Persistent discrepancies in orbital morphology after surgical treatment of unicoronal craniosynostosis: a critical image-based analysis

Abstract

OBJECTIVE Unicoronal craniosynostosis (UCS) is characterized by complex orbital deformity and is typically treated by asymmetrical fronto-orbital remodeling (FOR) during the 1st year of life. The aim of this study was to elucidate to what extent orbital morphology is corrected by surgical treatment. METHODS The extent to which orbital morphology was corrected by surgical treatment was tested by analysis of differences in volume and shape between synostotic, nonsynostotic, and control orbits at two time points. In total, 147 orbits were analyzed from patient CT images obtained preoperatively (mean age 9.3 months), at follow-up (mean age 3.0 years), and in matched controls. Semiautomatic segmentation software was used to determine orbital volume. For analysis of orbital shape and asymmetry, geometrical models, signed distance maps, principal modes of variation, and three objective parameters (mean absolute distance, Hausdorff distance, and dice similarity coefficient) were generated by statistical shape modeling. RESULTS Orbital volumes on both the synostotic and nonsynostotic sides were significantly smaller at follow-up than volumes in controls and significantly smaller both preoperatively and at follow-up than orbital volumes on the nonsynostotic side. Significant differences in shape were identified globally and locally, both preoperatively and at 3 years of age. Compared with controls, deviations were mostly found on the synostotic side at both time points. Asymmetry between synostotic and nonsynostotic sides was significantly decreased at follow-up, but not compared with the inherent asymmetry of controls. On a group level, the preoperative synostotic orbit was mainly expanded in the anterosuperior and anteroinferior regions and smallest on the temporal side. At follow-up, the mean synostotic orbit was still larger superiorly but also expanded in the anteroinferior temporal region. Overall, the morphology of nonsynostotic orbits was more similar to that of controls than to synostotic orbits. However, the individual variation in orbital shape was greatest for nonsynostotic orbits at follow-up. CONCLUSIONS In this study, the authors presented what is, to their knowledge, the first objective automatic 3D bony evaluation of orbital shape in UCS, defining in greater detail than has been done previously how synostotic orbits differ from nonsynostotic and control orbits, and how orbital shape changes from 9.3 months of age preoperatively to 3 years of age at the postoperative follow-up. Despite surgical treatment, both local and global deviations in shape persist. These findings may have implications for future directions in the development of surgical treatment. Future studies connecting orbital morphology to ophthalmic disorders, aesthetics, and genetics could provide further insight to enable better outcomes in UCS.