Yue CHENG Thesis defense

Cerebral palsy (CP), a common physical disability in childhood, often causes abnormal movement patterns and posture, resulting in an abnormality in musculoskeletal morphometry. However, the bone deformation occurring in the CP population remains unclear. To better understand the pathology and improve the rehabilitation of patients, the main objective of this thesis is to comprehensively analyze the morphological change in children with CP based on image studies. In the first step, the normal bone shape of typical developing (TD) children is modeled with statistical shape modeling (SSM). A mean shape template called atlas of TD population is created as a reference. To locate the possible abnormal area, we analyze the bone morphometry of the CP population in a multi-resolution manner. The bone shape is analyzed from the global, region-based and voxel-based scales. For voxel-based scale analysis, in order to cope with the limited size of the pediatric dataset, we introduce a voxel-based anomaly detection method adapted to the small dataset. Moreover, due to the variability between individuals and the requirement of personalized medical care, we study bone shape at both populational and personal levels. Besides, we compare the results of our voxel-based approach and one state-of-the-art SSM tool called ShapeWorks. The comparison demonstrates the consistency between methods, confirming the reliability of the obtained results.