Dual fluoroscopy accurately captures in vivo dynamic motion of human joints, which can be visualized relative to reconstructed anatomy (e.g., arthrokinematics). Herein, a detailed protocol to quantify hip arthrokinematics during weight-bearing activities of daily living is presented, including the integration of dual fluoroscopy with traditional skin marker motion capture.
Several hip pathologies have been attributed to abnormal morphology with an underlying assumption of aberrant biomechanics. However, structure-function relationships at the joint level remain challenging to quantify due to difficulties in accurately measuring dynamic joint motion. The soft tissue artifact errors inherent in optical skin marker motion capture are exacerbated by the depth of the hip joint within the body and the large mass of soft tissue surrounding the joint. Thus, the complex relationship between bone shape and hip joint kinematics is more difficult to study accurately than in other joints. Herein, a protocol incorporating computed tomography (CT) arthrography, three-dimensional (3D) reconstruction of volumetric images, dual fluoroscopy, and optical motion capture to accurately measure the dynamic motion of the hip joint is presented. The technical and clinical studies that have applied dual fluoroscopy to study form-function relationships of the hip using this protocol are summarized, and the specific steps and future considerations for data acquisition, processing, and analysis are described.
The number of total hip arthroplasty (THA) procedures performed on adults aged 45-64 years suffering from hip osteoarthritis (OA) more than doubled between 2000 and 20101. Based on the increases in THA procedures from 2000 to 2014, a recent study predicted that the overall number of yearly procedures may triple over the next twenty years2. These large increases in THA procedures are alarming considering that current treatment costs exceed $18 billion annually in the United States alone3.
Developmental dysplasia of the hip (DDH) and femoroacetabular impingement syndrome (FAIS), which describe an under- or over-constrained hip, respectively, are believed to be the primary etiology of hip OA4. The high prevalence of these structural hip deformities in individuals undergoing THA was initially described more than three decades ago5. Still, the relationship between abnormal hip anatomy and osteoarthritis is not well understood. One challenge to improving the working understanding of the role of deformities in the development of hip OA is that abnormal hip morphology is very common amongst asymptomatic adults. Notably, studies have observed morphology associated with cam-type FAIS in approximately 35% of the general population6, 83% of senior athletes7, and more than 95% of collegiate male athletes8. In another study of female collegiate athletes, 60% of participants had radiographic evidence of cam FAIS, and 30% had evidence of DDH9.
Studies demonstrating a high prevalence of deformities amongst individuals without hip pain point to the possibility that morphology commonly associated with FAIS and DDH may be a natural variant that only becomes symptomatic under certain conditions. However, the interaction between hip anatomy and hip biomechanics is not well understood. Notably, there are known difficulties with measuring hip joint motion using traditional optical motion capture technology. First, the joint is relatively deep within the body, such that the location of the hip joint center is difficult to both identify and track dynamically using optical skin marker motion capture, with errors on the same order of magnitude as the radius of the femoral head10,11. Second, the hip joint is surrounded by large soft tissue bulk, including subcutaneous fat and muscle, that moves relative to the underlying bone, resulting in soft tissue artifact12,13,14. Finally, using optical tracking of skin markers, kinematics are evaluated relative to generalized anatomy and thus do not provide insight into how subtle morphological differences might affect the biomechanics of the joint.
To address the lack of accurate kinematics in combination with subject-specific bone morphology, both single and dual fluoroscopy systems have been developed for analyzing other natural joint systems15,16,17. However, this technology has only recently been applied to the native hip joint, likely due to the difficulty in acquiring high-quality images through the soft tissue surrounding the hip. The methodology to accurately measure in vivo hip joint motion and display this motion relative to subject-specific bone anatomy is described herein. The resulting arthrokinematics provide an unparalleled ability to investigate the subtle interplay between bone morphology and biomechanics.
Herein, the procedures for acquiring and processing dual fluoroscopy images of the hip during activities of daily living have been described. Owing to the desire to capture whole-body kinematics with optical marker tracking simultaneously with dual fluoroscopy images, the data collection protocol requires coordination between several sources of data. Calibration of the dual fluoroscopy system utilizes plexiglass structures implanted with metallic beads that can be directly identified and tracked as markers. In contrast, dynamic bone motion is tracked using markerless tracking, which utilizes only the CT-based radiographic density of the bones to define orientation. Dynamic motion is then tracked simultaneously using dual fluoroscopy and motion capture data that are spatially and temporally synced.
The systems are synced spatially during calibration through concurrent imaging of a cube with both reflective markers and implanted metal beads and the generation of a common coordinate system. The systems are synced temporally for each activity or capture through the use of a split electronic trigger, which sends a signal to end the recording of the dual fluoroscopy cameras and interrupts a constant 5 V input to the motion capture system. This coordinated protocol enables the quantification of the position of body segments that fall outside the combined field of view of the dual fluoroscopy system, expression of kinematic results relative to gait-normalized events, and characterization of the soft tissue deformation around the femur and pelvis.
Procedures outlined in this protocol were approved by the University of Utah Institutional Review Board.
1. CT arthrogram imaging
2. Dual fluoroscopy imaging
3. Skin marker motion capture and instrumented treadmill
4. Image preprocessing
5. Bone motion tracking
6. Data analysis
Using dual fluoroscopy as a reference standard, the accuracy of skin-marker-based estimates of the hip joint center and the effect of soft-tissue artifact on kinematic and kinetic measurements were quantified22,23,24. The superior accuracy of dual fluoroscopy was then used to identify subtle differences in pelvic and hip joint kinematics between patients with FAIS and asymptomatic control participants25. Dual-fluoroscopy-based arthrokinematics were analyzed to quantify hip joint coverage, the relationship between morphology and kinematics, and bone-to-bone distances during dynamic motions26,27,28,29.
Before developing a protocol to investigate weight-bearing hip joint kinematics, the system was validated in cadaveric specimens with implanted metal beads during supine clinical exams to an accuracy within 0.5 mm and 0.6°30. Once validated, kinematics during clinical exams were measured using dual fluoroscopy in patients with FAIS and asymptomatic control participants. The results demonstrated that patients had altered motion in both internal rotation and adduction31.
Using weight-bearing dual fluoroscopy as a reference standard, the error in identifying the location of the hip joint center as well as the errors caused by soft-tissue artifact were then directly analyzed. Functional methods of identifying the hip joint center, i.e., the star-arc motion, were identified to be more accurate than predictive, landmark-based methods with errors of 11.0 and 18.1 mm, respectively32. Dynamic errors in the hip joint center were similar to those from standing; however, an additional 2.2 mm of spurious hip joint center movement was attributed to soft tissue artifact, with errors of more than 5 cm during dynamic movement for the greater trochanter marker23.
In addition to the errors in identification of the hip joint center, joint angles were underestimated by greater than 20° in internal-external rotation pivots23. While the underestimation of kinematics is cause for concern in itself, these errors reduced the measured range of motion and calculated kinetic variables during even a low range of motion activities, such as gait24. However, accurate dual fluoroscopy kinematic data can be difficult to incorporate into musculoskeletal models. Specifically, model marker errors were approximately 1 cm when running inverse kinematics with dual fluoroscopy-based landmark locations. While this error is relatively small compared to the 5 cm errors due to soft tissue artifact found for skin marker motion capture data, such error is an order of magnitude larger than that of bone positions measured by dual fluoroscopy.
In addition to the quantification of errors in traditional skin marker motion capture, the accuracy and methodology behind dual fluoroscopy provide the capability to evaluate even subtle differences in kinematics between cohorts, which may otherwise be hidden by the errors of the measurement technique. While differences in hip joint kinematics were not observed between patients with cam FAIS and asymptomatic control participants, differences in pelvic kinematics that would have been difficult to detect in the presence of soft-tissue artifact were identified25. This assessment required direct comparison between cohorts. Moreover, the potential relationship between kinematic variation and bone morphology, such as femoral anteversion, was also investigated27. These findings indicated the need for consideration of both morphology and biomechanics in the diagnosis of hip pathologies and the planning of conservative or surgical treatments.
A major hurdle in the use of biomechanical data in a clinical care setting is the difference in coordinate systems used by biomechanists and clinicians. In a biomechanics lab, the landmarks used to define coordinate systems of the femur and pelvis are driven by the ability to identify and track the landmarks from the skin surface during dynamic motion. In contrast, surgical coordinate systems are defined using bony landmarks identifiable during surgery with a patient supine or prone. The direct tracking of the femur and pelvis in dual fluoroscopy allowed for the evaluation of the influence of various coordinate system definitions on kinematic output29. The differences between coordinate system definitions resulted in kinematic offsets greater than 5°. However, these offsets were relatively consistent during motion and could be accounted for through bony landmark identification.
The combination of subject-specific bone morphology and kinematics — arthrokinematics — provides a joint-level assessment of form and function. For patients with DDH, femoral under-coverage is thought to be the cause of degeneration, and therefore, measurements of coverage are used heavily in diagnosis and surgical planning. Unfortunately, these measurements are often limited to static images, obtained with an individual supine, and only in two dimensions. Dual fluoroscopy-derived arthrokinematics were used to directly measure the variability in femoral coverage during dynamic activities26. Importantly, strong correlations between coverage in standing and coverage during gait when evaluated in entirety were found. Yet, regionalized coverage varied for both anterior and posterior regions of the femoral head even during the stance phase of gait.
Extra-articular impingement is a cause of pain at the hip and surrounding region and describes abnormal contact between the femur and regions of the pelvis outside the acetabulum, including the ischium and anterior inferior iliac spine. The dynamic nature of ischiofemoral impingement was evaluated through the comparison of clinical MRI-based measurements of ischiofemoral space and those during dynamic activities28. Therein, decreased space was observed dynamically in comparison to the standard clinical measures; sex-based differences, which could not be attributed to kinematic differences, were also identified. These methods could also be applied to evaluate joint space dynamically, providing insight into the variability of the position of the femoral head within the acetabulum and the variability across patient cohorts (Figure 8).
Figure 1: Overhead view of the dual fluoroscopy system positioned over the instrumented treadmill for a left hip. The system is positioned to minimize the effect of scatter and maximize the field of view. The image intensifiers are positioned approximately 100-110 cm from the source of the emitter and angled 50° from one another. Please click here to view a larger version of this figure.
Figure 2: View from the contralateral (right) side of a participant during dynamic activities. The participant is positioned between the two image intensifiers (II) such that the field of view of the dual fluoroscopy system is centered over the left hip joint. Level and incline walking, internal and external rotation pivots, and range of motion activities are performed on a treadmill platform. Abbreviation: FHJC = functional hip joint center. Please click here to view a larger version of this figure.
Figure 3: Overhead view of the motion capture system relative to the dual fluoroscopy system. The optical motion capture system includes 10 infrared cameras and a single video-based camera and is positioned on a frame hanging from the ceiling. Please click here to view a larger version of this figure.
Figure 4: Anterior and posterior view of the marker set used for skin marker motion capture. There are five plates with four markers each, which are positioned on the back, thighs, and shanks of the participants; all other markers are applied directly to the skin. Calibration markers are removed for dynamic motion capture. Marker labels prefaced with an R or L indicate markers on the right or left side of the body; marker labels suffixed with S, L, R, I, A, or P indicate marker locations on a marker plate, specifically superior, left, right, inferior, anterior, or posterior, respectively. Abbreviations: *SHO = shoulder; CLAV = center of clavicles; STRN = bottom of sternum; BACK_* = markers of plate placed on the lower back; *ILC = iliac crest; *ASI = anterior superior iliac spine; *PSI = posterior superior iliac spine; GRT_TRO = greater trochanter; *THI_* = markers of the respective plates placed on the thigh; *KNE_M = medial femoral condyle (knee); *KNE_L = lateral femoral condyle (knee); *TIB_* = markers of the respective plates placed on the shank (tibia); *ANK_M = medial malleolus (ankle); *ANK_L = lateral malleolus; *5TH = fifth metatarsophalangeal joint; *TOE = first metatarsophalangeal joint; *HEE = calcaneus (heel). Please click here to view a larger version of this figure.
Figure 5: Landmarks and coordinate systems of the femur and pelvis. Landmarks of bilateral anterior superior iliac spine (ASIS; magenta) and posterior superior iliac spine (PSIS; cyan) and their mid-points are used to define the coordinate system of the pelvis. The center of the femoral head (orange) and bilateral femoral condyles (green), their mid-point, and a cylinder fit of the condyles are used to define the coordinate system of the femur (shown for left femur). The third axis of each bone is determined from the cross-product of the two displayed axes. Please click here to view a larger version of this figure.
Figure 6: Dual fluoroscopy images and associated markerless tracking of a left hip. Images are shown for maximum rotation of the external and internal rotation pivots (center), with the image from the anterior fluoroscope (left) and the posterior fluoroscope (right). Markerless tracking solutions for the pelvis (top) and femur (bottom) for each dual fluoroscopy image. Please click here to view a larger version of this figure.
Figure 7: Dual fluoroscopy measured kinematics. Kinematics for 100 frames surrounding the maximum rotation (vertical dotted line) of external and internal rotation pivots for a representative participant. Please click here to view a larger version of this figure.
Figure 8: Arthrokinematics-based surface distance between a left hemi-pelvis and femur. Arthrokinematics are shown for maximum rotation of the external and internal rotation pivot (center) with respective bone models measured with dual fluoroscopy (outer). Please click here to view a larger version of this figure.
Dual fluoroscopy is a powerful tool for the investigation of in vivo kinematics, especially for the hip, which is difficult to accurately measure using traditional optical motion capture. However, fluoroscopy equipment is specialized, wherein a unique system setup may be required when imaging other joints of the human body. For example, several modifications were made to the mounting of the image intensifiers, positioning of the system, and settings of the beam energy in the application of dual fluoroscopy to the study of ankle kinematics32,33,34,35. In addition to requiring considerable study preparation, dual fluoroscopy requires the acquisition of additional data, including 3D medical imaging and potentially traditional skin marker motion capture to track whole-body kinematics, as well as lengthy post-processing, including CT image segmentation and markerless tracking of the acquired images. Fortunately, fully processed data from dual fluoroscopy can be used in various applications with capabilities reaching far beyond those available with traditional motion capture.
Optical motion capture utilizes the motion of markers on the skin to estimate body segment positions, while radiation-based dual fluoroscopy allows for direct measurement of only the bone positions. While significant effort has been dedicated to quantifying soft tissue dynamics relative to bone motion36, 37, it is inherently difficult to measure the motion patterns of the large mass of soft tissue between the outer layer of skin and the bones. However, for thinner tissues in direct contact with the bones, such as the cartilage and labrum of the hip, the combination of dual fluoroscopy and CT arthrogram imaging provides the ability to dynamically evaluate their spatial relationship. The data collected during supine clinical exams were used to show that the location of clinically observed damage to the acetabular labrum aligned with the position of contact between the femur and labrum during supine impingement exams38. Importantly, this analysis identified that the region of initial and greatest contact between the femur and labrum did not align with the location of the smallest distance between the bones.
Individuals with hip pathoanatomy are at risk of damage to the cartilage and labrum. However, the mechanisms responsible for chondrolabral injuries are not well understood. Conceivably, arthrokinematics data built from CT arthrogram data could be analyzed to study the mechanics of the cartilage and labrum. For example, the observed penetration between surface reconstructions representing soft tissue (e.g., labrum, cartilage) and bone could be analyzed and interpreted to approximate the strain experienced by these tissues. However, even slight errors in the tracking of kinematics or reconstruction of surfaces could result in drastic differences in estimated strains and joint loads. Thus, more advanced modeling methods, such as the FE method, may be required to comprehensively evaluate chondrolabral mechanics in the hip. Data from dual fluoroscopy, traditional skin marker motion capture of whole-body kinematics, and the instrumented treadmill can serve as input for models that estimate muscle forces and joint reaction loads and torques. These kinetic data can then serve as loading conditions to FE models that estimate chondrolabral stresses and strains.
In addition to the specific steps involved in the protocol, the scheduling of different aspects of the study is also relevant to successful data acquisition. First, in studies using arthrogram imaging, which is inherently invasive due to the injection of contrast into the hip capsule, the arthrogram must be performed either several days before or any time after the completion of motion capture experiments to avoid any effect on patient motion patterns. Second, all calibration must be performed prior to, but just before, the arrival of the participant to ensure that the system configuration is not altered between calibration and image acquisition. Third, the participant should be instructed to perform dynamic trials in a random order to eliminate any effect of ordering on the performance of tasks.
Another major consideration for the use of dual fluoroscopy for the measurement of hip kinematics is radiation exposure. It is important to note, however, that 80% of the estimated dose equivalent of radiation in the described protocol is from the CT scan. One solution to reduce exposure is the substitution of magnetic resonance imaging (MRI) for CT imaging. While MRI can be used for surface reconstruction, the tracking of dual fluoroscopy images also relies on the projection of bone densities from the digitally reconstructed radiographs. Although MRI cannot directly measure bone density, specific sequences, such as the dual echo steady state (DESS), provide some differentiation between the denser cortical bone and the less dense cancellous bone. These images can be transformed to have a similar appearance to CT images and could potentially reduce the radiation exposure of participants in dual fluoroscopy studies.
Owing to the large amount of soft tissue surrounding the hip joint, the specific positioning of the dual fluoroscopy system must be optimized to reduce X-ray scatter. The position of the participant relative to the X-ray emitters and the angle between the image intensifiers were found to be important factors. This protocol indicates the positioning of the dual fluoroscopy system used to study hip motion in participants during weight-bearing activities. It is, however, also relevant to note that the participant cohort was limited to individuals with a BMI less than 30 kg/m2. A similar BMI limit is recommended when capturing dual fluoroscopy images of joints surrounded by large masses of soft tissue.
The protocol described herein can be applied to various dual fluoroscopy system configurations and joints, including supine and weight-bearing hip kinematics, both treadmill and overground weight-bearing ankle kinematics, and sitting shoulder kinematics16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35. Owing to the minimal global motion of the hip joint during treadmill gait, an instrumented treadmill was used for the assessment of weight-bearing kinematics of the hip joint. Without a treadmill or a moving fluoroscope system, it would only be possible to capture the hip joint during activities performed in a confined field of view. However, the use of a treadmill is not appropriate for all joints. As an example, application of this protocol to the investigation of ankle kinematics during treadmill walking captured only a small portion of gait due to the inherent motion of the treadmill32,35, while overground gait was able to capture a larger portion of gait, spanning from prior to heel-strike to after toe-off33,40,41.
The authors have nothing to disclose.
This research was supported by the National Institutes of Health (NIH) under grant numbers S10 RR026565, R21 AR063844, F32 AR067075, R01 R077636, R56 AR074416, R01 GM083925. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Amira Software | ThermoFisher Scientific | Version 6.0 | |
Calibration Cube | Custom | 36 steel beads (3 mm diameter, spacing 6.35 cm, uncertainty 0.0036 mm) | |
Calibration Wand | Vicon | Active Wand | |
CT Scanner | Siemens AG | SOMATOM Definition 128 CT | |
Distortion Correction Grid | Custom | Acrylic plate with a grid of steel beads spaced 10 mm and 31 beads across the diameter (2 mm diameter) | |
Dynamic Calibration Plate | Custom | Acrylic plate with 3 steel beads spaced 30 mm (2 mm diameter, uncertainty 0.0013 mm) | |
Emitter (2) | Varian Interay; remanufactured by Radiological Imaging Services | Housing B-100/Tube A-142 | |
Epinephrine | Hospira | Injection, USP 10 mg/mL | |
FEBioStudio Software | FEBio.org | Version 1.3 | Mesh processing and kinematic visualization |
Graphical Processing Unit | Nvidia | Tesla | |
Hare Traction Splint | DynaMed | Trac-III, Model No. 95201 | |
High-speed Camera (2) | Vision Research, Inc. | Phantom Micro 3 | |
Image Intensifier (2) | Dunlee, Inc.; remanufactured by Radiological Imaging Services | T12964P/S | |
Iohexol injection | GE Healthcare | Omnipaque 240 mgI/mL | 517.7 mg iohexol, 1.21 mg tromethamine, 0.1 mg edetate calcium disodium per mL |
ImageJ | National Institutes of Health and Laboratory for Optical and Computational Instrumentation | ||
Lidocaine HCl | Hospira | Injection, USP 10 mg/mL | |
Laser and Mirror Alignment System | Custom | Three lasers adhered to acrylic plate that attaches to emitter, mirror attaches to face of image intensifier | |
Markless Tracking Workbench | Henry Ford Hospital, Custom Software | Custom | |
MATLAB Software | Mathworks, Inc. | Version R2017b | |
Motion Capture Camera (10) | Vicon | Vantage | |
Nexus Software | Vicon | Version 2.8 | Motion capture |
Phantom Camera Control (PCC) Software | Vision Research, Inc. | Version 1.3 | |
Pre-tape Spray Glue | Mueller Sport Care | Tuffner | |
Retroreflective Spherical Skin Markers | 14 mm | ||
Split Belt Fully Instrumented Treadmill | Bertec Corporation | Custom | |
Visual3D Software | C-Motion Inc. | Version 6.01 | Kinematic processing |