Prevalence and Anatomical Distribution of Accessory Cranial Bones in Scaphocephaly: A CT-Based Case–Control Study

Abstract Objectives To quantify the prevalence and topographic distribution of accessory cranial bones in children with sagittal craniosynostosis versus trauma controls, and to assess their incremental diagnostic value beyond cranial index, age and anterior fontanelle status. Methods We conducted a retrospective review of head computed tomography (CT) examinations [axial images and three-dimensional computed tomography (3D CT) reconstructions] in 442 infants aged 0–24 months (94 with sagittal craniosynostosis; 348 trauma controls) acquired between 2014 and 2024. Accessory bones were classified using predefined anatomical criteria. Group differences were tested using chi-squared or Fisher’s exact tests with Bonferroni correction for multiple comparisons. We compared two logistic regression models: a baseline model incorporating cranial index, age and anterior fontanelle status, and an extended model incorporating three prespecified ossicle variables. Results Accessory cranial bones were more prevalent in infants with craniosynostosis than in the trauma control group [79.8% vs. 41.0%; odds ratio (OR) 5.66; 95% confidence interval (CI) 3.28–9.78], with the largest between-group difference in infants under 4 months. Disparities were most pronounced for o. epiptericum , o. asterii and os preinterparietale bipartitum . Adding these markers improved model performance, with better fit and discrimination (AUC 0.96 vs 0.93; likelihood-ratio test p < 0.001), and a modest gain in sensitivity at high specificity. Conclusions Accessory cranial bones, particularly o. asterii , o. epiptericum and os preinterparietale bipartitum , are substantially more prevalent in sagittal craniosynostosis and provide supportive, not standalone, diagnostic information alongside standard cranial metrics. Their recognition is most useful in equivocal examinations and paediatric trauma imaging and may aid preoperative planning. Prospective, multicentre validation is warranted.