Modeling the future of shoulder arthroplasty

Abstract

Background: In patients with advanced shoulder pathology requiring arthroplasty, the goal is to restore optimal function and provide sustained pain relief. However, the capacity to select the ideal reconstruction of a specific patient's shoulder and to predict the resulting functional outcome is limited. Computational modeling of the musculoskeletal system has the potential to expand our foundational knowledge of both the native and prosthetic shoulder joint.

Methods: The aim of this review is to describe how computational modeling enables more detailed analyses of the interactions between musculoskeletal anatomy and function and to suggest ways in which this approach may be used to optimize outcomes in patients undergoing shoulder arthroplasty.

Results: Computational modeling has been used to study shoulder joint biomechanics for more than 30 years. There are specific limitations that need to be addressed to realize the full potential of computational modeling in shoulder arthroplasty. First, more realistic, patient-specific models of the glenohumeral and scapulothoracic joints must be developed. Second, shoulder models must be coupled with accurate in vivo measurements of joint motion to perform more comprehensive analyses of muscle and joint function. Third, model predictions of shoulder biomechanics must be validated against experimental data.

Conclusion: Patient-specific musculoskeletal modeling of shoulder joint biomechanics can contribute significantly to predicting pathology and optimizing postoperative function.

Keywords: Biomechanics; Computational modeling; Joint; Muscle architecture; Musculoskeletal modeling; Shoulder arthroplasty.

Figure 1

Musculoskeletal model of the shoulder used to calculate individual contributions of the shoulder muscles to glenohumeral joint stability. Eighteen bundles were used to represent 11 muscles crossing the shoulder in the model: anterior, middle, and posterior portions of deltoid; clavicular, sternal, and ribs portions of pectoralis major; supraspinatus; infraspinatus; subscapularis; teres major; teres minor; long head and short head of biceps brachii; long head of triceps brachii; thoracic, lumbar, iliac portions of latissimus dorsi; and coracobrachialis. Adapted from Yanagawa et al with permission.

Figure 2

Flowchart illustrating the various stages of a dynamic simulation of shoulder joint movement. The inputs to the model are the neural commands (excitation signals), u, sent from the central nervous system to muscle. The neural commands are converted to muscle forces, F, by a model of muscle-tendon dynamics, which consists of the force-length and force-velocity properties of muscle. Muscle forces are then converted to joint torques, T, by virtue of a model of musculoskeletal geometry (muscle moment arms). Finally, the joint torques are converted to joint motion (q) by integrating the equations of motion describing skeletal dynamics forward in time.