Back to 2026 Abstracts
Late Presenting Isolated Sagittal Craniosynostosis: Can We Use Cranial Morphology to Guide Surgical Decision Making?
Tiffany E. Jeong
1, Michael Hernandez
1, Michelle Antony
1, Alzbeta Novotna
*1, Viraj Govani
1, Victoria García Rodríguez
1, Zhazira Irgebay
1, Shelby Nathan
1, Mokshagna Karanam
2, Nawazish Khan
2, Shireen Elhabian
2, Jesse A. Goldstein
11Department of Plastic Surgery, University of Pittsburgh School of Medicine, Pittsburgh, PA; 2SCI, University of Utah, Salt Lake City, UT
Background Sagittal craniosynostosis is the most common form of isolated craniosynostosis, yet operative decision making can be challenging especially in patients presenting after 1 year of age. Advances in machine learning have enabled automated quantification of cranial shape abnormalities through validated metrics such as preoperative Cranial Morphology Deviation (CMD) and Sagittal Severity Score (SSS). However, clinically validated thresholds to guide surgical intervention are lacking. This study evaluates the association between CMD and SSS with operative management and proposes thresholds to aid surgical decision-making.
Methods We retrospectively reviewed patients over 1 year of age with isolated sagittal synostosis at a single institution from January 1993 to December 2024. Preoperative CT scans meeting CranioRate™ quality standards were analyzed to generate CMD and SSS. Groups were compared using the Wilcoxon rank-sum test, and multivariable logistic regression assessed association with operative status, adjusting for age and sex. Optimal thresholds were identified using the Youden index.
Results 231 patients presented after 1 year of age, of whom 56 (24.2%) underwent surgery. CMD was significantly higher in operative patients (median: 176; IQR: 152-207) versus non-operative patients (median: 139; IQR: 117-163; p<0.001). SSS was similarly elevated: 4.13 (IQR: 3.04-5.75) versus 2.29 (IQR: 0.99-3.25; p<0.001). Both metrics were independently associated with operative status. Each CMD unit increase raised surgical odds by 2% (OR=1.02, p<0.001); each SSS unit increase raised odds by 50% (OR=1.50, p<0.001). Age and sex were not significant. AUC was 0.76 for CMD and 0.70 for SSS, with optimal thresholds of CMD=138.96 and SSS=2.69.
Conclusion This study establishes evidence-based operative thresholds for machine learning derived cranial morphology metrics in sagittal craniosynostosis, providing objective criteria to guide surgical decision making for patients over 1 year with isolated sagittal craniosynostosis.
Back to 2026 Abstracts
