Identifying reproducible patterns of calvarial dysmorphology in nonsyndromic sagittal craniosynostosis may affect operative intervention and outcomes assessment.

Background: The authors tested the premise that there are four distinctive patterns of calvarial dysmorphology in nonsyndromic sagittal craniosynostosis that can be reproducibly recognized.

Methods: Twenty-nine computed tomographic scan data sets of infants met the following criteria: nonsyndromic sagittal craniosynostosis, age younger than 12 months, and satisfactory computed tomographic data. Osseous reformations were constructed in the anteroposterior, right lateral, and vertex projections for each patient. From these images, four templates--coronal constriction, occipital protuberance, bifrontal bossing, and bitemporal protrusion--were selected as prototypes of the specific dysmorphologies the authors observed in patients with sagittal craniosynostosis. Four residents assigned the 29 calvarial image sets to one of the four templates or, if they were unable to do so, to the group "other." The sortings were then assessed for clustering. The same patient computed tomographic data were reformatted with osseous color images, which were then sorted according to template group by eight senior craniofacial surgeons, who repeated the task approximately 3 months later. The repeatability and assessment of clustering of image sets using the templates was evaluated.

Results: In the residents' pilot study, 41 percent (12 of 29) of patients had 100 percent concordance rates, 31 percent (nine of 29) had 75 percent concordance, 24 percent (seven of 29) had 50 percent, and 3 percent (one of 29) had 25 percent concordance. In summary, greater than 70 percent of the patient image sets could be sorted with at least 75 percent concordance by residents. In the senior surgeons' study, 90 percent of patients could be identified as falling into two of five possible groups. Senior raters demonstrated nearly 70 percent repeatability between sortings.

Conclusion: These findings support the hypothesis that there are identifiable and reproducible patterns of varying calvarial dysmorphology in patients with sagittal craniosynostosis.