Abstract
BACKGROUND: Preoperative gait analysis plays a crucial role in determining the necessity and correction angle for femoral derotational osteotomy (FDO). However, conventional musculoskeletal models used in gait analysis often fail to reflect patient-specific musculoskeletal characteristics, such as femoral and tibial deformities. This study evaluates the impact of a personalized musculoskeletal model incorporating these deformities on gait analysis and investigates the surgical outcomes of FDO based on kinematic changes derived from this personalized musculoskeletal model. METHODS: A retrospective analysis was conducted on 254 limbs from 127 children with cerebral palsy (CP) who presented with increased femoral anteversion and underwent pre- and postoperative gait analyses. Kinematic data were generated using general and personalized musculoskeletal models developed in OpenSim. Patients were classified according to FDO status and the presence of excessive hip internal rotation (IR). Surgical outcomes were assessed based on postoperative changes in hip rotation. Subgroup analyses were performed to evaluate the model's impact on surgical outcomes. RESULTS: Of the 254 limbs, 92 underwent FDO. Patients with increased hip IR in the general model (Group A) had a higher good responder rate (88.2%) than those without (Group B, 17.2%). All limbs in Groups A1 and B1 (increased hip IR using personalized musculoskeletal models) had 100% favorable outcomes, whereas Groups A2 and B2 (not increased hip IR using personalized musculoskeletal models) showed favorable outcomes in 20% and 13.5%, respectively. Increased hip IR was more frequent in patients with external tibial rotation (p < 0.05). Surgical outcomes differed significantly between patients with and without increased hip IR in the personalized musculoskeletal model (χ(2) = 4.90, p = 0.027). CONCLUSION: Gait analysis using personalized musculoskeletal models improved surgical decision-making for FDO, leading to better outcomes in children with CP. Personalized musculoskeletal models better identified suitable FDO candidates and more accurately predict surgical outcomes than general models. TRIAL REGISTRATION: Retrospectively registered.