predictions of Cartilage Contact Pressure in the Human Hip Joint. Cooper. G. A. N. A. Maas. R." J Biomech.. C. and Silman. J. "Osteoarthritis of the Hip and Acetabular Dysplasia.. C. Inoue. Berlin. pp. C." Ann Rheum Dis. R. Croft. E. "Defining Osteoarthritis of the Hip for Epidemiologic Studies. K. 2007. C. Hutchinson. Peters... 1992. M.. Graichen. 2007. Andrew.. R. C. pp.190 [11] Anderson. Croft. Ellis. pp. pp. A. R. D. J. Ellis. L.. Tuttle. J. E. 132... Strauss." J Biomech Eng. 727-9. 2006.. 69-78.. 1991.. Wickham. A New Perspective. K.. J. pp. 2005. Bombelli. Validation and Sensitivity Studies. Rohlmann. A. P. 364-73. and Weiss." Br J Rheumatol. 859-71. J. Cooper. S. Peters. C. F. 2001. P. C. Berlin. A... "Hip Contact Forces and Gait Patterns from Routine Activities. 1983. 23.. Croft. "Prevalence of Hip Osteoarthritis and Acetabular Dysplasia in Kamitonda: From a Longitudinal Population-Based [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] . Deuretzbacher. 308-10.... G. 1998. and Poss. H. "Is the Hip Involved in Generalized Osteoarthritis?. L. 512-5... 1990.." Journal of Biomechanical Engineering." Clin Orthop.. B. 1996. B. Croft. A. J. R. J. Shichikawa. 31. 514-22. Bombelli. Heller. 88. 127. D. 325-8.. E.. Anderson.. F.. pp. D. Feb 2007.. D. Free University and Humboldt University." Am J Epidemiol.. and Coggon." Radiology. L. Bergmann. G.. Ali-Gombe. and Ota.. Cooper.." J Bone Joint Surg Br.. 34. Bergmann. and Weiss.. 50.. P. P. Santore. In Press. Wickham. and Allen. "Factors Influencing Cartilage Thickness Measurements with Multi-Detector Ct: A Phantom Study. pp. A.. B." J Rheumatol. pp.. G. P... J.... Germany. C. Springer-Verlag. "Osteoarthritis of the Hip and Acetabular Dysplasia in Nigerian Men. C. "Mechanics of the Normal and Osteoarthritic Hip. W. Osteoarthritis of the Hip. C. E. A. 1984. and Coggon. D. Peters.. and Weiss. Kim.. and Duda.. Y. Anderson. G. A. "The Relationship between Acetabular Retroversion and Osteoarthritis of the Hip. Wickham. A. 1999... Hip98: Data Collection of Hip Joint Loading on Cd-Rom.

S. B. A. M.. 1997. Malvitz. 745-8. 38.. and Pedersen. D. "Closed Reduction for Congenital Dysplasia of the Hip.. Wallensten. A. pp. 1025-1030. Siebenrock.. M. "Cartilage Contact Pressure Elevations in Dysplastic Hips: A Chronic Overload Model. "A Computer Simulation Study of Normal and Abnormal Hip Joint Contact Pressure.. R. 2000. 20. P. Validation and Sensitivity Studies in Computational Biomechanics.. B. 2006. "Mathematical Estimation of Stress Distribution in Normal and Dysplastic Hips." Rheumatology (Oxford). S. Shivanna. and Kralj-Iglic. J.... Hukuda. S. A. Michaeli. J." Rheumatology (Oxford). Daniel. Hasegawa. [24] Inoue. 134-43. J. B. Sugano. 1999. R. 1. 2006. Pompe. 81.. V. E. A Cause of Hip Pain. Grosland. "Prevalence of Hip Osteoarthritis and Acetabular Dysplasia in French and Japanese Adults. pp... pp. 2001. S." J Orthop Surg.. Russell. D. "Verification. pp.... pp. and Courpied. Antolic. pp. Murphy.. J. Leunig. Reynolds. 239-45... pp." Clin Orthop. 2002. 202-6.. K. Y. E. N. R. pp. 39. A.. D." J Bone Joint Surg Br. M... Cooperman.... 19. 1983." Arch Orthop Trauma Surg. B. and Klaue.. J..191 Epidemiological Study of Rheumatic Diseases in Japan. A. B. Mavcic. 1995.." Computer Methods in Biomechanics and Biomedical Engineering. "Retroversion of the Acetabulum. M. S.. D. 76. "Planning Acetabular Redirection Osteotomies Based on Joint Contact Pressures. Wicart." Instr Course Lect. and Hipp.. A. Genda.. "Acetabular Dysplasia in the Adult.. B.. and Miura. Kawasaki. In Press Dec 2006. M. A. 1994. and Ganz. and Stulberg. K. and Weinstein. R. pp. J." Journal of Orthopaedic Research. 114. Anderson. pp. V. Ushiyama. L. Konishi. D.. 1777-92. J. 180-6. K.. T. "Comparison of Predicted and Measured Contact Pressures in Normal and Dysplastic Hips. and Murphy." J Bone Joint Surg Am. Lucas." Clin Orthop. 50.." Med Eng Phys. T. H. pp. 281-8. 1999. T... Huang. 793-4. K. A. and Weiss. T. [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] . Functional and Radiographic Results after an Average of Thirty Years. N. "Periacetabular Osteotomy: The Bernese Experience. Ellis. 6. Hipp.. Millis. 79-85. Iglic. N..

S. R. Hughes. 185. 315-36. K.. N. "Fat-Suppressed 3d Spoiled Gradient-Echo Mri and Mdct Arthrography of Articular Cartilage in Patients with Hip Dysplasia. "A 3d Mortar Method for Solid Mechanics. pp. Hung. "Generalization of Selective Integration Procedures to Anisotopic and Nonlinear Media.. Buchler. and van Erning. 59. 2004. Park.. J.. Hestenes. R. and Liu. H. pp.." Clin Biomech (Bristol. 1987. M. and Yoshikawa.... and Laursen." Osteoarthritis Cartilage... M. 2002. C. K. Nakanishi. Miki. J. 15. pp. "Marching Cubes: A High Resolution 3d Surface Construction Algorithm. pp. A.. "Mechanical Response of Bovine Articular Cartilage under Dynamic Unconfined Compression Loading at Physiological Stress Levels.. P. M. A. 21. 630-9. 1995. 117. D. J. 2004. and Ateshian. 167-181. pp. H. "Nonliner Finite Element Analysis of Shells: Part I. Berlin. 1969. R. Puso.. "Multiplier and Gradient Methods. Puso. L. 12. Graph.. 379-85. 65-73.. Lorensen. Nishii. T. E." Journal of Optimization Theory and Applications. "A Finite Element Model of the Shoulder: Application to the Comparison of Normal and Osteoarthritic Joints. L.. Dalstra.. 9. N." Compuational Methods in Applied Mechanics. Hughes.... H. Avon). 303-320. T. 4. pp. 193. 17. E. Ramaniraka. pp. "A Mortar Segment-to-Segment Contact Method for Large Deformation Solid Mechanics. Tanaka. T. M.. "Development and Validation of a Three-Dimensional Finite Element Model of the Pelvic Bone. Congenital Dysplasia and Dislocation of the Hip in Children and Adults..." International Journal for Numerical Methods in Engineering.192 [35] Tönnis." Computer Methods in Applied Mechanics and Engineering. and Farron.. 2005. Iannotti. W. P. 601-29. 27. Two Dimensional Shells.. W. [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] . and Cline.. A. 163-169. Sugano. 1981. T." Comput. pp. A. pp. H. Rakotomanana." AJR Am J Roentgenol. T. G. A. 1980. Huiskes. 272-8." Interational Journal for Numerical Methods in Engineering. A. pp. 2004. Springer-Verlag. 1987." J Biomech Eng.

pp." Ann Rheum Dis. K. W. 658-66. and Shaw. Brown. R. D. W. W. "Contact Pressures in the Human Hip Joint.... Brown. D." Proc Natl Acad Sci U S A.. 58.. T. "Outcome of Periacetabular Osteotomy: Joint Contact Pressure Calculation Using Standing Ap Radiographs.. N.. Macirowski. 1994. "Contact Pressures in the Human Hip Joint Measured in Vivo. 315-23. 12 Patients Followed for Average 2 Years. "Incongruent Surfaces in the Human Hip Joint.O." J Bone Joint Surg Br.. 1290. P. D. 303-13. 1981. 373-84. Goodfellow. R. 1981. 1984. Adams. and Harris. Mann. J.. UCRL-MA. D. T.. Rushfeldt. pp.. Shepherd. S." Acta Orthop. and Mann. L... 2879-83. Ferencz.. "In Vitro Contact Stress Distributions in the Natural Human Hip. J. "A Contact-Coupled Finite Element Analysis of the Natural Adult Hip. Bullough. Greenwald. M. and Hutton. D. Fijan. and DiGioia. E. W. 76. and J.. T. Y. 1983.. W. S. 44." Lawrence Livermore National Laboratory Technical Report.. pp. 1986. and Seedhom.. 1990. 69.." J Biomech.. Sphericity and Cartilage Thickness. A. "Direct Measurement of Local Pressures in the Cadaveric Human Hip Joint During Simulated Level Walking. 14. 17. Hodge. D.... 1999. 1987." Nature. pp. Lepisto. R. M. P.. J. W. H. G. "Nike3d: A Nonlinear. and Harris.. Afoke. pp. Rushfeldt.. Three-Dimensional Finite Element Code for Solid and Structural Mechanics. S. 253-60. pp. 14. 116. D. H. A. and Swanson. R.. 16. H.. Tallroth. Implicit. P. S. "Cartilage Stresses in the Human Hip Joint. A. 2005. D. E. pp.. R. pp. B. P. pp. Carlson. Ultrasonic Measurement of Acetabular Surfaces.193 [46] Maker. Burgess." J Biomech.. Armand. A. B. R. H. Ii Instrumented Endoprosthesis Measurement of Articular Surface Pressure Distribution. [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] . K. 27-34." J Biomech. Byers. W. 1968. W. 3rd. 217. C. R. "Improved Techniques for Measuring in Vitro the Geometry and Pressure Distribution in the Human Acetabulum--I. and O'Connor. "Improved Techniques for Measuring in Vitro the Geometry and Pressure Distribution in the Human Acetabulum.. 437-48. 536-41. Harris." J Biomech. J. Mann. T. and Mann." J Biomech Eng. pp. 1985. B. 83." Ann Rheum Dis. Tepic. M. Elias.. and Chao. pp. W.. 10-8... N. "Thickness of Human Articular Cartilage in Joints of the Lower Limb.

pp. D.. D. E. F. "An in Vitro Investigation of the Acetabular Labral Seal in Hip Joint Mechanics. and Yoshikawa. Bryant.. Schneiderman. D. Nishii. Shepherd.. J. pp. 2000. "Mechanical Determinants of Osteoarthrosis. E.. T. B. 2003. Fyhrie. pp... 1999.. and Seedhom.. 1991. 21. B. Falcovitz.. 12-21. 183-90. B. D. 266. T. R. K. C. "The 'Instantaneous' Compressive Modulus of Human Articular Cartilage in Joints of the Lower Limb. G. S. 2004. pp. Sugano. pp. 6-9. Burr." Rheumatology (Oxford)." J Biomech.. C. Wolff. K. D. and Sah. C." Semin Arthritis Rheum. 36. Bryant. 561-9. Y." Osteoarthritis Cartilage.." J Biomech. "Articular Cartilage Abnormalities in Dysplastic Hips without Joint Space Narrowing. Hung." J Biomech. Das Gesetz Der Transformation Der Knochen. T. Tanaka. S. and Boyd. and Ateshian. Chen. L. and Ito. "Comparative Study of the Intrinsic Mechanical Properties of the Human Acetabular and Femoral Head Cartilage. K. A.. R. S.. J... 2001." J Orthop Res. 9.. L.. Ganz.. J. A. Chahine. 171-8. "Depth-Dependent Compressive Properties of Normal Aged Human Femoral Head Articular Cartilage: Relationship to Fixed Charge Density. J. J.... T. pp. 340-9. 953-60. N. "Osteoarthrosis--the Orthopedic Surgeon's Perspective.194 [58] Radin. A.. S.. 38. pp. Berlin. T. Maroudas. pp. K.." Clin Orthop Relat Res. K. Athanasiou. 1892. and Ito.. B. 33. Nakanishi. Ferguson.. Ohzono. Hirschwald." Acta Orthop Scand Suppl. 1995.. Radin.. R.. 12. 37. Ferguson... and Dzida. Ganz. Brown. 1994.. Caterson. O. Agarwal. R. H. "Anisotropic Strain-Dependent Material Properties of Bovine Articular Cartilage in the Transitional Range from Tension to Compression. N. Wang. E. H. J. L. A. 2001.. R. 124-32. "The Influence of the Acetabular Labrum on Hip Joint Cartilage Consolidation: A Poroelastic Finite Element Model. 1251-61.. H. pp.. [59] [60] [61] [62] [63] [64] [65] [66] [67] .

. and boundary conditions. motivating further study of this patient population. Experimental hip joint tests were conducted to measure pelvic cortical bone strains and cartilage contact stresses in separate studies. The ability of computed tomography (CT) to accurately reconstruct cortical bone and cartilage thickness was also determined using phantoms. The results of this proof of concept study indicated altered biomechanics in the dysplastic hip. geometry. Model sensitivity studies were performed to analyze the effects of altering assumed and measured inputs including material properties. Finally. Further. the feasibility of patient-specific FE modeling was demonstrated by developing and analyzing FE models of a normal and dyplastic hip joint. Detailed computational protocols were developed and validated by comparing finite element (FE) predictions directly with bone strains and cartilage contact stresses measured experimentally. the results of this dissertation may be applied to the study of other joints such as the knee and shoulder.CHAPTER 7 DISCUSSION SUMMARY The research described in this dissertation investigated the mechanics of the human hip joint and developed novel experimental and computational tools that will facilitate future studies of the hip.

assuming that subjects have cortical bone with sufficient thickness. Although previous models of the pelvis have been described [3. The results from the phantom imaging study .196 The results of the bone modeling study demonstrated the accuracy of the FE method for quantifying bone strains and has direct relevance to the study of hip dysplasia and OA as it is believed that bone mechanics play an important role in the progression of these pathologies [1. Prior to developing subject and patient-specific joint contact stress models it was important to demonstrate that cartilage thickness could be accurately reconstructed from CT image data. direct validation between computational predictions and experimental measurements was not performed. The pelvis analyzed in this study rarely had thickness below this critical threshold. the cortical bone thickness algorithm developed herein will be useful for constructing accurate FE models in the future.7 mm thick to obtain accurate geometrical reconstructions. Finally. The results of the parameter studies demonstrated that predictions of cortical bone strains were most sensitive to changes in cortical bone thickness. Thus.4]. the model detailed in this study may have the ability to serve as a template for studying general biomechanics of the pelvis or to investigate changes in bone stresses and strains due to surgical procedures such as peri-acetabular osteotomy (PAO) or total hip arthroplasty (THA).2]. Therefore. The accuracy of CT reconstructions of cortical bone was also quantified in this study and it was determined that cortical bone should be at least 0. our approach for modeling the cortex as a layer of thin shells has direct applicability to the analysis of patientspecific models.

many studies have made the assumption of rigid bones [12. the FE model demonstrated good qualitative and quantitative correspondence between computational predictions and experimental data. but the results of this study demonstrate that the direction of the error during CT arthrography is dependent on several parameters such as contrast agent concentration. Overall.13] yet the results of this study demonstrated that this model simplification may lead to erroneous estimates of contact stress and area. Additionally. It is generally assumed that CT underestimates the thickness of thin tissues due to volumetric averaging [5. In addition. most models have predicted pressures that are substantially less than those measured experimentally.6]. It is hoped that the results of this study will circulate to a larger audience of clinicians and scientists who may be unaware of the consequences of using higher concentrations. Modeling errors decreased when larger bounds of . and spatial resolution.197 suggested that cartilage should be at least 1. scanner direction.0 mm thick to ensure accurate (<10% error) reconstructions. it was shown that cadaveric hips could be imaged without contrast agent as errors in the “dry” scans were nearly identical to those when using a low concentration of contrast agent. we have found that the choice of contrast agent concentration generally depends on the particular radiologist who is performing the injection. Therefore. Prior hip joint contact models assumed spherical geometry and concentric articulation between opposing layers of cartilage [7-11]. From our recent experience working alongside radiologists. The FE modeling study of cartilage contact was the first to validate predictions directly with experimental pressures measurements.

Because the geometric pathologies associated with hip dysplasia are three-dimensional [16-19].198 pressure were compared. It will be important to analyze several additional patient-specific FE models to learn more about the pathological mechanics of hip dysplasia.14. Changing the direction of loading and the anatomical orientation using standard deviations reported in the literature did not alter the pattern of cartilage contact substantially.14. Although mathematical and computational models have been developed to study hip dysplasia [8. Hip dysplasia may be the primary etiology of OA in the hip joint [2.9. . Recent three-dimensional models [22] represent significant improvements over 2-D techniques however this particular protocol was not validated. and thus the protocol discussed herein may be more suitable for generating models that can predict highly stressed regions of cartilage as opposed to the entire pattern of contact.20-22]. Therefore. The patient-specific models described in this dissertation demonstrated differences in joint biomechanics between normal and dysplastic hips. it was important to develop techniques to analyze joint mechanics without simplifying assumptions regarding bone and cartilage geometry. the patient-specific models presented in Chapter 6 represent one of the first three-dimensional modeling efforts. it is likely that the topology of the joint surfaces primarily drives the patterns of hip joint contact rather than the underlying boundary and loading conditions.15]. Nevertheless it is possible that some intermediate kinematic position would provide a “path of least resistance” that would likely reduce pressures from the values that were predicted in this study.

Although it has been argued that absolute model validation is impossible [24. If a similar combined experimental and computational approach is adopted by analysts who wish to study other joints such as the knee and shoulder. Model credibility must be established before clinicians and scientists can be expected to extrapolate information and decisions based on model predictions [23]. peer acceptance will ultimately be improved. After review of the joint modeling literature. .25] it is hoped that the protocol developed herein has been subject to enough scrutiny that it will be accepted as an appropriate methodology for generating patient-specific hip joint models. it is clear that substantial effort has been directed towared model “development” yet the notion of model “validation” has been largely neglected.199 The primary purpose of this dissertation was to develop and validate protocols to generate patient-specific models of the hip joint.

. recent commercial segmentation programs have made the process of data segmentation nearly automated. The surface used to create the pelvic mesh in Chapter 3 took nearly one month to manually segment from the CT image data. it would be beneficial to predict trabecular bone stresses and strains in both normal and pathological hips since this could provide insight into why diseases such as OA develop.200 LIMITATIONS AND FUTURE WORK One of the primary critiques of using the FE method for studying joint biomechanics is that it is very time consuming to segment surface geometry from medical image data [9].1. predictions of cortical bone stresses corresponded well with experimental measures despite changes in the trabecular bone modulus and Poisson’s ratio. Is it believed that the architecture of trabecular bone is primarily dictated by the stress history [26]. However. Therefore. By using a commercial program (Amira 4. It was not possible to experimentally measure trabecular bone strains in the pelvic FE modeling study (Chapter 3). caution should be exercised when interpreting patientspecific trabecular bone mechanics using models formulated from the protocol discussed in Chapter 3. potentially limiting the amount of subjects that could be analyzed in a given study. Therefore. where) we were able to reduce the time to segment the pelvic surface from one month to less than three hours. Therefore. Nevertheless. this modeling approach for the analysis of patient-specific cortical bone mechanics does not require one to stipulate a preferred material direction for trabecular bone.

However. Therefore it is possible that the labrum may also serve some (albeit limited) load bearing function in the intact joint. To model the labrum it would first be necessary to perform material testing to ascertain its material properties as these have not been reported in the literature. To obtain FE model convergence it is often necessary to limit the degrees of freedom of the model. Therefore. it is likely that these models overestimated cartilage contact pressures in both the normal and dysplastic joint. This assumption is appropriate when the same boundary conditions have been applied experimentally. . only extreme positions were analyzed. During loading of both normal and pathological joints it is likely that the hip joint follows a path of least resistance in an effort to minimize energy [28].201 The labrum was removed in the cartilage contact stress study (Chapter 5) and was not segmented and meshed in the patient specific FE study (Chapter 6). The locations of contact stresses in the subject-specific model were primarily on the lateral aspect of the acetabulum during each loading scenario studied. thus eliminating rigid body modes. Given its preferred circumferential alignment of collagen is it quite possible that the transversely isotropic model described by Quapp and Weiss [27] would work well and this will be investigated in future experimental and computational studies. this could result in substantial elevations in patientspecific models since the exact boundary conditions are not known a-priori. Although the parameteric studies detailed in Chapter 6 indicated that cartilage contact pressures were not sensitive to changes in the kinetics and kinematics.

First. This ongoing work will more thoroughly characterize the biomechanics of hip dysplasia by developing additional patient-specific FE models. an algorithm could be developed to incrementally rotate and translate the femur relative to the acetabulum. . prior to model loading. Several subjectspecific FE hips will also be tested experimentally and subsequent FE models will be validated using methodologies very similar to those described in this dissertation. Dynamic measurements of cartilage contact pressures will be More realistic constitutive recording using a novel tactile pressure sensor. This technique would require an analyst to assume it is the differences in joint congruence rather than deformation that dictates the preferred kinematics of the hip joint. Then the model could be constrained to move only in the direction of applied load. an estimate of the joint reaction force could be based on a linear scaling of the force plate data. Based largely on the work described in this dissertation. The kinematic position at any point in the gait cycle could be quantified using surface markers and standard motion capture technology. which could also be used in conjunction with the former. our laboratory has recently obtained federal funding to continue the study of hip dysplasia. The peak reaction force at the hip measured in-vivo corresponds very well with peaks of force plate readings [29. Therefore. equations will be used to model the tension-compression behavior of cartilage.30]. A second approach. using the basis of minimization of energy.202 Several possible approaches could be adopted to resolve the possible need for patient-specific kinematics. is related to patient-specific gait analysis. until the distance between opposing layers of cartilage was minimized.

31]. Many orthopaedic surgeons are unaware of multiple facets of the dysplasia diagnosis and their potential implications for joint degeneration. It is hoped that this study will elucidate the mechanics of dysplasia and result in better diagnosis and treatment of young adults who are afflicted with this disease.203 Finally. Patients are roughly divided into those who manifest instability due to acetabular undercoverage (usually classic dysplasia) and those with acetabular overcoverage (retroversion [16-18]) or impingement [19. . In reality there is likely a spectrum of abnormal hip morphology in terms of severity and location of stress transmission [16-19]. which could fundamentally change the way that acetabular dysplasia is diagnosed and treated. Recognizing the mechanical consequences of different forms of dysplasia allows earlier identification of “at risk” hips so that earlier treatment can be initiated. The long-term goals of this research are to improve the diagnosis and treatment of acetabular dysplasia and to extend the applicability of hip joint FE modeling. Using patient-specific modeling techniques could delineate the true spectrum of this disease process by quantifying the degree of stress transfer from the acetabulum to the femoral head and primary location of contact. Treatment decisions for patients with acetabular dysplasia currently follow a simplified procedure with surgical decision making primarily based on empiric recommendations and past outcomes. hopefully delaying the need for THA. patient-specific models will be developed for two different manifestations of dysplasia (i.e. Improved diagnostic ability translates into more appropriate surgical treatment [31]. “traditional” dysplasia and acetabular retroversion).

Patient models could potentially be applied to quantify changes in mechanical loading due to surgical intervention. these techniques could be utilized in longer-term prospective studies in an effort to correlate surgical correction with changes in mechanical loading and outcome. . to understand how geometric and mechanical abnormalities affect cartilage mechanics.204 Patient-specific FE models of the hip joint have a number of potential longer-term uses and benefits. Hip joint FE modeling may also assist the development of other treatment methods. patient-specific approaches to treatment. including cartilage thickness and underlying bone geometry. autologous cartilage cell tissue engineering could offer the same type of treatment potential. and prediction of the long-term success rate of corrective surgeries based on pre. Allograft techniques require mapping of the articular cartilage anatomy. As tissue engineering techniques continue to improve. including improved diagnosis of pathology. Currently.and post-operative mechanics. such as osteochondral autografting of defective cartilage in patients with dysplasia. success is measured by avoidance of a total hip arthroplasty and is not correlated with any preoperative variable other than the relatively crude radiographic measurements [31]. In addition. allowing one to assess the efficacy of different approaches to osteotomy.

. "A Computer Simulation Study of Normal and Abnormal Hip Joint Contact Pressure. Y." Arch Orthop Trauma Surg. 1995. Prevrhal." Med Phys. E. pp. J. A. "Etiology of Osteoarthritis of the Hip. Hasegawa. Konishi.. M. 114. Rapperport. A. H. Germany. E." Clin Orthop. pp. Berlin." J Biomech. 16.205 REFERENCES [1] Bombelli. MacWilliams. pp.. T. Genda. Iwasaki. J. H. S.. 895-905. "Contact Finite Element Stress Analysis of the Hip Joint. A. 20-33. H. Engelke." J Biomech Eng. 1983. and Kalender. Oonishi... pp. D. Yoshida. 1984." J Biomech... 44. 3rd. pp. L. 17. Fox.. "Normal Hip Joint Contact Pressure Distribution in Single-Leg Standing--Effect of Gender and Anatomic Parameters. and DiGioia. J. R.. and Schurman. 1-8. and Hasegawa. A. D. Springer-Verlag... Y. H. M. and van Erning. 751-64. S. Carter. "Accuracy of Ct-Based Thickness Measurement of Thin Structures: Modeling of Limited Spatial Resolution in All Three Dimensions. B. Brown. J. G. J. N. E. 34.. Osteoarthritis of the Hip. K. 2003. and Genant. 1986.. T. D.. Barrance. R. K. 2006. pp.. pp. Isha. "Accuracy Limits for the Determination of Cortical Width and Density: The Influence of Object Size and Ct Imaging Parameters. 3. 427-44. Wilckens. A.. Shepherd... "A Contact-Coupled Finite Element Analysis of the Natural Adult Hip. "Mechanical Analysis of the Human Pelvis and Its Application to the Artificial Hip Joint--by Means of the Three Dimensional Finite Element Method.. Li.. 437-48.. 1985. Y. M. Genda. W. 30. T. and Miura." Phys Med Biol. Dalstra. 43546..." J Orthop Res. Huiskes. Fetto. Harris. "Three-Dimensional Dynamic Hip Contact Area and Pressure [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] . H. R. Kitagawa. pp. E. W. Faust." J Biomech. 117.. "Development and Validation of a Three-Dimensional Finite Element Model of the Pelvic Bone. 202-6. and Chao. pp... 1999. 272-8... 1995. J. and Chao. 1983.... 2001... C. P.. J. Prevrhal. N. D.

.. N. M. [12] Ferguson. Hutchinson. 81. 6. Pompe.. Ellis. M. A. and Ito.. Macirowski. R. N. Treatment with Periacetabular Osteotomy.. 134-43.. "The Influence of the Acetabular Labrum on Hip Joint Cartilage Consolidation: A Poroelastic Finite Element Model." Med Eng Phys.. S. 116. Ganz. 1999. 33. A.. 2006. Siebenrock. J. "Comparison of Predicted and Measured Contact Pressures in Normal and Dysplastic Hips.. "Retroversion of the Acetabulum. Shivanna. 2006. Lucas. C." J Bone Joint Surg Br. R. S.206 Distribution During Activities of Daily Living. A. M. 10-8. Y.." J Biomech. Bircher. V.. "Verification. R. 20. "Anterior FemoroAcetabular Impingement Due to Acetabular Retroversion... B. and Hipp. 2003. 1. D. pp. 727-9. B. "Cartilage Stresses in the Human Hip Joint. J. Daniel. J. Bryant." J Biomech Eng. 1025-1030. D. B. 1976.. Validation and Sensitivity Studies in Computational Biomechanics. and Kralj-Iglic. and Klaue. 1997. "Planning Acetabular Redirection Osteotomies Based on Joint Contact Pressures. pp.. 58.." J Bone Joint Surg Am. P.. K. K. H." J Bone Joint Surg Br. S. B. W." J Biomech. R. A. pp. "Mathematical Estimation of Stress Distribution in Normal and Dysplastic Hips. E. J.. Schoeniger.. K. pp.. 1999. 278-86. and Allen.. Millis. Michaeli. pp. 743-4. 85-A. J. B. "Patterns of Osteoarthritis of the Hip. pp. D. 281-8. Sugano.. Andrew. pp. Murphy.. T.. 39. "Retroversion of the Acetabulum. D. 2002. Antolic. S." J Orthop Surg. R. 19962004. A. J." Clin Orthop. pp." Computer [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] . M. and Weiss. Tepic.. T. L.. Solomon. B. 88. A. and Murphy. 176-83. 1999. V. 81. pp. and Ganz. J. and Pedersen. M. "The Relationship between Acetabular Retroversion and Osteoarthritis of the Hip. D. E. pp. G. Iglic.. Russell. A Cause of Hip Pain.. pp.. 19. Anderson." J Bone Joint Surg Br. "Cartilage Contact Pressure Elevations in Dysplastic Hips: A Chronic Overload Model.. 1994.... Hipp.." J Bone Joint Surg Br. Grosland.. 2006." Journal of Orthopaedic Research. J. Mavcic. 953-60.. W. K. 180-6. 2000.. pp... A. Reynolds. and Mann. Kim.

"The Etiology and Treatment of Hip Pain in the Young Adult. Hip98: Data Collection of Hip Joint Loading on Cd-Rom.." Science. K... Sahrmann. A. M. 641-646. Validation. K. Quapp.. "Material Characterization of Human Medial Collateral Ligament.. pp. Graichen. L. pp. 757-63." Phys Ther. A. J. "Verification. G. 88 Suppl 4. Heller.. K. Bergmann. 2006. 1998. Free University and Humboldt University. Hirschwald. 263. 34. Bergmann. pp. and Erickson. The Logic of Scientific Discovery. 120. and Confirmation of Numerical Models in the Earth Sciences. Peters.207 Methods in Biomechanics and Biomedical Engineering. and Weiss. G. Wolff. [25] [26] [27] [28] [29] [30] [31] . "The Twenty-Ninth Mary Mcmillan Lecture: Moving Precisely? Or Taking the Path of Least Resistance?. [24] Oreskes. London. pp. Berlin.. N. J. pp. M. 20-6. G. J. Deuretzbacher. 1998.. Popper. Rohlmann. F. 1998. A... and Duda." J Bone Joint Surg Am. Shrader-Frechette. N.. J.. "Hip Contact Forces and Gait Patterns from Routine Activities. 1892. Strauss. S." J Biomech Eng. 1992.. 1994. 1208-18. 2001. K.. and Belitz. Routledge (originally published 1959 by Hutchinson Education). Berlin. A. Das Gesetz Der Transformation Der Knochen. 859-71. C.. G. In Press Dec 2006." J Biomech. 78..