Musculoskeletal loads during daily motor activities in patients with pelvic custom-made implants following bone tumor: a biomechanical analysis using multibody musculoskeletal modeling

Treatment of pelvic bone sarcomas represents one of the most controversial areas of orthopedic oncology due to the size of these sarcomas and the complexity of the pelvic anatomy, which increase the difficulty of resection and reconstruction. Recent advances in orthopedic oncology have improved the pelvic surgical technique through the rapid prototyping, a manufacturing technology where computerized image data are used to create high-precision 3D models of the pelvis. Based on these models, cutting guides and custom-made prostheses are produced facilitating resection, reconstruction and implantation. In particular, 3D printed prostheses are promising, but as a young technology there have been few studies utilizing it and exhibiting short follow-up periods. To ensure that complications such as fracture and dislocation do not occur, a better understanding of the biomechanical interactions between the implant and bone is needed. However, there are no studies analyzing musculoskeletal loads during movement in patients with 3D printed pelvic recostructions. A method for improving this understanding is to use computational models of the musculoskeletal system and reconstructed pelvis to predict the forces and stresses on the implant in-vivo. Multibody-dynamics models allow to calculate muscle and joint forces as an alternative to the invasive experimental methods that would otherwise be required. The aim of this thesis is to analyze the musculoskeletal loads during common motor activities in a group of six patients who underwent hemipelvic reconstruction with trabecular-titanium custom-made implants following resection of bone tumor, with the same innovative and unique surgical technique at the Rizzoli Orthopaedic Institute. In particular, we quantified the variability in muscle and joint contact forces and analyzed the differences between the operated and the contralateral limb. We created personalized multibody models of the musculoskeletal system of the patients using ground reaction forces measurements and experimental data acquired during a gait analysis session where patients, covered with reflective markers in according to Leardini protocol, performed some dynamic trials: five walkings, three squats, three sittings and three stairs ascent and descent. These simulations were solved for an optimization-based inverse dynamics workflow to calculate joint angles, joint moments, muscle forces and joint contact forces, leveraging OpenSim. The results show that lower-limb muscle forces and joint contact forces were all, on average, higher for the contralateral limb. This suggests that, due to the surgery and to the presence of the implanted prosthesis, all patients prefer to use and to strain more the contralateral side than the operated side during motor activities. Our findings might contribute to the development of specific rehabilitation programs, improvement of long-term strength of the implants, and analysis of bone adaptation in the follow-up.