Computer-Based Quantitative Assessment of Skull Morphology for Craniosynostosis

In this paper we present a processing pipeline for the computational analysis of the craniosynostotic skull. Our fully automatic methodology uses a statistical shape model in order to produce diagnostic features tailored to the anatomy of the subject. We obtained an index of cranial suture closure and deformation and curvature averages across five bone segments and six suture regions automatically delineated on each subject skull. We show high correlation between these shape characteristics and our diagnostic ground truth, displaying significant differences between normal and craniosynostosis subjects, and thus suggesting the ability of our approach to provide new pathways towards the automatic diagnosis of cranysinostosis, and optimized surgical planning.