INTRODUCTION

The restoration of physiological kinematics is one of the goals of a total knee arthroplasty (TKA). Navigation systems have been developed to allow an accurate and precise placement of the implants. But its application to the intraoperative measurement of knee kinematics has not been validated. The hypothesis of this study was that the measurement of the knee axis, femoral rotation, femoral translation with respect to the tibia, and medial and lateral femorotibial gaps during continuous passive knee flexion by the navigation system would be different from that by fluoroscopy taken as reference.

Disorders of human joints manifest during dynamic movement, yet no objective tools are widely available for clinicians to assess or diagnose abnormal joint motion during functional activity. Machine learning tools have supported advances in many applications for image interpretation and understanding and have the potential to enable clinically and economically practical methods for objective assessment of human joint mechanics. We performed a study using convolutional neural networks to autonomously segment radiographic images of knee replacements and to determine the potential for autonomous measurement of knee kinematics. The autonomously segmented images provided superior kinematic measurements for both femur and tibia implant components. We believe this is an encouraging first step towards realization of a completely autonomous capability to accurately quantify dynamic joint motion using a clinically and economically practical methodology.

Background

Scapular notching is a complication after reverse shoulder arthroplasty with a high incidence up to 100%. Its clinical relevance remains uncertain; however, some studies have reported that scapular notching is associated with an inferior clinical outcome. There have been no published articles that studied positional relationship between the scapular neck and polyethylene insert in vivo. The purpose of this study was to measure the distance between the scapular neck and polyethylene insert in shoulders with Grammont type reverse shoulder arthroplasty during active external rotation at the side.

Musculoskeletal modeling techniques simulate reverse total shoulder arthroplasty (RTSA) shoulders and how implant placement affects muscle moment arms. Yet, studies have not taken into account how muscle-length changes affect force-generating capacity postoperatively. We develop a patient-specific model for RTSA patients to predict muscle activation.

Patient-specific muscle parameters were estimated using an optimization scheme calibrating the model to isometric arm abduction data at 0°, 45°, and 90°. We compared predicted muscle activation to experimental electromyography recordings. A twelve-degree of freedom model with experimental measurements created patient-specific data estimating muscle parameters corresponding to strength. Optimization minimized the difference between measured and estimated joint moments and muscle activations, yielding parameters corresponding to subjects' strength that can predict muscle activation and lengths.

Model calibration was performed on RTSA patients' arm abduction data. Predicted muscle activation ranged between 3% and 70% of maximum. The maximum joint moment produced was 10 Nm. The model replicated measured moments accurately (R² > 0.99). The optimized muscle parameters produced feasible muscle moments and activations for dynamic arm abduction when using data from isometric force trials. A normalized correlation was found between predicted and experimental muscle activation for dynamic abduction (r > 0.9); the moment generation to lift the arm was tracked (R² = 0.99).

Statement of Clinical Significance: We developed a framework to predict patient-specific muscle parameters. Combined with patient-specific models incorporating joint configurations, kinematics, and bone anatomy, they can predict muscle activation in novel tasks and, e.g., predict how RTSA implant and surgical decisions may affect muscle function.

Reverse Total shoulder arthroplasty (RTSA) has become an increasingly used solution to treat osteoarthritis and cuff tear arthropathy. Though successful there are still 10 to 65% complication rates reported for RTSA. Complication rates range over different reverse shoulder designs but a clear understanding of implant design parameters that cause complications is still lacking within the literature. In efforts to reduce complication rates (Implant fixation, range of motion, joint stiffness, and fracture) and improve clinical/functional outcomes having to do with proper muscle performance we have employed a computational approach to assess the sensitivity of muscle performance to changes in RTSA implant geometry and surgical placement. The goal of this study was to assess how changes in RTSA joint configuration affect deltoid performance.

An approach was developed from previous work to predict a patient's muscle performance. This approach was automated to assess changes in muscle performance over 1521 joint configurations for an RTSA subject. Patient-specific muscle moment arms, muscle lengths, muscle velocities, and muscle parameters served as inputs into the muscle prediction scheme. We systematically varied joint center locations over 1521 different perturbations from the in vivo measured surgical placement to determine muscle normalized operating region for the anterior, lateral and posterior aspects of the deltoid muscle. The joint center was varied according to previous published work from the RTSA subject's nominal surgical position ±4 mm in the anterior/posterior direction, ±12mm in the medial/lateral direction, and −10 mm to 14 mm in the superior/inferior direction (Walker 2015 et al. Table 2).

Overall muscle normalized operating length varied over 1521 different implant configurations for the RTSA subject. Ideal muscle normalized operating length variations were found to be in all the fundamental directions that the joint was varied. The anterior deltoid normalized operating length was found to be most sensitive with joint configurations changes in the anterior/posterior medial/lateral direction. It lateral deltoid normalized operating length was found to be most sensitive with joint configurations changes in the medial/lateral direction. It posterior deltoid normalized operating length was found to be most sensitive with joint configurations changes in the medial/lateral direction. Reserve actuation for all samples remained below 1 Nm. The most optimal deltoid normalized operating length was implemented by changing the joint configuration in the superior/inferior and medial/lateral directions.

Current shoulder models focus on predicting muscle moment arms. Although valuable it does not allow me for active understanding of how lengthening the muscle will affect its ability to generate force. Our study provides an understanding of how muscle lengthening will affect the force generating capacity of each of the heads of the deltoid. With this information improvements can be made to the surgical placement and design of RTSA to improve functional/clinical outcomes while minimizing complications.

For any figures or tables, please contact the authors directly.

The design of every post-surgical knee arthroplasty study begins with the question “How soon after surgery should we assess the patients?”. The consensus, based primarily upon clinical rating systems, is that patients' scores reach a plateau roughly one year after surgery, and that observations performed at that time should be indicative of the long-term behavior of the joint. This is satisfactory for long-term studies of clinical performance. However, when new devices are introduced there is a need to determine as quickly as possible if the device performs as designed. Waiting a year or more after surgery to characterize a device's performance may place additional patients at risk of receiving an inferior design, or may delay widespread availability of a superior design. The goal of this study was to assess knee arthroplasty patients at 6–12 weeks, 6 months and 1 year after surgery to determine if their tibiofemoral kinematics changed during functional activities.

A total of 13 patients (7 female) were recruited from an ongoing clinical study to participate in this IRB-approved sub-study. All subjects received fixed-bearing, cemented, posterior-cruciate-retaining total knee arthroplasty of the same design from a single surgeon. Subjects averaged 69 years, 169cm tall, and 28 BMI. Subjects were studied at 6–12 weeks, at 6 months and at 12 months post-surgery, when they showed an average clinical flexion of 106°, 113° and 115°, respectively. Subjects' knees were observed using pulsed-flat-panel-fluoroscopy during three activities: lunging to maximum flexion with their foot placed on a 20cm step, kneeling to maximum flexion on a padded bench, and step-up/down on a 20cm step without progression of the contralateral limb. Model-image registration was used to register 3D geometric models of the implants with their radiographic projections based upon measured projection parameters. 3D knee kinematics were derived from the registered models, including joint angles and the antero-posterior translation of the medial and lateral condyles relative to the tibial baseplate.

There were no statistically significant changes in knee kinematics between the 6–12 week and 6 month, and 6-month and 12-month visits during the kneel and lunge activities (Table 1). Similarly, there were no pair-wise differences in tibial rotation or condylar translation during the dynamic step activity at any flexion angle (Figure 1).

Traditional thinking suggests studies of knee mechanics should be performed at least one year after surgery to make observations that are predictive of long-term joint function. In three different functional activities, we could not demonstrate significant changes in knee kinematics between 6–12 weeks and 6 months, nor between 6 months and 12 months. If these results can be confirmed in a larger subject cohort, and for a range of TKA designs, then functional follow-up studies of novel knee arthroplasty designs might be justified as early as 6–12 weeks after surgery, making it possible to accelerate confirmation devices are performing in patients as designed.

For any figures or tables, please contact the authors directly.

Introduction

Current modeling techniques have been used to model the Reverse Total Shoulder Arthroplasty (RTSA) to account for the geometric changes implemented after RTSA [2,3]. Though these models have provided insight into the effects of geometric changes from RTSA these is still a limitation of understanding muscle function after RTSA on a patient-specific basis. The goal of this study sought to overcome this limitation by developing an approach to calibrate patient-specific muscle strength for an RTSA subject.

Reverse Total shoulder arthroplasty (RTSA) has become an increasingly used solution to treat osteoarthritis and cuff tear arthropathy. Though successful there are still 10 to 65% complication rates reported for RTSA. Complication rates range over different reverse shoulder designs but a clear understanding of implant design parameters that cause complications is still lacking within the literature. In efforts to reduce complication rates (Implant fixation, range of motion, joint stiffness, and fracture) and improve clinical/functional outcomes having to do with proper muscle performance we have employed a computational approach to assess the sensitivity of muscle performance to changes in RTSA implant geometry and surgical placement. The goal of this study was to assess how changes in RTSA joint configuration affect deltoid performance.

An approach was developed from previous work to predict a patient's muscle performance. This approach was automated to assess changes in muscle performance over 1521 joint configurations for an RTSA subject. Patient-specific muscle moment arms, muscle lengths, muscle velocities, and muscle parameters served as inputs into the muscle prediction scheme. We systematically varied joint center locations over 1521 different perturbations from the in vivo measured surgical placement to determine muscle activation and normalized operating region for the anterior, lateral and posterior aspects of the deltoid muscle. The joint center was varied from the RTSA subject's nominal surgical position ±4 mm in the anterior/posterior direction, ±12mm in the medial/lateral direction, and −10 mm to 14 mm in the superior/inferior direction.

Overall muscle activity varied over 1521 different implant configurations for the RTSA subject. For initial elevation the RTSA subject showed at least 25% deltoid activation sensitivity in each of the directions of joint configuration change(Figure 1). Posterior deltoid showed a maximal activation variation of 84% in the superior/inferior direction(Figure 1c). Deltoid activation variations lie primarily in the superior/inferior and anterior/posterior directions. An increasing trend was seen for the anterior, lateral and posterior deltoid outside of the discontinuity seen at 28°(Figure 1). Activation variations were compared to subject's experimental data. Reserve actuation for all samples remained below 4Nm(Figure 2). The most optimal deltoid normalized operating length was implemented by changing the joint configuration in the superior/inferior and medial/lateral directions(Figure 3).

Current shoulder models utilize cadaver information in their assessment of generic muscle strength. In adding to this literature we performed a sensitivity study to assess the effects of RTSA joint configurations on deltoid muscle performance in a single patient-specific model. For this patient we were able to assess the best joint configuration to improve the patients muscle function and ideally their clinical outcome. With this information improvements can be made to the surgical placement and design of RTSA on a patient-specific basis to improve functional/clinical outcomes while minimizing complications.

For any figures or tables, please contact authors directly (see Info & Metrics tab above).

Total knee arthroplasty (TKA) is an exceptionally successful and robust treatment for disabling knee disease, but many efforts continue to improve patient postoperative satisfaction and performance. One approach to improving performance is to restore TKA motions closer to those in healthy knees. Based upon an idealized model of knee motions, it is possible to design tibiofemoral articulating surfaces to promote natural kinematics and force transfer (Fiedler et al., Acta Bioeng Biomech, 2011). Such an asymmetric design is expected to promote rollback in stance phase that continues through deeply flexed activities. The purpose of this study is twofold: (1) To determine if a TKA designed on a theoretical basis achieves the proposed motions in vivo, and (2) To track postoperative kinematic patterns with examinations at 6–12 weeks, 6 months and one year postoperatively. This paper reports results of the initial cohort that has completed 6–12 week and 6-month examinations.

Eight patients, including 3 females, with unilateral TKA for varus osteoarthritis provided written informed consent prior to beginning the study. Patients averaged 66±9 years, 168±14cm, and 28±3 BMI. Patients performed three weightbearing activities observed using pulsed x-ray flat-panel imaging at 30Hz: stepping up from flexion to extension on a 20cm step, lunging to maximum flexion with the foot placed on a 20 cm step, and kneeling to maximum flexion with the shin placed on a padded support. Three-dimensional knee kinematics were quantified using model-image registration to determine flexion, tibial internal rotation, anteroposterior movement of the femoral condyles (relative to the tibial AP center) and average center of rotation (CoR) in the transverse plane.

During the maximum-flexion lunge and kneeling activities subjects exhibited average knee flexion of 104°–110° and tibial internal rotation of 2°–6° (Table 1). At 6–12 weeks, the medial/lateral condyles were at −3mm/−8mm and −1mm/−6mm during maximum flexion lunge and kneeling, respectively. During the stair activity from 0° to 70° flexion, there were small tibial internal rotations (1°/5°) and anterior medial (2mm/5mm) and lateral (3mm/3mm) condylar translations at both time points (Figure 1). The average CoRs for the stair activity were medial +18% and +5% for the 6–12 week and 6-month exams, respectively.

It has long been assumed knee kinematics change during a patient's first one or two postoperative years. In our early post-op cohort, changes in weight-bearing kinematics over the first 6 postoperative months are small. In maximal flexion activities, patients exhibited flexion similar to similar cohorts studied at least one year post-op (Clin Orthop, 410:131–138, 2003). Similarly, kinematics during the weight-bearing step activity were similar in pattern and magnitude to those previously reported for posterior cruciate-retaining (CR) TKA at least one year post-op (Clin Orthop, 426:187–193, 2004). The average CoRs were medial for the stair activity, which is normal for healthy knees but uncommon for CR TKA. Early post-op results with an asymmetric CR TKA implant intended to promote physiologic motion show flexion and stair kinematics similar to many successful CR designs at longer follow-up. The medial CoR indicates closer-to-physiologic motion than commonly is observed in CR TKA.

Introduction

Reverse total shoulder arthroplasty (RTSA) is a commonly performed operation for a variety of pathologies. Despite excellent short-term outcomes, complications are commonly encountered. Recurrent instability occurs in up to 31% of cases, often due to components placed with too little tension. Acromial stress fractures can occur in up to 7% of cases, often due to components placed in too much tension. Despite these concerns, there is little evidence evaluating the intraoperative tension and glenohumeral contact forces (GHCF) during RTSA. The purpose of this study was to measure the intraoperative GHCF during RTSA.

Intraoperative planning of knee replacement components, targeting a desired functional outcome, requires a calibrated patient-specific model of the patient's soft-tissue anatomy and mechanics. Previously, a surgical technique was demonstrated for measuring knee joint kinematics and kinetics consistent with modern navigation systems in conjunction with the development of a patient-customizable knee model. A data efficient approach for the model calibration task was achieved utilizing the sensitivity of the model to simulated clinical hand manipulations of the knee joint requiring 85% less computations.

For this numerical investigation a simplified knee joint model, based on the OpenKnee repository, consisting of bone (rigid), cruciate ligaments (single-bundle, nonlinear spring), collateral ligaments (multiple nonlinear springs), articular cartilage (rigid, pressure-over-closure relationship), and combined capsule/meniscus (linear springs) was created using a custom Matlab (MathWorks)-Abaqus (Dassault Systèmes) implicit finite element modeling framework (Figure 1). A sensitivity analysis was performed by applying constant loading along the anterior-posterior, medial-lateral, varus-valgus, and internal-external directions (30 N for forces and 3 Nm for moments) while perturbing each customizable parameter positively and negatively by 1 mm at 0, 25, 50, 75 and 100 degrees of flexion. A constant load of 150 N was maintained in compression. The change in static endpoint position was measured relative to the respective position without perturbation. Sensitivity results were then arranged by load direction and principal component analysis was subsequently performed (Table 1).

First a single optimization task was simulated including all model parameters and all loading sequences with the goal of minimizing the kinematic differences between the reference model and a perturbed model (Figure 2). Second, a piecewise optimization task was designed using only the sensitive parameters for a spanning set of loads for the same perturbed model. Parameters 3 and 4 were tuned using internal and external endpoints. Then parameters 1 and 5 were tuned using the anterior endpoints. Similarly, parameters 2 and 7 were tuned using the posterior endpoints. Finally, parameter 8 was tuned using the varus endpoints. All loadings were observed to be insensitive to parameter 6 (ACL-Y). The number of model evaluations required were 2520 and 390 for the single and piecewise optimizations, respectively. The single simulation task recovered all parameters within 0.57 mm on average compared to 0.64 mm on average for the piecewise task. Kinematic errors due to the calibration technique were within 0.001 mm and 0.18 deg compared to 0.001 mm and 0.04 deg.

Computational cost for the optimization task required to calibrate a patient-specific knee model was reduced while maintaining clinically relevant accuracy. This model reduction approach will further enable the rapid adoption of the technology for intraoperative planning of knee replacement components based on targeted functional outcomes.

Modern musculoskeletal modeling techniques have been used to simulate shoulders with reverse total shoulder arthroplasty and study how geometric changes resulting from implant placement affect shoulder muscle moment arms. These studies do not, however, take into account how changes in muscle length will affect the force generating capacity of muscles in their post-operative state. The goal of this study was to develop and calibrate a patient-specific shoulder model for subjects with RTSA in order to predict muscle activation during dynamic activities.

Patient-specific muscle parameters were estimated using a nested optimization scheme calibrating the model to isometric arm abduction data at 0°, 45° and 90°. The model was validated by comparing predicted muscle activation for dynamic abduction to experimental electromyography recordings. A twelve-degree of freedom model was used with experimental measurements to create a set of patient-specific data (three-dimensional kinematics, muscle activations, muscle moment arms, joint moments, muscle lengths, muscle velocities, tendon slack lengths, optimal fiber lengths and peak isometric forces) estimating muscle parameters corresponding to each patient's measured strength. The optimization varied muscle parameters to minimize the difference between measured and estimated joint moments and muscle activations for isometric abduction trials. This optimization yields a set of patient-specific muscle parameters corresponding to the subject's own muscle strength that can be used to predict muscle activation and muscle lengths for a range of dynamic activities.

The model calibration/optimization procedure was performed on arm abduction data for a subject with reverse total shoulder arthroplasty. Muscle activation predicted by the model ranged between 3% and 90% of maximum. The maximum joint moment produced was 20 Nm. The model replicated measured joint moments accurately (R2 > 0.99). The optimized muscle parameter set produced feasible muscle moments and muscle activations for dynamic arm abduction, when calibrated using data from isometric force trials.

Current modeling techniques for the upper extremity focus primarily on geometric changes and their effects on shoulder muscle moment arms. In an effort to create patient-specific models, we have developed a framework to predict subject-specific muscle parameters. These estimated muscle parameters, in combination with patient-specific models that incorporate the patient's joint configurations, kinematics and bone anatomy, provide a framework to predict dynamic muscle activation in novel tasks and, for example, predict how joint center changes with reverse total shoulder arthroplasty may affect muscle function.

Reverse total shoulder arthroplasty (RTSA) is an increasingly common treatment for osteoarthritic shoulders with irreparable rotator cuff tears. Although very successful in alleviating pain and restoring some function, there is little objective information relating geometric changes imposed by the reverse shoulder and arm function, particularly the moment generating capacity of the shoulder muscles. Recent modeling studies of reverse shoulders have shown significant variation in deltoid muscle moment arms over a typical range of humeral offset locations in shoulders with RTSA. The goal of this study was to investigate the sensitivity of muscle moment arms as a function of varying the joint center and humeral offset in three representative RTSA subjects that spanned the anatomical range from our previous study cohort. We hypothesized there may exist a more beneficial joint implant placement, measured by muscle moment arms, compared to the actual surgical implant configuration.

A 12 degree of freedom, subject-specific model was used to represent the shoulders of three patients with RTSA for whom fluoroscopic measurements of scapular and humeral kinematics during abduction had been obtained. The computer model used subject-specific in vivo abduction kinematics and systematically varied humeral offset locations over 1521 different perturbations from the surgical placement to determine moment arms for the anterior, lateral and posterior aspects of the deltoid muscle. The humeral offset was varied from its surgical position ±4 mm in the anterior/posterior direction, ±12mm in the medial/lateral direction, and −10 mm to 14 mm in the superior/inferior direction.

The anterior deltoid moment arm varied up to 20 mm with humeral offset and center of rotation variations, primarily in the medial/lateral and superior/inferior directions. Similarly, the lateral deltoid moment arm demonstrated variations up to 20 mm, primarily with humeral offset changes in the medial/lateral and anterior/posterior directions. The posterior deltoid moment arm varied up to 15mm, primarily in early abduction, and was most sensitive to changes of the humeral offset in the superior/inferior direction.

The goal of this study was to assess the sensitivity of the deltoid muscle moment arms as a function of joint configuration for existing RTSA subjects. High variations were found for all three deltoid components. Variation over the entire abduction arc was greatest in the anterior and lateral deltoid, while the posterior deltoid moment arm was mostly sensitive to humeral offset changes early in the abduction arc. Moment arm changes of 15–20 mm represent a significant amount of the total deltoid moment arm. This means there is an opportunity to dramatically change the deltoid moment arms through surgical placement of the joint center of rotation and humeral stem. Computational models of the shoulder may help surgeons optimize subject-specific placement of RTSA implants to provide the best possible muscle function, and assist implant designers to configure devices for the best overall performance.

Reverse Total shoulder arthroplasty (RTSA) has become an increasingly used solution to treat osteoarthritis and cuff tear arthropathy. Though successful there are still 10 to 65% complication rates reported for RTSA. Complication rates range over different reverse shoulder designs but a clear understanding of implant design parameters that cause complications is still lacking within the literature. In efforts to reduce complication rates (Implant fixation, range of motion, joint stiffness, and fracture) and improve clinical/functional outcomes having to do with proper muscle performance we have employed a computational approach to assess the sensitivity of muscle performance to changes in RTSA implant geometry and surgical placement. The goal of this study was to assess how changes in RTSA joint configuration affect deltoid performance.

An approach was developed from previous work to predict a patient's muscle performance. This approach was automated to assess changes in muscle performance over 1521 joint configurations for an RTSA subject. Patient-specific muscle moment arms, muscle lengths, muscle velocities, and muscle parameters served as inputs into the muscle prediction scheme. We systematically varied joint center locations over 1521 different perturbations from the in vivo measured surgical placement to determine muscle activation and normalized operating region for the anterior, lateral and posterior aspects of the deltoid muscle. The joint center was varied from the RTSA subject's nominal surgical position ±4 mm in the anterior/posterior direction, ±12mm in the medial/lateral direction, and −10 mm to 14 mm in the superior/inferior direction.

Overall muscle activity varied over 1521 different implant configurations for the RTSA subject. For initial elevation the RTSA subject showed at least 25% deltoid activation sensitivity in each of the directions of joint configuration change(Figure 1A–C). Posterior deltoid showed a maximal activation variation of 84% in the superior/inferior direction(Figure 1C). Deltoid activation variations lie primarily in the superior/inferior and anterior/posterior directions(Figure 1). An increasing trend was seen for the anterior, lateral and posterior deltoid outside of the discontinuity seen at 28°(Figur 1A–C). Activation variations were compared to subject's experimental data (Figure 1). Reserve actuation for all samples remained below 4Nm. The most optimal deltoid normalized operating length was implemented by changing the joint configuration in the superior/inferior and medial/lateral directions.

Current shoulder models utilize cadaver information in their assessment of generic muscle strength. In adding to this literature we performed a sensitivity study to assess the effects of RTSA joint configurations on deltoid muscle performance. With this information improvements can be made to the surgical placement and design of RTSA to improve functional/clinical outcomes while minimizing complications.

“How does the knee move?” is a question of fundamental importance for treatment of knee injuries and knee replacement design. Unfortunately, we lack unambiguous and comprehensive knee function data sets and/or consensus on how healthy knees move. One can just as easily find reports stating the natural knee has a center of axial rotation in the medial compartment of the knee as in the lateral. This is due to technical and practical issues: It is extremely difficult to accurately measure knee motions during ambulatory activities and, when that can be done, very few studies have examined a range of weightbearing activities in the same study cohort. The purpose of this study is to report knee kinematics in a cohort of healthy older subjects whose motions were examined during four different movements, three of them weightbearing ambulation, using a high-speed stereoradiographic system.

Six healthy consenting subjects (age = 61 ± 5 years, body mass = 75 ± 8 kg, BMI = 27 ± 4) were observed using a high-speed stereoradiographic system while completing four tasks. Subjects were instructed to perform an unloaded, seated knee extension from high flexion to full extension; to walk at a self selected pace; to step down from a 7 inch platform; and to walk and perform a 90° direction change (pivoting). Stereoradiographic images (1080 × 1080 pixels) were acquired at 100 images/second using 40cm image intensifiers and pulsed x-ray exposures. The three-dimensional knee kinematics were measured using the XROMM software suite (xromm.org, Brown University). Post-processing of the kinematics was performed in custom Matlab programs, and included fitting spheres to the posterior condylar surfaces of each knee, and then tracking the motions of the sphere centers relative to a fixed tibial reference frame (Figure 1). The motions of these flexion-facet centers, were used to determine an average center of axial rotation (CoR) over each activity as previously reported by Banks and Hodge.

Average CoRs for all four activities were in the posterior-medial quadrant of the knee, with the CoR for open-chain knee extension being the most medial and gait the most lateral (Table 1, Figure 2). One-way ANOVA showed average CoRs are different (p « 0.001). There was considerable variation in individual CoRs, for example, with two knees showing lateral CoRs for gait and the remaining knees having medial CoRs.

It should not be surprising that natural knee motions vary with dynamic activity, yet knee kinematics often are presented as being one stereotypic, monolithic pattern of motion. Our data show that the same healthy subjects performing different dynamic activities manifest different knee motions, with open-chain knee extension having the most medial CoR and gait the most lateral. This finding is consistent with previous reports comparing stair climbing and gait in knees with various implant designs. Additional experimental data and, ultimately, validated numerical simulations should facilitate an increasingly accurate process for designing improved treatments for diseased and damaged knees.

Background

Though many advantages of reverse total shoulder arthroplasty (RTSA) have been demonstrated, a variety of complications indicate there is much to learn about how RTSA modifies normal shoulder function. This study assesses how RTSA affects deltoid muscle moment arms post-surgery using a subject-specific computational model driven by in vivo kinematic data.

Introduction

Total knee arthroplasty (TKA) designs evolve as evidence accumulates on natural and prosthetic knee function. TKA designs based upon a medially conforming tibiofemoral articulation seek to reproduce essential aspects of normal knee stability and have enjoyed good clinical success and high patient satisfaction for over two decades. Fluoroscopic kinematic studies on several medially conforming knee designs show extremely stable knee function, but very small ranges of tibial axial rotation compared to healthy knees. The GMK Sphere TKA is a recent evolution in medially-conforming TKA designs that adopts a sagittally unconstrained lateral tibiofemoral articulation to allow more natural tibial rotation. This study was conducted to quantify motions in knees with this prosthesis to address two questions:

Does the medially conforming GMK Sphere design provide an AP-stable articulation that provides for tibiofemoral translations that are comparable to, but not larger than, translations measured in natural knees?

The history of knee mechanics studies and the evolution of knee arthroplasty design have been well reported through the last decade (e.g. [1],[2]). Through the early 2000's, there was near consensus on the dominant motions occurring in the healthy knee among much of the biomechanics and orthopaedic communities. However, the past decade has seen the application of improved measurement techniques to permit accurate measurement of natural knee motion during activities like walking and running. The results of these studies suggest healthy knee motion is more complex than previously thought, and therefore, design of suitable arthroplasty devices more difficult. The purpose of this paper is to briefly review the knee biomechanics literature before 2008, to present newer studies for walking and running, and to discuss the implications of these findings for the design of knee replacement implants that seek to replicate physiologic knee motions.

Many surgeons point to Brantigan and Voshell [3], an anatomic study of over one hundred specimens focusing on the ligamentous and passive stabilizers of the knee, as being an important influence in their thinking about normal knee function. M.A.R. Freeman and colleagues in London claim particular influence from this work, which motivated their extensive series of MR-based knee studies reported in 2000 [4,5,6]. These papers, perhaps more than any others, are responsible for the common impression that knee kinematics are well and simply described as having a ‘medial pivot’ pattern, where the medial condyle remains stationary on the tibial plateau while the lateral condyle translates posteriorly with knee flexion. Indeed, subsequent studies in healthy and arthritic knees during squatting and kneeling [7,8,9] and healthy and ACL-deficient knees during deep knee bends [10,11] show patterns of motion quite similar to those reported by Freeman and coworkers. These studies make a convincing case for how the healthy knee moves during squatting, kneeling and lunging activities. However, these studies are essentially silent on knee motions during ambulatory activities like walking, running and stair-climbing; activities which most agree are critically important to a high-function lifestyle.

In 2008 Koo and Andriacchi reported a motion laboratory study of walking in 46 young healthy individuals and found that the stance phase knee center of rotation was LATERAL in 100% of study participants [12]. One year later, Kozanek et al. published a bi-plane fluoroscopy study of healthy knees walking on a treadmill and corroborated the findings of Koo and Andriacchi, i.e. the center of rotation in healthy knees walking was lateral [13]. Isberg et al. published in 2011 a dynamic radiostereometric study of knee motions in healthy, ACL-deficient and ACL-reconstructed knees during a weight-bearing flexion-to-extension activity, and showed consistent anterior-to-posterior medial condylar translations with knee extension, accompanied by relatively little lateral condylar translation [14]. Hoshino and Tashman reported in 2012 another dynamic radiostereometric analysis of healthy knees during downhill running and concluded “While the location of the knee rotational axis may be dependent on the specific loading condition, during … walking and running … it is positioned primarily on the lateral side of the joint. ”[15] Finally, Claes et al. reported in late 2013 the detailed anatomy of the anterolateral ligament (ALL), another structure serving to stabilize the lateral knee compartment near extension, roughly in parallel with the anterior cruciate ligament (ACL) [16].

Studies since 2008 [9,12–16] show knee motions during walking, running and pivoting activities do not fit the “medial pivot” pattern of motion, but rather point to a “lateral pivot” pattern of knee motion consistent with the stabilizing roles of the ACL and ALL. Having a medial center of rotation in flexion and a lateral center of rotation in extension greatly complicates knee arthroplasty design if the goal is to reproduce kinematics approximating those observed in the natural knee. Consistent kinematics having a fixed center of rotation implies joint stabilizing structures or surfaces, not simply articular laxity allowing the knee to move as forces dictate. Thus, a total knee arthroplasty design seeking to reproduce physiologic motions may need to provide distinct means for controlling tibiofemoral motion in both extension and flexion. Recent studies of natural knee motions have made the implant designer's job more difficult!

Model-image registration types of measurements have profoundly changed capabilities for studying dynamic 3D joint and implant kinematics since their introduction in the early 1990's. Since that time, a variety of proprietary and open-source software packages have been developed and reported for performing these measurements. Model-image registration based measurements have been used to quantify motions in natural and replaced knees, hips, ankles, shoulders, elbows, and spines in both single- and stereo-projection radiographic measurement setups. In theory, with the same quality images and the same quality bone/implant models, any of the software developed to perform model-image registration has the potential to provide equivalent measurement accuracy. Hence, much of the effort to improve measurement capabilities has been to reduce human interaction requirements and make the measurements more automatic and objective. In this paper, we report a new open-source software program that requires a minimum of user input to automate the 3D kinematic measurement process from single- or bi-plane radiographic projections.

JointTrack Auto (JTA) is an open source (sourforge.net/projects/jointtrackauto) program for performing model-image registration of metallic implants with single- or bi-plane radiographic images (see image). A predominantly edge-based cost function is used with an adaptive partioning global optimization scheme for model-image registration. Although this method works without any human intervention, JTA allows users to roughly identify one ‘feature’ of each implant that is visible in all images, e.g. the tip of a peg, to very significantly reduce the search space and time required for numerical optimization. This makes for a very convenient and fast initialization process where a human user simply mouse-clicks on a few easily identifiable locations in each radiographic image, and then the automated registration process is begun.

Registration accuracy examples and a software demonstration will be included in this e-poster presentation to introduce attendees to the software and spur discussion about the various methods available to perform these important measurements.

Reverse total shoulder arthroplasty (RTSA) has had rapidly increasingly utilization since its approval for U.S. use in 2004. RTSA accounted for 11% of extremity market procedure growth in 201. Although RTSA is widely used, there remain significant challenges in determining the location and configuration of implants to achieve optimal clinical and functional results. The goal of this study was to measure the 3D position of the shoulder joint center, relative to the center of the native glenoid face, in 16 subjects with RTSA of three different implant designs, and in 12 healthy young shoulders.

CT scans of 12 healthy and 16 pre-operative shoulders were segmented to create 3D models of the scapula and humerus. A standardized bone coordinate system was defined for each bone (Figure 1). For healthy shoulders, the location of the humeral head center was measured relative to the glenoid face center. For the RTSA shoulders, a two-step measurement was required. First, 3D models of the pre-operative bones were reconstructed and oriented in the same manner as for healthy shoulders. Second, 3D model-image registration was used to determine the post-operative implant positioning relative to the bones. The 3D position and orientation of the implants and bones were determined in a sequence of six fluoroscopic images of the arm during abduction, and the mean implant-to-bone relationships were used to determine the surgical positioning of the implants (Figure 2). The RTSA center of rotation was defined as the offset from the center of the implant glenosphere to the center of the native glenoid face.

The center of rotation in RTSA shoulders varied over a much greater range than the native shoulders (Table 1 (Figure 3)). Lateral offset of the joint center in RTSA shoulders was at least 6 mm smaller than the smallest joint center offset in the healthy shoulders. The center of rotation in RTSA shoulders was significantly more inferior than in healthy shoulders. The range of anterior/posterior placement of the rotation center for RTSA shoulders was bounded by the range for normal shoulders.

How to best position RTSA implants for optimal patient outcomes remains a topic of great debate and research interest. We found that the 3D joint center position can vary over a supraphysiologic range in shoulders with RTSA, and that this variation is primarily in the coronal plane. By relating these geometric variations to muscle, shoulder and clinical function, we hope to establish methods and strategies for predictably obtaining the best clinical and functional outcomes for RTSA patients on a per-subject basis.