Cerebral palsy (CP) is a neural condition that impacts and impairs the musculoskeletal system. Skeletal muscles, particularly in the lower limb, have previously been shown to be significantly reduced in volume in CP compared to typical controls. Muscle volume is a gross measure, however, and does not capture shape characteristics which—if quantified—could offer a robust and novel assessment of how this condition impacts skeletal muscle form and function in CP. In this study, we used mathematical shape modelling to quantify not just size, but also the shape, of soleus muscles in CP and typically developing (TD) cohorts to explore this question.

Shape modelling is a mathematical technique used previously for bones, organs, and tumours. We obtained segmented muscle data from prior MRI studies in CP. We generated shape models of CP and TD cohorts and used our shape models to assess similarities and differences between the cohorts, and we statistically analysed shape differences.

The shape models revealed similar principal components (PCs), i.e. the defining mathematical features of each shape, yet showed greater shape variability within the CP cohort. The model revealed a distinct feature (a superior –> inferior shift of the broad central region), indicating the model could identify muscular features that were not apparent with direct observation. Two PCs dominated the differences between CP and TD cohorts: size and aspect ratio (thinness) of the muscle.

The distinct appearance characteristic in the CP model correspond to specific muscle impairments in CP to be discussed further. Overall, children with CP had smaller muscles that also tended to be long, thin, and narrow. Shape modelling captures shape features quantitatively, which indicate the ways that muscles are being impacted in CP. In the future, we hope to tailor this technique toward informing diagnosis and treatments in CP.