UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Open MRI investigation of contact mechanics in cam femoroacetabular impingement Buchan, Lawrence Lewis 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2015_may_buchan_lawrence.pdf [ 3.84MB ]
Metadata
JSON: 24-1.0215937.json
JSON-LD: 24-1.0215937-ld.json
RDF/XML (Pretty): 24-1.0215937-rdf.xml
RDF/JSON: 24-1.0215937-rdf.json
Turtle: 24-1.0215937-turtle.txt
N-Triples: 24-1.0215937-rdf-ntriples.txt
Original Record: 24-1.0215937-source.json
Full Text
24-1.0215937-fulltext.txt
Citation
24-1.0215937.ris

Full Text

    Open MRI Investigation of Contact Mechanics in  Cam Femoroacetabular Impingement  by Lawrence Lewis Buchan B.A.Sc., The University of Toronto, 2012   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in The Faculty of Graduate and Postdoctoral Studies (Biomedical Engineering)   THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) March 2015  © Lawrence Lewis Buchan, 2015  ii  Abstract Cam femoroacetabular impingement (FAI) is a mechanical process thought to cause of hip osteoarthritis (OA). In cam FAI, it is thought that a ‘cam deformity’ on the femoral head-neck junction intrudes into the intra-articular joint space, inducing elevated mechanical force on acetabular cartilage. However, few experimental studies have measured contact mechanics in FAI. Open MRI in functional positions has potential to directly and non-invasively assess cam FAI, but MRI measures have not been related to mechanics. This thesis asked, in cadaver hips positioned in a simulated anterior impingement posture: (1) Does open MRI show intrusion of a cam deformity into the intra-articular joint space? (2) Is a cam deformity associated with elevated acetabular contact force? (3) Are MRI measures of cam FAI related to acetabular contact force? Cadaver hips (cam, n=9; controls, n=3) were positioned in a simulated anterior impingement posture, then imaged using open MRI with multi-planar reformatting. The β-angle was measured at 72 locations about the circumference of the femoral neck, and a binary ‘MRI cam-intrusion sign’ was defined (positive if βmin<0°). Hips were then instrumented with a piezoresistive sensor before conducting six repeated impingement trials, measuring acetabular contact force (F), centroid location, and distribution. A binary ‘contact-force sign’ (positive when F>20N) defined elevated contact force. Minimum β-angle ranged from 1.4° to -28.5° in cams versus 4.6° to -0.2° in controls. Cam hips were significantly more likely than controls to have a positive MRI cam-intrusion sign (p=0.0182, Fisher’s exact test) and positive contact-force sign (p=0.0083). There was a significant relationship between the MRI cam-intrusion sign and contact-force sign (p=0.033). This thesis established that open MRI measures of cam FAI relate to contact mechanics, indicating that open MRI has significant potential to investigate the biomechanics of cam impingement. Open MRI can be used to establish treatment guidelines and understand why some hips develop OA and some do not.    iii  Preface  A version of Section 2.3: Biomechanics of Cam FAI was included in a book chapter submitted for publication [d’Entremont AD, Buchan LL, Wilson DR. Submitted. “Biomechanical Considerations in Arthritis of the Hip”]. I developed the image acquisition protocol in Chapter 3 in collaboration with Andrew Yung and Dr Honglin Zhang. Dr Ingrid Heaslip provided clinical insight that defined the optimal MRI sequence. I designed and constructed the positioning rig in Chapter 4. Mike Brenneman and Dr Jon Nakane provided key design feedback and assistance with fabrication. A version of Chapter 6 is being prepared for publication. Dr David Wilson was responsible for the study concept and significant editing. Dr Honglin Zhang contributed important feedback regarding the study concept, magnetic resonance imaging, and interpretation of β-angle findings. Dr Sujith Konan helped classify the specimens and interpret qualitative imaging findings. Dr Ingrid Heaslip provided key clinical insight in developing the imaging protocol. Dr Charles Ratzlaff helped frame the research questions and interpret results. I was responsible for study design, methodology (Chapters 3-5), conducting all experiments, analyzing data, presenting data, and writing the manuscript. The specimen screening protocol in Appendix A was conducted in collaboration with Dr Sujith Konan. The computer code in Appendix C was adapted from various scripts written by members of the Wilson lab (Jonathan Doucette, Christopher Bhatla, Honglin Zhang, Laura Given, Marianne Black). All cadaver experiments were covered by UBC Clinical Research Ethics Board certificate number H08-01931.    iv  Table of Contents Contents Abstract ........................................................................................................................................................ ii Preface ......................................................................................................................................................... iii Table of Contents ....................................................................................................................................... iv List of Tables ............................................................................................................................................. vii List of Figures ............................................................................................................................................. ix List of Acronyms ........................................................................................................................................ xi Acknowledgements ................................................................................................................................... xii Dedication ................................................................................................................................................. xiii 1 Introduction ......................................................................................................................................... 1 2 Literature Review ............................................................................................................................... 4 2.1 Hip Joint ........................................................................................................................................ 4 2.1.1 Anatomy .............................................................................................................................................. 4 2.1.2 Osteoarthritis ....................................................................................................................................... 6 2.2 Cam-type Femoroacetabular Impingement ................................................................................... 7 2.2.1 Clinical Presentation ............................................................................................................................ 7 2.2.2 Epidemiology & Etiology.................................................................................................................... 9 2.2.3 Quantification .................................................................................................................................... 10 2.2.4 Causal Relationship between Cam FAI and OA ............................................................................... 13 2.2.5 Prospective Studies ............................................................................................................................ 14 2.2.6 Treatment .......................................................................................................................................... 15 2.3 Biomechanics of Cam FAI .......................................................................................................... 17 2.3.1 Range of Motion ................................................................................................................................ 17 2.3.2 Impingement Location....................................................................................................................... 18 2.3.3 Pelvic Kinematics .............................................................................................................................. 21 2.3.4 Labral Seal & Joint Fluid Pressure .................................................................................................... 22 2.3.5 Joint Contact Pressure and Tissue Stress ........................................................................................... 22 2.3.6 Subchondral Bone Density ................................................................................................................ 23 2.3.7 Considerations in Pincer FAI ............................................................................................................ 23 2.4 Measuring Hip Contact Mechanics in Cam FAI ......................................................................... 24  v  2.4.1 Direct Imaging in Functional Positions ............................................................................................. 24 2.4.2 Direct Mechanical Measurements ..................................................................................................... 25 2.5 Summary ..................................................................................................................................... 29 3 Methods A: Quantifying Cam FAI using Open MRI .................................................................... 30 3.1 Introduction ................................................................................................................................. 30 3.2 Methods....................................................................................................................................... 31 3.2.1 Protocol Development ....................................................................................................................... 31 3.2.2 Repeatability ...................................................................................................................................... 38 3.3 Results ......................................................................................................................................... 40 3.4 Discussion ................................................................................................................................... 41 3.5 Summary ..................................................................................................................................... 44 4 Methods B: Positioning Hips in an Impingement Posture ............................................................ 45 4.1 Introduction ................................................................................................................................. 45 4.2 Methods....................................................................................................................................... 46 4.2.1 Protocol Development ....................................................................................................................... 46 4.2.2 Repeatability and Accuracy ............................................................................................................... 51 4.3 Results ......................................................................................................................................... 53 4.4 Discussion ................................................................................................................................... 56 4.5 Summary ..................................................................................................................................... 60 5 Methods C: Measuring Acetabular Contact Force in Cam FAI .................................................. 61 5.1 Introduction ................................................................................................................................. 61 5.2 Methods....................................................................................................................................... 62 5.2.1 Protocol Development ....................................................................................................................... 62 5.2.2 Repeatability ...................................................................................................................................... 70 5.3 Results ......................................................................................................................................... 70 5.4 Discussion ................................................................................................................................... 73 5.5 Summary ..................................................................................................................................... 76 6 Integrated Study: Open-MRI Measures of Cam Intrusion for Hips in an Anterior Impingement Posture Are Related to Acetabular Contact Force ................................................................................ 77 6.1 Introduction ................................................................................................................................. 77 6.2 Methods....................................................................................................................................... 78 6.3 Results ......................................................................................................................................... 84 6.4 Discussion ................................................................................................................................... 93 6.5 Conclusion .................................................................................................................................. 96  vi  7 Integrated Discussion ........................................................................................................................ 97 7.1 Motivation & Contributions ........................................................................................................ 97 7.2 Strengths ..................................................................................................................................... 98 7.3 Limitations .................................................................................................................................. 99 7.4 Significance ............................................................................................................................... 101 7.5 Future Directions ...................................................................................................................... 101 7.6 Conclusion ................................................................................................................................ 102 References ................................................................................................................................................ 103 Appendix A Specimen Screening .................................................................................................... 118 Appendix B Hip Joint Coordinate System ..................................................................................... 121 Appendix C Joint Position Measurement ....................................................................................... 123 Appendix D Analysis of the Anterior Impingement Test .............................................................. 127     vii  List of Tables Table 3.1: MRI sequence parameters. ......................................................................................................... 33 Table 3.2: Judgment using binary pincer scale between two trials by Reader A. ....................................... 40 Table 3.3: Joint judgment using binary pincer scale between two trials by Reader B. ............................... 40 Table 3.4: Joint judgment using binary pincer scale between Reader A and Reader B. ............................. 41 Table 4.1: Measured flexion angles in degrees (6 repeated positioning trials per hip). .............................. 54 Table 4.2: Measured external rotation angles in degrees (6 repeated positioning trials per hip). ............... 54 Table 4.3: Measured abduction angles in degrees (6 repeated positioning trials per hip). ......................... 54 Table 4.4: Measured lateral translation (mm) of the femoral head (6 positioning trials per hip). .............. 55 Table 4.5: Measured anterior translation (mm) of the femoral head (6 positioning trials per hip). ............ 55 Table 4.6: Measured superior translation (mm) of the femoral head (6 positioning trials per hip). ........... 55 Table 4.7: Measured accuracy error in degrees for flexion and abduction. ................................................ 56 Table 5.1: Tekscan K-Scan 4400 parameters (active sensing region only) [47]. ........................................ 62 Table 5.2: Resultant force and centroid location, presented as mean and SD (6 trials), for each hip. ........ 71 Table 5.3: Region-specific resultant force (N) for each specimen reported as mean (SD) from 6 trials. ... 71 Table 5.4: Repeatability results for resultant contact force from six trials each in 10 hips. ....................... 72 Table 5.5: Repeatability results for contact centroid (six repeated trials per hip). ..................................... 72 Table 5.6: Repeatability results (6 trials per hip) for region-specific contact force. ................................... 73 Table 6.1: Selected specimen screening information. ................................................................................. 79 Table 6.2: MRI sequence parameters .......................................................................................................... 80 Table 6.3: Summary of MRI results for minimum β and intrusion based on a 0° β threshold. .................. 89 Table 6.4: Resultant force and centroid location for each hip. ................................................................... 90 Table 6.5: Resultant force (N) at each of 6 regions for each hip. ............................................................... 90 Table 6.6: Contingency table of each hip, MRI cam-intrusion sign vs morphology group. ....................... 91 Table 6.7: Contingency table of each hip, contact-force sign vs morphology group.................................. 92  viii  Table 6.8: Contingency table for the MRI cam-intrusion sign and contact-force sign. .............................. 93 Table A.1: Complete specimen screening information. ............................................................................ 120     ix  List of Figures Figure 1.1: Schematic diagram of cam impingement. .................................................................................. 1 Figure 2.1: The hip joint as a rotational conchoid. ....................................................................................... 5 Figure 2.2: The anterior impingement test. ................................................................................................... 8 Figure 2.3: A schematic diagram of α-angle measurement in cam and control hips. ................................. 11 Figure 2.4: Radial slice planes generated using multi-planar reformatting (MPR). ................................... 12 Figure 2.5: Definition of the β angle on an MR image and on a diagram. ................................................. 13 Figure 2.6: Surgical resection simulation of the cam deformity for treatment of cam FAI. ....................... 16 Figure 2.7: Distribution of cartilage/labal damage and computed impingement zones in cam hips. ......... 19 Figure 2.8: An open MRI image of a normal hip in the W-sitting position. ............................................... 25 Figure 2.9: A schematic diagram of pressure transducers implanted into the acetabulum. ........................ 27 Figure 3.1: The 0.5T MROpen (Paramed; Genoa, Italy). ........................................................................... 31 Figure 3.2: The scan coil wrapped in protective plastic for biosafety. ....................................................... 33 Figure 3.3: A view of the proximal femur showing multi-planar reformatting. ......................................... 35 Figure 3.4: The β-angle defined by Wyss et al. .......................................................................................... 36 Figure 3.5: A pilot image of a cam hip demonstrating negative β-angle and cam intrusion. ..................... 37 Figure 3.6: Sequence of steps performed by readers measuring β-angle. ................................................... 38 Figure 4.1: Components of the positioning rig. .......................................................................................... 47 Figure 4.2: Joints of the positioning rig. ..................................................................................................... 48 Figure 4.3: Illustration of the hip joint coordinate system under the floating axis convention................... 49 Figure 4.4: A right hemi-pelvis with proximal femur positioned in a simulated anterior impingement posture. ........................................................................................................................................................ 51 Figure 5.1: The front and back of the Tekscan K-Scan 4400 contact force sensor. ................................... 63 Figure 5.2: Cross-sections of the Tekscan K-Scan 4400 hip contact force sensor. .................................... 63 Figure 5.3: Dissection of the hip for sensor insertion. ................................................................................ 66  x  Figure 5.4: The sensor aligned along the chondrolabral junction of the lunate surface. ............................. 67 Figure 5.5: The clock-face representation of the acetabulum. .................................................................... 68 Figure 5.6: The sensor map and associated clock-face orientation. ............................................................ 69 Figure 5.7: Mean vs SD (from 6 trials) for region-specific force (10 hips, 6 regions each). ...................... 73 Figure 6.1: Schematic diagram depicting the different slice planes after multi-planar reformatting.......... 81 Figure 6.2: Sample β-angle profile. ............................................................................................................ 82 Figure 6.3: Results from Hip #9, a control hip (with chondrocalcinosis). .................................................. 85 Figure 6.4: Results from two bilateral hips with pure cam morphology..................................................... 86 Figure 6.5: Results from Hip #7.................................................................................................................. 87 Figure 6.6: Results from Hip #5.................................................................................................................. 87 Figure 6.7: Visual inspection of joint space shown on MRI in two positions. ........................................... 88 Figure 6.8: Fisher’s exact test of cam-intrusion sign vs. cam deformity for various β-angle thresholds. .. 91 Figure 6.9: Fisher’s exact test of contact-force sign vs. cam deformity for various force thresholds. ....... 92 Figure 6.10: Open MRI for two hips that dislocated during the sensor experiment. .................................. 94 Figure A.1: Diagrams showing a standard geographic representation of the acetabulum. ....................... 119 Figure B.1: Illustration of the hip joint coordinate system. ...................................................................... 122 Figure C.1: Optotrak probe (Optotrak, Northern Digital; Waterloo, Ontario, Canada). ........................... 125 Figure D.1: A schematic of the anterior impingement test. ...................................................................... 127 Figure D.2: Free body diagram of the lower leg (coronal view) in the anterior impingement test. .......... 129 Figure D.3: A free body diagram of the thigh (axial view) in the anterior impingement test. .................. 130     xi  List of Acronyms 2D  Two dimensional 3D  Three dimensional A-P  Anterior-posterior ASIS  Anterior superior iliac spine CI  Confidence interval CT  Computed tomography CV  Coefficient of variation dGEMRIC Delayed gadolinium-enhanced magnetic resonance imaging of cartilage DOF  Degrees of freedom FAI  Femoroacetabular impingement FOV  Field of view ICC  Intra-class correlation coefficient LCPD  Legg-Calvé-Perthes Disease MPR  Multi-planar reformatting MRI  Magnetic resonance imaging OA  Osteoarthritis PSIS  Posterior superior iliac spine RMS  Root mean square ROM  Range of motion SCFE  Slipped capital femoral epiphysis SD  Standard deviation SNR  Signal to noise ratio TE  Echo time TR  Repetition time T1  Spin-lattice relaxation time T2  Spin-spin relaxation time   xii  Acknowledgements Thank you to Dr David Wilson for your invaluable guidance and support during my studies. Your mentorship has fostered my growth as a researcher. I am grateful for the experience. Thank you to all staff at the Centre for Hip Health and Mobility for helping me survive my studies to date. You are life-savers and the unsung heroes of research. Thank you to the UBC Biomedical Engineering program and Engineers in Scrubs for creating such a rich, collaborative learning environment, and for your financial support. Thank you to my lab mates, fellow students, and Drill Cover team for your mentorship, guidance, and constructive feedback. Finally, thank you to Genelle, my parents, family, and friends for your unconditional love, inspiration, and cooking. Your support is inimitable and the best motivation of all.   xiii  Dedication      In loving memory of C. Elaine Murray  Chapter 1: Introduction 1  1 Introduction Cam-type femoroacetabular impingement (FAI) is a mechanical process thought to be a leading cause of hip osteoarthritis (OA) [1]. Cam FAI can occur in hips with a ‘cam deformity’, which is characterized by a visible bump on the femoral neck or decreased concavity of the femoral head-neck junction [1–3]. In cam FAI, it is postulated that the cam deformity intrudes into the hip joint space and causes elevated shear forces on acetabular cartilage at the chondrolabral junction (Figure 1.1) [1]. The cam deformity is usually located on the anterosuperior femoral head-neck junction [4–8], which means that movements like combined flexion and internal rotation are involved with cam impingement because they bring the cam deformity closer to the anterosuperior acetabular margin [9]. Symptoms associated with cam FAI include reduced range of motion as well as pain during activity or prolonged sitting [10,11]. Symptomatic cam FAI presents in young patients, typically aged 20-40 years [12]. Cam deformities have been associated with up to 52% of hips that require total hip replacement before the age of 55 [13].  Figure 1.1: Schematic diagram of cam impingement. © Reproduced from [14] with permission from Wolters Kluwer Health. While there appears to be a causal relationship between some cam deformities and hip OA, there is a significant body of evidence to suggest that not all hips with cam deformities will develop OA. Cross-Chapter 1: Introduction 2  sectional studies have estimated that radiographic cam deformities are present in 5% to 35% of people who are asymptomatic and free of hip OA [15–18], and other retrospective studies of asymptomatic cam hips with 10 to 20 year follow-up have reported that OA did not progress in 33% to 82.3% of cases [19,20]. It is not yet clear why some cam hips develop OA and some do not.  Biomechanical studies of femoroacetabular impingement are required to explain why some cam hips develop OA and some do not, since the postulated etiology is mechanical. In general, few experimental studies have explored contact mechanics in FAI despite the central role of mechanics in the hypothesis that FAI causes hip OA. Ideally, researchers would measure tangential shear and compressive forces in hip cartilage during physiological cam impingement, but this would require implantation of sensors that would be far too invasive to be ethical in vivo [21]. A number of computer models have been used to simulate motion and predict impingement location [22–28], while finite element models have predicted joint contact pressure and cartilage stress [29–31], but model-based investigation requires researchers to make assumptions and simplifications of hip physiology. Such assumptions may limit the utility of computer models in detecting clinically important mechanical differences between symptomatic and asymptomatic cam hips in impingement positions. Direct three-dimensional imaging of postures suspected of causing impingement has strong potential to provide insight into the mechanics of FAI because it is non-invasive and provides direct visualization of hip anatomy. With imaging, no assumptions about hip movement and no simplifications of hip structures are required. In contrast, many models assume that the centre of hip rotation is fixed and that the acetabular labrum or cartilage is rigid, neither of which is true. Open MRI has been used in one study for qualitative investigation of normal hips in common impingement positions, but has never been applied to cam hips in impingement positions. A key limitation to both imaging and computer models is that they cannot provide direct measures of the mechanical properties of interest to explaining the etiology of osteoarthritis, such as contact force. Quantitative analysis using imaging or models would be strengthened by comparison to experimental Chapter 1: Introduction 3  findings, but to date no studies of cam hips in impingement positions have measured contact mechanics experimentally. However, contact mechanics of the hip joint have been studied extensively in non-impingement positions [32–46], and hip-specific piezoresistive sensors in particular have potential to be used intra-articularly in cadaver hips because they are less than 1 mm thick and able to make measurements of continuously changing forces [47]. To use an open MR imaging approach for in vivo studies of FAI, we must first establish whether we can directly observe cam FAI using MRI, and whether cam impingement can be quantified with measures that relate to direct measures of contact mechanics. Since it is not possible to simultaneously acquire MR images and take sensor-based measurements of contact force in the hip joint due to interference of the sensor with MRI, and since variability in joint position can have significant implications on the clearance between the cam deformity and acetabular margin [26], this work requires a standardized joint positioning technique whose precision is known. This thesis had the following aims: A. To develop a novel method for investigating and quantifying cam impingement using open MRI; B. To develop a method for positioning hips in a simulated anterior impingement posture such that there is limited error in joint angles and translations between repeat positioning trials in the same specimen; and C. To develop a method for measuring acetabular contact force directly in a simulated anterior impingement posture, in order to relate those measurements to MRI-based measures of cam FAI. After developing these methods, this thesis addressed the following questions in cadaver hips (with and without cam deformities) positioned in a simulated anterior impingement posture: 1. Does open MRI show intrusion of the cam deformity into the intra-articular joint space? 2. Are cam deformities associated with acetabular contact force? 3. Are MRI measures of cam intrusion related to acetabular contact force? Chapter 2: Literature Review 4  2 Literature Review 2.1 Hip Joint The hip is a major weight-bearing joint that connects the trunk to the lower limb via articulation between the acetabulum of the pelvis and head of the femur. 2.1.1 Anatomy The pelvis comprises three bones (ilium, ischium, pubis) that fuse together during adolescence and create a Y-shaped union; the acetabulum is located at the centre of the Y and projects laterally. The only surface of the acetabulum that is covered by articular cartilage is the horseshoe-shaped lunate surface. In the middle of the horseshoe is the cotyloid fossa, where the round ligament of the femoral head originates. The horns of the lunate surface point inferiorly and form the acetabular notch which is traversed by the transverse acetabular ligament, creating the acetabular foramen. The proximal femur has a round head that articulates with the acetabulum. Approximately half of the femoral head is covered by articular cartilage, except for the fovea capitis, a small indentation where the round ligament of the femoral head attaches. The femoral head is connected to the femoral neck by a (usually) concave head-neck junction. The femoral neck projects inferolaterally where it connects with the femoral shaft. The hip is commonly approximated as a ball-and-socket spherical joint that permits rotational gliding motion, but the femoral head may be better approximated by a rotational conchoid. Meridian cross-sections of a rotational conchoid have the general equation: r = a + b cosφ (Figure 2.1). The consequence of a rotational conchoid hip model is that relative motion between the femur and acetabulum involves both rolling (i.e. translation) and rotational gliding [48].  Chapter 2: Literature Review 5   Figure 2.1: The hip joint as a rotational conchoid. (A) Schematic diagram of a conchoid with the general equation, r = a + b cosφ. (B) The respective conchoids of the acetabulum and femoral head leave a distinct gap in the non-load bearing zone that is important for lubrication [48]. © Reproduced from [48] with permission from Elsevier. A fibrous joint capsule surrounds the hip joint which helps to hold the femoral head in the acetabulum, while an inner synovial membrane lines the joint capsule and contains the lubricating synovial fluid within the joint. The iliofemoral, ischiofemoral, and pubofemoral ligaments surround the hip joint capsule. The round ligament of the femoral head connects the acetabular foramen and the fovea capitis of the femoral head, providing a passageway for blood vessels and nerves to reach the femoral head. The acetabular labrum is a fibrocartilaginous structure that lines the acetabular rim. The labrum forms a continuous boundary, a ‘chondrolabral junction’, with the articular cartilage on the lunate surface. The labrum is thought to improve load distribution on acetabular cartilage in two ways: (1) by forming a seal against the head of the femur to pressurize synovial fluid within the intra-articular space, which allows fluid to evenly distribute contact pressure during weight bearing [49]; and (2) by increasing the acetabular contact area, which is thought to reduce mean contact pressure on the lunate surface [50]. Motion at the hip is actuated by 22 muscles that cross the hip joint. They generally operate in groups to initiate and control rotation (in flexion/extension, abduction/adduction, and internal/external rotation), and also control translation of the femoral head within the acetabulum. Chapter 2: Literature Review 6  2.1.2 Osteoarthritis Osteoarthritis (OA) is a progressive, debilitating disease that causes deep, aching joint pain during daily tasks and inhibits mobility [51]. Osteoarthritis involves the breakdown of articular cartilage and changes to subchondral bone, synovium, and ligaments [52]. Early in the OA process, the surface of articular cartilage becomes visibly rough and fibrillations form. Since articular cartilage is an avascular tissue with no nervous supply, and cartilage has limited ability to heal on its own, it can be difficult to intervene in the OA process [52]. Continued cartilage degeneration often leads to exposure of the underlying subchondral bone and osteophytes (bone spurs) in late-stage OA [52]. In 2010, an estimated 4.4 million Canadians were living with OA [53]. That number is expected to reach over 10.4 million by 2040 [53]. The associated economic burden is massive: in 2010, cumulative direct costs in Canada associated with OA (due to hospital visits and other treatment) were estimated at $10.2 billion, while indirect costs (due to disability and loss of productivity) were estimated at $17.3 billion, for a total of $27.5 billion [53]. The hip is commonly affected by OA; in a 2007 systematic review of 23 studies, radiographic hip OA prevalence ranged from 0.9% (population aged 20 years and over) to 27% (population aged 45 and over), with an 8.0% mean, 7.0% standard deviation, and 5.3% median [54]. Mechanics play a key role in the initiation and progression of OA [55], and it has been recently argued that nearly all cases of osteoarthritis are caused by abnormal mechanics rather than other factors like inflammation [56]. In the hip, the hypothesized mechanical causes of hip pain and OA have been grouped as static or dynamic [57]. Static causes (such as obesity, acetabular dysplasia, acetabular anteversion, or femoral valgus/coxa valga) increase contact pressure above normal levels in key weight-bearing cartilage zones during activities when the hip is in a relatively neutral position (i.e. when standing and during gait). Dynamic causes (cam or pincer FAI) are thought to induce abnormal forces on acetabular cartilage during activities that place the hip close to the limits of motion [1]. Dynamic causes are therefore associated with young athletic populations who frequently participate in sports that involve a wide range of hip motion [58]. Because FAI is associated with young patients, and the underlying causes are thought to be primarily Chapter 2: Literature Review 7  mechanical, there is hope that FAI-related hip OA may be preventable [59]. For this reason, FAI has become a prominent research topic in orthopaedic biomechanics. 2.2 Cam-type Femoroacetabular Impingement As described by Leunig and colleagues, “femoroacetabular impingement is not a disease per se, but rather a pathomechanical process by which the hip can fail” [60]. FAI is defined as ‘abnormal contact’ that occurs between the femoral head-neck junction and acetabulum during hip motion. Two types of impingement mechanisms are possible, depending on whether certain morphological aberrations are present: (i) cam impingement, and (ii) pincer impingement [1]. The definition of ‘abnormal contact’ in impingement varies by mechanism. The scope of this thesis is limited to cam-type FAI. Cam impingement occurs in hips that have decreased concavity at the femoral head-neck junction. Decreased concavity may appear as flattening of the head-neck junction or as a raised bump, usually located anterosuperiorly [61], and is often referred to as ‘cam deformity’. It is postulated that the cam deformity jams inside the acetabulum during forceful motion (particularly during internal rotation at high flexion), leading to high shear forces on acetabular cartilage at the chondrolabral junction, and eventually OA [1], although there is no experimental evidence to suggest that cartilage forces during cam impingement will be shear rather than compressive. 2.2.1 Clinical Presentation Patients with cam FAI pathology typically present with pain during activity or prolonged sitting, which may sometimes be accompanied by a sensation that patients and clinicians describe as ‘catching’ or ‘locking’, as well as restricted range of motion [62]. Hip pain related to FAI is typically focused in the anteromedial groin region, but is also associated with pain in the pubic symphysis, sacroiliac joint, lumbar spine, and posterior acetabulum [10,57]. Motion limitations are common in flexion, internal rotation, abduction, but most notably internal rotation combined with 90° flexion [63–66]. Labral tears are frequently Chapter 2: Literature Review 8  associated with cam deformities, with one study reporting that 76% patients with labral tears showed radiographic cam deformities on CT [67].  The anterior impingement test is one of the most commonly used tests during physical examination because it recreates what is thought to be a worst-case impingement scenario [11] by placing the anterosuperior region of the femoral head-neck junction (where cam deformities are most prominent [61,68,69]) close to the anterosuperior region of the acetabulum (where the acetabular margin is often most prominent [70,71]). First described by Klaue et al. [72], the examiner passively flexes the hip to 90°, then forces internal rotation, sometimes with added adduction (Figure 2.2). The test is ‘positive’ if the patient experiences pain, as is often the case in patients with cam FAI [1]. Due to its clinical relevance, the anterior impingement test is also commonly used in biomechanical studies of FAI [22–28,73–76].   Figure 2.2: The anterior impingement test. The hip is flexed passively to 90°, then placed into internal rotation manually by the examiner. © Reproduced from [11] with permission from Springer. Chapter 2: Literature Review 9  2.2.2 Epidemiology & Etiology FAI affects young patients, with most patients presenting with symptoms between the ages of 20 and 40 years [12]. Radiographic FAI has been associated with up to 52% of hips with OA that require total hip replacement before the age of 55 [13]. It is generally agreed that cam deformities appear on imaging more frequently in men than women, but reported ratios vary dramatically from 3:1 (male:female) [15], to 4:1 [16], to 14:1 [12]. Signs of pincer FAI are more prevalent in women [12], but prevalence of cam deformities in women was still reported as high as 11% in one study [17]. In males, prevalence of cam deformities is approximately 1-in-4 [15,17,77]. However, most populations that have been studied are highly homogenous and incidence may vary depending on the population. For instance, prevalence varies dramatically depending on activity levels, and also varies by ethnicity; one study of a Japanese population with hip OA reported that cam or pincer-related deformities were present in fewer than 1% of cases (n=946) [78]. The etiology of the cam deformity is not completely understood [79], but most cam deformities are now thought to be a consequence of growth plate remodelling that arises due to activity-related stress during closure of the growth plate [80,81]. This hypothesis is largely driven by evidence that cam deformities in young populations are significantly associated with high-frequency participation in sports [58]. One study found that the prevalence of radiographic cam morphology in collegiate American football players may be as high as 72% [82]. Other cross-sectional studies comparing athlete sub-groups to non-athletes made similar findings in adolescent football (soccer) players) [83]; elite ice hockey players [84,85]; elite capoeira athletes (a Brazilian martial art requiring extreme hip movement, jumping, and kicking) [86]; and elite track and field athletes [87]. One finite element modelling study of hips with un-fused growth plates reported that stress at the common cam deformity site is significantly increased during certain loading activities (i.e. weight-bearing with hip flexion) [88]. Other causes of cam deformities may include: (i) incorrect healing after a femoral neck fracture resulting in an anterosuperior prominence [64,89];  (ii) Legg-Calvé-Perthes disease (Perthes’, or LCPD) [90]; and (iii) slipped capital femoral epiphysis (SCFE) [91,92]. In LCPD, a loss of blood flow to the femoral head Chapter 2: Literature Review 10  leads to necrosis, which interrupts the bony modelling process of the femoral head during growth. If LCPD is mild, the condition may not need intervention and therefore might introduce a residual deformity during healing, affecting the congruence of the hip. In SCFE, the epiphysis (the portion of the femoral head proximal to the growth plate) slips along the not-yet-fused physis (grow plate). If the slip is mild and does not warrant intervention during adolescence, the growth plate will fuse resulting in a permanent slip, or ‘head tilt’, that resembles a cam deformity. The permanent slip can induce anterior impingement with the acetabulum and is thought to eventually lead to cartilage damage [93]. 2.2.3 Quantification The most widely used quantifier of the cam deformity is the α-angle. As it was first described by Nötzli et al., the α-angle was measured on an MRI slice through the middle of the femoral head that was normal to the coronal plane and also contained the femoral neck axis [3]. The α-angle was defined in that plane as the angle between the line along the centre of the femoral neck and the line joining the centre of the femoral head with the point at which the femoral head-neck junction first exceeded the radius of the perfect circle fit to the femoral head anteriorly [3] (Figure 2.3). When applied to a control group and a patient group with groin pain, decreased internal rotation, and a positive impingement test, the α-angle averaged 42.0 ± 2.2° (range 33 to 48) for controls and 74.0 ± 5.4° (range 55 to 95) for patients. The α-angle can also be measured using CT: Beaulé et al. used CT to measure α-angle in 20 controls and 30 symptomatic cam FAI patients (groin pain, positive anterior impingement test, and normal acetabula), and found an α-angle of 43.8 ± 4.5° (range 36.1 to 50.1) for controls and 66.4 ± 17.2° (range 41.1 to 95.0) [61]. These findings led to the suggestion that a threshold for α-angle of 50° be adopted to define the presence of cam morphology [3,61].   Chapter 2: Literature Review 11   Figure 2.3: A schematic diagram of α-angle measurement in cam and control hips. As first described by Nötzli et al., the reader first fits a circle of radius (r) and centre (hc) to the femoral head. The reader then selects the point (A) at which the femoral head-neck junction extends beyond the femoral head circle. The α-angle is measured between A-hc and hc-nc, where nc is a point along the femoral neck axis. A control hip is shown on the left, while a cam-type hip is shown on the right. © Reproduced from [3] with permission from the British Editorial Society of Bone and Joint Surgery. The original description of the α-angle was limited because it only considered the contour of the femoral head-neck junction in a single plane, when in reality cam morphology is very much a 3D entity. Multi-planar reformatting (MPR) of 3D images addresses this concern. In MPR, the original 3D image is used to generate a radially sliced image volume, revolved about the long axis of the femoral neck, so that clinicians can measure α-angle in multiple planes to capture the size of the cam deformity around the entire circumference of the head-neck junction (Figure 2.4) [4]. Studies employing MPR have demonstrated that:  It is necessary to consider the size of the cam deformity about the entire circumference of the femoral neck [4,6,7]  The α-angle is commonly largest in the anterosuperior region [6–8,69,94,95], so the true size of the cam deformity can be underestimated if only viewing the anterior plane [5]; and   An α-angle threshold of 60° is more appropriate when defining the presence/absence of a cam deformity on radially reformatted images [4]. CONTROL CAM Chapter 2: Literature Review 12   Figure 2.4: Radial slice planes generated using multi-planar reformatting (MPR). With radial reformats, α-angle and other measures can be made about the entire circumference of the femoral head-neck junction. © Reproduced from [4] with permission from the Radiological Society of North America. The β-angle was proposed in 2007 to measure the functional clearance between the cam deformity and acetabular rim (incorporating joint position in addition to both femoral and acetabular anatomy) [9], and has major advantages over the purely structural α-angle. The β-angle is defined by a line connecting the centre of the femoral head to the first point on the head-neck junction that extended beyond the femoral head radius, and by a line connecting the femoral head centre to the acetabular margin (Figure 2.5). Since the β-angle involves landmarks from both the femur and acetabulum, β-angle is dependent on the relative position between the two bones and provides information about the functional clearance between them. In its first description, β-angle was measured at 90° flexion in neutral rotation, and was shown to correlate with the maximum achievable internal rotation at 90° flexion [9]. Subsequent research has found that cartilage health (measured via delayed gadolinium-enhance magnetic resonance imaging of cartilage, dGEMRIC) was correlated with the β-angle, but not the α-angle, when measured in neutral positions using MPR reformats of MRI [95].  Chapter 2: Literature Review 13   Figure 2.5: Definition of the β angle on an MR image and on a diagram. The reader first defines a circle fit to the femoral head, with centre hc, and the point at which the femoral head-neck junction extends beyond the head circle, C. The β-angle is formed by the line hc-C and hc-D, where D is the lateral-most margin of the acetabular rim. © Reproduced from [9] with permission from Wolters Kluwer Health. 2.2.4 Causal Relationship between Cam FAI and OA More than 75 years ago it was accepted that many cases of hip arthritis were idiopathic (a disease process without a known cause), yet in the last 50 years, researchers have put together a convincing body of evidence that suggests that most idiopathic hip OA is actually secondary to processes such as FAI [96]. In 1965, Murray published a retrospective review of 200 A-P radiographs of patients with ‘primary’ hip OA and identified that 39.5% of cases actually showed a femoral head that had very subtle varus tilt relative to the femoral neck, which was given the term ‘tilt deformity’ [97]. In 1975, Stulberg et al. described the similar ‘pistol-grip deformity’, and in 1976 Solomon postulated that hip OA was secondary to such deformities [98]. In 1986, Harris suggested that hip OA as a primary disease occurs extremely rarely, if at Chapter 2: Literature Review 14  all [99]. In 2001, Ganz et al. published a safe surgical dislocation technique [100] which avoided avascular necrosis of the femoral head and permitted an open, full view of the hip joint. With this technique, surgeons were able to make observations about the interaction between the femur and acetabulum in hip OA patients, and postulated that most cases of idiopathic hip OA were in fact due to cam impingement [1]. Since then, FAI research has progressed rapidly, but it is arguably still not established that FAI-related deformities cause hip OA. A striking limitation in of the early FAI research by Stulberg, Murray, and Solomon leading to the FAI hypothesis is that the studies were each designed as retrospective reviews of radiographs. In 1976, Resnick had proposed that the tilt deformity could actually be a consequence of the OA process, rather than a cause [101]. In 2008, Ganz et al. suggested that, in order to establish a causal relationship between FAI and OA, researchers must: (i) conduct prospective studies to demonstrate the progression of OA in the presence/absence of FAI-related deformities; (ii) develop a treatment method that can slow or halt the progression of hip OA; and (iii) firmly establish the biomechanics of FAI. 2.2.5 Prospective Studies To the best of my knowledge, no prospective studies of hips with cam FAI with long-term follow-up to track OA progression have been conducted to date. Cross-sectional and retrospective studies of hips have shown associations between the presence of cam deformities and pain (University of Ottawa cohort [102]), MRI-detected cartilage damage (Sumiswald Swiss cohort [77]), clinical hip OA (Cohort Hip and Cohort Knee [103]; Washington University cohort [104]), or end-stage OA leading to total hip replacement (Chingford cohort [105]). However, other retrospective populations studies have reported that the presence of radiographic OA does not always mean that OA will progress (after follow-up of 10 years or greater [19,20]), and that imaging-based measures that quantify FAI-related deformities, such as α-angle, cannot predict the age at which patients reach end-stage OA requiring THA [106]. Furthermore, the estimated prevalence of radiographic OA in asymptomatic, pain-free populations ranges between 5% and 35% [15–18], indicating that more factors than just hip structure should be considered when predicting OA risk. Chapter 2: Literature Review 15  One animal model for cam FAI has been created in sheep and successfully produced rapid joint degeneration [107]. The sheep hip was selected because it naturally has greater asphericity of the head-neck junction and reduced anteversion compared to human hips. Siebenrock et al. used a 15° medial closed-wedge varus osteotomy (without disrupting the joint capsule) that induced cam-type impingement during flexion in 16 Swiss Alpine sheep, with the contralateral hip left intact (control). Macroscopically and histologically, 15 of 16 cam hips acquired chondrolabral lesions (after 14-38 weeks) in the weight-bearing zone versus only 1 of 16 control hips. One limitation of this study was that there was no sham surgery control group. Since no sheep received bilateral osteotomies, post-operative asymmetrical biomechanics may have played a role in joint degradation. 2.2.6 Treatment Surgical treatment of cam FAI has become widely used in the last decade. The surgical objective is to restore concavity to the femoral head-neck junction by resecting the cam deformity (Figure 2.6), with the goal of restoring range of motion, reducing the number of collisions that occur during daily motion, and eliminating the shear forces due to cam impingement [59,69,108–110]. The surgery can be performed with open dislocation [100,111–113], or arthroscopically [60,79,114], but open dislocations risk damage to neurovascular structures [79] and the arthroscopic approach is technically challenging [115]. Computer assisted surgery has the potential to improve arthroscopic outcomes [109,115], but it is not yet clear how effectively the surgical resection of cam deformities halts progression of hip OA in the long term [28]. Chapter 2: Literature Review 16   Figure 2.6: Surgical resection simulation of the cam deformity for treatment of cam FAI. Generated using Mimics 15.0 (Materialise; Leuven, Belgium). Physiotherapy is the most common non-operative treatment option. FAI patients typically present with altered gait kinematics and kinetics compared to controls [116–119], which has led to the hypothesis that internal joint loading due to muscle activation is vastly different between FAI patients and controls. The goal of physiotherapy is therefore to strengthen muscles that cross the hip, which are generally weak in FAI patients [120]. Therapists primarily focus on strengthening the abductor and external rotator muscles groups; when these muscles contract, a force is applied to the proximal femur that acts posteriorly/inferiorly/laterally, which counteracts anterior/superior/medial translation of the femoral head-neck junction and is thought to resist anterior impingement. Some case studies have reported success in reducing FAI-related symptoms [121], but this theory has yet to be tested in a broad population. Physiotherapy is also commonly used in conjunction with surgical cam decompression to improve rehabilitation, but its efficacy is not yet clear [122]. One study has published methods for a randomised control trial, where FAI patients undergoing surgical cam resection (n=100) will be assigned to a physiotherapy group or control group, in order to investigate the effect of physiotherapy on quality of life, physical function, and reduction of post-operative FAI-related symptoms [122]. The study results have not yet been published. Chapter 2: Literature Review 17  2.3 Biomechanics of Cam FAI There are surprisingly few research publications that provide quantitative, experimental evidence in support of the cam FAI hypothesis that shear forces occur at the acetabular chondrolabral junction. The literature to date has generally focused on measuring three main factors characteristic of cam-type impingement: 1. Reductions in hip range of motion (ROM), measured primarily using experimental methods; 2. The interaction between the cam deformity and acetabulum during motion, to date only measured by computer models; and 3.  Other kinematic changes that might be secondary to bony collision during cam impingement, often measured experimentally. 2.3.1 Range of Motion Patients with cam deformities typically present with restricted flexion, internal rotation, and abduction [63–66]. ROM limitations have been demonstrated experimentally in many studies and subsequently supported with computer models. In theory, a reduced ROM would mean a higher likelihood that cam impingement would occur during (and impede) daily activity, and provides a good indication of the joint position thought to cause impingement. ROM measurements tell us nothing directly about hip joint contact mechanics. An experimental in vivo study found that symptomatic cam hips (n=24) had significantly reduced ROM in all directions (external rotation, internal rotation, flexion, and internal rotation at 90° flexion) compared to controls (n=24. The study also found that symptomatic cam hips had reduced internal rotation at 90° flexion compared to asymptomatic cam hips (n=24) [25]. However it is not clear whether range of motion was dictated by bony impingement, soft-tissue constraints, or pain-related (compensatory) limits. Another in vivo study found that cam or pincer FAI hips (n=32) had on average 24° less internal rotation at 90° flexion compared to controls (n=40) [9]. This study also used open-configuration MRI to acquire images of each hip in 90° flexion, and defined the novel β-angle (Figure 2.5, Section 2.2.3 above) as a quantifier of the neck-rim relationship [9]. The β-angle was 5° ± 9° (mean ± standard deviation) in the cam Chapter 2: Literature Review 18  or pincer subjects compared to 30° ± 9° in the controls, and β-angle strongly correlated with internal rotation at 90° flexion (R=0.97) [9]. This work provided strong support for the hypothesis that linear abutment of the head-neck junction against the acetabular rim (impingement) terminates motion. One major limitation of this work was that the authors do not report their method for joint angle measurement.  Computer models relying on mathematically simulated hip motion have confirmed that cam hips are limited in flexion, internal rotation, abduction, and internal rotation at high flexion [22,23,25,26]. The basic principle for collision detection models is to create a rigid solid body of both the femur and acetabulum using patient-specific CT data, and then rotate the femur about a chosen centre of rotation (often representing a clinical exam like the anterior impingement exam). Motion continues in stepwise increments until collision between the two bodies is identified and located. A key advantage of this approach is that patient-specific models and computer automation allow a broad range of disorders and treatments to be simulated. However, computer models have been shown to over-predict ROM of cadaver hips [123]. The leading limitation of these computer models, and potential reason for ROM prediction error in cadaver hips, is that soft-tissue structures such as the labrum, joint fluid, stabilizing ligaments, and cartilage are usually not incorporated, and all bodies are assumed to be rigid. Since soft-tissue structures of the hip may provide resistance to motion, the model may over-predict ROM if only bony collisions are considered. 2.3.2 Impingement Location Computer models have been used to predict impingement contact location, but models (based on mathematical motion simulation) are incomplete reflections of hip biomechanics in vivo, so their predictions for contact location may be unreliable. One study by Tannast et al. found a relationship between the computer-model-predicted impingement location in one group of FAI hips (9 cam, 6 pincer) and the intraoperatively observed chondrolabral damage in a separate group (24 cam, 16 pincer) [24]. The cam group had chondrolabral damage that was focused anterosuperiorly, and the computer-detected impingement zones matched the locations of cartilage damage [24] (Figure 2.7). However, the areas of cartilage damage were much broader than predicted collisions zones. One explanation may be that computer Chapter 2: Literature Review 19  models assume collisions occur at a single point, whereas in real hips, there may be tissue deformation during contact which would result in wider contact patterns and cartilage damage compared to model predictions [24,124].   Figure 2.7: Distribution of cartilage/labal damage and computed impingement zones in cam hips. © Reproduced from [24] with permission from Springer. Another computer model found that collisions consistently occurred on the anterosuperior portion of the acetabulum after simulating 441 internal rotation tests at varying levels of high flexion (70° to 110°) and adduction (-20° to 20°) in 10 symptomatic cam hips, 10 asymptomatic cam hips, and 10 controls, and [25]. On the femur side, collisions of control femurs were localized on the femoral neck anteriorly. Collisions of cam femurs were localized more superolaterally where the cam deformity is usually largest, but were not exclusively located on the cam deformity – about 40% of collisions occurred on the femoral neck, away from the cam deformity [25]. Models can also be used to investigate impingement location without requiring mathematical simulation of hip motion; one common method is to measure patient-specific motion experimentally using optical or magnetic tracking, then directly register computer models to tracking data in order to position the models Chapter 2: Literature Review 20  into the measured joint positions [27,75,125]. A key strength of this method is that it is non-invasive, can be used with patient-specific models and kinematics, and minimizes the number of assumptions made about hip motion. Another method is to apply ‘clinically-achieved’ joint rotations to the computer models, which still requires assumptions about hip motion but ensures ROM is not over-predicted [26]. A key limitation of combined motion tracking with computer models in the prediction of impingement location is that the models cannot make true measurements of contact, but rather use the observed ‘intersection’ of two bodies as a surrogate indictor that collision has occurred. Also, when models are registered to kinematic data, the registration step can introduce error between the relative positions of two objects and potentially skew the final prediction for contact area and location. Four studies have employed this technique (two evaluating cam hips, two evaluating controls only) but none performed a mechanical validation to confirm that the surrogate measure of ‘intersection’ in fact has a relationship with true mechanical contact:  One study of 13 cam hips predicted that the cam deformity intruded medial to the acetabular rim, entering the intra-articular joint space, and ‘engaged’ with the lunate surface of the acetabulum during various passive tests (flexion, abduction, and flexion plus internal rotation). This study used the method of applying clinical achieved joint rotations (measured using magnetic tracking) to CT-derived computer models, assuming a static centre of hip rotation at the femoral head centre. The cam lesion engaged with the anterosuperior quadrant of acetabular cartilage before the terminal point of motion. The engaging portion of the cam lesion varied by position – the superior portion engaged during abduction, while the anterior and inferior portions engaged during the flexion activities.   One study of both cam hips (n=3) and control (n=9) predicted engagement of the inferomedial portion of the femoral head-neck junction with the anterosuperior portion of the labrum [27] using in vivo motion tracking (via dual-fluoroscopy video) that was directly registered to a CT-arthrography model during the anterior impingement exam. The study reported posteroinferior translation of the femoral head in cam hips, which is the first evidence for ‘levering’ of the femur about the acetabulum in cam Chapter 2: Literature Review 21  FAI [27]. It is theorized that posteroinferior translation of the femoral head can lead to elevated contact force on the posteroinferior acetabulum, in a mechanism known as the ‘contre-coup effect’.  Two studies used similar methods to the studies above but only evaluated control hips [75,125]. They registered computer models of the hip (derived from MRI or probe-based digitization) to optical tracking kinematic data in order to follow the models through patient-specific kinematic exams. Neither study performed a characterization of the registration accuracy. 2.3.3 Pelvic Kinematics One study of control hips found that pubic symphysis motion during internal rotation of the hip at 90° flexion was altered after the addition of a simulated cam deformity [74], which may provide an explanation for reported clinical links between cam FAI and athletic pubalgia [126]. Birmingham et al. used optical tracking to measure pubic symphysis rotation during passive internal rotation at 90° flexion, in 12 cadaver hips with normal morphology, both before and after the addition of a simulated cam deformity, and found that the added deformity caused changes in pelvic kinematics beyond 12 Nm of applied internal rotation torque. The cam deformity simulation was performed by trochanteric osteotomy plus a wood dome (5mm dome height, base diameter 25mm) fixed to the anterosuperior femoral head-neck junction.  The cam deformity simulation represented a major limitation to this work: it may not be representative of a true cam deformity. Additionally, the native state was tested prior to capsulotomy, so the native state may have had naturally stiffer soft-tissue preventing bony contact. In light of these limitations, this work suggests that a definite bony contact point related to cam impingement may alter kinematics at the pubic symphysis. One in vivo study found changes in pelvic kinematics during active squatting, a movement suspected to induce cam impingement, but it was not clear whether kinematic changes were due to a pain-related compensatory strategy, muscle weakness, soft-tissue restraint, or forced effect due to bony cam impingement [127]. One study of pelvic motion during squatting found that the cam hips had a decreased sagittal pelvic inclination (14.7±8.4°; n=15) compared to controls (24.2±6.8°; n=11). The differences in sagittal pelvic inclination were independent of squat depth [127]. Pelvic inclination brings the Chapter 2: Literature Review 22  anterosuperior portion of the acetabulum close to the femoral neck, so decreased pelvis inclination may be unavoidable and an effect of direct cam impingement. However, this type of kinematic exam is inherently unable to conclude that direct impingement-related joint contact had occurred. 2.3.4 Labral Seal & Joint Fluid Pressure One experimental study found that joint fluid sealing ability was reduced in 4 cam-type hips with chondrolabral damage compared to 6 normal controls during loaded pivoting (external rotation in neutral flexion/adduction), but not walking or stooping (internal rotation at high flexion) [128]. If indeed cam deformities reduce sealing ability during pivoting, cam hips may experience a drop in hydrostatic pressure during certain activities which could elevate contact force in weight-bearing cartilage zones.  2.3.5 Joint Contact Pressure and Tissue Stress Three finite element models have predicted that the cam deformity does not affect joint contact pressure or acetabular cartilage stress when the hip is loaded in neutral positions, but that contact pressure, cartilage stress, and subchondral bone stress are all raised during activities involving rotation and high flexion:  A finite element model based on idealized geometry predicted that cam deformity size had no effect on peak contact stress or acetabular cartilage stress during walking, but increased joint contact stress during the deep flexion phase of stand-to-sit (at α = 40°, peak contact stress was 3.66 MPa versus 8.84 MPa at α = 80°) [31].  A subject-specific finite element analysis of two hips from patients with large cam deformities (α = 83° for both) and FAI symptoms and two hips from matched controls (α = 42°, 45°) predicted that cartilage stresses were higher in cams versus controls in both standing and maximum squat depth using subject-specific squat kinematics that had been gathered in a separate study [127].  The biggest difference between cam and control hips was found in subchondral bone during squatting: in cam hips, peak maximum shear stress in the underlying subchondral bone was 13.4 and 16.9 MPa in the cam hips versus 4.4 and 4.5 MPa in the control hips [29]. Chapter 2: Literature Review 23   A patient-specific finite element model combined with subject specific ROM for one pathological pure-cam FAI hip (α = 98°) and one control (α = 48°) predicted raised contact pressure in the cam hip compared to the control hip: approximately 3.0 MPa greater in flexion, and 5.6 MPa greater in internal rotation [30].  A key strength of finite element models is their ability to provide direct information about mechanics, such as contact pressure and cartilage stress distribution, completely non-invasively and with patient-specific geometry. However, most models exclude key features of the hip such as pressurized synovial fluid, the labral seal, and ligaments, so it is difficult to know whether these finite element models can accurately predict true contact mechanics. 2.3.6 Subchondral Bone Density In an in vivo quantitative computed tomography study, symptomatic and asymptomatic cam-deformity groups (n=12 for all groups) had greater bone density in the anterosuperior acetabular region than controls by 14-35% and 15-34% respectively [129]. Bone mineral density had a mild positive correlation with α-angle [129]. The increase in bone density may reflect repeated impingement between the deformed femur and the acetabulum. 2.3.7 Considerations in Pincer FAI Pincer impingement is a mechanical concept that is distinct from cam impingement, but is important to consider because pincer impingement is also associated with OA [1,71] and can happen concurrently with cam impingement. The postulated mechanism of pincer impingement is characterized by linear contact of the femoral head-neck junction against the acetabular rim and labrum [1,100]. Chondrolabral damage patterns related to pincer morphology are widely distributed around the acetabulum, whereas damage patterns related to cam morphology are usually more focal and located in the anterosuperior acetabular region [24]. Chapter 2: Literature Review 24  Pincer impingement may occur if the clearance between the femoral neck and acetabular margin is limited. The pathomechanism of pincer impingement is therefore commonly associated with conditions that reduce clearance in the anterior-superior region, such as general and focal acetabular overcoverage, coxa vara, and femoral retroversion [130]. ROM is frequently reduced with pincer impingement [131]. Notably, it has also been suggested that linear contact between the femoral neck and acetabulum can occur at extreme joint positions in hips with normal morphology [1,125,132]. 2.4 Measuring Hip Contact Mechanics in Cam FAI There is very little biomechanics research to date that provides a direct experimental measure of contact mechanics during the impingement event. Two key methods, imaging in functional positions and direct mechanical measurement, have seen limited used in the exploration of cam FAI. In light of the need for further experimental research about impingement-related hip joint mechanics, this subsection considers imaging and direct mechanical measurement methods for their potential to elucidate more information about cam impingement-related contact mechanics. 2.4.1 Direct Imaging in Functional Positions Direct imaging has the advantage of providing non-invasive, direct assessment and can therefore be used in vivo to study the relationship between the cam deformity and acetabular cartilage or labrum.  MRI is one modality that has been used widely in other joints for quantification of joint contact area and location during weight-bearing because it can image soft tissue and does not require radiation exposure to the patient. MRI has also recently been used to measure cartilage strain [133–137]. While MRI has been used frequently in neutral hip positions, it has only once been applied once to hips in postures that induce impingement [132]. Yamamura et al. used open MRI at 0.5T to evaluate interaction of the femoral head with the acetabulum in extreme postures in 5 control subjects [132]. Impingement was thought to have been observed in the anterosuperior quadrant in all subjects (Figure 2.8) specifically during the ‘W-sitting’ position. The ‘W-sitting’ position induced on average 91.3° flexion, 1.0° adduction, and 37.2° internal rotation. Impingement Chapter 2: Literature Review 25  was positive if less than 3 mm of separation between the femoral neck and acetabulum was visible [132]. A key strength of this method compared to other methods was that no registration step is required, and that direct observation of joint contact could be made in situ, requiring no simulations or simplifications of hip physiology. However, it is not clear whether the images show true contact between femoral neck and acetabulum, or mere adjacency, a key limitation of imaging methods delineating visible contact.  Figure 2.8: An open MRI image of a normal hip in the W-sitting position. There is visible impingement between the anterior femoral head-neck junction and acetabular rim (circled). © Reproduced from [132] with permission from John Wiley and Sons. The study by Wyss et al. that first described the β-angle used open MRI to image hips in 90° of flexion [9], but did not acquire images of hips in the impingement position and stopped short of investigating impingement-related contact mechanics. 2.4.2 Direct Mechanical Measurements Hip joint mechanics have been studied ex vivo by instrumenting various tissues (usually cartilage, labrum, or bone) in cadaver hips, then subjecting the hips to simulated movement and loading. Direct measurement of joint contact mechanics have been made in vivo using instrumented prostheses [138], but instrumented prostheses cannot inform us about the native hip, cam or control. The standard for direct measurement of mechanics remains to be ex vivo studies. The key advantage of ex vivo studies is that many mechanical Chapter 2: Literature Review 26  parameters of interest – such as contact pressure, contact area, resultant force, force distribution, labral strain, and joint space distribution – can be measured directly. A central limitation to most direct mechanical measurement methods is that the joint must be compromised in order to instrument the tissue of interest; investigators often sacrifice ligamentous structures and/or the labral seal. Also, simulating muscular activity can be technically challenging, so the simulated loading scenario is often a dramatic simplification of physiological hip loading. For this reason, direct mechanical measurements are sometimes coupled with non-invasive methods like imaging or modelling to provide mechanical context to the measures made in imaging or modelling [32,33]. In one study, differential variable-reluctance transducers were implanted into the labrum to measure labral strain during a wide range of passive clinical hip exams, including the anterior impingement exam. When hips were positioned in the anterior impingement exam, strain in the anterosuperior portion of the labrum was approximately 2% higher than when hips were supine [76]. It was not clear if locations of high strain related to locations of bone-labrum or cartilage-labrum contact since contact force/area/distribution was not measured. Pressure sensitive film has been widely used to measure contact area, pressure, and force distribution in the hip joint in various postures and loading conditions, notably during simulated walking [33,34], simulated side-falling in high flexion (lateral loading through the greater trochanter) [35], or stair climbing [32,33]. Pressure sensitive film has also been used widely to assess the effects of pelvic trauma, soft-tissue damage, and corrective surgery on hip loading, usually in single-leg stance [36–42]. A key limitation of pressure sensitive film is that it is only valid for single time points – it cannot be used for testing that requires a continuous recording of changing contact pressures, which is one reason why pressure sensitive film has not been used to measure contact mechanics during simulated impingement scenarios. Miniature pressure transducers have been implanted directly into subchondral bone to measure stresses at the osteochondral boundary in a variety of postures and loading conditions. Piezoelectric ceramic transducers in acetabular subchondral bone (Figure 2.9) were used to study stresses during simulated Chapter 2: Literature Review 27  walking [43,44] and standing [45], while piezoresistive transducers in subchondral bone of the femoral head were used to study stresses during a loading pattern similar to the gait cycle [46]. One reason why this method is no longer widely used it that the implantation of transducers is highly technically challenging and time-consuming, and cannot be used to cover the entire surface due to the spatial constraints with the need to anchor transducers in underlying bone.  Figure 2.9: A schematic diagram of pressure transducers implanted into the acetabulum. Adams et al. used piezroresistive tranducers to measure pressure at the osteochondral junction. © Reproduced from [43] with permission from BMJ Publishing Group Ltd. Casting has been used to determine joint space distribution during loading in a simulated walking cycle [139–141]. The concept is to load the hip joint, with a casting material in the joint space, holding the load statically until the casting material sets. The absence of any joint space can be used to measure contact between cartilage surfaces. This method has revealed that joint space size and shape is highly variable depending on joint position and anatomy [140,141]. However, the casting method is dramatically invasive (requiring expulsion of all synovial fluid, negating the labral seal), time consuming, and cannot be used for varying applied loads or varying joint position. In 2014, Rudert et al. described a recently developed piezoresistive sensor array (Tekscan K-Scan 4400; South Boston, MA, USA) with curvature designed specifically for use in the hip joint [47]. The Tekscan hip sensor showed qualitative agreement with pressure sensitive film (Fujifilm Prescale; Valhalla, NY, USA) in contact area measurement and force distribution during simulated walking and stair climbing [47]. This sensor has the potential for applications to simulated impingement scenarios. The Tekscan sensor is Chapter 2: Literature Review 28  also thinner (0.1 mm thickness) than most Fujifilm Prescale (0.3mm thickness), meaning that it will interfere less with natural joint contact mechanics compared to pressure sensitive film [142]. Before 2014, piezoresistive sensor arrays had not been used in the hip joint because sensors were rectangular, and therefore were susceptible to crinkling in the highly curved hip joint space. In other joints, piezoresistive sensors are commonly used to measure contact area and force distribution, but have shown poor precision in force magnitude (variability between trials over 20% in the patellofemoral joint [142,143], 10% in the lumbar facet joint [144]). The strength of piezoresistive sensors in previous studies has been the ability to make measurements of contact area that are repeatable [142] and more accurate compared to pressure sensitive film [145]. One study that used piezoresistive sensors in the patellofemoral joint reported that the repeatability of contact area measurement ranged from 0.8 to 8.3% (3.0% mean, six repeat trials of continuous knee flexion from 20° to 90°), whereas the repeatability of force measure ranged from 1.9 to 16.2% (7.3% mean). Another study compared the use of a piezoresistive sensor (Tekscan K-Scan 4000; Boston, USA) with various pressure sensitive films (Fujifilm Prescale; Valhalla, NY, USA) for measuring contact forces in a total knee arthroplasty implant that was cemented to a Sawbones® knee joint, as well as in a phantom with known contact area during axial loading [145]. At the lowest load (1000N) in the phantom, when compared to the known contact area, the Tekscan sensor underestimated contact area by only 5%, whereas the Fujifilm Prescale underestimated the true contact area by 45% [145].  Capacitive sensor arrays offer an alternative to piezoresistive arrays. One studied compared the output of a capacitive sensor (Novel AJP; Munich, Germany) to a piezoresistive sensor (Tekscan I-Scan 5051; South Boston, MA, USA) during loading axial in phantoms of various shapes using a materials testing machine, and found that the capacitive sensor was more accurate for measurement of contact force: error ranged from -3 to 5%, compared to -12% to 20% error in the piezoresistive sensor [146]. However, the key limitation to capacitive sensors is the thickness – usually 1 mm or greater compared to 0.1 mm for most piezoresistive Chapter 2: Literature Review 29  sensors [146] – which means capacitive sensors will interfere more with natural hip contact mechanics, and also will be nearly impossible to use in a highly curved joint like the hip. 2.5 Summary  It is hypothesized that cam FAI causes hip OA. Up to 39.5% of primary hip OA cases are now thought to be secondary to a cam deformity [97].  The presence of a cam deformity alone does not predict progression of hip OA. As many as 10-74% of healthy, active, asymptomatic people have cam deformities and will never experience rapid OA progression.   There is no direct evidence that mechanical impingement occurs in patients with a cam deformity, despite the central role of mechanics in the hypothesis that FAI causes hip OA.  Imaging has the potential to provide a direct, visible description of the impingement event in vivo – non-invasively and without requiring simulations or simplifications of hip physiology – a major improvement on previous approaches that use computer models. Imaging is very widely used to describe the severity of the cam deformity but to date has not been used to visualize the impingement event directly in patients with a cam deformity because impingement postures cannot be imaged in traditional scanners. Open MRI makes it possible to image the impingement event, which represents a key research opportunity and direction for this thesis.  Recently, a piezoresistive sensor designed specifically for the hip has made it simpler to measure contact mechanics ex vivo and permits continuous force measurement during a passive impingement test. Yet piezoresistive sensors have shown poor accuracy for measuring resultant force magnitudes and require disruption of the intact hip, so may be limited to use as supporting data for other non-invasive techniques like imaging.  Imaging measures of FAI have not been linked to direct measurements of hip mechanics.  Chapter 3: Methods A  30  3 Methods A: Quantifying Cam FAI using Open MRI 3.1 Introduction The first aim of this thesis (Aim A) was to develop a method for investigating and quantifying cam impingement using open MRI. Magnetic resonance imaging of hips in impingement postures has strong potential to provide insight into the mechanics of FAI. In particular, MRI has the ability to non-invasively investigate impingement in vivo, whereas methods for making direct mechanical measures in hips in vivo are far too invasive to be ethical. Imaging also has the ability to directly visualize native hip structures, whereas computer models predicting impingement location [22–28] or finite element models predicting joint contact pressure [29–31] require many assumptions and simplifications about hip physiology (i.e. they ignore synovial fluid, or assume that the labrum is rigid). Traditionally, scanning hips in impingement postures has been impossible because common MRI scanners have a cylindrical bore that limits patients to the supine position. Recently, the advent of open MRI with an upright configuration made it possible to image hips in a variety of postures. To date, one open MRI study has imaged control hips in high-flexion postures suspected of causing impingement [132], while another open MRI study imaged hips with cam and pincer deformities in 90° flexion (without attempting to induce impingement) [9], but open MRI has never been applied to cam hips in impingement positions. In order to use an open MR imaging approach for in vivo studies of FAI, we must first establish whether we can directly observe and quantify cam FAI using MRI. Therefore, the objectives of this section were to: 1. Describe image acquisition methods for open MRI of hips in an impingement position, image processing methods to best visualize cam impingement, and a measure that can be used to quantify cam impingement (as well as pincer impingement, secondarily); and 2. Evaluate the repeatability of the impingement quantification method. Chapter 3: Methods A  31  3.2 Methods 3.2.1 Protocol Development Image Acquisition Images were acquired using 0.5T upright open MRI scanner (MROpen, Paramed; Genoa, Italy) that was available at our research facility (Centre for Hip Health and Mobility; Vancouver, BC, Canada). The scanner has a high-temperature superconducting magnet with a primary magnetic field that runs normal to the vertical bore opening (Figure 3.1). While a 0.5T magnet limits the possible signal and subsequent signal-to-noise ratio during imaging, the open configuration of the MROpen has significant advantages because it permits the positioning of patients in functional, weigh-bearing postures that are impossible in traditional closed-bore scanners.  Figure 3.1: The 0.5T MROpen (Paramed; Genoa, Italy). The primary magnetic field direction, 𝑩𝟎⃑⃑⃑⃑  ⃑, is normal to the upright bore opening, as shown. In order to optimize the image acquisition protocol for visualizing the relationship between the femoral head-neck junction and acetabular margin, I conducted a protocol development process with an experienced 𝑩𝟎⃑⃑ ⃑⃑ ⃑⃑   Chapter 3: Methods A  32  MR physicist. Development of the acquisition protocol required four pilot scanning sessions. In general, protocol development is an iterative process that begins with requirements for the output images, then requires the MR physicist to vary multiple imaging parameters while making trade-offs that affect the ‘quality’ of the output image. For example, echo time (TE) and repetition time (TR) can be varied to change the contrast of the image (i.e. whether it is T1-weighted, T2-weighted, or proton-density weighted). The number of repetitions, n, can be increased to improve signal-to-noise ratio (SNR) by a factor of √𝑛, but more repetitions will increase scan time. The field of view (FOV) can be varied both in-plane and in the normal direction, with a larger FOV and thicker slices providing more signal for the image but decreasing the resolution, given the same matrix size and number of slices. The matrix size can be increased for finer image resolution, but finer resolution results in decreased SNR. The following requirements were used to define an ‘acceptable output image’ for this study:  Isotropic resolution to permit post-acquisition multi-planar reformatting (as done in previously published methods used for hip MRI [95,147]);  No slice gap, so that all tissue was included in the image; and  ‘Adequate’ signal-to-noise ratio, as judged qualitatively by an experienced radiologist, to distinguish between two structures that might be in contact, particularly between the femur and acetabulum at the location of impingement. There was no risk of motion artefact in the cadaveric hips, so it was not important to minimize time except for efficiency and cost. The finalized imaging sequence was a 3D gradient echo sequence with no slice gap and sub-millimetre isotropic resolution (Table 3.1). TE and TR were minimized to create a T1-weighted sequence that could best visualize the bony structures. Pilot images suggested that the margin of the femoral head-neck junction would induce impingement against the acetabular side of the hip in both cam and pincer impingement cases. Therefore, since the head-neck junction is bony rather than cartilaginous, a T1-weighted image would provide the best contrast to visualize the head-neck margin. Chapter 3: Methods A  33   Table 3.1: MRI sequence parameters. Parameter Value Matrix size 256 x 256 FOV 28.5cm x 28.5cm Slices 102 Slice thickness 0.90mm (no gap) TE/TR 8ms/19ms (T1-weighted) n repetitions 4 Flip angle 45° Scan time 27m 19s Output resolution 0.9 x 0.9 x 0.9 mm3  Before scanning, a one-channel send-receive coil was placed around the cadaver hip (Figure 3.2). The coil was chosen because it had the smallest possible radius to fit around the pilot hemi-pelvis (with all soft tissue intact). It was critical to minimize the amount of ‘free-space’ within the coil in order to optimize SNR (because free-air supplies noise but not signal). Some specimens were larger than the pilot specimen and required some dissection of skin and subcutaneous fat for the coil to be positioned around the hip joint centre. The hip joint centre was positioned at the magnetic isocentre during imaging.  Figure 3.2: The scan coil wrapped in protective plastic for biosafety. Chapter 3: Methods A  34   For each scan, the acquisition plane was defined by the femoral neck axis and a vector pointing anteriorly (i.e. normal to the coronal plane). After MR image acquisition, there were 102 image intensity files (one for each slice) exported for processing. Image Processing Each image volume was reconstructed into 36 radial slices about the femoral neck axis using multi-planar reformatting (MPR) to visualize the clearance between the femur and acetabulum about the entire circumference of the femoral neck and capture the three-dimensional nature of impingement. In MPR, a rectangular image volume is transformed into a pack of slices that revolve around a common vector at angular increments. Each radial slice provides a different cross-section through the centre of the femoral neck (Figure 3.3). With 36 slices through the femoral neck, there are 72 visible locations around the circumference of the femoral neck that can be used to evaluate clearance with the acetabular margin, which provides more information than a simple two-dimensional image where only two locations on the femoral neck are visible.  The first step of the MPR process was to define the femoral neck axis. Each image set was imported into image processing software (Mimics 15.0, Materialise; Leuven, Belgium). The proximal femur was segmented using a 3D semi-automated gradient-based spline method, and then the proximal femur surface was rendered by creating a stereolithography file. The femoral neck portion of the proximal femur was manually delineated (3-Matic, Materialise; Leuven, Belgium) by marking the region on the femur that spanned the femoral head-neck junction (proximally) to the inter-trochanteric ridge (distally). A cylinder was fit to the femoral neck (non-negative linear least-squares method) to calculate the longitudinal femoral neck axis (MATLAB 2012a, MathWorks; Natick, MA, USA). Starting with 0° as the anterior plane, I revolved planes about the neck axis at 5° increments (spanning 180° to avoid duplicates). The 36 new radial slices (spaced 5° apart, intersecting at the femoral neck axis, with a resultant voxel size of 0.9 x 0.9 x 0.9mm) were created from the original 3D image volume using an affine Chapter 3: Methods A  35  transformation with trilinear interpolation (MATLAB 2012a, MathWorks; Natick, MA, USA). Notably, the MPR process means that there was significant overlap of voxels near the femoral neck axis between adjacent radial slices. At the femoral neck margin, given an average male femoral head diameter of 48.4 mm [148] and a radial slice spacing of 5°, the centre to centre distance between voxels on adjacent slices was approximately 2 mm.  Figure 3.3: A view of the proximal femur showing multi-planar reformatting. The view shown is directed with the femoral neck axis normal to the page. The 0/180° plane points anteriorly while the 90/-90° plane points superiorly. In total, each image was reformatted into 36 radial slice planes separated by 5° each (not shown - only 30° increments are shown). Note that the medial-lateral direction is approximately directed into the page, but is not exactly normal to the page and dependent on femoral version. Quantifying Impingement The β-angle, which describes the relationship between the acetabular margin and femoral head-neck junction, was selected to quantify cam impingement (refer to Section 2.2.3; Figure 3.4) [9]. In its first description, the β-angle was measured on images of hips in 90° of flexion, neutral adduction, and neutral internal rotation [9], but the β-angle has never before been reported in an impingement position. After Chapter 3: Methods A  36  reviewing pilot images of hips in the anterior impingement test position, it was apparent that a negative β-angle represented intrusion of the cam deformity into the intra-articular joint space – medial to the acetabular rim – as suggested in the early description of cam impingement by Ganz et al. [1] (Figure 3.5). As such, the β-angle is a measure for ‘cam intrusion’ (i.e. angular depth of cam projection into the intra-articular joint space). Measurement of β-angle on multiple slices provides an ‘intrusion profile’ about the circumference of the entire femoral head-neck junction.   Figure 3.4: The β-angle defined by Wyss et al. (Top) The β-angle as shown on an MRI slice. (Bottom) The first line connects C, the first deviation of the femur from a femoral head circle, to hc, the femoral head centre, The second line connects D, the acetabular bony margin, to hc. The angle between the two lines is defined as ‘β’. © Reproduced from [9] with permission from Wolters Kluwer Health. Chapter 3: Methods A  37   Figure 3.5: A pilot image of a cam hip demonstrating negative β-angle and cam intrusion. The dotted line, connecting the head-neck junction to femoral head centre, lies medial to the solid line connecting the head centre to the acetabular margin, resulting in a negative β-angle that is indicative of cam intrusion. To measure the β-angle, image readers followed a sequence of steps (Figure 3.6) in order to specify: 1. A circle that fits the femoral head; 2. The point at which the femoral head-neck junction begins to deviate from the circle; and  3. The lateral-most bony margin of the acetabulum. Using custom code in MATLAB, each radial slice appeared for readers on a large digitizing tablet and the reader could manually select eight points on the circular profile of the femoral head using a stylus. Based on the eight points, a circle was fit to the femoral head and displayed on the image along with the upper 98% confidence interval (CI) for the circle fit, as well as the circle’s centre. The reader could then visualize the femoral head-neck profile on the image after the mean circle and upper 98% confidence interval had been superimposed. The reader was instructed to select the point at which the femoral head-neck junction deviated from the 98% CI, as well as to select the lateral-most point on the bony acetabular margin. The custom MATLAB code then calculated the resulting β-angle automatically.  β Chapter 3: Methods A  38   Figure 3.6: Sequence of steps performed by readers measuring β-angle. A) For each image, the reader picked eight points around the femoral head to define a circle. B) Based on the eight selected points, a circle was automatically fit (solid green line) to the femoral head using a non-negative least-squares fitting method. The eight points were used to calculate a 98% confidence interval (blue dashed line). C) The reader then used the 98% CI to select the femoral head neck junction, both anteriorly and posteriorly, and the two bony margins of the acetabulum. The presence of pincer impingement was defined using a binary contact scale. Image reading was performed manually – on each slice, the reader would delineate the locations where no visible space between the bony rim of the acetabulum and femur were present. This method had been used in the only open MRI that has imaged hips in postures suspected of causing impingement [132], and other studies that used a binary delineation method to measure contact area [149,150]. 3.2.2 Repeatability There were 12 cadaver hips acquired for study (Appendix A Specimen Screening). Hips were frozen and stored at -18°C before use. Each specimen included all soft-tissue, hemi-pelvis and proximal femur to mid-shaft. Half of the sacrum was included (sectioned sagitally at the midline). After thawing, each specimen was dissected of skin and subcutaneous fat. The hip joint capsule, surrounding ligaments, and muscular attachments were left intact. MR images of all 12 cadaver hips in a simulated anterior impingement posture were acquired using the above MRI protocol. Two experienced readers quantified impingement independently for all 12 images (each with 36 radial slices, and 72 corresponding locations about the circumference of the femoral neck) Chapter 3: Methods A  39  using both the β-angle (cam impingement) and binary pincer delineation scale (pincer impingement). Each reader repeated the measurements twice.  For β-angle reading, a two-way random average measures absolute agreement intra-class correlation coefficient ICC (2,k) was used to evaluate both inter- and intra-rater reliability (SPSS Statistics, IBM; Armonk, NY, USA) for all slices (i.e. ‘slice-wise β measurement’, at 72 β-angle locations per image). Inter-rater reliability was calculated using each reader’s mean reading from each of the two trials. The root-mean-square (RMS) average standard deviation (SD) between all four readings was also calculated for slice-wise β-angle measuement. The ICC (2,k) and RMS average SD were calculated again for the minimum β-angle, βmin, per hip (i.e. the minimum β-angle from all 72 locations per hip). For binary pincer delineation, intra-reader agreement was evaluated by only looking at a single reader’s results, and inter-reader agreement was evaluated by using an individual reader’s consensus result from the two trials. In cases where a reader’s two ratings disagreed, the reader was asked to review that slice a third time for a final decision. Since it was expected that a high proportion of the readings would be negative, a large degree of agreement would be expected simply by chance. Therefore it was important to characterize the positive agreement probability. Positive agreement, ppos, is the number of positive readings that both readers agree on, divided by all of the positive readings given by both readers [151]. Cohen’s κ was used to adjust for positive agreement by chance. The index κ was calculated by subtracting the proportion of agreement expected by chance, called pe, from the overall agreement, po, all divided by the proportion of agreement not expected by chance (1 – pe): 𝜅 =𝑝𝑜 − 𝑝𝑒1 − 𝑝𝑒 The 95% CIs for κ were calculated and interpreted according to the guidelines for strength of agreement by Landis and Koch [152]. Chapter 3: Methods A  40  3.3 Results For slice-wise β-angle measurement, Reader A had a two-way random average measures absolute agreement ICC (2,k) of 0.987 based on 864 total β readings. Reader B had an ICC (2,k) of 0.961. Overall agreement between the two readers, using the average from each reader, was 0.934. The RMS average SD was 9.7° for slice-wise β-angle measurement. For β-angle between -10° and 10°, the RMS average SD was 5.7°. For βmin measurement, the ICC (2,k) was 0.940 for Reader A and 0.875 for Reader B. The inter-reader ICC (2,k), based on each reader’s average βmin from two readings, was 0.988. The RMS average SD for βmin (from all four readings) was 3.5°. For pincer measurement, reader A had a ppos of 0.89, while κ was 0.89 (95% CI: 0.87 – 0.91)(Table 3.2). According to the guidelines by Landis and Koch, κ greater than 0.81 can be considered ‘almost perfect’ agreement [152]. Reader B had a ppos of 0.89, and κ was 0.90 (95% CI: 0.87 – 0.92) (Table 3.3). For inter-reader agreement in pincer delineation, ppos was 0.84, and κ was 0.80 (95% CI: 0.75 – 0.85) (Table 3.4).  Table 3.2: Judgment using binary pincer scale between two trials by Reader A.  Trial 2 Totals Positive Negative Trial 1 Positive 34 4 38 Negative 4 822 826 Totals 38 826 864   Table 3.3: Joint judgment using binary pincer scale between two trials by Reader B.   Trial 2 Totals Positive Negative Trial 1 Positive 29 3 32 Negative 4 828 832  Totals 33 831 864  Chapter 3: Methods A  41  Table 3.4: Joint judgment using binary pincer scale between Reader A and Reader B.   Reader B Totals Positive Negative Reader A Positive 26 8 34 Negative 2 828 830  Totals 28 836 864  3.4 Discussion In this chapter, I addressed Aim A of this thesis by presenting a novel method for acquiring images of cadaver hips in an impingement position with a quantifiable measure to represent cam intrusion, the β-angle. The β-angle was measured by two readers with high inter- and intra-observer repeatability: for slice-wise β measurement ICC was greater than 0.97 for both readers individually (between two repeat trials) and 0.93 between readers, and for βmin the ICC was greater 0.87. These values are consistent with the 0.90 to 0.93 ICC range reported by one MRI study (two readers measuring both α- and β-angle) [95] and the 0.86 to 0.97 range reported by a radiographic study (four readers measuring both α- and β-angle) [153]. The RMS average SD between readings was 3.5° for βmin, which was significantly lower than the RMS average SD from slice-wise β-angle measurement of 9.7°. Notably, the RMS average SD for slice-wise measurement was 4° lower when looking only at β-angles between -10° and 10°, indicating that β-angle error is higher when the magnitude of β-angle is large. The fact that error was smaller near β=0° is important clinically because it suggests that readers have better agreement near the critical point of cam intrusion. Furthermore, reliable measurement of βmin is arguably more important than reliable measurement of slice-wise β-angle because βmin defines the presence/absence of intrusion of the cam deformity into the intra-articular joint space. The RMS average SD in βmin measurement, 3.5°, is within the range of error reported in the study that first used β-angle [9], and the inter-reader ICC of 0.988 indicates that reader-specific average measures produce βmin agreement between readers at the upper limits for what can be expected. A key strength of this method was that the β-angle provided a simple and direct measure compared to complex automated methods while still maintaining repeatability equivalent to previous clinical studies. Chapter 3: Methods A  42  Direct image-based measures are used commonly in practice by radiologists and surgeons, and measurements of interest can guide clinicians to look at certain images in more depth. For example, a negative β-angle might indicate to a clinician that a particular image warrants further investigation for other qualitative observations that might be related to cam impingement. Since the outputs of the image reading method showed equivalent or higher repeatability compared to previous studies [95,153] (β-angle measurement ICC > 0.9), it was not necessary to pursue automated quantification methods that have been used in other studies for detecting cam deformities  [8,69,154,155] but are computationally intensive. Another key strength of this method was that the 3D image acquisition, sub-millimeter isotropic resolution, and multi-planar reformatting (MPR) permitted visualization of 72 locations around the entire circumference of the femoral head-neck junction – and therefore 72 possible sites of cam impingement. Past studies evaluating cam deformity size have shown that there is a risk of underestimating the size of the cam deformity if only a single two-dimensional plane is assessed [4,7,156], so it was important to visualize the femoral head-neck junction in multiple planes to avoid underestimating the extent of cam intrusion. Furthermore, by using MPR to generate 36 radial slices (72 head-neck locations) revolved around the femoral neck axis, our method provides more information than previous MPR studies that only generated 5 to 7 slices (10 to 14 head-neck locations) [4,95,132]. Ideally an MR acquisition would directly generate radial slices without requiring post-processing, but such acquisition sequences were not available for use on the 0.5T MROpen. The final MPR protocol employed an affine transformation with a trilinear interpolation, which might have introduced into the intensity of some voxels. This may be pertinent when manually defining tissue boundaries. Qualitatively, when comparing linear, nearest-neighbor, and cubic spline interpolation methods on pilot slices, there was no visible loss of information or differences between methods, and the benefits of MPR outweighed the potential risk of interpolation error. A secondary aim of this chapter was to present a measure for pincer impingement as well, for descriptive purposes, and also quantify its repeatability. The two readers used a binary pincer rating to evaluate the presence of pincer abutment at each head-neck location; a similar binary measure had been used in the only Chapter 3: Methods A  43  previous open MRI study looking at hips in suspected impingement postures [132]. Both the positive prediction probability and Cohen’s κ were 0.80 or higher (Reader A: ppos = 0.89, κ = 0.89; Reader B: ppos = 0.89, κ = 0.90; inter-reader: ppos = 0.84, κ = 0.80) and classified as ‘excellent’ inter- and intra-observer agreement according to the guidelines for interpretation by Landis and Koch [152]. However, an inter-reader positive prediction probability (ppos) of 0.84 indicates that the two readers disagreed 16% of the time for slices where at least one of the readers observed the presence of pincer abutment. This suggests that the images may not have been clear enough for readers to confidently decide whether mechanical contact was present. The ‘partial volume effect’ and resulting ‘grey’ voxels at the potential impingement zone may have contributed to disagreement in pincer rating between readers. The partial volume effect means that at the site of pincer abutment, tissue from both the femur and acetabulum, as well as cartilage, labrum, and synovial fluid, are contributing signal to a single voxel which results in a ‘grey’ intensity that is not representative of any one tissue.  Therefore, an observer who delineates ‘absence of joint space’ may not be selecting locations of true mechanical contact (‘impingement’) – they may be locating two merely adjacent structures separated by an imperceptible distance (i.e. less than a single voxel). For this reason, the binary pincer abutment scale should be used with caution. Another potential cause for disagreement between readers in binary pincer abutment rating was that the MRI sequence did not permit visualization of the labrum, which is a limitation of the imaging method, although this limitation did not affect the β-angle measurement. Without visualizing the labrum, it was difficult to conclude whether pincer abutment was present on some slices. Qualitatively, some slices had a clear bone-bone contact, but in some slices the difference between joint space and bone-bone contact was unclear and appeared as grey regions. While the grey regions may represent an invisible labrum being compressed by the femoral neck, we cannot conclude that the labrum is being compressed if it cannot be visualized. Since the labrum is a small structure and provides little MR signal, it is difficult to see on non-contrast MR images [157], so most MRI techniques evaluating the labrum employ either contrast injection Chapter 3: Methods A  44  [158,159] or high field strength (7T or greater) MRI [160,161]. Since this research was only possible with open MRI at 0.5T field strength due to spatial constraints, the only available option for visualizing the labrum is to use contrast-injection. We decided that future in vivo work in a large patient cohort would not be able to ethically employ contrast injection since it is very uncomfortable for patients, particularly patients with existing hip pain, so this cadaveric study pursued an imaging protocol that could not visualize the labrum. 3.5 Summary  The novel methods for MRI-based quantification of impingement were able to produce images with sub-millimetre isotropic resolution.  Multi-planar reformatting made it possible to visualize 72 locations around the circumference of the neck, representing 72 possible sites of cam impingement.  The β-angle, a measure representing the extent of cam intrusion into the intra-articular joint space, was repeatable between two readers who made two readings each (ICC > 0.9; better or equivalent to previous studies measuring β-angle).  A binary pincer abutment scale was used as a measure for pincer impingement, similar to one previous open MRI study, with ‘excellent’ repeatability (κ > 0.80) according to the guidelines by Landis and Koch. However, the inter-reader positive prediction probability was only 0.84, indicating frequent disagreement between readers. Due to the partial volume effect and lack of labrum visibility, the pincer abutment measures should be used with caution.   Chapter 4: Methods B 45  4 Methods B: Positioning Hips in an Impingement Posture  4.1 Introduction The second aim of this thesis (Aim B) was to develop a method for positioning hips in a simulated anterior impingement posture that approximates the end-point of the anterior impingement test. Since it is not possible to simultaneously acquire MR images and take sensor-based measurements of contact force in the hip joint, and since variability in joint position can have significant implications on the clearance between the cam deformity and acetabular margin [26], this work requires a standardized joint positioning technique with known precision. It is important to evaluate the error in joint angles and translations between repeat positioning trials in the same specimen so that we can address Question 3 of this thesis (are MRI measures of cam intrusion related to acetabular contact force in a simulated anterior impingement posture?) without requiring a simultaneous sensor measurement and MRI measurement.  Therefore, the objectives of this section were to: 1. Describe the design of a mechanical rig and positioning protocol that could rigidly hold a cadaver hip in a position approximating the end-point of the anterior impingement test; 2. Evaluate the repeatability of the positioning method in terms of joint rotations and translations; and 3. Evaluate how accurately hips can be positioned in the desired position of 90° flexion and neutral adduction. Chapter 4: Methods B 46  4.2 Methods 4.2.1 Protocol Development Rig Description The mechanical rig design process began by establishing design requirements. Before constructing the rig, it was specified that the rig must:  Rigidly hold hips in 90° flexion, 0° adduction, and maximum internal rotation;  Accommodate hips of different shapes and sizes;  Allow for ordered application of rotations, to approximate the clinical anterior impingement test [11,72]; and  Be constructed from MRI-compatible materials. The rig was bench-top prototyped and then constructed from acrylic, polycarbonate, and nylon hardware (Figure 4.1). The rig fixed the pelvis rigidly in a supine position and allowed hip joint angles to be changed by moving the femur in space. The key components of the rig were: a horizontal ‘base plate’, a mobile ‘femur-gripper’, a ‘flexion-arm’, and an ‘adduction hand’. Chapter 4: Methods B 47   Figure 4.1: Components of the positioning rig. The main components included: a horizontal ‘base plate’ to fix the pelvis, a ‘femur-gripper’ to hold the femoral shaft, a ‘flexion arm’ to induce hip flexion, and an adduction hand to connect the flexion arm to the femur gripper. The components formed a kinematic chain, with three one-degree of freedom (DOF) rotation joints at the junction of each component, to facilitate movement of the femur-gripper. The rotational axes of each joint intersected at a single, fixed point defined as the ‘origin’ of the rig’s coordinates. The joints included:  A ‘flexion pivot’, permitting rotation of the flexion arm relative to the base construct;  An ‘adduction slider’, permitting sliding of the adduction hand relative to the flexion arm in an arc trajectory centred at the origin; and  An ‘internal rotation pivot’, permitting rotation of the femur gripper relative to the adduction hand. The base plate, holding the pelvis, could be translated along three axes and then locked in place to ensure that the hip joint centre was positioned to coincide with the rig’s origin. Base plate translation was achieved Base plate Femur gripper Flexion arm Adduction hand Chapter 4: Methods B 48  via two 1-DOF sliding mechanisms (with locks) for translation in the horizontal plane, and spacers that permit incremental vertical translation of the pelvis (Figure 4.2).  Figure 4.2: Joints of the positioning rig. The flexion pivot permits rotation of the flexion arm relative to the base construct. The adduction slider permits translation of the adduction hand relative to the flexion arm along an arc. The internal rotation pivot allows rotation of the femur gripper relative to the adduction hand. The three rotational axes intersect at the origin. A lateral-medial slider and superior-inferior slider permit translation of the base-plate in the horizontal plane. Spacers (not shown; 6mm thickness) can be added underneath the base plate to adjust vertical height of the base plate. Positioning Protocol The objective of the positioning protocol was to mount the pelvis onto the base plate such that the hip joint centre aligns with the rig origin, and the specimen’s anatomical planes align with the rig’s horizontal/vertical planes. The first step in the mounting procedure was to align the coronal pelvic plane to the rig’s horizontal plane. The coronal pelvic plane was approximated using the pubic tubercle and anterior superior iliac spine as had been done in previous studies [22,23]. The second step was to align the sagittal Translation sliders Internal rotation pivot Flexion pivot Adduction slider Origin Chapter 4: Methods B 49  pelvic plane to a vertical plane normal to the rig’s flexion pivot axis. The pelvic sagittal plane was approximated by using the cut-section through the sacrum that had been created by the specimen procurement company when splitting the pelvis into two hemi-pelvis specimens. When the pelvic planes and rig’s normal planes were aligned, the pelvis was fixed rigidly by anchoring the ischial tuberosity, sacrum, iliac crest, and pubic ramus to the base plate using zip-ties. The next step was to translate the pelvis and base plate in the horizontal plane using the two sliding mechanisms, and add/remove spacers in the vertical direction, until the estimated hip joint centre aligns with the rig origin. In the fourth step, the femoral shaft is potted in an acrylonitrile butadiene styrene (ABS) pipe using poly-methyl-methacrylate (PMMA) and rigidly connected to the femur gripper. After the pelvis was mounted to the rig, the rig three axes of rotation were coincident with the three axes of hip rotation according to a standard hip joint coordinate system (based on the floating axis convention [162] and consistent with the International Society of Biomechanics recommendation for standardizing joint coordinate systems [163]; Appendix B Hip Joint Coordinate System; Figure 4.3).  Figure 4.3: Illustration of the hip joint coordinate system under the floating axis convention. The fixed pelvic coordinate system (XYZ) and the fixed femoral coordinate system (xyz) are used to define the hip joint coordinate system (e1 e2 e3). © Reproduced from [163] with permission from Elsevier. Chapter 4: Methods B 50  Rotation of the flexion arm about the rig’s flexion pivot was equivalent to hip flexion about the e1 (flexion) axis. Rotation of the femur gripper relative to the adduction hand was equivalent to internal rotation of the hip about the e3 (long) axis. Sliding of the adduction hand along its arc relative to the femur arm approximates hip adduction about the e2 (floating) axis, the common normal, since the adduction hand’s trajectory always lies in a plane that is perpendicular to e1 and e3. At each joint, there were angular tick-mark indicators etched into the plastic so that the hip could be positioned to an estimated angle in a repeatable way, similar to a goniometer. To define the end-point of internal rotation, a torque wrench was used to apply an internal rotation torque of 14±1 Nm (Figure 4.4). The torque wrench was connected to a rigid bolt on the internal rotation gripper, coincident with the axis of internal rotation, and oriented such that the lever arm lay in the coronal plane and projected inferiorly (in a similar fashion to how the lower leg projects inferiorly during the anterior impingement test when both the knee and hip are at 90° flexion). A linear force was applied to the handle of the torque wrench, oriented orthogonally to the torque wrench, while the flexion and adduction rig joints were locked in place to avoid the femur moving in any direction other than in internal rotation. The value of 14 Nm was selected based on the finding of a previous study that pelvic kinematics are altered in the presence of a cam deformity beyond an applied internal rotation torque of 12 Nm [74]. Chapter 4: Methods B 51   Figure 4.4: A right hemi-pelvis with proximal femur positioned in a simulated anterior impingement posture. The star denotes the location of the hex bolt that was connected to the torque wrench. The torque wrench was aligned in the coronal plane, approximating the position of the lower leg during the clinical anterior impingement test. A force was applied to the handle of the torque wrench, oriented orthogonal to the torque wrench, while the flexion and adduction angles were held rigidly in place. 4.2.2 Repeatability and Accuracy There were six cadaver hips acquired for study (Appendix A Specimen Screening). Hips were frozen and stored at -18°C before use. Each specimen included all soft-tissue, hemi-pelvis and proximal femur to mid-shaft. Half of the sacrum was included (sectioned sagitally at the midline). After thawing, each specimen was dissected of skin and subcutaneous fat. The hip joint capsule, surrounding ligaments, and muscular attachments were left intact. To evaluate the repeatability and accuracy of the positioning protocol, I performed six positioning trials in each of the six cadaver hip specimens (36 total trials) in a laboratory setting outside of the MRI scanner.  Torque wrench Applied force Direction of applied internal rotation torque Cranial Caudal Chapter 4: Methods B 52   The objective of each trial was to position the hip at 90° flexion, neutral adduction, and maximum internal rotation. I measured the joint angles and translations relative to the supine position with an optical tracking method (Appendix C Joint Position Measurement) adapted from previous studies [75,124] using an infrared motion capture system (Optotrak, Northern Digital, Waterloo, Ontario; manufacturer specified marker accuracy of 0.1mm, spatial resolution of 0.01mm).  ‘Precision error’ was used as a measure for repeatability based on recommendations in a past study by Glüer et al. that presented a standardized approach for using precision errors to characterize repeatability [164]. In a study of repeated bone mineral density measurements, Gluer et al. recommended that precision errors be calculated using root-mean-square (RMS) averages of standard deviations (SD) of repeated measurements, with appropriate adjustment for the degrees of freedom in each measure. The study compared the values of precision error that had been calculated using different methods, and found that inadequate adjustment for degrees of freedom or calculation of precision error using an arithmetic mean (versus RMS average), caused underestimation of the true precision error by 41% and 25% respectively [164]. While the objective of the repeatability testing in the current thesis was to characterize the repeatability of the hip positioning method, rather than the repeatability of a diagnostic technique, the same statistical analysis still applied and therefore I followed the recommendation laid out by Gluer et al. for standardized reporting of precision error [164]. Repeatability of the hip joint positioning protocol was therefore defined as the precision error between the six repeated trials of joint angle/translation measurements for each individual specimen (irrespective of the desired values, which were 90° flexion, neutral adduction, and specimen-specific-maximum internal rotation). First, standard deviation from the six trials for each specimen (5 degrees of freedom) was calculated for each of flexion, adduction, internal rotation, and the three translations. Then, the overall precision error of the sample was given for each by the RMS average of the SD for all six subjects. This resulted in 30 total degrees of freedom for calculating precision error; Glüer et al. had shown in their Chapter 4: Methods B 53  simulation that 27 degrees of freedom were sufficient to establish precision errors within an upper 95% confidence limit that is within 30% of the mean precision error (a ‘reasonable’ guideline suggested by the authors; e.g. an upper 95% CI of 2.6° given a mean precision error of 2.0°) [164]. The 95% confidence interval for the true SD was calculated using an asymmetric chi-square (χ2) distribution. The accuracy of the positioning method was calculated to evaluate how closely the measured hip positions matched the desired ‘true’ hip position of 90° flexion, neutral adduction, and specimen-specific maximum internal rotation. Since the true value for maximum internal rotation was not known, nor the corresponding joint translations, internal rotation accuracy and joint translation accuracy was not reported. Flexion/adduction accuracy error was calculated for each specimen, Ej, using the general equation: 𝐸𝑗 = ∑(𝑥𝑖𝑗 − 𝑥𝑡𝑟𝑢𝑒)𝑛𝑗 − 1 Where xtrue was 90° for flexion, and 0° for adduction, nj was the number of trials per specimen (6 for each specimen), and xij was the measured value in degrees (either in flexion or adduction) for each trial i. Overall accuracy error was reported as the numerical mean for all specimens. 4.3 Results The overall mean joint angles (36 trials total; 6 specimens, 6 trials each) were 99.5° flexion, -10.2° external rotation, and 7.5° abduction. Overall precision error, represented by the RMS average standard deviation of the sample (95% CI in brackets), was 2.0° (1.6, 2.7) for flexion (Table 4.1), 1.3° (1.0, 1.7) for external rotation (Table 4.2), and 1.6° (1.3, 2.1) for abduction (Figure 4.3).    Chapter 4: Methods B 54  Table 4.1: Measured flexion angles in degrees (6 repeated positioning trials per hip). Hip # Trial 1 Trial 2 Trial 3 Trial 4 Trial 5 Trial 6 Mean, 𝒙?̅? SDj 5 103.9 105.4 107 103.8 102.9 103.1 104.4 1.6 7 89.9 86.3 91.3 90.7 94.7 91.5 90.7 2.7 8 114.6 115.2 118.4 119.9 118 121.5 117.9 2.7 9 93.9 93.9 93.8 93.5 91.3 90.5 92.8 1.5 11 93.8 96.1 92.3 95.8 94.3 96.8 94.9 1.7 12 94.1 97.4 94.5 96.1 97.2 97.3 96.1 1.5          OVERALL 99.5° 2.0°  Table 4.2: Measured external rotation angles in degrees (6 repeated positioning trials per hip). Hip # Trial 1 Trial 2 Trial 3 Trial 4 Trial 5 Trial 6 Mean, 𝒙?̅? SDj 5 -2.2 -4.6 -4.2 -5.7 -6.0 -6.6 -4.9 1.6 7 -5.6 -8.3 -9.0 -9.7 -8.1 -10.1 -8.5 1.6 8 -23.1 -22.6 -23 -23.2 -23.2 -22.1 -22.9 0.4 9 -3.4 -3.7 -4.4 -5.1 -5.6 -5.4 -4.6 0.9 11 -5.7 -6.3 -8.6 -8.2 -8.4 -8.4 -7.6 1.3 12 -10.3 -12 -13 -13.2 -13.2 -14.3 -12.7 1.4          OVERALL -10.2° 1.3°  Table 4.3: Measured abduction angles in degrees (6 repeated positioning trials per hip). Hip # Trial 1 Trial 2 Trial 3 Trial 4 Trial 5 Trial 6 Mean, 𝒙?̅? SDj 5 0.7 3.8 2.2 2.4 2.7 3.8 2.6 1.2 7 4.0 4.3 4.6 5.7 5.3 6.9 5.1 1.1 8 12.1 13.1 12.2 14.6 13.3 11.2 12.8 1.2 9 11.7 12.1 13.9 11.4 12.0 9.4 11.8 1.4 11 3.2 3.8 9.6 9.2 8.0 8.2 7.0 2.8 12 5.0 6.3 3.7 6.3 6.4 7.1 5.8 1.2          OVERALL 7.5° 1.6°  The overall mean translation of the femoral head (relative to the supine position) was 6.2 mm laterally, 4.6 mm anteriorly, and -2.4 mm superiorly. Overall precision error, represented by the RMS average SD of the sample (95% CI in brackets), was 4.6 mm (3.7, 6.2) for lateral translation (Table 4.4), 0.56 mm (0.44, 0.74) for anterior translation (Table 4.5), and 0.40 mm (0.32, 0.53) for superior translation (Table 4.6).  Chapter 4: Methods B 55  Table 4.4: Measured lateral translation (mm) of the femoral head (6 positioning trials per hip). Hip # Trial 1 Trial 2 Trial 3 Trial 4 Trial 5 Trial 6 Mean, 𝒙?̅? SDj 5 1.7 2.4 2.8 2.8 2.9 2.8 2.5 0.4 7 -0.5 -1.8 20.0 20.3 20.6 21.2 13.3 11.2 8 11.6 12.2 13.0 13.7 13.6 13.9 13.0 0.9 9 -1.0 -0.9 -0.5 -0.2 -0.4 -0.5 -0.6 0.3 11 5.2 6.9 6.7 7.0 7.3 7.5 6.8 0.8 12 0.7 2.0 2.0 2.3 2.6 3.0 2.1 0.8          OVERALL 6.2mm 4.6mm  Table 4.5: Measured anterior translation (mm) of the femoral head (6 positioning trials per hip). Hip # Trial 1 Trial 2 Trial 3 Trial 4 Trial 5 Trial 6 Mean, 𝒙?̅? SDj 5 3.4 4.7 5.3 5.1 5.2 5.4 4.8 0.74 7 5.9 6.0 6.4 6.4 6.5 6.3 6.3 0.2 8 10.6 10.4 11.3 11.4 11.4 11.9 11.2 0.6 9 -0.1 -0.1 0.0 0.5 0.1 0.1 0.1 0.2 11 1.9 2.9 1.8 3.2 2.8 3.7 2.7 0.7 12 1.7 2.7 2.6 2.9 3.0 3.4 2.7 0.6          OVERALL 4.6mm 0.6mm  Table 4.6: Measured superior translation (mm) of the femoral head (6 positioning trials per hip). Hip # Trial 1 Trial 2 Trial 3 Trial 4 Trial 5 Trial 6 Mean, 𝒙?̅? SDj 5 1.1 0.9 0.7 1.0 1.2 1.2 1.0 0.2 7 -3.6 -3.7 -4.6 -4.7 -5.2 -5.3 -4.5 0.7 8 -6.9 -7.2 -7.7 -8.3 -7.9 -8.2 -7.7 0.6 9 0.8 0.9 0.7 0.7 0.7 0.8 0.7 0.1 11 -2.4 -2.8 -2.9 -2.9 -2.9 -3.0 -2.8 0.2 12 -1.2 -0.9 -0.9 -0.8 -0.8 -0.6 -0.9 0.2          OVERALL -2.4mm 0.4mm  The flexion accuracy error was (average for all hips) 11.4° greater than the 90° target.  Standard deviation for flexion error was 11.1°, and resulted in 95% CIs for flexion error of -10.5 to 33.2°. Overall abduction accuracy error was on average 9.0° away from the 0° target. Standard deviation for abduction error was 4.3° and resulted in 95% CIs for abduction error of -0.5° to 17.5° (Table 4.7).   Chapter 4: Methods B 56  Table 4.7: Measured accuracy error in degrees for flexion and abduction. Hip # Flexion error (°) Abduction error (°) 5 17.2 3.1 7 0.9 6.2 8 33.5 15.3 9 3.4 14.1 11 5.8 8.4 12 7.3 7.0 OVERALL Mean [95% CI] 11.4 [-10.5, 33.2] 9.0 [0.5, 17.5] SD 11.1 4.3 4.4 Discussion In this investigation, I addressed aim B of this thesis by presenting a method for positioning hips in a posture that approximates the end-point of the anterior impingement test, and established: (i) the repeatability of the method between six repeated trials in six specimens (36 trials total); and (ii) determined the overall accuracy error with which hips could be positioned in the target position of 90° flexion and 0° abduction. The repeatability of the positioning method (precision error, upper 95% CI limits for RMS average standard deviation between all trials) was 2° or less for flexion, abduction, and external rotation, which is an order of magnitude smaller than the reported repeatability of goniometry [165], a common method for clinical ROM assessment. Specifically, one study reported that the standard deviation between repeat goniometer tests ranged from 4.0° in adduction to 15.7° in flexion [165] when measuring ROM in vivo. Furthermore, the study also compared electromagnetic tracking and goniometry, and found that the percentage difference between the two systems was 18% for flexion, 6% for abduction, 13% for adduction, 12% for internal rotation, and 20% for external rotation when measuring ROM in vivo [165]. Another electromagnetic tracking study found that the reported joint angles varied by 1.91° to 5.81° depending on whether bone-anchored (common for ex vivo studies) or skin-anchored markers (necessary for in vivo studies) were used [166]. These past studies indicate that the error due to measurement method alone is likely greater than the 2° precision error of this positioning method. Finally, the fact that I have quantified the repeatability of the Chapter 4: Methods B 57  positioning method at all is an improvement over FAI imaging studies that do not report any method for controlling joint position or do not quantify positioning repeatability [9,95,132,153].  The repeatability (precision error upper 95% CI limit) for both anterior and superior joint translation were under 1 mm, but lateral translation showed a higher repeatability error (upper 95% CI limit) of 6.2 mm, which may be due to impingement-related levering in one hip due to compromised soft tissue structures. Notably, the standard deviation for lateral translation was less than 1mm for all hips except one hip (hip #7) that had 11.2mm in lateral translation standard deviation between trials. In hip #7, the standard deviation in the other translations and joint rotations fell within the standard deviation range of the other five hips, so it is not clear why only lateral translation was affected. One explanation is that stabilizing structures in that specific hip may have failed or been damaged between tests, which could have caused the femoral head to experience exaggerated levering (causing lateral translation) in subsequent tests. Levering is hypothesized to be a consequence of impingement-related contact between the femoral neck and acetabular margin and has been described in one previous study [27]. The lateral translation was near 0 mm for the first two tests, but reached approximately 20 mm for the final four tests. In addition, that particular hip had signs of both cam and pincer FAI anatomy (Appendix A Specimen Screening). Damage to stabilizing structures such as the labral seal could have reduced the hip’s soft-tissue resistance against levering of the femoral head laterally. It is unlikely that the process of applying internal rotation torque through the handle of the torque wrench led to any lateral reaction force in the hip that could have caused variability in lateral translation. Application of translational force through the torque wrench handle should be counteracted equally and oppositely (anti-parallel) by the rigidly-held femur-gripper. The equal and anti-parallel reaction force should be applied at the femur gripper in a similar manner to how a clinician opposes their own manually applied force to generate internal rotation force (usually applied on the lower leg, near the ankle) with another manually applied force to keep the knee in a fixed location and hold the femur in 90° of flexion and neutral adduction (Appendix D Analysis of the Anterior Impingement Test). Equal and opposite Chapter 4: Methods B 58  translational reaction force at the rigidly-held femur gripper would mean that no additional translational forces should be present at the hip joint due to application of the eccentric torque wrench load. In flexion, the overall accuracy error (six hips, six trials each) was 11.4° greater than the 90° target on average, which may have been caused by misalignment of the pelvis into a supine position when mounting the pelvis on the rig’s base plate, but the range of flexion values reported here are within the range reported in two previous studies that placed hips in impingement positions [27,132]. One explanation for variation is that the orientation of the ASIS and pubic symphysis relative to the coronal plane may have varied from specimen to specimen, depending on anatomy, which could have caused between-specimen variability. Actual flexion values ranged from 90.7° to 117.9°. For comparison, one past open MRI study that positioned in vivo hips (n=5) in a ‘W-sitting’ position (closely resembling the anterior impingement test) reported a wider range of flexion angles during imaging from 66.4° to 108.9° [132]. Another study in which a clinician applied the anterior impingement test during dual fluoroscopy video reported that the actual flexion angles ranged from 86.2° to 107° [27]. Both studies were conducted in vivo and therefore the entire body was available for reference, which makes it easier to estimate global anatomical planes compared to cadavers specimens where only small portions of the body are available. In the context of other studies, the range of flexion values reported here (90.7° to 117.9°) does not seem unreasonable, but precludes the conclusion that this positioning method consistently and exactly replicates the clinical anterior impingement test.  In abduction, the overall accuracy error in abduction was 9° greater than the 0° target (range of actual abduction values: 2.6 to 12.8°), a range that is similarly wide but with more abduction compared to previous studies that placed hips in impingement positions [27,132]. The open MRI study that positioned hips in a ‘W-sitting’ position reported a range of abduction angles of 3.3° to -5.2° [132] (more neutral than the angles I have reported) while the dual fluoroscopy study of the clinical anterior impingement test reported a range of abduction angles from -2.2° to 14.4° [27] (more adduction than the angles I have reported). The implication of more abduction than adduction when testing maximum internal rotation at 90° flexion is not Chapter 4: Methods B 59  clear; many computer modelling studies conduct repeated impingement tests over a range of abduction angles (usually between -20° and 20°) to predict impingement zones without distinguishing the specific abduction angle between repeated tests [22,23,25], suggesting that the difference in abduction between 10° and -10°, when studying impingement mechanics, has not been clinically important to date. The 10° range in abduction values could have been caused by inaccurate alignment of the sagittal pelvic plane to the rig’s vertical plane that varied from specimen to specimen. Since estimation of the sagittal plane during pelvis mounting relied on the physical section that had been created in the specimen’s sacrum during procurement into a hemi-pelvis, error in the physically cut section may have led to error in sagittal plane estimation. Other potential sources of error include the registration between the optical tracking-derived digitized surfaces and the CT atlas (Appendix C Joint Position Measurement). There may have been inaccuracy in the manual surface digitization, which would manifest as registration error due to surface sets with imperfect point-to-point correspondence. Inaccurate landmark selection also may have led to registration error and subsequent error in the measured joint angles. As previously discussed, any methods that rely on using anatomical landmarks to generate coordinate systems will be subject to anatomical variation and potential error since relative distances/orientation between landmarks may vary from specimen to specimen. The precision error for internal rotation of 1.6° (95% CI: 1.3 to 2.1) likely has minimal implication with respect to cam intrusion because one modelling study predicted that cam intrusion occurs at internal rotation angles more than 10° before the terminal position of the anterior impingement test is reached [26]. Audenaert et al. studied cam hips (n=13) using clinical ROM measurement and a registered computer model, in the anterior impingement test and reported that the earliest cam deformity intrusion occurred at an internal rotation angle of 7.2 ± 10.4° (mean ± SD) whereas the end-point of the anterior impingement test was 18.9 ± 6.2° [26]. Given that the range of internal rotation angles that produce cam intrusion is greater than 10° wide, a precision error of 1.6° for internal rotation should not affect whether cam intrusion can be observed on open MRI nor should it affect whether joint contact is present or absent. Chapter 4: Methods B 60  4.5 Summary  A novel positioning method was developed, including the design of an MR-compatible rig that could rigidly hold hips of varying shapes and sizes in a simulated anterior impingement test posture.  The repeatability of the positioning method was 2° or less for flexion, abduction, and external rotation, an order of magnitude smaller than the reported repeatability of goniometry. The repeatability was less than 1 mm for anterior, lateral, and superior joint translation except for a single specimen with larger lateral translation precision error. Precision error of 2° provides a point of reference when comparing position trials for open MRI to positioning trials performed for sensor-based experiments.  A precision error of 1.6° for internal rotation should not affect whether cam intrusion can be observed on open MRI since the range of internal rotation angles that produce cam intrusion is greater than 10° wide (in the anterior impingement test).  This positioning method produced flexion angles that are within the ranges reported in two previous impingement studies. The final joint positions had more abduction than previous studies, but still within the abduction angle range of -20° to 20° commonly used for computer modelling studies that study impingement location. This positioning method is able to produce a simulate anterior impingement posture, but does not consistently replicate the exact clinical anterior impingement test.Chapter 5: Methods C 61  5 Methods C: Measuring Acetabular Contact Force in Cam FAI 5.1 Introduction The third aim of this thesis (Aim C) was to develop a method for directly measuring acetabular contact force in a simulated anterior impingement posture in order to provide mechanical context for the open MRI measures of cam intrusion. To date, no imaging or model-based studies that investigate the biomechanics of cam impingement have related their findings to mechanical force measures despite the central role of contact mechanics in the cam impingement hypothesis. Piezoresistive contact force sensors are a strong candidate for measuring contact force in the hip joint: they have been used before in the hip joint [47]; can be used to record continuously changing contact pressures; and are thinner than both pressure-sensitive film and capacitive sensors [47,144,167]. While piezoresistive sensors have been reported to make highly accurate and repeatable measurements of contact area and force distribution [142,144–146,168], past studies have reported coefficients of variation between 3.1 and 23.5% in the patellofemoral joint [142,143] indicating poor repeatability for resultant force measurement. To date, no studies have quantified the repeatability of piezoresistive force sensor measurements in the hip joint in simulated impingement postures, so it is not yet clear how appropriate piezoresistive sensors will be in assessing MRI-based measures of cam impingement. Therefore, the objectives of this section were to: 1. Describe methods for measuring the resultant force, force centroid location, and force distribution for each hip; and 2. Evaluate the repeatability between trials for measurements of resultant force, force centroid, and force distribution in multiple hips, in a simulated anterior impingement posture. Chapter 5: Methods C 62  5.2 Methods 5.2.1 Protocol Development Sensor Description Pressure was measured using a recently developed hip sensor (Tekscan K-Scan 4400; South Boston, MA, USA). The K-Scan 4400 hip sensor is a piezoresistive sensor, 0.10mm thick, rated for maximum pressure up to 20.0 MPa. In piezoresistive sensors, the application of a normal force to the sensor surface results in a change to the resistance of the sensing elements (sensels) inversely proportional to the applied force. Therefore by definition, resistive sensors measure force at each sensel location. When a resistive sensor is described as being ‘rated to 20.0’ MPa, it means that an applied pressure of 20.0 MPa or above in any area will produce a force that would cause the sensing elements output to saturate (i.e. reach a maximum output, beyond which the output cannot increase even if the force increases). The sensor forms an arc-shaped segment that spans 150° of an annulus. The dimensions for the inner and outer radius are 44.0 and 25.0 mm respectively. The sensor utilizes a radial ‘ring-and-spoke’ array layout in polar coordinates. The ‘spoke’ electrical leads run normal to the annulus, while the ‘ring’ electrical leads run radially along the annulus (Table 5.1, Figure 5.1).  Table 5.1: Tekscan K-Scan 4400 parameters (active sensing region only) [47]. Outer radius (mm) Inner radius (mm) Span (°) Rings Spokes Sensels Thickness (mm) Area (mm2) Sensel density (sensels/cm2) 44.0 25.0 150 20 52 1040 0.10 1716 61  Chapter 5: Methods C 63    Figure 5.1: The front and back of the Tekscan K-Scan 4400 contact force sensor. The ‘spoke’ electrical leads run normal to the arc, while the ‘ring’ electrical leads run radially along the arc forming an array in polar coordinates. (Front) The ‘ring’ electrical leads run on the front of the sensor. (Back) The ‘Spoke electrical leads run on the back of the sensor. Adapted from [47]. In the sensing region, the spokes and rings overlap and are separated by a thin layer of piezoresistive ink, forming a sandwiched configuration (Figure 5.2A). In the non-sensing region, the spokes and rings are separated by biaxially-oriented polyethylene terephthalate (BoPET, or trade-name Mylar; Mylar is generally used for applications requiring high tensile strength, chemical stability, and electrical insulation) (Figure 5.2B). The coating and sensor matrix regions are also comprised of Mylar.   Figure 5.2: Cross-sections of the Tekscan K-Scan 4400 hip contact force sensor. (A) In the sensing region, piezoresistive ink is sandwiched between the rings and spokes. (B) In non-sensing regions, the rings and spokes are separated by a layer of insulating Mylar. A B Piezoresistive ink Mylar insulation Spokes Rings BACK FRONT ‘Spoke’ electrical leads ‘Ring’ electrical leads ‘Ring’ electrical leads ‘Spoke’ electrical leads Chapter 5: Methods C 64  Sensor Preparation Sensors were conditioned, equilibrated, and calibrated immediately prior to use using a pneumatic pressure bladder unit that covered the entire sensing area. During loading using the pressure bladder, the sensor rested on a flat rigid board that was lined with 1/16” (roughly 1.6mm) thick 90A durometer hardness polyurethane to simulate cartilage surface compliance [47]. A new sensor was used for each hip. While the sensor is rated up to 20.0 MPa of contact pressure for use in simulated weight bearing, it was anticipated that the resulting intra-articular contact pressures would be significantly lower since the aim was to conduct a passive anterior impingement exam, not a simulated weight-bearing exam. I estimated a maximum expected pressure on acetabular cartilage of approximately 2.0 MPa. The first step was to condition the sensors according to the manufacturer’s guidelines, as was the standard in similar literature [47,144,145]. The manufacturer recommends conditioning the sensors to 20% greater than the expected load, which in this case was 2.4 MPa; the sensors were loaded using a ramp to 2.4 MPa for five repetitions, with each loading phase lasting 30 seconds and a one-minute no-load rest period. Each sensor was equilibrated following conditioning. Equilibration ensures uniformity of the readout for all sensing elements. To perform equilibration, known uniform pressures were applied to the entire sensing area, at multiple pressure intervals, to equalize variability in each sensing element’s output. The sensor was equilibrated at 10 separate intervals, equally spaced by 0.2±0.01 MPa, from 0.2 to 2.0 MPa. The 10-interval equilibration was chosen to match the 10-interval calibration recommended by Brimacombe et al. for Tekscan sensors [169]. Pressure was only ever increased to each equilibration interval to avoid hysteresis (which is the difference in sensor output response during loading and unloading at the same applied force). If the pressure exceeded the desired equilibration point, all air was released from the bladder, the sensor was unloaded for 30 seconds, and then air was added so that pressure increased to the equilibration interval.  The duration of total loading never exceeded 30 seconds to avoid drift effects (which is the change in sensor output when a constant force is applied over time). Chapter 5: Methods C 65  Each sensor was calibrated following equilibration at the same 10 intervals from 0.2 to 2.0 MPa. At each calibration interval, 10 raw output frames of the Tekscan sensor data were acquired at 5 Hz (2 seconds total). The 10 frames were averaged at each applied pressure. For each sensel, average raw sensel output was plotted as a function of the applied bladder pressure in MPa. While previous work using rectangular sensors had recommended calibration using a cubic polynomial [169], the single publication related to the design of the K-Scan 4400 sensor model recommended use of a second-order polynomial for calibration [47]. Therefore I fit a second order polynomial using the polyfit function in MATLAB, a non-negative least squares fitting method. The same guidelines for avoiding drift and hysteresis effects were applied during the calibration procedure. Sensor Insertion Each hip was dissected using a standard protocol that was developed based on pilot testing. During pilot testing, I had observed that the impingement positioning was not possible if the joint capsule and ligamentum teres were completely severed – after complete dissection, the hip would dislocate before reaching the predefined internal rotation limit (based on 14 Nm applied torque). Therefore, previously used specimen preparation techniques that remove nearly all soft-tissue including the entire joint capsule [46,47,141,170], common when applying vertical loads to the hip with flexion below 60°, were not possible and it was clear that a partial dissection strategy should be used. The first dissection step was to make an anterior incision down to the level of the joint capsule which included severing the iliopsoas. Next, the anterior portion of the joint capsule, the iliofemoral ligament, was completely severed. This represented roughly the anterior half of the joint capsule. Traction was applied to the femur manually to dislocate the femur from the acetabulum. Force required for traction varied depending on the quality of the labral suction seal. After dislocation, the ligamentum teres and associated neurovascular structures were severed which permitted significant dislocation and access for the acetabulum (Figure 5.3). Chapter 5: Methods C 66   Figure 5.3: Dissection of the hip for sensor insertion. Dissection was directed by an anterior approach, permitting dislocation of the femoral head while maintaining the entire posterior portion of the hip joint capsule. Key anatomical landmarks are shown for orientation: pubic tubercle, femoral shaft (potted in ABS pipe), femoral head-neck junction, and anterior superior iliac spine (ASIS). The lunate surface of the acetabulum was dried using gauze to prepare the surface for sensor fixation using cyanoacrylate. A thin layer of cyanoacrylate was applied to the back side of the sensor using a small brush to ensure even distribution. The sensor was then aligned to the chondrolabral junction and anterior horn of the lunate surface to allow for simple location-based interpretation of the resulting contact force measurements (Figure 5.4). Manual pressure was applied and the sensor stood for 10 minutes to let the cyanoacrylate set, after which a petroleum jelly lubricant was applied to the sensor’s top surface and the femoral head was reduced into the hip joint. Pubic tubercle Femoral shaft ASIS Femoral head-neck junction Chapter 5: Methods C 67   Figure 5.4: The sensor aligned along the chondrolabral junction of the lunate surface. The pubic tubercle and ASIS are shown for reference, along with anatomical directions. Note that the femur has been completely removed in order to acquire this image; during testing, the femur was still held in place by soft tissue and posterior joint capsule. Data Acquisition & Processing Once the hip was reduced, the pelvis was mounted rigidly to the base plate in order to conduct repeated manual positioning trials into a simulated anterior impingement posture that is common clinically and used in similar biomechanics studies [74,76,171]. The positioning trials were conducted by first applying 90° flexion at neutral adduction manually, then maximum internal rotation using a torque wrench to 14 ± 1 Nm based on findings in a previous study that estimate cam impingement occurs at 12 Nm of applied internal rotation torque [74]. The force sensor’s raw output was acquired at 5 Hz for 30 seconds. The sensor recording frame that corresponded to the point of maximum internal rotation was noted during testing and exported. Raw force units were saved to an output file and the calibration polynomial was applied to the output in each sensing element. Pubic tubercle ASIS Superior Anterior Chondrolabral junction Anterior horn of lunate surface Chapter 5: Methods C 68  Since one objective was to investigate a location-specific contact area profile, an important step was the registration of the sensor to the acetabular cartilage. To do so, the anterior horn of the sensor’s annulus was transformed to the clock-face location (Figure 5.5) of the anterior horn of the lunate surface (roughly 5:00 o’clock in most specimens). Then, using a CT model of the lunate surface (Appendix C), a sphere was fit to the acetabulum to evaluate the radius of the acetabulum. The outer margin of the sensor ring spanned 150 degrees and therefore its circumference was given by 150 divided by 360 multiplied by twice the outer radius multiplied by π. Assuming the outer ring of the sensor aligned perfectly to the rim of the acetabulum, the sensor location relative to the acetabular clock face was determined by calculating the span of the sensor based on the circumference of the acetabulum (Figure 5.6).   Figure 5.5: The clock-face representation of the acetabulum. © Reproduced from [24] with permission from Springer.  Chapter 5: Methods C 69   Figure 5.6: The sensor map and associated clock-face orientation.  (A) A sample readout of the sensor map qualitatively shows a contact pattern indicative of cam-type impingement, before registration. (B) After registration, the force distribution can be interpreted relative to the acetabulum using the clock-face method. The pink lines show the areas of the sensor that correspond to 3:00 (anterior) and 9:00 (posterior) on the acetabular lunate surface. Generated using MATLAB 2012a (MathWorks; Natick, MA, USA). Since the size of acetabula varied, the coverage of each sensor varied so only the sensing elements covering 9:00 o’clock to 3:00 o’clock were evaluated. The sensing elements were binned into 6 regions between each number on the clock-face (9:00 to 10:00, 10:00 to 11:00, and so on). Total resultant force was the sum of all force readouts in the entire 9:00 to 3:00 range. Location-based force distribution was determined by calculating the sum of force for each specific region. The radial component of force centroid, ?̅?, was calculated for each sensor map using the following equation: ?̅? =∑𝐹𝑖𝜃𝑖∑𝐹𝑖 Where 𝐹𝑖 is the force at each sensel i, and 𝜃𝑖 is the location. B A Pressure (MPa) 0 2.5 3:00 (Anterior) 9:00 (Posterior) Chapter 5: Methods C 70  5.2.2 Repeatability Cadaver hips (n=12) were obtained from six donors (age range: 62 to 74; two female) using a specimen screening protocol via collaboration with a donor company (Science Care; Phoenix, AZ, USA). The aim of the screening process was to acquire hips without cam deformities, as well as hips that had natural cam deformities (rather than simulate them as had been done in previous studies [74,107,124]). The hips had varying femoral and acetabular morphology. There were three control hips and nine cam hips (Appendix A Specimen Screening). Hips were frozen and stored at -18°C before use. Each specimen included all soft-tissue, hemi-pelvis and proximal femur to mid-shaft. Half of the sacrum was included (sectioned sagitally at the midline). After the experiments in Chapter 3 and Chapter 4 had been conducted, each specimen was dissected according to the protocol in Section 5.2.1 (above). Each hip was moved manually into a simulated anterior impingement posture (Chapter 4: Methods B) for six repeat trials to measure acetabular contact force. Mean resultant force, mean force centroid location, and mean force per region, plus the standard deviation and 95% confidence interval for each, were calculated for each specimen individually. Measurement repeatability (six repeat trials each for 12 hips, 72 trials total) was determined by calculating the root-mean-square (RMS) average standard deviation (SD) and coefficient of variation (CV) for each measure (resultant force, force per region, force centroid location). The RMS average for SD and CV were calculated according to the recommendations for quantifying reproducibility by Glüer et al. [164].  5.3 Results Two hips from the cam group dislocated during the positioning test and were excluded from analysis (hips #11 and #12 in Appendix A Specimen Screening). The range for resultant force was 1.3 N to 14.8 N in the control group, compared to 143.4 N to 768.9 N in the cam group. The overall mean contact centroid location was 58.4° (95% CI: -40.3° to 157.0°) in controls Chapter 5: Methods C 71  (i.e. 1:00 to 2:00 region), compared to 69.3° (95% CI: 54.8° to 83.8°) in the cam group (i.e. 12:00 to 1:00 region) (Table 5.2). Region-specific contact force was highest between 12:00 and 2:00 (Table 5.3).  Table 5.2: Resultant force and centroid location, presented as mean and SD (6 trials), for each hip. Hip Group Resultant Force (N) Centroid (°) Mean SD  Mean SD  1 Cam 248.7 154.1  62.1 12.8  2 Cam 381.1 57.1  92.8 5.5  4 Cam 400.5 156.9  82.3 3.0  5 Cam 768.9 169.9  74.6 3.9  6 Cam 143.4 43.1  55.5 7.7  7 Cam 271.6 181.1  78.4 8.5  8 Cam 197.5 69.9  39.5 5.7  Avg Cam 344.5 131.0  69.3 7.4          3 Control 14.8 9.3  98.3 24.2  9 Control 1.3 2.1  72.5 83.6  10 Control 6.9 2.0  4.2 4.8  Avg Control 7.7 3.7  58.4 50.3  Table 5.3: Region-specific resultant force (N) for each specimen reported as mean (SD) from 6 trials. Hip Group Region 3:00-2:00 2:00-1:00 1:00-12:00 12:00-11:00 11:00-10:00 10:00-9:00 1 Cam 6.0 (6.6) 138.9 (108.2) 87.3 (61.4) 3.4 (3.8) 6.3 (6.6) 6.9 (13.7) 2 Cam 30.3 (17.6) 130.1 (35.1) 30.9 (13.2) 32.0 (15.8) 71.9 (12.2) 85.8 (19.9) 4 Cam 9.5 (14.2) 55.7 (46.3) 227.6 (46.2) 65.4 (24.9) 17.6 (24.2) 24.7 (24.5) 5 Cam 44.6 (27.5) 168.3 (61.6) 365.7 (127.4) 135.4 (31.4) 50.0 (9.6) 4.8 (1.8) 6 Cam 23.9 (22.3) 71.2 (17.7) 29.3 (6.6) 13.7 (5.0) 1.6 (1.0) 3.7 (1.6) 7 Cam 18.0 (15.8) 91.8 (61.1) 72.0 (27.6) 32.7 (32.3) 16.2 (13.1) 40.8 (35.7) 8 Cam 67.5 (57.2) 118.3 (41.0) 2.7 (2.0) 2.2 (3.1) 2.8 (3.8) 4.0 (6.1) Mean (RMS Avg) Cam 28.5 (27.6) 110.6 (59.3) 116.5 (57.5) 40.7 (20.5) 23.8 (12.3) 24.4 (18.9)         3 Control 3.3 (2.2) 3.5 (3.6) 0.5 (0.3) 2.4 (1.3) 0.0 (0.1) 5.1 (2.8) 9 Control 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.1) 0.0 (0.0) 1.3 (2.2) 10 Control 6.8 (2.0) 0.1 (0.1) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) Mean (RMS Avg) Control 3.3 (1.7) 1.2 (2.0) 0.2 (0.2) 0.8 (0.7) 0.0 (0.0) 2.1 (2.0)  Chapter 5: Methods C 72  The RMS average SD from all repeated measurements (six trials each for 10 hips) was two orders of magnitude higher in the cam group than in the control group. Due to the small total force in the control group, the RMS average CV in the control group was more than double than in the cam group (Table 5.4).  Table 5.4: Repeatability results for resultant contact force from six trials each in 10 hips. Standard deviation (SD) is reported in N. Coefficient of variation (CV) is reported in %. Group N SD (CV) RMS Avg Lower 95% CI Upper 95% CI Control 3 3.7 (99.8) 3.0 (73.7) 4.8 (154.4) Cam 7 131.0 (42.6) 106.2 (34.5) 170.8 (55.5) Total 10 109.6 (65.2) 917 (54.6) 136.2 (85.1)  The RMS average SD of the force centroid measurement (six trials each for ten hips) was 50.3° in the control group, representing 28% of the sensor’s radial 180° span. In the cam group, the RMS average SD of the force centroid was 7.4° representing 4% of the sensor’s radial span (Table 5.5).  Table 5.5: Repeatability results for contact centroid (six repeated trials per hip). Standard deviation (SD) is reported in °. Group N SD RMS Avg Lower 95% CI Upper 95% CI Control 3 50.3 37.2 77.9 Cam 7 7.4 6.0 9.6 Total 10 28.3 23.6 35.1  For location-based force measurement, the RMS average SD over six trials in 10 specimens (with six regions each), was minimal in the control group (1.4 N) and an order of magnitude lower than the RMS average SD in the cam group (37.7 N) (Table 5.6). The magnitude of SD was greater for higher forces (Figure 5.7). The RMS average coefficient of variation (CV) was not calculated for 12 of the 60 possible regions since mean force in those locations was less than 0.5 N, causing CV to approach infinity.   Chapter 5: Methods C 73  Table 5.6: Repeatability results (6 trials per hip) for region-specific contact force. Standard deviation (SD) is reported in N. Coefficient of variation (CV) is reported in %. Group N SD (CV) RMS Avg Lower 95% CI Upper 95% CI Control 3 1.4 (52.0)  1.2 (45.4) 1.6 (60.8) Cam 7 37.7 (84.0) 34.4 (76.7) 41.6 (92.9) Total 10 31.5 (75.8) 29.2 (70.2) 34.3 (82.4)   Figure 5.7: Mean vs SD (from 6 trials) for region-specific force (10 hips, 6 regions each). SD increased as force magnitude increased demonstrating that readings were highly variable at higher force magnitudes. 5.4 Discussion In this chapter, I addressed Aim C of this thesis by describing a novel method for measuring acetabular contact mechanics in a simulated anterior impingement posture using a piezoresistive Tekscan sensor. The method was successfully applied to seven hips with cam morphology and three controls in order to quantify the measurement repeatability over six trials for resultant force, force centroid location, and force distribution in each hip. Chapter 5: Methods C 74  The finding of 42.6% (95% CI: 34.5 to 55.5) coefficient of variation for repeated resultant force measurements in cam hips demonstrated significant variability between trials, but is not surprising for Tekscan sensors. Past studies have reported coefficients of variation between 3.1 and 23.5% in the patellofemoral joint [142,143]. One study used a Tekscan sensor to measure pressure in shoes of human subjects during walking, and reported that the variability between steps per subject was greater than the variability between subjects, indicating very poor between-trial repeatability (the study did not report values for CV but instead reported intraclass correlation) [172]. Considering that the hip joint is more curved than previously studied joints, an average CV for resultant force of 42.6% is not unexpected.  Large variation between repeated measures may be due to shear forces on the sensor. Shear force is thought to be characteristic of cam impingement: clinicians hypothesize that the femoral head-neck junction jams into the acetabulum, inducing shear force at the chondrolabral junction (note that this has only been postulated and not yet demonstrated experimentally). Shear forces distort sensor output, so if the force experienced by the sensor had a substantial shear component it cannot be expected that the sensor will respond with a repeatable output. Also, the shear force on the sensor may have been increased by cementing the sensor to acetabular cartilage, since the rigid bond between cartilage and sensor means that shear loading can occur at the bond interface [142]. While fixing the sensor in place may have increased variability of resultant force measurements, one advantage of fixing the sensor along the acetabular rim was that it provides the capacity to measure contact force location in a repeatable way. Error between repeated measurements was 7.4° (RMS average SD from 7 cam hips, 6 trials each; 95% CI: 6.0 to 9.6°) for force centroid location measurement, which represents less than 5% of the sensor’s 180° radial span. For context, when using the clock-face representation (common in previous simulation studies or intra-operative cartilage assessments [22–24,173–176]) the clinically relevant ‘resolution’ is 30° (i.e. a one hour range between, or 12:00 to 1:00), so a variation of 7.4° affects less than 25% of the effective clock-face resolution. Furthermore, the finding that contact force centroid was located in the 12:00 to 1:00 region for cam hips is consistent with past computer simulation Chapter 5: Methods C 75  studies [22,24,147], and helps confirm that our method was able to measure contact force in the critical impingement zones that have been shown to correlate with cartilage damage [24]. Likewise, for region-specific force measurement in cam hips, force was highest in the 12:00 to 1:00 and 1:00 to 2:00 regions; force in those two regions was twice the force from the other four regions combined. Fixing the sensor to the acetabulum also has the advantage of reducing crinkling, a common problem for Tekscan sensors due to their 0.1mm thickness [142,146] that may have contributed to variability in resultant force measurement. It is not surprising that the two hips that were excluded from analysis (due to unavoidable dislocation) had large cam deformities and additional pincer-related deformities. In these two hips (Hips #11 and #12 in ), the α angle was high (99°, 82°), coxa vara was present in both, and one had a lateral centre-edge angle greater than 39° which is indicative of acetabular overcoverage. It is possible that the large cam and pincer deformities cause dramatic levering during impingement, requiring the joint capsule, ligaments, and/or labral seal to provide stability and prevent translation of the femoral head. After the dissection to insert the sensor, impingement may have caused the femoral head to easily lever out of the acetabulum since it was no longer opposed by the stabilizing soft-tissue structures. One limitation of the methods presented in this chapter is that the radial ring-and-spoke layout of the sensor means that spacing between sensing elements varies slightly by location along the spokes. Sensing elements from rings near the inner radius of the annulus are spaced more closely compared to sensels from rings near the outer radius. However, a large number of sensing elements (61 sensels/cm2) means that potential error in area measurement might be negligible with respect to other error contributions already discussed.  Another limitation was that I performed a flat surface calibration when in reality the sensor is designed for use in an approximately spherical joint, although previous work has shown that there is negligible difference between flat and conical calibrations. In their description of the 4400 sensor, Rudert et al. compared a flat-platen calibration to a cup-and-cone-platen calibration up to 10.4 MPa contact stress [47]. Both the flat and conical plates were lined with 90A polyurethane to simulate cartilage loading. They qualitatively compared calibration curves by plotting applied contact stresses (performed using a materials testing system) versus Chapter 5: Methods C 76  the raw output at 250N load increments up to roughly 10.4MPa. A second-order polynomial was fit to the raw output versus applied contact stress data for both calibrations. Qualitatively there was negligible difference between the two curves indicating that the sensor behaviour was not affected by the conical surface curvature [47]. While a conical surface does not represent the hip joint, it provides good reference for the effect of single-dimension curvature on force measurement.  5.5 Summary  I developed a novel method for measuring contact force, centroid, and distribution on acetabular cartilage in cam and control hips that are positioned in a simulated anterior impingement posture.  The variability of the method for resultant force measurement was 42.6% (RMS average CV; seven cam hips, six repeat trials each), which was large but not surprising for Tekscan sensors since past studies have reported coefficients of variation between 3.1 and 23.5%.  In light of the highly variable force measurement in the cam group, and the finding that resultant force in the cam group was one to two orders of magnitude greater than in the control group, it may make sense to represent contact force with a binary measure.  The variability between repeated contact centroid measurements was 7.4° (RMS average SD; 95% CI: 6.0 to 9.6°), less than 5% of the sensor’s 180° radial span of interest, and less than 25% of the resolution associated with a typical ‘clock-face’ representation of the acetabulum.  Contact force centroid was located in the 12:00 to 1:00 region for cam hips, consistent with past computer simulation studies, which supports the cam FAI hypothesis.   Chapter 6: Integrated Study 77  6 Integrated Study: Open-MRI Measures of Cam Intrusion for Hips in an Anterior Impingement Posture Are Related to Acetabular Contact Force 6.1 Introduction Cam-type femoroacetabular impingement (FAI) is a proposed pathomechanism for hip osteoarthritis (OA) [1,130]. It has been hypothesized that the cam femur, with its reduced head-neck offset, intrudes into the intra-articular joint space and compresses the acetabular chondrolabral junction during combined flexion and internal rotation [27,57,107,130,177]. Abnormal shear forces due to cam intrusion are thought to directly cause cartilage and labral damage [24,77,178]. However, not all hips with cam deformities will develop OA [19,106] or hip pain [16,179], indicating that the biomechanics of cam FAI are not fully understood [177]. To establish treatment guidelines, assess the importance of specific activities in FAI, and understand why only some hips with deformities become symptomatic, we require a method to assess impingement directly in vivo. Most biomechanical evidence about cam FAI comes from studies using computer models [22–24,26,180] or intraoperative observation [1,100]. Both methods have significant limitations. Computer models exclude soft-tissue and require assumptions about hip motion, while intraoperative observations are invasive and may not reflect hip behaviour in conscious patients. Imaging in functional positions offers the potential to directly assess impingement without disrupting or simplifying hip physiology. Dual fluoroscopy was used to study patterns of joint contact in hips with FAI morphology during the anterior impingement test [27]. However, the radiation exposure produced by dual fluoroscopy limits its wide use in vivo and three dimensional visualization with this method requires a registration step between the fluoroscopy images and a high-resolution model that limits the method’s accuracy for predicting joint contact.  Open magnetic resonance imaging (MRI) has been used to image femoroacetabular relationships directly in supine hip flexion [9] and in a set of functional postures [132]. Chapter 6: Integrated Study 78  However, a key limitation to image-based studies of impingement biomechanics is that there is not an established link between imaging findings and mechanical force at the site of impingement [56].The objective of this study was to establish whether there is a link between MRI findings and mechanical contact force in hips with a cam deformity. This section addressed the following research questions, in cam and control specimens positioned in a simulated anterior impingement posture: 1. Does open MRI show intrusion of the cam deformity into the intra-articular joint space? 2. Is the presence of a cam deformity associated with elevated acetabular contact force? 3. Are MRI measures of cam intrusion related to acetabular contact force? 6.2 Methods We obtained 12 cadaver hips from 6 donors (Table 6.1) and classified them as ‘cam’ or ‘control’ using a 3D imaging protocol. The specimen procurement company (Science Care; Phoenix, AZ, USA) imaged each donor’s full pelvis (both left and right hips) using helical CT (Toshiba Aquilion; Tustin, CA, USA) with 1 mm slice thickness and 0.80 mm in-plane voxel size. The raw image volumes were used to create radial reformats with the long axis of the femoral neck as the rotation axis, using a standard multi-planar reformatting (MPR) technique [4,6,7]. The femoral neck axis was determined by segmenting the femur (Mimics 15.0, Materialise; Leuven, Belgium) and fitting a least-squares cylinder (MATLAB, Mathworks; Natick, MA, USA).  A rigid transformation plus trilinear interpolation defined the new radially-sliced image volume. The reformatted image volume had 18 total slices (1.0 mm isotropic voxels), corresponding to 36 positions circumferentially along the femoral head-neck junction. We evaluated the α-angle [3] at all 36 positions, and used the maximum, αmax, to define the presence of a cam deformity. We classified hips as having a cam deformity if αmax > 60° [4], rather than αmax > 50° or 55°, since cam deformities appear larger when looking at the entire cam profile via MPR [156,181]. In total, there were three hips without cam morphology that were classified as ‘controls’, and nine hips with cam deformities classified as ‘cams’ (Table 6.1). Chapter 6: Integrated Study 79  A trained reader assessed femoral head congruency, acetabular version (cross-over sign), depth (profunda/protrusio), coverage (lateral centre-edge angle, LCE), and caput-collum-diaphyseal angle (CCD angle) to identify coxa vara/valga, in all hips, using a plain A-P x-ray (Appendix A Specimen Screening). Femoral condyles were not available to classify femoral version/torsion. In the ‘control’ hips without cam morphology, one hip had coxa valga (Hip #3), two hips had chondrocalcinosis (Hips #9-10), and one hip had additional anterior femoral neck osteophytes (Hip #10). In the hips with cam morphology, five of nine hips had additional pincer-type acetabular morphology (Hips #5-8, #11) and four had coxa vara (Hips #5-6, #11-12). One cam hip had coxa valga and an incongruent head (Hip #4). Two cam hips had significant anterior neck osteophytes (Hips #5-6). Labral damage (fraying or tears of any kind) was present in 1/3 control hips (Hip #3), and 7/9 cam hips (Hips #1-2, 4-7, 12). All hips, cams and controls, had macroscopic cartilage damage upon dissection.  Table 6.1: Selected specimen screening information. Full screening information: see Appendix A. Hip 1 2 3 4 5 6 7 8 9 10 11 12 Group CAM CAM CON CAM CAM CAM CAM CAM CON CON CAM CAM αmax (°) 71 84 57 67 84 82 70 79 54 53 99 82 CCD angle (°) 128 129 150 143 112 114 132 130 133 129 112 115 Lateral centre- edge angle (°) 30 28 34 30 48 46 41 39 36 37 42 34 Profunda/protrusio - - - - - - - - - - - - Crossover sign - - - - + + + + - - - - Labral damage + + + + + + + - - - - + Age (years) 62 72 73 75 64 78 Gender M F M M F M Height (cm) 185.4 162.6 182.9 175.3 152.4 177.8 Weight (kg) 93.0 40.8 75.7 125.6 46.2 127.0   We used a custom MR-safe positioning rig to position hips a in a posture approximating the anterior impingement test (Chapter 4: Methods B). The rig positioned each hip passively in 90° flexion and 0° adduction, then 14±1 Nm of torque was applied in the internal rotation direction with a torque wrench. To Chapter 6: Integrated Study 80  establish the precision of the positioning technique, we performed six repeat positioning trials in each of six specimens, for a total of 36 trials. We measured joint position using an optical tracking system (Optotrak Certus, Northern Digital; Waterloo, Canada) and represented it using the floating axis joint angle convention [162] [163]. Precision for each joint rotation, reported as the root-mean-square (RMS) standard deviation (SD) over the 36 trials, was 2.0° (1.6 – 2.7; 95% confidence interval) for flexion, 1.6° (1.3 – 2.1) for adduction, and 1.3° (1.0 – 1.7) for internal rotation (Chapter 4: Methods B). Each hip was imaged in two positions, (i) supine and (ii) the simulated anterior impingement exam position, with a 0.5T upright open MR scanner (Paramed MROpen; Genoa, Italy) using a T1-weighted 3D gradient echo sequence (Table 3.1) (Chapter 3: Methods A). A one-channel send-receive coil was placed around each cadaver hip. Each image was then reconstructed into radial reformats via the same MPR technique used with CT, above. The reformatted image volume had 36 total slices (0.9 mm isotropic voxels), corresponding to 72 positions circumferentially along the femoral head-neck junction at which we could evaluate clearance with the acetabular margin.   Table 6.2: MRI sequence parameters Parameter Value Matrix size 256 x 256 FOV 28.5cm x 28.5cm n Slices 102 Slice thickness 0.90mm (no gap) TE/TR 8ms/19ms (T1-weighted) n Excitations 4 Flip angle 45° Scan time 27m 19s Output resolution 0.9 x 0.9 x 0.9 mm3  The β-angle [9] was used to define the relationship between the head-neck junction and acetabulum at all 72 positions around the head-neck junction (Chapter 3: Methods A). First, a trained reader selected 8 points on the femoral head margin to define a circle. A least-squares circle was automatically fit to the 8 selected points and the mean circle was displayed along with a circle based on the 98% confidence interval of the Chapter 6: Integrated Study 81  radius to inform the reader of circle fit error. The reader then selected the point at which the femoral head-neck junction deviated from the mean circle, and also the lateral-most bony margin of the acetabulum. Finally, with the centre of the circle fit known, β was calculated automatically. Two readers measured β on each slice for all images, and reader agreement was evaluated using the intra-class correlation coefficient. We took the slice-wise average from all readings to yield a location-specific profile for β-about the entire femoral head-neck junction (0° pointing anteriorly, 90° pointing superiorly, Figure 6.1).   Figure 6.1: Schematic diagram depicting the different slice planes after multi-planar reformatting. In total, there were 36 slice planes, each separated by 5° (not shown), allowing the visualization of 72 locations, θ, around the circumference of the femoral head-neck junction. The 0° location pointed anteriorly, while the 90° location pointed superiorly. Adapted from [4]. For each hip’s β-angle profile (Figure 6.2), we reported the minimum β-angle and corresponding location, θβmin. We used a threshold of β < 0° to define ‘intrusion’. We reported the intrusion span (number of slices with β < 0°, multiplied by 5° per slice), intrusion area (5° per slice multiplied by β, if β < 0°), and intrusion centroid θ̅, given by the equation: θ̅ =∑βiθ𝑖∑β𝑖 Chapter 6: Integrated Study 82  Where βi is the β-angle at slice i showing intrusion (i.e. slices with β > 0° were excluded from this calculation), and θ𝑖 is the circumferential slice location in degrees. We then defined a positive “MRI cam-intrusion sign” as any posture/hip with β < 0° at two or more locations, based on all individual β-angle readings as well as the average β-angle profile from all four readings. We assessed the presence of pincer abutment qualitatively for each posture/hip.  Figure 6.2: Sample β-angle profile. Each image reading results in a β-angle profile, or 72 slice-wise β-angles that corresponded to a location around the circumference of a femoral head. The β-angle profile could then be used to extract βmin, θβmin, intrusion span/centroid/area, and the MRI cam-intrusion sign. Following imaging, we dissected each hip by severing the iliofemoral ligament and ligamentum teres, dislocated the femoral head, and attached a 0.10 mm thick arc-shaped piezoresistive force sensor [47] (Tekscan K-scan 4400; Boston, USA) to the acetabulum (Chapter 5: Methods C). The sensor was fixed rigidly to the lunate surface along the chondrolabral junction with cyanoacrylate, and the exposed surface was lubricated with petroleum jelly. The size of each acetabulum varied, so we analyzed only the sensing elements covering the superior acetabular region (9:00 to 3:00 using a clock-face representation). -40-20020406080100120-180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180β-angle (°)Location about femoral neck, θ (°) βmin Slices showing intrusion Intrusion span Intrusion threshold Chapter 6: Integrated Study 83  Before insertion, each sensor was conditioned, equilibrated, and calibrated using a pneumatic pressure bladder that covered the entire sensing area. The calibration mounting board was lined with 1/16” thick 90A durometer hardness polyurethane to simulate cartilage surface compliance [47]. Conditioning was executed by loading five times to 2.4 MPa (20% greater than the maximum expected load) with 30 second rest intervals per the manufacturer’s recommendations. Equilibration and calibration were conducted at 10 evenly spaced intervals between 0.2 and 2.0 MPa, with increasing load only to avoid hysteresis effects, and with 30 second rest intervals to avoid drift effects. Calibration curves were calculated using a second order least-squares polynomial fit as recommended by the developers of the sensor [47]. Acetabular contact force was measured by moving each hip manually into a simulated anterior impingement posture for six trials to replicate the position held during MR imaging. Maximum internal rotation was applied using a torque wrench to 14±1 Nm. Total resultant force in the entire 9:00 to 3:00 sensing area was assessed for the position of maximum internal rotation. The sensing area was also divided into 6 regions (each representing a one-hour interval on the acetabular clock-face between 9:00 and 3:00) to calculate region-specific force distribution. The force centroid, θ̅𝑓, was calculated using the following equation: θ̅𝐹 =∑𝐹𝑖θ𝐹,𝑖∑𝐹𝑖 Where 𝐹𝑖 is the force at each sensel i, and θ𝐹,𝑖 is the circumferential location in polar coordinates. Precisions for total resultant force, region-specific force distribution, and force centroid were reported as the RMS average SD from all trials for all specimens.  A binary measure, the ‘contact-force sign’, was defined using a threshold value of 20 N. The contact-force sign was applied to the mean resultant force for all hips.   Chapter 6: Integrated Study 84  We used Fisher’s exact test to evaluate three null hypotheses: 1. The probability of a positive MRI cam-intrusion sign is the same whether a hip has cam or control morphology; 2. The probability of a positive contact-force sign is the same whether a hip has cam or control morphology; and 3. The probability of a positive MRI cam-intrusion sign is the same whether the hip has a positive or negative contact-force sign. To explore the dependency of each result and p-value on the β < 0° threshold for MRI-cam intrusion and 20 N threshold for contact-force, all analysis was repeated using multiple thresholds. For MRI analysis, thresholds were varied between 5° and -5° at 0.1° intervals. For force measurement, thresholds were varied between 1N and 500N at 1N intervals, respectively. The statistical hypotheses were re-tested at each threshold. 6.3 Results Qualitative findings In general, there were visible differences between cam and control hips at the impingement position based on both open MRI and sensor measurements. In the three control hips, there was no cam intrusion visible on MRI, no direct pincer abutment visible, and contact force was negligible during sensor testing (Figure 6.3). In hips with pure cam morphology, there was clear cam intrusion on MRI and a distinct acetabular contact pattern with force concentrated in the anterosuperior region (Figure 6.4). In the cam group, pincer abutment was visible in six hips with open MRI. Hips with pure cam morphology showed cam intrusion but no pincer abutment, whereas all hips with mixed morphology showed simultaneous cam intrusion and pincer abutment. In most mixed-morphology hips, cam intrusion and pincer abutment were both visible on a single slice (Figure 6.5). However, in one hip, we observed that the region of maximum cam intrusion (anterosuperior femoral head-neck junction interacting with the acetabulum Chapter 6: Integrated Study 85  between 11:00 and 12:00 o’clock) was distinct from the region of pincer abutment (directly anterior femoral head-neck junction interacting with the acetabulum between 12:00 and 1:00 o’clock) and therefore cam and pincer impingement were visible on separate slices (Figure 6.6). Changes in joint space width, contre-coup femoral head translation, and posterior instability were visible in all hips when comparing supine images to impingement images (Figure 6.7).    Figure 6.3: Results from Hip #9, a control hip (with chondrocalcinosis). The open MR slice in the 0° plane (anterior) showed the femoral head-neck junction in close proximity with the acetabular margin, but no direct pincer abutment was visible. The head-neck junction did not intrude into the intra-articular joint space. Contact force was negligible. Chapter 6: Integrated Study 86   Figure 6.4: Results from two bilateral hips with pure cam morphology. (Top) Hip #1. The MR image shown is a slice in the 0° plane. (Bottom) Hip #2. The MR image shown is a slice in the -5° plane.  Chapter 6: Integrated Study 87   Figure 6.5: Results from Hip #7. Hip #7 had a cam deformity and acetabular retroversion. On open MRI, the radial slice showing maximum cam intrusion was the 20° plane, in the anterosuperior region, and also showed simultaneous pincer abutment. Contact force was concentrated near the sensor margin/chondrolabral junction.  Figure 6.6: Results from Hip #5. Hip #5 had a cam deformity, coxa vara, acetabular retroversion, and anterior femoral neck osteophytes. (Top) An open MR slice in the 35° plane showed maximum cam intrusion anterosuperiorly. (Bottom) An open MR slice in the -20° plane showed pincer abutment, indicating that both pincer abutment and cam intrusion occurred simultaneously, but in separate locations/slices (55° apart) relative to the femoral head-neck junction. In this specimen, contact force was highest of all specimens and widely distributed radially. Chapter 6: Integrated Study 88   Figure 6.7: Visual inspection of joint space shown on MRI in two positions.  (Top) Open MRI slice (0° plane) of hip #1 in a supine position. Joint space appeared evenly distributed in all planes. (Bottom) Open MRI slice (0° plane) of hip #1 in an anterior impingement posture. Joint space appeared widened medially and narrowed laterally, near the site of impingement. The imaging findings suggest inferolateral translation of the femoral head.  Supine Anterior Impingement Test Chapter 6: Integrated Study 89  Quantitative findings The minimum β-angle ranged from 1.4° to -28.5° in cam hips versus 4.6° to -0.2° in control hips (Table 6.3). The intrusion centroid location relative to the femur varied from -3.0° to 35.6°, where 0° points anteriorly (see Figure 6.1 for reference). The ICC (2,k) for slice-wise β measurement for two trials each for two readers was 0.93. Table 6.3: Summary of MRI results for minimum β and intrusion based on a 0° β threshold. Group Hip # βmin (°) θβmin Intrusion Span (°) Intrusion area (°2) Intrusion centroid Cam 1 -14.3 -5 60 475 -3.0 2 -6.2 25 25 180 14.6 4 1.4 -10 - - - 5 -14.5 35 75 433 22.8 6 -1.8 5 15 27 3.8 7 -12.5 -20 80 628 3.6 8 -5.9 15 15 72 2.2 11 -28.5 55 85 1262 35.6 12 -6.5 30 45 169 21.7 Control 3 4.6 -100 - - - 9 -0.2 30 5 1 5 10 0.1 -90 - - -  Two cam hips (Hips #11-12) were excluded from contact-force analysis because they dislocated before any torque could be applied during the impingement positioning. The range of mean resultant contact force in the remaining 7 cam hips was 143 to 769 N, at least one order of magnitude greater than in control hips, whose resultant contact force ranged from 1.3 to 14.8 N. In cam hips, the force centroid was located between 12:00 and 1:00 relative to the acetabular clock-face (mean 69°, 95% confidence interval 55° to 84°, where 60° represents 1:00 and 90° represents 12:00) (Table 6.4). The 12:00 to 2:00 regions (60° circumferential span) accounted for 66% of the total force, on average, while the other four regions combined (120° circumferential span) accounted for the remaining 34%, indicating a concentration of force in anterosuperior region (Table 6.5). The RMS average SD in force centroid location from all repeated trials was 7.4° in cam hips, less than 5% of the span of the sensing region. Chapter 6: Integrated Study 90  Table 6.4: Resultant force and centroid location for each hip. Data is presented as mean and standard deviation (SD) from 6 repeat trial per hip. Hip Group Resultant Force (N) Centroid (°) Mean SD  Mean SD  1 Cam 248.7 154.1  62.1 12.8  2 Cam 381.1 57.1  92.8 5.5  4 Cam 400.5 156.9  82.3 3.0  5 Cam 768.9 169.9  74.6 3.9  6 Cam 143.4 43.1  55.5 7.7  7 Cam 271.6 181.1  78.4 8.5  8 Cam 197.5 69.9  39.5 5.7  Avg Cam 344.5 131.0  69.3 7.4  3 Control 14.8 9.3  98.3 24.2  9 Control 1.3 2.1  72.5 83.6  10 Control 6.9 2.0  4.2 4.8  Avg Control 7.7 3.7  58.4 50.3    Table 6.5: Resultant force (N) at each of 6 regions for each hip. Data is reported as mean (standard deviation) from 6 repeat trials per hip. Hip Group Region 3:00-2:00 2:00-1:00 1:00-12:00 12:00-11:00 11:00-10:00 10:00-9:00 1 Cam 6.0 (6.6) 138.9 (108.2) 87.3 (61.4) 3.4 (3.8) 6.3 (6.6) 6.9 (13.7) 2 Cam 30.3 (17.6) 130.1 (35.1) 30.9 (13.2) 32.0 (15.8) 71.9 (12.2) 85.8 (19.9) 4 Cam 9.5 (14.2) 55.7 (46.3) 227.6 (46.2) 65.4 (24.9) 17.6 (24.2) 24.7 (24.5) 5 Cam 44.6 (27.5) 168.3 (61.6) 365.7 (127.4) 135.4 (31.4) 50.0 (9.6) 4.8 (1.8) 6 Cam 23.9 (22.3) 71.2 (17.7) 29.3 (6.6) 13.7 (5.0) 1.6 (1.0) 3.7 (1.6) 7 Cam 18.0 (15.8) 91.8 (61.1) 72.0 (27.6) 32.7 (32.3) 16.2 (13.1) 40.8 (35.7) 8 Cam 67.5 (57.2) 118.3 (41.0) 2.7 (2.0) 2.2 (3.1) 2.8 (3.8) 4.0 (6.1) Avg Cam 28.5 (27.6) 110.6 (59.3) 116.5 (57.5) 40.7 (20.5) 23.8 (12.3) 24.4 (18.9) 3 Control 3.3 (2.2) 3.5 (3.6) 0.5 (0.3) 2.4 (1.3) 0.0 (0.1) 5.1 (2.8) 9 Control 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.1) 0.0 (0.0) 1.3 (2.2) 10 Control 6.8 (2.0) 0.1 (0.1) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) Avg Control 3.3 (1.7) 1.2 (2.0) 0.2 (0.2) 0.8 (0.7) 0.0 (0.0) 2.1 (2.0)   Chapter 6: Integrated Study 91  MRI Cam-intrusion and Contact-force Signs We found a significant association between the presence of cam morphology (from MPR CT) and the MRI cam-intrusion sign (Table 3; p=0.0182, Fisher’s exact test). After re-testing with varying cam-intrusion thresholds, there was significant association between cam morphology and the MRI cam-intrusion sign for all thresholds between -2.7° and 0.2° (p < 0.05; Figure 6.8). For each reader’s individual readings, 47/48 cases resulted MRI cam-intrusion signs were shown to agree with the MRI cam-intrusion sign that was based on the average β-angle profile (from all four readings).   Table 6.6: Contingency table of each hip, MRI cam-intrusion sign vs morphology group. A β-angle threshold of 0° was used to determine cam-intrusion sign in this table.   MRI    Cam-intrusion positive Cam-intrusion negative Totals Morphology Cam 8 1 9 Control 0 3 3  Totals 8 4 12     Figure 6.8: Fisher’s exact test of cam-intrusion sign vs. cam deformity for various β-angle thresholds. Statistical analysis was repeated (Fisher’s exact test; cam-intrusion sign vs. cam deformity) at thresholds from -5° to 5° to investigate dependency of p-value on the β intrusion threshold. Chapter 6: Integrated Study 92  We also found a significant association between the presence of cam morphology and the acetabular contact-force sign (Table 6.7; p=0.0083, Fisher’s exact test). After re-testing with varying contact-force thresholds, there was significant association between cam morphology and the contact-force sign for all thresholds between 15 N and 143 N (p < 0.01; Figure 6.9).   Table 6.7: Contingency table of each hip, contact-force sign vs morphology group. A contact force threshold of 20 N was used to determine the contact-force sign in this table.   Sensor    Contact-force positive Contact-force negative Totals Morphology Cam 7 0 7 Control 0 3 3  Totals 7 3 10    Figure 6.9: Fisher’s exact test of contact-force sign vs. cam deformity for various force thresholds. Statistical analysis was repeated (Fisher’s exact test, contact-force sign vs. cam deformity) using thresholds at 1 N intervals from 1 N to 500 N to depict the dependency of p-value on the contact force threshold. Chapter 6: Integrated Study 93  Finally, we found a significant association between the MRI cam-intrusion sign and the contact-force sign, given a cam-intrusion threshold of 0° and a contact-force threshold of 20 N (Table 6.8; Fisher’s exact test, p=0.033).  Table 6.8: Contingency table for the MRI cam-intrusion sign and contact-force sign.   MRI    Cam-intrusion positive Cam-intrusion negative Totals Sensor Contact-force positive 6 1 7 Contact-force negative 0 3 3  Totals 6 4 10   6.4 Discussion We assessed impingement visible on MRI and forces on acetabular cartilage in 12 cadaver hips that were positioned in an anterior impingement posture to assess links between MRI assessment of impingement and biomechanics.  We found that: (i) cam morphology was significantly associated with MRI-observed cam intrusion; (ii) cam morphology was significantly associated with experimentally measured contact force; and (iii) experimentally measured contact force was significantly associated with MRI-observed cam intrusion. Our findings that the femoral head-neck junction intruded medial to the acetabular margin in 8/9 cam hips, and that intrusion was significantly associated with cam morphology, represent some of the first direct experimental evidence for cam intrusion during impingement. Intra-operative observations have qualitatively demonstrated cam intrusion [1,100], but intra-operative observations cannot necessarily be extrapolated to intact hips. Our results are consistent with findings from a computer simulation study that quantified cam intrusion by positioning 3D CT-derived models in impingement positions using patient-specific range of motion data that had been measured separately using magnetic tracking [26]. In this study, 10/13 cam hips experienced cam intrusion during maximal internal rotation at 90° flexion [26], which is consistent with our finding of cam intrusion in 8/9 of the hips we studied. Chapter 6: Integrated Study 94  Our finding that the MRI cam intrusion centroid ranged from -3.0° to 35.6° (where 0° points toward the anterior femoral head-neck junction, 90° superior; see Figure 6.1 for reference) is consistent with computer simulations and 3D imaging studies [22,26,27,69] which demonstrated that the anterior and anterosuperior regions of cam deformities are likely to induce impingement, and that cam deformity collision zones  are specimen-specific. Furthermore, our finding that the acetabular contact force centroid was located near 1:00 (clock-face localization) is consistent with five studies that reported that either intraoperative cartilage damage [61,130], computer-simulated impingement zones [22,23], or both [24] were localized (mode) at 1:00 on the acetabulum in cam hips. Our finding that two cam hips dislocated during the sensor-in positioning trials (which led to their exclusion from our analysis) is not surprising. Both hips has a large α-angle (99°, 82°) and coxa vara, and one had a lateral centre-edge angle greater than 39°, indicative of acetabular overcoverage. The open MR images showed dramatic cam intrusion in both hips (Figure 6.10), indicating that impingement combined with compromised ligaments and labral seal may have led to dislocation.   Figure 6.10: Open MRI for two hips that dislocated during the sensor experiment. (Left) MR slice through the 30° plane of the femoral neck of hip #12 (β = -6.5°). (Right) MR slice through the 55° plane of the femoral neck of hip #11 (β = -28.5°).   Chapter 6: Integrated Study 95  A strength of this study was that we standardized the joint positioning technique, using a method with documented precision (RMS average SD for flexion, adduction, and internal rotation of less than 2° for repeated trials). Relative joint position and hip anatomy determine clearance between the acetabulum and femoral head-neck junction, so high precision in joint positioning is essential when studying hips in functional positions, although positioning precision is rarely reported. A second strength of this study was that the significance of the relationship between the MRI cam-intrusion sign and presence of a cam deformity was not affected by selection of the β-angle threshold for intrusion (for β-angle between 0.2° and -2.7°). A range of possible β-angle thresholds will still be sensitive at detecting presence/absence of intrusion, which affords clinical readers a margin of error when measuring β-angle.  One limitation was that the sensor-based contact force measurements were highly variable (RMS average SD of 131 N, 42.6% RMS average CV in cam hips). This is consistent with poor variability findings from  past studies that reported coefficients of variation between 3.1 and 23.5% in the patellofemoral joint [142,143], and between 4 and 8% in the lumbar facet joint [144]. Considering that the curvature of the hip joint is extreme in comparison with previously studied joints and sensor performance is affected by curvature, the CV of 42.6% in cam hips is not surprising. Large variations in force magnitude between repeated measures may be due to shear forces, which might have resulted from cementing the sensor in place [142], or from the shearing nature of cam impingement itself. However, our finding that the standard deviation in contact-force centroid measurement was less than 5% of the span of the sensor region indicates that shear did not affect location-specific contact force measurement. Importantly, high variability between repeated force measurement trials did not affect the binary contact-force sign analysis and the final conclusion that cam hips experienced elevated contact force, while control hips experienced negligible force.  Furthermore, it was qualitatively clear that cam hips experienced intra-articular contact force while control hips did not. A binary sign proved to be a robust method for representing the presence or absence of contact force, as supported by our finding that the contact-force Chapter 6: Integrated Study 96  sign was independent of the selected contact-force sign threshold between 15 N and 143 N, almost a full order of magnitude. Elevated contact force in cam hips suggests that contact between the cam deformity and lunate surface contribute to resisting the forced internal rotation in the approximated anterior impingement exam, but it is not clear what tissue structures resisted contact force in control hips. Since the anterior hip joint capsule was dissected, the ligamentum teres was severed, the labral seal was disrupted due to sensor insertion, and the cadaveric model meant there was no muscular activity, we can eliminate those structures as sources of resistance terminating motion. Some possible tissues that may have been resisting the internal rotation torque in control hips include the posterior joint capsule/ligaments, or extra-articular impingement of the femur against the labrum or rim of the acetabulum. We are limited in drawing further conclusions because this experiment only controlled applied torque, and not ‘end-feel’ like a clinician would in the anterior impingement test. The open MRI method we have presented here has significant potential to investigate cam intrusion in vivo. Potential applications of this method include investigating which morphological parameters predict the magnitude of cam intrusion during the anterior impingement test and how posture affects cam intrusion. 6.5 Conclusion We found a relationship between an MRI-based measure of cam intrusion during hip impingement and sensor-based measurement of acetabular contact force. Our work supports the use of direct hip imaging postures suspected of producing impingement for in vivo studies of FAI. Open MRI can provide direct image-based assessment of FAI mechanics, and has advantages over previous approaches that rely on assumptions or simplification of hip joint biomechanics.  97  7 Integrated Discussion 7.1 Motivation & Contributions Mechanics are central to the hypothesis that cam-type femoroacetabular impingement causes hip osteoarthritis, but impingement-related contact mechanics are not well understood. It is proposed that cam deformities induce abnormal shear forces on acetabular cartilage during motion [1], but this claim has not been confirmed with direct, experimental measures of force, and most supporting evidence is based on computer model-based investigation with no mechanical validation. The work in this thesis was done to address the lack of experimental evidence about the mechanics of cam impingement and made three key contributions: 1. This thesis described the first direct, non-invasive, non-model-based representation of the cam-type hip joint in situ in an actual impingement posture, which allowed for quantitative description of cam deformity intrusion into the acetabulum. 2. This thesis described the first mechanical measurement of acetabular contact force in a simulated anterior impingement posture, and demonstrated that cam hips experience contact force that is one to two orders of magnitude higher than the force measured in control hips. 3. This work found that open MRI measures of cam intrusion were linked to contact mechanics in a simulated anterior impingement posture, which represents the first time that any study of FAI has related an imaging measure to an experimentally measured mechanical property. The established relationship between open MRI-observed cam intrusion and contact mechanics supports the use of open MRI for investigation of cam FAI patients in vivo.  98  7.2 Strengths The first key strength of this work was that the imaging approach provided direct images of hips in an impingement posture, which is an improvement over indirect computer-model based approaches. Direct 3D imaging has many advantages over previously used methods, most of which used computer models to predict the locations where impingement-related joint contact will occur during motion [22–27,75,125]. Computer model impingement predictions only provide surrogate measures of joint contact such as overlap or intersection between two rigid bodies. The imaging approach inherently avoids several pitfalls associated with computer models:  Imaging avoids assumptions about physiological hip movement, whereas models that mathematically simulate hip motion often assume a static centre of hip rotation (which is a simplification).  No simplifications of hip tissue structures are required. Many models assume that the labrum/cartilage is rigid.  Imaging directly provides a visual representation of the hip, avoiding a registration step between computer models and kinematic data that can introduce error in measures of the relative position between the cam deformity and the acetabulum. A second strength of this research was that the open MRI measures of cam intrusion in the impingement test were related to sensor-measured acetabular contact force. No previous studies have related imaging measures to experimental mechanical measurements; yet without mechanical validation it is difficult to know whether computer models predict joint mechanics reliably [55]. A third strength of this study was the high resolution (0.9 mm isotropic voxels) that was achieved during imaging, which permitted multi-planar reformatting (MPR) and a full 3D investigation of clearance between the femoral head-neck junction and acetabulum. It is well established that using a single slice through the femoral neck runs the risk of underestimating the size of the cam deformity [4,6–8]; MPR is necessary to visualize the clearance between acetabulum and cam deformity about the entire circumference  99  of the femoral neck. The methods presented in this thesis are improvement on a past study using open MRI that did not report resolution, and used a bone-bone proximity of 3mm or less to define pincer-impingement ‘contact’ [132]; and on another that examined clearance between the acetabulum and cam deformity in a single plane only [9], and not in an impingement position. A fourth strength of this research was that the ex vivo hips obtained for this study ranged widely in femoral/acetabular shape, size, and orientation, including nine hips that had natural cam morphology. Other studies investigating impingement have used ‘simulated’ cam deformities, either via plastic appendages to the femoral head-neck junction [74,124] or varus osteotomy [107]. Simulated deformities permit the study of the same hip before and after addition of a cam lesion, but are severely limited because they do not necessarily lead to physiological hip impingement. This thesis used a lengthy screening process and collaboration with the specimen procurement company (Science Care; Phoenix, AZ, USA) to obtain the group of native cam hips. 7.3 Limitations One limitation of this study is that the contact force measurements were not carried out simultaneously with the MRI scans. Therefore the conditions of the hip joint were different when making sensor measurements (dissected hip, with sensor inserted) compared to when making MRI measurements (intact hip). However, the positioning technique did have good repeatability between repeat positioning trials compared to other studies that did repeated manual positioning trials and measured angles using goniometry. Furthermore, it would have been impossible to acquire images after dissecting the hip to instrument the acetabular cartilage with a piezoresistive sensor. The electrical leads in the sensor would have introduced imaging artifact at the region of interest, which would have made image reading near impossible. Also, dissection and disruption of the labral seal would have introduced air artifacts on the MR images, which would have also rendered image reading impossible.  100  A second limitation was that the acetabular contact force measurements were highly variable (43% CV in cam hips), so we could not define a quantitative relationship between β-angle on MRI and contact force, or make comparison in contact force between cam hips. High variability in force measurement was expected, since past studies using Tekscan sensors in cadaver joints showed variability in the 4-24% range [142–144], and the hip joint highly curved in comparison with previously studied joints, which accounts for our even higher variability. Assessing the force distribution data qualitatively, it was clear that cam hips experienced intra-articular contact force, while control hips did not. The measured contact force in cam hips was one to two orders of magnitude greater than contact force measured in control hips, and therefore it seemed reasonable to use a binary representation for contact force measurement. Furthermore, the binary representation was robust to large changes in the contact force threshold, lending support to the conclusion that there was a distinct separation between hips that experienced contact force and hips that experienced negligible contact force. A third limitation was that we studied hips in the anterior impingement test position, which is a passive exam (i.e. there is not muscle activity); it is not yet clear whether the findings for MRI-observed cam intrusion or elevated contact force  readings would hold in active postures. Externally applied load is not necessarily representative of physiological active loading. Muscle activation may drastically change the loading conditions in the hip joint, but simulation of muscular loading was not in the scope of this ex vivo model. A potential direction for future in vivo cam FAI research would be to explore the effects of muscle loading on cam intrusion using open MRI. Since this was the first study to explore cam intrusion using imaging, it made sense to begin with a posture most likely to produce direct impingement, which is the passive anterior impingement test. Our finding that the cam lesion intrudes medial to the acetabular rim during the anterior impingement test provides an important point of reference for future in vivo open MRI investigation of any variety of active postures.  101  7.4 Significance The finding that the femoral head-neck junction intruded medial to the acetabular margin in cam hips, but not controls, represents the first direct non-computer-model evidence for cam intrusion in an impingement position. Cam intrusion is a relatively simple concept central to the cam FAI hypothesis, but to the best of my knowledge only one computer model study has demonstrated the cam intrusion event [26]. Furthermore, this thesis found that acetabular contact force was elevated in cam hips compared to controls, which supports the hypothesis that cam FAI causes abnormal joint mechanics that lead to hip OA. This thesis also found that cam intrusion was linked with acetabular contact force, which suggests that the cam intrusion sign on open MRI is a useful surrogate measure of direct impingement. This is valuable for in vivo studies which can investigate the many questions that we still must answer in order to confirm that cam FAI causes hip OA. For example, the cam intrusion sign could be used to determine the importance of deformity size and role of activity on contact mechanics in vivo. The open MRI technique could also be used to evaluate how effectively certain treatment options (such as surgical resection of the cam deformity or physical therapy-based strengthening of muscles that cross the hip) alter contact mechanics. 7.5 Future Directions My first recommendation is that the open MRI method presented in this thesis be used to investigate cam intrusion in vivo, in a broad population (i.e. one that that includes both symptomatic and asymptomatic cam hips). This thesis showed that cam intrusion occurs in cam hips in a passive, ex vivo model; the next step is to assess whether cam intrusion occurs in vivo. Such a study could also investigate the effect of different postures on cam intrusion, particularly ones that are relevant to sports and daily activity, in order to determine high-risk activities (and potentially recommend activity modification). Since it clear that impingement patterns are highly subject-specific [25–27], an important component of this study may be to use patient-specific computer models to predict MRI postures that might be worst-case scenarios for each  102  individual hip, rather than standardize the impingement scenarios into a single posture like the anterior impingement exam. My second recommendation is that this open MRI method be used to investigate the etiology of the cam deformity, one area that remains under-studied in cam FAI. Open MRI could be used to track cam intrusion longitudinally in a population of asymptomatic adolescents that has higher risk for cam FAI (i.e. most sports-playing populations [82,84–87,182]). Each hip could be evaluated for cam intrusion in an impingement position at regular follow-up intervals, and monitored for changes in shape of the femoral head-neck junction, to see if the cam deformity becomes larger over time and leads to progressively worse cam intrusion. A study of this nature could also track risk factors in the adolescent population, such as activity levels and body mass index, to better understand the development of cam deformities and cam impingement, and ultimately predict as early as possible which hips will develop FAI-related OA. 7.6 Conclusion This thesis presented an open MRI technique that detected intrusion of femoral head-neck junction medial to the acetabular rim during the anterior impingement test in cam hips, but not in controls. The open MRI measures of cam intrusion were related to direct, sensor-based measurement of acetabular contact force, which supports the use of open MRI to investigate cam impingement in vivo. Open MRI can provide direct image-based assessment of cam FAI which has advantages over previous approaches that rely on simplifications or assumptions about hip joint biomechanics. Open MRI has significant potential to explore key questions about contact mechanics in femoroacetabular impingement and the progression of hip osteoarthritis.    103  References [1] Ganz, R., Parvizi, J., Beck, M., Leunig, M., Nötzli, H., and Siebenrock, K. A., 2003, “Femoroacetabular Impingement: A Cause for Osteoarthritis of the Hip,” Clin. Orthop. Relat. Res., (417), pp. 112–20. [2] Ito, K., Minka, M., Leunig, M., Werlen, S., and Ganz, R., 2001, “Femoroacetabular impingement and the cam-effect,” J. Bone Jt. Surg. Br. Ed., 83(B), pp. 171–6. [3] Nötzli, H. P., Wyss, T. F., Stoecklin, C. H., Schmid, M. R., Treiber, K., and Hodler, J., 2002, “The contour of the femoral head-neck junction as a predictor for the risk of anterior impingement,” J. Bone Jt. Surg. Br. Ed., 84(4), pp. 556–60. [4] Sutter, R., Dietrich, T. J., Zingg, P. O., and Pfirrmann, C. W. A., 2012, “How Useful is the Alpha Angle for Discriminating between Symptomatic Patients with Cam-type Femoroacetabular Impingement and Asymptomatic Volunteers,” Radiology, 264(2), pp. 514–21. [5] Meyer, D. C., Beck, M., Ellis, T., Ganz, R., and Leunig, M., 2006, “Comparison of Six Radiographic Projections to Assess Femoral Head/Neck Asphericity,” Clin. Orthop. Relat. Res., (445), pp. 181–5. [6] Dudda, M., Albers, C., Mamisch, T. C., Werlen, S., and Beck, M., 2009, “Do normal radiographs exclude asphericity of the femoral head-neck junction?,” Clin. Orthop. Relat. Res., 467(3), pp. 651–9. [7] Barton, C., Salineros, M. J., Rakhra, K. S., and Beaulé, P. E., 2011, “Validity of the alpha angle measurement on plain radiographs in the evaluation of cam-type femoroacetabular impingement.,” Clin. Orthop. Relat. Res., 469(2), pp. 464–9. [8] Audenaert, E., Baelde, N., Huysse, W., Vigneron, L., and Pattyn, C., 2011, “Development of a three-dimensional detection method of cam deformities in femoroacetabular impingement,” Skeletal Radiol., 40(7), pp. 921–7. [9] Wyss, T. F., Clark, J. M., Weishaupt, D., and Nötzli, H. P., 2007, “Correlation between internal rotation and bony anatomy in the hip.,” Clin. Orthop. Relat. Res., 460(460), pp. 152–8. [10] Clohisy, J. C., Knaus, E. R., Hunt, D. M., Lesher, J. M., Harris-Hayes, M., and Prather, H., 2009, “Clinical presentation of patients with symptomatic anterior hip impingement.,” Clin. Orthop. Relat. Res., 467(3), pp. 638–44. [11] Philippon, M. J., Maxwell, R. B., Johnston, T. L., Schenker, M., and Briggs, K. K., 2007, “Clinical presentation of femoroacetabular impingement,” Knee Surgery, Sport. Traumatol. Arthrosc., 15, pp. 1041–1047. [12] Tannast, M., Siebenrock, K. A., and Anderson, S. E., 2007, “Femoroacetabular Impingement: Radiographic Diagnosis - What the Radiologist Should Know,” Am. J. Roentgenol., 188(6), pp. 1540–52.  104  [13] Lung, R., O’Brien, J., Grebenyuk, J., Forster, B. B., De Vera, M., Kopec, J., Ratzlaff, C., Garbuz, D., Prlic, H., and Esdaile, J. M., 2012, “The prevalence of radiographic femoroacetabular impingement in younger individuals undergoing total hip replacement for osteoarthritis,” Clin. Rheumatol., 31, pp. 1239–1242. [14] Beck, M., Leunig, M., Parvizi, J., Boutier, V., Wyss, D., and Ganz, R., 2004, “Anterior Femoroacetabular Impingement: Part II. Midterm Results of Surgical Treatment,” Clin. Orthop. Relat. Res., 418, pp. 67–73. [15] Gosvig, K. K., Jacobsen, S., Sonne-Holm, S., Palm, H., and Troelsen, A., 2010, “Prevalence of malformations of the hip joint and their relationship to sex, groin pain, and risk of osteoarthritis: a population-based survey.,” J. Bone Joint Surg. Am., 92, pp. 1162–1169. [16] Hack, K., Di Primio, G., Rakhra, K., and Beaulé, P. E., 2010, “Prevalence of cam-type femoroacetabular impingement morphology in asymptomatic volunteers,” J. Bone Jt. Surg., 92(14), pp. 2436–44. [17] Laborie, L. B., Lehmann, T. G., Engesæter, I. Ø., Eastwood, D. M., Engesæter, L. B., and Rosendahl, K., 2011, “Prevalence of Radiographic Findings Thought to Be Associated with Femoroacetabular Impingement in a Population-based Cohort of 2081 Healthy Young Adults,” Radiology, 260(2), pp. 494–502. [18] Kang, A. C. L., Gooding, A. J., Coates, M. H., Goh, T. D., Armour, P., and Rietveld, J., 2010, “Computed tomography assessment of hip joints in asymptomatic individuals in relation to femoroacetabular impingement.,” Am. J. Sports Med., 38, pp. 1160–1165. [19] Hartofilakidis, G., Bardakos, N. V, Babis, G. C., and Georgiades, G., 2011, “An examination of the association between different morphotypes of femoroacetabular impingement in asymptomatic subjects and the development of osteoarthritis of the hip.,” J. Bone Joint Surg. Br., 93(5), pp. 580–6. [20] Bardakos, N. V, and Villar, R. N., 2009, “Predictors of progression of osteoarthritis in femoroacetabular impingement: a radiological study with a minimum of ten years follow-up.,” J. Bone Jt. Surg., 91-B(2), pp. 162–169. [21] Wilson, D. R., McWalter, E. J., and Johnston, J. D., 2009, “The measurement of joint mechanics and their role in osteoarthritis genesis and progression.,” Med. Clin. North Am., 93(1), pp. 67–82. [22] Kubiak-Langer, M., Tannast, M., Murphy, S. B., Siebenrock, K. a, and Langlotz, F., 2007, “Range of motion in anterior femoroacetabular impingement.,” Clin. Orthop. Relat. Res., 458, pp. 117–24. [23] Tannast, M., Kubiak-langer, M., Langlotz, F., Puls, M., Murphy, S. B., and Siebenrock, K. A., 2007, “Noninvasive Three-Dimensional Assessment of Femoroacetabular Impingement,” J. Orthop. Res., (January), pp. 122–131. [24] Tannast, M., Goricki, D., Beck, M., Murphy, S. B., and Siebenrock, K. a, 2008, “Hip Damage Occurs at the Zone of femoroacetabular impingement.,” Clin. Orthop. Relat. Res., 466(2), pp. 273–80.  105  [25] Audenaert, E. A., Peeters, I., Vigneron, L., Baelde, N., and Pattyn, C., 2012, “Hip morphological characteristics and range of internal rotation in femoroacetabular impingement,” Am. J. Sports Med., 40(6), pp. 1329–36. [26] Audenaert, E. A., Mahieu, P., and Pattyn, C., 2011, “Three-dimensional assessment of cam engagement in femoroacetabular impingement,” Arthroscopy, 27(2), pp. 167–71. [27] Kapron, A. L., Aoki, S. K., Peters, C. L., and Anderson, A. E., 2014, “Subject-specific Patterns of Femur-labrum Contact are Complex and Vary in Asymptomatic Hips and Hips With Femoroacetabular Impingement,” Clin. Orthop. Relat. Res., 472(12), pp. 3912–22. [28] Bedi, A., Dolan, M., Magennis, E., Lipman, J., Buly, R., and Kelly, B. T., 2012, “Computer-assisted modeling of osseous impingement and resection in femoroacetabular impingement,” Arthroscopy, 28(2), pp. 204–10. [29] Ng, K. C. G., Rouhi, G., Lamontagne, M., and Beaulé, P. E., 2012, “Finite Element Analysis Examining the Effects of Cam FAI on Hip Joint Mechanical Loading Using Subject-Specific Geometries During Standing and Maximum Squat.,” HSS J., 8(3), pp. 206–12. [30] Jorge, J. P., Simões, F. M. F., Pires, E. B., Rego, P. A., Tavares, D. G., Lopes, D. S., and Gaspar, A., 2014, “Finite element simulations of a hip joint with femoroacetabular impingement,” Comput. Methods Biomech. Biomed. Engin., 17(11), pp. 1275–84. [31] Chegini, S., Beck, M., and Ferguson, S. J., 2009, “The effects of impingement and dysplasia on stress distributions in the hip joint during sitting and walking: a finite element analysis,” J. Orthop. Res., 27(2), pp. 195–201. [32] Michaeli, D. A., Murphy, S. B., and Hipp, J. A., 1997, “Comparison of predicted and measured contact pressures in normal and dysplastic hips.,” Med. Eng. Phys., 19(2), pp. 180–6. [33] Anderson, A. E., Ellis, B. J., Maas, S. a, Peters, C. L., and Weiss, J. a, 2008, “Validation of finite element predictions of cartilage contact pressure in the human hip joint.,” J. Biomech. Eng., 130(5), p. 051008. [34] Afoke, N. Y. P., Byers, P. D., and Hutton, W. C., 1987, “Contact Pressures in the Human Hip Joint,” J. Bone Jt. Surg., 69-B(4), pp. 536–541. [35] Sparks, D. R., Beason, D. P., Etheridge, B. S., Alonso, J. E., and Eberhardt, A. W., 2005, “Contact pressures in the flexed hip joint during lateral trochanteric loading,” 23, pp. 359–366. [36] Hak, D. J., Hamel, A. J., Bay, B. K., Sharkley, N. A., and Olson, S. A., 1998, “Consequences of Transverse Acetabular Fracture Malreduction on Load Transmission Across the Hip Joint,” J. Orthop. Trauma, 12(2), pp. 90–100. [37] Konrath, G. A., Hamel, A. J., Sharkey, N. a, Bay, B. K., and Olson, S. A., 1998, “Conseqeucnes of Anterior Column Fracture of the Acetabulum,” J. Orthop. Trauma, 12(8), pp. 547–52.  106  [38] Konrath, G. a, Hamel, A. J., Guerin, J., Olson, S. A., Bay, B. K., and Sharkey, N. a, 1999, “Biomechanical Evaluation of Impaction Fractures of the Femoral Head,” J. Orthop. Trauma, 13(6), pp. 407–13. [39] Konrath, G. A., Hamel, A. J., Sharkley, N. A., Bay, B. K., and Olson, S. A., 1998, “Biomechanical Evaluation of a Low Anterior Wall Fracture: Correlation with the CT Subchondral Arc,” J. Orthop. Trauma, 12(3), pp. 152–8. [40] Konrath, G. a, Hamel, a J., Olson, S. a, Bay, B., and Sharkey, N. a, 1998, “The role of the acetabular labrum and the transverse acetabular ligament in load transmission in the hip.,” J. Bone Joint Surg. Am., 80, pp. 1781–1788. [41] Olson, S. A., Bay, B. K., Pollak, A. N., Sharkey, N. a, and Lee, T., 1996, “The Effect of Variable Size Posterior Wall Acetabular Fractures on Contact Characteristics of the Hip Joint,” J. Orthop. Trauma, 10(6), pp. 395–402. [42] Olson, S. a, Bay, B. K., Chapman, M. W., and Sharkey, N. a, 1995, “Biomechanical consequences of fracture and repair of the posterior wall of the acetabulum.,” J. Bone Jt. Surg., 77, pp. 1184–1192. [43] Adams, D., Kempson, G. E., and Swanson, S. A. V, 1978, “Direct Measurement of Local Pressures in the Cadaveric Human Hip Joint,” Med. Biol. Eng. Comput., (January), pp. 113–115. [44] Adams, D., and Swanson, S. A. V, 1985, “Direct measurement of local pressures in the cadaveric human hip joint during simulated level walking,” Ann. Rheum. Dis., 44(March), pp. 658–66. [45] Mizrahi, J., Solomon, L., Kaufman, B., and Duggan, T. O., 1981, “An Experimental Method for Investigating Load Distribution in the Cadaveric Human Hip,” J. Bone Jt. Surg., 63-B(4), pp. 610–613. [46] Brown, T. D., and Shaw, D. T., 1983, “In vitro contact stress distributions in the natural human hip,” J. Biomech., 16(6), pp. 373–84. [47] Rudert, M. J., Ellis, B. J., Henak, C. R., Stroud, N. J., Pederson, D. R., Weiss, J. a, Brown, T. D., and Pedersen, D. R., 2014, “A new sensor for measurement of dynamic contact stress in the hip.,” J. Biomech. Eng., 136(3), pp. 035001–1–8. [48] Menschik, F., 1997, “The Hip Joint as a Conchoid Shape,” J. Biomech., 30(9), pp. 971–973. [49] Ferguson, S. J., Bryant, J. T., Ganz, R., and Ito, K., 2003, “An in vitro investigation of the acetabular labral seal in hip joint mechanics,” J. Biomech., 36(2), pp. 171–178. [50] Tan, V., Seldes, R. M., Katz, M. A., Freedhand, A. M., Klimkiewicz, J. J., and Fitzgerald, R. H., 2001, “Contribution of acetabular labrum to articulating surface area and femoral head coverage in adult hip joints: an anatomic study in cadavera.,” Am. J. Orthop., 30(11), pp. 809–812. [51] Fautrel, B., Hilliquin, P., Rozenberg, S., Allaert, F.-A., Coste, P., Leclerc, A., and Rossignol, M., 2005, “Impact of osteoarthritis: results of a nationwide survey of 10,000 patients consulting for OA,” Jt. Bone Spine, 72(3), pp. 235–40.  107  [52] Aigner, T., and Schmitz, N., 2011, “Pathogenesis and pathology of osteoarthritis,” Rheumatology, Elsevier Inc., pp. 1741–1759. [53] Bombardier, C., Hawker, G., and Mosher, D., 2011, The Impact of Arthritis in Canada: Today and Over the Next 30 Years. [54] Dagenais, S., Garbedian, S., and Wai, E. K., 2009, “Systematic review of the prevalence of radiographic primary hip Osteoarthritis,” Clin. Orthop. Relat. Res., 467, pp. 623–637. [55] Wilson, D. R., McWalter, E. J., and Johnston, J. D., 2013, “The Measurement of Joint Mechanics and Their Role in Osteoarthritis Genesis and Progression,” Rheum. Dis. Clin. North Am., 39, pp. 21–44. [56] Felson, D. T., 2013, “Osteoarthritis as a disease of mechanics,” Osteoarthr. Cartil., 21(1), pp. 10–5. [57] Bedi, A., Dolan, M., Leunig, M., and Kelly, B. T., 2011, “Static and Dynamic Mechanical Causes of Hip Pain,” Arthroscopy, 27(2), pp. 235–51. [58] Nepple, J. J., Vigdorchik, J. M., and Clohisy, J. C., 2014, “What is the Association Between Sports Participation and the Development of Proximal Femoral Cam Deformity? A Systematic Review and Meta-analysis,” Am. J. Sports Med. [59] Bedi, A., Dolan, M., Hetsroni, I., Magennis, E., Lipman, J., Buly, R., and Kelly, B. T., 2011, “Surgical treatment of femoroacetabular impingement improves hip kinematics: a computer-assisted model.,” Am. J. Sports Med., 39 Suppl, p. 43S–9S. [60] Leunig, M., Beaulé, P. E., and Ganz, R., 2009, “The concept of femoroacetabular impingement: current status and future perspectives.,” Clin. Orthop. Relat. Res., 467(3), pp. 616–22. [61] Beaule, P., Zaragoza, E., Motamedi, K., Copelan, N., and Dorey, F., 2005, “Three-dimensional computed tomography of the hip in the assessment of femoroacetabular impingement,” J. Orthop. Res., 23(6), pp. 1286–1292. [62] Beck, M., Chegini, S., and Ferguson, S., 2012, “Mechanism of Femoroacetabular Impingement,” Femoroacetabular Impingement, Ó. Marín-Peña, ed., Springer Berlin Heidelberg, Berlin, Heidelberg, pp. 9–22. [63] Jäger, M., Wild, A., Westhoff, B., and Krauspe, R., 2004, “Femoroacetabular impingement caused by a femoral osseous head-neck bump deformity: clinical, radiological, and experimental results,” J. Orthop. Sci., 9(3), pp. 256–63. [64] Strehl, A., and Ganz, R., 2005, “[Anterior femoroacetabular impingement after healed femoral neck fractures].,” Unfallchirurg, 108(4), pp. 263–73. [65] Krekel, P. R., Vochteloo, A. J., Bloem, R. M., and Nelissen, R. G., 2011, “Femoroacetabular impingement and its implications on range of motion: a case report.,” J. Med. Case Rep., 5(1), p. 143.  108  [66] Kapron, A. L., Anderson, A. E., Peters, C. L., Phillips, L. G., Stoddard, G. J., Petron, D. J., Toth, R., and Aoki, S. K., 2012, “Hip internal rotation is correlated to radiographic findings of cam femoroacetabular impingement in collegiate football players.,” Arthroscopy, 28(11), pp. 1661–70. [67] Dolan, M. M., Heyworth, B. E., Bedi, A., Duke, G., and Kelly, B. T., 2011, “CT reveals a high incidence of osseous abnormalities in hips with labral tears,” Clin. Orthop. Relat. Res., 469(3), pp. 831–8. [68] Siebenrock, K. a, Ferner, F., Noble, P. C., Santore, R. F., Werlen, S., and Mamisch, T. C., 2011, “The cam-type deformity of the proximal femur arises in childhood in response to vigorous sporting activity.,” Clin. Orthop. Relat. Res., 469(11), pp. 3229–40. [69] Audenaert, E., Vigneron, L., and Pattyn, C., 2011, “A method for three-dimensional evaluation and computer aided treatment of femoroacetabular impingement,” Comput. Aided Surg., 16(3), pp. 143–8. [70] Dandachli, W., Islam, S. U., Liu, M., Richards, R., Hall-Craggs, M., and Witt, J., 2009, “Three-dimensional CT analysis to determine acetabular retroversion and the implications for the management of femoro-acetabular impingement,” J. Bone Jt. Surg., 91(8), pp. 1031–6. [71] Siebenrock, K. A., Schoeniger, R., and Ganz, R., 2003, “Anterior Femoro-Acetabular Impingement Due to Acetabular Retroversion: Treatment with Periacetabular Osteotomy,” J. Bone Jt. Surg., 85-A(2), pp. 278–286. [72] Klaue, K., Durnin, C. W., and Ganz, R., 1991, “The Acetabular Rim Syndrome: A Clinical Presenation of Dysplasia of the Hip,” J. Bone Jt. Surg. Br. Ed., 73B(3), pp. 423–429. [73] Kapron, A. L., Aoki, S. K., Peters, C. L., and Anderson, A. E., 2014, “In Vivo Hip Kinematics During Clinical Exams Using Dual Fluoroscopy and Model-Based Tracking: Application to the Study of Femoroacetabular Impingement,” Orthopaedic Research Society Annual Meeting, New Orleans, LA, p. 242. [74] Birmingham, P. M., Kelly, B. T., Jacobs, R., McGrady, L., and Wang, M., 2012, “The Effect of Dynamic Femoroacetabular Impingement on Pubic Symphysis Motion: A Cadaveric Study,” Am. J. Sports Med., 40(5), pp. 1113–8. [75] Signorelli, C., Lopomo, N., Bonanzinga, T., Marcheggiani Muccioli, G. M., Safran, M. R., Marcacci, M., and Zaffagnini, S., 2013, “Relationship between femoroacetabular contact areas and hip position in the normal joint: an in vitro evaluation,” Knee Surgery, Sport. Traumatol. Arthrosc., 21(2), pp. 408–14. [76] Safran, M. R., Giordano, G., Lindsey, D. P., Gold, G. E., Rosenberg, J., Zaffagnini, S., and Giori, N. J., 2011, “Strains Across the Acetabular Labrum During Hip Motion: A Cadaveric Model,” Am. J. Sports Med., 39(Supp 1), p. 92S–102S. [77] Reichenbach, S., Leunig, M., Werlen, S., Nüesch, E., Pfirrmann, C. W., Bonel, H., Odermatt, A., Hofstetter, W., Ganz, R., and Jüni, P., 2011, “Association between cam-type deformities and magnetic resonance imaging-detected structural hip damage: a cross-sectional study in young men.,” Arthritis Rheum., 63(12), pp. 4023–30.  109  [78] Takeyama, A., Naito, M., Shiramizu, K., and Kiyama, T., 2009, “Prevalence of femoroacetabular impingement in Asian patients with osteoarthritis of the hip,” Int. Orthop., 33(January 1998), pp. 1229–1232. [79] Beaulé, P. E., Allen, D. J., Clohisy, J. C., Schoenecker, P. L., and Leunig, M., 2009, “The Young Adult with Hip Impingement: Deciding on the optimal intervention,” J. Bone Jt. Surg., 91A(1), pp. 209–21. [80] Siebenrock, K. A., Wahab, K. H. A., Werlen, S. A., Kalhor, M., Leunig, M., and Ganz, R., 2004, “Abnormal Extension of the Femoral Head Epiphysis as a Cause of Cam Impingement,” Clin. Orthop. Relat. Res., 418, pp. 54–60. [81] Kienle, K. P., Keck, J., Werlen, S., Kim, Y. J., Siebenrock, K. A., and Mamisch, T. C., 2012, “Femoral morphology and epiphyseal growth plate changes of the hip during maturation: MR assessments in a 1-year follow-up on a cross-sectional asymptomatic cohort in the age range of 9-17 years,” Skeletal Radiol., 41, pp. 1381–1390. [82] Kapron, A. L., Anderson, A. E., Aoki, S. K., Phillips, L. G., Petron, D. J., Toth, R., and Peters, C. L., 2011, “Radiographic prevalence of femoroacetabular impingement in collegiate football players: AAOS Exhibit Selection.,” J. Bone Joint Surg. Am., 93, pp. e111(1–10). [83] Agricola, R., Bessems, J. H. J. M., Ginai, a. Z., Heijboer, M. P., van der Heijden, R. a., Verhaar, J. a. N., Weinans, H., and Waarsing, J. H., 2012, “The Development of Cam-Type Deformity in Adolescent and Young Male Soccer Players,” Am. J. Sports Med., 40, pp. 1099–1106. [84] Ayeni, O. R., Banga, K., Bhandari, M., Maizlin, Z., de Sa, D., Golev, D., Harish, S., and Farrokhyar, F., 2014, “Femoroacetabular impingement in elite ice hockey players.,” Knee Surg. Sports Traumatol. Arthrosc., 22, pp. 920–5. [85] Siebenrock, K. a, Kaschka, I., Frauchiger, L., Werlen, S., and Schwab, J. M., 2013, “Prevalence of cam-type deformity and hip pain in elite ice hockey players before and after the end of growth.,” Am. J. Sports Med., 41, pp. 2308–13. [86] Mariconda, M., Cozzolino, A., Di Pietto, F., Ribas, M., Bellotti, V., and Soldati, A., 2014, “Radiographic findings of femoroacetabular impingement in capoeira players,” Knee Surgery, Sport. Traumatol. Arthrosc., 22, pp. 874–881. [87] Lahner, M., Bader, S., Walter, P. A., Duif, C., Pellengahr, C. V. S., Lukas, C., Ficklscherer, A., Fickert, S., and Hagen, M., 2014, “Prevalence of femoro-acetabular impingement in international competitive track and field athletes,” pp. 2571–2576. [88] Roels, P., Agricola, R., Oei, E. H., Weinans, H., Campoli, G., and Zadpoor, A. A., 2014, “Mechanical factors explain development of cam-type deformity,” Osteoarthr. Cartil., 22(12), pp. 2074–2082. [89] Eijer, H., Myers, S. R., and Ganz, R., 2001, “Anterior femoroacetabular impingement after femoral neck fractures,” J. Orthop. Trauma, 15(7), pp. 475–81.  110  [90] Snow, S. W., Keret, D., Scarangella, S., and Bowen, J. R., 1993, “Anterior Impingement of the Femoral Head: A late phenomenon of Legg-Calve-Perthes’ disease,” J. Paediatr. Orthop., 13(3), pp. 286–9. [91] Leunig, M., Casillas, M. M., Hamlet, M., Hersche, O., Noetzli, H., Slongo, T., and Ganz, R., 2000, “Slipped capital femoral epiphysis: early mechanical damage to the acetabular cartilage by a prominent femoral metaphysis,” Acta Orthop Scand, 71(4), pp. 370–375. [92] Rab, G. T., 1999, “The Geometry of Slipped Capital Femoral Epiphysis: Implications for Movement, Impingement, and Corrective Osteotomy,” J Pediatr Orthop, 19(4), pp. 419–424. [93] Rhyu, K. H., Kim, Y. H., Park, W. M., Kim, K., Cho, T.-J., and Choi, I. H., 2011, “Application of finite element analysis in pre-operative planning for deformity correction of abnormal hip joints - a case series,” Proc. Inst. Mech. Eng. Part H J. Eng. Med., 225(9), pp. 929–936. [94] Pfirrmann, C. W. A., Mengiardi, B., Dora, C., Kalberer, F., Zanetti, M., and Hodler, J., 2006, “Cam and Pincer Femoroacetabular Impingement: Characteristic MR Arthroghraphic Findings in 50 patients,” Radiology, 240(3), pp. 778–85. [95] Zilkens, C., Miese, F., Krauspe, R., and Bittersohl, B., 2013, “Symptomatic femoroacetabular impingement: does the offset decrease correlate with cartilage damage? A pilot study.,” Clin. Orthop. Relat. Res., 471(7), pp. 2173–82. [96] Lavigne, M., Jean-Michel, L., and Pascal-Andre, V., 2012, “Historical Evolution of the Concept of Femoroacetabular Impingement as a Cause of Hip Osteoarthritis,” Femoroacetabular Impingement, Ó. Marín-Peña, ed., Springer Berlin Heidelberg, Berlin, Heidelberg, pp. 3–7. [97] Murray, R. O., 1965, “The aetiology of primary osteoarthritis of the hip.,” Br. J. Radiol., 38(455), pp. 810–24. [98] Solomon, L., 1976, “Patterns of Osteoarthritis of the Hip,” J. Bone Jt. Surg., 58-B(2), pp. 176–183. [99] Harris, W. H., 1986, “Etiology of Osteoarthritis of the Hip,” Clin. Orthop. Relat. Res., 213, pp. 20–33. [100] Ganz, R., Gill, T. J., Gautier, E., Ganz, K., Krügel, N., and Berlemann, U., 2001, “Surgical dislocation of the adult hip,” J. Bone Jt. Surg., 83-B(8), pp. 1119–1124. [101] Resnick, 1976, “The ‘Tilt Deformity’ of the Femoral Head in Osteoarthritis of the Hip: A Poor Indicator of Previous Epiphysioloysis,” Clin. Radiol., 27(3), pp. 355–63. [102] Allen, D., Beaulé, P. E., Ramadan, O., and Doucette, S., 2009, “Prevalence of associated deformities and hip pain in patients with cam-type femoroacetabular impingement.,” J. Bone Jt. Surg., 91-B(5), pp. 589–594. [103] Agricola, R., Heijboer, M. P., Bierma-Zeinstra, S. M. a., Verhaar, J. a. N., Weinans, H., and Waarsing, J. H., 2012, “Cam impingement causes osteoarthritis of the hip: a nationwide prospective cohort study (CHECK),” Ann. Rheum. Dis., pp. 1–6.  111  [104] Clohisy, J. C., Dobson, M. a, Robison, J. F., Warth, L. C., Zheng, J., Liu, S. S., Yehyawi, T. M., and Callaghan, J. J., 2011, “Radiographic structural abnormalities associated with premature, natural hip-joint failure.,” J. Bone Jt. Surg., 93-A(S2), pp. 3–9. [105] Nicholls, A. S., Kiran, A., Pollard, T. C. B., Hart, D. J., Arden, C. P. a, Spector, T., Gill, H. S., Murray, D. W., Carr, A. J., and Arden, N. K., 2011, “The association between hip morphology parameters and nineteen-year risk of end-stage osteoarthritis of the hip: A nested case-control study,” Arthritis Rheum., 63(11), pp. 3392–3400. [106] Audenaert, E. A., Peeters, I., Van Onsem, S., and Pattyn, C., 2011, “Can we predict the natural course of femoroacetabular impingement?,” Acta, 77, pp. 188–196. [107] Siebenrock, K. A., Fiechter, R., Tannast, M., Mamisch, T. C., and von Rechenberg, B., 2013, “Experimentally Induced cam impingement in the sheep hip.,” J. Orthop. Res., 31(4), pp. 580–7. [108] Albers, C. E., Steppacher, S. D., Ganz, R., Siebenrock, K. A., and Tannast, M., 2012, “Joint-preserving surgery improves pain, range of motion, and abductor strength after Legg-Calvé-Perthes disease.,” Clin. Orthop. Relat. Res., 470(9), pp. 2450–61. [109] Almoussa, S., Barton, C., Speirs, A. D., Gofton, W., and Beaulé, P. E., 2011, “Computer-assisted correction of cam-type femoroacetabular impingement: a Sawbones study,” J. Bone Jt. Surg. Am. Ed., 93 Suppl 2, pp. 70–5. [110] Kelly, B. T., Bedi, A., Robertson, C. M., Dela Torre, K., Giveans, M. R., and Larson, C. M., 2012, “Alterations in internal rotation and alpha angles are associated with arthroscopic cam decompression in the hip.,” Am. J. Sports Med., 40(5), pp. 1107–12. [111] Leunig, M., Beck, M., Dora, C., and Ganz, R., 2005, “Femoroacetabular Impingement: Etiology and Surgical Concept,” Oper. Tech. Orthop., 15(3), pp. 247–255. [112] Naal, F. D., Miozzari, H. H., Wyss, T. F., and Nötzli, H. P., 2011, “Surgical hip dislocation for the treatment of femoroacetabular impingement in high-level athletes.,” Am. J. Sports Med., 39(3), pp. 544–50. [113] Naal, F. D., Miozzari, H. H., Schär, M., Hesper, T., and Nötzli, H. P., 2012, “Midterm results of surgical hip dislocation for the treatment of femoroacetabular impingement.,” Am. J. Sports Med., 40(7), pp. 1501–10. [114] Philippon, M. J., Stubbs, A. J., Schenker, M. L., Maxwell, R. B., Ganz, R., and Leunig, M., 2007, “Arthroscopic management of femoroacetabular impingement: osteoplasty technique and literature review.,” Am. J. Sports Med., 35(9), pp. 1571–80. [115] Ecker, T. M., Puls, M., Steppacher, S. D., Bastian, J. D., Keel, M. J. B., Siebenrock, K. A., and Tannast, M., 2012, “Computer-assisted femoral head-neck osteochondroplasty using a surgical milling device an in vitro accuracy study,” J. Arthroplasty, 27(2), pp. 310–6. [116] Hunt, M. A., Gunether, J. R., and Gilbart, M. K., 2013, “Kinematic and kinetic differences during walking in patients with and without symptomatic femoroacetabular impingement,” Clin. Biomech., 28(5), pp. 519–523.  112  [117] Kennedy, M. J., Lamontagne, M., and Beaulé, P. E., 2009, “Femoroacetabular impingement alters hip and pelvic biomechanics during gait. Walking biomechanics of FAI,” Gait Posture, 30, pp. 41–44. [118] Rylander, J., Shu, B., Favre, J., Safran, M., and Andriacchi, T., 2013, “Functional testing provides unique insights into the pathomechanics of femoroacetabular impingement and an objective basis for evaluating treatment outcome,” J. Orthop. Res., 31(SEPTEMBER), pp. 1461–1468. [119] Rylander, J. H., Shu, B., Andriacchi, T. P., and Safran, M. R., 2011, “Preoperative and postoperative sagittal plane hip kinematics in patients with femoroacetabular impingement during level walking.,” Am. J. Sports Med., 39 Suppl, p. 36S–42S. [120] Casartelli, N. C., Maffiuletti, N. a., Item-Glatthorn, J. F., Staehli, S., Bizzini, M., Impellizzeri, F. M., and Leunig, M., 2011, “Hip muscle weakness in patients with symptomatic femoroacetabular impingement,” Osteoarthr. Cartil., 19, pp. 816–821. [121] Yazbek, P. M., Ovanessian, V., Martin, R. L., and Fukuda, T. Y., 2011, “Nonsurgical treatment of acetabular labrum tears: a case series.,” J. Orthop. Sports Phys. Ther., 41(5), pp. 346–353. [122] Bennell, K. L., O’Donnell, J. M., Takla, A., Spiers, L. N., Hunter, D. J., Staples, M., and Hinman, R. S., 2014, “Efficacy of a physiotherapy rehabilitation program for individuals undergoing arthroscopic management of femoroacetabular impingement - the FAIR trial: a randomised controlled trial protocol.,” BMC Musculoskelet. Disord., 15, p. 58. [123] Puls, M., Ecker, T. M., Tannast, M., Steppacher, S. D., Siebenrock, K. A., and Kowal, J. H., 2010, “The Equidistant Method - a novel hip joint simulation algorithm for detection of femoroacetabular impingement,” Comput. Aided Surg., 15(4-6), pp. 75–82. [124] Given, L. E., 2010, “Effect of Cam-Type Femoroacetabular Impingement on Hip Joint Kinematics,” University of British Columbia. [125] Charbonnier, C., Assassi, L., Volino, P., and Magnenat-Thalmann, N., 2009, “Motion study of the hip joint in extreme postures,” Image Vis. Comput., 25(9), pp. 873–882. [126] Larson, C. M., Pierce, B. R., and Giveans, M. R., 2011, “Treatment of Athletes with Symptomatic Intra-articular Hip Pathology and Athletic Pubalgia/Sports Hernia: a case series,” Arthroscopy, 27(6), pp. 768–75. [127] Lamontagne, M., Kennedy, M. J., and Beaulé, P. E., 2009, “The effect of cam FAI on hip and pelvic motion during maximum squat.,” Clin. Orthop. Relat. Res., 467(3), pp. 645–50. [128] Dwyer, M. K., Jones, H. L., Field, R. E., McCarthy, J. C., and Noble, P. C., 2014, “Femoroacetabular Impingement Negates the Acetabular Labral Seal During Pivoting Maneuvers but Not Gait,” Clin. Orthop. Relat. Res. [129] Speirs, A. D., Beaulé, P. E., Rakhra, K. S., Schweitzer, M. E., and Frei, H., 2013, “Increased acetabular subchondral bone density is associated with cam-type femoroacetabular impingement.,” Osteoarthr. Cartil., 21(4), pp. 551–8.  113  [130] Beck, M., Kalhor, M., Leunig, M., and Ganz, R., 2005, “Hip morphology influences the pattern of damage to the acetabular cartilage: femoroacetabular impingement as a cause of early osteoarthritis of the hip,” J. Bone Jt. Surg., 87-B(7), pp. 1012–8. [131] Siebenrock, K. A., Kalbermatten, D. F., and Ganz, R., 2003, “Effect of Pelvic Tilt on Acetabular Retroversion: A Study of Pelves,” Clin. Orthop. Relat. Res., (407), pp. 241–248. [132] Yamamura, M., Miki, H., Nakamura, N., Murai, M., Yoshikawa, H., and Sugano, N., 2007, “Open-Configuration MRI Study of Femoro-Acetabular Impingement,” J. Orthop. Res., 25(12), pp. 1582–1588. [133] Butz, K. D., Chan, D. D., Nauman, E. a., and Neu, C. P., 2011, “Stress distributions and material properties determined in articular cartilage from MRI-based finite strains,” J. Biomech., 44(15), pp. 2667–2672. [134] Eckstein, F., Lemberger, B., Stammberger, T., Englmeier, K. H., and Reiser, M., 2000, “Patellar cartilage deformation in vivo after static versus dynamic loading,” J. Biomech., 33, pp. 819–825. [135] Song, Y., Greve, J. M., Carter, D. R., Koo, S., and Giori, N. J., 2006, “Articular cartilage MR imaging and thickness mapping of a loaded knee joint before and after meniscectomy,” Osteoarthr. Cartil., 14, pp. 728–737. [136] Herberhold, C., Faber, S., Stammberger, T., Steinlechner, M., Putz, R., Englmeier, K. H., Reiser, M., and Eckstein, F., 1999, “In situ measurement of articular cartilage deformation in intact femoropatellar joints under static loading,” J. Biomech., 32, pp. 1287–1295. [137] Nishii, T., Shiomi, T., Tanaka, H., Yamazaki, Y., Murase, K., and Sugano, N., 2010, “Loaded Cartilage T2 Mapping in Patients with Hip Dysplasia,” Radiology, 256(3), pp. 955–965. [138] Bergmann, G., Deuretzbacher, G., Heller, M., Graichen, F., Rohlmann, a., Strauss, J., and Duda, G. N., 2001, “Hip contact forces and gait patterns from routine activities,” J. Biomech., 34, pp. 859–871. [139] Eisenhart, R. Von, Adam, C., Steinlechner, M., and Eckstein, F., 1999, “Quantitative Determination of Joint Incongruity and Pressure Distribution during Simulated Gait and Cartilage Thickness in the Human Hip Joint,” (13), pp. 532–539. [140] Afoke, N. Y. P., Byers, P. D., and Hutton, W. C., 1980, “The Incongruous Hip Joint: A Casting Study,” J. Bone Jt. Surg., 62-B(4), pp. 511–514. [141] Afoke, N. Y. P., Byers, P. D., and Hutton, W. C., 1984, “The Incongruous Hip Joint: A Loading Study,” Ann. Rheum. Dis., 43, pp. 295–301. [142] Wilson, D. R., Apreleva, M. V., Eichler, M. J., and Harrold, F. R., 2003, “Accuracy and repeatability of a pressure measurement system in the patellofemoral joint,” J. Biomech., 36(12), pp. 1909–1915.  114  [143] Van Haver, A., Quintelier, J., Beule, M. De, Verdonk, P., Almqvist, F., and De Baets, P., 2010, “Patellofemoral Contact during Simulated Weight Bearing Squat Movement: A Cadaveric Study,” IFMBE Proc., 29, pp. 639–642. [144] Wilson, D. C., Niosi, C. A., Zhu, Q. A., Oxland, T. R., and Wilson, D. R., 2006, “Accuracy and repeatability of a new method for measuring facet loads in the lumbar spine,” J. Biomech., 39(2), pp. 348–53. [145] Harris, M. L., Morberg, P., Bruce, W. J. M., and Walsh, W. R., 1999, “An improved method for measuring tibiofemoral contact areas in total knee arthroplasty : a comparison of K-scan sensor and Fuji " lm,” 32, pp. 951–958. [146] Martinelli, L., Hurschler, C., and Rosenbaum, D., 2006, “Comparison of capacitive versus resistive joint contact stress sensors.,” Clin. Orthop. Relat. Res., 447(447), pp. 214–20. [147] Bittersohl, B., Zilkens, C., Kim, Y.-J., Werlen, S., Siebenrock, K. a, Mamisch, T. C., and Hosalkar, H. S., 2011, “Delayed gadolinium-enhanced magnetic resonance imaging of hip joint cartilage: pearls and pitfalls.,” Orthop. Rev. (Pavia)., 3(2), p. e11. [148] Milner, G., and Boldsen, J., 2012, “Humeral and femoral head diameters in recent white American skeletons,” J. Forensic Sci., 57(1), pp. 35–40. [149] Gold, G. E., Besier, T. F., Draper, C. E., Asakawa, D. S., Delp, S. L., and Beaupre, G. S., 2004, “Weight-bearing MRI of patellofemoral joint cartilage contact area,” J. Magn. Reson. Imaging, 20(3), pp. 526–30. [150] McWalter, E. J., 2010, “In vivo Assessments of Patellofemoral Kinematics and Contact Areas with Applications to Osteoarthritis,” University of British Columbia. [151] Kundel, H. L., and Polansky, M., 2003, “Measurement of Observer Agreement,” Radiology, 228, pp. 303–308. [152] Landis, J. R., and Koch, G. G., 1977, “The Measurement of Observer Agreement for Categorical Data,” Biometrics, 33(1), pp. 159–174. [153] Khan, M., Ranawat, A., and Williams, D., 2014, “Relationship between the alpha and beta angles in diagnosing CAM-type femoroacetabular impingement on frogeg lateral radiographs,” Knee Surgery, Sport. Traumatol. Arthrosc. [154] Kang, R. W., Yanke, A. B., Orias, A. E., Inoue, N., and Nho, S. J., 2013, “Emerging ideas: novel 3-d quantification and classification of cam lesions in patients with femoroacetabular impingement.,” Clin. Orthop. Relat. Res., 471(2), pp. 358–62. [155] Kang, X., Zhang, H., Garbuz, D., Wilson, D. R., and Hodgson, A. J., 2013, “Preliminary evaluation of an MRI-based technique for displaying and quantifying bony deformities in cam-type femoroacetabular impingement,” Int. J. Comput. Assist. Radiol. Surg., 8, pp. 967–975.  115  [156] Rakhra, K. S., Sheikh, A. M., Allen, D., and Beaulé, P. E., 2009, “Comparison of MRI alpha angle measurement planes in femoroacetabular impingement.,” Clin. Orthop. Relat. Res., 467(3), pp. 660–5. [157] Carty, F. L., Crosbie, I., Ryan, J., and Cashman, J. P., 2013, “Imaging of the acetabular labrum : a review,” 2(1), pp. 1–6. [158] Czerny, C., Hofmann, S., Neuhold, A., Tschauner, C., Engel, A., Recht, M., and Kramer, J., 1996, “Lesions of the acetabular labrum: accuracy of MR imaging and MR arthrography in detection and staging,” Radiology, 200(1), pp. 225–30. [159] Leunig, M., Werlen, S., Ungersbock, A., Ito, K., and Ganz, R., 1997, “Evaluation of the Acetabular Labrum by MR Arthography,” J. Bone Jt. Surg. Br. Ed., 79B(230-4), pp. 230–234. [160] Greaves, L. L., 2008, “Effect of Acetabular Labral Tears , Repair and Resection on Hip Cartilage Strains: A 7T MR Study,” University of British Columbia. [161] Greaves, L. L., Gilbart, M. K., Yung, A. C., Kozlowski, P., and Wilson, D. R., 2010, “Effect of acetabular labral tears, repair and resection on hip cartilage strain: A 7T MR study.,” J. Biomech., 43(5), pp. 858–63. [162] Grood, E. S., and Suntay, W. J., 1983, “A joint coordinate system for the clinical description of three-dimensional motions: application to the knee.,” J. Biomech. Eng., 105(2), pp. 136–44. [163] Wu, G., Siegler, S., Allard, P., Kirtley, C., Leardini, A., Rosenbaum, D., Whittle, M., D’Lima, D., Cristofolini, L., Witte, H., Schmid, O., and Stokes, I., 2002, “ISB recommendation on definitions of joint coordinate system of various joints for the reporting of human joint motion - part I: ankle, hip, and spine,” J. Biomech., 35, pp. 543–8. [164] Gluer, C., Blake, G., Lu, Y., Blunt, B. A., Jergas, M. I., and G, H. K., 1995, “Accurate Assessment of Precision Errors: How to Measure the Reproducibility of Bone Densitometry Techniques,” Osteoporos. Int., 5, pp. 262–270. [165] Nussbaumer, S., Leunig, M., Glatthorn, J. F., Stauffacher, S., Gerber, H., and Maffiuletti, N. A., 2010, “Validity and test-retest reliability of manual goniometers for measuring passive hip range of motion in femoroacetabular impingement patients,” BMC Musculoskelet. Disord. [166] Audenaert, E. A., Vigneron, L., Van Hoof, T., D’Herde, K., van Maele, G., Oosterlinck, D., and Pattyn, C., 2011, “In vitro validation and reliability study of electromagnetic skin sensors for evaluation of end range of motion positions of the hip.,” Med. Biol. Eng. Comput., 49(12), pp. 1405–12. [167] Cottrell, J. M., Scholten, P., Wanich, T., Warren, R. F., Wright, T. M., and Maher, S. a, 2008, “A new technique to measure the dynamic contact pressures on the Tibial Plateau.,” J. Biomech., 41(10), pp. 2324–9. [168] Bachus, K. N., DeMarco, A. L., Judd, K. T., Horwitz, D. S., and Brodke, D. S., 2006, “Measuring contact area, force, and pressure for bioengineering applications: using Fuji Film and TekScan systems.,” Med. Eng. Phys., 28(5), pp. 483–8.  116  [169] Brimacombe, J. M., Wilson, D. R., Hodgson, A. J., Ho, K. C. T., and Anglin, C., 2009, “Effect of calibration method on Tekscan sensor accuracy.,” J. Biomech. Eng., 131(3), p. 034503. [170] Harris, M. D., Anderson, A. E., Henak, C. R., Ellis, B. J., Peters, C. L., and Weiss, J. A., 2012, “Finite element prediction of cartilage contact stresses in normal human hips,” J. Orthop. Res., 30(7), pp. 1133–9. [171] Kapron, A. L., Aoki, S. K., Peters, C. L., Maas, S. A., Bey, M. J., Zauel, R., and Anderson, A. E., 2014, “Accuracy and Feasibility of Dual Fluoroscopy and Model-Based Tracking to Quantify in vivo Hip Kinematics During Clinical Exams,” J. Appl. Biomech., In press. [172] McPoil, T. G., Cornwall, M. W., and Yamada, W., 1995, “A Comparison of Two In-Shoe Plantar Pressure Measurement Systems,” Low. Extrem., 2(2), pp. 95–103. [173] Puls, M., Steppacher, S. D., Tannast, M., and Kowal, J., 2008, “Simulation of Hip Motion for Impingement Detection: A Comparison of Four Strategies,” J. Biomech., 41(S1), p. 2008. [174] Ross, J. R., Nepple, J. J., Philippon, M. J., Kelly, B. T., Larson, C. M., and Bedi, A., 2014, “Effect of Changes in Pelvic Tilt on Range of Motion to Impingement and Radiographic Parameters of Acetabular Morphologic Characteristics.,” Am. J. Sports Med. [175] Mamisch, T. C., Kain, M. S. H., Bittersohl, B., Apprich, S., Werlen, S., Beck, M., and Siebenrock, K. A., 2011, “Delayed gadolinium-enhanced magnetic resonance imaging of cartilage (dGEMRIC) in Femoacetabular impingement.,” J. Orthop. Res., 29(9), pp. 1305–11. [176] Bittersohl, B., Hosalkar, H. S., Apprich, S., Werlen, S. A., Siebenrock, K. A., and Mamisch, T. C., 2011, “Comparison of pre-operative dGEMRIC imaging with intra-operative findings in femoroacetabular impingement: preliminary findings,” Skeletal Radiol., 40(5), pp. 553–61. [177] Ganz, R., Leunig, M., Leunig-Ganz, K., and Harris, W. H., 2008, “The etiology of osteoarthritis of the hip: an integrated mechanical concept.,” Clin. Orthop. Relat. Res., 466(2), pp. 264–72. [178] Beaulé, P. E., O’Neill, M., and Rakhra, K., 2009, “Acetabular labral tears.,” J. Bone Joint Surg. Am., 91(3), pp. 701–10. [179] Jung, K. A., Restrepo, C., Hellman, M., AbdelSalam, H., Morrison, W., and Parvizi, J., 2011, “The prevalence of cam-type femoroacetabular deformity in asymptomatic adults,” J. Bone Joint Surg. Br., 93(10), pp. 1303–7. [180] Audenaert, E., Van Houcke, J., Maes, B., Vanden Bossche, L., Victor, J., and Pattyn, C., 2012, “Range of motion in femoroacetabular impingement,” Acta Orthop. Belg., 78(3), pp. 327–32. [181] Pollard, T. C. B., Villar, R. N., Norton, M. R., Fern, E. D., Williams, M. R., Simpson, D. J., Murray, D. W., and Carr, A. J., 2010, “Femoroacetabular impingement and classification of the cam deformity: the reference interval in normal hips.,” Acta Orthop., 81(1), pp. 134–41. [182] Nepple, J. J., Brophy, R. H., Matava, M. J., Wright, R. W., and Clohisy, J. C., 2012, “Radiographic findings of femoroacetabular impingement in National Football League Combine  117  athletes undergoing radiographs for previous hip or groin pain.,” Arthroscopy, 28(10), pp. 1396–403. [183] Konan, S., Rayan, F., Meermans, G., Witt, J., and Haddad, F. S., 2011, “Validation of the classification system for acetabular chondral lesions identified at arthroscopy in patients with femoroacetabular impingement.,” J. Bone Joint Surg. Br., 93, pp. 332–336. [184] Ilizaliturri, V. M., Byrd, J. W. T., Sampson, T. G., Guanche, C. a., Philippon, M. J., Kelly, B. T., Dienst, M., Mardones, R., Shonnard, P., and Larson, C. M., 2008, “A Geographic Zone Method to Describe Intra-articular Pathology in Hip Arthroscopy: Cadaveric Study and Preliminary Report,” Arthrosc. - J. Arthrosc. Relat. Surg., 24(5), pp. 534–539. [185] McDowell, M. A., Fryar, C. D., Ogden, C. L., and Flegal, K. M., 2008, Anthropometric Reference Data for Children and Adults: United States, 2003-2006. [186] Toogood, P. A., Skalak, A., and Cooperman, D. R., 2009, “Proximal femoral anatomy in the normal human population,” Clin. Orthop. Relat. Res., 467, pp. 876–885. [187] Myronenko, A., and Song, X., 2010, “Point set registration: coherent point drift,” IEEE Trans. Pattern Anal. Mach. Intell., 32(12), pp. 2262–75.   118  Appendix A Specimen Screening The objective of specimen screening was to acquire a group of hips with a cam deformity and a group without. Chapters 3-6 studied some or all of the hips listed in this Appendix. Ultimately, hips were grouped as cam (n=9) or control (n=3) based on the presence/absence of a cam deformity. Donor Screening and Imaging The pre-acquisition screening process (donor information, x-ray, and CT) was conducted in collaboration with a donor procurement company (Science Care; Phoenix AZ, USA) as follows: 1. Review of donor information. Donors less than 80 years were included (it was suggested by the donor company that narrower age inclusion criteria would make the order difficult to fill). Any donors with history of bone or joint disease were excluded. 2. Anterior-posterior (A-P) x-ray of the donor’s full pelvis (both hips included) according to the guidelines by Tannast et al. [12]. An orthopaedic surgeon reviewed the images for anatomy related to pincer or cam FAI. 3. CT image of the donor’s full pelvis (both hips included), acquired using helical CT (Toshiba Aquilion; Tustin, CA, USA) with 1 mm slice thickness and 0.80 mm in-plane voxel size. The raw image volumes were used to create 18 radial reformats using a standard multi-planar reformatting (MPR) technique [4,6,7]. We evaluated the α-angle [3] on all slices, and used the maximum, αmax, to define the presence of a cam deformity if αmax > 60° [4]. 4. Final grouping as ‘cam’ or ‘control’ based on the presence of a cam deformity (Table A.1).    119  Post-Dissection Evaluation A post-dissection screening process (macroscopic evaluation of cartilage/labral damage) was conducted by an experienced orthopaedic surgeon. The surgeon reported the severity of cartilage damage according to the University College Hospital (UCH) cartilage classification scale [183] and the location of damage according to a standardized six-region geographic reporting convention for the acetabulum [184] (Figure A.1). The UCH cartilage classification system classifies cartilage from grade 0 (healthy) to grade 4. Grade 1 is classified as loss of fixation to subchondral bone resulting in a ‘wave sign’. Grade 2 is classified as a cleavage tear (without delamination). Grade 3 is classified as delamination from subchondral bone. Grade 4 is classified as exposed bone. Furthermore, each cartilage lesion is grouped as ‘A’, ‘B’, or ‘C’ depending on whether the damage was located less than one third (‘A’), between one and two thirds (‘B’), or more than two thirds (‘C’) of the distance from the acetabular rim to cotyloid fossa [183].  Figure A.1: Diagrams showing a standard geographic representation of the acetabulum. (A) A right hip with zone 1 located anteriorly. (B) A left hip (with the geographic zones mirrored about the sagittal plane) with zone 1 located anteriorly. Zone 6 encompasses the cotyloid fossa only. © Reproduced from [184] with permission from Elsevier.    120Table A.1: Complete specimen screening information. Hip 1 2 3 4 5 6 7 8 9 10 11 12 Group CAM CAM CON CAM CAM CAM CAM CAM CON CON CAM CAM Donor Info Side L R L R L R L R L R L R Age (years) 62 72 73 75 64 78 Sex M F M M F M Height (cm) 185.4 162.6 182.9 175.3 152.4 177.8 Weight (kg) 93.0 40.8 75.7 125.6 46.2 127.0 X-ray CCD angle (°) 128 129 150 143 112 114 132 130 133 129 112 115 Coxa vara - - - - + + - - - - + + Coxa valga - - + + - - - - - - - - Head offset (mm) 0.7 0.5 0.8 0.5 1.0 1.1 0.5 0.4 0.8 0.9 0.1 0.1 Lateral centre- edge angle (°) 30 28 34 30 48 46 41 39 36 37 42 34 Profunda/protrusio - - - - - - - - - - - - Crossover sign - - - - + + + + - - - - CT αmax (°) 71 84 57 67 84 82 70 79 54 53 99 82 Location (°) of αmax (Figure 6.1) -5 80 -115 45 35 55 95 40 55 -130 0 35 Incongruent head - - - + - - - - - - - - Post-dissection Cartilage damage Zone Grade Depth 2,3 3 A 3,4 4 A 3,4 1 A 2 3 A 2,3 3 A 2 2 AB 2 3 A 3,4 4 A 2,3 3 A 3 3 A 2,3 3 AB 3 4 A Zone Grade Depth 2 2 B 3,4 3 A 2 3,4 AB 2,3,4 2 AB 4 4 A 3,4 2 A 3,4 4 A 3 3 AB 2,3 2 AB 2,3 2 AB 3 4 A - Zone Grade Depth - 2,3,4 2 AB 3 3 A - - - - - - - - - Labral Damage Zone Type 2 Radial tear  3 Partial thickness tear 2,3 Partial thickness Tear 2 Fraying, radial tear  3 Bucket handle tear 2 Fraying   2,3 Fraying   - - - 2 Fraying, partial thickness tear - Zone Type - - - - - 3,4 Labral loss - - - - - - Other secondary changes - - - - Anterior neck osteophyte Anterior neck osteophyte - - Chondrocalcinosis Chondrocalcinosis - -  121  Appendix B Hip Joint Coordinate System A joint coordinate system (JCS) defines the origin and clinical rotation axes of a joint so that rotations and translations can be reported. A variety of conventions for defining the JCS can be used, but in 2001 Wu et al., on behalf of the International Society of Biomechanics, recommended that the field of biomechanics adopt the ‘floating axis’ convention [163], originally presented by Grood and Suntay for the knee [162]. The key advantage of the floating axis convention is that the order of rotation does not matter. The first step in establishing a hip JCS is to create rigid coordinate systems for each body, femur (x, y, z) and acetabulum (X, Y, Z), using anatomical landmarks. The axes are formed as follows (Figure B.1): Pelvis  Origin: hip joint centre  Z-axis: parallel to a line connecting the right and left anterior superior iliac spines (ASIS), pointing to the right  X-axis: parallel to a line that lies in the plane defined by two ASISs and the mid-point of the two posterior superior iliac spines (PSIS), orthogonal to the Z-axis, pointing anteriorly  Y-axis: line perpendicular to both X and Z  Femur  Origin: hip joint centre  y-axis: line connecting the origin to the midpoint of the right and left femoral epicondyles, pointing superiorly  z-axis: parallel to a line that lies in the plane defined by two femoral epicondyles and the origin, orthogonal to the y-axis, pointing to the right  x-axis: line perpendicular to both z and y  122   Figure B.1: Illustration of the hip joint coordinate system. The fixed pelvic coordinate system (XYZ) and the fixed femoral coordinate system (xyz) are used to define the hip joint coordinate system (e1 e2 e3). © Reproduced from [163] with permission from Elsevier. The second step in establishing the hip JCS is to define two axes that are ‘body-fixed’, one from the pelvis and one from the femur, and one floating axis that is orthogonal to both and therefore ‘floats’ when the femur moves relative to the pelvis. The hip’s axes are given as follows:  Flexion/extension axis, e1: pelvic Z-axis  Internal/external rotation axis, e3: femoral y-axis  Abduction/adduction axis, or floating axis, e2: common axis perpendicular to both e1 and e3  Translation is represented as motion of the femur relative to the pelvis along each of these axes. Medial-lateral translation occurs along e1, superior-inferior translation occurs along e3, and anterior-posterior translation occurs along e2.   123  Appendix C Joint Position Measurement The joint angles and translations in Chapter 4 were measured by: 1. Creating a CT-based 3D model of each donor’s full pelvis and left and right proximal femurs, to generate a joint coordinate system; 2. Acquiring motion data during the anterior impingement test using optical tracking, with surface digitization to track key  landmarks and joint surfaces; and 3. Registering the digitized surfaces to the CT model so that the CT-derived pelvic/femoral axes can be used to calculate joint angles and translations during the anterior impingement test.  Creating the Joint Coordinate System A custom joint coordinate system for each hip was generated using CT-based models of each pelvis and proximal femur. CT images were acquired using helical CT (Toshiba Aquilion; Tustin, CA, USA; 1 mm slice thickness, 0.80 mm in-plane voxel size) by the donor agency at the time of specimen screening.  The field of view for each image included each donor’s full pelvis, and both left and right femurs up to mid-shaft. The images were segmented in the axial plane using a standard bone threshold technique (Mimics 15.0, Materialise; Leuven, Belgium) and automatically reconstructed into a three-dimensional model. The 3D models of the femur and pelvis were used to extract landmarks required for calculating the axes of the hip joint coordinate system described above (Appendix B, Figure B.1). For the pelvis, the three pelvic axes X, Y, and Z required the coordinates of the left and right ASISs, and left and right PSISs, so those landmarks were manually delineated from the 3D surface model of the pelvis using image processing software (3-Matic, Materialise; Leuven, Belgium). For the femur, because the cadaver specimen’s femoral condyles were required but not available, the femoral anatomical axes were calculated by approximating the y and z-axes using the proximal femur anatomy and average femoral version. Approximations were performed as follows:  124   The centre of the femoral head was defined as the origin. I segmented the portion of the femoral head that was covered by articular cartilage (excluding the femoral head-neck junction or cam deformity, and fovea capitis), and fit a least-squares sphere to estimate the femoral head centre.   The y-axis of the femur was defined by the origin and an estimated location of the midpoint of the epicondylar line. First, the femoral neck and shaft were segmented and each fit to a least squares cylinder to calculate the femoral neck and shaft axes, which were used to define a plane that contained the y-axis. The mid-point of the femoral epicondyles was determined by selecting a point on the femoral shaft axis that was one ‘hip-knee length’ in distance from the femoral head centre (anthropometric data was used to estimate the hip-knee length of each specimen [185]).   The z-axis of the femur (line connecting the two epidcondyles) was estimated by first determining a line that lied in the neck-shaft plane orthogonal to the y-axis,  then by rotating that line 10° anteriorly about the y-axis in order to simulate the 10° population average for femoral version [186]. The femoral x-axis was then calculated by the taking the cross-product between y and z. All calculations were carried out in custom software that was previously written in our lab (MATLAB 2012a, MathWorks; Natick, MA, USA). Motion Tracking and Surface Digitization Joint angles and translations in Chapter 4 were measured using an infrared motion capture system (Optotrak, Northern Digital; Waterloo, Ontario, Canada). Optotrak has a manufacturer-specified accuracy for detecting marker positioning of 0.1mm, with a resolution of 0.01 mm. I connected one Optotrak 3-marker rigid body to the each of the femur (attached to a rigid Steinman pin anchored in the femoral shaft) and pelvis (anchored in the iliac wing). During the anterior impingement test, the location data of each rigid body relative to the global origin was recorded at 5 Hz. Following test data acquisition, the hip was dissected (ensuring the rigid marker position was not disrupted) to digitize the joint surfaces/bony landmarks required for registration to a CT-based joint coordinate system, a method used previously [75]. Before use, a digitizing probe (Figure C.1) was calibrated by a pivoting  125  procedure to determine the location of the centre of probe tip and account for the diameter of the probe tip when digitizing surfaces. Probe-tip location data was collected at 20 Hz for 90 seconds per landmark. Surfaces or landmarks that were digitized on the femur included the femoral head cartilage, femoral neck, greater trochanter, inter-trochanteric ridge, and lesser trochanter. Surfaces or landmarks that were digitized on the hemi-pelvis included the lunate surface of the acetabulum, the anterior inferior iliac spine, the anterior superior iliac spine, and pubic tubercle, and ischial spine.   Figure C.1: Optotrak probe (Optotrak, Northern Digital; Waterloo, Ontario, Canada).  For each landmark, probe tip data was transformed into the coordinates of the associated rigid body (i.e. femur or pelvis) at each time point, meaning that each digitized surface was registered to its actual position for each time point during the impingement test. Registration Joint Angle Calculation The CT-based model was registered to the digitized landmarks in order to calculate joint angles/translations for the actual anterior impingement test positions. I manually delineated the landmarks from the CT model that corresponded to the digitized surface using commercial image processing software (3-Matic,  126  Materialise; Leuven, Belgium) with the goal of achieving a one to one correspondence between digitized and CT-derived surfaces. I used a rigid registration based on a coherent point drift algorithm [187]). Once the CT-derived anatomical axes were aligned to the actual joint position measured during motion tracking, I calculated joint angles and translations using custom software adapted from previous work [124] that was based on the conventions for joint angle calculation initially described by Grood and Suntay [162].    127  Appendix D Analysis of the Anterior Impingement Test Background The anterior impingement test is widely used in the clinical examination of hip pathology. The maneuver involves applying maximal internal rotation to the hip in 90° flexion [72] with a patient supine on the table (Figure D.1). The clinician applies two forces to internally rotate the hip until the hip reaches its limit of motion: (i) a linear force at the foot, Fclin foot, in the coronal plane, normal to the tibia; and (ii) a linear force at the knee, Fclin knee, acting antiparallel and equal to Fclin foot. Together, Fclin knee and Fclin foot keep the knee at a fixed location in space and induce a torque on the femur in the internal rotation direction.    Figure D.1: A schematic of the anterior impingement test. The two forces, 𝐹 clin foot and 𝐹 clin knee. induce a torque of the femur about the hip joint long axis in the internal rotation direction, indicated by the black arrow. Joint contact mechanics in the hip at the terminal position of the anterior impingement test are a result of reactions to the clinician-applied forces that generate torque about the internal rotation axis of the hip. In hips with FAI anatomy, it is believed that the maximal internal rotation causes the anterosuperior femoral head-neck junction (where cam lesions are typically located) to come in contact with the anterosuperior acetabular rim (labrum and cartilage). 𝐹 clin foot  𝐹 clin knee   128  It was therefore important to consider if the clinician-applied forces, which are applied linearly and eccentric to the hip joint, led to any additional forces (such as translational force on the femur relative to the acetabulum) that may affect our measurements of joint mechanics. The objective of this Appendix was to present a simplified free-body analysis of the anterior impingement test and ask: does the clinical anterior impingement test apply any translational reaction forces at the hip joint? Simplified Model In order to evaluate forces at the hip, the following assumptions were made:  With both the hip and knee at 90° flexion respectively, the lower leg lies in the coronal plane  The terminal position of the anterior impingement test is static A free body diagram (FBD) from a coronal view of the lower leg at the terminal position of the anterior impingement test is shown in Figure D.2, with reaction forces from the thigh that act on the knee.    129   Figure D.2: Free body diagram of the lower leg (coronal view) in the anterior impingement test. Fclin foot and Fclin knee are applied by the clinician and lie in the coronal plane. ?⃑⃑? knee,z is a reaction moment at the knee that acts about the long axis of the femur (normal to the page). The x-axis is defined by the axis of the tibia, and therefore is orthogonal to the clinician-applied forces (i.e. Fclin foot and Fclin knee act only in the y-direction). Evaluating static equilibrium at the knee, we get Eq. D.1-3 as shown in Figure D.2:  ∑𝐌z = 0  ?⃑⃑⃑? knee,z = −𝐅 clin foot ∗ Ltibia  (Eq D.1) ∑𝐅y = 0   𝐅 knee,y + 𝐅 clin knee + 𝐅 clin foot = 0  (Eq. D.2) ∑𝐅x = 0   𝐅 knee,x = 0    (Eq. D.3) We can assume that no other forces are present (i.e. translational reaction force at the knee in the z-direction, moments at the knee about the x- or y-axes) because of the assumption that both clinician-applied forces act in the coronal plane. A free body diagram of the thigh is shown in Figure D.3 from an axial view. KNEE FOOT  𝐅 foot ?⃑⃑⃑? knee,z x y (z positive out of page) Ltibia 𝐅 knee,x  𝐅 knee,y 𝐅 clin knee  𝐅 clin foot  CORONAL  130   Figure D.3: A free body diagram of the thigh (axial view) in the anterior impingement test. Note that because both applied clinician-applied forces act in the coronal plane, there is no linear reaction force at the knee in the z-direction. The hip and knee both lie in the same axial plane because the hip is flexed to 90°. Evaluating static equilibrium at the hip about the internal rotation axis (z-axis, Figure D.3):    ∑𝐌z = 0  ?⃑⃑⃑? hip,z = −?⃑⃑⃑? knee,z  (Eq. D.4) Eq. D.4 + Eq. D.1  ?⃑⃑⃑? hip,z = 𝐅 clin foot ∗ Ltibia (Eq. D.5) Eq. D.5 states that the internal rotation torque (moment about the z-axis, Figure D.3) is dictated by (i) the force that the clinician applies to the foot and (ii) the length from the knee to the applied force at the foot. Evaluating static equilibrium at the hip about the adduction axis (x-axis, Figure D.3): ∑𝐌x = 0  ?⃑⃑⃑? hip,x = −𝐅 knee,y ∗ Lthigh (Eq. D.6) KNEE HIP z y (x positive into page) ?⃑⃑⃑? knee,z (about z-axis) ?⃑⃑⃑? hip,z (about z-axis) 𝐅 knee,y 𝐅 hip,y AXIAL 𝐅 hip,z Lthigh ?⃑⃑⃑? hip,x (about x-axis)  131  Assuming that the hip is free to rotate about the adduction axis, there is no reaction moment about the adduction axis (i.e. there are no tissues resisting adduction moment when the hip is at 90° flexion, neutral adduction, and maximum internal rotation): ?⃑⃑⃑? hip,x = 0   (Eq. D.7) Then we can demonstrate the following: Eq. D.6 + Eq. D.7  𝐅 knee,y = 0   (Eq. D.8) Eq. D.8 + Eq. D.2  𝐅 clin knee + 𝐅 clin foot = 0  (Eq. D.9) Eq. D.9 tells us that in order for the position to be static (i.e. for the hip to remain at 90° flexion and not to abduct/adduct), the clinician-applied forces, Fclin foot and Fclin knee, must be equal in magnitude and antiparallel in direction. Evaluating static equilibrium at the hip along the z- and y-axes (Figure D.3), and also the x-axis (note: not shown in Figure D.3): ∑𝐅𝑧 = 0   𝐅 hip,z = 0   (Eq. D.10)    ∑𝐅y = 0   𝐅 hip,y + 𝐅 knee,y = 0   (Eq. D.11) Eq. D.11 + Eq. D.8  𝐅 hip,y = 0   (Eq. D.12)    ∑𝐅x = 0   𝐅 hip,x + 𝐅 knee,x = 0   (Eq. D.13) Eq. D.13 + Eq. D.3  𝐅 hip,x = 0   (Eq. D.14) Eq. D.10, Eq. D.12, and Eq. D.14 indicate that no linear reaction forces are present at the hip joint at the end-point of the anterior impingement test.  132  Conclusion Based on this simplified model, we can make the following statements about the clinical anterior impingement test:  Clinicians apply two equal antiparallel forces at the knee and foot to induce an internal rotation. Clinician-applied forces must be equal in magnitude and antiparallel in order to keep the knee stationary and not ‘collapse’ the hip joint position of 90° of hip flexion and neutral adduction.  If the clinician-applied forces are equal in magnitude and antiparallel, they do not create any linear reaction forces within the hip joint.

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0215937/manifest

Comment

Related Items