Musculoskeletal Models to Aid in Clinical Decision-Making in Children with Cerebral Palsy

Clinical decision-making benefits from having accurate and complete information to make objective and justified treatment and surgical intervention recommendations as well as ensuring higher likelihoods of positive outcomes for patients. In Cerebral Palsy (CP), a neurological condition causing movement disability in children, 3D clinical gait analysis provides an important diagnostic and evaluation tool to achieve this to a large extent. It is somewhat limited with regards to certain surgical interventions, particularly when they are complex and involve information that cannot be obtained directly or without the use of invasive measurements. The advent of musculoskeletal (MSK) models have shown that this additional information such as muscle-tendon lengths, moment arms, muscle forces and joint contact forces, can be obtained, while permitting interrogative analysis as well. Although efforts have been made to verify the accuracy and reliability of MSK modelling with regards to the current measures obtained from clinical gait analysis, these methods have not seen much uptake. The general aim of this thesis was therefore to investigate the suitability and use of MSK models to quantify and predict longitudinal changes and surgical intervention outcomes in children with cerebral palsy using an example case of the Femoral Derotation Osteotomy (FDO) and thereby make a case for their use in supporting the clinical decision-making process.
The results from this work showed that generic MSK models can perform as well as currently employed techniques in estimating longitudinal changes over time of clinical measures and outcomes after surgical intervention of FDO. Potential predictors of positive outcome after the FDO from muscle-tendon lengths that are usually not included in the clinical decision-making process were also identified, lending support to the available literature on the utility of musculoskeletal models use in this manner and for clinical gait analysis. Additionally, the results discussed showed that the accuracy of outputs from the generic model can be improved with minimal personalisation although additional work is required to achieve improvements in aspects related to the muscle-tendon parameters.
Future work will be targeted at investigating and validating the identified muscles with predictive value for FDO outcomes as well as improving the personalisation of the bone and muscle-tendon properties using less time and cost intensive techniques such as ultrasound.